TWENTY - SEVENTH ANNUAL INTERNATIONAL
PITTSBURGH COAL CONFERENCE FINAL TECHNICAL PROGRAM COAL - ENERGY, ENVIRONMENT AND SUSTAINABLE DEVELOPMENT October 11 - 14, 2010 Hilton Istanbul Istanbul, TURKEY Co-Sponsors and Organizers
University of Pittsburgh Swanson School of Engineering
Istanbul Technical University
Sponsors
gold b
Turkish Coal Enterprises
2
GENERAL INFORMATION
WELCOME! On behalf of the Conference Advisory Board, Conference Committees, and the University of Pittsburgh we welcome you to the Twenty-Seventh Annual International Pittsburgh Coal Conference held October 11 - 14, 2010 at the Hilton Istanbul in Istanbul, Turkey. The Conference is hosted by the University of Pittsburgh. The theme of this year’s conference is “Coal - Energy, Environment and Sustainable Development” which covers a wide spectrum of important topics on coal technology, synfuel and environmental issues. The topics cover energy and environmental issues and technologies related to coal and its byproducts. Over 300 technical papers and posters will be presented throughout the conference. The Poster Sessions will be held on Tuesday, October 12 from 18:20 - 20:20. For detailed information on technical sessions, papers and speakers, please turn to page 6 in the Technical Program. The invited Plenary Speakers include: Charles Taylor, Director, Chemistry and Surface Science Division, U.S. Department of Energy, National Energy Technology Laboratory, USA; Carlos A. Cabrera, President and CEO, National Institute of Clean and Low Carbon Energy (NICE), CHINA; Selahaddin Anaç, General Director, Turkish Coal Enterprise (TKI), TURKEY; Richard Winschel, Director, Research Services, CONSOL Energy Inc., USA; John Topper, Managing Director, IEA Clean Coal Centre, UNITED KINGDOM; Ekrem Ekinci, Rector, Isık Universty, TURKEY; John Wheeldon, Manager of Science and Technology, National Carbon Capture Center, EPRI, USA; Volkan Ediger, İzmir Technology Institute, TURKEY; and Robert Beck, Executive Vice President and Chief Operating Officer, The National Coal Council, Inc., USA. We express our sincere gratitude to the contributors for their support and involvement, to all the authors and co-authors of the technical papers and to all the members of the Program Committee, Awards Committee, International Committee and Membership Committee. Special thanks go to our Turkish host and coordinator Dr. Güven ÖNAL, Technical Program Chairs, Evan Granite of NETL-DOE, USA, and Jim Hower of the University of Kentucky, CAER, USA as well as to all session chairs, speakers and international delegates for their contributions to the 2010 technical program. As the chair and vice chair of the Advisory Board of the Conference, we deeply appreciate your participation and interest in this year’s Conference and we invite you to join us next year for the Twenty-Eighth Annual International Pittsburgh Coal Conference, which will be held in Pittsburgh, PA, USA at the David L. Lawrence Convention Center. Sincerely,
Robert Beck, Chair The National Coal Council, Inc., Washington, DC, USA
Richard Winschel, Vice -Chair CONSOL Energy Inc., Pittsburgh, PA, USA
CONFERENCE REGISTRATION On-Site Registration will begin Monday, October 11, from 18:00 - 20:00 and continues Monday, Tuesday, and Wednesday from 08:00 until 17:00. Please check in even if you have Pre-Registered!
Tours and reservations for the 2010 Pittsburgh Coal Conference are provided by: Aktuel Tourism Company Contact Person: Süleyman Göncü Ortaklar Cad. Unal Apt. No.7/Daire. 8 Mecidiyeköy/Istanbul/Turkey Tel. : +90(212)273. 06 90 Fax : +90(212)213. 63 95 E-mail:
[email protected] http://www.aktueltours.com/
The International Pittsburgh Coal Conference EXECUTIVE DIRECTOR: Dr. Badie I. Morsi CONFERENCE ORGANIZER: Ms. Heidi M. Aufdenkamp University of Pittsburgh Swanson School of Engineering 1249 Benedum Hall Pittsburgh, PA 15261 USA Tel: +1-412-624-7440 FAX: +1-412-624-1480 Email:
[email protected] www.engr.pitt.edu/pcc
PITT AWARD The Award for Innovation in Coal Conversion was founded by the Chemical and Petroleum Engineering Department, University of Pittsburgh in 1983 with industrial support. Since 1992, it has been fully funded by CONSOL Energy Inc.
3
GENERAL INFORMATION PLENARY SPEAKERS CONFERENCE OVERVIEW MONDAY, OCTOBER 11, 2010 Technical Tour Registration Reception
07:00 - 22:30 18:00 - 20:00 19:30 - 21:30
TUESDAY, OCTOBER 12, 2010 Registration Opening Ceremony Plenary Session – 1 Concurrent Tech. Sessions Conference Luncheon Concurrent Tech. Sessions Poster Session Dinner
08:00 - 17:00 09:00 - 09:30 09:30 - 11:15 11:30 - 13:10 13:10 - 14:25 14:25 - 18:20 18:20 - 20:20 20:20 - 22:20
WEDNESDAY, OCTOBER 13, 2010 Registration Plenary Session – 2 Concurrent Tech. Sessions Conference Luncheon Concurrent Tech. Sessions Gala Dinner
08:00 - 17:00 09:00 - 10:45 11:00 - 12:40 12:40 - 13:55 13:55 - 17:50 20:30 - 22:30
THURSDAY, OCTOBER 14, 2010 Registration Plenary Session – 3 Concurrent Tech. Sessions Awards Luncheon Concurrent Tech. Sessions Advisory Board Meeting
08:00 - 17:00 09:00 - 10:45 11:00 - 12:40 12:40 - 13:55 13:55 - 17:50 18:00 - 21:00
TUESDAY, OCTOBER 12, 2010 Energy Production/Policy Speakers Charles Taylor Director, Chemistry and Surface Science Division U.S. DOE/NETL, USA “U.S. Department of Energy’s Carbon Capture and Storage Efforts and Results”
Carlos A. Cabrera President and CEO National Institute of Clean and Low Carbon Energy (NICE), CHINA “A Refined Approach to Coal”
Selahaddin Anaç General Director Turkish Coal Enterprise (TKI), TURKEY “Coal in Turkish Energy Policy and Clean Coal Technologies”
WEDNESDAY, OCTOBER 13, 2010 International Issues Richard Winschel Director, Research Services CONSOL Energy Inc., USA “FutureGen – the World’s First Near-Zero Emission Coal-Based Power Plant”
John Topper Managing Director IEA Clean Coal Centre, UNITED KINGDOM “Sustainable Low Emissions Coal for our Grandchildren?”
Ekrem Ekinci Rector Isık Universty, TURKEY “Realities and Constraints of Coal in Energy and Environmental Perspectives”
THURSDAY, OCTOBER 14, 2010 Environmental Issues John Wheeldon Manager of Science and Technology National Carbon Capture Center, EPRI, USA “CO2 Capture and Storage for Coal-Based Power Generation”
Volkan Ediger İzmir University of Economics, TURKEY “Geopolitics of Coal and Global Climate”
Robert Beck Executive Vice President and Chief Operating Office The National Coal Council, Inc., USA “Low Carbon Coal: 21st Century Technologies and Policies”
4
TECHNICAL PROGRAM SCHEDULE
5
LOCATION GUIDE Hilton Istanbul Convention Center
ROOM DIRECTORY Registration Hyde Park Monday Evening Reception Convention Center Opening Ceremony Convention Center Plenary Sessions Convention Center Conference Luncheons Convention Center Poster Presentations Convention Center Exhibits Convention Center A/V & Speaker Preparation Room Orman Park
SESSION MEETING ROOMS Convention Center Sessions 1, 7, 13, 19, 25, 31, 37, 43, 49 Convention Center Sessions 2, 8, 14, 20, 26, 32, 38, 44, 50 Convention Center Sessions 3, 9, 15, 21, 27, 33, 39, 45, 51 Saturn Sessions 4, 10, 16, 22, 28, 34, 40, 46, 52 Lalezar Sessions 5, 11, 17, 23, 29, 35, 41, 47, 53 Jupiter Sessions 6, 12, 18, 24, 30, 36, 42, 48, 54
6
TECHNICAL PROGRAM ORAL SESSIONS
Tuesday, October 12, 2010 11:30 - 18:20 SESSION 1 COMBUSTION: OXY-COAL DEVELOPMENT – 1 John Wheeldon and İskender Gokalp Development of OxycoalTM Technology Resulting from Testing Conducted at Doosan Power Systems’ Clean Combustion Test Facility (CCTF), Peter Holland-Lloyd, David Fitzgerald, Doosan Power Systems, UNITED KINGDOM Co-Firing of Coal and Wood Biomass in Oxy-Fuel Combustion, Seong Yool Ahn, Jae woo An, Yon Mo Sung, Cheor Eon Moon, Gyung Min Choi, Duck Jool Kim, Pusan National University, SOUTH KOREA NO x Reburning in Oxy-Fuel Combustion - An Experimental Investigation, Daniel Kühnemuth, Fredrik Normann, Klas Andersson, Filip Johnsson, Bo Leckner, Chalmers University of Technology, SWEDEN Oxy-Combustion of Pulverized Coal: Modeling of Char-Combustion Kinetics, M. Geier, C. R. Shaddix, Sandia National Labs, USA; B. S. Haynes, University of Sydney, AUSTRALIA
SESSION 3 CARBON MANAGEMENT: GHG MANAGEMENT STRATEGIES AND ECONOMICS – 1 Leslie Ruppert and Ender Okandan
Design and Operational Strategies for IGCC with CO2 Capture, Chris Higman, Higman Consulting GmbH, GERMANY; George Booras, Electric Power Research Institute; Dan Kubek, Gas Processing Solutions LLC; Jim Sorensen, Sorensenergy LLC; Doug Todd, Process Power Plants, LLC, USA
The Creation of Georeactor Global Scientific Network, Jan Rogut, Jozef Dubinski, Aleksandra Tokarz, GIG, Central Mining Institute, POLAND; Marc Steen, Institute for Energy, Joint Research Centre; Hans Bruining, Delft University of Technology, THE NETHERLANDS; Hema J. Siriwardane, West Virginia University; Tomasz Wiltowski, Southern Illinois University; Elizabeth Burton, Lawrence Livermore National Laboratory; Subhas K. Sikdar, US EPA, USA; Thomas Kempka, German Research Centre for Geosciences (GFZ), GERMANY; Sohei Shimada, University of Tokyo, JAPAN
Envisioning CO2 Distribution Networks for Carbon Capture and Storage (CCS) in the United States: Strategies for CO2 Pipeline Deployment at a Regional Scale, Nils Johnson, Joan Ogden, University of California, Davis, USA
SESSION 6 COAL-DERIVED PRODUCTS: CHEMICALS/MATERIALS Belma Demirel and Bekir Zühtü Uysal
Processes to Produce Value Added Products from CO2, Belma Demirel, Deena Ferdous, Rentech, Inc., USA
Linking Economic and Technological Modeling of CCS and Legislative Policy for Coal Mining Companies, Pratt Rogers, Sean Dessureault, University of Arizona, USA SESSION 4 COAL SCIENCE: COAL CHEMISTRY – 1 Ashok K. Singh and Karol Koster
Study on Thermodynamic Calculation for O2/CO2 Flue Gases Recycled Combustion Boiler, Li-Qi Zhang, Can-Zhi Li, Fang Huang, Ji-Hua Qiu, Chu-Guang Zheng, Huazhong University of Science and Technology, CHINA
Carbonaceous Emissions Reflected in Deposits on Building Stones: Case Study in Prague Castle, Ivana Sýkorová, Martina Havelcová, Hana Trejtnarová, Petra Matysová, Alexandr Šulc, Institute of Rock Structure and Mechanics AS CR, v.v.i.; Antonín Zeman, Institute of Theoretical and Applied Mechanics, AS CR, v.v.i., CZECH REPUBLIC
SESSION 2 GASIFICATION: GENERAL SESSION – 1 Ke Liu and Hasancan Okutan
Visualization of Coal Conversion Using X-Ray Computed Tomography, QP Campbell, HWJP Neomagus, North-West University, SOUTH AFRICA
Coal: Biomass Gasification - A Pathway for New Technology Development of Oxygen Blown Co-fired Gasification with Integrated Electrolysis, Tana Levi, R. Witney, Y Iwasaki, Tony Clemens, CRL Energy Ltd.; S Pang, Q Xu, University of Canterbury; AI Gardiner, Industrial Research Limited, NEW ZEALAND Hydrogen Generation from Water by Using Plasma, Beycan İbrahimoğlu, İbrahim İbrahimoğlu, Anadolu Plazma Teknoloji Merkezi; Fırat Şen, Vestel Defence Industry R&D Department; Şahika Yürek, Türkiye Kömür İşletmeleri; Orhan Demirel, Turkish Coal Enterprises (TKI), TURKEY Technology and Operational Experience – The Shell Perspective, Jay Wang, Shell Global Solutions International BV, THE NETHERLANDS Controlling the Synthesis Gas Composition from Catalytic Gasification of Hypercoal and Coal by Changing the Gasification Parameters, Atul Sharma, Toshimasa Takanohashi, National Institute of Advanced Industrial Science and Technology, JAPAN A Technico-Economical Feasibility Study of Plasma Assisted Coal Gasification Compared to a Reference Auto-Thermal Gasifiction Process, Nazim Merlo, Iskender Gökalp, ICARE-CNRS, FRANCE
Reducing Greenhouse Gas Emissions from Coal Combustion by Adding Micro-Algal Biomass, Jaco Brink, Sanette Marx, North-West University, SOUTH AFRICA
Organic Sulphur Functionality Changes in Biotreated Coals, Lenia Gonsalvesh, Stefan Marinov, Maya Stefanova, Institute of Organic Chemistry, Bulgarian Academy of Sciences, BULGARIA; Robert Carleer, Jan Yperman, Hasselt University, BELGIUM Surface Coal Mine Planning Against Large Landslides, Celal Karpuz, Levent Tutluoglu, Arman Kocal, Middle East Technical University; Mustafa Ozdingis, Kıvanc Het, Turkish Coal Enterprises (TKI), TURKEY SESSION 5 SUSTAINABILITY AND ENVIRONMENT – 1 Brenda Pierce and Fevzi İsbilir A Solution to Water Crisis in Energy Production: Feasibility of Using Impaired Waters for Coal-Fired Power Plant Cooling, Radisav D. Vidic, Heng Li, ShihHsiang Chien, Jason D. Monnell, University of Pittsburgh; David Dzombak, Ming-Kai Hsieh, Carnegie Mellon University, USA Leaching Characteristics of Waste from PF Utilities and Transitional Technologies using Australian Coal, D.H. French, K.W. Riley, CSIRO Energy Technology; C.R. Ward, L.G. Stephenson, L.W. Gurba, University of New South Wales, AUSTRALIA Borovac Coal Cleaning Process, Branislav Grbovic, Borovac International Pty Ltd, AUSTRALIA; Miloljub Grbovic, Borovac International Pty Ltd; Jelenko Micic, Mining Basin Kolubara d.o.o.; Miroslav Spasojevic, “Nikola Tesla” Power Plants, SERBIA
Converting Brown Coal into Chemicals and Hydrogen by Steam Cracking and Gasification, Nozomu Sonoyama, Idemitsu Kosan Co., Ltd; Kazunari Nobuta, Tokuji Kimura, Sou Hosokai, Teruoki Tago, Takao Masuda, Hokkaido University; Jun-ichiro Hayashi, Kyusyu University, JAPAN Pilot Scale Production of Humic Substances from Turkish Leonardites, Bekir Zühtü Uysal, Ufuk Gündüz Zafer, Ö. Murat Dogan, Duygu Öztan, Gazi University; Zeki Olgun, Mustafa Ozdingis, Selahaddin Anac, Turkish Coal Enterprises (TKI), TURKEY Heavy Metal Removal by Using Chemically Crosslinked Turkish Coal Based Humic Acid, Tulay Inan, Hacer Dogan, Murat Koral, TUBITAK Marmara Research Center; Selahattin Anaç, Zeki Olgun, TKI(Turkish Coal Enterprises), TURKEY Analysis of the Effect of Internal Defect on Coke Fracture Behavior by Rigid Bodies-Spring Model, Kenichi Hiraki, Hayashizaki Hideyuki, Yoshiaki Yamazaki, Tetsuya Kanai, Xiaoqing Zhang, Masakazu Shoji, Hideyuki Aoki, Takatoshi Miura, TOHOKU University, JAPAN SESSION 7 COMBUSTION: CHEMICAL LOOPING DEVELOPMENT – 1 John Wheeldon and Mehmet Tombul Ilmenite as an Oxygen Carrier in a Chemical Looping Combustion System: Reaction Kinetics and Fluidized Bed Performance, Christopher K. Clayton, Gabor Konya, Edward M. Eyring, Kevin J. Whitty, The University of Utah, USA Application of Inorganic Remains Originating from Water Purification and Sewage Sludge Ashes in Chemical Looping Combustion Process, Ewelina Ksepko, Grzegorz Łabojko, Marek Sciazko, Institute for Chemical Processing of Coal, POLAND Chemical Looping with Oxygen Uncoupling: Design Calculations and Process Engineering Simulations Using Kinetic Data, JoAnn S. Lighty, Adel F. Sarofim, Asad H. Sahir, Edward Eyring, Gabor Konya, University of Utah, USA Chemical Looping Combustion and Gasification – A Novel Technique to Produce Concentrated Stream of Hydrogen and Carbon Dioxide from Victorian Lignites, Chiranjib Saha, Ali Akhavan, Sankar Bhattacharya, Monash University, AUSTRALIA
7
TECHNICAL PROGRAM SESSION 8 GASIFICATION: UNDERGROUND COAL GASIFICATION – 1 Rohan Courtney and Ömer Sezgin Bloodwood Creek UCG Pilot 2008 – 2010, Cliff Mallet, Carbon Energy Pty. Ltd., AUSTRALIA; Burl E. Davis, Carbon Energy Pty Ltd, USA Studies on Gasification of Turkish Lignite via Underground Coal Gasification, Şahika Yürek, Kıvanç Het, Directorate of Turkish Coal Enterprises (TKİ), TURKEY Underground Coal Gasification and Applicability to Thrace Basin Lignite in Turkey, Ayşe Yildirim, Serdar Dogan, Turkish Petroleum Company, TURKEY SESSION 9 CARBON MANAGEMENT: GHG MANAGEMENT STRATEGIES AND ECONOMICS – 2 Leslie Ruppert and Bahtiyar Unver CO2-Reduction through Biomass Co-Firing in Coal Fired Power Plants, Klaus-Dieter Tigges, Roland Jeschke, Alfred Gwosdz, Alfons Leisse, Hitachi Power Europe GmbH, GERMANY Ventilation Air Methane Abatement at CONSOL Energy’s Enlow Fork Mine, Richard A. Winschel, Deborah A. Kosmack, William P. Fertall, CONSOL Energy Inc.; Jerry Gureghian, Green Holdings Corp., USA Novel Methods of Coal Seam Gas Content Determination for Estimation of Greenhouse Gas Emissions from Mining, Abouna Saghafi, CSIRO Energy Technology, AUSTRALIA Swelling of Moist Coal in Carbon Dioxide and Methane, Richard Sakurovs, Robyn Fry, Stuart Day, CSIRO Energy Technology, AUSTRALIA Evaluation of Total Porosity and the Amount of Inaccessible Pores in Coal Using Small-Angle Neutron Scattering, Yuri B. Melnichenko, L. He, Oak Ridge National Laboratory; M. Mastalerz, Indiana University, USA; R. Sakurovs, CSIRO Energy Technology; T. Blach, Griffith University, AUSTRALIA SESSION 10 COAL SCIENCE: COAL FIRES Susan J. Tewalt and Levent Ergun Early Stage Detection of Coal Spontaneous Combustion in View of Pretreatment of the Coal, Boleslav Taraba, Zdenek Pavelek, Jiri Janek, Ostrava University, CZECH REPUBLIC SEM Study of Some Indian Natural Cokes (Jhama), Ashok K. Singh, Nandita Choudhury, Central Institute of Mining & Fuel Research, CSIR; Mamta Sharma, National Metallurgical Laboratory, CSIR; Mahendra P. Singh, Banaras Hindu University, INDIA Examination of Low Temperature Air Oxidation Mechanism of Brown Coal for Supressing Self Ignition Tendency, Kouichi Miura, Ryota Okajima, Mitsunori Makino, Ryuichi Ashida, Kyoto University, JAPAN Beneficiation Prospects of Baked Coking Coals from Seam XV, Jharia Coalfield, Damodar Valley, India, Ashok K Singh, N. K. Shukla, N. Choudhury, Central Institute of Mining & Fuel Research, CSIR; Mamta Sharma, National Metallurgical Laboratory, CSIR, INDIA
SESSION 11 SUSTAINABILITY AND ENVIRONMENT – 2 Brenda Pierce and Aysel Atimtay Emissions from Cofiring Coal with Renewable Materials Such as Biomass and Sewage, Lesley Sloss, IEA Clean Coal Centre, UNITED KINGDOM Element Leaching from Coal Stockpiles – Case Studies from the Sydney and Collie Basins, Australia, Colin R. Ward, Leanne Stephenson, Zhongsheng Li, University of New South Wales; David French, Ken Riley, Owen Farrell, CSIRO Energy Technology, AUSTRALIA Coal Resource Estimation in Isiklar-Kisrakdere (Soma, Manisa, Turkey), A. Erhan Tercan, Bahtiyar Ünver, Mehmet Ali Hindistan, Hacettepe University; Perihan Çorbacı, Kıvanç Het, Turkish Coal Enterprises, TURKEY Coal Explorations in Turkey: New Projects and New Reserves, İlker Şenguler, MTA, TURKEY SESSION 12 COAL-DERIVED PRODUCTS: ACTIVATED CARBON PRODUCTION – 1 Belma Demirel and Hasan Heperkan Activated Carbon from Brown Coal by Chemical Activation, Luguang Chen, Sankar Bhattacharya, Monash University, AUSTRALIA
Effect of Coal Volatile Matter on Emissions of Boiler Combustion, Hasancan Okutan, Nalan Erdöl Aydın, Erhan Böke, İstanbul Technical University, TURKEY SESSION 14 GASIFICATION: UNDERGROUND COAL GASIFICATION – 2 Rohan Courtney and Sibel Özdogan The Improvement of UCG Processes, Karol Kostur, Technical University of Košice, SLOVAK REPUBLIC An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment, John J. Nitao, David W. Camp, Souheil M. Ezzedine, Thomas A. Buscheck, S. Julio Friedmann, Lawrence Livermore National Laboratory, USA Quantification of the Effects of Various Thermal Boundary Conditions in the Underground Coal Gasification Cavities Using a Compartment Model, Sateesh Daggupati, Ramesh Naidu Mandapati, Sanjay M Mahajani, Anuradda Ganesh, Preeti Aghalayam, IIT Bombay; Sapru R.K., Sharma R.K., ONGC, INDIA Estimation of Chemical Reaction Occurred in Underground Coal Gasification, Osamu Yamada, Mamoru Kaiho, National Institute of Advanced Industrial Science and Technology (AIST); Sohei Shimada, The University of Tokyo; Shouji Fujioka, Japan Coal Energy Center; Jie Liang, China University of Mining and Technology, JAPAN
Low-Temperature Catalytic Graphitization of Carbon Material, Ch. N. Barnakov, A.P. Kozlov, V.Yu.Malysheva, Institute of Coal and Coal Chemistry SB RAS; S.K. Seit-Ablaeva, Kemerovo Technological Institute of Food Industry; Z. R. Ismagilov, M.A.Kerzhentsev, Boreskov Institute of Catalysis SB RAS, RUSSIA
Computational Flow Modeling of Underground Coal Gasification (UCG) Process, Sateesh Daggupati, Ramesh Naidu Mandapati, Sanjay M Mahajani, Preeti Aghalayam, Anuradda Ganesh, IIT Bombay; Sapru R.K, Sharma R.K., UCG Group, IRS, ONGC, INDIA
Research on the Preparation of Highthermalconductivity Carbon Block by the Ordered Growth of Self-assembled Mesophase, Ming-Lin Jin, Rong-Hua Liu, Qing-Zhong Cheng, Jingxia Hu, Shanghai Institute of Technology; Zong-Hong Bao, Nanjing University of Technology, CHINA
SESSION 15 CARBON MANAGEMENT: CO2 SEQUESTRATION Leslie Ruppert and Hanzade Acma
SESSION 13 COMBUSTION: MERCURY AND TRACE ELEMENTS John Wheeldon and Mustafa Ziypak UNEP Coal Combustion Partnership Area Activities Prior to 2013 Global Mercury Treaty, Wojciech Jozewicz, ARCADIS, USA; Lesley Sloss, IEA Clean Coal Centre, UNITED KINGDOM; Gunnar Futsaeter, United Nations Environment Programme, SWITZERLAND Direct Measurement of Mercury in Simulated Flue Gas, Bihter Padak, Jennifer Wilcox, Stanford University, USA Modeling Trace Element Release from Included and Excluded Pyrite during Pulverized Coal Combustion, Wayne S. Seames, Esam I. Jassim, Steven A. Benson, University of North Dakota, USA Online Monitoring of Boron in Flue Gas Desulfurization Effluents by Fully Automated Measuring Equipment, Seiichi Ohyama, Keiko Abe, Hitoshi Ohsumi, Hirokazu Kobayashi, Central Research Institute of Electric Power Industry; Naotsugu Miyazaki, Koji Miyadera, Kin-ichi Akasaka, DKK-TOA Corporation, JAPAN Mercury Sorption on Brominated Activated Carbon, Erdem Sasmaz, Jennifer Wilcox, Stanford University, USA
CO2 Sequestration in Unminable Coal with Enhanced Coal Bed Methane Recovery: The Marshall County Project, Richard A. Winschel, James E. Locke, Ravi S. Srivastava, CONSOL Energy Inc.; Richard A. Bajura, Tom Wilson, Hema J. Siriwardane, Henry Rauch, Douglas Patchen, Brad D. Hega, Raj K. Gondle, West Virginia University; Arthur W. Wells, NETL/DOE, USA A GIS-DSS for a CO2-ECBM Project Feasibility Study: Case of Sulcis Coal Basin (Sardinia, Italy), Raimondo Ciccu, Alessandro Mazzella, Caterina Tilocca, University of Cagliari; Paolo Deiana, Sezione Impianti e Processi ENEA - Agenzia Nazionale per le Nuove Tecnologie, ITALY Effect of Rock Composition on Mineralization in Sequestration, Prashanth Mandalaparty, Milind Deo, Joseph Moore, University of Utah, USA Geological CO2 Storage in Coal-Bearing Formation, Sohei Shimada, Zhenjie Chai, Naoto Sakimoto, The University of Tokyo, JAPAN Experimental Study on Carbon Dioxide Sequestration by Mineral Carbonation, Jun-Ying Zhang, Heng Yan, Yong-Chun Zhao, Chu-Guang Zheng, Huazhong University of Science and Technology, CHINA CO2 Sequestration for the Shenhua DCL Plant in China, Qingyun Sun, Jerald J. Fletcher, US-China Energy Center, West Virginia University, USA
8
TECHNICAL PROGRAM SESSION 16 COAL SCIENCE: COKING Susan J. Tewalt and Erdoğan Yuzer
Clean Fuel Production Works from Canakkale-Can Coals, Oguz Altun, Akan Gulmez, Ayşe Erdem, Selami Toprak, Mineral Research and Exploration Directorate in Turkey; Zeki Olgun, Turkish Coal Enterprises, TURKEY Thermoplasticity Improvement of Coal Blends by Adding Solvent-Extracted Coal, Noriyuki Okuyama, Hiroki Shishido, Koji Sakai, Maki Hamaguchi, Nobuyuki Komatsu, KOBE STEEL, Ltd.; Haruo Kumagai, Hokkaido University, JAPAN SESSION 17 SUSTAINABILITY AND ENVIRONMENT – 3 Brenda Pierce and Mustafa Cetin Carbon Capture and Integration: An Alternative Perspective to CO2 Emissions and Carbon Capture and Sequestration, Catherine A. McGanity, University of Richmond, USA Future Coal Production Outlooks in the IPCC Emission Scenarios: Are They Plausible?, Mikael Höök, Uppsala University, SWEDEN The European Coal Market, a Prosperous Future?, Manfred Rumberger, ER-Consult GmbH, GERMANY Methane Enrichment from Anaerobic Digestion Gas (ADG) Using Polymeric Hollow Fiber Membrane, Hyung-Keun Lee, Dae-Hoon Kim, Ki-Hong Kim, YoungMo An, Hang-Dae Jo, Korea Institute of Energy Research; Gang-Woo Lee, Yoo Sung Co. R&D Center, KOREA; KiJun Baik, Yanbian University of Science and Technology, CHINA
Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences, RUSSIA Research on the Preparation of Thermal Conductivity C/C Composites by One-step Hot Press Molding, Jin Ming-Lin, Qingzhong Cheng, Shanghai Institute of Technology; Yan-Wen Zhang, Xiao-Long Zhou, East China University of Science and Technology, CHINA
ORAL SESSIONS Wednesday, October 13, 2010 11:00 - 17:50 SESSION 19 COMBUSTION: OXY-COAL DEVELOPMENT – 2 Steven Carpenter and Yucel Ozden Comparative Study of Coal Ash and Deposits from Air and Oxy-Fuel Combustion, Jost O.L. Wendt, Dunxi Yu, William J. Morris, University of Utah; Andrew Fry, Constance L. Senior, Reaction Engineering International, USA Oxy-Fuel Combustion: A Technological Option for Retrofitting Existing Pulverized Lignite Fired Power Plants in Turkey, İskender Gökalp, CNRS-ICARE, FRANCE; Mücella Ersoy, Turkish Coal Enterprises, TURKEY Oxy-Fuel Combustion Chemistry – Implications on Corrosion Related Issues, Klas Andersson, Daniel Fleig, Fredrik Normann, Filip Johnsson, Chalmers University of Technology, SWEDEN
Effect of Introduction of Clean Coal Technology on Future Asian Energy Supply, Sohei Shimada, Yuta Koyama, The University of Tokyo, JAPAN
SESSION 20 GASIFICATION: FUNDAMENTALS – 1 Ke Liu and Hüsnü Atakul
SESSION 18 COAL-DERIVED PRODUCTS: ACTIVATED CARBON PRODUCTION – 2 Belma Demirel and Özcan Gulsoy
Modeling of Coal Char Gasification in Coexistence of CO2 and H2O, Satoshi Umemoto, Shiro Kajitani, Saburo Hara, Central Research Institute of Electric Power Industry (CRIEPI), JAPAN
Investigation of Carbonization Kinetic of Tunçbilek Lignite Used for the Preparation of Activated Carbon, Burcu Özdemir, Nilgün Karatepe, Reha Yavuz, Istanbul Technical University, TURKEY
Investigation of Component Release During Pressurized, High Heating Rate Devolatilization of Coal and Petroleum Coke, David Wagner, Kevin J. Whitty, University of Utah; Glenn L Shoaf, Paul Fanning, Eastman Chemical Company, USA
Synthesis of Nitrogen-Doped Carbon Materials from Coal-Tar and Petroleum Pitches and Nitrogen Containing Organic Precursors, Z. R. Ismagilov, M.A.Kerzhentsev, I.Z.Ismagilov, Boreskov Institute of Catalysis SB RAS; Ch. N. Barnakov, A.P. Kozlov, Institute of Coal and Coal Chemistry SB RAS; E.I.Andreikov, Institute of Organik Synthesis UB RAS, RUSSIA Effect of Mineral Matter of Brown Coals on the Reactivity of Char Steam Gasification and on the Properties of Activated Carbons, P.N.Kuznetsov, Kolesnikova S.M., L.I.Kuznetsova, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences; Yu.F.Patrakov, Institute of Coal and Coal Chemistry of Siberian Branch of Russian Academy of Sciences, RUSSIA Study of the Properties of Coal from Mongolian SaikhanOvoo Deposit and the Char and Carbons Produced, B.Purevsuren, Ya.Davaajav, Kh.Serikjan, S.Batbileg, Institute of Chemistry and Chemical Technology, Mongolian Academy of Sciences, MONGOLIA; P.N Kutsnezov,
High Pressure TGA Studies on German Brown Coal under Carbon Dioxide Atmosphere, Abhishek Bhargava, Patrick J. Masset, Freiberg University of Mining and Technology, GERMANY Studies on Gasification of Char in Fixed Bed Reactor, Ramesh Naidu Mandapati, Preeti Aghalayam, Sateesh D, Narseh D, Sanjay M Mahajani, Anuradda Ganesh, IIT Bombay; Sharma R.K., IRS, ONGC, INDIA SESSION 21 CARBON MANAGEMENT: PRE-COMBUSTION CO2 CAPTURE Richard Sakurovs and Mustafa Ozdingis Development of Dry Regenerable CO2 Sorbent and WGS Catalyst for SEWGS Process, Joong Beom Lee, Tae Hyoung Eom, Jungho Ryu, Jeom-In Baek, Dong-Hyeok Choi, Keun Woo Park, Seong Jegarl, Seug-Ran Yang, Korea Electric Power Research Institute, KOREA
Development of Dry Regenerable CO2 sorbent for Fluidized-Bed CO2 Capture Process from Coal Power Plant, Chong Kul Ryu, Joong Beom Lee, Tae Hyoung Eom, Bok Suk Oh, Jeom-In Baek, Kyeongsook Kim, Young Ho Wi, Jegarl Seong, Won Sik Jeon, Korea Electric Power Research Institute, KOREA SESSION 22 COAL SCIENCE: CBM/CARBON DIOXIDE Richard Winschel and Işık Özpeker A Comparison of Experimental and Theoretical High Pressure CO2 Isotherms on Coals from the Upper Silesian Basin, Czech Republic, Zuzana Weishauptová, Oldřich Přibyl, Martina Švábová, Institute of Rock Structure and Mechanics, Academy of Sciences of the Czech Republic, CZECH REPUBLIC Conversion of Ukrainian Low Grade Solid Fuels with CO2 Capture, О.М. Dudnyk, I.S. Sokolovska, Coal Energy Technology Institute of National Academy of Sciences of Ukraine, UKRAINE Coalbed Gas Potential in the Miocene Soma Basin (Western Turkey), Sedat İnan, Aynur Dikbaş, Semih Ergintav, Fırat Duygun, TÜBİTAK Marmara Reasearch Center; M.Namık Yalçın, Kübra Tırpan, Korhan Yaşar, İstanbul University; Ender Okandan, Mustafa Baysal, Middle East Technical University; Yuda Yürüm, Sabancı University; Ruhi Saatçılar, Sakarya University; Murat Yılmaz, Ali Rıza Toygar, Turkish Petroleum Corporation; Ayhan Kösebalaban, Selahaddin Anaç, Hakkı Duran, Mehmet Onbaşı, Mücella Ersoy, Mehmet Atasayar, İsmail Ergüder, Turkish Coal Enterprises, TURKEY Underground Coal Determination by Integrated (Reflection & WVSP) Seismic in the Miocene Soma Basin (Western Turkey), Ruhi Saatçılar, Sedat İnan, Fırat Duygun, Semih Ergintav, Aynur Dikbaş, TÜBİTAK Marmara Reasearch Center; Murat Yılmaz, Ali Rıza Toygar, Turkish Petroleum Corporation; Ayhan Kösebalaban, Selahaddin Anaç, Hakkı Duran, Mehmet Onbaşı, Mücella Ersoy, Mehmet Atasayar, İsmail Ergüder, Turkish Coal Enterprises; M.Namık Yalçın, İstanbul University; Ender Okandan, Middle East Technical University; Yuda Yürüm, Sabancı University; Emin Demirbağ, İstanbul Technical University, TURKEY SESSION 23 SUSTAINABILITY AND ENVIRONMENT – 4 Brenda Pierce and Orhan Kural The Thermal Treatment Study of Pyrite from South Brazil Coal Mining Industry, Michael Peterson, Jussara P. Pizzolo, Morgana Bom, Deise Tramontin, Gabriela B. Vieira, Keli V.S. Damin, Universidade do Extremo Sul Catarinense; Adilson Oliveira, Formula Chemical Company, BRAZIL Pyrolysis of the Various Types of Fuels - Credit Cards, Dagmar Juchelkova, Helena Raclavska, Adela Cízkova, Vaclav Roubicek, VSB-Technical University of Ostrava, CZECH REPUBLIC SESSION 24 COAL-DERIVED PRODUCTS: DIRECT COAL-TO-LIQUIDS Rachid Oukaci and Fatma Arslan Alliance DCL Technology for Producing Ultra Clean Transportation Fuels, Theo L.K. Lee, Headwaters CTL, LLC.; John Duddy, Axens North America, Inc., USA
9
TECHNICAL PROGRAM Direct Coal to Liquids (DCL): High Quality Jet Fuels, W. Weiss, H. Dulot, A. Quignard, N. Charon, M. Courtiade, IFP New Energy, FRANCE Extraction of Brown and Sapropelitic Coals with Toluene and Water Containing Fluids under Supercritical Conditions, P.N.Kuznetsov, S.M.Kolesnikova, L.I.Kuznetsova, E.S.Kamensky, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences; V.A.Kashirtsev, Trofimuk Institute of Petroleum Geology and Geophysics SB RAS, RUSSIA Research and Development of New Type Reactor for Direct Coal Liquefaction, Hu Fating, Li Peilin, Shi Shidong, Liu Min, Beijing Research Institute of Coal Chemistry, China Coal Research Institute, CHINA Characterization and Distribution of Phenolics in Direct Coal Liquefaction Oils, Zhennan Gao, Xuefeng Mao, Beijing Research Institute of Coal Chemistry, China Coal Research Institute; Wenhua Li, GE China Technology Center, CHINA SESSION 25 COMBUSTION: CHEMICAL LOOPING DEVELOPMENT – 2 Steven Carpenter and Murat Özbayoglu Development of Real-Time Dynamic Simulation of Chemical Looping Process for Advanced Controls, Xinsheng Lou, Hao Lei, Abhinaya Joshi, Carl Neuschaefer, Alstom Power Inc., USA Effect of H2S on Chemical Looping Combustion of CoalDerived Synthesis Gas over Fe2O3 - MnO2 Supported on ZrO2/Sepiolite/Al2O3, Ewelina Ksepko, Marek Sciazko, Institute for Chemical Processing of Coal, POLAND; Ranjani V. Siriwardane, Hanjing Tian, Thomas Simonyi, James A. Poston Jr, U.S. DOE, USA Effect of Coal Blending Method on Combustion Characteristics and NOx Emission in a Drop Tube Furnace, Song-gon Kim, Chun-sung Lee, Byoung-Hwa Lee, Ju-Hun Song, Young-June Chang, Chung-Hwan Jeon, Pusan National University, SOUTH KOREA TGA and DTF Studies on Coal Blends to Assess Combustion Performance, P. Sarkar, A. Mukherjee, S. G. Sahu, A. Choudhury, A. K. Adak, M. Kumar, N. Choudhury, S. Biswas, Central Institute of Mining and Fuel Research; S. Ghosal, Jadavpur University, INDIA SESSION 26 GASIFICATION: FUNDAMENTALS – 2 Ke Liu and Ayhan Sirkeci Effect of Operation Parameters on Gasification for the Production of Synthesis Gas, Atilla Biyikoğlu, Bekir Zühtü Uysal, Gazi University; Afşin Güngör, Niğde University; Coşku Kasnakoğlu, Murat Özbayoğlu, TOBB University of Economics and Technology, TURKEY High-Pressure and High-Temperature Gasification of Upgraded Brown Coal Char Using a Mini Direct Heating Reactor, Kouichi Miura, Ryuichi Ashida, Mitsunori Makino, Xian Li, Kyoto University, JAPAN C F D Si mulation of Proces s - dr i v e n Pa r t i c l e Fragmentation in a Coal Bed Gasifier, Franz Holzleithner, Roland Eisl, Markus Haider, Institute for Energy Systems and Thermodynamics, Vienna University of Technology; Georg Aichinger, Siemens VAI Metals Technologies Gmbh&Co, AUSTRIA Experimental Investigation into Primary Fragmentation of Large Coal Particles, Adam Luckos, Roelof L.J. Coetzer, Ed L. Koper, Sasol Technology R&D, SOUTH
AFRICA; Monika Kosowska-Golachowska, Częstochowa University of Technology, POLAND
Joubert, Trudie Brittz, Sasol Technology R&D, SOUTH AFRICA
Investigation of the Pyrolysis and Gasification of a Turkish Coal Using Thermal Analysis Coupled with Mass Spectrometry, Sibel Özdoğan, Ugur Özveren, Mehmet Beypınar, Marmara University; Aylin Boztepe, Yildiz Technical University, TURKEY
Dry Coal Cleaning Using the FGX Separator, Mehmet Saracoglu, Rick Q. Honaker, University of Kentucky, USA
SESSION 27 CARBON MANAGEMENT: POST-COMBUSTION CO2 CAPTURE – 1 Richard Sakurovs and Mehmet Cagil Development of Post Combustion Carbon Capture Technology, Matthew Hunt, F. D. Fitzgerald, S. Black, R. A. Gardiner, Doosan Power Systems, UNITED KINGDOM Lignite Derived Carbons for CO2 Capture, Alan L Chaffee, Seamus Delaney, Gregory Knowles, Monash University, AUSTRALIA Physical Properties and Reactivities of Mg-Based Dry Regenerable CO2 Sorbents Prepared by Spray-Drying Method, Jeom-In Baek, Tae Hyoung Eom, Joong Beom Lee, Won Sik Jeon, Chong Kul Ryu, Korea Electric Power Research Institute, KOREA SESSION 28 COAL SCIENCE: COAL CHEMISTRY – 2 Frans Waanders and Dündar Ergunalp Carbonaceous Particles from the Incomplete Combustion, Martina Havelcová, Ivana Sýkorová, Hana Trejtnarová, Alexandr Šulc, Institute of Rock Structure and Mechanics AS CR,v.v.i., CZECH REPUBLIC Trace Element Partitioning and Leaching in Solids Derived from Gasification of Australian Coals, Alexander Ilyushechkin, Daniel Roberts, David Harris, Ken Riley, CSIRO Energy Technology, AUSTRALIA Effect of Coal Rank on Carbon Oxides Formation via the Low Temperature Atmospheric Oxidation Process, Uri Green, Zeev Aizenstat, Hebrew University of Jerusalem; Haim Cohen, Ariel University Center at Samaria and Ben-Gurion University of the Negev, ISRAEL; Christoph Wiedner, Sven Stark, Ariel University Center at Samaria, ISRAEL and TU Bergakademie Freiberg, GERMANY Quantitative Determination of Minerals in Coal by CQPAC Method, Zdeněk Klika, VŠB-Technical University Ostrava, CZECH REPUBLIC The Catalytic Effect of Added Sodium- and Potassium Carbonate to an Acid Treated Inertinite Rich South African Bituminous Coal Char, Christien A Strydom, Lucinda Klopper, John R Bunt, North-West University, SOUTH AFRICA; Harold H Schobert, Penn State University, USA SESSION 29 COAL SCIENCE: BENEFICIATION – 1 B.K. Parekh and Selçuk Buyurgan Pre-Combustion Cleaning of Pulverized Fine Coal at Power Plant Using Novel RTS Dry Separation Technology, Daniel Tao, Ahmed Sobhy, Qin Li, Rick Honaker, University of Kentucky, USA The Prediction of Caking Propensity of Gondwanaland Coals Using Petrography, Daniel Van Niekerk, Johan
Improving the Efficiency of Lignite Drying, Wayne S. Seames, Carlos J. Bucaram, Steven A. Benson, University of North Dakota; Srivats Srinivasachar, Envergex, LLC, USA SESSION 30 COAL-DERIVED PRODUCTS: COAL-TO-LIQUIDS/FISCHER-TROPSCH – 1 Rachid Oukaci and Alp Gurkan Deactivation of Iron Based Fischer-Tropsch Catalyst: A Critical Problem, Belma Demirel, Deena Ferdous, Rentech, Inc., USA Product Distribution and Reaction Pathways during Fischer-Tropsch Synthesis on an Iron Catalyst, Dragomir B. Bukur, Texas A&M University at Qatar, QATAR; Lech Nowicki, Technical University of Lodz, POLAND SESSION 31 COMBUSTION: FLUIDIZED-BED COMBUSTION AND CO-FIRING – 1 Steven Carpenter and Gündüz Atesok Smartsheet Tool Applied to Boiler Performance Analysis and Economic Optimization of a Circulating Fluidized Bed Boiler, Abhinaya Joshi, Xinsheng Lou, Carl Neuschaefer, Paul Panos, Alstom Power Inc.; Weikko Wirta, AES Thames, USA A Model of Primary Fragmentation of Coal Particles in Fluidized-Bed Combustion, Adam Luckos, Sasol Technology R&D, SOUTH AFRICA; Monika KosowskaGolachowska, Częstochowa University of Technology, POLAND Main Problems Concerning Co-Firing Biomass Mixture with Hard Coal in Pulverized-Fuel Boilers, Krzysztof Jesionek, Henryk Karcz, Marcin Kantorek, Wrocław University of Technology, POLAND Te c h n i c a l a n d E c o n o m i c E v a l u a t i o n o f t h e Desulphurization Processes at Power Stations Using Lignite, Hasancan Okutan, Bülent D. Çift, İstanbul Technical University, TURKEY Desulfurization Characteristics of Powdered Hydrated Lime for Flue Gas Sorbent Injection Process, Hyok Bo Kwon, Kyungnam University; Sang Whan Park, KIST; Hyung Taek Kim, Ajou University, KOREA Status of Large Circulating Fluidized Bed Boiler Operation in China, Jie Yu, Beijing Research Institute of Coal Chemistry, China Coal Research Institute, CHINA SESSION 32 GASIFICATION: FUNDAMENTALS – 3 Ke Liu and Gülhan Özbayoglu Performance of a 500 KWTH Pressurized EntrainedFlow Coal Gasifier, Kevin J. Whitty, Randy Pummill, David R. Wagner, Travis Waind, David A. Wagner, The University of Utah, USA Characterization of a Small Scale Slurry-Fed, OxygenBlown Entrained Flow Gasifier: How Injector Geometry Affects Flame Stability and Performance, Travis Waind, Kevin Whitty, University of Utah, USA
10
TECHNICAL PROGRAM
Analysis of Fines Produced from Non-Slagging Coal Gasifier and Evaluation of Economic Usage, Yongseung Yun, Seok Woo Chung, Na Rang Kim, Institute for Advanced Engineering, KOREA Development of Gas, Power and Tar Co-Generation System with Circulating Fluidized Bed Technology, Qinhui Wang, Mengxiang Fang, Zhongyang Luo, Mingjiang Ni, Kefa Cen, Zhejiang University, CHINA
Research and Exploration Directorate in Turkey; Zeki Olgun, Turkish Coal Enterprises, TURKEY
Tests on Drop Tube, Radim Paluska, Marian Bojko, VSB – Technical University of Ostrava, CZECH REPUBLIC
Drying Kinetics of Çanakkale – Çan Lignite, Ufuk Gündüz Zafer, Ö. Murat Dogan, Duygu Öztan, Bekir Zühtü Uysal, Gazi University; Zeki Olgun, Mustafa Ozdingis, Ömer Sezgin, Selahaddin Anac, Turkish Coal Enterprises (TKI), TURKEY
SESSION 38 GASIFICATION: FUNDAMENTALS – 4 Johan van Dyk and Mevlüt Kemal
Thermal Chemical Process Study on Chemical Reaction Network of Jet-Fluidized Bed Gasifer Reaction System, Jie Feng, Xuecheng Hou, Wenying Li, Xiao-Hui Chen, Taiyuan University of Technology, CHINA
Innovative High Energy Efficiency Brown Coal Drying based on Self-Heat Recuperation Technology, Muhammad Aziz, Chihiro Fushimi, Yasuki Kansha, Kazuhiro Mochidzuki, Shozo Kaneko, Atsushi Tsutsumi, The University of Tokyo, JAPAN
Optimization of Canadian Petroleum Coke, Coal and Fluxing Agent Blends via Slag Viscosity Measurements and Models, Marc A. Duchesne, Arturo Macchi, University of Ottawa; Ben Anthony, CanmetENERGY, CANADA; Alexander Y. Ilyushechkin, CSIRO Energy Technology, AUSTRALIA
SESSION 33 CARBON MANAGEMENT: POST-COMBUSTION CO2 CAPTURE – 2 Richard Sakurovs and Mustafa Tiris
SESSION 36 COAL-DERIVED PRODUCTS: COAL-TO-LIQUIDS/FISCHER-TROPSCH – 2 Rachid Oukaci and Ali İhsan Arol
An Efficient Membrane Process to Capture Carbon Dioxide From Power Plant Flue Gas, Bilgen Firat Sercinoglu, Tim Merkel, Xiaotong Wei, Haiqing Lin, Jenny He, Richard Baker, Karl Amo, Hans Wijmans, Membrane Technology and Research, Inc., USA
Using Pyrolysis Tar to Meet Fuel Specifications in Coal-to-Liquids Plants, Jaco Schieke, Foster Wheeler, UNITED KINGDOM
Effect of Dense Medium Separation of a South African Coal Source on Slag-Liquid Formation: An Experimental and Factsage Approach, JC van Dyk, SASOL Technology; FB Waanders, North West-University, SOUTH AFRICA
CO2 Capture by Condensed Rotational Separation, R.J. van Benthum, H.P. van Kemenade, J.J.H. Brouwers, M. Golombok, Eindhoven University of Technology, THE NETHERLANDS Influence of Pressure on Dry Reforming of Methane over Carbonaceous Catalyst, Bingmo Zhang, Yongfa Zhang, Guojie Zhang, Fengbo Guo, Taiyuan University of Technology, CHINA SESSION 34 COAL SCIENCE: COAL CHEMISTRY – 3 Frans Waanders and Neset Acarkan Predetermination of the Fault Crossing the Underground Coal Mine Galeries by Seismic Reflection Method: An Application at a Longwall Coal Mine in Turkey, G.G.U. Aldas, B. Kaypak, B. Ecevitoglu, Ankara University; A. Can, General Directorate of Mineral Research and Exploration, TURKEY Alberta’s 2 Trillion Tonnes of ‘Unrecognized’ Coal, R.J. Richardson, CanZealand Geoscience Ltd., NEW ZEALAND Temperature as a Factor Affecting Adsorption Behavior of Coal to Lead (II) Ions, Boleslav Taraba, Petra Vesela, Roman Marsalek, Zuzana Navratilova, Ostrava University, CZECH REPUBLIC Bioflotation of Coal, Peter Fecko, Tana Kantorkova, Radomir Michniak, Lukas Koval, Alena Kasparkova, VSB - Technical University of Ostrava, CZECH REPUBLIC SESSION 35 COAL SCIENCE: BENEFICIATION – 2 B.K. Parekh and Ali Güney Suitability of the Sulcis Coal for CWS Preparation, Raimondo Ciccu, Giovanni Mei, Caterina Tilocca, University of Cagliari; Paolo Deiana, Sezione Impianti e Processi ENEA - Agenzia Nazionale per le Nuove Tecnologie, ITALY
Overview of the Rentech Process, Belma Demirel, Harold Wright, Rentech, Inc., USA
Compositional Variations in Pilot Gasifier and Laboratory-Produced Slags and their Impacts on Slag Viscosity and Coal Assessment, Alexander Ilyushechkin, D. Roberts, D. Harris, CSIRO, Energy Technology, AUSTRALIA
Comparative Evaluation of Different Coal to Liquid Process Conditions via Fischer-Tropsch Synthesis, Serhat Gul, Atilla Ersoz, Murat Baranak, Omer Faruk Gul, Fehmi Akgun, TUBİTAK Marmara Research Centre, Energy Institute, TURKEY
Shaping Slag Flow in an Entrained Flow Gasifier: Numerical Simulation and Physical Experiments, Randy Pummill, Gabriel Hansen, Kevin Whitty, University of Utah, USA
Technoeconomic and Environmental Life Cycle Analysis of Coal and Coal/Biomass to Liquids Facilities, Anastasia M Gandrik, Idaho National Laboratory; Vivek P. Utgikar, University of Idaho, USA
Influence of Gasification Conditions on the Properties of Fly Ash in a Bench-Scale Opposed Multi-Burner Gasifier, Qinghua Guo, Guangsuo Yu, Fuchen Wang, Zhenghua Dai, East China University of Science and Technology, CHINA
Conversion of Waste Biomass to Transportation Fuels: Energy for the Future, S.K.Srivastava, S.R.K.Rao, Amlendu Sinha, Central Institute of Mining and Fuel Research, INDIA
ORAL SESSIONS Thursday, October 14, 2010 11:00 - 17:50 SESSION 37 COMBUSTION: OXY-COAL DEVELOPMENT – 3 John Wheeldon and Ersan Putun The Effect of Coal Composition on Ignition and Flame Stability in Co-Axial Oxy-Fuel Turbulent Diffusion Flames, Dadmehr M. Rezaei, Eric G. Eddings, Kerry E. Kelly, Jingwei Zhang, Jost O.L. Wendt, University Of Utah, USA; Yuegui Zhou, Shanghai Jiao Tong University, CHINA Study on the In-Furnace Desulfurization in Oxy-Fuel Combustion using Drop Tube Furnace with Limestone, Hyung-Keun Lee, Wook Choi, Hang-Dae Jo, Won-Kil Choi, Korea Institute of Energy Research; Sang-In Keel, Korea Institute of Machinery & Materials, KOREA Ignition Loss and Ultrafine Particle and Soot Emissions From Air and Oxy-Coal Flames, William J. Morris, Dunxi Yu, Jost O. L. Wendt, University of Utah, USA
Comparative Study of Oil Agglomeration and Flotation of Low Grade Coals, Feridun Boylu, Fırat Karakas, Istanbul Technical University, TURKEY
High Speed Video Analysis of Oxycoal Combustion in 40kw Coaxial Turbulent Diffusion Flames, Terry A. Ring, Jingwei Zhang, Husam el Gendy, Jost O.L. Wendt, Kerry Kelly, Eric G. Eddings, University of Utah, USA
Briquetting Studies of Canakkale-Can Coals, Oguz Altun, Akan Gulmez, Ayşe Erdem, Zafer Gencer, Mineral
Comparison of the Mathematical Model of Pulverized Coal Burnout with Results Gained from Experimental
SESSION 39 GASIFICATION: GENERAL SESSION – 2 Ting Wang and Muammer Bulut GTI’s Sampling and Analysis Systems for Gas Streams of Gasification and Downstream Processes, Osman M. Akpolat, Tanya S. Tickel, Rachid B. Slimane, Chun W. Choi, Gas Technology Institute (GTI), USA Design of Comminution Unit for the Gasification Pilot Plant, N. Acarkan, G. Onal, A. A. Sirkeci, G. Atesok, Istanbul Technical University, TURKEY Preparation of Coal Water Mixture with High Concentration from Low Rank Coals and Lignite by Dry Fine Coal with Optimum Particle Distribution, Baoqing Li, Institute of Coal Chemistry, Chinese Academy of Sciences; Feng He, Yulin Western Coal Technology Research Center, CHINA Fluidised Bed Co-Gasification of Coal and Biomass Under Oxy-Fuel Conditions, Marcos Millan, Nicolas Spiegl, Nigel Paterson, Cesar Berrueco, Imperial College London, UNITED KINGDOM Co-Gasification of Footwear Leather Waste and High Ash Coal: A Thermodynamic Analysis, Rodolfo Rodrigues, Nilson R. Marcílio, Jorge O. Trierweiler, Federal University of Rio Grande do Sul (UFRGS); Marcelo Godinho, University of Caxias do Sul (UCS); Adriene M. S. Pereira, Pontifical Catholic University of Rio Grande do Sul (PUC-RS), BRAZIL
11
TECHNICAL PROGRAM SESSION 40 COAL SCIENCE: COAL CHEMISTRY – 4 Ashok K. Singh and Omer Unver Mineralogical, Petrographic and Geochemical Features of the Achlada and Mavropigi Lignite Deposits, NW Macedonia, Greece, Colin R. Ward, Zhongsheng Li, University of New South Wales; Stavros P. Kalaitzidis, BHP Billiton Mitsubishi Alliance, AUSTRALIA; Nikolaos Koukouzas, Centre for Research and Technology Hellas, Institute for Solid Fuels Technology and Applications, GREECE X-Ray Computer Tomography on Coal Particles, Patrick J. Masset, Freiberg University of Mining and Technology, GERMANY; Heikki Suhonen, European Synchrotron Radiation Facility, FRANCE Mineralogy, Geochemistry, and Petrography of Upper Permian Bituminous and Carboniferous Anthracite Coals from Xuanwei County, Eastern Yunnan Province, China, Harvey E. Belkin, U.S. Geological Survey; James C. Hower, Jordan W. Drew, University of Kentucky, CAER, USA; Linwei Tian, Chinese University of Hong Kong, CHINA SESSION 41 COAL SCIENCE: BENEFICIATION – 3 B.K. Parekh and Steven Carpenter Pyrolysis Residue from Waste Materials in Black Coal Flotation, Peter Fecko, Alena Kasparkova, Vlastimil Kriz, Josip Isek, Tien Pham Duc, VSB – Technical University of Ostrava, CZECH REPUBLIC Studies of a Multi Gravity Separator (MGS) to Produce Clean Coal from Turkish Lignite and Hard Coal Fine Tailings, Eyüp Sabah, Selçuk Özgen, M. Fatih Can, Afyon Kocatepe University, TURKEY Petrographic Characterisation of Beneficiated Material of Tailings from Soma Işıklar Derekoy Coal Washing Plant (TURKEY),by Multi Gravity Seperator (MGS), Selami Toprak, Ayşe Erdem, Akan Gulmez, Oguz Altun, Mineral Research and Exploration Directorate in Turkey; Sarper Alyildiz, Turkish Coal Enterprises, TURKEY Soma Region’s Coals Washing at Dereköy Washery with Working 800 TPH and it’s Washing Performance Evaluation, S.I. Alyildiz, S. Gurkan, S. Tuncer, Directorate of the Aegean Lignite Establishment, TURKEY SESSION 42 COAL-DERIVED PRODUCTS: H2 PRODUCTION/SNG Dragomir Bukur and Vedat Arslan Development of Hydrogen Transport Membranes for Separating Hydrogen from Coal Gasification Stream, U. (Balu) Balachandran, T. H. Lee, C. Y. Park, Y. Lu, S. E. Dorris, Argonne National Laboratory, USA HTGR-Integrated Coal to Liquids Production Analysis, Anastasia M Gandrik, Rick A. Wood, Idaho National Laboratory, USA Methane Production from Coal, Coal-Biomass Mixtures, Arun SK Raju, Christophe Capelli, Viresco Energy LLC, USA Methanation of Syngas over Coral Reef-Ni/Alumina Catalysts, Yizhuo Han, Yisheng Tan, Institute of Coal Chemistry, Chinese Academy of Science; Shengli Ma, Graduate University of the Chinese Academy of Sciences, CHINA
SESSION 43 COMBUSTION: FLUIDIZED-BED COMBUSTION AND CO-FIRING – 2 John Wheeldon and Daman Wallia
Entrained Flow Slagging Slurry Gasification and the Development of Computational Fluid Dynamics Models at CanmetENERGY, Robin Hughes, Dennis Lu, Adrian Majeski, Ben Anthony, CanmetENERGY; Andrew Corber, National Research Council, CANADA
Sulphur Capture Under Fluidised Bed Combustion Conditions Using Coal Ashes as Sorbents, Rufaro Kaitano, Dursman Mchabe, Raymond C Everson, Hein W J P Neomagus, North-West University, SOUTH AFRICA
Numerical Simulation Analyses of an Entrained-Bed Gasification Reactor, Ming-Hong Chen, Tsung Leo Jiang, National Cheng Kung University; Yau-Pin Chyou, ChangBin Huang, Institute of Nuclear Energy Research Atomic Energy Council, TAIWAN, ROC
CO 2 Reduction Potential and Co-Combustion Possibilities of the FBC-Boilers on the Czech Conditions, Dagmar Juchelkova, Helena Raclavska, Jiri Bilik, Pavlina Pustějovská, VSB-Technical University of Ostrava, CZECH REPUBLIC
SESSION 46 COAL SCIENCE: COAL CHEMISTRY – 5 Quentin Campbell and Frantisek Verbich
Co-Combustion of Various Biowastes with a HighSulfur Turkish Lignite in a Circulating Fluidized Bed Combustor, Aysel T. Atimtay, Murat Varol, Middle East Technical University; Hayati Olgun, Alper Unlu, Berrin Bay, Ufuk Kayahan, TUBITAK-MRC, Energy Institute; Hüsnü Atakül, Mustafa C. Çelebi, Istanbul Technical University, TURKEY Co-Combustion Performance of Oil Shale and Biomass Fuels, Emre Özgür, Mustafa Verşan Kök, Middle East Technical University, TURKEY; Sharon Falcone Miller, Bruce G. Miller, Penn State University,USA SESSION 44 GASIFICATION: FUNDAMENTALS – 5 Johan van Dyk and Deniz Üner Influence of Steam on the Release of Alkali Metal, Chlorine, and Sulphur Species During High Temperature Gasification of Lignite, Marc Bläsing, Michael Müller, Institute of Energy Research (IEF-2), GERMANY Kinetics of Char and Catalyzed Char Gasification under High H2 and Steam Partial Pressure, Katsuhiro Nakayama, Yoshizo Suzuki, National Institute of Advanced Industrial Science and Technology; Shiying Lin, Japan Coal Energy Center, JAPAN Modeling of Steam Gasifier in Dual Fluidized Bed Gasification, Toshiyuki Suda, Zhihong Liu, Makoto Takafuji, Masahiro Narukawa, IHI Corporation, JAPAN Steam Gasification of Low Rank Coals with IonExchanged Sodium Catalysts Prepared Using Natural Soda Ash, Yasuo Ohtsuka, Yuu Hanaoka, Enkhsaruul Byambajav, Takemitsu Kikuchi, Naoto Tsubouchi, Tohoku University, JAPAN Separation of Pyrolysis from Fluidized Bed Steam Gasification: Its Conception and Application, Masahiro Narukawa, Makoto Takafuji, Toshiyuki Suda, IHI Corporation, JAPAN
Comparison of Measured and Calculated Viscosities of German Lignite Based Slags, Arne Bronsch, Patrick J. Masset, Freiberg University of Mining and Technology, GERMANY Drying Mechanism of Low Rank Coal with Different Reacting Conditions: Fixed Bed vs. Fluidized Bed, Hyungtaek Kim, Taejin Kang, Doman Jeon, Ajou University; Sihyun Lee, Sangdo Kim, Korea Institute of Energy Research, SOUTH KOREA The Natural Technology for Pretreatment and Utilization of the Energetic Fly Ash, Maria Kusnierova, Maria Prascakova, Institute of Geotechnics of Slovak Academy of Sciences, SLOVAK REPUBLIC; Peter Fecko, Rudolf Matysek, VSB-Technical University of Ostrava, CZECH REPUBLIC Physical Structure and Chemical Properties of Organic Matter of Brown Coals from Different Fields in Relation to the Composition of Mineral Components, P.N.Kuznetsov, L.I.Kuznetsova, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences, RUSSIA SESSION 47 COAL SCIENCE: BENEFICIATION – 4 B.K. Parekh and Ekrem Yuce A Study to Recover Coal from Turkish Lignite Fine Coal Tailings: Comparison of Falcon Concentrator and Multi Gravity Separator (MGS), Eyüp Sabah, M. Fatih Can, Selçuk Özgen, Afyon Kocatepe University, TURKEY Evaluation of Dense Medium Separation Performance of Imbat Coal Preparation Plant, G.Özbayoğlu, Atilim University; Ü. Atalay, Ali İ.Arol, O.Sivrikaya, Middle East Technical University, TURKEY Effect of Shape Factor on Coal Flotation, G. Bulut, O. Güven, K.T. Perek, Istanbul Technical University, TURKEY
SESSION 45 GASIFICATION: MODELING – 1 Ting Wang and Ziya Cosar
SESSION 48 COAL-DERIVED PRODUCTS: GENERAL SESSION – 1 Dragomir Bukur and Hüseyin Özdag
Process Simulation - The Way from Pilot Plant to a Training-Centre for a 500 MW Gasifier System, Friedemann Mehlhose, Julia Kittel, S. Stoye, H. Kotthaus, Siemens Fuel Gasificaiton Technology GmbH & Co.KG, GERMANY
Reforming of Low Rank Coal by Solvent Treatment at Around 350OC, Xian Li, Ryuichi Ashida, Hiroyasu Fujitsuka, Kouichi Miura, Kyoto University, JAPAN
A Dynamic Simulator of a Commercial-Scale IGCC Plant, Mi-Yeong Kim, Yong-Jin Joo, In-Kyu Choi, JoongWon Lee, Si-Moon Kim, Min-Churl Lee, Korean Electrical Power Corporation, KOREA
An Experimental Investigation of Factors Related to Coke Strength Degradation in Coke Milli-Structure, Tetsuya Kanai, Yoshiaki Yamazaki, Kenichi Hiraki, Xiaoqing Zhang, Masakazu Shoji, Hideyuki Aoki, Takatoshi Miura, TOHOKU University, JAPAN
12
TECHNICAL PROGRAM
An Experimental Study on the Effect of Metallic Iron Particles on Strength Factors of Coke after CO2 Gasification Reaction, Yoshiaki Yamazaki, Kenichi Hiraki, Tetsuya Kanai, Xiaoqing Zhang, Masakazu Shoji, Hideyuki Aoki, Takatoshi Miura, Tohoku University, JAPAN SESSION 49 COMBUSTION: ASH DEPOSITION AND HEAT TRANSFER John Wheeldon and Oktay Erbatur Spectral Emissivities of Ni and Fe based Boiler Tube Materials with Varying Chromium Content at High Temperature Atmospheres, Miki Shimogori, BabcockHitachi K.K. Kure Research Laboratory, JAPAN; Fabian Greffrath, Viktor Scherer, Ruhr University of Bochum; Alfred Gwosdz, Christian Bergins, Hitachi Power Europe GmbH, GERMANY Effect of MGO Additive on the Reduction of Ash Deposition of Upgraded Brown Coal, Katsuya Akiyama, Haeyang Pak, Kobe Steel, Ltd.; Yasuaki Ueki, Ryo Yoshiie, Ichiro Naruse, Nagoya University, JAPAN Modeling and Optimization of NOx Emission and Pulverized Coal Flame in Utility Scale Furnaces, Srdjan Belosevic, Miroslav Sijercic, Branislav Stankovic, Nenad Crnomarkovic, Institute of Nuclear Sciences Vinca, Laboratory for Thermal Engineering and Energy; Slobodan Djekic, Electric Power Industry of Serbia, SERBIA Observation of Heat Release Region as Functions of Coal Properties in Turbulent Jet Pulverized Coal Flames, Yon Mo Sung, Cheor Eon Moon, Seong Yool Ahn, Jae Woo An, Gyung Min Choi, Duck Jool Kim, Pusan National University, SOUTH KOREA
SESSION 51 GASIFICATION: MODELING – 2 Ting Wang and Volkan S. Ediger
SESSION 54 COAL-DERIVED PRODUCTS: GENERAL SESSION – 2 Dragomir Bukur and Mehmet Canbazoglu
Investigation of Coal Gasification Process under Various Operating Conditions Inside a Two-Stage Entrained Flow Gasifier, Ting Wang, Armin Silaen, University of New Orleans, USA
Natural-Gas-Level Emissions when Burning Naphtha (without Water Injection) in a Commercial Gas Turbine using the LPP Technology, Creating a “Clean Power” Alternative for an Integrated Gasification Combined Cycle (IGCC) Polygen Plant, Leo D. Eskin, LPP Combustion, LLC., USA
Start-Up Behavior of a Fixed Bed Gasifier: One Dimensional Modeling, Giampaolo Mura, Mariarosa Brundu, University of Cagliari, ITALY Entrained Flow Coal Gasification: Modeling, Simulation & Experimental Uncertainty Quantification for a Laboratory Reactor, Philip J. Smith, Charles Reid, Julen Pedel, Jeremy Thornock, Institute for Clean and Secure Energy, The University of Utah, USA Numerical Simulation of the Hydrodynamics of a Fluidized Bed Combined with an Entrained Bed Gasifier, Jiantao Zhao, Jiejie Huang, Yitian Fang, Yang Wang, Institute of Coal Chemistry, Chinese Academy of Sciences, CHINA Numerical Simulation of the Gasification Process inside a Cross-Type Two-Stage Gasifier, Yau-Pin Chyou, ChangBin Huang, Yan-Tsan Luan, Institute of Nuclear Research, Atomic Energy Council, TAIWAN, ROC; Ting Wang, University of New Orleans, USA SESSION 52 COAL SCIENCE: COAL CHEMISTRY – 6 Quentin Campbell and Mustafa Aktas
Mathematical Model of the Low-Temperature Oxidation of Coal in Coal Stockpiles and Dumps, Marian Bojko, Milada Kozubkova, VŠB-Technical University; Zdeněk Michalec, Institute of Geonics AS CR, v. v. i., CZECH REPUBLIC
Uranium and Some Other Trace Metal Element Concentration of Some Turkish Coal Ashes, Isik Ozpeker, Fikret Suner, Mehmet Maral, Tahsin Aykan Kepekli, Istanbul Technical University, TURKEY
SESSION 50 GASIFICATION: GAS CLEANUP Johan van Dyk and İsmail Boz
Transformations of Karaman -Ermenek Lignites of Turkey under Accelerated Electrons Impact, Islam Mustafayev, Fethullah Chichek, Azerbaijan National Academy of Sciences, AZERBAIJAN; Guven Onal, Istanbul Technical University, TURKEY
Slipstream Tests of Palladium Sorbents for High Temperature Capture of Mercury, Arsenic and Selenium from Fuel Gas, Hugh G.C. Hamilton, Liz Rowsell, Stephen Poulston, Andrew Smith, Johnson Matthey Technology Centre, UNITED KINGDOM; Tony Wu, Subhash Datta, Robert C. Lambrecht, John Wheeldon, National Carbon Capture Center; Evan J. Granite, Henry W. Pennline, U.S. DOE/NETL, USA Mercury Measurement and Removal from an Entrained Flow Slagging Coal Gasifier, Dennis Lu, Robin Hughes, Ben Anthony, CanmetENERGY/Natural Resources Canada; Karl Abraham, Environment Canada, CANADA Performance Improvement of a Desulfurization Sorbent for Warm Synthesis Gas Cleanup, Jeom-In Baek, Jungho Ryu, Tae Hyoung Eom, Joong Beom Lee, Yong-Ho Wi, Chong Kul Ryu, Korea Electric Power Research Institute, KOREA
Desulfurization and Kinetics of Removal of Sulfur from High Sulfur Coal under Hydrogen Atmosphere, Guojie Zhang, Yongfa Zhang, Fengbo Guo, Bingmo Zhang, Taiyuan University of Technology, CHINA
Laminar Flame Speed Study of Syngas Mixtures (H2CO) with Straight and Nozzle Burners, İskender Gökalp, Nicolas Bouvet, Christian Chauveau, CNRS-Institut de Combustion, FRANCE; Seong-Young Lee, Robert J. Santoro, The Pennsylvania State University, USA Structural Changes in Bituminous Coal Fly Ash Due to Treatments with Aqueous Solutions, Roy Nir Lieberman, Ariel Goldman, Ariel University Center of Samaria; Haim Cohen, Ariel University Center at Samaria and Ben-Gurion University of the Negev, ISRAEL; Roy Nitzsche, TU Bergakademie Freiberg, GERMANY Influence Factors on Density and Specific Surface Area (Blaine Value) of Fly Ash from Pulverized Coal Combustion, Hiromi Shirai, Michitaka Ikeda, Kenji Tanno, Central Research Institute of Electric Power Industry, JAPAN
POSTER SESSIONS Tuesday, October 12, 2010 18:20 - 20:20 POSTER SESSION 1 COMBUSTION Ionic Liquids with Amine Functional Group: A Shortcut to Improve the Performance of Ionic Liquids for CO2 Scrubbing, Jelliarko Palgunadi, Jin Kyu Im, Antonius Indarto, Hoon Sik Kim, Minserk Cheong, Kyung Hee University, KOREA Absorption of Sulfur Dioxide in Task Specific Ionic Liquids Containing SO2-Philic Groups on the Cation, Sung Yun Hong, Jelliarko Palgunadi, Hoon Sik Kim, Minserk Cheong, Kyung Hee University, KOREA
SESSION 53 COAL SCIENCE: BENEFICIATION – 5 B.K. Parekh and Mehmet S. Celik
Reaction Characteristics of New Oxygen Carriers for Chemical Looping Combustion, Ho-Jung Ryu, Jaehyeon Park, Gyoung-Tae Jin, Korea Institute of Energy Research; Moon-Hee Park, Hoseo University, KOREA
Preparation of Alternative Fuel from Compost and Coal Slurries, Dagmar Juchelkova, O. Zajonc, H. Skrobankova, H. Raclavska, K. Raclavsky, VSB – Technical University Ostrava, CZECH REPUBLIC
Combustion Reactivity of Char Derived from Solvent Extracted Coal, Sihyun Lee, Hokyung Choi, Sangdo Kim, Jeongwhan Lim, Youngjoon Rhim, Korea Institute of Energy Research (KIER); Woosik Park, Hanyang University, KOREA
Investigation of Effect of Reagents on the Coal Recovery from Coal Washing Plant Tailings by Floatation, Oktay Bayat, Huseyin Vapur, Cukurova University; Metin Ucurum, Nigde University, TURKEY Column Flotation of Fine and Coarse Coal using a Novel Approach, B.K. Parekh, D.P. Patil, University of Kentucky CAER; Edgar B. Klunder, NETL,USA Study of the Lignite Qualitative Parameters Modification, During its Storage, Sanda Krausz, Nicolae Cristea, University of Petrosani; Ion Bacalu, Mihail Dafinoiu, Daniel Burlan, National Society of Lignite, Oltenia, ROMANIA
Forced Flame Response Measurement in a Gas Turbine Combustor with High Hydrogen Fuel, Kyu Tae Kim, University of Cambridge, UNITED KINGDOM; Jong Guen Lee, Bryan D. Quay, Dom A. Santavicca, Pennsylvania State University, USA Development of Commercial CFBC Boiler for Refuse Derived Fuel, Dowon Shun, Dal-Hee Bae, Jaehyeon Park, Seung Yong Lee, Korea Institute of Energy Research, KOREA Simplified Quantification of Tetrafluoroborate Ion in Flue Gas Desulfurization Effluent for Management of Fluorine Emission, Seiichi Ohyama, Hiroyuki Masaki,
TECHNICAL PROGRAM Shinji Yasuike, Kazuo Sato, Central Research Institute of Electric Power Industry, JAPAN Enhancing Thermal Efficiencies in Steam Power Plants by Utilizing the “W2” Prime Mover as Auxillary Equipment, Jerry F. Willis, Admiral Air, Inc., USA Comparing Efficiencies of the Steam Turbine Versus the “W2” Prime Mover, Jerry F. Willis, Admiral Air, Inc., USA Optimization of Fuel Properties with Utilization of Biodegradable Municipal Wastes for Combustion Units, Dagmar Juchelkova, Martina Hajkova, Helena Raclavska, VSB – Technical University Ostrava; L. Tararik, Frydecka skladka, a.s., CZECH REPUBLIC Sulfur Retention in the Ash During Combustion of Tuncbilek Briquettes, Ayfer Parlak, Mustafa Ozdingis, H. Köksal Mucuk, Selahaddin Anac, Turkish Coal Enterprises (TKI); Bekir Zühtü Uysal, Gazi University, TURKEY Development of an Analytical Solution for Jet Diffusion Flame Equations, F. N. Pereira, A. L. de Bortoli, N. R. Marcilio, UFRGS – Universidade Federal do Rio Grande do Sul, BRAZIL
POSTER SESSION 2 GASIFICATION Suitability of a South-African High Ash Content and High Ash Flow Temperature Coal Source for Entrained Flow Gasification, JC van Dyk, SASOL Technology, SOUTH AFRICA; R Stemmer, Corus Technology, THE NETHERLANDS Continuous Experiments of Hot Gas Desulfurization Process Using Zn-Based Solid Sorbents in a Pressurized Condition, Sung-Ho Jo, Young Cheol Park, Ho-Jung Ryu, Chang-Keun Yi, Korea Institute of Energy Research; Jeom-In Baek, Korea Electric Power Research Institute, KOREA Biomass Gasification in Dual Fluidized Reactors: Process Modeling Approach, Thanh D. B. Nguyen, Young-Il Lim, Hankyong National University; Byung-Ho Song, Kunsan National University; Won Yang, Uendo Lee, Young-Tai Choi, Korean Institute of Industrial Technology; Jae-Hun Song, Gi-Chul Myoung, Yong-Soo Cho, SeenTec Co., Ltd., KOREA The Role of O2/COG Ratio on Non-Catalytic Partial Oxidation Process of Coke Oven Gas, Haizhu Cheng, Sufang Song, Yongfa Zhang, Taiyuan University of Technology, CHINA Numerical Simulation of Carbon Catalytic Reforming Reactor, Haizhu Cheng, Yongfa Zhang, Sufang Song, Taiyuan University of Technology, CHINA A Study on the Temperature Profile and Heat Transfer Coefficients in Underground Coal Gasification Cavities, Sateesh Daggupati, Ramesh Naidu Mandapati, Sanjay M Mahajani, Anuradda Ganesh, Preeti Aghalayam, IIT Bombay; Pal AK., Sharma R.K., UCG Group, IRS, ONGC, INDIA Some Results of UCG Ex-Situ Trials from HBP Company Point of View, Peter Cicmanec, Frantisek Verbich, Jaroslav Belacek, Hornonitrianske bane Prievidza, a.s.; Karol Kostur, Technical University of Kosice, SLOVAK REPUBLIC
13
POSTER SESSION 3 SUSTAINABILITY AND ENVIRONMENT
POSTER SESSION 5 COAL-DERIVED PRODUCTS
The Coagulation in Electric Field of the Argillaceous Suspensions from the Wastewater Resulted in Coal Processing, Romulus Sarbu, Diana Marchis, University of Petrosani; Adrian Corui, SC AQUATIM SA Timisoara, ROMANIA
Reactions of Coal Structures with Polymers Leading to Hydrogen Production, Pavel Straka, Institute of Rock Structure and Mechanics, v.v.i, Academy of Sciences of the Czech Republic, CZECH REPUBLIC
Thar Coal Mining Challenges, Farid A. Malik, EMRConsult, PAKISTAN Submerged Sequencing-Batch Membrane Bioreactor to Treat the Coke Wastewater, Wen-Ying Li, Jingwen Wu, Baojie Zhao, Jie Feng, Taiyuan University of Technology, CHINA
POSTER SESSION 4 CARBON MANAGEMENT Options of CO2 Capture in Oxyfuel Coal Combustion Technologies, C. Clemente-Jul, J. Rodrigo -Naharro, Universidad Politécnica de Madrid, SPAIN Installation and Operation of 0.5 MW-Scale Dry Sorbent CO2 Capture Pilot Plant Integrated with Real Coal-Fired Power Plant, Chang-Keun Yi, Sung-Ho Jo, Young Cheol Park, Korea Institute of Energy Research; Chong Kul Ryu, Korea Electric Power Research Institute, KOREA Reaction Characteristics of Water Gas Shift Catalysts for SEWGS Process in a Bubbling Fluidized Bed, SeungYong Lee, Ho-Jung Ryu, Dowon Shun, Dal-Hee Bae, Korea Institute of Energy Research, KOREA Carbon Dioxide Capture of Flue Gases from Coal-Fired Power Plant Using Enzymes Originated Marine Life, Sihyun Lee, Soonkwan Jeong, Kyungsoo Lim, Jeonghwan Lim, Mari Vinoba, Korea Institute of Energy Research (KIER); Daehoon Kim, Korea University, KOREA Preparation and Characteristics of Formed Active Carbons for Natural Gas Storage, Grzegorz Łabojko, Aleksander Sobolewski, Institute for Chemical Processing of Coal; Leszek Czepirski, AGH – University of Science and Technology, POLAND The Kinetics of the CO 2 Reforming of CH 4 over Carbonaceous Catalyst, Fengbo Guo, Yongfa Zhang, Guojie Zhang, Bingmo Zhang, Taiyuan University of Technology, CHINA
Catalytic Performance in Fixed-Bed and Bubbling Fluidized-Bed Reactor during Fischer-Tropsch Synthesis on the Iron-Based Catalysts, Jong Wook Bae, Ki-Won Jun, Yun-Jo Lee, Kyoung Su Ha, Korea Research Institute of Chemical Technology (KRICT), KOREA Operation of Slurry Reactor for Fischer-Tropsch Synthesis, Ho-Tae Lee, Jung-Il Yang, Jung Hoon Yang, Dong-Hyun Chun, Hak-Joo Kim, Heon Jung, Korea Institute of Energy Research, KOREA Process Simulation of Steam Hydrogasification to Produce F-T Products and Electricity, Xiaoming Lu, Chan S Park, Joseph M Norbeck, University of California, Riverside, USA Further development of the PSRK Model for the Prediction of the Vapor-Liquid-Equilibria of Direct Coal Liquefaction System at High Temperatures and High Pressures, Xuefeng Mao, Shidong Shi, Wenbo Li, Zhennan Gao, China Coal Research Institute, CHINA Arsenic and Mercury Removal by Using Iron Humate Prepared from Turkish Coal Based Humic Acid, Hacer Dogan, Murat Koral, Tulay Inan, TUBITAK Marmara Research Center; Selahattin Anaç, Zeki Olgun, TKI(Turkish Coal Enterprises), TURKEY Heavy Metal Adsorption of Turkish Coal Based Humic Acid/Epoxy Composites, Emel Yildiz, Hacer Dogan, Murat Koral, Tulay Inan, TUBITAK Marmara Research Center; Selahattin Anaç, Zeki Olgun, TKI(Turkish Coal Enterprises), TURKEY Numerical Simulation of Syngas Production by Partial Oxidation of Coke Oven Gas under Non-Premixed Condition, Honggang Chen, Kai Zhang, North China Electric Power University; Hui Zhao, China University of Petroleum; Yongfa Zhang, Taiyuan University of Technology, CHINA Coal Supply Agreements and Competition, Değer Boden Akalın, Boden Law Office, TURKEY Chemicals from Turkish Lignites, Vedat Mihladiz, Turkish Coal Enterprises, TURKEY
Pd-Free Composite Membrane for Pre-Combustion Capture, Jung Hoon Park, Sung Il Jeon, Korea Institute of Energy Research; Young Jong Choi, Innowill Corporation, KOREA
Methane Cracking over De-ashed Coal Chars and the Effect of the De-ashing Conditions, Yizhuo Han, Yisheng Tan, Jiantao Zhao, Hongjuan Xie, Institute of Coal Chemistry, Chinese Academy of Science; Ling Wei, Graduate School of the Chinese Academy of Sciences; Jinhu Wu, Qingdao Institute of Bioenergy and Bioprocessing Technology, Chinese Academy of Sciences, CHINA; Dongke Zhang, The University of Western Australia, AUSTRALIA
Composite Ceramic Membrane for Oxygen Separation, Jung Hoon Park, Soo Hwan Son, Korea Institute of Energy Research; Jong Pyo Kim, Chungnam National University, KOREA
CeO2-K2O Promoted Co-Mo Sulfur-Tolerant Shift Catalyst for the Shift Reaction of CO in Coke Oven Gas, Yuqiong Zhao, Yongfa Zhang, Guojie Zhang, Taiyuan University of Technology, CHNA
Fixed-Bed Adsorption of Carbon Dioxide-Nitrogen Mixtures onto Activated Carbon: Characteristics of CO2 Adsorption and Modeling, Regina F.P.M. Moreira, Tirzhá L.P. Dantas, Federal University of Santa Catarina; Francisco Murilo T. Luna, Ivanildo J. Silva Jr., Diana C. S. de Azevedo, Federal University of Ceará, BRAZIL; Carlos A. Grande, Alírio E. Rodrigues, University of Porto, PORTUGAL
Effects of Preparation Conditions on Ru/Al2O3 Catalyst for Coal-Based Syngas Methanation Reaction, Liping Wang, Yongfa Zhang, Yaling Sun, Xianglan Li, Taiyuan University of Technology, CHINA
A Study on the Absorption Characteristics of CO2 with a Vortex Tube Type Absorber, Keun-Hee Han, Woo-Jung Ryu, Jong-Ho Park, Won-Kil Choi, Jong-Sub Lee, ByoungMoo Min, Korea Institute of Energy Research, KOREA
The Production of Organic Fertilizers from Göynük, Ilgin and Elbistan Lignites with H2SO4 Oxidation, M.Çöteli, A.Güntürk, A.Yavuz, A.Köker, G.Yıldırım,
14
TECHNICAL PROGRAM
S.Atlıhan, General Directorate of Mineral Research and Exploration, TURKEY Modeling, Scaleup and Optimization of Slurry Bubble Column Reactors for Fischer-Tropsch Synthesis, Laurent Sehabiague, Badie I. Morsi, University of Pittsburgh, USA Biogasification of Soma Lignite (A Preliminary Study), Mustafa Baysal, Yuda Yürüm, Sabanci University; Sedat İnan, TÜBİTAK Marmara Research Centre, TURKEY
POSTER SESSION 6 COAL SCIENCE Modeling of Coal Drying in a Pneumatic Dryer, Sihyun Lee, Sangdo Kim, Kyoungsoo Lim, Soonkwan Jeong, Youngjoon Rhim, Korea Institute of Energy Research (KIER), KOREA Characterization of Chars Made of Solvent Extracted Coals, Sihyun Lee, Jiho Yoo, Hokyung Choi, Sangdo Kim, Jeongwhan Lim, Thiruppathi Raja, Korea Institute of Energy Research (KIER); Wantaek Jo, Yonsei University, KOREA Upgrading of Low Rank Coal by Hybrid Flash Dryer, Sihyun Lee, Sangdo Kim, Hokyung Choi, Kyoungsoo Lim, Sangyoung Lee, Korea Institute of Energy Research (KIER), KOREA Drying Kinetics of Low Rank Coal in Multi-Chamber Fluidized Bed, Jaehyeon Park, Dowon Shun, Dal-Hee Bae, Sihyun Lee, Jeong Hak Seo, Korea Institute of Energy Research; Jaehyeok Park, Hanyang University, KOREA The Influence of the Temperature on Adsorption of SDS on Coals, Boleslav Taraba, Roman Marsalek, Ostrava University, CZECH REPUBLIC Assessment of Elemental Sulphur in Biodesulphurized Coals, Lenia Gonsalvesh, Stefan Marinov, Maya Stefanova, Institute of Organic Chemistry, Bulgarian Academy of Sciences, BULGARIA; Robert Carleer, Jan Yperman, Hasselt University, BELGIUM Study of Biodesulphurized High Sulphur Coals from Bulgaria, Stefan Marinov, Maya Stefanova, Lenia Gonsalvesh, Nadezda Kazakova, Institute of Organic Chemistry, Bulgarian Academy of Sciences; Veneta Groudeva, M.Iliev, University of Sofia; Petyo Gadjanov, Technical University of Sofia, BULGARIA; Robert Carleer, Jan Yperman, Hasselt University, BELGIUM Samples in the World Coal Quality Inventory - A USGS Compilation on Global Coal, Susan J. Tewalt, Harvey E. Belkin, U.S. Geological Survey, USA Model Structure of High Sulphur N.E. Region Indian Coals, Sunil Kumar Srivastava, Atma Ram Singh, Central Institute of Mining and Fuel Research, INDIA Influence of the Surface Treatment with O3 And NH3 on the Physical and Chemical Characteristics of Dried Low Rank Coal, Gi Bo Han, Yongseung Yun, Changsik Choi, Institute of Advanced Engineering, KOREA Energy, Natural Gas, Türkiye & Ankara, İbrahim Halil Kirsan, Başkent Doğalgaz Dağıtım A.Ş., TURKEY An Alternative Application to the Centrifugal Dryer at a Coal Preparation Plant, Ahmet Gitmez, Mustafa Yılmaz, Western Lignite Establishment (GLI), TURKEY The Evaluation of the Contributions to the Productivity of the Process Changes at Tuncbilek Coal Preparation Plant, Ahmet Gitmez, F. Zehra Taksuk, Fatih Albayrak, Western Lignite Establishment (GLI), TURKEY
The Dump Truck Requirement Planning Studies of Turkish Coal Corporation, Mustafa Ziypak, Turkish Coal Corporation, TURKEY Investigation of Radioactive Contents Soma Coals, İsmail Demir, İlgin Kurşun, İstanbul University, TURKEY Effect of Triboelectrostatic Separation on Coal Desulfurization and Deashing, Byoung-Gon Kim, HoSeok Jeon, Sang-Ho Back, Chong-Lyuck Park, Korea Institute of Geoscience and Mineral Resources (KIGAM), KOREA Remove of Ash and Sulfur Minerals from Coal by Triboelectrostatic Separation, Ho-Seok Jeon, ByoungGon Kim, Korea Institute of Geoscience and Mineral Resources (KIGAM); Woo-Zin Choi, The University of Suwon, KOREA Recovery of Valuable Metallic and Non-Metallic Minerals from Coal Mine Wastes, Sang-Bae Kim, SooBok Jeong, Ho-Seok Jeon, Chong-Lyuck Park, Korea Institute of Geoscience and Mineral Resources (KIGAM), KOREA Manufacture of Fired Clay Brick from Coal-Preparation Refuse and its Characteristics, Soo-Bok Jeong, Ho-Seok Jeon, Chong-Lyuck Park, Byoung-Gon Kim, Korea Institute of Geoscience and Mineral Resources (KIGAM), KOREA
15
GENERAL INFORMATION / PROCEEDINGS
2010 Session Chairs ADVISORY BOARD
PROCEEDINGS Orders can be placed online. Credit cards, checks and money orders are accepted. Shipping: CD-ROM (Vol. 14 - 25): $5.00 Domestic and International Paper proceedings (Vol. 1 - 13): Domestic - $5.00 International - $25.00
Year Edition and Cost 1st Annual - 1984 2nd Annual - 1985 3rd Annual - 1986 4th Annual - 1987 5th Annual - 1988 6th Annual - 1989 7th Annual - 1990 8th Annual - 1991 9th Annual - 1992 10th Annual - 1993 11th Annual - 1994 12th Annual - 1995 13th Annual - 1996 14th Annual - 1997* 15th Annual - 1998* 16th Annual - 1999* 17th Annual - 2000* 18th Annual - 2001* 19th Annual - 2002* 20th Annual - 2003* 21st Annual - 2004* 22nd Annual - 2005* 23rd Annual - 2006* 24th Annual - 2007* 25th Annual - 2008* 26th Annual - 2009* * CD-ROM
$30.00 $30.00 $30.00 $30.00 $40.00 $40.00 $50.00 $50.00 $50.00 $50.00 $50.00 $50.00 $50.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $50.00
Robert Beck, Chair, The National Coal Council, Inc., USA Richard Winschel, Vice Chair, CONSOL Energy Inc., USA Richard Bajura, West Virginia University, USA Francois Botha, Illinois Clean Coal Institute, USA Francis Burke, CONSOL Energy Inc., USA Vann Bush, Gas Technology Institute, USA Tarunjit Butalia, The Ohio State University, USA Shiao-Hung Chiang, University of Pittsburgh, USA Dan Duellman, American Electric Power, USA Shannon Fraser, U.S. Dept. of Commerce, USA Evan Granite, U.S. DOE/NETL, USA Yizhou Han, Chinese Academy of Sciences, CHINA Gerald Holder, University of Pittsburgh, USA Jim Hower, University of Kentucky - CAER, USA Mike Jones, Univeristy of North Dakota EERC, USA Francis Lau, Synthesis Energy Systems, USA Ke Liu, National Institute of Clean & LowCarbon Energy, CHINA Chuck McConnell, Battelle, USA Robert Miller, Air Products & Chemicals, USA Kouichi Miura, Kyoto University, JAPAN Badie I. Morsi, University of Pittsburgh, USA Masakatsu Nomura, Osaka University, JAPAN Guven Onal, Istanbul Technical University, TURKEY Brenda Pierce, U.S. Geological Survey, USA Massood Ramezan, Leonardo Technologies, Inc. (LTI), USA Leslie Ruppert, U.S. Geological Survey, USA Richard Sakurovs, CSIRO, AUSTRALIA Alan Scaroni, Pennsylvania State University, USA Alan Singleton, Energy Technology Partners LLC, USA Chunshan Song, Pennsylvania State University, USA Gary Stiegel, U.S. DOE/ NETL, USA Johan van Dyk, SASOL, SOUTH AFRICA Frans Waanders, North-West University, SOUTH AFRICA Ting Wang, University of New Orleans, USA John Wheeldon, EPRI, USA
Combustion
John Wheeldon, National Carbon Capture Center, USA Evan Granite, U.S. DOE/NETL, USA Steven Carpenter, Marshall Miller & Associates, USA
Gasification
Gary Stiegel, U.S. DOE/NETL, USA Jenny Tennant, U.S. DOE/NETL, USA Johan van Dyk, SASOL, SOUTH AFRICA Ke Liu, NICE, CHINA Rohan Courtney, UCG Partnership, UNITED KINGDOM Ting Wang, University of New Orleans, USA
Sustainability and Environment
Jim Hower, University of Kentucky, USA Brenda Pierce, U.S. Geological Survey, USA
Carbon Management
Bob Miller, Air Products, USA Leslie Ruppert, U.S. Geological Survey, USA Richard Sakurovs, CSIRO, AUSTRALIA
Coal-Derived Products
Rachid Oukaci, Energy Technology Partners, LLC, USA Belma Demirel, Rentech, Inc., USA Dragomir Bukur, Texas A&M University at Qatar, QATAR
Coal Science
Jim Hower, University of Kentucky, USA B.K. Parekh, University of Kentucky, USA Frans Waanders, North-West University, SOUTH AFRICA Susan J. Tewalt, U.S. Geological Survey, USA Quentin Campbell, North-West University, SOUTH AFRICA Richard Winschel, CONSOL Energy Inc., USA Ashok K. Singh, Central Institute of Mining & Fuel Research, CSIR, INDIA
INTERNATIONAL VICE CHAIRS Duke Du Plessis, Alberta Energy Research Institute, CANADA Juan Jose Garcia, Vice Minster Energy, VENEZUELA Yizhou Han, Chinese Academy of Sciences, CHINA Hung-Taek Kim, Ajou University, SOUTH KOREA Evgeny Kuzmin, Moscow State Mining University, RUSSIA Bernd Meyer, Freiberg University Mining & Technology, GERMANY Kouichi Miura, Kyoto University, JAPAN Geoff Morrison, IEA Clean Coal Centre, UNITED KINGDOM Selahaddin Anac, Turkish Coal Enterprise (TKI), TURKEY Richard Sakurovs, CSIRO, AUSTRALIA Harry Schreurs, NOVEM, THE NETHERLANDS Marek Sciazko, Institute of Chemical Processing of Coal, POLAND Sunil Srivastava, Central Fuel Research Institute, INDIA Johan van Dyk, SASOL Technology, SOUTH AFRICA
ANNOUNCING:
TWENTY - EIGHTH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE DAVID L. LAWRENCE CONVENTION CENTER PITTSBURGH, PA, USA SEPTEMBER 12 - 15, 2011
Abstracts must be submitted by March 1, 2011. Please forward paper title, intended topic area, authors, affiliations, contact information with valid email address and a one-page abstract to: Conference Secretary
[email protected] Please visit the PCC WEBSITE: www.engr.pitt.edu/pcc
DEVELOPMENT OF OXYCOALTM TECHNOLOGY RESULTING FROM TESTING CONDUCTED AT DOOSAN POWER SYSTEMS’ CLEAN COMBUSTION TEST FACILITY (CCTF)
F D Fitzgerald and P Holland-Lloyd Doosan Babcock United Kingdom
Abstract Oxyfuel combustion technology is one of several Carbon Abatement Technologies (CATs) currently being developed. The technology offers a means of generating carbon dioxide rich flue gas requiring minimal treatment prior to sequestration or beneficial application. Doosan Power Systems are aiming to develop a competitive oxyfuel firing technology suitable for full plant application post 2015, and is taking a phased approach to the development and demonstration of oxyfuel technology. Doosan Power Systems is leading a number of collaborative projects that are investigating ‘Oxyfuel Combustion Fundamentals and Underpinning Technologies’, the ‘Demonstration of an Oxyfuel Combustion system’, the ‘Modelling and Testing of the 40 MWt OxyCoal™ Burner’, and the ‘Optimisation of Oxyfuel PF Power Plant for Transient Behaviour’. This paper outlines progress on the ‘Demonstration of an Oxyfuel Combustion system’ project, which is demonstrating an oxyfuel combustion system of a type and size (40MWt) applicable to new build and retrofit advanced supercritical boiler plant. Installation and commissioning are complete and testing is in progress. Preliminary results have shown safe and smooth transitions between air firing and oxyfuel operation, with economiser outlet CO2 concentrations greater than 85% v/v dry being achieved. Background To achieve the global target reduction in CO2 emissions of some 60%, compared to 1990 levels, by 2050 Carbon Capture and Storage (CCS) will be necessary. Carbon Abatement Technologies (CATs) will be required for retrofit to existing power plant and installation on new build power plant. Doosan Power Systems have taken a proactive approach to developing CATs, leading and supporting techno-economic studies on both Track 1 and Track 2 approaches of the UK DTI’s Carbon Abatement Technology Strategy
(1)
. The Track 1 approaches are available now and
reduce CO2 emissions per unit of electricity generated by means of improved cycle efficiency and biomass co-firing (i.e. carbon neutral generation). The Track 2 approach, CCS, achieves much larger reductions, of up to 95%, by means of oxyfuel combustion, Post Combustion Capture (PCC) or Pre-Combustion capture via Integrated Gasification Combined Cycle (IGCC) technologies. Doosan Power Systems and Doosan Heavy Industries are active in all three CCS approaches.
Oxyfuel Combustion The oxyfuel process is based on excluding the inert components (mainly nitrogen) of air from the combustion process. In oxyfuel combustion, nitrogen is largely absent from the flue gas, since the fuel is combusted with a mixture of nearly pure oxygen (~95% O2 - separated from air in an air separation unit {ASU}) and CO2 rich recycled flue gas. It is recognised as a leading Carbon Capture technology for new and retrofit power plant. Figure 1 presents the comparison between conventional air and oxyfuel combustion on a utility boiler.
© Doosan Power Systems 2010 Figure 1: Schematic Diagram of Oxyfuel versus Air Firing The combustion temperature of a fuel with high purity oxygen is too high (3500°C) for any conventional combustion process used in utilities, therefore the temperature is controlled by recycling a portion of the flue gas to the furnace. The impact of varying flue gas recycle (FGR) rate on oxyfuel combustion efficiency, boiler thermal performance and pollutant formation is of major interest to manufacturers and utility operators alike. Demonstration of an Oxyfuel Combustion System Doosan Power Systems is leading a number of UK Government supported collaborative projects that are developing oxyfuel combustion technology, including the £7.4M UK Department of Energy and Climate Change (DECC) Hydrogen Fuel Cells and Carbon Abatement Technologies (HFCCAT) Demonstration Programme collaborative project: Demonstration of an Oxyfuel Combustion System (OxyCoal 2), which is demonstrating the Doosan Power Systems’ 40MWt OxyCoalTM burner. The project partners comprise: Doosan Power Systems (Lead), Imperial College London and the University of Nottingham.
Scottish and Southern Energy plc, Air Products plc,
ScottishPower Limited, E.ON UK plc, EDF Energy plc, Drax Power Limited, DONG Energy
Power, UK Coal plc and Vattenfall AB are sponsor participants, with Scottish and Southern Energy plc acting as prime sponsor. The phases of the project comprise development of a purpose designed oxyfuel demonstration facility (Doosan Power Systems’ 90MWt Clean Combustion Test Facility (CCTF)) including detailed design and HAZOP study, installation and commissioning of the oxyfuel equipment and design, manufacture and parametric testing of the 40MWt OxyCoal™ burner.
The
specific test objectives are demonstration of safe operation and successful performance of the full-scale 40MWt OxyCoal™ burner firing at conditions pertinent to the application of an oxyfuel combustion process in a utility power plant, in terms of flame stability, NOX, flame shape and heat transfer characteristics over a reasonable operational envelope with respect to start-up, turndown, shutdown and the transition between air and oxyfuel firing. Clean Combustion Test Facility The 90MWt Clean Combustion Test Facility (CCTF) is designed primarily for the development of burners for fossil fuel firing applications and is one of the largest and most modern single burner test rigs in the world. The CCTF was commissioned in 2000 and has been designed to enable burners to be developed, optimised and performance tested at fullscale prior to application in industrial plant. The front of the CCTF furnace is shown in Figure 2.
© Doosan Power Systems 2010 Figure 2: Clean Combustion Test Facility
The main component is a horizontal, water-jacketed test furnace that is partly lined with high temperature refractory. The furnace is 17 m long and 5.5 m square in section. Observation ports are arranged along one side wall of the furnace on the burner centerline and flame probing access ports, are located on the opposite side wall.
Photographic and video
equipment are fitted for flame observation, monitoring and image recording. An adaptable windbox is fitted at the front of the furnace to accommodate single test burners with throat diameters up to a maximum of 2 m. The facility is able to fire a wide variety of fuels; coals (ranging from low volatile semianthracites – 8% volatile matter to high volatile bituminous coals with 50% volatile matter, up to 35% ash, and up to 20% inherent moisture), heavy fuel oil and natural gas. A storage silo, loss-in-weight feeder and pneumatic transport system is fitted for the supply of pulverised coal to the burner. The system has a maximum feed rate capability of 12 to 14 tonnes per hour depending on the bulk density of the material. The draught plant consists of forced draught (FD), transport air, primary air (PA), core air and induced draught (ID) fans and blowers. Gas-fired airheaters are installed to raise fuel transport stream and combustion air temperatures to plant representative values. A multi-cyclone grit collection system is fitted for cleaning of flue gases prior to the ID fan. A tri-drum boiler with superheater and economiser is installed to cool the flue gases from approximately 1200°C at the furnace exit to approximately 230°C at the grit collector inlet. Two-stage combustion has been implemented on the CCTF by means of installing overfire air ports to the furnace sections to allow the testing of burners operating under staged combustion conditions. OxyCoalTM Clean Combustion Test Facility To allow testing of an OxyCoalTM burner in oxyfuel firing mode the CCTF has been upgraded with the addition of equipment and instrumentation required for oxyfuel firing.
The
conversion to oxyfuel included the addition of an oxygen storage facility comprising three liquid oxygen (LOX) storage tanks each with a capacity of approximately 50 tonnes, and eight ambient vaporisers to supply gaseous oxygen for injection into the primary and secondary flue gas recycle (FGR) streams. The primary and secondary FGR streams replace the primary air and main combustion air respectively, each having a dedicated fan.
A
transport FGR stream replaces the transport air stream. The transport and primary FGR streams have additional flue gas cooling systems fitted to condense moisture, followed by in-
duct heating as a means to mitigate against PF feeding problems in a high moisture flue gas. A proportion of the secondary FGR stream can be redirected to an overfire FGR system for two-stage combustion.
A schematic diagram of the CCTF, including the oxyfuel
configuration is presented in Figure 3.
© Doosan Power Systems 2010 Figure 3: Schematic Diagram of CCTF OxyCoal™ Configuration The construction and installation of the oxyfuel equipment have been completed, with OxyCoalTM burner testing being carried out during late 2009 and early 2010. OxyCoal™ Burner The first generation of oxyfuel burners will most likely be based on current low NOX air-fired burner technology in order to ensure compatibility with existing plant for retrofit purposes. Additionally, a uniform ‘simulated air’ flue gas composition and a design flue gas recycle rate, based on the consideration of radiant and convective heat transfer being theoretically similar to air firing, is a logical first operating condition. With these points in mind, a 40 MWt OxyCoalTM burner was designed to best exploit a range of potential operating conditions for both oxyfuel and air firing. For oxyfuel operation the volumetric flow rate and molar oxygen content of the primary gas is maintained as per air firing. The design overall stoichiometric ratio is 1.2 and the flue gas recycle rate has been chosen to give an adiabatic flame temperature equivalent to air operation.
Burner Modelling Prior to testing, computational fluid dynamics (CFD) modelling was undertaken to support the burner design process. The CFD model of the 40 MWt OxyCoal™ burner extends from the windbox and primary gas inlets to the outlet of the CCTF furnace, and consists of 3.6 million cells. An image of the modelled geometry is shown in Figure 4. Two design operating conditions were considered for this burner arrangement: 1. Air Firing, Existing Plant. 2. Oxyfuel Firing.
© Doosan Power Systems 2010 Figure 4: Modelled Geometry of 40 MWt OxyCoal™ Burner in CCTF Figure 5 shows the predicted gas temperature profiles in the furnace. Although Furnace Exit Gas Temperature (FEGT) is predicted to be broadly similar under oxyfuel and air firing combustion, the peak temperature is predicted to be higher under the particular oxyfuel firing conditions modelled. Despite the flame being well rooted to the flame holder, it is seen that oxyfuel firing is predicted to lead to a narrower flame. This may be due to the reduced momentum resulting from the lower volumetric flow rate of FGR when operating in oxyfuel mode, arising from the higher density of CO2 compared with N2 (1.8 kg/m3 versus 1.14 kg/m3 at 25°C).
Air Firing
Oxyfuel Firing
© Doosan Power Systems 2010 Figure 5: Predicted Gas Temperature (°C) under Air and Oxyfuel Operation Figure 6 presents the predicted CO concentrations under air and oxyfuel operation. It is seen that local CO concentrations within the flame are also predicted to be much higher under oxyfuel firing, which agrees with the findings of a number of experimental studies (2, 3).
Air Firing
Oxyfuel Firing
© Doosan Power Systems 2010 Figure 6: Predicted CO Concentration (ppm) under Air and Oxyfuel Operation
OxyCoal™ Burner Commissioning and Testing The OxyCoalTM burner is mounted in a windbox which supplies either combustion air or secondary FGR to the burner outer annuli.
Combustion air enters the windbox
perpendicularly from the right side wall, looking at the front of the windbox/burner back plate, and the FGR enters perpendicularly from the opposing wall (Figure 3). Before any testing of the burner could commence, the existing and new equipment related to the oxyfuel modifications required cold and hot commissioning.
As part of the cold
commissioning activities, the OxyCoalTM burner underwent isothermal testing to determine swirl numbers and pressure loss factors of the burner outer annuli. The hot commissioning was split into three parts. The first part successfully proved that the test facility was still capable of operating under air firing conditions and provided baseline air firing OxyCoalTM burner, furnace and boiler performance results. Figure 7 shows a typical OxyCoalTM burner air firing flame. The flame is anchored to the burner, within the quarl, and has a bright root.
© Doosan Power Systems 2010 Figure 7: Typical Coal Air firing Flame on OxyCoalTM Burner The second part successfully proved the test facility in oxyfuel oil firing operation. The safe transition from oil firing on air to oxyfuel and establishment of a stable oxyfuel oil flame were demonstrated. The third and final part successfully proved the test facility under oxyfuel operation. The safe transition from coal firing on air to oxyfuel and establishment of a stable oxyfuel flame were demonstrated. Figure 8 shows a typical OxyCoalTM burner oxyfuel flame. The oxyfuel flame
is visually practically indistinguishable from the flame obtained during air firing, again exhibiting a bright root that is strongly anchored to the burner.
© Doosan Power Systems 2010 Figure 8: Typical Coal Oxyfuel Firing Flame on OxyCoalTM Burner The testing of the OxyCoalTM burner was split into two phases. Phase 1 concentrated on ‘burner proving’ in terms of burner characterisation, flame stability and control/operability, while Phase 2 aimed to characterise the achievable performance of the burner. Phase 1 of testing was successfully completed in January 2010 with the following main outcomes: Three air to oxyfuel transition methods were investigated: 1. Transition from Oxyfuel Oil to Oxyfuel Coal Combustion This is the preferred transition method on the CCTF, since it maximises the test time on coal. Oil air firing is switched to oxyfuel oil firing: secondary air is switched to SFGR followed by the introduction of coal with primary air. The coal flow rate is increased, with corresponding reductions in oil flow until coal is at full load and oil support is off. Primary air is then switched to PFGR. This method of transition on the CCTF is reasonably quick to complete and is tolerant to fairly large fluctuations in fuel flow, without affecting the flame stability. 2. Transition from air, starting with SFGR stream, to Oxyfuel This approach starts with coal air firing (primary and secondary air), with a low level of oil support at reduced load. Secondary air is switched to SFGR, followed by increases in coal flow, with corresponding reductions in oil flow until coal is at full load and oil support is off. Primary air is then switched to PFGR. A longer period is required to complete the transition using this method, since small changes in coal flow
during the coal load increments have a significant impact on economiser outlet oxygen and therefore SFGR oxygen concentration. As a result, smaller and slower increments in coal load are required, which consume more test coal and, therefore, reduce the coal test time available. 3. Transition from air, starting with PFGR stream, to Oxyfuel This approach also starts with reduced load coal air firing (primary and secondary air) with a low level of oil support. Primary air is switched to PFGR. Coal flow is increased, with corresponding reductions in oil flow until coal is at full load and oil support is off. The secondary air is then switched to SFGR. This method proved to be the most difficult to perform, since even the smallest possible coal load increments had an appreciable impact on PFGR oxygen concentration.
Flame stability was
adversely affected, since PFGR is supplied at the flame root and provides oxygen at the initial combustion stage. Although a preferred method of CCTF air to oxyfuel transition has been selected, to maximise coal oxyfuel test time, all three of the above methods have been safely demonstrated and are perceived to be applicable to oxyfuel boiler plant. The sensitivity of Methods 2 and 3 to small changes in coal flow rate on the CCTF’s single burner is likely to be less of a concern in multi-burner plant. The importance of air ingress has been investigated. Initial testing, with substantial air ingress, resulted in a CO2 concentration of around 50% v/v dry at full load OxyCoalTM operation. Air ingress minimisation resulted in CO2 concentrations of around 85% v/v dry at full load OxyCoalTM operation. At the time of writing, Phase 2 of testing is in progress and is providing positive results. The remaining test programme aims to establish an operational envelope for the OxyCoal™ burner design, investigating the effects of parameters including FGR rate and stoichiometric ratio on flame stability, heat release and flue gas composition. Conclusions •
Modelling of a full scale OxyCoal™ burner in air and oxyfuel firing modes shows acceptable flame characteristics and emissions performance for both conditions.
•
The 90 MWt CCTF oxyfuel conversion has been successfully completed, all equipment has been commissioned and testing is in progress.
•
Safe and smooth transitions between air firing and oxyfuel operation have been demonstrated.
•
Economiser outlet CO2 concentrations greater than 85% v/v dry are achievable during oxyfuel operation.
•
Successful demonstration of the OxyCoalTM burner will form the foundation for the development of an OxyCoalTM boiler reference design.
References [1]
A Strategy for Developing Carbon Abatement Technolgies for Fossil Fuel Use, Department for Trade & Industry, DTI/Pub URN 05/844, June 2005
[2]
Goh, B, ‘E.ON-UK’s Pilot-Scale Oxy-Fuel Combustion Experience: Development, Testing and Modelling’, IEA GHG International Oxy-Combustion Network, Third Workshop, Yokohama, Japan, March 2008
[3]
Hjärtstam, S et al., ‘Combustion Characteristics of Lignite-fired Oxy-fuel Flames’, presented at the 32nd International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, Florida, USA, 2007
Acknowledgements The authors gratefully acknowledge the grant funding provided by the UK Department of Energy and Climate Change (DECC) and the technical and financial contributions made by the project collaborators: University of Nottingham, Imperial College London, Scottish and Southern Energy plc (Prime Sponsor), Air Products plc, DONG Energy Power, Drax Power Limited, EDF Energy plc, EON UK plc, ScottishPower Limited, UK Coal plc and Vattenfall AB. The authors also acknowledge the support of the members of Doosan Power Systems’ oxyfuel research and development team.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
PROGRAM TOPIC: COMBUSTION NOX REBURNING IN OXY-FUEL COMBUSTION AN EXPERIMENTAL INVESTIGATION
Daniel Kühnemuth, Fredrik Normann, Klas Andersson, Filip Johnsson, Bo Leckner Division of Energy Technology, Department of Energy and Environment, Chalmers University of Technology, SE – 412 96 Göteborg, Sweden This work investigates the reburning reduction of nitric oxide (NO) in a 100 kW propane-fired oxyfuel flame. The conducted experiments include a comprehensive parameter study: NO was injected into the recycled flue-gas, the inlet oxygen concentration was varied between 25 and 37 vol. % and the stoichiometric ratios at the burner inlet ranged from 0.7 and 1.15. The respective influence of inlet oxygen concentration and burner stoichiometry on once-through and total reduction of NO was measured. Furthermore, concentration and temperature in the furnace were mapped to identify important differences between oxy and air-fired conditions. The furnace measurements show that the peak concentration of carbon monoxide may be more than twice as high as in air-fired conditions. The formation paths of CO and its influence on the NOx chemistry are therefore discussed. The results of the parameter study show that reburning is favored by decreased burner stoichiometry. The effect of inlet oxygen concentration on once-through NO reduction is of minor importance. Changes in stoichiometry and oxygen inlet concentration are associated with changes in recycle ratio. The influence of the recycle ratio on the NO reduction is of great importance and is investigated as separate parameter. Keywords: Reburning, Oxy-fuel, NOx, CO2/O2.
Introduction Oxy-fuel combustion is a promising near-term technology for CO2 capture in power plants. Presently, this technology is in a stage where the first pilot-scale plants (tens of MW) are in operation and demo-scale plants (hundreds of MW) are being planned for commissioning (Toftegaard et al. 2010). Also in oxy-fuel power plants combustion of coal will be associated with
1
NOx formation, and the presence of NOx may have to be considered both in the gas entering the flue gas treatment and in the gas emitted to the atmosphere, as well as in the storage gas (Normann et al. 2009). Thus, it is important to examine the possibilities for NOx removal. The changed conditions within the oxy-fuel process compared to air-combustion present new opportunities for NOx control, both by methods applied in the furnace (e.g. reburning of recycled NOx (Andersson et al. 2008, Mackrory 2008, Okazaki and Ando 1997) and high-temperature NOx reduction (Normann et al. 2008)), as well as during the flue-gas treatment (e.g. absorption of NO2 in water (White et al. 2010, Kühnemuth et al. 2008)). Further work is needed to identify strategies to be applied in commercialscale oxy-fuel plants.
In oxy-fuel combustion, pure oxygen is added to a stream of flue gas recirculated to the furnace to control the combustion temperature (among other parameters). NO, which is also contained in the flue gases is recirculated to the combustion zone and a reburning situation is inherent without any special arrangements. The present work investigates the reduction of recycled NO by reburning reactions in oxy-fuel combustion. The reburning mechanism is a route for reduction of NOx by hydrocarbon radicals, whose global reaction can be expressed as, 𝑁𝑂 + 𝐶𝐻𝑖 ↔ HCN + ⋯
(1)
Reduction of NO by reburning has been extensively investigated for air-fired conditions (see e.g. Dagaut and Lecomte 2003, Frassoldati, Faravelli and Ranzi 2003, Bilbao, Alzueta and Millera 1995) with the general conclusion that fuel-mixing, temperature, stoichiometry, and gas residence-time need to be optimized to successfully apply reburning. Two main factors change the conditions for reburning in oxy-fuel combustion: the recycle of NO-containing flue gas, which increases the total residence time of NO in the flame zone, and the changed combustion environment, which affects combustion and nitrogen chemistry.
Previous experimental work has shown that oxy-coal combustion reduces the emission of NOx with around 70%(in mg/MJ of fuel supplied), compared to a corresponding air-fired unit (Andersson et al. 2008). The main part of the reduction achieved is commonly considered to be caused by reburning of NO, recycled with the flue gas to the flame (Andersson et al. 2008, Mackrory 2008, Okazaki and Ando 1997). The most obvious indication of the change in combustion chemistry in oxy-fuel combustion is the elevated CO concentration in the flame, which is reported in several experimental studies (Hjärtstam et al. 2009, Dhungel 2009, Tan et al. 2006, Liu, Zailani and Gibbs 2005). However, literature disagrees on the reasons to the increased CO level and its importance to nitrogen 2
chemistry. Three reaction routes for CO formation have been discussed. Firstly, for coal-fired experiments, char gasification by CO2, Reaction (2), is commonly mentioned (Zhang et al. 2010, Li et al. 2009, Toporov et al. 2008), 𝐶 + 𝐶𝑂2 ↔ 2CO + 𝑂
(2)
However, there is also work (Mackrory 2008) indicating this reaction to be of minor importance for CO formation. Instead, Mackrory suggests that CO is formed by thermal dissociation of CO2 at temperatures above 1500 K, Reaction 3, 𝐶𝑂2 ↔ CO + 𝑂
(3)
However, thermal dissociation of CO2 is commonly claimed to be of importance at much higher temperatures (the conversion of CO2 is around 1% at 2000K) (Jin et al. 2006, Fan et al. 2003, Itoh et al. 1993, Nigara 1986), but it is concluded that the conversion can be significantly improved by removal of the final reaction products CO and O2.
Another formation route of CO is the water-gas-shift reaction, Reaction (4), as e.g. suggested by Dhungel et al. (Dhungel 2009) 𝐶𝑂 + H2 O ↔ CO2 + 𝐻2
(4)
Reaction (5), which is an elementary reaction included in the mechanism of the water-gas shift reaction, is shown to be reversed to produce CO, when CO2 is added to an air-fired flame (Liu et al. 2001). For the high CO2 concentrations in oxy-fuel combustion, similar conclusions have been made, e.g. (Glarborg and Bentzen 2008) CO + 𝑂𝐻 ↔ 𝐶𝑂2 + H
(5)
Two possible routes have been proposed for the impact of elevated CO concentration in oxy-fuel combustion on the nitrogen chemistry: an increased direct catalytic effect on the heterogeneous reaction of NO and char (Dhungel 2009), as earlier observed for air combustion (Levy et al. 1981), and an indirect effect on the radical pool involved in the nitrogen chemistry. The latter effect is explained by the impact of formation as well as oxidation of CO by Reaction (5) on the critical chain branching Reaction (6) (Mendiara and Glarborg 2009b, Normann et al. 2010), 𝐻 + O2 → OH + O
(6) 3
the resulting change in the radical pool under oxy-fuel conditions leads to an altered oxidation of hydrocarbon radicals (Normann et al. 2010), which are the most important reburning agents according to Reaction (1). Furthermore, changes in the radical pool may have an important influence on the oxidation of intermediate nitrogen species in the flame. For instance, Mendiara and Glarborg (Mendiara and Glarborg 2009a) mention that the oxidation behavior of NH3, is significantly influenced by the OH/H ratio, and Giménez-López et al. (Giménez-López et al. 2010) make similar conclusions regarding the oxidation of HCN. Thus, the elevated in-flame concentration of CO may have several in-direct effects on the nitrogen chemistry in oxy-fuel combustion; in this work the homogenous reactions are captured while heterogeneous effects are omitted. The practical implication of the changed combustion chemistry for the reduction of recycled NO during oxy-fuel combustion has been investigated recently. Commonly, experiments are conducted with coal, which contains fuel-bound nitrogen that contributes to the formation of NO and, hence, do not permit separating reburning from NO formation. In such experiments the emission of NO usually decreases with increasing oxygen concentration in the oxidizer (Liu and Okazaki 2003, Croiset and Thambimuthu 2001, Hu et al. 2000). Reduction of the stoichiometric ratio at the burner inlet is reported to be an effective measure for controlling NOx emission also in oxy-fuel combustion (Mackrory 2008), (Liu et al. 2005). It is important to note that both inlet oxygen concentration and burner stoichiometry affect combustion temperature, and that the NOx emission usually increases with temperature (Normann et al. 2010, Hu et al. 2000). Mendiara and Glarborg performed benchscale experiments focused on reburning of NO during oxy-methane combustion, thus excluding the effects of fuel-bound nitrogen. Their once-through experiments showed that oxy-fuel combustion has higher NO-reduction efficiency under fuel-lean conditions than air firing, but that the difference is negligible under fuel-rich conditions. The aim of the present investigation is to complement the present knowledge on reburning in oxyfuel combustion. A parameter study is performed to analyze the sensitivity of reburning reduction to changes in stoichiometric ratio of the burner, inlet oxygen concentration and recycle ratio. All results are compared to air-fired conditions.
Experiments The experiments were performed in a 100 kWth oxy-fuel combustion test unit, shown in Figure 1. The unit can operate in air or oxy-fuel combustion mode, the latter employing dry flue-gas recycling. The furnace is a cylindrical, top-fired, combustion chamber measuring 800 mm in inner diameter and 2400 mm in height. The swirl burner is equipped with two oxidant registers: a primary register
4
with guide vanes of an angle of 45° and a secondary oxidizer-register with 15° angle. A bluff body stabilizes the flame.
Figure 1. Schematic of the 100 kWth oxy-fuel test facility, including the ports for sampling (R2-R5; L1-L3) and NO injection. Measurement Ports R2 – R5 are located at 215, 384, 553 and 800 mm from the burner. Gases were extracted from the flame zone by a water-cooled gas-sampling probe presented in Figure 2. The gas-sampling probe consists of a heated inner tube with controlled temperature to prevent condensation inside the probe and an outer protecting water-cooled jacket with a diameter of 45 mm. The gas is sucked through a 4 mm thick ceramic filter at the tip of the probe. The extracted gases are transported through hot sampling lines (200°C) to gas analysis. Cold (ambient temperature) measurements of NO, O2, CO2, CO were performed with on-line gas analyzers, specified in Table 1. Gas temperature in the furnace was measured with a water-cooled suction pyrometer with a Type-B thermocouple, shielded by a ceramic tube to prevent radiative heat loss from the thermocouple junction, Figure 3.
5
Figure 2. Gas sampling probe.
1900 mm Cooling water outlet
A Thermocouple Type B: D=3mm
Suction inlet flue gas: D=6mm
200mm Cooling water inlet
A
Section A-A
Ø45,0/41,0mm Ø30,0/26,0mm Ø17,2/13,0mm 10,0x1,0 mm Ø13,0mm
Figure 3. Suction Pyrometer
Table 1. Gas analyzers Manufacturer/model NO O2 CO2 CO
Measurement principle Chemoluminescence
Range
Detection limit
Span gas
0 - 0.1 %
≤ 1 %*
417 ppm
Fisher Rosemount (NGA 2000)
Paramagnetic
0 - 25 %
≤ 1 %*
9.00 %
Binos (100 2M)
Paramagnetic
0 - 50 %
≤ 1 %*
9.00 %
Sick Maihak (Sidor)
NDIR
0 - 20 %
≤ 1 %*
18.00 %
Binos (100 2M) Sick Maihak (S 710)
TC NDIR
0 - 100 %
≤ 2 %*
90.00 %
0 - 1.0 %
≤ 1 %*
0.90 %
Sick Maihak (S 710)
NDIR
0 - 20 %
≤ 1 %*
9.00 %
ECO (CLD 700 EL)
* Percent of the full range
Propane was used as fuel in all experiments (Table 2), to exclude heterogeneous effects. The fuel input was 81 kW. Two stoichiometric ratios are defined: the stoichiometric ratio of the burner (λburner), and the global stoichiometric ratio (λglobal). The global stoichiometric ratio was kept at 1.15 6
in all tests by injecting pure oxygen for burn-out (comparable to over-fire air) in Port R5 (Figure 1), 800 mm downstream of the burner inlet. Thereby, a region with adjustable stoichiometry is created between the burner and Port R5, while the burnout of the fuel was secured, and accumulation of CO and hydrocarbons in the recycle loop was avoided. Table 2. Fuel composition Fuel analysis Propane
98.0
mol %
Butane
1.1
mol %
Ethane
0.7
mol %
Pentane
<0.1
mol %
Olefins
<0.1
mol %
Methane
<0.01
mol %
Lower heating value
46.4
MJ/kg
In oxy-fuel operation, oxygen (99.5 vol. % purity) was premixed with the recycled flue gases. The fraction of flue gases recycled (RFGR) to dilute the injected oxygen to its desired concentration in the feed gas is defined by the oxy-fuel case (OF case; e.g. a case with an oxidizer having 30 vol. % O2 is denoted OF 30). The RFGR was adjusted by two frequency-controlled fans, one situated in the recycle loop and one before the stack (Figure 1). The recycling ratio decreases with increasing oxygen concentration as well as with decreasing burner stoichiometry. NO (99.0 vol. % purity) was supplied to the oxidizer upstream of the O2 injection port (Figure 1) to maintain a constant inlet concentration of 500 ppm in the oxidizer (air as well as O2+RFG). The composition of the feed-gas stream in the two burner registers is identical. The flow through each register is adjustable and measured by differential pressure flow-meters. Table 3 presents the test cases, including the range of oxygen concentration in the feed gas between 25 and 37 vol. % (on dry basis). The lower concentration (25 vol. %) is the low limit before flame stability problems appear. The upper concentration of 37 vol. % is the limit above which extensive soot formation takes place. The stoichiometric ratio of the burner (λburner) was varied between 0.7 and 1.15. In the air-fired and OF 30 cases with λburner = 1.15, gas composition and temperature were measured through the measurement ports R2 to R5 (Figure 1) in which radial profiles in the furnace (0, 25, 50, 75, 100, 150, 200, 250 and 350 mm from the centerline) were sampled. Table 3: Test cases. The NO inlet concentration
7
was 500 ppm and λglobal=1.15 in all cases. L denotes global measurements in the recycle loop, and R detailed profile measurements in the reactor. λburner Air OF 25 OF 27 OF 30 OF 32 OF 35 OF 37
1.15 R, L L L R, L L L L
1.00 L
0.90 L
L
L L
0.80 L
0.70 L
L
L
Two NO reduction efficiencies are illustrated in Figure 4: once-through (ηonce) and total (ηtotal) reduction. Once-through reduction is measured over the furnace and concerns the molar flows of NO at the furnace inlet 𝑛̇ (𝑁𝑂)𝑓𝑖 and outlet 𝑛̇ (𝑁𝑂)𝑓𝑜 , 𝜂𝑜𝑛𝑐𝑒 = 1 −
𝑛̇ (𝑁𝑂)𝑓𝑜 𝑛̇ (𝑁𝑂)𝑓𝑖
(7)
The total reduction efficiency includes the influence from the flue-gas recycling on the total residence time of NO before emitted to the stack and concerns the molar flow of NO added to the oxidizer 𝑛̇ (𝑁𝑂)𝑖𝑛 and the molar flow emitted through the stack 𝑛̇ (𝑁𝑂)𝑠𝑡 , 𝜂𝑡𝑜𝑡𝑎𝑙 = 1 −
𝑛̇ (𝑁𝑂)𝑠𝑡 𝑛̇ (𝑁𝑂)𝑖𝑛
(8)
The molar flows of NO are obtained from concentration measurements in port L2 (amount at furnace inlet) and L3 (amount at furnace outlet) in Figures 1 and 4, combined with volume-flow measurements of the oxidizer and the calculated flue-gas volume flow. In the air-combustion case, only the once-through efficiency for reduction of NO added to the entering air stream is formulated, since no recycling is applied.
8
Figure 4. Definition of once-through and total NO reduction efficiency; 𝑛̇ (𝑁𝑂)𝑥
[mol/s]
denotes the molar flow of nitric oxide at positions x. L1-3 are the positions of the measurement ports given in Figure 1.
RESULTS The results are presented in two parts. The first part gives the measured compositions and temperature profiles of the air and the OF 30 flames. Based on these measurements, the implications of oxy-fuel conditions for, especially, CO concentration and nitrogen chemistry in the furnace are discussed. The second part deals with the influence of oxygen concentration in the feed gas and stoichiometric ratio of the burner on the reburning efficiency for once-through and total reduction of NO. FLAME CHARACTERIZATION Figure 5 shows contour maps of temperature and concentrations of CO and NO in the flame zone for OF 30 and air-fired conditions (at λburner = 1.15 and 500 ppm NO inlet concentration). The maps are drawn by interpolation of inlet and furnace concentrations (the measurement positions are indicated by white crosses). The OF 30 and the air flames have similar temperatures (Figure 5a) and are, therefore, suitable for direct comparison. The maximum measured flame temperature is around 1570°C in the OF 30 case and 1530°C in the air case. Figure 5b shows the measured CO concentration in the OF 30 flame compared to that in the air flame. The flame zone with elevated CO concentration has similar extension in the air and in the OF 30 flames. The highest CO
9
concentration measured in the OF 30 case exceeds the measurement limit of the analyzer, which is 20 vol. %. The maximum CO concentration in the air case is considerably lower (9.5 vol. %). As the measurements are conducted with a gaseous fuel, homogenous reaction routes for the production of CO are shown to be sufficient to explain the large difference. The difference in CO concentration is observed up to an axial distance of over 400 mm from the burner, which is a substantial part of the flame-zone in all cases. Thus, the combustion chemistry in oxy-fuel combustion differs from that in air combustion in extensive parts of the flame. As discussed earlier, this has possible consequences on the nitrogen chemistry. Figure 5c presents the NO concentration profiles of the two reference flames. The main part of the NO reduction by reburning takes place in the near-burner region. The measured minimum concentration of NO (Measurement Port R2, 215 mm downstream of the burner outlet) is comparable in the two cases. Downstream of the minimum, some fraction of the produced intermediate nitrogen species is again oxidized to NO. This NO formation is one of the main effects leading to increasing NO concentration towards the furnace exit. The outlet NO concentration is around 400 ppm for air-firing and around 460 ppm for oxy-fuel combustion. Due to internal recirculation, the concentration in the major part of the combustor volume is similar to the combustor outlet concentration. a)
b)
c)
Figure 5. Contour maps of a) temperature, b) CO-concentration and c) NO-concentration in the OF 30 (left-hand part of the furnace) and the air flame (right-hand part) cases. White crosses mark measurement positions and black crosses the contour lines. PARAMETER STUDY
10
The burner stoichiometry and inlet oxygen concentration together determine the amount of flue-gas recycled to the burner, and thereby the amount of NO recycled to the flame. Depending on the recycle ratio, the total residence time of NO in the reducing part of the flame is varied compared to once-through combustion. The total reduction efficiency can be expressed as a function of the oncethrough reduction and the recycle ratio and was first given by Liu and Okazaki (2003). They described coal-fired experiments, but in the present work fuel-bound nitrogen is excluded, since propane is used as fuel, and the correlation between once-through and total reduction of NO becomes 𝜂𝑡𝑜𝑡𝑎𝑙 =
𝜂𝑜𝑛𝑐𝑒 1 − R FGR ∙ (1 − 𝜂𝑜𝑛𝑐𝑒 )
(9)
Equation (9) shows that when the recycle ratio increases the total reduction of NO also increases. However, the influence of RFGR depends on the once-through reduction; at low once–through reduction efficiency RFGR has a large impact on the total reduction efficiency and vice versa for high reduction efficiency. Figure 6a compares the once-through and total reduction for different inlet oxygen concentrations at a burner stoichiometry of 1.15. The influence of feed-gas concentration of oxygen (including the effect of temperature changes) and of the flue-gas recirculation are illustrated. The once-through NO reduction in the OF-cases is smaller than in the air-fired case (21% O2), which is indicated by the grey horizontal line (λburner =1.15). The sensitivity of once-through reduction (white bars) to changes in inlet oxygen concentration is small. The total reduction of NO (black bars) is distinctly higher in oxy-fuel combustion than in air-fired conditions. More flue gas is recycled in the oxy-fuel cases with lower oxygen concentration (the amount of O2 in the feed gas was constant to yield λburner =1.15). Consequently, the difference between once-through and total reduction increases with decreasing concentration of oxygen in the feed gas. The influence of inlet oxygen concentration on NO reburning for sub-stoichiometric conditions (λburner =0.9) is shown in Figure 6b. At the same inlet oxygen concentrations as in Figure 6a, the recycle ratio is lower, because the amount of oxygen supplied to the burner is lower (lower burner stoichiometric ratio). Two differences can be noted: the influence of inlet oxygen concentration on once-through reduction is higher, and, the advantage of oxy-fuel compared to air combustion in terms of total reduction is lower than in the over-stoichiometric cases. The latter effect is explained by two variables included in Equation (9): the recycle ratio RFGR is lower in the sub-stoichiometric cases, and, additionally, there is less impact of RFGR on the increase of total reduction, because ηonce is higher than in the over-stoichiometric cases. a)
b)
11
Figure 6. Comparison between once-through NO reduction and total NO reduction for different inlet oxygen concentrations a) λburner = 1.15; b) λburner = 0.9. Figure 7 shows the influence of burner stoichiometry (including temperature effects) on NO reburning in the OF 35 case and compares oxy-fuel combustion with air combustion operated at the corresponding burner stoichiometries (horizontal lines). The once-through NO reduction (white bars) varies from 80 % at low burner stoichiometric ratio to about 20 % at high stoichiometric ratio in both air and oxy-fuel cases. However, at burner stoichiometries above 0.8, the reduction is lower under oxy-fuel conditions. As shown in Figure 6, the total reduction (black bars) is greater than in the corresponding air case for all investigated stoichiometries. The total reduction approaches the oncethrough reduction with decreasing stoichiometric ratio, both due to decrease in recycle ratio and to the fact that the impact of recycle ratio is lower at higher once-through reduction (c.f. Equation (9)). For this reason, at all investigated stoichiometric ratios, the OF 35 case shows a similar advantage in total NO reduction as in the corresponding air case.
Figure 7. Comparison between once-through NO reduction and total NO reduction at different burner stoichiometries (OF 35 case)
12
Conclusions This work investigates the influence of inlet oxygen concentration, oxygen-to-fuel ratio and recycle ratio on the efficiency of the reburning reduction under oxy-fuel conditions. Experiments were conducted in a 100 kWth oxy-fuel combustion unit with propane as a fuel. The oxidant was composed of oxygen diluted by dry recycled flue gas, into which NO was injected. In the investigated oxy-fuel cases (25 to 37 vol. % O2) the once-through reduction by reburning is lower than in air combustion. The peak concentration of CO in oxy-fuel combustion may be more than twice of that during air-firing when similar temperature conditions are applied. Homogenous reaction routes for the production of CO are shown to be sufficient to explain the large differences in CO concentration seen in the experiments. A significant difference in CO concentration was measured over large parts of the flame. This indicates that he region of the flame, where the combustion environment has significant influences on the nitrogen chemistry, may have a larger extension in oxy-fuel combustion than in air-firing. Stoichiometric ratio is an important parameter for efficient NO reduction by reburning in oxy-fuel combustion, while the inlet oxygen concentration is of minor importance. However, under fuel-rich conditions, low oxygen concentration in the feed gas influences NO reduction significantly. Both burner stoichiometry and inlet oxygen concentration determine the flue-gas recycle ratio. Recirculation of flue gas increases the total residence time of NO in the flame zone, so the total reduction of NO is improved during oxy-combustion compared to that seen in air firing. The influence of the recycle ratio on the total reduction is important at low reduction efficiency, but it becomes less significant at increasing once–through reduction efficiencies. Hence, if high reduction (>80%) is desired, the recycle of flue gases is less important. On the other hand, oxy-fuel combustion does not need to be designed for low-NOx emission to achieve a fairly good reduction, as already low recycle ratios lead to significant improvement compared to air firing.
Acknowledgement. This work has been financed by ALSTOM Power Systems GmbH and Vattenfall AB.
13
References Andersson, K., F. Normann, F. Johnsson & B. Leckner (2008) NO emission during oxy-fuel combustion of lignite. Industrial and Engineering Chemistry Research, 47, 1835-1845. Bilbao, R., M. U. Alzueta & A. Millera (1995) Experimental study of the influence of the operating variables on natural gas reburning efficiency. Industrial and Engineering Chemistry Research, 34, 4531-4539. Croiset, E. & K. V. Thambimuthu (2001) NOx and SO2 emissions from O2/CO2 recycle coal combustion. Fuel, 80, 2117-2121. Dagaut, P. & F. Lecomte (2003) Experiments and kinetic modeling study of NO-reburning by gases from biomass pyrolysis in a JSR. Energy and Fuels, 17, 608-613. Dhungel, B. 2009. Experimental investigations on combustion and emission behaviour during oxy-coal combustion. In IFK. Stuttgart: Stuttgart University. Fan, Y., J.-y. Ren, W. Onstot, J. Pasale, T. T. Tsotsis & F. N. Egolfopoulos (2003) Reactor and Technical Feasibility Aspects of a CO2 Decomposition-Based Power Generation Cycle, Utilizing a High-Temperature Membrane Reactor. Industrial & Engineering Chemistry Research, 42, 2618-2626. Frassoldati, A., T. Faravelli & E. Ranzi (2003) Kinetic modeling of the interactions between NO and hydrocarbons at high temperature. Combustion and Flame, 135, 97-112. Giménez-López, J., A. Millera, R. Bilbao & M. U. Alzueta (2010) HCN oxidation in an O2/CO2 atmosphere: An experimental and kinetic modeling study. Combustion and Flame, 157, 267-276. Glarborg, P. & L. L. B. Bentzen (2008) Chemical effects of a high CO2 concentration in oxyfuel combustion of methane. Energy and Fuels, 22, 291-296. Hjärtstam, S., K. Andersson, F. Johnsson & B. Leckner (2009) Combustion characteristics of lignite-fired oxy-fuel flames. Fuel, 88, 2216-2224. Hu, Y., S. Naito, N. Kobayashi & M. Hasatani (2000) CO2, NOx and SO2 emissions from the combustion of coal with high oxygen concentration gases. Fuel, 79, 1925-1932. Itoh, N., M. A. Sanchez, W.-C. Xu, K. Haraya & M. Hongo (1993) Application of a membrane reactor system to thermal decomposition of CO2. Journal of Membrane Science, 77, 245-253. Jin, W., C. Zhang, P. Zhang, Y. Fan & N. Xu (2006) Thermal decomposition of carbon dioxide coupled with POM in a membrane reactor. AIChE Journal, 52, 2545-2550. Kühnemuth, D., F. Normann, F. Johnsson & K. Andersson. 2008. Evaluation of Concepts for secondary SOx and NOx Removal from the Oxy-Fuel Process. In 15th International Conference on the Properties of Water and Steam. Berlin. Levy, J. M., L. K. Chan, A. F. Sarofim & J. M. Beer (1981) Nitric oxide/char reactions at pulverized coal flame conditions. Symp. (Int.) Combust., [Proc.], 18th, 111-20. Li, Q., C. Zhao, X. Chen, W. Wu & Y. Li (2009) Comparison of pulverized coal combustion in air and in O2/CO2 mixtures by thermo-gravimetric analysis. Journal of Analytical and Applied Pyrolysis, 85, 521-528.
14
Liu, F., H. Guo, G. J. Smallwood & O. L. Gülder (2001) The chemical effects of carbon dioxide as an additive in an ethylene diffusion flame: Implications for soot and NOx formation. Combustion and Flame, 125, 778-787. Liu, H. & K. Okazaki (2003) Simultaneous easy CO2 recovery and drastic reduction of SOx and NOx in O2/CO2 coal combustion with heat recirculation. Fuel, 82, 1427-1436. Liu, H., R. Zailani & B. M. Gibbs (2005) Comparisons of pulverized coal combustion in air and in mixtures of O2/CO2. Fuel, 84, 833-840. Mackrory, A. J. 2008. A mechanistic investigation of nitrogen evolution in pulverized coal oxyfuel combustion. In Department of Mechanical Engineering. Brigham Young University. Mendiara, T. & P. Glarborg (2009a) Ammonia chemistry in oxy-fuel combustion of methane. Combustion and Flame, 156, 1937-1949. Mendiara, T. & P. Glarborg (2009b) Reburn chemistry in oxy-fuel combustion of methane. Energy and Fuels, 23, 3565-3572. Nigara, Y. (1986) Production of carbon monoxide by direct thermal splitting of carbon dioxide at high temperature. Bulletin of the Chemical Society of Japan, 59, 1997-2002. Normann, F., K. Andersson, F. Johnsson & B. Leckner (2010) Reburning in oxy-fuel combustion. Submitted for publication. Normann, F., K. Andersson, B. Leckner & F. Johnsson (2008) High-temperature reduction of nitrogen oxides in oxy-fuel combustion. Fuel, 87, 3579-3585. Normann, F., K. Andersson, B. Leckner & F. Johnsson (2009) Emission control of nitrogen oxides in the oxy-fuel process. Progress in Energy and Combustion Science, 35, 385-397. Okazaki, K. & T. Ando (1997) NO(x) reduction mechanism in coal combustion with recycled CO2. Energy, 22, 207-215. Tan, Y., E. Croiset, M. A. Douglas & K. V. Thambimuthu (2006) Combustion characteristics of coal in a mixture of oxygen and recycled flue gas. Fuel, 85, 507-512. Toftegaard, M. B., J. Brix, P. A. Jensen, P. Glarborg & A. D. Jensen (2010) Oxy-fuel combustion of solid fuels. Progress in Energy and Combustion Science, In Press, Corrected Proof. Toporov, D., P. Bocian, P. Heil, A. Kellermann, H. Stadler, S. Tschunko, M. Förster & R. Kneer (2008) Detailed investigation of a pulverized fuel swirl flame in CO2/O2 atmosphere. Combustion and Flame, 155, 605-618. White, V., L. Torrente-Murciano, D. Sturgeon & D. Chadwick (2010) Purification of oxyfuelderived CO2. International Journal of Greenhouse Gas Control, 4, 137-142. Zhang, L., E. Binner, Y. Qiao & C.-Z. Li (2010) In situ diagnostics of Victorian brown coal combustion in O2/N2 and O2/CO2 mixtures in drop-tube furnace. Fuel, In Press, Corrected Proof.
15
Oxy-Combustion of Pulverized Coal: Modeling of CharCombustion Kinetics M. Geier1, C. R. Shaddix1, and B. S. Haynes2 1
2
Combustion Research Facility, Sandia National Labs, Livermore, CA 94550, USA School of Chemical and Biomolecular Engineering, University of Sydney, Sydney, NSW 2006, Australia
In this study, char combustion of pulverized coal under oxy-fuel combustion conditions was investigated on the basis of experimentally observed temperature-size characteristics and corresponding predictions of numerical simulations. Using a combustion-driven entrained flow reactor equipped with an optical particlesizing pyrometer, combustion characteristics (particle temperatures and apparent size) of pulverized coal char particles was determined for combustion in both reduced oxygen and oxygen-enriched atmospheres with either a N2 or CO2 bath gas. The two coals investigated were a low-sulfur, high-volatile bituminous coal (Utah Skyline) and a low-sulfur subbituminous coal (North Antelope), both size-classified to 75– 106 m. A particular focus of this study lies in the analysis of the predictive modeling capabilities of simplified models that capture char combustion characteristics but exhibit the lowest possible complexity and thus facilitate incorporation in existing computational fluid dynamics (CFD) simulation codes. For this purpose, char consumption characteristics were calculated for char particles in the size range 10–200 m using (1) single-film, apparent kinetic models with a chemically “frozen” boundary layer, and (2) a reacting porous particle model with detailed gas-phase kinetics and three separate heterogeneous reaction mechanisms of char-oxidation and gasification. A comparison of model results with experimental data suggests that single-film models with reaction orders between 0.5 and 1 with respect to the surface oxygen partial pressure may be capable of adequately predicting the temperature-size characteristics of char consumption, provided heterogeneous (steam and CO2) gasification reactions are accounted for.
1. Introduction Coal combustion using pure oxygen as the oxidizer appears to be an economically promising process for CO2 separation for carbon capture and storage and has thus received considerable attention of the scientific community in recent years (e.g. [1-3] and references therein). Temperature control and material safety issues generally require a substantial degree of flue gas recycling, which yields elevated concentrations of carbon dioxide in the char combustion environment. Depending on whether the flue gas is dried before being recycled, higher moisture contents may be present as well. Flue gas recycling is particularly necessary for boiler retrofit applications, in which gas and particle temperatures similar to air-blown combustion are required to achieve the necessary heat transfer in different regions of the boiler. Due to the higher molar heat capacity of carbon dioxide
relative to nitrogen, flame temperatures for a given initial oxygen concentration are lower. Therefore, retrofitted boilers have to be operated at enhanced oxygen levels to maintain similar heat transfer rates [4,5]. Application of O2-recycle combustion can cause differences in furnace operation parameters such as burner stability, char burnout, heat transfer and gas temperature profiles, as various pilot-scale tests have shown [4-12]. Computational fluid dynamics (CFD) simulations have become an indispensable tool during the process of designing and optimizing coal-fired boilers. CFD codes are frequently used to predict temperatures, heat transfer rates and species profiles in full-scale facilities, posing high demands on computational performance. For computations to provide results within practically acceptable time frames, reduced spatial resolution of the computational domain and simplified coal combustion models have to be used. The challenge in
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey Oct. 11-14, 2010
2 this context is to develop models that are computationally inexpensive while still capturing all relevant information to produce realistic predictions. As most kinetics models and appropriate rate parameters have been derived for conventional, airblown combustion systems, application of existing models for prediction of char burnout characteristics (particle temperatures and consumption rates) may yield erroneous results if char consumption in the oxy-combustion environment is influenced by additional reactions. Recent computational results suggest that the the endothermic char gasification reaction C(s) + CO2 → 2CO significantly reduces the temperature of pulverized coal chars burning at high temperatures in oxy-combustion environments [13]. The reaction affects the gas composition in the particle interior and, for 130 µm particles, leads to enhanced char consumption rates in environments with less than 24% oxygen. Similarly, simulations including the steam gasification reaction C(s) + H2O → H2 + CO show a reduction of char particle combustion temperatures, but a slight enhancement of the char consumption rate [14]. Simulation results further suggest that the partial conversion of CO to CO2 in the particle boundary layer has an important influence on the char particle temperatures and burning rates under oxygenenriched combustion conditions [15]. This is generally expected to be more important for larger particle sizes, as the characteristic diffusion time through the boundary layer increases for larger particles and thus provides more time for oxidation of CO to proceed. The purpose of this study is to identify approaches for modification of existing single-film, nth-order apparent char kinetic models in order to accommodate differences in char consumption characteristics between conventional and oxy-fuel conditions. To this end, particles of two coals (a high-volatile bituminous and a subbituminous coal), sieved to 75-106 μm, are combusted in N2 and CO2 baths with different contents of O2, and experimentally observed temperature-size characteristics are compared with predictions of simple single-film models. Results of simulations with the porous particle combustion code SKIPPY are used to provide complementary information.
2.Experiment 2.1. Gas composition and temperature Char consumption characteristics of pulverized coal was studied experimentally using Sandia’s optical entrained flow reactor, described in more detail in [16]. The combustion-driven reactor relies on a diffusion-flamelet-based Hencken burner to produce a high-temperature gas flow at ambient pressure (1 atm). Solid fuel particles are introduced at the furnace centerline through a 0.75 mm stainless-steel tube by means of a small flow of N2 or CO2, depending on the diluent used in the main reactant flow. For the experiments reported here, the coal feed rate was kept sufficiently low (~0.6 g/hr) to assure that injected particles burned in isolation from one another. The experimental setup further consists of a 46 cm tall 5 cm 5 cm quartz chimney that isolates the hot, postcombustion gas flow from the surrounding air and enables in situ optical measurements on the particles injected into the flow.
Figure 1 Schematic of Sandia’s laminar entrained flow reactor for coal combustion studies. The particle-sizing pyrometry diagnostic is shown to the right of the flow reactor. Data collection is triggered by a particle passing through the focus of the HeNe laser beam along the centerline of the reactor.
For comparisons with predictions of different char combustion models, sieve-classified coal particles were entrained into mixtures with 12% (by mole), 24% or 36% O2, 16% moisture, and CO2 or N2 balance gas. Particle temperature-size data were collected at locations between 25 and 125 mm above the burner face, where gas temperatures were 1680±40 K (see Figure 2). The gas temperature profile peaked at about 1700±10 K at
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
3 50 mm above the burner surface and then decreased nearly linearly with increasing height (temperature profiles remained in a band of ±10 K from the average gas temperature profile). The total burner product flow rate was 60 slpm (liters per minute at 0°C and 1 atm), which includes 0.03 slpm of diluent flow used for entraining the particles.
steady-state species concentration and temperature profiles of a single, porous, spherical particle placed in a quiescent, chemically reacting environment. Based on user-specified mechanisms for heterogeneous (gas-solid) and homogeneous reactions (e.g. GRIMECH 3.0 [19]), species concentrations and temperature inside the porous particle, at the external surface, and in the gas layer outside the particle are computed. Table 1 Proximate and Ultimate Analysis of Coals Coal Type Utah Skyline North Antelope wt%, as wt% dry wt%, as wt% dry Proximate rec’d rec’d moisture 3.18 23.69 ash
8.83
9.12
4.94
6.47
volatiles
38.60
39.87
33.36
43.72
fixed C
49.39
51.01
38.01
49.81
Ultimate C Figure 2 Radiation-corrected thermocouple measurements of gas temperature profiles in the entrained flow reactor. Measurements used a 25-µm, type-R fine-wire thermocouple and were corrected for radiative losses following [17].
2.2. Fuels The two coals investigated were a high-volatile bituminous coal (Utah Skyline) and a subbituminous coal from the Powder River Basin (North Antelope). Proximate and ultimate analyses of the pulverized coals are summarized in Table 1. For particle-sizing pyrometry, the coals, which are both considered low-sulfur coals, were sieved to 75-106 µm to aid data acquisition. Additional analysis of three coal size fractions (53-75 µm, 75106 µm and 106-125 µm) confirmed sizeindependent elemental composition of ash contents. 3.Numerical Modeling 3.1. Detailed Char Consumption Model (SKIPPY) SKIPPY (Surface Kinetics in Porous Particles) [18] is a FORTRAN program that calculates 1-D,
wt% dry wt% DAF wt% dry wt% DAF 70.60 77.44 53.72 56.51
H
5.41
5.93
6.22
6.54
O (by diff.)
13.21
14.49
34.11
35.88
N
1.42
1.56
0.78
0.82
S
0.53
0.58
0.23
0.24
The necessity of specifying reaction mechanisms makes SKIPPY a well-suited tool for studying the effects of reactions in the boundary layer (e.g. [15, 20]) or the importance of individual heterogeneous reactions (e.g CO2 gasification reaction on char consumption during oxy-fuel combustion of pulverized coal [13]). For the study presented here, we explored the effect of adding steam and CO2 gasification reactions (R5 and R6 in Table 2) to the surface oxidation mechanism originally employed by Molina et al. [21]. In previous work (without gasification reactions), reactive specific char surface area was adjusted to bring observed and predicted temperatures of char particles burning in N2 at different oxygen concentrations into agreement [15]. This approach resulted in a 24-fold reduction in surface area for char particles burning in 36% O2 relative to those burning in 12% O2. While there may be differences in surface areas of char formed in situ at differ-
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
4 ent bulk oxygen concentrations (due to differences in devolatilization characteristics), the variation in surface area that was required in those computations seems too large to be attributable to this phenomenon only. More likely, the simple heterogeneous reaction mechanism that was employed (char reaction with O2 as the sole reactant) does not adequately model char surface reactions. Allowing for contributions by gasification reactions seems a reasonable approach, particularly for oxyfuel combustion systems with flue gas recycling, where CO2 (and possibly H2O) concentrations can be considerably higher than in conventional combustion systems. The mechanism summarized in Table 2 is clearly an over-simplification (e.g. inhibition effects of CO and H2 on the gasification reactions [22], rate-limiting effects of product desorption [23], and dependence of the CO/CO2 production ratio on the concentration of O2 at the surface [24] are not captured), but the goal here was to shed some light onto the relative importance of oxidation and gasification reactions in oxy-fuel versus conventional pulverized-coal combustion systems. Table 2 Reaction rate constant parameters for the heterogeneous reactions used for SKIPPY simulations. The rate constants are specified in terms of Arrhenius expressions with pre-exponential factors A and activation energy E. The use of “_s” denotes species bond to surface sites, and “C(b)” denotes (inaccessible) bulk carbon atoms. Reaction Heterogeneous oxidation: (R1) C_s + O2 => CO + O_s (R2) O_s + 2C(b) => CO + C_s (R3) C_s + O2 => O2_s+ C(b) (R4) O2_s + 2C(b) => C_s + CO2 CO2 gasification reaction: (R5) C_s + CO2 => CO + O_s + C(b) Steam gasification reaction: (R6) C_s + H2O => H2 + O_s + C(b)
A (g/cm2 s)
E (kJ/mol)
3.3E+15 1.0E+08 9.5E+13 1.0E+08
167.4 0. 142.3 0.
3.60E+15
251.0
4.35E+14
222.8
In order to compare predictions and observations, char particle temperatures were computed for the size range 10 - 200 μm. The SKIPPY simulations used GRIMECH 3.0 [19] without nitrogen reactions and used the heterogeneous reactions listed in Table 2, with the rate parameters of R5 and R6 varied to assess the importance of these reactions. Internal specific surface area was held
constant for all particle sizes and O2 concentrations; a value of 10 m2/g was used to crudely reproduce temperatures for 100 μm particles burning in 12% O2. At this condition, particle temperatures are around 1900 K, which is relatively low to expect appreciable effect of the high activation energy gasification reactions [13-14]. Further simulation input parameters include bulk density = 560 kg/m3, tortuosity = 5, void fraction = 0.4, particle thermal conductivity = 1.33 W/m·K, emissivity = 0.8, wall temperature (for radiative heat transfer) = 500 K, pressure = 101 kPa and gas temperature = 1690 K.
3.2. Simplified Char Consumption Model Kinetic parameters for calculating the rate of char consumption can be obtained from measured data and an appropriate char consumption model. In the case of optical measurements on individually burning char particles carried by a hot gas flow, the char consumption process is governed by mass and heat transfer to and from the particle, as well as chemical reactions, both on the particle surface and in the gas phase. Mathematical models of considerable complexity can certainly be formulated to fully describe the overall process. However, using these models to estimate the rate parameters of heterogeneous reactions (a) requires accurate specification of imperfectly known model details (such as gas phase reaction rates, transport properties, pore structure, etc.), and (b) yields kinetics rate parameters that can produce poor predictions if utilized directly with simplified models of CFD codes. For those reasons, we compare measured temperature-size data with predictions from a relatively simple burnout model, as outlined below. A simplified model, similar to those typically employed in CFD codes, was derived from the instantaneous energy balance on a homogeneous, chemically reacting, spherical particle. Assuming negligible effect of reactions in the boundary layer, average gas-mixture properties, and the overall char reaction C + (1+ψ)/2 O2 → ψ CO2 + (1−ψ) CO, the instantaneous energy balance is given by (see [25] for more detail):
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
5 d p p c vp dTp 6
dz
(Tp4 Tw4 )
2 2 (Tp Tg ) q h d p e / 2 1
The left-hand term in this equation represents the thermal inertia of the particles, with particle diameter dp (m), particle bulk density ρp (kg/m3), specific heat c (J/kg·K), particle speed vp (m/s), surface temperature Tp (K), and spatial coordinate z (m). For the current set of measurements, the particle temperatures are nearly independent of the distance from the burner face, allowing this term to be neglected. The first term on the right represents radiative heat loss, with char emissivity ε and wall temperature Tw (K). The second term on the right represents convective heat loss of the particle, with mixture thermal conductivity λ (W/m·K), free-stream gas temperature Tg (K) and κ = (−q·dp/λ) ∑iνi cg,i, which characterizes the heat transfer correction due to Stefan flow [25] (νi are stoichiometric coefficients, with νO2 = (1+ψ)/2, νCO = −(1−ψ), and νCO2 = −ψ; cg,i are the corresponding specific heat capacities (J/kg·K)). The chemical heat release is represented by the far right-hand term, where q denotes the overall burning rate per unit external surface area (kg/m2·s), and Δh = (1−ψ) ΔhCO + ψ ΔhCO2 is the overall heat of reaction (J/kg). The CO2/CO production ratio at the char particle surface is modeled as CO2/CO ≡ ψ/(1−ψ) = 0.02 p(O2,s)0.21 exp(3070/Tp) [24], where p(O2,s) is the oxygen partial pressure at the particle surface (atm). p(O2,s) follows from solution of the gas-phase diffusion equation [25] as p(O2,s)/p = γ + [p(O2,∞)/p − γ] exp[−q/(γ·kd·p)], where γ = −(1+ψ)/(1−ψ), p(O2,∞) is the free-stream oxygen partial pressure (atm), kd = 2·WC·DO2,mix/ (dp·R·Tf·νO2) is the oxygen mass transfer coefficient (kg/m2·s·atm) with WC = 12 g/mol, DO2,mix is the mixture-averaged oxygen diffusion coefficient (m2/s), R = 8.3144 J/mol·K is the universal gas constant, and Tf = (Tp + Tg)/2 is the film temperature (K). Gas properties (λ, DO2,mix, and cg) were calculated for T = Tf, and the free-stream gas composition as outlined in [26] (κ was approximated as κ ≈ −q·dp·cg·WC/(γ·λ·νO2)). The surface-specific burning rate q, is typically expressed as q k s (Tp ) pon2 ,s
where ks(Tp) = WC·A·exp(−E/RTp) in units of kg/m2·s·atmn, n is the reaction order, and the preexponential factor A is expressed in units of kmol/m2·s. This is the “nth-order Arrhenius” expression of global char kinetics. For Zone-II combustion conditions, as are typically found for pulverized coal combustion, the diffusion of oxygen through the boundary layer has some influence on the burning rate of the particle, but the influence is not controlling the rate. However, the oxygen diffusion profile needs to be accounted for when calculating the surface concentration of oxygen. For the study presented here, temperature-size characteristics were calculated with the described model. The required rate parameters A and E to specify q were determined for three reaction orders n = 1, 0.5 and 0.1 such that observed temperatures of 100 μm particles burning in N2 with 12% and 36% O2 in the free stream were reproduced. Particle temperatures were then computed for sizes from 10 to 200 μm for the experimental conditions studied (either CO2 or N2 diluent, 16% H2O and three O2 bulk concentrations (12%, 24%, 36%)). A gas temperature Tg = 1690 K, radiative boundary Tw = 500 K, and char particle emissivity ε = 0.8 were assumed. The obtained rate parameters are summarized in Table 3. Table 3 Estimated rate parameters for North Antelope and Utah Skyline for the simplified steady-state burning model. Fixed n = 1 0.5 0.1
North Anthelope: A E (kmol/m2·s) (kJ/mol) 0.44 0.00 0.67 36.50 1.70 76.26 Utah Skyline:
1 0.5 0.1
0.39 1.05 2.31
11.95 51.75 83.54
4. Results and Discussion 4.1. Single-film char consumption models Figure 3 shows the results from the simplified model calculations for different reaction orders together with measured data for char particles
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
6 burning in N2 diluent. Temperatures are significantly higher for the subbituminous North Antelope char (top), for which considerably smaller sizes were also measured (most likely due to a stronger tendency to fragment than the bituminous Utah Skyline char (bottom)). The overall spread of temperatures of Utah Skyline char particles is larger for a given reactor condition, which may reflect a higher variability in mineral contents of the coal particles. For both North Antelope and Utah Skyline char, the calculated temperatures agree well near the “design” diameter 100 μm, but the diameter range for which predictions are reasonably good appears to be wider for reaction orders of n = 0.5 and n = 1. According to classical Thiele analysis, apparent reaction orders are constrained to lie between 0.5 and 1 for Zone II combustion [27,28], even though smaller reaction orders have been obtained in regression procedures for estimation of rate parameters (e.g. [25]) and ash-inhibition or similar processes that decrease reactivity with burnout can result in low effective reaction order [29]. The results shown in Figure 3 suggest that the predictions with n = 0.1 hold for sizes larger than 100 μm, and thus could be used in CFD models as long as the kinetics rates are not applied to particle sizes outside that range. Choosing a smaller particle size when estimating the rate parameters would have improved the predictions at smaller sizes, but at the cost of over-predicting temperatures for larger sizes. Given the observed nearly size-independent temperatures, it is recommended to keep reaction orders in the range between 0.5 and 1, to obtain more realistic profiles over the entire size range. However, the activation energies that are required when fitting the simulations to the data for high reaction orders can be very small (e.g. for n =1, E = 11.95 kJ/mol and 0 kJ/mol for North Antelope and Utah Skyline, respectively) and should be viewed as pure fitting parameters without practical meaning. But fitted parameters that strongly deviate from a physically realistic range indicate that the fitted model does not fully capture all relevant physical processes. Figure 4 shows predicted and measured particle temperatures as a function of particle diameter for char particles burning in CO2-dominated gas mix-
tures. Compared with the data for N2 diluent (Figure 3), the measured temperatures for both chars are lower in the CO2 environment, consistent with previous findings [15]. Interestingly, for each free-stream oxygen concentration the char temperatures show a much wider spread about an average temperature-size profile than for the N2 environment. This may be caused by differences in properties of in situ char produced in N2 and CO2, but further investigation is necessary to clarify this point.
Figure 3 Predicted and measured (symbols) particle surface temperatures for three free-stream O2 concentrations, 16% moisture and N2 bath gas. Solid lines n = 0.5, dotted lines n = 1, dashed lines n = 0.1.
As shown in Figure 4, predicted temperatures for reaction orders of n = 0.5 and n = 1 are too high for North Antelope char particles but agree reasonably well for Utah Skyline char. Although
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
7 the predictions are slightly better for n = 1, both models yield similarly good predictions. In interpreting the results for Utah Skyline char, two issues must be kept in mind. First, the measured temperatures for both N2 and CO2 experiments are ~ 150 K lower than for North Antelope char, thus potential contributions by gasification reactions are less significant for Utah skyline char. Second, due to the higher variability in measured temperatures (presumably a consequence of higher variability of mineral contents), average temperatures of 100 μm particles seem biased towards lower temperatures (see Figure 3). The increased spread of temperatures of char burning in a CO2 bath further obscures clear interpretation of discrepancies between model and experimental data. The over-prediction of temperatures for North Antelope char shown in Figure 4 clearly points to inadequacy of the simplified model. From a CFD modeling perspective, several modifications of the model without the need to include time-consuming calculations of temperature and multi-species profiles in the boundary layer could be conceived. The simplest option may be to use the simplified model as it is, but with a different set of rate parameters for CO2 environments. While this is straightforward, the model may remain valid only in a narrow range of moisture and CO2 concentrations if gasification reactions considerably affect particle temperatures and/or char consumption rates. The next level of complexity might thus be to include char gasification reactions in CFD codes.
between 30 and 70 μm, with peak temperatures shifting towards smaller sizes with increasing freestream oxygen concentration.
Figure 4 Predicted and measured (symbols) particle surface temperatures for three free-stream O2 concentrations, 16% moisture and CO2 bath gas. Solid lines n = 0.5, dotted lines n = 1, dashed lines n = 0.1.
4.2. SKIPPY Simulations Simulated particle temperatures for combustion in N2-dominated environments and three different heterogeneous reaction mechanisms are shown in Figure 5. The first mechanism consists of oxidation reactions R1-R4 only, which, as mentioned earlier, results in vast over-prediction of particle temperatures at elevated oxygen concentrations. The temperatures of smaller particles are controlled by convective heat loss, whereas radiative losses cause the temperature to fall at the largesize end of the studied size range. Highest temperatures are predicted in the intermediate size range
The second mechanism includes R1-R4 and the steam gasification reaction, R6; however, in this case a different set of rate parameters was used for R6 than listed in Table 2: A = 1.16E16 g/cm2·s, and E = 251.0 kJ/mol. This parameter set was found to produce the best match of predicted and observed temperatures of 100 μm particles and is still well within the range of activation energy and overall reaction rate that appears in the literature for steam gasification [14]. The relatively good fit for all oxygen bulk concentrations suggests that the steam gasification reaction plays an important
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
8 role in determining the char particle temperature and thereby the char oxidation rate. The third model adds the CO2 gasification reaction, R5, and thus uses the full mechanism specified in Table 2. Although the fit for 100 μm particles is slightly better than for the second model, the necessity of including this reaction is not obvious, based on the data shown in Figure 5 for N2dominated environments.
when adding the heterogeneous CO2 reaction. This implies that CO2 gasification plays an important role also when insignificant amounts of CO2 are present in the bulk gas. Under those conditions, carbon dioxide formed at the char surface and in the boundary layer participates in gasification reactions (because of relative slow diffusion of CO2 in N2, appreciable concentrations of CO2 can be present at the particle surface despite low CO2 production rates). For application to CFD codes, the suggested relevance of steam and CO2 gasification reactions implies that lumped kinetics to model char consumption as a single-step oxidation process may only be valid for narrow ranges of combustion environments (in terms of steam and carbon dioxide concentrations). Therefore, it may be necessary to account for both steam and CO2 concentrations to produce appropriate estimates of char temperatures and burning rates.
Figure 5 SKIPPY predictions of particle surface temperatures for three free-stream O2 concentrations, 16% moisture and N2 diluent compared with measured data for North Antelope char. Model I: oxidation only (R1-R4), Model II: oxidation with steam gasification (R6 with E = 251 kJ/mol, A = 1.16E16 g/cm2 s, Model III: R1-R6 as specified in Table 2.
Comparison of experimental data and model prediction for combustion in CO2 diluent (Figure 6) clearly shows that both steam and CO2 gasification reactions have to be included to produce an acceptable fit for sizes near 100 μm. Both reduced models (oxidation alone and oxidation together with steam gasification) over-predict temperatures for that size range. The suggested importance of CO2 gasification implies that CO oxidation in the boundary layer, which produces additional CO2 in the vicinity of the particle, may be more important then previously assumed. Similarly, the water-gas shift reaction (CO + H2O → CO2 + H2) might play a subtle role due to its impact on the gas composition near the particle surface. For combustion in conventional environments, the rate parameters for the steam gasification reaction had to be reduced
Figure 6 SKIPPY predictions (solid lines) of particle surface temperatures for three free-stream O2 concentrations, 16% moisture and CO2 diluent compared with measured data for North Antelope char. Model I: oxidation only (R1-R4), Model II: oxidation with steam gasification (R6 with E = 251 kJ/mol, A = 1.16E16 g/cm2 s, Model III: R1-R6 as specified in Table 2.
5. Summary Combustion of char particles of bituminous (Utah Skyline) and subbituminous (North Antelope) coal in N2 and CO2 dominated gas enviroments with 12 to 36 mole-% O2 and 16% H2O were investigated both experimentally and through numerical simulations. Particle temperatures from
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
9 two-color pyrometer measurements were compared with predictions from (1) a simplified single-film apparent kinetics model, and (2) a more complex model that includes both homogeneous reactions of gas phase species and heterogeneous reactions in pores and external particle surface. Single-film model predictions suggest that optimized rate parameters for any reaction order between 0.5 and 1 produce reasonably good fits to the data and may thus be employed in CFD codes for wide size and temperature ranges. Particle temperatures for CO2-dominated environments are generally over-predicted when using rate parameters derived with N2 data, i.e. the simplified model investigated does not fully capture the characteristics of char consumption in oxy-fuel combustion conditions. As an alternative to the option of estimating separate sets of rate parameters for CFD applications in N2-dominated and CO2-dominated environments, heterogeneous steam and CO2 gasification reactions could be included to improve the quality of the predictions. Using the reacting particle simulation code SKIPPY (Surface Kinetics in Porous Particles), temperatures for particles around 100 μm were reasonably well predicted for both N2 and CO2 environments if both steam and CO2 gasification reactions were included and a fixed reactive surface area of 10m2/g assumed. The results suggest that oxidation and both steam and CO2 gasification reactions contribute to the char consumption process. An important conclusion therefore is that simplified (single-film) models must account for gasification reactions to maintain predictive capability in both combustion environments. To further substantiate these conclusions, experiments with externally prepared char (to remove effects of the combustion environment on devolatilization characteristics (and hence char properties)) will be conducted. In addition, a more sophisticated oxidation mechanism will be implemented to improve the quality of SKIPPY predictions. Acknowledgement This research was sponsored by the U.S. Department of Energy’s Carbon Sequestration Program under award number DE-NT0005288. This work is part of a project led by Reaction Engineering International and is managed by Mr.
Timothy Fout of the National Energy Technology Laboratory. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under Contract DE-AC04-94AL85000.
Disclaimer This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
References [1] B.J.P. Buhre, L.K. Elliott, C.D. Sheng, R.P. Gupta, T.F. Wall, Prog. Energy Combust. Sci. 31 (4) (2005) 283-307. [2] R. Tan, G. Corragio, S. Santos, Oxy-Coal combustion with flue gas recycle for the power generation industry. A literature review, Report No. IFRF Doc. No. G 23/y/1, International Flame Research Foundation, 2005. [3] M.B. Toftegaard, J. Brix, P.A. Jensen, P. Glarborg, A.D. Jensen, Prog. Energy Combust. Sci. 36 (2010) 581-625. [4] R. Payne, S.L. Chen, A.M. Wolsky, W.F. Richter, Combust. Flame 67 (1989) 1-16. [5] H. Liu, R. Zailani, B.M. Gibbs, Fuel 84 (16) (2005) 2109-2115. [6] T. Kiga, S. Takano, N. Kimura, K. Omata, M. Okawa, T. Mori, M. Kato, Energy Conv. Managmt. 38 (0) (1997) S134. [7] T. Nozaki, S. Takano, T. Kiga, K. Omata, N. Kimura, Energy 22 (2-3) (1997) 199-205. [8] H. Farzan, S.J. Vecci, F. Chatel-Pelage, P. Pranda, A.C. Bose, Pilot-scale evaluation of coal combustion in an oxygen-enriched recycled flue gas. The 30th international technical conference on coal utilization & fuel systems. Clearwater, FL, USA, 2005. [9] H. Liu, R. Zailani, B.M. Gibbs, Fuel 84 (7-8) (2005) 833-840. [10] S.P. Khare, T.F. Wall, A.Z. Farida, Y. Liu, B. Moghtaderi, R.P. Gupta, Fuel 87 (2008) 1042-1049. [11] X. Huang, X. Jiang, X. Han, H. Wang, Energy & Fuels 22 (2008) 3756-3762. [12] P. Heil, D. Toporov, H. Stadler, S. Tschunko, M. Förster, R. Kneer, Fuel 88 (2009) 1269-1274. [13] E. S. Hecht, C. R. Shaddix, A. Molina, B. S. Haynes Proc. Combust. Inst., in press.
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
10 [14] C.R. Shaddix, E.S. Hecht, M. Geier, A. Molina, B.S. Haynes, Effect of gasification reactions on oxy-fuel combustion of pulverized coal char, Proceedings of the 35th International Technical Conference on Clean Coal & Fuel Systems, Clearwater FL, June 6–10, 2010. [15] C.R. Shaddix, A. Molina, Effect of O2 and High CO2 Concentrations on PC Char Burning Rates during Oxy-Fuel Combustion, Proceedings of the 33rd International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, FL, June 1-5, 2008. [16] D.A. Tichenor, S. Niksa, K.R. Hencken, R.E. Mitchell, Proc. Combust. Inst. 20:1213-1221 (1984). [17] C.R. Shaddix, “Correcting thermocouple measurements for radiation loss: A critical review,” 33rd National Heat Transfer Conference, Albuquerque, NM, USA, 1999, p. 1150. [18] P.J. Ashman, B.S. Haynes, Improved techniques for the prediction of NOx formation from char nitrogen, Project No. C4065, Australian Coal Association. CSIRO Energy Technology, 1999. [19] G.P. Smith, S.D. Golden, M. Frenklach, N.W. Moriaty, B. Eitener, M. Goldenberg, C.T. Bowman, R.K. Hanson, S. Song, W.C. Gardiner, V.V. Lissianski, Z. Qin, (2001) GRIMECH 3.0. http://www.me.berkeley.edu/gri_mech (accessed December 2001). [20] M. Geier, E.S. Hecht, C.R. Shaddix, 26th Annual International Pittsburgh Coal Conference, Pittsburgh PA Sept. 20-23, 2009 [21] A. Molina, A.F. Sarofim, W. Ren, J. Lu, G. Yue, J.M. Beer, B.S. Haynes, Combust. Sci. Technol.174 (2002) 43-71. [22] N.M. Laurendeau, Prog. Energy Combust. Sci. 4 (1978) 221-270. [23] R.H. Essenhigh, Energy & Fuels 5 (1991) 41-46. [24] L. Tognotti, J.P. Longwell, A.F.Sarofim, Proc. Combust. Inst. 23 (1990) 1207-1213. [25] J.J. Murphy, C.R. Shaddix, Combust. Flame 144:710729 (2006). [26] P.H. Paul, J. Warnatz, Proc. Combust. Inst. 27 (1998), 495–504. [27] E.W. Thiele, Ind. Eng. Chem. 31 (1939) 916–920. [28] M.F.R. Mulcahy, I.W. Smith, Rev. Pure. Appl. Chem. 19 (1969) 81–108. [29] J.J. Murphy, C.R. Shaddix, Combust. Flame 157: (2010) 535-539
27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 STUDY ON THERMODYNAMIC CALCULATION FOR O2/CO2 FLUE GASES RECYCLED COMBUSTION BOILER Li-Qi Zhang*, Can-Zhi Li, Fang Huang, Ji-Hua Qiu, Chu-Guang Zheng State Key Lab. of Coal Combustion, Huazhong University of Science and Technology 1037 Luoyu Road,Wuhan, Hubei, 430074, P.R.CHINA *Corresponding author. Tel: +86 27 87542417-8316 Fax: +86 27 87545526 Email:
[email protected]
Abstract: Oxy-fuel combustion technology is considered as an effective approach to capture CO 2 from the combustion of fossil fuels. The traditional boiler must be retrofitted in order to adapt to this new technology and the conventional thermodynamic calculation method also needs to be modified for obtaining higher accuracy. A 50MW traditional pulverized coal boiler was calculated in this paper, which adopts modified thermal calculation method to calculate the thermodynamic in the different O2/CO2 recycle mode and in different recycle rate, compares the results with that of traditional air and optimizes the recycle rate in O2/CO2 recycle mode. The results indicated that the best operation condition will be achieved when the oxygen volume of oxidant is 30% (recycle rate 0.633) in hot recycle mode. KEY WORDS:O2/CO2; thermal calculation; recycle rate
1. Introduction O2/CO2 recycle combustion is considered as the most potential new combustion technology on CO 2 capture currently. The main characteristic of this technology is that the air oxidants is replaced by the pure oxygen and recycled part of the flue gas, which can greatly increase the concentration of CO2 in flue gas and the CO2 can be used and disposed without separation, consequently, sharply decrease the emission of CO2 into the atmosphere. Meanwhile, the recycled flue gas significantly reduce the exhaust (only 1/5 of traditional air), which greatly reduce exhaust heat loss and apparently increase the boiler efficiency; besides, this new technology has a good ability of removing SOx and NOx. A lot of studies on this new technology have been done until now, including combustion characteristic, pollutant emission, mineral behavior and the theory of radiant heat transfer et al at laboratory scale; besides, including the technological and economic feasibility of using O2/CO2 recycle combustion on retrofitted traditional boilers, pilot scale study and numerical simulation are also included [1-12]. In the near 5 years (before 2015) O2/CO2 recycle combustion will be soon accomplished from 30MW to full industrial scale, 565MW [13]. Due to the change of combustion atmosphere, the thermal calculation method of traditional boiler is no longer suitable for industry application as for the application of this new combustion technology, thus in urgent need of developing new calculation theory and method. This paper has modified the thermal calculation method of traditional boiler and made it useful in O2/CO2 recycle mode compared the results between recycle mode and air in order to find out the best O2/CO2 recycle mode and the best recycle rate.
2. The Calculated Boiler and Coal The calculated boiler is a pulverized coal boiler whose model is F220/100-W and its parameters are showed in tab.1. Along with the flow direction of flue gas, the components of the boiler are furnace, screen superheater, dregs prevent tubes, high temperature superheater hot section and cold section, low temperature superheater, high temperature economizer, high temperature pre-heater, low temperature economizer and low temperature pre-heater. The coal used in this paper is HeGang bituminous coal and its properties are showed in tab.2. Tab.1 The parameters of the boiler F220/100-W Pulverized coal boiler Foursquare once-through burner, Dry-bottom furnace, Natural circulation Rated load 220 t/h Superheat steam pressure 9.8MPa Superheat steam temperature 540 ℃ Feedwater temperature 215 ℃ Feedwater pressure 11.57 MPa Steam drum pressure 11.08 MPa Discharge rate 2% Ambient temperature 30 ℃ Drum-type ball mill, intermediate storage bunker hot air milling system Tab.2 Proximate and ultimate analysis of bituminous coal Proximate analysis/% Ultimate analysis/% Qnet,ar/(kJ/kg) Aar Mar Var Car Har Oar Nar 20.07 9.00 34.73 22676 58.14 3.86 8.09 0.62
Sar 0.22
3. Thermal Calculation Method in O2/CO2 Recycle Mode Fig.1 is the schematic of boiler in air, fig.2,3 are schematic of boiler in cold and hot recycle mode respectively. Air separation unit, condenser and compressor are added to cold recycle mode compared to air but the original boiler structure and parameters remain unchanged. Reversely, the boiler structure in hot recycle mode change a lot. Behind low temperature superheater, part of the flue gas recycled after got through the high temperature precipitator and delivered into the furnace as part of the secondary oxidant. A high temperature precipitator is added because the hot recycle flue gas need to remove the dust, meanwhile, the recycled of hot flue gas result in the reduction of the flue gas that get through the economizer, so it can replace the high and low temperature economizers by using only one economizer; the amount of primary air and oxygen in O 2/CO2 recycle mode which get through the pre-heaters is less than that in air's, so the high and low temperature pre-heaters can be replaced by one pre-heater as well.
Fig.1 Schematic of boiler in air
Fig.2 Schematic of boiler in cold recycle mode
Fig.3 Schematic of boiler in hot recycle mode
The volume difference of each component in the flue gas between O2/CO2 recycle mode and air is quite large, especially CO2. The volume differences of the components definitely change the properties of flue gas in O2/CO2 recycle mode when compared to that in air. Also, the oxidant in O2/CO2 recycle mode is composed of recycle flue gas and pure oxygen, so that the volume of the components has great distinction with air, hence the thermal calculation method in air needs to be modified in order to be used in O2/CO2 recycle mode. The modifications are mainly caused by the different components volume of the flue gas and oxidant, such as the calculation formula of density and enthalpy of flue gas and oxidant, the variation of heat transfer caused by the change of physical parameters et al. Among all of the changes the heat transfer is the most influential, including radiant and convection heat transfer [14,15].
4 Results and Analysis The oxidant in O2/CO2 recycle mode is composed of recycle flue gas and pure oxygen, when operating in recycle mode the newly rejoin gas is only oxygen and the amount of flue gas is controlled by recycle rate, the recycle rate also decides the combustion atmosphere in O2/CO2 recycle mode. In order to stabilize the operation of the boiler in certain recycle rate, it needs to find out each component volume of gas in a state of equilibrium in this certain recycle rate. This paper has set up corresponding oxidant-flue gas cycle calculation program according to different O2/CO2 recycle mode. In the cycle calculation program the basic calculated unit is the component of gas, so each component volume can be attained when the state of equilibrium will be achieved.
4.1 Definition of Oxygen Purity, Excessive Oxygen Coefficient and Air Leakage Coefficient The relation between oxygen purity and cost of the ASU is inverse, the higher the purity the higher the cost, but it will be harmful for the separation of CO2 if the purity is too low. The oxygen purity chosen here is 95% after having considered the factors above. The excessive oxygen coefficient chosen here is 1.05[11]. The N2 volume in flue gas is greatly influenced by air leakage coefficient in O 2/CO2 recycle mode. Fig.4 is the relation between air leakage coefficient and N2 volume in flue gas in a state of equilibrium in O2/CO2 recycle mode when the recycle rates are 0.7 and 0.8, respectively. It can be found out that the N2 volume is linear increased with the increasing of air leakage coefficient and the linear
slope increase with the increasing of recycle rate. The linear slope is up to 209 and 322 respectively by fitting the straight lines, and the N2 volume is up to 44.45% and 67.06% when air leakage coefficient is 0.2 (approximates the air leakage coefficient in air). The excessive high of N2 volume obviously is not the original intention of O2/CO2 recycle combustion technology, because this will enhance the difficulty of separating CO2. The air leakage coefficient chosen here is 0.02, under which the N 2 volume can be controlled within 10%, approximately. 32
70
recycle rate 0.7 recycle rate 0.8
50
Oxygen volume in oxidant/%
N2 volume in flue gas/%
60
40 30 20 10 0 0.00
30
cold recycle mode hot recycle mode
(0.633,30)
28
26
24
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0.62
Air leakage coefficient in cold recycle mode
Fig.4 Relation between air leakage and N2 volume in flue gases
0.64
0.66
0.68
0.70
Recycle rate
Fig.5 Relation between oxygen volume in oxidants and recycle rate
4.2 The Adiabatic Temperature and Boiler Exhaust Temperature Fig.5 is the relation between oxygen volume in oxidant and recycle rate, the trend of this two recycle mode is linear and the lines are basically coincided with each other, when the oxygen volume in oxidant is 30%, the recycle rate is 0.633. Fig.6 is the relation between adiabatic temperature and boiler exhaust temperature of different recycle rates in cold and hot recycle mode. As can be seen from the figure, in both recycle modes the adiabatic temperature decreases with the increasing of recycle rate, and when the recycle rates are 0.679 and 0.683 respectively the theoretical combustion temperature equal to that of in air. This is mainly because small recycle rate generates less flue gas, but fuel consumption changes little along with the change of recycle rate. Although the specific heat capacity of CO2 is higher than N2, the specific heat capacity impact can not offset the impact of the decreased flue gas, resulting in the lower adiabatic temperature in recycle mode compared to that in air; the higher CO2 specific heat capacity start to work when the recycle rate increase gradually and decrease the adiabatic temperature. The boiler exhaust temperature increase first and decrease later along with the increasing of recycle rate in both recycle mode, but the maximum temperature is basically the same as that in air, the reason of this is very similar to that of adiabatic temperature, when adiabatic temperature is high, radiant heat transfer will be strengthened but the amount of flue gas will be few. On the contrary, when the adiabatic temperature is low, radiant heat transfer will be decreased but the amount of flue gas will be increased, which is a counter-balance, interactive process.
2400
2400
adiabatic temperature in cold recycle mode boiler exit temperature in cold recycle mode
adiabatic temperature in hot recycle mode boiler exit temperature in hot recycle mode
2200
2200
2000
(0.679,2034)
1800
adiabatic temperature in air
1100
boiler exit temperature in air
(0.683,2034)
1800
Temperature/℃
Temperature/℃
2000
(0.706,1061) 1050
adiabatic temperature in air
1600
boiler exit temperature in air
1080
(0.759,1061)
(0.700,1061)
1060 1040 1020
1000 0.60
0.62
0.64
0.66
0.68
0.70
0.72
1000 0.60
0.74
0.62
0.64
0.66
0.68
0.70
0.72
0.74
0.76
0.78
Recycle rate in hot recycle mode
Recycle rate in cold recycle mode
a) cold recycle mode b) hot recycle mode Fig.6 Comparison of adiabatic temperature and boiler exhaust temperature in different recycle mode
4.3 H2O volume in flue gas and CO2 volume in flue gas after condensation Fig.7 is the comparison of H2O volume in different recycle mode and in air, it can be found that in hot recycle mode is above 10% higher than in air and in cold recycle mode the increment is about 5%. This is because of the accumulation of H2O in recycle mode. Part of the flue gas is recycled after high temperature precipitator without getting through the condenser in hot recycle mode, which results in higher H2O volume in oxidant when compared to that in cold recycle mode, ultimately increases the H2O volume in flue gas. Fig.8 shows the CO2 volume in flue gas after condensed in different recycle mode and recycle rate, in hot recycle mode the CO2 volume is up to about 89%, while in cold recycle mode is only about 86%. The higher CO2 volume in hot recycle mode is because the higher H2O volume in hot recycle mode which can reduce the impact of N 2, so after the condensation of flue gas the effect is reflected in the increasing of CO2. 89
CO2 volume in flue gas after condensation/%
20
Steam volume in flue gas/%
18
cold recycle mode hot recycle mode
16 14 12
10
air
8 0.61
0.62
0.63
0.64
0.65
Recycle rate
Fig.7 Comparison of H2O volume in flue gas after condensed
88
87
air recycle rate 0.614 recycle rate 0.633 recycle rate 0.648
86
85 15 10 5 0
cold recycle mode
recycle rate
hot recycle mode
Fig.8 Comparison of CO2 volume in flue gas
4.4 Amount of Flue Gas, Exhaust Heat Loss and Boiler Efficiency The air leakage coefficient is usually reached 20% or more in air, but in the O2/CO2 recycle mode the coefficient must be controlled in a rather small range, so the amount of flue gas in air is 40% higher than that in recycle mode; The exhaust is just a little in recycle mode, for most of the flue gas is recycled, thus reducing the exhaust heat loss to about 2%, almost 4% lower than that in air, as shown in Fig.9. The reduction of exhaust heat loss raise the boiler efficiency from about 92% in air to above 95% in recycle mode, and at the same time reduce the boiler fuel consumption.
6
Exhaust heat loss/%
5
air recycle rate 0.614 recycle rate 0.633 recycle rate 0.648
4
3
2
1
0 cold recycle mode
hot recycle mode
Recycle rate
Fig.9 Comparison of exhaust heat loss in difference recycle mode
4.5 Flue Gas Enthalpy and Heat Absorption of Every Part of the Boiler Fig.10 is the relation between flue gas enthalpy and flue gas temperature in different recycle mode. When the recycle rate is 0.633 (oxygen volume is 30%), the changing curve of flue gas enthalpy is coincident to that in air. In this recycle rate the heat transfer of flue gas rate is more closed to that in air. The results show that the amount of hearth heat absorption accounted for the proportion of the total boiler heat absorption is 62% in air, while the proportion in recycle mode is a few percentage points higher than that in air. This is because after N2 is replaced by CO2 the flue gas total emissivity and the adiabatic temperature will increase, thus increasing the radiant heat transfer of the furnace. As the radiant heat transfer has a dominant place in heat transfer in furnace, it will increase the heat absorption in furnace. The increment of heat absorption in furnace will inevitably lead to lower heat absorption in economizer. For a certain boiler, the absorbed heat that makes the feedwater change into steam can be divided into pre-heat, heat of vaporization and overheat. Usually the ratio of this three part is fixed, but when the ratio changes, it will often lead to operate unstable and even cannot operate normally, so in recycle mode the ratio should remain in a relatively unchanged level. The increasing heat absorption in furnace reduces the heat absorption in other parts of the boiler accordingly, the reduction of the economizer pre-heat absorption will enhance the gap between the economizer exit temperature and the saturation temperature in steam drum, which is harmful to the operating stability of a boiler. 25000
25000
air recycle rate 0.614 recycle rate 0.633 (oxygen volume30%) recycle rate 0.648
15000
air recycle rate 0.616 recycle rate 0.633 (oxygen volume30%) recycle rate 0.648
20000
Enthalpy of flue gas/kJ/kg
Enthalpy of flue gas/kJ/kg
20000
10000
5000
15000
10000
5000
0
0 0
500
1000
1500
2000
Temperature of flue gas in cold recycle mode/℃
2500
0
500
1000
1500
a) cold recycle mode b) hot recycle mode Fig.10 Relation between flue gas enthalpy and temperature in different recycle mode
4.6 The Total Water of the Spray-type Attemperator
2000
Temperature of flue gas in hot recycle mode/℃
2500
The role of the spray-type attemperator is to regulate the temperature of superheated steam. Due to the limited range and small amount that it will affect the safety of the boiler operation; besides, as the amount is too large, it will not only affect its safety but also will be wasteful. Hence, usually the total spraying water is about 3% of the boiler capacity (in this paper the capacity is about 6000 kg/h). As can be seen from Fig.11, the total spraying water increase in both recycle rate along with the increasing of recycle rate, the total spraying water in hot recycle mode is lower than that in cold recycle mode in the same recycle rate but the difference is not too much. When the recycle rate is 0.633(oxygen volume is 30%), the total spraying water is about 3% of boiler capacity, in this recycle rate a more reasonable temperature reduction effectiveness will be achieved by using the spray-type attemperator.
锅炉减温喷水总量/kg/h
20000
cold recycle mode hot recycle mode
15000
10000
recycle rate 0.633 oxygen volume 30%
5000
0 0.61
0.62
0.63
0.64
0.65
0.66
0.67
0.68
Recycle rate
Fig.11 Relation between total spraying water and recycle rate
5. Conclusion The combustion atmosphere in O2/CO2 recycle mode is mainly controlled by regulating recycle rate. In recycle mode, the increment of tri-atom gas volume (mainly CO2 volume) and boiler efficiency are very obvious, the reduction of the exhaust heat loss is obvious as well; besides, the fuel consumption decreases a little, the furnace blackness increases because of the increment of the emissivity of flue gas; the air leakage coefficient has great effect on CO2 volume; adiabatic temperature will decrease as the recycle rate increases and equal to that in air when recycle rate is 0.679 (cold recycle mode) or 0.683 (hot recycle mode). The boiler structure remains basically unchanged except a few apparatus are added in cold recycle mode. Though the total results meet the requirements, some of them are manifestly unreasonable in practical operation. For example, the heat absorption of the economizer is decreased which result in much lower exit temperature and make the thermal stress of junction with the steam drum too excessive, affecting the boiler’s safety. Meanwhile, the exhaust gas temperature will be reduced to about 100℃ due to the enhancement of convection heat transfer, which is manifestly unreasonable. The original boiler economizer and air pre-heater have been changed a lot in hot recycle mode. The changed version will adjust better according to different recycle rate. Therefore, hot recycle mode is the better choice to O2/CO2 combustion technology. Through comparing two recycle modes in different recycle rate, it is found that it can get the best results when in hot recycle mode and the oxygen volume in oxidant is about 30%(recycle rate is 0.633).
Acknowledgement:
The authors express their great thanks to the National Natural Science Foundation of China (Grant No. 50704017, 50936001, 50721005), and National Basic Research Program of China (“973” Program) (Grant No. 2006CB705806).
Refrences: [1] Zheng CG, Zheng Y, Li F, et al. Greenhouse effect and its control measures [M]. Beijing, China Electric Power Press, 2001. [2] Zhang LQ, Huang ZJ, Zou C, Zheng CG. Study on changes in pore structure of pulverized coal combustion in O 2/CO2 Atmosphere. International Pittsburgh Coal Conference,2007,Johannesburg,South Africa, September 10-14, 2007. [3] Zhou C, Huang ZJ, Chu K, Gui XL, Qiu JH, Zhang LQ, Zheng CG.A pilot scale study on SO2 and NOx emission control in O2/CO2 recycled coal combustion. Proceedings of the Chinese Society for Electrical Engineering, 2009, 20(2): 20-24. [4] Payne R, Chen SL, Wolsky AM, Richter WF. CO2 recovery via combustion in mixtures of oxygen and recycled flue gas [J]. Combustion Science & Technology, 1989, 67: 833-40. [5] Kimura N, Omata K, Kiga T, Tanako S, Shikisima S. The characteristics of pulverized coal combustion in O2/CO2 mixtures for CO2 recovery [J]. Energy Conversion & Management. 1995, (36): 805-8. [6] Kiga T, Tanako N, et al. Characteristic of pulverized-coal combustion in the system of oxygen/recycled flue gas [J]. Energy Conversion & Management. 1997, (38): S129-134. [7] Liu H, Ramlan Z, Bernard M G. Comparisons of pulverized coal combustion in air and in mixtures of O 2/CO2[J]. Fuel, 2005, 84(7-8): 833-840. [8] Klas A,Robert J, Stefan H, Filip J, Bo L. Radiation intensity of lignite-fired oxy-fuel flames[J].Experimental Thermal and Fluid Science. 2008, 33(1): 67-76. [9] Stefan H, Klas A, Filip J, Bo L. Combustion characteristics of lignite-fired oxy-fuel flames[J].Fuel.2009, 88(11): 2216-2224. [10] Buhre B J P, Elliott L K, et al. Oxy-fuel combustion technology for coal-fired power generation [J]. Progress in Energy and Combustion Science,2005,31(4):283-307. [11] Wal1 T. Combustion processes for carbon capture [J].Proceedings of the Combustion Institute, 2007: 31-47. [12] Wall T, Yinghui Liu, et al. An overview on oxyfuel coal combustion-State of the art research and technology development [J]. Chemical Engineering Research and Design,2009,87(8): 1003-1016. [13] WALL Terry, TU Jianglong. Coal-fired oxyfuel technology status and progress to deployment [C].34th International Technical Conference on Coal Utilization & Fuel Systems, Florida, USA: Coal Technology Association,2009. [14] Leckner B. Spectral and total emissivity of water vapor and carbon dioxide [J]. Combustion and Flame. 1972, (19): 133-148. [15] Feng JK, Shen YT, editor in chief. Principle and calculation of boiler[M].Beijing, Science Press. 1992.7.
Coal:Biomass Gasification - A Pathway for New Technology Development of Oxygen Blown Co-fired Gasification with Integrated Electrolysis Presenting Author: Dr Tana Levi, Technology Operations Manager,
[email protected] CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand A I Gardiner,
[email protected], Manager Hydrogen and Distributed Energy, Industrial Research Limited, 5 Sheffield Crescent, Bishopdale, Christchurch 8053, New Zealand R. Whitney, CEO,
[email protected] CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand Y Iwasaki, Scientist,
[email protected] CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand S Pang, Professor,
[email protected] Department of Chemical and Process Engineering, University of Canterbury, Christchurch, New Zealand . Q Xu, PhD Student,
[email protected] Department of Chemical and Process Engineering, University of Canterbury, Christchurch, New Zealand . and The Late Dr Tony Clemens CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand
This paper is dedicated to the memory of Tony Clemens who sadly recently passed away. At the time of his death Tony was a highly respected and greatly esteemed energy scientist. He had a strong belief that it was important the world have alternative, sustainable energy technologies available and led the Government supported project to build a gasifier to produce hydrogen from coal and biomass. Over the years Tony has won several awards at the Pittsburgh International Coal Conference, the most recent in 2009 as a member of a team for their scientific paper on the atomic level processes that occur when carbon dioxide is injected into various coal types. Tony died unexpectedly on 19 February 2010, he was a valued colleague and friend and his dedication to research and his helpfulness to fellow scientists will continue to be an inspiration to us all.
Abstract: This paper describes progress in new research that builds on the air blown gasification of lignite for hydrogen production technology that has been developed over the past several years. The new technology package is designed to assist New Zealand meet the challenges of peak oil and global climate change. The programme uses the results from gasifying a range of pellet blends of lignite or sub-bituminous coal with P.radiata or E.nitens, in bench scale and then small pilot scale gasifiers, to develop a fundamental understanding of the chemical processes underpinning gasification behaviour of New Zealand lignite and coal / woody biomass blends. Significant attention is given to the development of a new technology comprised of oxygen blown co-fired gasification with the integration of high efficiency electrolysis for production of low carbon footprint syngas and/or and hydrogen and ultimately synfuels from New Zealand’s coal, biomass and renewable electricity resources.
Syngas produced can be used for Fischer Tropsch production of liquid transport fuels. If shown to be successful the new technology would offer: i) Low carbon footprint transport fuels for the present and hydrogen for future transport fleets, ii) Increased flexibility in
tailoring syngas for Fischer Tropsch production, iii) A new option for matching the uptake of New Zealand intermittent renewable generation (including wind, marine and hydro) with a new deferrable load in the form of stored hydrogen, and iv) in the extreme case of critical global oil constraints the possibility of protecting New Zealand’s energy security by providing an alternative hydrogen feedstock for hydro-cracking of low grade oil supplies.
1.
Introduction
As a result of environmental and other policy considerations, there is increasing world-wide interest in the use of biomass resources as feedstocks for producing power, fuels, and chemicals. Biomass resources are a major component of strategies to mitigate global climate change. Plant growth recycles CO2 from the atmosphere, and the use of biomass resources for energy and chemicals results in low net emissions of carbon dioxide. The emissions of NOx and SOx from biomass facilities are also typically low. The use of these locally produced energy resources also results in new markets for agricultural and forestry products and provides a mechanism for rural economic development[1].
Gasification technologies provide the opportunity to convert renewable biomass feedstocks into clean fuel gases or synthesis gases. These gaseous products can be burned to generate heat or electricity, or they can potentially be used in the synthesis of liquid transportation fuels, hydrogen, or chemicals. Gasification offers a combination of flexibility, efficiency, and environmental acceptability that is essential in meeting future energy requirements.
Currently, the majority of biomass-based power generation plants are small and relatively inefficient, (< 2 MW) with efficiencies generally in the ~20-25% range. These plants use several feed stocks such as agricultural wastes, landfill methane, wood wastes in various energy conversion devices such as steam turbines, internal combustion engines, and, more recently, fuel cells. Larger biomass combustion systems used in industrial plants, generally in the wood products industry, and in municipal solid waste disposal can range from 5 - 75 MW, but again are not very efficient. With these factors in mind, research programmes to identify methods to more efficiently use the biomass resources to generate power have developed[2] , requiring that gasification of biomass be integrated with the power plant into a system that offers significant improvements in thermal efficiency and environmental performance. The cost of power from these facilities must be competitive with local circumstances.
When looking at biomass conversion it is instructive to look at coal conversion, as there are many similarities. For coal gasification the minimum temperature required is about 900oC. About the same maximum temperature of 800-900oC is required to gasify the most refractory part of almost any biomass, i.e the temperature required for the complete thermal gasification of biomass is of the same order of magnitude as for coal. This high temperature, in combination with the impurities present, whether sulphur or ash components, is why indirectly heated coal and biomass gasification processes in which external heat has to be transferred via a metal surface have not yet achieved any commercial success[3].
There are however a number of significant differences between coal gasification and biomass gasification, which are directly attributable to the nature of the feedstock. These include the quality of biomass ash, which has a comparatively low melting point but
in the molten state is very aggressive; its reactivity; and particularly with vegetable biomass, it’s fibrous characteristics. Finally, particularly in the lower temperature range, biomass gasification has a very high tar make[3].
Although an entrained-flow process might seem an attractive option for generating a clean, tar-free gas as required for chemical applications, where the low melting point of the ash would keep the oxidant demand low, there are problems with this approach. The aggressive quality of the molten slag negates such a solution, whether using a refractory or a cooling membrane for containment protection. Furthermore the short residence times of entrained-flow reactors require a small particle size, to ensure full gasification of the char. No method of size reduction has yet been found, which will perform satisfactorily on fibrous biomass.
A number of fixed-bed processes have been applied to lump wood, but they are limited to this material. They would not work on straw, miscanthus or other materials generally considered for large-scale biomass production unless these were previously briquetted. Furthermore in a counter-flow gasifier, the gas would be heavily laden with tar. The alternative of co-current flow could reduce the tar problem substantially, but the necessity to maintain good control over the blast distribution in the bed restricts this solution to units of very small size.
With this background it is probably not surprising that most processes for biomass gasification use fluidized beds and aim at finding a solution to the tar problem outside the gasifier. In co-firing applications where the syngas is fired in an associated large-scale fossil fuel boiler, the problem can be circumvented by maintaining the gas at a temperature above the dewpoint of the tar. This has the added advantage of bringing the heating value of the tars and the sensible heat of the hot gas into the boiler. There are many biomass processes at various stages of development[3]. Recent research shows increasing interest in co-gasification of biomass with coal[4] and for large scale applications, blending coal with biomass is becoming a preferred option. It not only reduces the cost of the raw materials for energy supply but also increases the overall conversion of coal and increases the calorific value of the feedstock. The blended feed is also considered as a more sustainable energy resource due to its reduced global environmental impact.
Some of the benefits of using oxygen instead of air in the gasification process are: the heating value of the syngas is higher, the volume of gas is approximately half that required for an air blown system for a similar amount of gasification energy, gas handling and cleanup units can be smaller, as can the heat exchangers. The advantages of using pure or enhanced oxygen content in the combustion process are well recognized. However the energy costs associated with producing oxygen in an air separation unit are in the order of 15-25% of the electrical output or 5-10% of the total energy input[5] . This has kindled interest in the use of electrolysis to produce oxygen [6,7] for the gasification process. In this study[6] electrolysis is used to produce oxygen at 64% Higher Heating Value (HHV) efficiency and the study acknowledges that efficiency can be improved and costs lowered further. Another study[7] proposes methanol production from a combination of energy resources including electrolysis to produce oxygen for gasification and hydrogen to increase the H:CO ratio.
The benefits of and problems with existing electrolyser technology are well known. Electrolysis provides an elegant and relatively simple means of producing high purity oxygen and hydrogen in a ratio of 1:2, but the technology is expensive, the cost of feedstock (electricity) is high, and the production efficiency is at present of the order of only 60% HHV. Recent advances in materials technology can potentially reduce each of these barriers. Combined with the changing environment of electricity supply, improvements could alter the economics of electrolysis in applications where both oxygen and hydrogen have a high value. In[8], a number of hydrogen production technologies including electrolysis are assessed for potential cost reduction based on technology learning rates. For electrolysis this is estimated as 18%, which is substantial, and similar to growth energy technologies such as photovoltaics and wind.
A lot of effort is applied within conventional gasification processes to enhance or upgrade the hydrogen content of the syngas. Coproduced hydrogen from electrolysis provides a potential feedstock for more flexible blending or upgrading, reducing the criticality of the water gas shift process. Pilot scale integration of electrolysis, and analysis of how it fits into the overall co-production process to improve syngas quality is an aspect of the research. This approach is particularly relevant to New Zealand since there is already a high content of renewable resources in the electricity mix (60-70% depending on rainfall patterns) and intermittent wind energy will progressively become a significant component of the generation capacity. An overview of the process is shown in Figure 1.
Hydrogen-syngas co-production diagram Lignite and biomass energy sources
Advanced Cogasification
Flexible syngas production for low GHG content F-T liquid fuels
Peaking electricity gen. Chemical energy buffer storage
Intermittently available renewable electricity eg wind
O2
H2
Electrolyser Plant
WGS and CO2 Capture
Carbon sequestration
Hydrogen fuel for FCVs and stationary use
Figure 1. The overall co-fired co-production gasification-electrolysis process.
Hydrocarbon feedstock is applied to the gasifier. Hydrogen and oxygen are produced in the electrolyser at an average rate sufficient to supply the gasifier with oxygen. If the electrolyser has a relatively fast response and overdrive capacity, buffer storage for the product gases could be provided and use of high cost peak electricity could be reduced. The buffer store of hydrogen is effectively a fuel gas, as is the hydrocarbon feedstock. In a large plant, and if the payback could justify the investment, fuel (in the form of hydrogen enriched synthetic gas) could be burnt in a gas turbine for peaking electricity production. The combined system represents a
form of flow battery, with hydrogen and oxygen as the working fluids. The main emission from the plant is carbon dioxide, which along with the carbon monoxide in the syngas contains both the biomass and lignite sourced carbon. If the plant carbon dioxide emissions are sequestered, then the ratio of the biomass/lignite carbon input, the emissions factor for the electricity production, and the plant efficiency all affect how “green” the resulting syngas is.
2.
Initial Objectives
Before the full pilot scale fluidized bed gasifier and electrolysis co-production plant can be brought together and evaluated as an integrated system several individual objectives must be implemented. These are: •
Develop a process in which pellets of a known coal:biomass ratio could be made
•
Conduct fundamental studies on the char activity and catalytic effect on the co-gasification of coal:biomass pellets
•
Develop a model of the co-gasification processes
•
Understand how these pellets perform in an already commissioned 50kw air or air with 30% O2 blown fluidized bed gasifier
•
Analyse electrolysis needs and develop an electrolysis process matched to the requirements of the integrated system
•
Construct and commission a pilot scale oxygen-hydrogen production plant
•
Design, build and commission an integrated O2 blown fluidized bed gasifier where the O2 is provided by an electrolyser.
3.
Experimental Details
3.1
Pellitizing
Prior to pelletizing the coal and biomass in the required percentages the raw materials had to be prepared. Figure 2 shows some of the raw materials used. Preparation of E. nitens was time consuming and required air drying prior to several passes through a knife mill, followed by several hours in a ball mill to break it to a size where it was possible to pass it through the pellitizer. P. radiata, lignite and sub-bituminous coal did not present such a problem and only required ball milling in the appropriate percentages. Once all components were < 1mm in diameter the pelletization procedure could proceed. This consisted of the milled feed being passed through the pellitizer twice. The first pass heated the feed, this was found necessary for the pellets to bind, and the second pass formed compact pellets of the required size for fluidization. Figure 3 details some of the equipment used to prepare the pellets. The pellets need to be robust enough to be fed into a 50kw gasifier via a drop lock hopper and auger. To meet this criteria the pellets were subjected to a quality control “drop-shatter” test.
a) E. nitens Figure 2. Starting materials for pellitizing.
b) P. radiata
c) lignite
a) Ball mill
b) Pellitizer
c) 20% lignite:80% P. radiata pellet
Figure 3. Some of the equipment used during the pellitization processes.
3.2
Bench scale gasifier
The bench scale gasifier comprised a vertical tube into which char samples derived from the pelletised blends were placed and exposed to temperatures ranging from 700 to 950°C whilst being subjected to a gas flow containing entrained steam. The assembly is integrated with an online gas chromatograph for real time analysis of the syngas produced and accurate determination of carbon consumption. This design recognises the fact that while gasification is a complex process involving a number of stages including devolatilisation, the rate determining step, and therefore the factor of critical importance, is most commonly the endothermic reaction between the char arising from devolatilisation and steam: C + H2O => CO + H2. Samples of char (1.00g) were placed into a small, bench scale gasifier, under a flow of nitrogen and then heated to a pre-selected temperature. Chars derived from pellets comprised of E. nitens or P. radiata with (0 , 20, 50, 80 or 100%) added lignite or subbituminous coal were used. After stabilizing at the selected temperature, degassed, distilled water was injected at 1.77 – 1.81 ml/min through a peristaltic pump into the heated nitrogen stream and passed through the char. After exiting the bench scale gasifier, the gas mixture passed through a condenser assembly, drying tower and particulate filters before being drawn into an MTI M200 gas chromatographic gas analyzer. The concentrations of product gases (hydrogen, carbon monoxide, methane and carbon dioxide) were measured. Samples were withdrawn and measured every 90 seconds. Gasification was continued until at least 60% carbon was consumed. 3.3
50kw gasifier
The gasification suite is made up of several integrated modules as shown in Figure 4 and Figure 5. The main modules of interest for this work are are the gasifier, cyclone, venture scrubber and gas cleaning tower.
In the first stage of development air is used as the oxidation medium, with higher content of oxygen progressively introduced later. The combustion mode can be started using a variety of feed from 100% lignite or sub-bituminous coal to a blend of 80% coal : 20% woody biomass pellets. A feed rate is maintained at an average of 4 - 5 kg/h and an air feed of 85 – 90 m3/h. – these parameters will
vary depending of feed stock. The bed depth remains relatively constant at just below 300 mm and the temperature generally settles at 900oC. Combustion was held in a stable state for 1 h prior to switching to gasification. A Delta V system is the main control system. Transition from combustion mode to gasification is normally easily achieved in a matter of minutes. This is done by increasing the coal feed to 17 – 19 kg/h and decreasing the air to 60 m3/h and adding a feed of 3kg/h steam. Temperature is generally maintained by adjusting the feed rather than the air or oxygen. Gas from the gasifier passes through a heat exchanger and cyclone prior to the venturi scrubber. Next it is further scrubbed in a counter current alkaline wash tower where H2S is removed completely (cannot be detected on 0.2ppm draegger tubes). Depending on the test requirements the gas is then either flared, or sent through the WGS and membrane systems. A more detailed description of the operation can be found in a previous publication[9].
Figure 4., Gasification suite (not to scale).
4.
Figure 5. CRL Energy 50kw Gasifier.
Results
The results detailed here are from the first phase of the work and only provide a brief overview of the results to date. Numerous experiments have been carried out over the course of programme, most of which were repeated several times for statistical validation and for use in the modelling. Only a small representation of these results are presented here.
4.1
Bench scale gasifier
The char consumption rates of the pure matrials and blends were ascertained. It was found that the time to 20% and 50% char consumption increased with increasing temperature. This was due to the increasing temperatures enhanceing the active kinetic energy of the reactions. The fastest rate for the 100% pure materials was observed for 100% lignite followed by E. nitens > P. radiata > subbituminous coal > acid washed lignite. For the blended chars it was found that adding lignite to E. nitens or P. radiata increased the reactivities at 850 and 900oC with no change observed at 950oC. In comparison, adding sub-bituminous coal to E. nitens or P.radiata decreased the observed char reactivities at all temperatures.
The effect of alkali and alkaline earth metallic (AAEM) in lignite gasification expressed as the gas production rate (L/s) is shown in Figure 6a for H2 and in Figure 6b for CO. From these Figures it is clearly seen that with the same gasification temperature, the peak values of both H2 and CO generation from lignite gasification were much higher than that from gasification of washed lignite (AAEM-free lignite). When the curves decayed exponentially from the peak value, AAEM-free lignite gasification took approximately twice as long to tend to zero than the gasification of original lignite.
CO production in gasification
H2 production in gasification
0.002000
0.007 0.0065
H2 (L 850)
0.001800
H2 (AWL 850)
CO (L 850)
CO (AWL 850)
CO (L 900)
CO (AWL 900)
CO (L 950)
CO (AWL 950)
0.006 0.001600
0.0055 H2 (L 900)
H2 (AWL 900)
H2 (L 950)
H2 (AWL 950)
0.001400
0.0045
CO amount (L/s)
H2 amount (L/s)
0.005 0.004 0.0035 0.003 0.0025
0.001200 0.001000 0.000800 0.000600
0.002 0.0015
0.000400
0.001 0.000200
0.0005 0
0.000000
0
1000
2000
3000
4000 Time (Sec)
5000
6000
a) Hydrogen
7000
8000
0
1000
2000
3000 4000 Tim e (Sec)
5000
6000
7000
8000
b) Carbon monoxide
Figure 6. The effect of AAEM in lignite gasification.
4.2
Modelling co-gasification
Some of the initial results obtained for the modeling work are described. The model developed for char gasification was firstly solved to predict the gasification rate and producer gas composition at various operational conditions and varying coal/biomass ratios. Gas phase equilibrium was used in calculating the simulated gas components. The resultant gas profiles and carbon consumption rate are then compared with the experimental data to validate the model. Some of the initial results are shown in Figure 7a for the gasification of pure wood and in Figure 7b for pure coal.
a) 100% E. nitens
b) 100% lignite o
Figure 7. Gas production profiles at 900 C.
4.3
50kw air blown gasifier
Good regular quality syngas can be produced from the system. The use of sized coal (+3mm, -10mm) or pellets improved steadiness of operation whilst the injection of steam into the bed slightly increased the hydrogen concentration in the syngas. Typical results for the gasifier running on lignite or sub-bituminous coal and biomass pellets are shown in Table1.
Table 1. Typical syngas compositions.
Fuel 100% lignite 80% lignite – 20% P.radiata 80% lignite – 20% E. nitens 100% subbituminous coal 80% subbituminous coal – 20% P. radiata 80% subbituminous coal – 20% E. nitens
5.
% Gas H2 12 9
CO 12.5 11
CO2
CH4
15
1.5
8
10
14
1.5
11
15
12
1
14
16
13
2
11
13
14
1.5
O2 blown fluidized bed gasifier and electrolyser
The anticipated H2:CO ratio from the gasifier (without WGS promotion) is 1.2:1. In order to upgrade it to 2:1, 0.8 additional mole of H2 is required per mole of CO. Oxygen production from the electrolyser provides 2 moles of hydrogen for each mole of oxygen used in the gasifier. This oxygen is used to produce heat by combusting some of the carbon in the fuel to CO2. The CO in the syngas mostly comes from injected steam. We estimate (using earlier results from the air blown gasifier) that 4Nm3/hr of oxygen is required per 100kW HHV of fuel. This provides 8 Nm3/hr of hydrogen, which based on the above raw syngas composition fairly closely matches the amount required to increase the molar ratio to 2:1. Hydrogen surplus to this requirement can be sold. The inclusion of a modest level of WGS would allow for larger amounts of surplus hydrogen.
5.1
Design of O2 gasifier
The new fluidized bed gasifier has been designed as a 50kw unit capable of running on 100% O2, air or a combination. It has also been designed to have the capability of using different feedstock from a high percentage of woody biomass to 100% lignite or subbituminous coal. A schematic of the system is shown in Figure 9. From the experience with the initial pilot gasifier this unit has been designed in a modular form so that maintenance, repair or installation of new units is easily achieved. The gasifier will run at ambient pressure and have a maximum temperature of 1150oC. As part of the government funded research programme the O2 will be supplied from an electrolyser designed and commissioned by Industrial Research Limited and described later in this paper.
tar removal system
he a t exchanger
cyclone b ypass line
lo c k hoppe r fe e d
venturi scrubber
existing systems
g a s i fi e r gas preh e a te r
LPG preh e a te r H2 electrolyser
N2
a ir
O2
flare
- counter flow caustic wash - WGS reactor - H2 separation membrane systems
s te a m
Figure 9. Schematic of the O2 fluidized bed 50kw gasifier with integrated electrolysis.
5.2
Electrolysis Requirements
For the purposes of this pilot system, the electrolyser must be able to respond to half hour electrcity pricing signals such as might result from load-following of the power output profile of a wind farm. It is required to produce both oxygen and hydrogen at low pressure. The estimated oxygen requirement for the pilot gasifier is 4 Nm3/hr O2. High overall electrolysis plant efficiency is vital to minimize feedstock cost. In a scale up implementation, low maintenance costs and durability are also very important. Basic electrolyser theory shows that the HHV electrical efficiency of an electrolyser can potentially be above 100%. At 20oC the required energy to split the water molecule to form the hydrogen and oxygen gases is made up from 83% electrical energy (the reversible potential) and 17% endothermic heat of reaction. In existing practice all of this heat (and more) is produced from electrical losses within the cell. Most of the loss is due to the sum of the electrode surface activation energy for both gas evolution reactions. For the electrolysis reaction in alkaline electrolytes the oxygen evolution reaction (OER) provides the majority. There are a number of other loss mechanisms including electrode and electrolyte ohmic conduction and bubble blocking that particularly impact on high current performance. To achieve a very low loss design these also cannot be ignored. Recent advances in materials fabrication and processing have raised the possibility of increasing practical electrolyser efficiency, without compromising durability and cost requirements. The project intention is to demonstrate and evaluate oxygen and hydrogen production based on an electrolysis process developed from first principles with these needs in mind.
5.3
Electrolyser Technology
It is not intended that the electrolyser technology developed for this pilot scale system will be scaled up to full production plant level. If the system concept proves to be attractive and can subsequently be shown to be commercially viable, international suppliers of large electrolyser technologies would most likely be engaged. The purpose of building this electrolyser technology is two fold – similar characteristics are required for an alternative application in a different field, and hands-on experimental development is a means by which to builds a deeper understanding of the potential for of electrolyser technology in other hydrogen energy applications. The planned first step is to produce a module operating at nominal 50Vdc, which will produce 0.4 Nm3/hr O2 (0.8Nm3 H2). This module is currently going through prototype testing and the main parts are illustrated in Figure 10. The module is fully self contained and produces oxygen and hydrogen at the required quality without further clean-up. A combination of design features has been introduced to achieve a wide operating range and fast turn-up and turndown, with very low peripheral power demand. The initial prototype achieves modest efficiency of 70% HHV without any special electrode surface preparation. This module is to now be used as a platform for an efficiency enhancement programme with a target module
Figure 10. Exploded assembly diagram showing the main parts in the integrated electrolyser module
level efficiency of over 80% HHV. The 4Nm3/hr O2 electrolyser design for the full pilot plant will consist of a scale up system based on replication of the prototype module.
6.
Summary
There is a significant role for coal to play in the future eceonomy of New Zealand. The research carried out to date within this programme has enabled a significant step in the coal to hydrogen proof of concept technology package to be developed. The facility has grown and developed and become a focus for other hydrogen research initiatives such as the integrated gasifier-electrolyser system. Future scenarios encompass the laboratory as being a test-bed for other hydrogen conversion technologies and gas clean-up options.
The cost of electricity has generally been seen as a barrier to the use of electrolysis for hydrogen production. If modestly priced electricity is available and electrolysis efficiencies are improved, electrolytically produced oxygen may be a realistic alternative to other methods of oxygen production such as air separation units, which also consume substantial amounts of electricity. This may be further supported by lower carbon emission costs from the process.
In the New Zealand context, the production costs appear competitive, and the H2:CO ratio appears suitable for F-T production of liquid fuels. The effects of lignite to biomass ratio on syngas composition and process carbon emissions is as yet unknown.
7.
Future Work
The overall programme raises several interesting energy system integration tradeoffs that will be explored later in the research. If for example there is a strong need to address the greenhouse gas impact of synthetic liquid fuels, by using CCS at the gasification plant it might be possible to sequester most of the fossil component of carbon, resulting in much of the carbon in the liquid fuel being biomass sourced. If the added hydrogen is renewable electricity sourced, the liquid fuel produced becomes relatively green.
8.
Acknowledgements
We are grateful to the New Zealand government through the Foundation for Research, Science and Technology for their investment in this research. We would also like to acknowledge the support and input from our colleagues at Industrial Research Limited and The University of Canterbury.
9. 1.
References Stevens, D.J., “Hot gas conditioning: Recent progress with larger-scale biomass gasification systems,” NREL/SR-510-29952 August 2001
2.
Liscinsky, D., “Biomass gasification and power generation using advanced gas turbine systems,” Final Report United Technologies Research Centre DE-FC26-01NT41354 September 2002
3.
Gasification, C. Higman and M. van der Burgt, Elsevier 2003
4.
Robert C. Brown, Qin Liu, Glenn Norton, “Catalytic effects observed during the co-gasification of coal and switchgrass”, 1999
5.
F. Starr, E. Tzimas, S. Peteves, “Critical factors in the design, operation and economics of coal gasification plants: The case of the flexible co-production of hydrogen and electricity”, Hydrogen Energy 32 (2007)
6.
P C Hulteberg, H T Karlsson, “A study of combined biomass gasification and electrolysis for hydrogen production”, IJHE 34 (2009)
7.
Katayama Y and Tamaura Y, “Development of new green-fuel production technology by combination of fossil fuel and renewable energy” , Energy, Vol30, Aug.-Sep. 2005
8.
K Schoots, F Ferioli, G J Kramer, BCC van der Zwaan, “Learning curves for hydrogen production technology: an assessment of observed cost reductions”, IJHE 33 (2008)
9.
Anthony Clemens, Tana Levi, Alister Gardiner, Ruben Smit and Jonathan Leaver, “Development and Testing of a Coal to Fuel Cell Grade Hydrogen Technology Package and a Pathway for Hydrogen Uptake in New Zealand”, Pittsburg Coal Conference, Pittsburg, USA (2008)
Manuscript Not AVAILABLE
HYDROGEN GENERATION FROM WATER BY USING PLASMA Beycan İbrahimoğlu1, İbrahim İbrahimoğlu1, Fırat Şen2, Şahika Yürek3 , Orhan Demirel4 1
1Anadolu Plazma Teknoloji Merkezi, Ankara Vestel Defence Industry R&D Department, Ankara 3 Türkiye Kömür İşletmeleri, Ankara
2
Abstract Plasma is the key to the development of new advanced technologies of producing hydrogen from different sources – water, hydrogen sulfide, a variety of hydrocarbons (including natural gas) and even coal. Plasma processes are characterized by extremely high specific productivity (more than 100 times in comparison with catalytic processes), low metal capacity and absence of inertia, they are ecology friendly. 21’st era’s the new energy carrier hydrogen can be produced from different sources. Today though different methods have been used for hydrogen production from water; it has not been produced in great amount and cheaply. Recently, to produce hydrogen from water, plasma method is used.
Keywords: Hydrogen, Hydrogen Generation, Plasma
E-mail:
[email protected]
2nd International Conference on Nuclear and Renewable Energy Resources 4-7 July 2010 Ankara TURKEY 1. Introduction Plasma is the fourth state of matter and in quasi-neutral field is formed of neutral and charged particles. It is possible to generate plasmas of solid, liquid and gasses by thermal, electromagnetic, laser and other methods. It can exist over an extremely wide range of temperature and pressure. It can be produced at low-pressure or atmospheric pressure by coupling energy to a gaseous medium by several means such as mechanical, thermal, chemical, radiant, nuclear, or by applying a voltage, or by injecting electromagnetic waves and also by a combination of these to dissociate the gaseous component molecules into a collection of ions, electrons, charge-neutral gas molecules, and other species [1]. It is more or less an electrified gas with a chemically reactive media that consists of a large number of different species such as electrons, positive and negative ions, free radicals, gas atoms and molecules in the ground or any higher state of any form of excited species [1]. Plasma is the key to the development of new advanced technologies of producing hydrogen from different sources – water, hydrogen sulfide, a variety of hydrocarbons (including natural gas) and even coal. Plasma processes are characterized by extremely high specific productivity (more than 100 times in comparison with catalytic processes), low metal capacity and absence of inertia, they are ecologically friendly. 2. Principle Of Hydrogen Generation From Water By Plasma Today, in order to produce hydrogen several methods are used [2]. However, these methods are not satisfactory to produce hydrogen in a cheap and easy way. Producing hydrogen from water is the most popular method among others [3]. In this manner, ecologically clean method, plasma is used as the hydrogen production method. Related to this topic, new test apparatus is developed from us as seen in figure 1. With the help of this test apparatus, every kind of water can be used in the hydrogen production with plasma method.
1. Propellant 2. Main Body 3. Spring 4. Valve 5. Quartz Pipe 6. Sponge
7. Anode 8. Cathode 9. Vapor Room 10. Nozzle 11. High Temperature Vapor Room 12. Magnet Figure 1. Test Apparatus
In this test apparatus, electric arc is produced between wire cathode and porous anode when electrical load is applied at this poles and very high pressure is obtained while water is evaporating. Pressurized water vapor changes state to the plasma and is ejected to the outside. This ejected matter contains hydrogen and oxygen ions and water vapor. With the help of the magnetic field, this ejected matter is splitted into the hydrogen and oxygen. 4. Conclusions
2
2nd International Conference on Nuclear and Renewable Energy Resources 4-7 July 2010 Ankara TURKEY As a result, great amount of hydrogen can be produced by using plasma method. Via this method hydrogen splitting process is fast and clean. References [1] Vijay N., Kumar, A. and Dwivedi, H. K., “Atmospheric Non-Thermal Plasma Sources,” International Journal of Engineering. 2 (1),(2008). [2] B. İbrahimoğlu, Ş.Yürek ve O. Demirel, (2009), “Plazma Yöntemi ile Kömürün Gazlaştırılması” IV. Ulusal Hidrojen Enerjisi Kongresi ve Sergisi (Uluslararası Katılımlı) 15–16 Ekim Kocaeli Üniversitesi, Kocaeli, Türkiye. [3] M. D. Kozlu, S.Aksongur, B. İbrahimoğlu, (2009), “Plazma Teknolojilerinin Hidrojen ve Yakıt Pilleri Üzerinde Uygulamaları”, IV. Ulusal Hidrojen Enerjisi Kongresi ve Sergisi (Uluslararası Katılımlı) 15–16 ekim 2009, Kocaeli Kocaeli Üniversitesi, Kocaeli, Türkiye.
3
Manuscript Not AVAILABLE
Technology and operational experience – the Shell perspective In our presentation we will discuss Shell’s experience of deploying SCGP in China, and how learning from our experience has enabled new levels of operational excellence. Up to date, Shell has sold nineteen licences in China, some of which are repeat customers. We will illustrate how we have learnt from experience on the ground and were able to make a wave of several start-ups a success: in one plant less than 12 hours elapsed between initial coal feeding and actual production of methanol products. We will show that we are not just focusing on deployment enablers for today, but continue to invest in both R&D and deployment capabilities that will enable the growth of coal gasification over the coming decades. We’re harnessing excellence in technology and design – especially in India and China – while constantly seeking out new opportunities to deploy gasification. Jay Wang Senior Engineer Shell Global Solutions International BV P.O. Box 38000, 1030 BN Amsterdam, The Netherlands Tel: +31 20 630 2024 Mobile: +31 65 209 7693 Fax: +31 20 630 3964 Email:
[email protected] Internet: www.shell.com/globalsolutions
CONTROLLING THE SYNTHESIS GAS COMPOSITION FROM CATALYTIC GASIFICATION OF HYPERCOAL AND COAL BY CHANGING THE GASIFICATION PARAMETERS
Atul Sharma* and Toshimasa Takanohashi Advanced Fuel Group, Energy Technology Research Institute, National Institute of Advanced Industrial Science and Technology, 16-1, Onogawa, Tsukuba, Ibaraki, JAPAN. Email: [email protected]
ABSTRACT. Catalytic gasification of coal is an efficient way to achieve high gasification rates at as low as 700 oC temperatures. The problem of deactivation of catalyst due to the interaction of catalyst with mineral matter in the coal was overcome by using HyprCoal, an ash less product of solvent extraction process as feed coal for catalytic process. Synthesis gas is the main desirable product and its composition H2/CO ratio is important for its use in the downstream FT process. However, in a catalytic gasification process it is difficult to control the gas composition because of the effect of catalyst on water-gas shift reaction. Effect of temperature and gasifying agent composition on gasification rate and synthesis gas composition were investigated. Experiments were carried out with pure steam, pure CO2 and mixture of steam and CO2 as gasifying agents in the 600~700 oC temperature range to investigate the effects. Results showed that by adjusting the steam to CO2 ratio of the gasifying agent it is possible to control the synthesis gas composition. Effect of CO2 addition on reaction kinetics was discussed along with the calculated gas compositions. A new single step process to produce a desired synthesis gas from catalytic gasification has been proposed. KEYWORDS. Coal gasification, Catalyst, Steam, CO2, Synthesis gas. 1
Introduction Gasification is a primary conversion route to produce synthesis gas (H2 and CO) from coal1-12. Gasification is an endothermic reaction and requires temperature above 1000 °C to achieve acceptable rates for commercial application1-6. The product gas obtained at such high temperatures usually has low H2/CO (0.5~0.7) and it is also difficult to control the product gas composition due to equilibrium constraints at such high temperature. Introduction of steam at some stage of the gasifier can increase H2/CO of the product gas but only marginally (~0.9). A more common process approach is to upgrade the synthesis gas from the gasifier to the desired H2/CO (~1, 2) by water gas-shift reaction at a lower temperature (300~400 °C) in a second stage often called as sweet shift process. Catalytic gasification of coal has been widely considered as an effective mean to decrease the gasification temperature8-12. Exxon mobile1 developed a K2CO3 catalyzed steam coal gasification process at high pressure to produce methane. Nearly all catalytic gasification processes developed were to produce either H2 or CH4. Not many attempts were made to produce synthesis gas by catalytic gasification12. This was partly because alkali catalysts catalyzed both gasification and water-gas shift reaction under atmospheric pressure condition leading to formation of H2 and CO2 as main products while higher pressure favors CH4 formation. The most favored catalysts are alkali metal salts especially K2CO3
1-16
. The major
drawback in catalytic gasification of coals is the interaction of catalyst with the mineral matter (ash) present in the coals leading to the formation of compounds from which recovery of catalyst is difficult1-8,12. To overcome the problem of loss of catalyst, our research group has developed a process to remove mineral matter from coal by solvent extraction13-17. The solvent extracted coal from hereon called HyperCoal (HPC) has less than 500 ppm of ash. Because of its ashless characteristics, a catalytic gasification process for coal may be 2
developed by using HyperCoal as a feed material leading to low gasification temperature, easy recovery and recycling of catalyst. In our first study13 we reported high gasification rates at temperatures as low as 775 °C, no catalyst deactivation, feasibility of catalyst recovery and recycling, and H2 selectivity from catalytic gasification of Oaky Creek HyperCoal. In a second study14, effect of catalyst addition on gasification reactivity of HyperCoal and parent coal was compared at 700 and 775 °C and it was found that HyperCoal and coal have nearly same rates at 700 and 775 °C but at different catalyst loadings. In the following study15, gasification rates of HyperCoals prepared from three different ranks of parent coals in the 600~775 °C range were compared. In a subsequent study16, effect of catalyst mixing procedure and catalyst loading ratio on gasification rate was examined. Production of synthesis gas (H2/CO) at such low temperatures would also be an attractive application for HyperCoal. Results of our previous studies13-16 showed that irrespective of gasification temperature and coal type, the product gas from the K2CO3-catalysed steam gasification of HyperCoal contained H2 and CO2 as the major products gases with little CO and was suitable for H2 production. In the most recent study17 we reported production of synthesis gas from catalytic gasification of HyperCoal at 700 °C by changing the steam partial pressure. In this study, we report production and control of H2/CO ratio of synthesis gas by gasifying coal and HyperCoal in steam and carbon dioxide mixed environment. Coal and HyperCoal were gasified in a steam and CO2 mixed environment as a gasifying agent at 700, 650 and 600 °C with K2CO3 as a catalyst. Effect of temperature and ratio of steam to CO2 on reaction rate and on composition of the gas produced was investigated. A single step process to produce and control the composition of synthesis gas from coal at 700~600 °C has been proposed. 3
Experimental Procedures A subbituminous coal, Pasir (PAS) from Indonesia was selected for the investigation. HyperCoal production method18 has been described in detail elsewhere. Briefly, HyperCoal (HPC) was produced by the solvent extraction of the coal with 1-methylnaphthalene at 360 °C and subsequently separating the extract (HyperCoal) from the solvent. The extraction yield was 51 % for Pasir coal. The properties of the Pasir coal and HyperCoal produced from Pasir coal are shown in Table 1. HyperCoal has nearly no mineral matters. Because of almost no mineral matter, all the inorganically associated sulfur will be removed. The only sulfur in HyperCoal will be the organically associated sulfur. A detailed characterization of catalyst, HyperCoal, original coal, and chars using XRD, NMR, and SEM-EDX mapping techniques had been carried out and reported elsewhere13,15-16. The experimental setup is shown in Figure 1. Samples for catalytic gasification experiments were prepared with 50 % catalyst loading. Catalyst loading was on dry and ash free wt % basis of coal and HPC. Both dry mixing and wet mixing methods were investigated16. Catalyst mixing method has been described in detail elsewhere16. Briefly, a desired amount of K2CO3 was added on the top of a measured sample already loaded into a test crucible as solid particles and stirred with a small spatula until white K2CO3 disappears by capturing moisture from the air. Common procedure is to mix K2CO3 as an aqueous solution for homogeneous dispersion7-11 but could not be applied as HyperCoal does not mix with water. The particle size of coal and HyperCoal sample was under 75 µm. The gasification experiments were carried out with and without K2CO3 as a catalyst at 700, 650 and 600 °C with different steam to carbon dioxide (H2O/CO2) ratios as gasifying agent. Experiments were carried out in a thermogravimetric (TG-DTA 2020S, MAC) apparatus with about 20 ml/min argon (Ar) as TG carrier gas flowing from the bottom 4
(Figure 1). At the start of the experiment, 100 ml/min Ar was mixed with 20 ml/min oxygen (O2) and flowed from the top into the TG-DTA. Pre-oxidation was required for HyperCoal samples because of their extremely high swelling propensity. To keep the same experimental conditions coal samples were are subjected to pre-oxidation. In actual process, pre-oxidation may not be necessary depending on the type of feeding system. A desired amount of water was pumped by a HPLC pump to a steam generator held at 250 °C. CO2 was flowed to the steam generator as a steam carrier gas. By changing the amount of water pumped by the HPLC pump to the steam generator and the flow rate of CO2 as the carrier gas, different steam to CO2 ratio were achieved. For pure steam gasification, argon gas instead of CO2 was used as carrier gas. In case of steam and CO2 only conditions, partial pressure of steam and CO2 were kept at 0.5 atm by mixing argon gas. A 4-way valve at the inlet of the TG-DTA was used to change (Ar+O2) flow to (CO2+steam) flow. The flow lines were kept at 250 °C by using ribbon heaters. First, a desired amount of sample was heated in (Ar+O2) flow up to 200 °C and held for 5 min to remove moisture and reduce the swelling propensity of the HyperCoal by mild pre-oxidation. After 5 min hold at 200 °C, the gas was switched to pure argon flow for 60 min to remove the O2 from the reaction zone. After 60 min hold, the sample was heated to the desired temperature at 20 °C /min in pure argon. When the desired temperature was reached without any hold time the pure argon gas was switched to the preset steam/CO2 gas mixture. The steam+CO2 mixture flowing from top comes into contact with the sample in the crucible. The evolved gases flow out together with the purge gas from the side into an ice cooled tar trap to remove tar before injecting to the micro gas chromatograph (Agilent 3000A). The total gas flow rate at the outlet was measured every 3 min by a film flow meter.
Results and Discussion 5
Figure 2 shows a typical weight loss curve for HPC+50 % K2CO3 sample pyrolyzed in argon up to 700 °C followed by gasification with [50% steam+50% carbon dioxide (vol/vol)] as gasifying agent (from hereon in this manuscript steam to carbon dioxide ratio will be addressed as H2O/CO2). In a previous study16 gasification rate and gas composition was investigated at 10, 20, 40, and 50 % catalyst mixing ratio. It was reported that gasification rate was affected by the catalyst amount up to 50 % loading and above this catalyst loading rates were almost independent of catalyst amount. In addition, unlike gasification rate, gas composition was not found to be affected by the catalyst amount. Therefore, in the present study only one catalyst mixing ratio, 50 % catalyst loading has been selected. In actual process 40-50 % catalyst loadings would be very high as it may cause problem to the access of reactant gas to the reaction surface at medium to high conversion level and lower loadings such as 20-30 % would be more appropriate. All experiments were carried out at atmospheric pressure. The weight loss curve can roughly be divided into three stages; moisture removal or drying stage, devolatilization stage and fixed-carbon gasification stage. The initial weight loss (up to 200 °C) during heating from room temperature to 200 °C in O2+Ar mixture is mainly due to moisture captured from air. Pre-oxidation was done to reduce the extremely high swelling propensity of HyperCoal. After pre-oxidation stage, sample was switched to 100 % argon for 60 min. The coal/HPC conversion on dry, ash, catalyst and volatile free basis (dacvf) (from hereon called char conversion) was calculated during the fixed-carbon gasification stage by the following equation: X (char conversion, % dacvf) =
W0 − W × 100 W0 × (1 − Wash − Wcat )
[1]
where W0 is the weight when the gasification begins (db, mg) (weight at t= 45, 41 and 39 min for T= 700, 650 and 600 °C, respectively), W is weight at any gasification time (db, mg, >39
6
min), Wash is weight fraction of ash content in coal or HPC, Wcat is weight fraction of catalyst content. Results of gasification rate and gas composition only in the char gasification stage will be discussed further. Figure 3 shows the gasification profiles and gas composition of produced gas from K2CO3 catalyzed HPC at 700, 650 and 600 °C as a function of H2O/CO2 ratio. In general, at any given temperature within the temperature range investigated, rate decreased with increasing CO2 fraction in the gas mixture. Similarly at any given temperature, H2 decreased and CO increased with increasing CO2 fraction in the gas mixture. Under H2O/CO2 mixed gas environment, three reactions as shown below: C-H2O reaction (1), C-CO2 reaction (2) and water-gas shift (WGS) reaction (3) are expected to take place. While the rate of carbon loss or gasification rate can be predicted primarily by reaction (1) and (2), for proper prediction of gas composition, inclusion of WGS reaction (3) is essential. This is because WGS reaction is one of oxygen exchange reactions which are known to be catalyzed by alkali-carbon system and therefore plays a major role in determining the gas composition12. It is well known that C-H2O reaction (1) is faster than C-CO2 reaction (2)2-5. If overall carbon loss due to gasification is assumed to be the sum of individual contributions of carbon loss by reaction (1) and reaction (2), a decrease in overall gasification rate with increasing CO2 fraction in the mixed gas can be expected due to the increased contribution of reaction (2) to the overall rate. The observed gasification trend in Figure 3 (a,c,e) is in accordance with this explanation. As mentioned before, reaction (3) is essential for predicting the gas composition. It is well known1-5,7-17 and also observed in this study that under pure steam (H2O/CO2=100/0) H2 and CO2 are the main gases and very little CO is produced. This is because the WGS reaction is catalyzed by alkali-carbon system resulting in conversion of nearly all CO produced by reaction (1) to H2 and CO2. In case of pure CO2 (H2O/CO2=0/100), because of absence of 7
steam WGS reaction does not take place and only reaction (2) occurs resulting in CO as main gas produced. The little H2 produced is mainly from the inherent hydrogen in the coal. However, in a H2O+CO2 mixed gas, all three reactions take place and gas composition is determined by these competing reactions12. Ideally, if reaction (3) is the only reaction controlling the gas composition, CO produced by reaction (1) and (2) is expected to be consumed by reaction (3) as WGS reaction can not differentiate between CO from reaction (1) or reaction (2). However, the observed gas composition showed that CO is produced and H2 decreased when CO2 is introduced and fraction of CO increased with increasing CO2 fraction in the H2O+CO2 mixed gas. This may be explained by considering [1] contribution of reaction (2), [2] change in equilibrium state of the reaction (3) on addition of CO2 and [3] steam becoming rate limiting reactant with increasing CO2 ratio. As discussed before, with increasing CO2 fraction the contribution of reaction (2) to overall reaction increases thus increase in CO and decrease in H2 due to decreased contribution of reaction (1) can be expected. However, the above explanation will be valid only if the three reactions are independent. Since reaction (3) is also taking place, only reaction (1) and (2) could be considered as independent but not reaction (3). Therefore, effect of CO2 addition on WGS reaction equilibrium should also be considered. Addition of CO2 may shift the equilibrium of WGS reaction (3), which has been reported to achieve the equilibrium state above 650 °C under catalyzed conditions12, towards left thus reducing the consumption of CO to CO2 and H2. In this case one can expect increase in CO and decrease in H2 which is observed in the present study. The reason [3] is also possible as partial pressure of steam is decreased with increasing CO2 fraction. In our previous report17, we reported the effect of steam partial pressure on gas composition. However, the partial pressure of steam below which any significant change in gas composition was observed (<0.1 atm) were far lower than those used 8
in this study (0.3 atm is minimum in this study). Therefore, the observed increase in CO and decrease in H2 with increasing CO2 fraction can be explained by considering the first two effects although reason [3] can not be ruled out. C
+
H2O
→
CO
C
+
CO2
⇌
2CO
CO +
H2O
⇌
H2
+
H2
[1] [2]
+
CO2
[3]
The effect of CO2 fraction in the gas mixture on the gasification rate and H2/CO ratio with temperature can be more appropriately discussed by plotting rate constant and H2/CO ratio against the H2O/CO2 ratio. For simple understanding and comparison of gasification reactivity, time needed to reach 50 % conversion (at 25% conversion where 50% conversion could not achieved) was obtained from the gasification profiles from Figure 3 (a,c,e). Rate constants were obtained by inversing this time value (except for 25% conversion where rate constant was obtained by [(1/(1-x))(dx/dt)]. The rate constants thus obtained were plotted against the CO2 fraction in the H2O+CO2 mixed gas. Similarly H2/CO (mol/mol) was obtained from Figure 3 (b,d,f) and plotted against CO2 fraction in the H2O+CO2 mixed gas. Figure 4a&b shows the effect of CO2 fraction in the H2O+CO2 mixed gas on the gasification rate and H2/CO ratio of the synthesis gas at 700, 650 and 600 °C. Gasification rate was affected by both temperature and CO2 fraction in the H2O+CO2 mixed gas. At a given temperature, gasification rate decreased with increasing CO2 fraction. Similarly, rate decreased with decreasing temperature at a given CO2 fraction. The extent of decrease in rate was large at lower CO2 fraction and becomes smaller as CO2 fraction increases as shown in Figure 4a. The results in Figure 4a suggest that an initial addition of CO2 lowered the observed gasification rate significantly but with further addition the rate was not so much affected. A possible explanation for could be that when CO2 was initially added most of the active sites in 9
the active alkali cluster would be chemisorbed by gas phase CO2 making fewer sites available for fast oxygen transfer reaction thus lowering the overall reaction12. With further addition of CO2, the decrease in rate due to above reason would not happen as most of the sites were already occupied by CO2. This further decrease in rate may be due to the increased contribution of the reaction (2) to the overall gasification rate. On the other hand unlike gasification rate, H2/CO ratio was primarily affected by CO2 fraction in H2O+CO2 mixed gas only and little or negligible effect was observed with temperature as shown in Figure 4b. For commercial application a high gasification rate or short gasification time is required. Considering the scale of operation of a coal gasifier typically about thousand tons per day, a long gasification time means large residence time leading to huge gasifier size in addition to additional energy requirements. Figure 5 shows a correlation between H2/CO ratio of synthesis gas, gasification time and H2O/CO2 ratio of the gasifying agent at 700, 650 and 600 °C. Takarada et al.11 suggested gasification rate of about 0.3 h-1 (18 min) at low temperatures such as 700 °C and below would be necessary for commercial application. From Figure 5 it can be seen that synthesis gas with H2/CO=1~3 can be produced in a single step from catalytic coal gasification at 700 °C with gasification time <20 min by changing H2O/CO2 ratio from 50/50~30/70. Figure 6 shows a schematic of such a process for producing synthesis gas from coal. HyperCoal is first produced from coal by HyperCoal production process not included in this process as a bench scale 0.1 ton/day plant is already operated by Kobe Steel Ltd18. The process flow shows that gasifier and combustor are separated. Combustor can be used to produce steam and supply necessary heat to gasifier by combusting HyperCoal residue. HyperCoal is mixed with catalyst and fed to gasifier where synthesis gas with desired H2/CO is produced in a single step by catalytic gasification of HyperCoal with H2O/CO2 as gasifying 10
agent. Catalyst is recovered from the bottom and recycled. As HyperCoal is used, catalyst will not be deactivated13. Coals especially with low ash contents such as Victorian brown coal can also be used instead of HyperCoal although their use will depend on the cost of the catalyst. The produced gas is then used as a feedstock for FT synthesis process to produce DME, methanol, methane and other chemicals. CO2 is produced as end product of FT synthesis process. CO2 is first separated and mixed with steam (H2O) before recycled to the gasifier as gasifying agent. The recycled CO2 can also act as fluidizing agent in case a fluidized bed gasifier is used. The proposed process is only a simplified overview of a new single step low temperature synthesis gas production and control process. There are several technical and operating challenges such a tar formation, heat supply to the gasifier, operating pressure of the gasifier, energy and commercial economics of the process, to list a few to be solved before a pilot scale process can be conceived. However, the results from the present study do suggest that there is a possibility to develop a new commercial process to produce synthesis gas as a feedstock for FT synthesis process from coal in single step at low temperature.
Conclusions Effect of gasifying agent on gasification rate and composition of product gas from catalytic gasification of coal and HyperCoal at 700, 650, and 600 °C was investigated. Following conclusions were drawn. 1. Synthesis gas can be produced from low temperature catalytic coal gasification by using H2O/CO2 mixture as gasifying agent. 2. Synthesis gas composition (H2/CO ratio) can be controlled in a single step by changing H2O/CO2 ratio of the gasifying agent. 3. Gasification rate was affected by both temperature and CO2 fraction in the H2O+CO2 11
mixed gas. H2/CO ratio was only affected by CO2 fraction in the H2O+CO2 mixed and little or negligible with temperature. 4. A new process to produce synthesis gas as a feedstock for FT synthesis process from coal in single step at low temperature was proposed. Gasifier and combustor must be separated in order to produce and control the synthesis gas composition in a single step. References 1.
Nahas, N. C. Fuel 1983, 62, 239-241.
2.
Veraa, M. J.; Bell, A. T. Fuel 1978, 57, 194-200.
3.
McKee, D. W.; Chatterji, D. Carbon 1975, 13, 381-390.
4.
Kayembe, A; Pulsifer, A. H. Fuel 1976, 55, 211-216.
5.
Wigmans, T; Elfring, R; Moulijn, J. A. Carbon 1983, 21(1), 1-12.
6.
Formella, K.; Leonhardt, P.; Sulimma, A.; van Heek, K. H.; Juntgen, H. Fuel 1986, 65, 1470-72.
7.
Miura, K.; Aimi, M.; Naito, T.; Hashimoto, K. Fuel 1986, 65, 407-11.
8.
Kwon, T. W. ; Kim, J. R.; Kim, S. D.; Park, W. H. Fuel 1989, 68, 416-421.
9.
Lee, J. W.; Kim, S. D. Fuel 1995, 74, 1387-1383.
10. Schumacher, W.; Muhlen, H. J.; van Heek, K. H.; Juntgen, H. Fuel 1986, 65, 1360-1363. 11. Takarada, T.; Tamai, Y.; Tomita, A. Fuel 1985, 64, 1438-42. 12. Meijer, R.; Kapteijn, F.; Moulijn, J. A. Fuel 1994, 73(5), 723-30. 13. Sharma, A.; Takanohashi, T.; Morishita, K.; Takarada, T.; Saito, I. Fuel 2008, 87, 491-97. 14. Sharma, A.; Takanohashi, T.; Saito, I. Fuel 2008, 87, 2866-2690. 15. Sharma, A.; Saito, I.; Takanohashi, T. Energy & Fuels 2008, 22(6), 3561-3565. 16. Sharma, A.; Kawashima, H.; Saito, I.; Takanohashi, T. Energy & Fuels 2009, 23(4), 1888-1895. 12
17. Sharma, A.; Saito, I.; Takanohashi, T. Energy & Fuels 2009, 23(10), 4887-4892. 18. Okuyama, N.; Komatsu, N.; Shigehisa, T.; Kaneko, T.; Tsuruya, S. Fuel Processing Technology 2004, 85, 947-967.
TABLE 1: Properties of coal and HyperCoal.
Coal
Ash (wt %, db)
Elemental analysis (wt % daf) C
H
N
S
O
Pasir coal
4.2
68.2
5.4
1.1
0.16
25.14
Pasir HyperCoal
0.01
79.5
5.9
1.1
0.29
13.21
13
Ar/O2 mixer
Purge gas /(Ar+O2) gas Steam generator 3-way valve
4-way valve
Gas out Flow meter
CO2
Sample
Ar
water Ice Trap
CO2
Ar
O2
HPLC pump
TG carrier gas Flow controller
TGA
µGC
Figure 1. Schematic of experimental set up for catalytic steam gasification of HyperCoal.
800
20
Weight [ mg]
16
700
Ar (100%) Steam + CO2 50 + 50
12
600 500 400
8
300
Temperature [C]
Ar + O2
200
4
100 0
0 0
30
60
90
120
150
Time [min]
Figure 2. Weight loss profile of HPC pyrolyzed in argon up to 700 °C followed by steam+CO2 gasification.
14
150
100/0
80
Gas yield, [mmol/g-char]
Char conversion [%, davcf]
100
30/70 50/50
60
70/30 40
0/100 (a) T=700
20
125 100 75 50 25 0
0 0
50
100
100/0
150
100/0
150
70/30
80
Gas yield, [mmol/g-char]
Char conversion [%, davcf]
100
50/50 30/70
60 40 0/100 20
70/30
50/50
30/70
0/100
Gasifying agent, H2O/CO2, [vol/vol]
Char gasification time, [min]
(c) T=650
125
(d) T=650
CO H2
100 75 50 25 0
0 0
50
100
100/0
150
150
100 Gas yield, [mmol/g-char]
100/0 80 70/30 50/50
60 40
30/70
20 0/100
70/30
50/50
30/70
0/100
Gasifying agent, H2O/CO2, [vol/vol]
Char gasification time, [min]
Char conversion [%, davcf]
CO H2
(b) T=700
(e) T=600
125
(f) T=600
CO H2
100 75 50 25 0
0 0
50
100
100/0
150
70/30
50/50
30/70
0/100
Gasifying agent, H2O/CO2 , [vol/vol]
Char gasification time, [min]
Figure 3. Effect of H2O/CO2 ratio of the gasifying agent on (a) gasification profiles and (b) gas yield of K2CO3 catalyzed HyperCoal at different temperatures.
15
6
700 0.32
650 (a)
700
600
5
650 (b)
600
4
0.24
H2/CO
Rate constant, [1/min]
0.4
0.16
3 2
0.08
1
0
0
0
25
50
75
100
0
CO2 fraction in (H2O+CO2) mixed gas, [vol %]
25
50
75
100
CO2 fraction in (H2O+CO2) mixed gas, [vol %]
Figure 4. Effect of CO2 fraction in the H2O+CO2 mixed gas on (a) gasification rate and (b) H2/CO ratio of the produced gas at different temperatures.
6 H2O/CO2=70/30 5
H2/CO
4
700 650 600
H2O/CO2=50/50
3 2
H2O/CO2=30/70
1
H2 O/CO2 =0/100
0 0
50
100 150 200 Gasification time, [min]
250
Figure 5. Correlation between H2/CO ratio, gasification time and H2O/CO2 ratio of the gasifying agent as a function of temperature.
16
Synthesis gas H2/CO = 1, 2, 3
Coal/HPC
5
6
8
DME Methanol Methane
2
Coal/HPC residue Steam
1
1. Boiler 2. Feeder 3. Gasifier 4. Steam/CO2 mixer 5. Heat recovery unit 6. Gas cleaning 7. Catalyst recovery 8. Compressor 9. FT process
3
4
CO2 Catalyst recycle
7
9
Recycled CO2
Water
Figure 6. Schematic process flow of a new low temperature single step coal to synthesis gas production process.
17
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
GASIFICATION A TECHNICO-ECONOMICAL FEASIBILITY STUDY OF PLASMA ASSISTED COAL GASIFICATION COMPARED TO A REFERENCE AUTO-THERMAL GASIFICTION PROCESS Nazim Merlo ICARE-CNRS, 1C Av. de la Recherche Scientifique, 45071 Orléans cedex 2 (France) Tel : 33-(0)2 38 25 54 63 Fax: 33-(0)2 38 25 78 75 [email protected] Iskender Gökalp ICARE-CNRS 1C Av. de la Recherche Scientifique, 45071 Orléans cedex 2 (France) Tel : 33-(0)2 38 25 54 63 Fax: 33-(0)2 38 25 78 75 [email protected] Abstract: The paper explores the technical and economic feasibility of a plasma assisted coal gasification facility. Allothermal gasification processes provide higher gas mass yield compared to auto-thermal processes but their relevancy in terms of energy balance, global economics and environmental, namely CO2, impact remains still to be demonstrated. In this work, we first select a reference auto-thermal process of coal gasification compatible with plasma generator integration. The concept chosen is based on the NETL study entitled Baseline Technical and Economic Assessment of a Commercial Scale FischerTropsch Liquids Facility (2007). The coal of reference in this study is an Illinois No. 6 bituminous coal. The gasifiers studied concern two-stage, oxygen-blown, entrained flow, refractory-lined gasifiers with continuous slag removal. Dedicated air separation units supply 95 % purity oxygen to the gasification process. Gasification units are the heaviest investment part with 41 % of the total investment for the considered plant. The commercial aim of the modeled plant is to produce 50 000 barrels per day of commercialgrade diesel (56 %) and naphta liquids (44 %). In a second step, we design a new gasification process where plasma generators are integrated, as realistically as possible, to the reference process. For this work we essentially use the AspenPlus software. Plasma torches chosen belong to the non-transferred arc technology with a hot cathode up to 200 kWe and a cold one beyond. Generally, the power range of plasma torches selected for industrial uses is 200-4000 kWe. Basically, current technological constraints concern cathode materials in relation to the plasma gas and the gasifier pressure range which is 0.1-0.5 MPa for these plasma torches. We also define the operating conditions of syngas for a Fischer-Tropsch synthesis: with iron catalyst, the H2 / CO ratio required being between 0.7 and 1.5 with the temperature range 493613 K and the pressure range 1.5-3 MPa. Note that the quality of syngas is also determined by the content of tars usually accepted as < 0.1 mg / Nm3. This upper limit could be tightened if the generated syngas should be compressed for Fischer-Tropsch treatment. Several studies about coal gasification in hydrogen, air and steam plasmas were carried out since 1990’s [1-3]. Plasma generators integration in a coal gasification process is considered in [4,5], especially for coal ignition assisted by plasma generators. Equilibrium thermodynamic models and chemical kinetics approaches were also developed in [6,7]. A similar study conducted for liquid biomass gasification with a plasma assisted process has shown an improved ratio of H2 / CO but of course at the expense of electric power consumption [8].
Using all this available information, a comparison of the mass and energy balances with an assessment of carbon footprint for both autothermal and allothermal coal gasification processes are presented. The additional energy consumption should not ideally exceed the equivalent of energy released by coal combustion to entertain the auto-thermal gasification in the reference process. We also compare investment and operating costs for the two processes including the Fischer-Tropsch units, compression units and the gas cleaning systems. Acknowledgments: This work is supported by the General Directorate of Turkish Coal Enterprises References: [1] B. Mikhailov, In: Solonenko OP, Zhukov MF (Editors) Plasma gasification of coal, thermal plasma and new materials technology, Vol. 2. Cambridge, Interscience, 1995. p. 345–69 [2] IM. Zasypkin, GV. Nozdrenko. Production of acetylene and synthesis gas from coal by plasma chemical methods. In: Solonenko OP. (Editor).Thermal plasma torches and technologies, Vol II.. Cambridge, Interscience, 2001. p. 234–43 [3] J. Qiu, X. He, T. Sun, Z. Zhao, Y. Zhou, S. Gou, J. Yhang, T. Ma, Coal gasification in steam and air medium under plasma conditions: a preliminary study, Fuel Processing Technology 85 (2004) 969–982 [4] A. N. Bratsev, V.E. Popov, A. F. Rutberg, and S. V. Shtengel’, A facility for plasma gasification of waste of various types, High Temperature, 44 (2006), 823-828 [5] M. Sugimoto, K. Maruta, K. Takeda, O.P. Solonenko, M. Sakashita, M. Nakamura, Stabilization of pulverized coal combustion by plasma assist, Thin Solid Films 407 (2002) 186–191 [6] A. Mountouris, E. Voutsas, D. Tassios, Solid waste plasma gasification: Equilibrium model development and exergy analysis, Energy Conversion & Management, 47 (2006) 1723-1737 [7] M. Minutillo, A. Perna, D. Di Bona, Modelling and performance analysis of an integrated plasma gasification combined cycle (IPGCC) power plant, Energy Conversion & Management, 50 (2009) 2837-2842 [8] J. Peybernes, D. Guenadou, H. Lorcet / CEA, Évaluation des potentialités de la gazéification allothermique de ligno-cellulose à la production de bio-carburant de synthèse (GALACSY), N° ANR-06-BIOE-006
PROCESSES TO PRODUCE VALUE ADDED PRODUCTS FROM CO2
Deena Ferdous and Belma Demirel Rentech Inc., Rentech Energy Technology Center, 4150 East 60th Avenue, Commerce City, CO 80022, USA [email protected] Abstract The possible ways to convert CO2 to valuable products are (1) catalytic conversion, (2) electrocatalytic/electrochemical process, (3) plasma processes, (4) photocatalytic /photochemical process, and (5) enzymatic/biochemical process. Hydrocarbon, hydrogen and oxygenates are mainly produced by the catalytic conversion of CO2 at the temperature and pressure of 250‐1000°C and 140‐3500 psi, respectively. On the other hand, carbon monoxide, syngas, hydrocarbons and alcohols are the main products produced at low pressure and high temperature from plasma process. CO2 is converted to hydrogen and hydrocarbons by photocatalytic/ photochemical process at the temperature and pressure of 5‐400°C and 485‐1,470 psi, respectively. Electrocatalytic/electrochemical process produces alcohol, aldehydes, ketones, other oxygenated, alkenes and alkanes at very low temperature (room – 150°C). Furthermore, bi‐carbonate and alcohols are also produced from CO2 at low temperatures and pressures using enzymatic/biochemical process. Successful implementation of the above mentioned processes could be a possible solution to reduce the amount of CO2 emission into the atmosphere. 1. Introduction The rapid growth in world population and industrial development has lead to massively increased usage of fossil fuels such as coal, petroleum and natural gas which resulted in the formation of carbon dioxide. Furthermore, the dynamic equilibrium of carbon dioxide formation and conversion has been seriously destroyed due to the deforestation and the reduction of rain forest. DEO report indicated that the annual worldwide production of carbon dioxide is ~6 billion tons among these US produces ~ 2 billion tons that represent the energy content of about 11 million barrels of oil per day or roughly the US daily import.1 Suggestion has been made to sequester carbon dioxide2. Moreover, the ability to store several billion tons of carbon emitted in the form of carbon dioxide is questionable as there are environmental consequences. Recycling of carbon dioxide via conversion into high energy content fuel is an alternative option. However this process is energy intense.3 The CO2 reduction technology can be divided into two methods (1) physics‐based; and (2) chemistry based. In physics‐based methods, carbon dioxide is captured from the atmosphere and then stored underground or under the sea bed using high pressure compression. From the viewpoint of equilibrium and cyclic of carbon dioxide on earth, the amount of carbon dioxide has not been reduced.4 Therefore, a significant effort is given to the use of chemistry–based methods to convert carbon dioxide to valuable products. This paper will discuss all possible routes of chemical conversion of CO2 to valuable products. 2. Different Processes to Convert CO2 to Value added products The possible ways to convert CO2 to valuable products are (1) catalytic conversion, (2) electrocatalytic/electrochemical process, (3) plasma processes, (4) photocatalytic/photochemical process and (5) enzymatic/biochemical process.
2.1. Catalytic conversion The carbon dioxide reforming of methane has received considerable attention from both academic and industrial sectors. This reaction converts the cheapest carbon containing gases (CH4 and CO2) to synthesis gas with a H2/CO ratio of 1. This process is particularly interesting environmentally when the gas fields contain a significant amount of methane and carbon dioxide. During CO2 methane reforming the following reaction occurs.5 CH4 + CO2 2CO + 2H2 (carbon dioxide reforming) (1) H2 + CO2 CO + H2O (reverse water gas shift) (2) (3) CO2 + 4H2 CH4 + 2H2O (CO2 methanation) CO + 3H2 CH4 + H2O (CO methanation) (4) CH4 + 2H2O CO + 3H2 (steam reforming) (5) 2CO C + CO2 (Boudouard reaction) (6) CH4 C + 2H2 (methane decomposition) (7) (8) C + H2O CO + H2 (coke gasification) The main drawback of this process are5 (1) it requires temperature as high as 800 °C for high conversion; and (2) catalytic deactivation due to carbon deposition. The two main reactions of coke formation are methane decomposition (Reaction 7) and the Boudouard reaction (Reaction 6). The first reaction is favored at high temperatures and low pressures, whereas the second one is favored at low temperatures and high pressures. Both low and high temperature steam reforming have been reported in the literature. Tomishige et al.6 studied the performance of Ni0.03Mg0.97O catalyst for dry reforming reaction (Reaction 1) at atmospheric pressure. This catalyst was very resistant to carbon deposition. On the hand, Song et al.7 reported a decrease in CO2 and methane conversion and H2/CO ratio and increase in carbon deposition when dry reforming was performed at high pressure on Ni/Na‐Y and Ni/Al2O3 catalysts. Similar trends were observed by Shamsi and Johnson8 when the reaction was performed at high pressure over Pt/ZrO2, Pt/Ce‐ZrO2 and Rh/Al2O3 catalysts. Itkulova et al.9 studied CO2 reforming of methane to syngas over Co‐Pd supported on aluminum catalyst and then syngas was converted to oxygenates. A two step process was used which was not very attractive from the economic point of view. The disadvantages of this process presently at the industrial scale is the utilization of extra energy in the form of hydrogen, which at industrial level is obtained by decomposition of natural gas or electrolysis of water and the use of high temperature and pressure. This makes the whole technology economically unfeasible for the industries. It is a thought that the rapid CO2‐methanation could be very useful to convert inactive CO2 to high energy products10. Rapid CO2 methanation mainly occurs in the temperature range of 250‐450°C. This reaction is very exothermic. The heat generated from this reaction can be utilized further. On the other hand, the methane formed can be used as a fuel to obtain high temperature (~1000°C). But the disadvantage of this process is that it requires large consumption of hydrogen per unit CO2. Ni‐La2O3‐Ru or Ni‐Ce2O3‐Pt supported on alumina catalyst is mainly used for this type of reaction11. From the point of view of energy balance, methanol synthesis by CO2 hydrogenation is regarded as the equivalent of the hydrocarbon synthesis from syngas. Methanol synthesis from CO2 over heterogeneous catalyst mainly occurs according to the following equations12. CO2 + 3H2 CH3OH + H2O (9) Conversion to methanol is favored by increasing the pressure and decreasing the temperature.12 Inui and co‐workers11 performed significant study on the catalyst development for CO2 hydrogenation. They found that the yield of CH3OH is very low when conventional Cu‐Zn catalyst was used. The reaction temperatures and pressures were as identical as the methanol synthesis from the syngas. However, a higher activity of the catalyst for methanol synthesis was obtained when Cu‐ZnO‐Cr2O3‐Ga2O3 catalyst was used. This catalyst also retarded the formation of CO.
2.2. Electrochemical conversion In the electrochemical process, chemical reactions take place in a solution at the interface of an electron conductor (the electrode, composed of a metal or a semiconductor) and an ionic conductor (the electrolyte), which involve electron transfer between the electrode and the electrolyte or species in solution. Electrochemical reduction of CO2 appears to be one of the effective methods for conversion and fixation of CO213. This reaction mainly occurs at low temperature and high pressure. Some low temperature and low pressure electrocatalysis reaction was also reported14. In the past, the electrochemical reduction of CO2 in water15 has been reported. However, the product distribution of electro catalytic reaction is influenced by the nature of noble metal. Methanol is considered as a much better solvent for CO2 than water because the solubility of CO2 in methanol is about five times or more than in water at ambient temperature and pressure16. This type of electrolyte is mainly used to produce formic acid. Mizuno and co workers17 have investigated the electrochemical reduction of CO2 on various metal electrodes using ethanol at ordinary temperature and pressure. In the electro reduction system using benzalkonium chloride/methanol electrolyte, no formic acid was produced on Ag, Au, Co, Cu, Fe, Ni, Pt, Sn, Ti and Zn. In potassium hydroxide/methanol electrolyte, the formation of formic acid was observed at Au, Cu, Pd, hydrogen storing Pd, Ti, and hydrogen storing Ti electrodes. This group observed that even under optimum conditions, the faradaic efficiency for formic acid formation was relatively low (21% on Cu electrode). Koneco et al.18 also studied the electrochemical reaction of CO2 at a Cu electrode in a methyl based electrolyte using lithium supported salts as LiCl, LiBr, LiI, LiClO4 and CH3COOLi at low temperature (‐ 30°C). The main products were methane, ethylene, carbon monoxide and formic acid. A maximum faradic efficiency of methane was 71.8% in LiClO4/methanol based electrolyte at ‐3.0 V vs Ag/AgCl saturated KCl. This group also studied the electrochemical reaction of CO2 in water+methanol (8:2) with copper‐electro‐ deposited nickel electrode. Under optimum experimental conditions, 20.2% methane Faradaic efficiency and 7.5% ethylene were obtained by electrochemical reduction of CO2. A reaction mechanism of the electrochemical reduction of CO2 at copper modified Ni electrode in water + methanol solution was proposed.
CO2
+e-
CO2-(ads) +H+ , +e-
CO (ads) + OH+4H+ , +4e-H2O
+eCO2
CO2-(ads)
+H+ , +eCO2, -CO32-
CH2 (ads) +:CH2 (ads)
HCOO-(ads)
CH4 C2H4 HCOO-(des) CO
Ogura et al.19 studied the electrolysis of CO2 at the three phase interface on a copper mesh electrode in acidic halide solutions containing cupric ions. A selective formation of ethylene with a maximum Faradaic efficiency of 73.5% vs Ag/AgCl in a KBr solution was obtained. The probable reaction scheme is considered as: CO2 (g) is first reduced to CO (g) at the three‐phase interface, CO absorbs on a newly formed copper surface, and the electrochemical hydrogenation of copper to :CH2 and the coupling of :CH2 to C2H4 were followed20.
2.3. Plasma process Yang et al.4 proposed that a plasma technology could be more effective approach for the conversion carbon dioxide. Plasma process is a simple and fast process compared to other process. In the plasma process molecules are entered into electric fields where they are excited and ionized by collision with accelerated electrons to generate various species such as atoms, electrons, ions, free radicals etc. The mixtures of these species are plasma. There are two types of plasmas: low energy (1‐20 eV) sometimes termed glow discharge, low temperature or cold plasmas and high temperature plasmas where the mean temperature of the gas can range from hundreds to millions of degrees. In the thermal plasma all the species, i.e free electrons, free radicals, neutral molecules and ions, are very hot (5,000‐50,000°C). On the other hand, in cold plasma only free electrons are hot, while heavier ions and neutral molecules remain close to room temperature. The activated species generated by plasma bombardments can recombine to form new products. The bond dissociation of CO2 can be achieved at lower energy level in the plasma process. Liu et al.22 have proposed a non‐thermal plasma approach for the formation of lower olefins from methane and CO2: 2CO2 + 2CH4 2CO + C2H4 + 2H2O (10) 18 At 800°C, equilibrium yields of ethylene were fairly high at 57%. However, the experimental yield over metal oxide catalysts was not sufficiently high (less than 9%). This group found that CO2 plasmas can generate many active species. Some of them can be utilized to produce other valuable chemicals such as ethylene, propylene and oxygenates. Their experimental results showed that CO2 is an excellent “catalyst” for the conversion of low alkanes to alkenes (especially ethylene and propylene). Yun et al.23 studied the thermal decomposition of CO2. Their results showed higher conversion of CO2 when it was reacted with other reactants such as CH4 and H2 compared to the direct decomposition of CO2. The method of quenching or cooling the hot product stream in thermal plasma is often a vital factor to determine the final product distribution during thermal plasma studies. For many plasma reactions, the final product distribution depends on the rates of competing reactions during the quenching step. Therefore, the quenching technology in controlling temperature is very important in the thermal plasma process. They have indicated that the quenching is not effective if the speed of quenching is less than 105 K/s. Yang et al.4 reported the formation of different organic products using plasma method. In this method, carbon dioxide and counterpart molecules were mixed in reaction chamber and the plasma excitation was utilized to trigger chemical reactions. Depending on the magnitude of the input power and molecular structure of counter parts, the final products were polymer, oligo‐polymer or low molecular weight small molecules. They reported that the conversion of carbon dioxide and chemical structures of the resulting products are strongly dependent on the selection of the counterpart molecules. They claimed that through this plasma process carbon dioxide can be converted to useful materials such as plastic or fuels. Among the effective plasma systems, dielectric barrier discharge (DBD) plasma which is characterized by the presence of one or more insulating layers in the current path has received significant attention. But the plasma discharge energy is still too high to meet economic requirements. Further improvements are needed to get a better understanding of the plasma kinetics, better discharge control, and the simultaneous use of catalysts. 2.4. Photoelectric process Photocatalytic reaction is one of the most available reactions at low temperature, because it can promote thermodynamically unfavorable reaction under mild conditions24. Photocatalysis is the acceleration of a photo reaction in the presence of a catalyst. In the catalyzed photolysis, light is absorbed
by an absorbed substrate. Generally, photon reaction is the primary source of energy for electron hole‐ pair formation at ambient temperature. In photo‐generated catalysis the photocatalytic activity (PCA) depends on the ability of the catalyst to create electron‐hole‐pairs, which generate free radicals (hydroxyl radicals: .OH) able to undergo secondary reactions. Photocatalysis occurs under relatively mild conditions with lower energy input, especially where the reaction is activated by solar energy or other easily obtained light source. The schematic of the photocatalysis reaction of CO2 is given below.
Figure 1. The schematic of the photo catalysis reaction of CO2 Photocatalysis reduction of CO2 was first demonstrated by Honda and co‐worker25 where they have reported to produce formaldehyde (HCHO), formic acid (HCOOH), methyl alcohol (CH3OH) and trace of amount of methane from this reaction using different semiconductors such as tungsten trioxide (WO3), titanium dioxide (TiO2), zinc oxide (ZnO), cadmium sulfide (CdS), gallium phosphide (GaP) and silicon carbide. In their work, xenon‐ and mercury lamp irradiation was used to activate these semiconductors. Literature indicates that the product distribution of this reaction can be changed by using specific semiconductors26. The production of methane, C2H5OH and C2H6 were also reported in the literature.26 Photoreduction of CO2 can be performed using different reductant such as water, H2, propanol, isopropyl alcohol, CH4 etc dependent on the photocatalyst. The photoreduction of CO2 by water is readily available and inexpensive. This process is also very promising and can form a useful carbon cycle. There are two important species such as H. (hydrogen atom) and .CO2‐ (carbon dioxide anion radical) are mainly involved in CO2 photoreduction. These species are produced by electron transfer according to the following reactions27. H+ + e‐ H. (11) ‐ ‐ CO2 + e CO2 (12) In photogenerated catalysis, the photocatalytic activity (PCA) depends on the ability of the catalyst to create electron‐hole‐pairs, which generate free radicals (hydroxyl radicals: .OH) which undergo secondary reactions. These radicals then form other stable substances. It should be noted that the stability of CO2 in water is low and CO2 photoreduction process competes with H2 and H2O2 formation which consumes H+ and e‐. Due to this limitation, water has been replaced by other reductants by some researchers in order to improve the desired product yields. Both low28 and high temperature29 photo‐catalytic reaction have been reported in the literature. Sometimes high temperature is used to increase the reaction rate by increasing collision frequency and diffusion rate. Saladin et al.30 reported that high temperatures are beneficial to the rate of thermally activated steps such as desorption, which occurs after photochemical reaction. The maximum temperature in their work was 427°C. On the other hand, Kohno et al.31 noted that the photo excitation step of CO2 with hydrogen over ZrO2 can be deactivated with increased temperature. On the hand, Gokon et al.29 indicated that the photo catalytic reaction cannot be induced when temperature is greater than 800°C in the presence of ZnO due to the dominating effect of carbothermal reduction of ZnO. However, they have observed enhanced CO evaluation rate when the experiment was performed at 600°C.
Photocatalytic reaction of CO2 at room temperature with gaseous water on powder TiO2 was also reported32, 33. But the production of CH4 was very low. Literature indicates that small TiO2 particles have higher activity29 and the highly dispersed TiO2 exhibits a higher reductivity than bulk catalyst32. Kaneco et al.14 stated that the formation of methane increased during photocatalysis reduction of CO2 using TiO2 suspension in isopropyl alcohol solution at 28 atm. On the other hand, Adachi et al.34 indicated that ethylene cannot be produced at relatively high pressure because due to the accelerated formation of methane. Photocatalysis is a potentially economical and environmental CO2 removal process. It is very important to understand the parameters that influence the photocatalysis reaction to develop this for process for practical use. Improvement can be done by choosing semiconductor with appropriate band‐ gap energy, by developing reductants and by optimizing operating conditions such as temperature, pressure, light intensity and operating wavelength. 2.5. Enzymatic/biochemical process Recently biocatalytic reduction of carbon has been gained significant attention because it requires mild reaction conditions and does not require high purity reactant. Biocatalytic reaction is capable to handle many impurities such as sulfur compounds which are common in flue gas. Biochemical reduction of carbon dioxide has been used to produce formate, phenol, ethanol, fuels, bi carbonate and menthol among them methanol production has received more attention because of its use in wide variety of application such as production of formaldehyde, in automotive anti‐freeze, in a variety of chemical synthesis, as a general solvent, as an activation fuel (for water injection), as a denaturant for ethyl alcohol and a dehydrator for natural gas35. Certain microorganisms such as E‐coli, formalose, CLEC Antarctica Lipase B, carbonic anhyhydrase can reduce CO2 into valuable organic compounds via several steps. The enzymatic reduction of CO2 requires an electron donar, which can be a coenzyme such as nicotinamide adenite dinucleotide (NADH) or a synthetic chemical like methyl viologen, depending on the specificity of the enzymes. The NADH‐mediate reduction of CO2 for production of methanol can be described as follows. FateDH + CO2 +NADH +H+ HCOOH + NAD (13) FaldDH + + HCOOH + NADH + H CH2O + NAD + H2O (14) + ADH CH2O + NADH + H+ CH (15) 3OH + NAD But the actual process is much more complicated because the solvation of CO2 involves the formation of several species including CO32‐, HCO3‐, and H2CO3. The thermodynamic equilibrium and the concentration distribution of these species depend on the pH and the other physicochemical properties of the solution. The enzyme can take all of these species as substrates for methanol synthesis other than the free state CO2 only.36 The enzyme carbonic anhydrase is known as the fastest CO2 catalyst. Conventionally, enzymatic reactions are performed in aqueous solutions, but recently, supercritical reactions in organic solvents and supercritical CO2 have received extensive attention. Mobilized and immobilized both enzyme can be used for enzymatic reaction. Immobolized enzyme is easy to separate and reusable. But most immobilized enzymes show lower activity than mobilized enzyme37. Wu et al.38 studied CO2 enzymatic conversion in methanol using formate dehydrogenase (FateDH), formaldehyde dehydrogenase (FaldDH) and alcohol dehydrogenase (ADH) co‐encapsulated in silica gel prepared by modified silica process as catalysts. In their work TEOS was used as the precursor and NADH was used as the electron donar. They indicated that each dehydrogenase had its own optimum reaction temperatures. For example, for FateDH and FaldDH, the optimum temperature was 37°C whereas for ADH, the optimum temperature was 25°C. The maximum methanol yield was 92.1% at the temperature, pressure and pH of 37°C, 0.3 MPa and 7, respectively. A wide variety of enzymes retain their characteristics reactivities and chemical functions when they are confined within the pores of the silica sol‐gel derived matrix. The porosity of silica gel glasses allows small
molecules and ion to diffuse into the matrix while the enzymes remain physically trapped in the pores, and thus resulting in an enhanced probability of reaction due to an increase in local concentration of substrates within the nanopores39. 3. Conclusion Carbon dioxide reforming of methane is very attractive with a view to enhance natural resources of CO2 and CH4 into valuable chemicals. However, the reaction for syngas (CO+H2) formation is processed at high temperature using various catalysts but the high temperature and the catalyst makes the process very costly. Catalytic conversion of CO2 to methane or methanol is also attractive but the methane production requires large consumption of hydrogen per unit CO2 and the methanol selectivity is very low even in the presence of catalyst. Electrochemical, photo electric, low temperature plasma or enzyme process could be effective methods for the conversion and fixation of CO2. 4. References 1. Feld, L. (2000),”Bulletin of the United States Energy Information Administration”, U.S. Department of Energy, http://www.eia.doe.gov/emeu/cabs/usa.html” 2. Buesseler, O. K., Doney S. C., Karl D. M., Boyd ,P. W., Caldeira, K., Chai, F., Coale, K. H., de Baar,J. W. H., Falkowski, P. G., Johnson, K. S., Lampitt, R. S., Michaels, A. F., Naqvi, W. A., Smetacek, V., Takeda, S., and Watson, A. J. (2008), “Ocean Iron Fertilization—Moving Forward in a Sea of Uncertainty,” Science, January, pp.162. 3. Varghese, O. K., Paulose, M., LaTempa, T. J. and Grimes, C. A. (2009), “High‐Rate Solar Photocatalytic Conversion of CO2 and Water Vapor to Hydrocarbon Fuels,” Nano Lett., 9(2), pp. 731–737. 4. Yang, A. C.‐M., Chang, Yi‐H., and Chang, C.‐C. (2009) “Conversion of Carbon Dioxide into Useful Organic Products by Using Plasma technology,” IPC8 Class: AB01J1908FI. 5. Gallego, S. G., Mondragon, F., Tatibouet, J.‐M., Barrault, J., and Batiot‐Dupeyrat, C. (2008), “Carbon Dioxide Reforming of Methane over La2NiO4 as Catalyst Precursor—Characterization of Carbon Deposition”, Catal. Today, 133‐155, pp. 200–209. 6. Tomishige, K.; Himeno, Y.; Mastuo, Y.; Yoshinaga and Fujimoto K. (2000), “Catalytic Performance and Carbon Deposition ; Behavior of a NiO‐ MgO Solid Solution in Methane Reforming with Carbon Dioxide under Pressurized Conditions”, Ind. Eng. Chem. Res., 39(6), pp.1891‐1897. 7. Song, C., Srinivas, S. T., Sun, L. and Armor, J. N. (2000), “Comparison of High‐Pressure and Atmospheric‐Pressure Reactions for CO2 Reforming of CH4 over Ni/Na‐Y and Ni/Al2O3 Catalysts. Am. Chem. Soc. Div. Petrol. Chem. Prep., 45 (1), pp. 143‐148. 8. Shamsi, A. and Johnson, C D. (2007), “The Effect of CO2 Reforming of Methane and the Carbon Deposition Route Using Noble Metal Catalyst”, ACS Symposium Series, Vol. 959, Chapter 8, pp. 87‐101. 9. Olah, G. A., Prakash, G. K. S. (2008), “Conversion of Carbon Dioxide to Methanol and/or Dimethyl Ether using Bi‐Reforming of Methane or Natural Gas”, IPC8 Class: AC07C2700F1, USPC Class: 518700. 10. Itkulova, S. S., Zhunsova, K. Z. and Zakumbaeva, G. D. (2005), “CO2 Reforming of Methane over CO‐ Pd/Al2O3 Catalysts”, Bull. Korean Chem. Soc., 26, pp. 2017‐2022. 11. Inui, T. (1996), “Highly Effective Conversion of Carbon Dioxide to Valuable Compounds on Composite Catalysts,” Catalysis Today, 29, pp. 329‐337. 12. Inui, T., Funabiki, M., and Takegami, Y., (1980), “Simultaneous Methanation of CO and CO2 on Supported Ni‐Based Composite Catalysts”, Ind. Eng. Chem. Prod. Res. Dev., 19, pp. 385–388 13. Halmann, M. M. (1993), “Chemical Fixation of Carbon Dioxide: Mrthods of Recycling CO2 into useful Products”, CRC Press, Florida, pp. 67‐127.
14. Kaneco, S., Iwao, R., Iiba, K., Ohta, K., and Mizuno, T. (1998), “Electrochemical Conversion of Carbon Dioxide to Formic Acid on Pb in KOH/Methanol Electrolyte at Ambient Temperature and Pressure,” Energy, 23(12), pp. 1107‐1112. 15. Hori, Y., Kikuchi, K., and Sizuki, S. (1985), “Production of CO and CH4 in Electrochemical Reduction of CO2 at Metal Electrodes in Aqueous Hydrogencarbonate Solution", Chemistry Letters, 11, pp. 1695‐1698. 16. Lide, D. R. (eds) (1984), “Handbook of Chemistry and Physics”, 72nd Ed. CRC Press, Boston, pp. 4‐6. 17. Mizuno, T., Kawamoto, M., and Koneco, M. (1998), “Electrochemical Reduction of Carbon Dioxide at Ti and Hydrogen‐Storing Ti Electrodes in KOH‐Methanol”, Electrochem. Acta, 43(8) pp. 899‐907. 18. Kaneco, S., Hiei, N., Xing, Y., Katsumata, H., Ohnishi, H., Suzuki, T., and Ohta, K. (2002), “Electrochemical Conversion of Carbon Dioxide to Methane in Aqueous NaHCO3 Solution at Less than 273 K,” Electrochem. Acta , 41(21), pp. 51‐55. 19. Ogura, K. , Yano, H., and Shirai, F. (2003), “Selective Conversion of CO2 to Ethylene by the Electrocatalysis at a Three‐Phases (Gas/Liquid/Solid) Interface in an Acidic Solution Containing Cupric Ions”, Fuel. Chem. Div. Pre., 48(1), pp. 264‐265. 20. Huang, H. Y., Padin, J., and Yang, R. T. (1999) “Comparison of π‐Complexations of Ethylene and Carbon Monoxide and Cu+ and Ag+”, Ind. Eng. Chem. Res., 38, pp. 2720‐2725. 21. Chapman, B. (1980) “Glow Discharge Processing”, Wiley, New York. 22. Liu,C‐J., Xu, G.‐H. and Wang, T. (1999), “Non‐thermal Plasma Approaches in CO2 Utilization”, Fuel Processing Technology, 58, pp. 119‐134. 23. Yun, S.‐H., Kim, G.‐J., and Park, D.‐W. (1997), “Decomposition and Conversion of Carbon Dioxide into Synthesis Gas using Thermal Plasma”, J. Ind. Eng. Chem. 3(4), pp. 293‐297. 24. Shi, X. D., Feng, Q. Y., and Zhong, H. S. (2005), “Photocatalytic Synthesis of Oxygenates from Gaseous CO2 and CH4 over Semiconductor”, Chi. Chem. Lett. 16(5), pp. 685‐687. 25. Inoue, T., Fujishima, A., Konishi, S., and Honda, K. (1979) “Photoelectric Catalytic Reduction of Carbon Dioxide in Aqueous Suspensions of Semiconductor Powders”, Natur, 277, pp. 637‐638. 26. Subrahmanyam, M., Kaneco, S. and Alonso‐Vante, N. (1999), “A Screening for the Photo Reduction of Carbon Dioxide Supported on Metal Oxide Catalysts for C1 ‐C3 Selectivity”, Appl. Catal. B: Environ., 23(2‐3), pp. 169‐174. 27. Usubharatana, P., McMartin, D., Veawab, A. and Tontiwachwuthikul, P. (2006), “Photocatalytic Process for CO2 Emission Reduction from Industrial Flue Gas Streams”, Ind. Eng. Chem. Res., 45, pp. 2558‐ 2568. 28. Guan G., Kida T., and Yoshida A. (2003), “Reduction of Carbon Dioxide with Water under Concentrated Sunlight using Photocatalyst Combined with Fe‐based Catalyst”, 41, pp. 387‐396. 29. Gokon, N., Hasegawa, N., Kaneko, H., and Aoki, H., Tamaura, Y., Kitamura, M. (2003), “Photocatalytic Effect of ZnO on Carbon Gasification with CO2 for High‐Temperature Solar Thermochemistry”, Sol. Energy Mater. Sol. Cells, 80, pp. 335‐341. 30. Saladin, F., and Alxneit, I. (1997), “Temperature Dependence of the Photosynthestic Reduction of CO2 with H2O at the Solid/Gas Interface of TiO2”, Chem. Soc. Faraday Trans, 93 (23), PP. 4159‐4163. 31. Kohno, Y., Hayashi, H., Takenaka, S., Tanaka, T., Funabiki, T., and Yoshida, S. (1999), “Photoenhanced Reduction of Carbon Dioxide with Hydrogen over Rh/TiO2”, J. Photochem. Photobiol. A, 126, pp. 117‐123. 32. Anpo, M., and Takeuchi, M. (2003), “The Design and Development of Highly Reactive Titanium Oxide Photocatalysts Operating Under Visible Light Irradiation”, J. Catal., 216, pp. 505‐516. 33. Yamashita, H., Nishiguchi, H., Kamada, N., and Anpo, M. (1994), “Photocatalytic Reduction of CO2 with H2O on TiO2 and Cu/TiO2 Catalysts”, Res. Chem. Intermed., 20(8), pp. 815‐823. 34. Adachi, K., Ohta, K., and Mizuno, T. (1994), “Photocatalytic Reduction of Carbon Dioxide to Hydrocarbon using Copper‐Loaded Titanium Dioxide”, Sol. Energy, 53(2), pp. 187‐190. 35. Dave, C. B., Rao, S. M., and Burt, C. M. (2007), “Conversion of Carbon Dioxide to Methanol in Silica Sol‐ Gel Matrix”, US patent, US2007/0042479AO, January.
36. Baskaya, F. S., Zhao X., Flickinger C. M., and Wang, P. (2010), “Thermodynamic Feasibility of Enzymatic Reduction of Carbon Dioxide to Methanol”, Appl Biochem Biotechnol, September pp. 8758‐8772. 37. Obert, R., and Dave, B. (1999), “Enzymatic Conversion of Carbon Dioxide to Methanol: Enhanced Methanol Production in Silica Sol‐Gel Matrices”, J. Am. Chem. Soc., 121, pp. 12192‐12193. 38. Wu, H., Jiang Y. Z., Xui, W. S., and Huang, F. S. (2003) “A New Biochemical Way for Conversion of CO2 to Methanol via Dehydrogenases Encapsulated in SiO2 Matrix”, Chinese Chemical Letters, 14(4), pp. 423 – 425. 39. Liu, C., Mallinson, R., Aresta, M., Jiang, Z., Wu, H., Xu, S., and Huang, S. (2002), ”Enzymatic Conversion of Carbon Dioxide to methanol by Dehydrogenases Encapsulated in Sol‐Gel Matrix”, Fuel Chemistry Division Preprints, 47(1), pp. 306
Design and Operational Strategies for IGCC with CO2 Capture George Booras, Chris Higman, Dan Kubek, Jim Sorensen, Doug Todd Electric Power Research Institute -- Palo Alto, CA & Charlotte, NC, USA
1 Introduction In Europe, Japan, Australia, the USA and other countries actively developing plans for CO2 capture, transport, and geological storage, specifications for the purity of the CO2 are still in the development stage. Specifications for existing CO2 pipelines in the USA vary according to the history of individual systems and CO2 usage. The EU directive on CCS calls for the CO2 stream to consist overwhelmingly of CO2 and addresses the makeup of the remainder in terms of outcomes to avoid, rather than in any specific detail. In addition to meeting CO2 specifications for transport and storage, designers of CO2 capture systems must address the environmental and safety impacts of low-concentration components in the CO2 when it must be vented during startup, shutdown, and situations when send-out to transport and storage is interrupted. Specifications for impurities such as sulfur species, CO, H2, acid gas removal solvent, etc. may be dictated by environmental regulations and safety requirements, rather than by pipeline materials compatibility and considerations related to underground storage or use for enhanced oil recovery (EOR). IGCC possesses inherent advantages in meeting varying CO2 purity requirements, and offers flexibility and optimization potential for mitigating the cost of satisfying these requirements. This paper presents the results of the Electric Power Research Institute (EPRI) CoalFleet-for-Tomorrow Program’s technical review of the impact that these issues have on the design and operation of IGCC units with CO2 capture.
2 Design Basis 2.1 Pipeline Purity Specifications Historically pipeline purity specifications historically have been set for reasons attributable mostly to the particular source of the CO2 and/or its use for EOR, rather than as a result of any specific analysis of potential contaminants. With only a few exceptions, specifications have not been set for gasification plants and thus could impact IGCC design and operation. Table 1 provides a summary of some existing pipeline purity specifications.
th
PCC, Istanbul, 12 October 2010
-1-
Table 1 Representative CO2 Pipeline Purity Specifications Kinder Morgan
Dakota Gasification
Hydrogen Energy California (HECA)
FutureGen
CO2
> 95%
> 95%
>97%
95%
H 2S
20 ppmw (~25 ppmv)
< 20,000 ppmv
Total S
35 ppmw (~48 ppmv as COS)
<30 ppmv
N2
< 4 mole %
< 2%
O2
< 10 ppmw
(25-50ppm)
HCs
< 5 mole %
< 2%
Water
< 30 lbs/MMscf No free water
Pressure
100 ppmv
< 0.5% <10 ppmv
0.61lb/MMSCF,13 ppmv o 11 F dew point o -100 F dew point No free water <15 lbs/MMscf
100 ppmv max
2700 psi (186 bar)
2200 psi (152 bar)
Temperature < 120ºF (< 50°C)
≤ 120ºF (≤ 50°C)
95º F (35°C)
The origin of these specifications varies. The first three are all connected with an EOR end use and the FutureGen specification is for geological storage in a saline aquifer. The Kinder Morgan specification is derived at least in part from the quality of CO2 obtained from natural occurrence or from separation at natural gas wellheads. The Dakota Gasification specification is simply the quality of CO2 that was available (and being blended with fuel gas for combustion in a boiler) prior to being exported to the Weyburn oil field. The FutureGen specification was arrived at on the basis of a risk 1 assessment prepared as part of the Environmental Impact Statement . There are other existing specifications for other pipeline systems or specifications under development such as that prepared by 2 the European DYNAMIS project . The low water requirement is common to all these specifications even if the numbers are slightly different. This requirement is needed to avoid corrosion in carbon steel pipelines. There are also 3 suggestions that the potential for hydrate formation may also impose a limit . There is currently no common approach for sulfur species. While for EOR the Weyburn experience indicates that a high concentration of H2S at the injection site is not an issue, the FutureGen study places a much lower limit as a result of potential safety issues in connection with pipeline transportation. Interestingly, the DYNAMIS figure for H2S is 200 ppm, double that determined by the FutureGen study. Kinder Morgan’s values are assumed to be based on their current sources of CO2 and commercial commitments to maintain these specifications. Limitations on total sulfur need more
1
th
www.netl.doe.gov/technologies/coalpower/futuregen/EIS, viewed 27 August, 2010.
2
http://www.dynamis-hypogen.com/publications/Deliverables/Year_2/D3-1-3 DYNAMIS CO2 quality th recommendations.pdf, viewed 27 August, 2010
3
Chapoy, A., Burgass, R., Tohidi, B., Austell, J.M. and Eikhoff, C. “Effect of Common Impurities on the Phase Behavior of Carbon Dioxide Rich Systems: Minimizing the Risk of Hydrate Formation and TwoPhase Flow.” Presented at SPE Offshore Europe Oil & Gas Conference, Aberdeen, 8-11 September 2009. th
PCC, Istanbul, 12 October 2010
-2-
careful review. In CO2 derived from natural gas operations, the principal non-H2S sulfur components are likely to be mercaptans. In the IGCC scenario it is likely to be COS. Oxygen is an important issue for EOR operations, since it may assist the propagation of bacteria present in the underground reservoir. For saline aquifers specifications may be relaxed, but this is generally not an issue for an IGCC plant. Non-condensables (in some specifications erroneously described as “inerts”) include O2, N2, Ar, CH4, CO and H2. Transport considerations limit the total non-condensables to about 4% to ensure that the CO2 stream remains in the supercritical phase in the pipeline for a reasonable input of CO2 compression power.. Note that none of the specifications in Table 1 contain any constraint on the levels of carbon monoxide, nor do they address hydrogen. Similar safety considerations need to be applied as for H2S. The DYNAMIS recommendations suggest limiting CO to 2000 ppm.
2.2 Constraints Imposed by Venting Venting a CO2 stream which adheres to these pipeline specifications will need to address a number of additional health and safety issues. The principal components that will need addressing are H2S, COS and CO. A Best Available Control Technology (BACT) analysis may be required to determine if further treatment (such as oxidation of the stream) is required. This is a case-by-case analysis, and the regulatory agency determines whether such treatment is cost-effective and required. Further, a dispersion analysis is required to assure adherence to air quality standards in the surrounding area. If the first dispersion calculations show that the air quality standards are not met, the designer has a number of potential corrective measures at his disposal: •
Increase the height of the vent stack.
•
Increase the temperature of the CO2 stream. Typically the stream emitted by the AGR will be at a lower temperature than ambient. In the case of Selexol it may be of the order of magnitude of 5°C. For Rectisol, the lower operating temperatures make refrigeration recovery essential, but the temperature of the CO2 stream is still likely to be about 20°C. Given that CO2 is in any case heavier than air, considerable heating may be necessary to achieve the required ground level air quality standards.
•
Oxidize the components to SO2 and CO2, for which higher concentrations are allowed.
•
Reduce the levels of the contaminants in the CO2 stream. This is discussed further in Section 3.2.
•
Limit the hours that the CO2 stream can be vented, so as to stay under the value that triggers the need for the Best Available Control Technology analysis and potential additional treatment.
•
Consider building in a fuel switching capability. This is explained and discussed further in Sections 2.4 and 4.3.
2.3 Capture Rate “Carbon capture” can refer to a number of different scenarios in which CO2 is captured and compressed for use or for storage in deep geologic formations. The percent capture can vary from a case where almost all of the carbon in the syngas is captured, to cases where much less is captured; for example, where only the CO2 naturally present in the syngas stream from a particular gasifier is captured, sometimes known as CO2 skimming. In the latter example, the percent capture can vary widely depending on the gasifier and coal specification. If the CO in the syngas is shifted to CO2, then the percent of CO2 in the shifted syngas will increase. The percent capture of whatever amount of CO2 is in the syngas to the AGR will also vary based on the plant designer’s choice, with the usual practical maximum extending up to a nominal 90%, depending upon gasifier type, gasifier operation, and equipment break points. th
PCC, Istanbul, 12 October 2010
-3-
Further, the percent capture of syngas carbon required to achieve a specific CO2 Emission Intensity (lb. CO2/net MWh); e.g., natural gas combined cycle equivalence, varies with coal rank and other design basis and plant performance parameters. Table 2 provides an overview of typical ranges of capture rate, CO2 emissions and technology applied to achieve these rates. The ranges given are fairly broad to take account of the difference between, say, Illinois #6 bituminous and Powder River Basin sub- bituminous coals. Table 2 CO2 Capture Rate Options Capture Case
% of Syngas Carbon that is Captured
Approximate CO2 Emissions from Plant lb CO2/ net MWh
Method
IGCC with Uncontrolled CO2
0%
1800 – 2000
None
Syngas CO2 Capture Only (Skimming)
3 % – 29 %
1400 – 1700
No Shift (highly dependent on gasifier type)
NGCC Equivalent CO2 Capture
40 % – 60 %
720 – 1100
Single Stage Shift / Syngas Combustors
Single Stage Shift Max
60 % – 80 %
400 – 720
Single Stage Shift / H2 Combustors
Practical Maximum CO2 Capture
80 % – 90 %
180 – 400
2-Stage Shift / Selexol / H2 Combustors
Examples for different design bases can be seen in projects currently in planning. Hydrogen Energy 4 California’s IGCC plant is designed for 90% (or full) carbon reduction . Mississippi Power’s Kemper 5 County IGCC Project is designed for 65% capture or natural gas equivalency. Duke Energy’s Edwardsport IGCC plant is being constructed without CO2 capture, but “capture ready”. Preliminary studies for retrofitting the capture facility reviewed several of these options and a FEED study is 6 currently underway . Given the quality of raw syngas to be expected of the GE gasifiers being used, skimming is in this case also a possibility.
2.4 Operational Requirements For IGCC configurations with CO2 capture there will be a need for a CO2 vent system to accommodate start-up and shutdown operations and interruptions in CO2 export capability. The CO2 vent specifications may be tighter than the CO2 pipeline specifications as discussed earlier. Thus the CO2 vent stack may need costly environmental controls to meet local specifications for trace contaminants such CO, COS, H2S, etc. If required, the BACT analysis may call for controls and an oxidizer or possibly a CO2 Purification/Scrubber unit may need to be foreseen in the design. Alternatively the owner may incorporate vent duration limits into the permit and operating manual. Another alternative is to consider the possibility of modifying the gas treatment system so that the gas turbine fuel can be conventional (un-shifted) syngas instead of raw hydrogen. Incorporation of such “fuel switching” capability for the gas turbine can also have a significant value to the power company 4
www.hydrogenenergycalifornia.com, viewed at 27th August 2010.
5
www.mississippipower.com/kemper/EnvironmentalBenefits.asp, viewed at 27 August 2010.
th
6
Zupan, D. M., “Edwardsport IGCC – Moving Forward.” Paper presented at Gasification Technologies Conference, Colorado Springs, CO. October 2009. th
PCC, Istanbul, 12 October 2010
-4-
allowing reserve credits for the potential power increase available when the CO2 capture and compression system is not being operated.. These credits can be available by demonstration of the capability rather than by actual venting of the CO2. Some of these issues are examined in more detail later in this paper.
3 Design Options 3.1 Basic Flow Scheme The typical, but simplified block diagram of an IGCC with CO2 capture shown in Figure 1 emphasizes those features that might be the subject of discussion in evaluating options for addressing scenarios when CO2 cannot be exported.
Figure 1
Block Flow Diagram of IGCC with CCS
Although there are many examples of using a syngas cooler and sweet shift configuration in chemical applications such as ammonia plants where 90% or more CO2 capture is required, it is assumed for the purposes of this discussion that a quench gasifier (with or without radiant cooling) and sour shift has been selected. Similarly a physical solvent will be assumed for the acid gas removal. Whether the solvent chosen is Selexol as in the Edwardsport and Kemper County IGCC plants or Rectisol as selected for HECA, most of the issues will be similar and for this example we have chosen Selexol.
3.2 Influences on CO2 Quality Influence of CO Shift on CO in CO2 Stream The characteristics of a physical wash AGR are such that the more CO there is in the feed stream, the more CO will appear in the CO2, unless countermeasures are taken in the AGR. Thus increasing the amount of shift will decrease the CO concentration in the CO2. This is, however, equivalent to increasing the overall capture rate, which will be fixed by the CO2 off-take contract and thus is an option that is unlikely to be attractive.
th
PCC, Istanbul, 12 October 2010
-5-
Influence of CO Shift on COS in CO2 Stream In the basic design of an IGCC plant with CO2 capture the use of a raw gas or sour shift will generally provide sufficient conversion of COS to H2S to eliminate the necessity for a dedicated COS hydrolysis unit as is typical of non-capture designs using Selexol or amine-based AGRs. If the design is to incorporate fuel switching, bypassing the entire syngas stream around the shift catalyst will be required. This will require incorporating a COS hydrolysis unit in the bypass stream so that the COS conversion effect of the shift catalyst is maintained in order to meet the sulfur specification of the AGR. In order to be able to react quickly to a loss of CO2 export capability, the COS hydrolysis catalyst must be maintained at operating temperatures during normal operation. EPRI has reviewed alternative methods for this. One could locate the COS hydrolysis in the bypass, but this requires a small stream in the bypass to maintain the necessary standby temperature. This has the advantage that it would not create additional pressure drop, but it does have negative implications for the capture rate. The preferred solution is to locate the COS hydrolysis unit in the main path downstream of the shift. This has the advantage of further reducing the COS slip into the AGR and thence into the CO2 stream even during normal operation. The penalty is the additional pressure drop.
Influence of AGR on Sulfur Species in CO2 Stream A typical two-stage Selexol unit as applied to an IGCC with CO2 capture is shown in Figure 2.
Figure 2
Typical 2-stage Selexol unit with HP CO2 flash highlighted
In the Selexol solvent H2S is absorbed much more easily than COS. Thus the predominant sulfur compound in the gas leaving the H2S absorber is COS. The solvent circulation rate in the CO2 absorber is much higher than in the H2S absorber and most of this residual sulfur is therefore removed from the clean gas (now mostly hydrogen) and is released in the CO2 stream. Given that the feed gas contains 30% or more CO2, the COS concentration in the CO2 stream can be as much as three times what it is in the desulfurized gas. Thus, the better the COS conversion upstream the AGR, the lower the total sulfur in the CO2. th PCC, Istanbul, 12 October 2010 -6-
Influence of AGR on CO and H2 in CO2 Stream The CO2 flash system shown in Figure 2 provides the designer with the most influence on the CO and H2 contents in the CO2 stream. Both H2 and CO are to some degree co-absorbed with the CO2 in the CO2 absorber. This represents a potential loss of hydrogen to the gas turbine as well as a contamination of the CO2 stream. In the first (highlighted) flash stage H2 and CO are preferentially flashed out of the solvent leaving most of the CO2 in solution. The flash gas is recycled to the H2S absorber and thus recovered. The pressure at which this first flash stage is operated is decisive for the quality of the CO2 – but also for the utility demand of the AGR. As can be seen in Figure 3, which was calculated for a typical example, flashing at high intermediate pressures leaves a considerable quantity of CO in the CO2 stream. The compressor load is however low. Not only is the pressure ratio of the compressor small, but more importantly only a small amount of CO2 is flashed out with the H2 and CO. Reducing the pressure of this first flash stage can in the extreme case reduce the CO concentration by a factor of 10 – but at a 10-fold increase in required compressor power. Actually the utility effect is even worse than this because the pressure for the second flash stage will also be reduced, further increasing the load in the main CO2 compressor.
Figure 3
Recycle compression energy and CO2 purity as a function of HP flash pressure.
4 Operating Scenarios 4.1 Start and stop A certain amount of CO2 venting will inevitably be required. At start up the vent will be operated until the CO2 compressor and pipeline system are functional. Normal operating procedures may be able to keep this to a minimum, but at the first start up, it may be some time before the whole transport and storage system is bedded in. Again, normal operating procedures will minimize venting during a planned shutdown. However when there is an unplanned shutdown in the compression and transport system, it may be several hours before the cause is determined. It is likely that venting would be required during this time, until it is determined whether a complete plant shutdown is required or not. th
PCC, Istanbul, 12 October 2010
-7-
4.2 Vent controls If needed, an oxidizer can be used in the CO2 vent to reduce CO, H2S and other trace pollutants in the stream being vented to atmosphere. There are three types of oxidizers available: •
Catalytic Oxidizers
•
Regenerative Thermal Oxidizers
•
Recuperative Thermal Oxidizers
The thermal oxidizers for a CO2 vent operate at considerably higher temperatures and require a significant amount of natural gas for operation and standby. Catalytic oxidizers operate at lower temperature ranges and are very good at CO destruction but not for higher hydrocarbons. (While the hydrocarbons issue may be important in oil and gas applications, it is unlikely to be a issue for IGCC plants.) Other considerations include the shorter heat up time for the catalytic oxidizer (a benefit) and the need for catalyst bed replacement (a deficit). A summary of the three options for oxidizing the CO in the vented CO2 stream is provided in Table 3. Table 3
Summary of CO2 vent options for CO removal Catalytic Oxidizer
Regenerative Thermal
Recuperative Thermal
Design Types
Fixed Bed (Pt or Mg), High-efficiency Heat Exchanger
Alternating hot-cold ceramic beds
Feed / Prod. Multi-pass Heat Exchanger
Op. Temp. Range, °F
500 - 600
1400-1600
1400-1600
Standby Temp, °F
450-500
1200-1300
1200-1300
HX Thermal Efficiency
98%
95%
70%
CO Destruction
>99%
>99%
>99%
Limits
Poor destruction of C1+ Bed replacement
Steady state, not suited for temperature swings O2 in feed gas
HX replacement
Switchover time
15-30 min. hot 24 hrs. cold
1 hour hot 24 hrs. cold
1 hour hot 24 hrs. cold
Irrespective of the technology applied, considerable capital cost is involved, but more important is the operating cost involved in maintaining such systems at standby temperature, should the need to vent occur.
4.3 Process Interruption and Fuel Switching In an IGCC plant designed for CO2 capture, situations will arise where the operation of the CO2 compression, transport or sequestration facilities are interrupted and the CO2 may have to be vented. An alternative to venting the CO2 stream with its contaminants is the possibility of operating the IGCC plant in conventional syngas mode, whereby the gas turbine fuel is clean, un-shifted syngas. The carbon would then be rejected as CO2 in the gas turbine exhaust. Motivations for reviewing this option include: •
Venting, if allowed, still involves much of the economic penalty for carbon capture (only CO2 compression is avoided) without the benefit of CO2 reduction. In particular this includes the higher heat rate that occurs due to the addition of the CO2 capture system. th
PCC, Istanbul, 12 October 2010
-8-
•
Subject to a possible short period of venting allowed by the permit, the only alternative may be a plant shutdown.
•
Switching to syngas firing improves the heat rate substantially – almost to the level of a noncapture IGCC. Carbon and sulfur species are oxidized in the gas turbine, possibly obviating the need for costly emission control equipment in the CO2 vent.
The latter point is illustrated by the following example of a current generation two-train IGCC using modern F class gas turbines. Operation with CCS provides a power output of 539 MW e at a heat rate of 11,254 BTU/kWh (~30.3% efficiency, HHV basis). With this performance one would receive the economic benefit of avoided CO2 in whatever form the regulatory environment had dictated. If the CO2 were vented, the net power output would increase by eliminating the CO2 compression power. Depending on other variables such as the necessity for an oxidizer, running the first stage flash at a lower pressure or being able to use extraction air from the gas turbine in the air separation unit, the power output might increase to about 580 MW e. For the same fuel input the heat rate would drop to about 10,500 BTU/kWh (32.5% HHV). Since under these conditions the CO2 benefit would no longer be available, this operation becomes extremely unattractive and the plant is not likely to dispatch anyway. On the other hand, switching the gas turbine operation to syngas firing would remove losses in the CO shift and the CO2 removal units. Output would increase to the 620-640 MW e range and the heat rate would drop to about 9100 BTU/kWh (37.5% HHV). For future plants the efficiencies would be higher, but the differentials would remain in the same order of magnitude. Issues requiring review during the plant design include the following: •
COS Conversion In the basic configuration with CO-shift, COS hydrolysis may not be required, because the CO-shift will perform this duty – even if not quite to the same degree as a pure selective COS hydrolysis. When bypassing the CO-shift upstream of a Selexol or MDEA AGR, a COS hydrolysis unit will probably be necessary to maintain the sulfur slip to the gas turbine at an acceptably low level, even for a limited operation period.
•
AGR Design The AGR must be designed specifically for operation in both two-stage (H2S and CO2) and single-stage (only H2S) modes. In the two-stage mode the operation of the H2S absorber relies on the CO2 pre-loading of the semi-lean solvent for efficient operation. This must be addressed differently during single stage operation. The likely result is that at least one additional solvent chiller would be required.
•
Gas Turbine Fuel Flexibility and Switching The gas turbine must be capable of accepting the change of fuel. In particular the switch-over control system must ensure that there are no excursions in the Modified Wobbe Index (MWI) = LHV/√(sg.T) of the fuel at the burner. The following capabilities/considerations need to be taken into account to switch from highhydrogen to syngas fuel: •
Fixed orifice fuel nozzles in gas turbine combustors generally have flexibility for +/10% Modified Wobbe Index (MWI).
•
Estimates show that MWI varies by about 10–15% across capture rate scenarios from 0–90%.
•
Fuel heating value must also be controlled above the minimum.
Gas turbine manufacturers usually limit the range to ±10% and the difference between normal syngas and high hydrogen syngas for low CO2 capture rates is slightly more than this range. In many cases the gas turbine designer can build in some flexibility to accommodate the larger range. th
PCC, Istanbul, 12 October 2010
-9-
There are several ways to control the MWI including those which will work for a conventional IGCC configuration: •
Fuel heating (T) if necessary can be used to ease Wobbe issues.
•
N2 and water diluent (sg) can be varied.
•
Normally N2 is the diluent for NOx control and power augmentation in CO2 capturecapable IGCC plants, but water saturation can be used for non-capture IGCC plants. Varying the mix (sg) can overcome MWI limitations.
•
N2 can be injected upstream and/or in the combustors .
•
Varying the mix may be able to overcome any MWI limitations and balance the heating value. Each manufacturer has its own rules.
Several projects have looked at simply injecting captured CO2 into the gas turbine but so far none have adopted this option. The following considerations have been identified:
•
•
Injection of CO2 requires compressor equipment. For a 630 MW e plant this requires about 4 MW and continued operation of the CO2 capture system requires about 10 MW along with other losses. Most of the N2 needs to be vented with a penalty as it typically exits the ASU coldbox at a somewhat elevated pressure. Furthermore, this configuration may require a separate CO2 compressor.
•
The CO2 amount is less than the N2 diluent so heating value can be controlled by backing out N2 and controlling the Modified Wobbe Index.
•
CO2 is more effective at controlling NOx than nitrogen.
•
CO2 injection may be feasible from the gas turbine point of view but probably only as an interim action.
System Dynamics The above considerations consider initially only the start and end points for such a switchover. Clearly the real advantage is to be able to perform the switchover online within a period of time that allows the IGCC plant to continue to generate power uninterruptedly in a cost-effective manner, while maintaining environmental compliance. EPRI has made a preliminary review of such a system and the necessary controls required. The conclusion is that such an online switchover could be achieved. Further work in this area is continuing.
•
Costs The principle cost involved is the inclusion of a COS hydrolysis unit into the system. Including the ability to bypass the CO2 wash will also involve some additional CAPEX. The total for a nominal 630 MW 60Hz IGCC is estimated to be less than US$ 10 million. This is a small penalty to pay to obtain the benefits of a fuel switching capability.
5 Conclusions CO2 composition specifications for EOR or geologic storage vary and there is no clear standard yet. The need to deal with planned and unplanned interruptions in the ability to export CO2 may impose more restrictions on the designer than the pipeline specification. The designer has a number of options available. Including the capability of fuel switching into the plant design appears to be an attractive alternative.
th
PCC, Istanbul, 12 October 2010
- 10 -
Manuscript Not AVAILABLE
Abstract Submission
PROGRAM TOPIC: Carbon Management: Transportation Infrastructure and Issues ENVISIONING CO2 DISTRIBUTION NETWORKS FOR CARBON CAPTURE AND STORAGE (CCS) IN THE UNITED STATES: STRATEGIES FOR CO2 PIPELINE DEPLOYMENT AT A REGIONAL SCALE Mr. Nils Johnson PhD Candidate Institute of Transportation Studies University of California, Davis One Shields Avenue Davis, CA 95616 415-307-7535 [email protected] Dr. Joan Ogden Professor, Environmental Science and Policy Institute of Transportation Studies University of California, Davis One Shields Avenue Davis, CA 95616 530-752-2768 [email protected]
If CO2 emissions are regulated in the future to address climate change, technologies for reducing the emissions associated with coal use must be implemented. The leading approach for achieving reductions is carbon capture and storage (CCS), where CO 2 is captured at the source, compressed, and transported for storage in onshore or offshore underground geologic reservoirs. CCS infrastructure will require large financial investments and research is needed to identify cost-effective deployment strategies. This paper uses a CO2 pipeline optimization tool and regional geographic data to examine the viability of regional CO 2 disposal networks for various CCS adoption scenarios in the United States. Geographic specific case studies are conducted for each of the U.S. Regional Carbon Sequestration Partnership territories and aggregated into a national assessment. We examine several approaches for building up regional CO2 sequestration systems. In one case we look at independent pipelines, connecting individual CO 2 sources with sinks. In the other case we look at the development of interconnected regional CO2 pipeline networks. Regional networks have been proposed as a means for reducing transportation costs through scale economies. However, it is unclear when and if these regional networks will be competitive with independent pipelines connecting individual sources and sinks. The CCS adoption scenarios will specify the adoption of capture technologies at different types of point sources over time (e.g., sites with low cost capture will adopt CCS first). The pipeline optimization tool will use the location of sources and sinks within each region, the scale of CO2 flows, and the capacity and type of geologic disposal sites to identify the CO2 pipeline network that minimizes both transport and injection costs. Not only will the proximity of sources and sinks impact the viability of regional networks, but also the capacity and type of available disposal sites. For example, it is possible that a more extensive regional network will be developed to take advantage of low-cost disposal sites (e.g., enhanced oil recovery sites) in cases where the reduction in injection costs exceed the increase in transport costs. Our optimization tool is uniquely capable of examining the trade-offs between transport and disposal costs in designing CO2 distribution networks in real geographic regions.
It is important to determine whether and when an organized regional pipeline network will become economically viable in order to identify whether there is a business case for regional CO2 disposal services. Potential cost advantages of a regional disposal network are scale economies in CO2 pipeline transport, and the ability to utilize a variety of higher quality disposal sites, rather than just the closest sites. This study will provide maps showing optimal strategies for the deployment of CO2 disposal infrastructure in the United States over time and quantify the cumulative required capital investment, levelized cost of disposal ($/tonne CO2), cumulative CO2 stored, and regional limits on CO2 storage capacity.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: LEGAL AND REGULATORY ISSUES EXACT TITLE OF PAPER: Linking Economic and Technological Modeling of CCS and Legislative Policy for Coal Mining Companies
Pratt Rogers, Research Assistant, University of Arizona Contact Information: [email protected] Sean Dessureault, Associate Professor, University of Arizona Contact Information: [email protected]
Abstract:
The combustion of coal releases the greatest percentage of CO2/kWh of all fossil fuels used in the generation of electricity. According to the Energy Information Agency (EIA), under proposed carbon legislation, the cost of coal steam generation increases dramatically. CCS will be one component of the energy generation options within this highly complex regulated energy sector and will be an integral part of coal’s future as a source of electricity generation. There are a multitude of economic models projecting the impact of proposed legislation targeted at reducing carbon emissions. These models offer broad generation projections based on macroeconomic, supply, demand, and conversion assumptions. Due to the effects of these different legislative proposals and CCS technology developments, coal mining companies will need to understand the price and fuel product (mineralogical) implications. An integration of these many different economic models would be needed. The first step in designing an integrated model is to characterize the existing models by determining their inputs, outputs and feedback loops. The proposed paper will review model characterization results from
the key existing public, and to the degree permitted, private economic models. Development of an integrated economic and technological model targeted for coal-mining companies will help them develop investment strategies under the different proposed legislation and mineralogical requirements of CCS technology.
27th Annual International Pittsburgh Coal Conference, October 11-14, 2010 Istanbul, Turkey
CARBONACEOUS EMISSIONS REFLECTED IN DEPOSITS ON BUILDING STONES: CASE STUDY IN PRAGUE CASTLE Ivana Sýkorováa, Martina Havelcováa, Antonín Zemanb, Hana Trejtnarováa, Petra Matysováa, Alexandr Šulca a) Institute of Rock Structure and Mechanics AS CR, v.v.i., V Holešovičkách 41, 182 09 Prague, Czech Republic b) Institute of Theoretical and Applied Mechanics, AS CR, v.v.i., Prosecká 76, 190 00 Prague, Czech Republic
INTRODUCTION Formation of thin black surface layers or black coloured gypsum crusts can be observed on stones of many buildings and sculptures around the world. Since the last century, degradation of natural and artificial stones has been accelerated by an increase of air pollution, essentially caused by sulphur and nitrogen oxides as well as hydrocarbons, which are emitted into the atmosphere by sources related to industry, transportation and heating (Grimm and Schwarz 1985; Bonazza et al. 2009). Also biological activity in surface layers of the material contributes to the weathering of building stones and monuments (Cappitelli et al. 2010). One aspect of the weathering process is formation of thin black surface layers or black coloured gypsum crusts on building stones and sculptures. These black surfaces not only impair the aesthetic aspect of (historic) sights, but they also initiate destructions and weathering of stone surfaces. Very important results were obtained from the extensive study of degradation of building stones (Dudková 1977; Pospíšil et al. 2006, Přikryl and Smith 2007) and biodegradation of surface layers (Wasserbauer 2000) of buildings in the Czech Republic. Sandstones of the sculptures from the medieval Charles Bridge in Prague were studied in relation to the air pollution and weathering effect using detailed mineralogical study and geochemistry of stable isotopes. It exhibits varying weathering phenomena that are manifested mostly by surface blackening and locally by gypsum crusts formation, strongly supported by anthropogenic atmospheric sources of sulphur (Přikryl et al. 2004). Interesting research results brought the study of impacted mortars, which are an integral part of the historic buildings, where the stones are bonded or repaired by them (Zeman and Růţičková 2004). Wasserbauer (2000) systematically dealt with biological impoverishment of buildings, a role of microorganisms in the weathering of building stones and methods of disposal of biological crusts. He found that black sulphate crusts on the Charles Bridge spans are spots of strong corrosion activity of chemolithotrophic and chemoorganotrophic bacteria and fungi. Therefore, the principal aim of this study was to identify the carbonaceous particles and isolate the organic compounds present in the black weathered stone surfaces coming from chosen parts of Prague Castle. SAMPLES AND METHODS Sampling and sample treatment for analyses was carried out uniformly and gently with regard to the limited amount of study material and multiple methods under the expert guidance of the staff from the Heritage Conservation Department of the Office of the President of the Republic, namely, Dr. Peter Chotěbor and Mgr. Petr Měchurka. Two samples
of black layers were collected from Saint Vitus’s Cathedral (Fig. 1) and one sample from a balcony of the historical building on the 3rd ward in Prague Castle. Limited amount of material puts demands on the modification of microscopic and chemical methods to gain reproducible results in accordance with the preliminary study of mortars (Zeman and Růţičková 2004). Identification and characterization of dispersed organic matter in black crusts and weathered layers of stones from studied buildings and sandstones from mined deposits were based on optical microspectrophotometry of polished blocks in reflected, polarized and fluorescence light, and on optical microscopy of thin sections in transmitted white and blue light in accordance with the methodology published by Taylor et al. (1998) and Sýkorová et al. (2009). Chemical analyses of organic compounds of anthropogenic and microbial origin were used for detailed characteristics of surface layers of the monitored sites and buildings. Gas chromatography with mass spectroscopy (GC/MS) and pyrolysis gas chromatography (PyGC/MS) provide a relatively detailed information on compounds of an anthropogenic and biological origin, including their degradation products based on biomarkers present (Nord et al. 1994; Machill et al. 1997; Sýkorová et al. 2009). GC/MS analysis is used for detecting and identifying substances soluble in organic solvents (Orecchio 2010) and thus results depend on extractability of samples tested. Greater emphasis was placed on Py-GC/MS method that allows rapid heating of the sample and its thermal decomposition to obtain compounds, and those directly identify by GC/MS. The need of a minimum quantity of analyte and its preparation for the determination is particularly suitable for these purposes when only a small amount of material is available. RESULTS Prague Castle is located in the centre of Prague – the capital of the Czech Republic – and is exposed to natural weathering influences, such as rain, wind, temperature or pressure. However, the main factors accelerating stone decay processes are miscellaneous pollutants to which stone surface is exposed. The approach applied on sandstone surfaces used here was a combination of non-destructive (optical microscopy) and destructive (gas chromatography and pyrolysis-gas chromatography/mass spectrometry, Py-GC/MS) analytical techniques. Microscopically, we found an authigenic gypsum formation of a thickness up to 1 mm at the surface of porous sandstone with the upper layer formed by admixtures of fine grains of quartz, clay minerals, bricks, fly ashes and organic particles. High reflecting combustion residues, namely soot and chars, prevail in all samples and contain a lower portion of the low reflecting modern organic matter such as wood tissues, plant remains, pollen grains, algae and fungal spores (Fig. 2).
Fig. 1. Irregular black crust on the building element of St. Vitus Cathedral.
Fig. 2. Mixture of carbonaceous particles and minerals in black crust - detail of sandstone building element. Reflected light, oil immersion.
Total organic carbon content (TOC) varied from 0.08% to 4.28%. GC/MS data of the dichloromethane extract did not show any significant amount of the organics typical of pollution such as polycyclic aromatic hydrocarbons or other organic compounds. The only substances found were phthalates. They are used in a large variety of industrial products and they are being released in the environment. Pyrolysis-GC/MS analysis was also used as a technique to determine organic substances. The pyrograms of the samples showed a range of alkenes, polyaromatic hydrocarbons and other molecular markers. Benzonitrille and nitrogen heterocyclic compounds were the most abundant compounds in the pyrolysis products. Traces of microflora in black crusts of indicative samples from selected parts of Saint Vitus’s Cathedral might be documented by benzonitrile, which was the dominant compound in the pyrolysis products of Py-GC-MS analysis. Benzonitrile and nitrogen heterocyclic compounds are pyrolysis products of proteins containing phenylalanine, which is found in all organisms. Some bacteria and algae can produce benzenonitriles. Absence of aliphatic nitriles, amides, and other biodegradation products such as furans or pyrroles indicates that the action of microorganisms was less significant. Benzonitrile and many other hydrocarbons such as naphthalene, biphenyle, and polycyclic aromatic hydrocarbons (Fig. 4) are also products of incomplete combustion and a source of relatively stable forms of ‘black carbon’ (Sýkorová et al. 2009). Identified compounds are derived mostly from the exhaust fumes. Benzonitrile is a product of vulcanization or asphalt processing, but this substance is also a marker for Black Carbon – a kind of carbonaceous particles released into the atmosphere mainly from an incomplete combustion of fossil fuels and biomass. Black Carbon particles are resistant in the environment and are able to absorb other organic particles, for example polycyclic aromatic hydrocarbons, styrene or dibenzofurane. These organic compounds were identified in the pyrolysates of the analysed samples and served as an evidence of fuelburning pollution. The most contaminated or polluted sample was from the south part of Saint Vitus’s Cathedral.
CONCLUSIONS Samples of weathered surfaces of sandstones from Prague Castle contained organic particles and compounds, which are known to result from human activities, particularly from combustion processes. Biological production played a significant role on the complex weathering process.
ACKNOWLEDGEMENTS The research was funded by the Grant Agency of the Academy of Sciences of Czech Republic (project No. IAA300460804). REFERENCES Bonazza A., Mesina P., Sabbioni C., Grossi C.M., Brimblecombe P., 2009. Mapping the impact of climate change on surface recession of carbonate buildings in Europe. Science of The Total Environment, 407, 2039-2050. Cappitelli F., Toniolo L., Sansonetti A., Gulotta D., Ranalli G., Zanardini E., Sorlini C., 2010. Advantages of using microbial technology over traditional chemical technology in the removal of black crusts from stone surface of historical monuments. Appl. Environ. Microbiol. doi: 10.1128/AEM.00394-07.
Dudková I., 1977. „Nemoci“ kamenů a jejich hlavní příčiny. Časopis pro mineralogii a geologii, 22, 3, 225-237. Grimm W.D., Schwarz U., 1985. Naturwerksteine und ihre Verwitterung an Münchener Bauten und Denkmällern. Arbeitsheft. Bayerisches Landesamt Denkmalpflege, 31, 27118. Machill S., Althaus K., Krumbein W.E., Stegew W.E., 1997. Identification of organic compounds extracted from black weathered surfaces of Saxonean sandstones, correlation with atmospheric input and rock inhabiting microflora. Organic Geochemistry, 27, 1/2, 79-97. Nord A.G., Svärdh A., Tronner K., 1994. Air pollution levels reflected in deposits on building stone. Atmospheric Environment, 28, 16, 2615-2622. Orecchio S., 2010. Analytical method, pattern and sources of polycyclic aromatic hydrocarbons (PAHs) in the stone of the Temples of Agrigento (Italy), J. Hazard. Mater., 176, 339-347. Pospíšil S., Drdácký M., Slíţková Z., Knotková D., Deplech Ph., 2006. Surface degradation of complex architectural form due to atmospheric pollution, International Conference „Heritage, Weathering and Concervation, Madrid, 21-24 June 2006, 447-452. Přikryl R., Smith B.J. (Eds), 2007. Building Stone Decay: From Diagnosis to Conservation. Geological Society Special Publication, 271, 9-21, 330 pp. Přikryl R., Svobodvá J., Ţák K., Hradil D., 2004. Anthropogenic origin of crusts on sandstone sculptures of Prague´s Charles Bridge (Czech Republic): Evidence of mineralogy and stable isotope geochemistry. European Journal of Mineralogy, 16, 4, 609-617. Sýkorová I., Havelcová M., Trejtnarová H., Matysová P., Vašíček M., Kříbek B., Suchý V., Kotlík B., 2009. Characterization of organic matter in dusts and fluvial sediments from exposed areas of downtown Prague, Czech Republic. International Journal of Coal Geology, 80, 2, 69-86. Taylor G.H., Teichmüller M., Davis A., Diessel C.F.K., Littke R., Robert P., 1998. Organic Petrology. Gebrüder Borntraeger, Berlin-Stuttgart, 704 pp. Wasserbauer R., 2000. Biologické znehodnocení staveb. ABF, Praha, 280pp. Zeman A., Růţičková. E., 2004. Mineralogicko petrografický výzkum malt – část I. Spektra nátěrových hmot, 1, 86-87.
Manuscript Not AVAILABLE
Visualization of coal conversion using X-ray computed tomography QP Campbell and HWJP Neomagus School of Chemical and Minerals Engineering, North-West University, South Africa
Abstract Coal conversion processes, specifically pyrolysis and gasification, are often characterized on a laboratory scale using thermogravimetric analysis (TGA) methods. From these experiments, information about kinetics, conversion, and even structural features may be deduced. When using kinetic models, some assumptions have to be made about the structure and mechanisms of conversion. These assumptions should always be confirmed with information about the coal structure, like porosity, pore structures and surface area. Still, it remains difficult to visualise the conversion process spatially in a coal particle, which is in any event almost never homogenous. Three dimensional X-ray computed tomography is a technique traditionally associated with the medical field, but industrial tomography is already a mature technique. With more sophisticated equipment, resolutions of images have increased to such an extent, that when the technique is used for scanning coal particles, a wealth of information can be gained about the structure of the individual particles. Features like lithotype layering, structures and extent of mineral distributions, crack distributions and cleat orientations can clearly be observed. In this study, both vitrinite- and inertinite rich coal particles were selected and these particles (approximately 20 mm) were first individually subjected to 3D X-ray CT scanning, and then subjected to pyrolysis. The same particles were then scanned again after complete pyrolysis, to determine any structural changes due to the process. These particles were then subjected to CO2 gasification. The process was interrupted a number of times, and after cooling in an inert atmosphere, the individual particles were scanned again. The remaining particles, after complete conversion, were scanned for a final time. The data obtained from the TGA were then compared with structural information from the CT images, to obtain a better description of the conversion process.
For Presentation at the 27thAnnual International Pittsburg Coal Conference, October 11-14, 2010, Istanbul, Turkey
ORGANIC SULPHUR FUNCTIONALITY CHANGES IN BIOTREATED COALS L. Gonsalvesh*a, S.P. Marinova, M. Stefanovaa, R. Carleerb, J. Ypermanb a b
Institute of Organic Chemistry, Bulgarian Academy of Sciences, bl. 9, Sofia 1113, Bulgaria
Research group of Applied and Analytical Chemistry, CMK, Hasselt University, Agoralaan – gebouw D, B3590, Diepenbeek, Belgium *correspondence author – [email protected]
Abstract: During the late century, rising global concerns over the effects of acid rain led to the development and utilisation of technologies to reduce SO2/NOx emissions. The formation of SO2 occurs during the combustion of sulphur containing coals and can lead to acid rain but also acidic aerosols (extremely fine air-borne particles). A number of technologies, collectively known as a desulphurization, have been developed to reduce SO 2 emissions. Recently, biological approaches – biodesulphurization, becomes more popular, in view of the fact that processes are performed under mild conditions with no harmful reaction products and the characteristics of coal are hardly affected. Since there are sufficient chemical, physical and biological desulphurization methods for inorganic sulphur removal, the aim of the present study is to apply biodesulphurization treatments with a desulphurization potential toward organic sulphur functional forms in coal. Another target of the study is to trace the changes in sulphur. Two Bulgarian high sulphur containing coal samples, sub-butiminious (Pirin) and lignite (Maritza East), and one Turkish lignite (Cayirhan-Beypazari) are used in the experiments. In order to concentrate our efforts on a deeper research on organic sulphur, samples are preliminary demineralized and depyritized. The white rot fungi “Phanerochaeta Chrysosporium” – ME446 and the thermophilic and acidophilic archae “Sulfolobus Solfataricus” – ATCC 35091 are the microorganisms applied in the biodesulphurization processes. A requirement of any research on desulphurization is an availability of accurate method for sulphur functional forms tracing. It is important to evaluate the effect of the treatment and to select proper coals for sulphur specific desulphurization processes. Temperature programmed reduction at atmospheric pressure (APTPR) together with its variation in detection mode proves to be an effective technique. To specify the organic sulphur forms in coal and to assess the changes in organic sulphur that occurs as a result of applied biotreatments AP-TPR coupled ―on-line‖ with mass-spectrometry (AP-TPR-MS) in different gas media, i.e.H2 and He, is used. The sulphur volatile organic compounds, neither reduced in the AP-TPR condition nor captured in the tar fraction, are quantitatively determined by ―off-line‖ TD-GC/MS technique applying deuterated sulphur compounds as inner standards. Oxygen bomb combustion followed by ion chromatography is used to evaluate quantitatively organic sulphur compounds captured in the tar and char residue.
Keywords: coal; biodesulphurization; sulphur functionalities; Pyrolysis; Mass Spectrometry.
1
1.
Introduction High sulphur content hinders coal exploration. The main problem with coal burning for whatever purpose is
environment protection due to generation of highly polluted gases. Combustion of sulphur-containing compounds in fossil fuels emits sulphur oxides, which can cause adverse effects on health, environment and economy. SO2 can be the cause of sulfate aerosol formation, which leads to respiratory illnesses [1]. SO2 can react with moisture in the air and cause acid rain or low pH fogs. The acid formed in this way can be transferred to soil, damage the foliage, depress the pH of the lakes with low buffer capacity and endanger the marine life [2,3]. In general, two different methods can be used to lower the impact of sulphur oxide emissions: precombustion and post-combustion treatment methods. Several methods have been proposed to reduce sulphur content as SO2 emission before coal combustion. The applied techniques include physical, chemical and biological processes [4] and usually most of the inorganic sulphur in fossil fuel can be removed easily. However, there is one part, known as refractory organic sulphur, which is very difficult to be removed. The current methods, which can affect the refractory part, operate under extremely invasive conditions. They are very costly and produce considerable amount of carbon dioxide. Therefore lately, there is a keen interest on biodesulphurization as an efficient method with much greater potential towards organic sulphur decrease. Its benefits are that treatment requires lower capital and operating costs, it is performed under mild conditions and additional calorific value of the products are only slightly affected. There are many bacteria, mold and mixed cultures held responsible for the desulphurization of coal. Research is still carried on new generations in view to increase simultaneously the reaction time rate and removal percentage. The most studied microorganisms for coal desulphurization and sulfide ore leaching are Thiobacillus ferrooxidans and Thiobacillus thiooxidans [4-6] easily removing the inorganic sulphur up to 80-90 %. Subsequently, increasing attention is paid to the possibility of organic sulphur removal in coal by bacteria and fungi, and even 48 % organic sulphur reduction is achieved by using Sulfolobus solfataricus [7]. While biological coal desulphurization has a great deal of promise, there is also a considerable amount of research to be done before industrial application. In this study among the different options for coal desulphurization, biodesulphurization is selected for organic sulphur forms removal. The aim is to trace the changes that occur with the sulphur, its different species and forms bounded up with organic compounds and to clarify the mechanisms of biodesulphurization process in coal by using sulphur species-specific bacteria and fungi. 2.
Material and Methods 2.1. Coal Two Bulgarian coal samples - Humovitrain Maritza East (M) and Pirin sub-bituminous coal (P), and
one Turkish lignite - Cayirhan-Beypazari
(B) are considered. Coal samples under study are preliminary
demineralized [8, 9] and subsequently treated by diluted nitric acid for depyritization [10]. After that, treated samples are consecutively extracted by chloroform and cyclohexane for further investigations on elemental sulphur changes as a results of the biotreatments. Demineralized and depyritized samples are marked as APF (Ash and Pyrite Free). The data characterizing the initial and demineralized, depyritized coal samples as well as biodesulphurized coals are given in Table 1. In this table, organic sulphur (So) is calculated by the difference from total sulphur (St) and the sum of pyritе sulphur (Sp) and sulphate sulphur (Ss).
2
Table 1 Characteristics of the coal samples Sample
Proximate analysis (%)
S content (%), daf
Calorific value
Ash
VM
W
Fix C
St
Sp
Ss
So
(MJ kg-1)
P-initial
10.36
30.97
6.45
52.22
5.49
0.57
0.54
4.37
22.68
P-APF
0.01
39.66
6.16
54.17
4.10
0.23
0
3.87
23.65
P-APF-PC
0.53
39.8
6.6
53.05
3.11
0.16
0
2.95
23.08
P-APF-SS
1.69
38.13
5.34
54.85
3.53
0.20
0
3.33
26.3
M-initial
8.47
40.76
8.06
42.71
7.06
1.64
1.54
3.88
19.29
M-APF
1.2
46.97
6.25
45.56
4.01
0.37
0
3.64
19.93
M-APF-PC
0.89
46.34
7.37
45.4
3.04
0.20
0
2.84
20.15
M-APF-SS
1.43
45.65
6.99
45.93
3.87
0.38
0
3.49
24.24
B-initial
29.1
23.55
12.34
35.01
7.62
1.68
1.23
4.72
16.51
B-APF
0.64
40.56
6.49
52.31
3.49
0.38
0
3.11
22.03
B-APF-PC
0.96
40.87
6.19
51.98
3.13
0.11
0
3.02
22.58
B-APF-SS
1.14
39.21
6.37
53.28
2.90
0.36
0
2.53
23.00
2.2. Biodesulphurization of coals The white rot fungi Phanerochaeta Chrysosporium – ME446 and thermophilic bacteria Sulfolobus solfataricus are the used microorganisms. Biotreated coal samples are separeted by filtration from the media and then washed with 5% HCl (aq.) solution and hot distilled water. 2.3. Analysis 2.3.1. AP-TPR “on-line” MS analysis To specify the organic sulphur forms in coal and to assess the changes in organic sulphur that occurs as a result of applied biotreatments, AP-TPR coupled ―on-line‖ with mass-spectrometry (AP-TPR-MS) in different gas media (H2 and He) is used. Apparatus and experimental procedure of the AP–TPR are described elsewhere [11]. In each experiment, approximately 40 mg of sample and 25 mg of fumed silica are placed in the reactor under a 100 ml min -1 flow of pure hydrogen or helium. A linear temperature program of 5 °C min-1 from ambient temperature up to 1025 °C is followed. TPR reactor is coupled ―on-line‖ with a mass spectrometer (FISONS-VG Thermolab MS) through a capillary heated at 135°C. The mass spectrometer equipped with a quadrupole analyzer is set at an ionising voltage of 70 eV. The MS signals m/z 10 ÷ 160, are monitored ―on-line‖. 2.3.2. AP-TPR “off-line” TD-GC/MS analysis The above described AP-TPR system can also be used in adsorption/desorption mode in order to explore the sulphur volatile organic compounds, neither reduced in the AP-TPR condition nor captured in the tar/char residue. The outlet of the AP-TPR reactor is connected to an ice-cooled metal adsorption tubes with SilcoSteel Coating, filled with Tenax/Carbopack B/Carbosieve SIII (Marks) as adsorbents. The adsorption tubes are desorbed systematically and analyzed by a TD-GC/MS apparatus using He as carrier gas at 85 kPa at the
3
following conditions: a) Unity thermal desorber (Marks): primary desorption 20 min at 320 ºC; Cold trap at -8 ºC, heated at maximum heating rate to 320 ºC, hold time 15 min; flow path temperature 200 ºC; b) Trace GC Ultra-Gaschromatography (Thermo Instruments): capillary column 30 meters ZB 5-MS 0.25 mm x 0.25 µ Phenomenex; temperature program – 3 min at 30 ºC, heated 8 ºC/min to 100 ºC, heated 12 ºC/min to 310 ºC, hold time 5 min; c) DSQ-Massaspectrometer (Thermo Instruments): EI spectra; Ionization energy – 70 eV; Scan range – m/z 33-480 in 0.4 s. Each tube is spiked with 200 ng d4-thiophene to quantify results for the target sulphur species. 2.3.3. Oxygen bomb combustion followed by Ion Chromatography Oxygen bomb combustion followed by ion chromatography is used to evaluate quantitatively organic sulphur compounds captured in the tar and char residue. Sulphate produced after the combustion is determined by ion chromatography instrument Dionex (DX-120), supplied with column (AS4A-SC) and conducting detector. 3.
Results and discussion Samples under study are preliminary demineralized and depyritized in order to focus our attention on the
changes that occur with organic sulphur due to biodesulphurization and to improve the information obtained by the AP-TPR technique concerning organic sulphur functionalities. Since the purpose of the study is to appreciate biodesulphurization, comments on the effect of chemical treatment will not be made. Two types of microorganisms are applied – Phanerochaete Chrysosporium (PC) and Sulfolobus solfataricus (SS). It is found that biotreatments with PC demonstrated maximum total sulphur desulphurization effect, about 25 % for P-APF-PC coal, followed by 24 % for M-APF-PC coal. Maximum organic sulphur biodesulphurization effect is achieved toward the same coal samples – P-APF-PC and M-APF-PC, respectively 24 % and 20 %. Concerning the biotreatment with SS, maximum total sulphur and organic sulphur desulphurizations effects of about 18 % and 19 % respectively, are achieved in the case of B-APF-SS coal sample. The changes that occur in heating value of the samples after biotreatment with PC are insignificant, while increasing in the heating value in the range of 5-20 % is registered after biotreatment with SS. 3.1. AP-TPR experiments coupled “on-line” with MS detection (AP-TPR/MS) 3.1.1. H2S monitoring in reducing atmosphere The normalized H2S kinetograms of AP-TPR/MS in H2 of initial and biotreated coal samples are visualized in Fig. 1. Because m/z = 34 (H2S+) and m/z = 33 (HS+) exhibit the same evolution, only the sum of both ion profiles is shown. There are always two dominant peaks: the first one at lower temperature with T max up to 400 ºC assigned to the presence of di-alkyl and alkyl-aryl sulphides [12] and second peak with Tmax after 500 ºC assigned to the presence of di-aryl sulphides, and complex thiophenic structures. This conclusion is based on the model compound approach and also on AP-TPR/MS profiles of typical aliphatic and aromatic fragments for the demineralized, depyritized coal sample and for the biotreated ones. AP-TPR/MS evolution profiles in H2 atmosphere for H2S, give us information that ratios between aliphatic and aromatic sulphur are in favour of aromatic sulphur for all investigated ash/pyrite free samples. After biotreatment with SS changes occur clearly in this ratio for M-APF-SS coal. The new peak maximum due to the biotreatments, observed at around 250ºC for M-APF-SS can be attributed either to elemental sulphur or polysulphides. After biotreatment with PC no sufficient changes in the ratios could be registered for P-APF-PC and M-APF-PC coal samples, while for B-APF-PC aromatic sulphur is more affected by the applied fungi.
4
For the initial coals compared to the biotreated ones (by PC and SS), differences in the intensity of H2S profiles during pyrolysis in H2 are also found. It should be kept in mind that the intensity changes must be handled with care as MS-detection system is a semi quantitative technique. Nevertheless for P-APF-PC and MAPF-PC, a decrease in the intensity of both peaks can be seen, demonstrating a removal of aliphatic and aromatic sulphur compounds. Different is the case for P-APF-SS and B-APF-SS. An increase in the intensity of both peaks can be noticed. Perhaps this treatment leads to the formation of sulphur containing by-products resulting from C-S and C-C bond cleavage.
Fig. 1. AP-TPR/MS (H2), m/z (33+34) kinetograms of investigated coal samples: (A) initial and biotreated with PC; (B) initial and biotreated with SS. AP-TPR/MS (H ) 2
AP-TPR/MS (H ) 2
Pirin coal treated with SS
Pirin coal treated with PC + + P-APF - mass (33 +34 )
5.00E-012
+ + P-APF - mass (33 +34 )
A
+ + P-APF-PC - mass (33 +34 )
B
+ + P-APF-SS - mass (33 +34 )
1.50E-011
1.00E-011
Intensity (mbar)
Intensity (mbar)
4.00E-012
3.00E-012
2.00E-012
5.00E-012
1.00E-012 0.00E+000
0.00E+000
100
200
300
400
500
600
700
800
900
1000
100
200
300
400
Temperature (°C)
AP-TPR/MS (H ) 2
800
900
+ + M-APF - mass (33 +34 )
7.00E-012
A
+ + M-APF-PC - mass (33 +34 )
1000
B
+ + M-APF-SS - mass (33 +34 )
6.00E-012 5.00E-012
Intensity (mbar)
4.00E-012
Intensity (mbar)
700
Maritza East coal treated with SS
+ + M-APF - mass (33 +34 )
5.00E-012
600
AP-TPR/MS (H ) 2
Maritza East coal treated with PC
6.00E-012
500
Temperature (°C)
3.00E-012
2.00E-012
4.00E-012 3.00E-012 2.00E-012 1.00E-012
1.00E-012
0.00E+000
0.00E+000
-1.00E-012 100
200
300
400
500
600
700
800
900
100
1000
200
300
AP-TPR/MS (H ) 2
500
600
700
800
900
1000
AP-TPR/MS (H ) 2
Baypazar coal treated with PC
8.00E-012
400
Temperature (°C)
Temperature (°C)
1.00E-011
+ + B-APF - mass (33 +34 )
Bay Pazar coal treated with SS + + B-APF - mass (33 +34 )
A
+ + B-APF-PC - mass (33 +34 )
B
+ + B-APF-SS - mass (33 +34 )
8.00E-012
Intensity (mbar)
Intensity (mbar)
6.00E-012
4.00E-012
2.00E-012
4.00E-012
2.00E-012
0.00E+000
100
6.00E-012
0.00E+000
200
300
400
500
600
700
Temperature (°C)
800
900
1000
100
200
300
400
500
600
700
800
900
1000
Temperature (°C)
5
3.1.2. SO2/SO monitoring It is known that oxidized sulphur forms present in the sample, are not entirely reduced in AP-TPR experiments into H2S [20]. Their SO2 and SO fragments can also be identified and so their sulphur functionalities using AP-TPR with MS detection. Therefore the profiles of both m/z 64 (SO2+) and m/z 48 (SO+) are traced (figures are not shown). The first peak (about 250°C) detected in all samples under consideration is attributed to decomposition of a part of organic sulphonic groups formed during nitric acid oxidation of coal (depyritization). The second peak appearing at about 390 °C could be assigned to disulfones and/or polysulfones [13]. This peak is absent in Maritza East coal samples, i.e. APF and APF treated with PC and SS. The third peak at about 650 °C is attributed to the presence of diaryl sulphoxides and/or diaryl sulfones [14]. Its intensity decreases for all samples except of M-APF-SS. More information for the oxidized sulphur species can be obtained by using pyrolysis in an inert atmosphere. In Fig. 2 AP-TPR/MS kinetograms in He atmosphere for ion fragments with m/z 48 and 64 are shown: (A) for initial and biotreted with PC coals; (B) for initial and treated with SS samples. All profiles visualize two main peaks. The first one at 250-280 oC refers to organic sulphonic groups decomposition. As regards tо the second peak, it could refer to diaryl sulphoxides and/or diaryl sulfones. But tracing the profiles of typical aliphatic and aromatic fragments showed that this peak must be related rather to inorganic sulphur, namely to inorganic sulfates or elemental sulphur. It is noticed that the intensity of the second peak increases with biotreatments.
Fig. 2. AP-TPR/MS (H2), m/z 64 and m/z 48 kinetograms of investigated coal samples: (A) initial compared to biotreated with PC; (B) initial compared to biotreated with SS.
6,00E-012
AP-TPR/MS (He) Pirin coal sample
AP-TPR/MS (He) Pirin coal sample +
A P-APF mass 64 + P-APF mass 48 + P-APF-PC mass 64 + P-APF-PC mass 48
5,00E-012
3,00E-012 2,00E-012 1,00E-012
B
+
P-APF mass 64 +
1.10E-011
P-APF mass 48 +
1.00E-011
P-APF-SS mass 64
9.00E-012
P-APF-SS mass 48
+
Intensity (mbar)
Intensity (mbar)
4,00E-012
1.20E-011
8.00E-012 7.00E-012 6.00E-012 5.00E-012 4.00E-012 3.00E-012
0,00E+000
2.00E-012
100
200
300
400
500
600
700
800
900
100
1000
200
300
400
500
Temperature ( C)
1,00E-011
600
700
800
AP-TPR/MS (He) Maritza coal sample
1000
AP-TPR/MS (He) 2.00E-011
Maritza coal sample
+
M-APF mass 64 A + M-APF mass 48 + M-APF-PC mass 64 + M-APF-PC mass 48
8,00E-012
6,00E-012
+
4,00E-012
2,00E-012
M-APF mass 64
1.80E-011
B
+ M-APF mass 48
1.60E-011
+ M-APF-SS mass 64
+
1.40E-011
Intensity (mbar)
Intensity (mbar)
900
Temperature ( C)
M-APF-SS mass 48
1.20E-011 1.00E-011 8.00E-012 6.00E-012 4.00E-012
0,00E+000
2.00E-012
-2,00E-012 100
0.00E+000
200
300
400
500
600
700
Temperature ( C)
800
900
1000
100
200
300
400
500
600
700
800
900
1000
Temperature ( C)
6
AP-TPR/MS (He) Bay Pazar coal sample
AP-TPR/MS (He) Bay Pazar coal sample +
+
B-APF mass 64 A + B-APF mass 48 + B-APF-PC mass 64 + B-APF-PC mass 48
B
+ + B-APF-SS mass 64
1.00E-011
+ B-APF-SS mass 48
4,00E-012
8.00E-012
2,00E-012
0,00E+000
100
B-APF mass 64 B-APF mass 48
Intensity (mbar)
Intensity (mbar)
6,00E-012
1.20E-011
6.00E-012
4.00E-012
2.00E-012
200
300
400
500
600
700
800
900
1000
100
200
300
Temperature ( C)
400
500
600
700
800
900
1000
Temperature ( C)
3.2. AP-TPR experiments coupled “off-line” with TD-GC/MS AP-TPR technique with its extension ―off-line‖ TD-GC/MS is also applied to receive additional information for the sulphur containing volatile organic compounds. The pyrolysis experiments are performed in a inert (He) atmosphere and GC/MS spectra are quantitatively interpreted. In the chromatograms of the samples under study typical products of coal pyrolysis, i.e. alkylbenzenes, alkylnaphthalenes, phenols, etc, accompanied by their sulphur containing analogues are detected. The last ones are in dominated quantity due to special sorbents applied. Some of the determined sulphur containing organic volatiles are shown in Fig. 3. The recovery of the sulphur determined by AP-TPR ―off-line TD-GC/MS and calculated on dry, ash free base by using St as reference value, varied in the range 8.33 – 14.82 %. Maximum desulphurization effect toward volatile sulphur containing organic compounds is 44.2 %. It is achieved for P-APF-SS coal sample. Most heavily attacked are the compounds with sulphur groups containing oxygen. This is judged by the reduced amount of sulphur dioxide, which even reached 66% in the case of P-APF-SS. 3.3. Oxygen bomb combustion followed by Ion Chromatography Total sulphur in the collected tar and char fractions is determined by oxygen bomb combustion technique. The obtained values show decrease in sulphur amount retained in the tar/char fractions for all treated with PC coals, while an increase in total sulphur is registered for the treated with SS coals. 4.
Conclusions Our data demonstrate that better desulphurization effect is achieved for coal samples treated with white
rot fungi “Phanerochaeta Chrysosporium”. For P-APF-PC maximum biodesulphurization due to total sulphur decreasing with 25% and organic sulphur decreasing with 24% is registered. By “Sulfolobus Solfataricus” a maximum desulphurization effect due to organic sulphur decreasing is 19 %, revealed in B-APF-SS coal sample. After biotreatments with PC no sufficient changes in the ratios between aliphatic and aromatic sulphur are found, indicating that organic sulphur desulphurization is performed by simultaneous removal of both forms. An increase in the intensity of the peaks for aliphatic and aromatic sulphur for treated with SS samples is registered. Perhaps this treatment leads to the formation of sulphur containing by-products results from C-S and C-C bond cleavage. Qualitative and quantitative interpretation of AP-TPR ―off-line‖ GC/MS study inclined us to suppose biodegradation process of complex sulphur species accompanied by removal of compounds with sulphur groups containing oxygen. The amount of sulphur dioxide released during pyrolysis is reduced with 66% in the case of P-APF-SS coal.
7
Fig. 3. Structure of S-containing organic sulphur compound determined by AP-TPR “off-line” TD-GC/MS
8
Acknowledgements The study was performed within the framework of Cooperation agreement for joint supervision and award of a doctorate between Hasselt University, Belgium and Bulgarian Academy of Sciences, BAS-Bulgarian bilateral project with Hasselt University, FWO-Flanders.
References: [1] World Bank Group. Pollution prevention and abatement handbook 1998: towards cleaner production. Washington DC: World Bank Group Publishers; 1999. [2] EPA—U.S. Environmental Protection Agency, 2006; http://www.epa.gov/air/urbanair/so2. [3] EEA
–
EU
Environment
Agency,
2001;
http://www.eea.europa.eu/publications/environmental_issue_report_2001_22/issue-22-part-10.pdf [4] Kawatra SK and Eisele TC. Coal desulphurization: High-efficiency preparation methods. 1st ed. New York: Taylor&Francis INC; 2001. [5] Güllü G, Durusoy T, Őzbaş T, Tanyolaç A, Yürüm Y, Biodesulphurization of coal, In: Y. Yürüm (ed), Clean utilization of coal, Kluwer Academic Publishers, Netherlands, 1992. [6] Rossi G. Biodepyritization of coal: achievements and problems. Fuel 1993; 72:1581-92. [7] Durusoy T, Özbas T, Tanyolaç A, and Yürüm Y. Biodesulphurization of some Turkish lignites by Sulfolobus solfataricus. Energy and Fuels 1992;6:804-8. [8] Radmacher W, Mohrhauer P. Brennstoff-Chemie 1956;37:353. [9] Jorjani E, Yperman J, Carleer R, Rezai B. Reductive pyrolysis study of sulphur compounds in different Tabas coal samples (Iran). Fuel 2006;85:114–20. [10] Marinov SP, Stefanova M, Stamenova V, Carleer R, Yperman J. Sulphur functionality study of steam pyrolyzed ‗‗Mequinenza‖ lignite using reductive pyrolysis technique coupled with MS and GC/MS detection systems. Fuel Process. Technol. 2005;86:523-34. [11] Mullens S, Yperman J, Reggers G, Carleer R, Buchanan III AC, Britt PF, et al. A study of the reductive pyrolysis behaviour of sulphur model compounds. J.Anal. Appl.Pyrolysis 2003;70:469-91 [12] Yperman J, Maes II, Van den Rul H, Mullens S, Van Aelst J, Franco DV, et al. Sulphur group analysis in solid matrices by atmospheric pressure–temperature programmed reduction. J Anal Chim Acta 1999;395(1–2):143–55. [13] Rutkowski P, Mullens S, Yperman J, Gryglewicz G. AP-TPR investigation of the effect of pyrite removal on the sulphur characterization of different rank coals. Fuel Processing Technology 2002;76:121-138. [14] Aelst JV, Yperman J, Franco DV, Van Poucke LC, Buchanan III AC, Britt PF. Energy Fuels 2000;14:1002.
9
Manuscript Not AVAILABLE
Surface Coal Mine Planning Against Large Landslides Prof.Dr. Celal Karpuz, Assoc. Prof. Dr. Levent Tutluoglu, Dr. Arman Kocal Department of Mining Engineering, Middle East Technical University, Ankara, TURKEY 066531 [email protected], [email protected], [email protected], B.Sc.Mustafa Ozdingis, M.Sc. Kıvanç Het Head, Department of Planning and Project, Turkish Coal Enterprises, TKI, Yenimahalle, Ankara, TURKEY [email protected], [email protected] Abstract Can Lignite Surface Mine of Turkish Coal Enterprises (TKI) is located on the northwestern Turkey. Lignite with about 9.3 billion tons of reserve is a major source for energy production in Turkey. Turkish Coal Enterprises produces about 60% of the country’s yearly production. Can Surface Lignite Mine which is planned to produce approximately 2.5 million tons of 3000 kcal/kg coal, will be one of the main production areas of TKI. Slope failures in the form of large landslides can cause serious interruptions in production of lignite. Major landslides commonly occur in TKI Can Lignite Mine. Back analysis of such slides can provide information about the mechanisms of such slides, so that preventive measures can be taken to reduce the future risks of landslides in the lignite basin. A large slide in Panel Can 5 involved movement of a mass of overburden extending about 1.5 km from the slope face. Production activities were stopped here. To provide information for slope planning in the other important panels such as Panel Can B, analyses were carried out in order to determine the mechanisms of this slope failure. Using a 2D limit equilibrium method of slices program SLIDE and FLAC 2D finite difference program, it was found that very low friction angle associated with a weak clay layer underneath the coal seam triggered this large slide. Using overall slope angles as low as 17° did not prevent slides in the basin, since this layer was found to have a much lower friction angle and dominate the sliding mechanism. Towards the edges of this lignite basin, inclination of the sedimentary units including this weak layer and local fault formations increase, and this situation contributes to the occurrence of landslides. In designing production slopes especially around the edges of the basin, new production advancement directions and slope stability improvement alternatives were to be considered. Another major landslide interrupted the production at the southern part of the open pit lignite mine. Landslide was back analyzed by numerical modeling in order to determine mobilized shear strength parameters during the slide. This information was believed to be useful in planning the safe slope angles for future production areas and safe production advancement directions in developing new production panels in the mine considering the inclined nature of the coal bed and the basin especially around lignite basin edges. In numerical modeling efforts, a 2D limit equilibrium method of slices program SLIDE and a 3D finite difference program were used to determine the shear strength parameters that were mobilized during the slide. After the slide natural slope of the loose material that moved on the foundation rock of the basin was around 13 . This friction angle was assumed to be the natural trend of the mechanical behavior of the weak material on the base rock, and thus it was kept constant during the analysis. In order to assign a cohesion value which was active during the slide, a circular failure was initiated, and the cohesion for the sliding material was determined first with 2D method of slices. Accepting that even a few degrees improvement in the overall pit slope angles can bring significant savings in stripping costs, further back analysis modeling was carried out with a 3D program FLAC 3D program. 3D modeling effort resulted in an activated cohesion value almost twice as much compared to the 2D back analysis results. Using this value in designing the overall slope angles future production plans is obviously going to lead to significant savings.
A SOLUTION TO WATER CRISIS IN ENERGY PRODUCTION: FEASIBILITY OF USING IMPAIRED WATERS FOR COAL-FIRED POWER PLANT COOLING Radisav D. Vidic, University of Pittsburgh [email protected] Heng Li, University of Pittsburgh, [email protected] Shih-Hsiang Chien, University of Pittsburgh, [email protected] Jason D. Monnell, University of Pittsburgh, [email protected] David Dzombak, Carnegie Mellon University, [email protected] Ming-Kai Hsieh, Carnegie Mellon University, [email protected]
Abstract The freshwater demand has increased globally due to rapid population growth and economic expansion. According to the U.N. Environment Program, the global freshwater withdrawal reached 4430 km3 in 2000, and was projected to reach 5240 km3 in 2030, a 38% increase from year 1995. Simultaneously, the global energy demand will increase by 77% from 2006 to 2030. Of the total energy production, thermoelectric power generation accounts for 80-90%. To operate a thermoelectric power plant, such as a coal-fired power plant, a great amount of water is needed for cooling system. Since 2005, thermoelectric energy generation has become the number one freshwater use in the U.S. With the growing demand for freshwater and water shortages in many parts of the U.S., a number of proposed power plants were terminated or suspended due to the lack of freshwater for cooling. Use of alternative water sources will inevitably be a critical solution to ensure sufficient power production in the future. Many types of nontraditional water sources already exist, with drastically varying availability and quality. This study, sponsored by the U.S. Department of Energy, demonstrated the feasibility of using nontraditional water sources, many of which are of impaired quality, for cooling in coal-fired power plants. Three types of impaired water were evaluated: secondary-treated municipal wastewater, passivelytreated abandoned mine drainage, and ash settling pond effluent from coal-fired power plants. A systematic review of the existing regulations in the U.S. revealed that using impaired waters for power plant cooling is generally allowed by the state/local governments, with some environmental quality control requirements. A national statistical analysis showed that secondary-treated municipal wastewater is a very promising alternative in terms of its quantity and geographical proximity to power plants. The effluent from only 1 or 2 fairly large publicly owned wastewater treatment facilities will satisfy the total cooling water needs of a coal-fired power plant that is located within a 10-mile radius. Half of the existing coal-fired power plants in the U.S. can benefit from their proximity to secondary effluent discharge points. Despite its widespread availability for use in power plant cooling, municipal wastewater can pose several technical difficulties in cooling systems, including corrosion, scaling, and biofouling. Results from both bench-scale and pilot-scale tests showed that the appropriate chemical treatment regimen can address the problems of corrosion, scaling, and biofouling in cooling systems when secondary-treated municipal wastewater is used as makeup water.
1
Introduction Water and energy are inseparable in many ways. For instance, water supply and wastewater treatment both require energy input. On the other hand, water is needed in power plant cooling to drive the thermoelectric energy generation. According to the latest survey, total water withdrawal in the US reached 410 billion gallons per day (BGD) in 2005, of which 349 BGD was freshwater and the rest (15%) was saline water (1). Among the major freshwater uses, thermoelectric power generation has surpassed agriculture to become the number one use (41%) since 2005. Reliable and abundant water sources are required to ensure sufficient power generation in the future. According to the U.N. Environment Program, the global energy demand will increase by 49% from 2007 to 2035 (2). However, the increasing water shortages and the global warming effect exacerbate the fierce competition for water among various users (3). In the effort to reduce green house gas emissions, implementation of carbon capture and sequestration/reuse (CCS/CCR) has been the focus of numerous research and policy development efforts. The adoption of CCS/CCR will increase water consumption by 76-90% in a fossil-fuel power plant due to the decrease in power generation efficiency and the need for water for the CCS/CCR processes (4). Coal, natural gas, and nuclear power, added together, account for nearly 90% of the world’s electricity produced in thermoelectric power plants (5). Regardless of the cooling technology used in these power plants, either recirculating cooling or once-through cooling, water is critical to cool and condense the steam in the boiler water loop (6). Several cases in arid regions, such as Arizona and Texas, have shown that the lack of available cooling water resulted in suspension for existing power plants and permit denial for constructing new plants (7, 8). Rising demand for electricity in those areas will lead to difficult decisions in water allocating priorities and finding reliable water sources for electricity production. The use of alternative water sources for power plant cooling will likely be inevitable to ensure sufficient energy production in the future. Moreover, tightening regulation for water intake structures imposed by the Clean Water Act (CWA) to protect fish and other aquatic life (CWA section 316(b)) has put the power plants and other industrial facilities in a challenging position for natural water withdrawal. Indeed, the ability to use alternative water sources for cooling may provide a competitive advantage for new power plants with respect to siting and construction permits (9). 2
Many types of nontraditional water sources may be considered for power plant cooling but their use may be restricted by drastically varied availability and quality. For instance, passively-treated coal mine drainage is most abundant in Pennsylvania and West Virginia, while the most severe water constraints for cooling are more likely to exist in western US (10). Water is used to transport either fly ash or bottom ash generated from coal combustion and the effluent from ash settling pond is readily available for reuse. However, the ash pond water can only serve in emergency situations due to its limited amount as it can satisfy only 25% of the need for cooling water makeup (11-12). Among all the alternatives, treated municipal wastewater (MWW) is considered as the most promising water owing to its ubiquitous availability and fairly steady quality (13). Although using treated municipal wastewater to replace freshwater withdrawal for power plant cooling seems to be feasible based on its availability, use of this impaired water can pose several technical difficulties in cooling systems because of its lower quality compared to typical freshwater sources. Metal corrosion, mineral scaling, and biofouling are the major technical challenges. Besides the technical challenges with the use of MWW, legal basis for wastewater reuse for power plant cooling needs to be considered. Whether the cooling systems are new or being retrofitted, any water reuse program needs to comply with federal, state, and local regulations. The issue of water ownership and right of use may also complicate the use of treated MWW if interstate or interbasin water transportation is involved. In addition to legal considerations, potential air-borne pollution and proper blowdown discharge from the cooling systems using concentrated wastewater are among the public concerns that must be addressed. A number of power plants have already blended MWW with freshwater as a cooling makeup (12, 14), but the blend ratio varies significantly. Only a few power plants have operated their cooling towers with treated MWW as the dominant makeup water. One notable example is the Redhawk Power Plant in Arizona that uses over 90% of MWW as its makeup water. The 6.46 MGD of wastewater used by the facility is transported 40 miles from a wastewater treatment facility, which is located at a higher elevation than the power plant. The MWW received at the power plant is treated further before addition to the cooling towers. Power plants usually do not use 100% MWW as cooling system makeup due to a lack of technical knowledge and legal framework to support such practices. 3
Availability of MWW for Power Plant Cooling Treated municipal wastewater holds great potential to be the most promising alternative water source for power plant cooling in the US. The analysis of its quantity, availability, and geographical proximity for power plants is critically needed and should precede technical assessment of the feasibility of MWW reuse for cooling. It is estimated that a total volume of 30-40 billion gallons of MWW is treated per day in the US (15). However, this amount of water is not evenly distributed and the effluent discharge rates may vary significantly from site to site. For the analysis of MWW availability for power plant cooling, a comparison must be performed between the location and amount of water supply and prospective water consumption. The publicly owned treatment works (POTWs), which are the most reliable and accessible point water sources regulated by the National Pollutant Discharge Elimination System (NPDES) permit program (14), are considered here as the major water providers. Based on NPDES, MWW effluents must reach at least the secondary treatment level and be disinfected prior to discharge from the treatment facilities. A database containing water flowrates and geographical locations of 33,852 NPDES permitted discharges has been constructed by the US EPA since 1996, of which 17,864 are POTWs in the lower 48 states and can be considered as potential water providers for power plant cooling. The amount of water withdrawal for power plant cooling system makeup can be very different, depending on the particular cooling technology used in the power plant. For a oncethrough cooling system, 20-50 gallons of water are used to generate one kWh of electricity. On the other hand, modern recirculating cooling towers need 0.2-0.6 gallons of water to generate each kWh of electricity (12). However, the construction of once-through cooling systems is no longer permitted by CWA 316(b), and the focus of this study was on recirculating cooling systems. Among all thermoelectric power plants in the lower 48 states, 407 existing coal-fired power plants are selected to represent potential users of treated MWW for cooling. It is assumed that the cooling systems in these 407 power plants are recirculating systems, regardless of their current actual configurations. A total of 110 proposed power plants, including renewal or new units as proposed in 2007 and to be constructed in five years, are also selected to represent the potential new water users (Figure 1). The cooling water demand is then estimated by using a conversion factor of 0.6 gallons of water per kWh of electricity produced. 4
The correlation between the water providers and water users in terms of the number of POTWs required to satisfy cooling water needs of a power plant within a specified radius is determined using geospatial analysis (ArcGIS version 9.2, ESRI, Redlands, CA) (16). Coverage of 10 and 25 mile radius is examined to ensure reasonable transportation distance of the treated MWW from the POTWs to power plant. The GIS data show that in Pennsylvania, for example, there are on average 11.7 POTWs within a 10-mile radius around each existing power plant, and 62.5 POTWs when the range is extended to 25 miles. Nationwide, about 50% of existing power plants can obtain sufficient amount of cooling water from POTWs located within a 10-mile radius. Furthermore, an average of only 1.14 POTWs is needed to supply enough water to meet their water demand. On the other hand, 76% of the power plants can have sufficient cooling water supply from an average of 1.46 POTWs if the radius is extended to 25 miles. As for proposed power plants, 81% of them can obtain sufficient cooling water from POTWs located in a 10-mile radius if freshwater is not readily available. With an increase in coverage to 25 miles, nearly all proposed power plants (97%) can operate regardless of freshwater availability (Figure 2). In summary, only 1 or 2 water conveyance pipes of reasonable length (less than 25 miles) are required between a power plant and the POTWs to supply sufficient amount of cooling water to replace more than 3/4 of current freshwater withdrawal for power plant cooling. The use of MWW as an alternative makeup water for power plant cooling is particularly meaningful for the development of coal-fired power plants in regions where freshwater sources are not readily available.
Technical Challenges Although the three major technical challenges, namely, biofouling, scaling, and corrosion, are discussed separately in this section, they can be interconnected and antagonistically impact each other. This may be particularly true in cooling systems using MWW because of the impaired quality and complex chemistry of the water. Biological Fouling. Although the treatment processes employed in most POTWs remove large portion of biodegradable organic material, treated MWW still contains appreciable levels of phosphorus, nitrogen, and residual organic matter. Warm, moist, and nutrient-rich environment in recirculating cooling systems is conducive to continued biological growth, including a wide range of microorganisms such as bacteria, fungi, and algae (17). Typical 5
cooling operation, such as temperature at 35-45°C, pH 6-9, continuous aeration, and sunlight exposure, makes the cooling system a favorable habitat for biological growth. The CoC effect concentrates the organic matter and other nutrients in the bulk cooling water, which renders the biological growth control even more challenging. Biological growth can quickly lead to biomass/slime accumulation on equipment surfaces, a process known as biological fouling (biofouling). For example, biofouling is able to develop on heat exchanger surfaces within 4-8 hr (18). The resulting biofilm can bind with suspended solids, silt, corrosion products, and organic and inorganic deposits, which greatly exaggerates mineral scaling and corrosion (microbial induced corrosion underneath biofilm) problems (19). Control of biofouling can be achieved through chemical or physical methods. Application of proper chemical biocides is a widely adopted approach based on cooling system design and the makeup water quality. The disinfection efficiency is affected by water temperature, pH, redox condition, chemicals added for corrosion and scaling control (18, 20). A general guideline of acceptable biofouling control is 104 CFU/mL (CFU: Colony Forming Unit) for planktonic bacteria or 104 CFU/cm2 for sessile bacteria measured by the standard heterotrophic plate counts (HPC) (19). Chlorination is the most common chemical program used in cooling systems to control biogrowth because of its high efficiency, easy implementation, and low cost (17). Biofouling can be well mitigated by maintaining a proper disinfectant residual concentration in the bulk water. However, elevated organic content and ammonia concentration in cooling systems using MWW can greatly increase the free chlorine dose required to achieve breakpoint chlorination. Moreover, free chlorine tends to react with natural organic matter to form undesirable disinfection byproducts (DBPs), which can endanger public health when they are discharged into surface waters. Monochloramine was comparably effective as free chlorine at similar doses after the same contact times in disinfecting preformed biofilms when seawater is used in cooling systems (21). However, application of monochloramine was more effective than free chlorine in preventing biofilm formation in a cooling system (22). Although chloramination may require higher residual concentrations than chlorination (1-3 ppm higher) to achieve similar disinfection efficacy, monochloramine is more stable with lower decomposition rates compared to free 6
chlorine (23, 24).Consequently, the overall chemical consumption can be potentially lower with chloramination than chlorination. In conclusion, monochloramine can be used to replace free chlorine as an effective disinfectant in cooling systems using secondary-treated MWW as makeup water. It is essential to continuously dose pre-formed monochloramine to maintain target residual disinfectant (2-3 ppm) and keep the bioactivity under control. Periodic free chlorine shock dosing may be needed to sustain effective biofouling control. Inorganic Scaling. Given the relatively high levels of both dissolved and suspended solids, alkalinity, and hardness in typical secondary-treated MWW (Table 1), a significant concern when using the wastewater for recirculating cooling is the potentially severe mineral scaling problems. Scale formation in cooling systems causes a multitude of operational problems, chief among which are the hindrance of heat transfer, the obstruction of pipe flow, pump failures, and potential for cooling tower fills fall due to excessive weight (25). The bottom line is that scaling causes large economic losses in virtually all water-based industrial processes. The control of scaling is an imperative to ensure proper operations of these processes. Cycles of Concentration, or CoC, is an important operational parameter in recirculating cooling systems and is the primary cause of mineral precipitation and scaling. Power plant cooling using freshwater may generally encounter only minor scaling problems, especially when the system is operated at low CoC. However, in the case of using MWW as makeup water, the cooling tower performance could be significantly compromised by scaling problems when the water chemistry is not properly managed. A better understanding of the mineral deposition processes and mechanisms of MWW in cooling systems is therefore required before sound control technologies can be developed and implemented to inhibit mineral scaling. Characterization of single mineral precipitation (e.g., calcium carbonate, calcium sulfate, iron oxides) in relatively simple solutions is well documented in the literature (e.g., 28-34). However, the knowledge acquired from these studies can be inadequate for elucidating mineral precipitation in MWW, which contains a whole array of inorganic and organic chemical constituents, as well as certain amount of particulate matter. The effects of co-existing species on the precipitation behavior of a particular mineral, being either synergistic or antagonistic, have continued to be the focus of nucleation research. Attempts to apply the specific data obtained from simple aqueous systems to more complex ones containing multiple chemical constituents is 7
likely to achieve limited success. For example, thermodynamic calculations indicate the formation of dolomite (CaMg(CO3)2) in a solution supersaturated with calcium, magnesium, and carbonate. However, direct dolomite precipitation in laboratories and natural waters under normal conditions of temperature, pressure, and salinity has rarely been observed; the phases that actually form consist of aragonite and magnesian calcite (a solid solution) (28, 29). Thus, great care must be taken in using simple thermodynamic calculations to predict mineral precipitation in complex water systems, such as concentrated MWW, where carbonate minerals may be dominated by metastable phases. The presence of Mg2+ appears to cause the precipitation of beta-tricalcium phosphate rather than apatite and the precipitation of magnesian calcium carbonate rather than calcite (30-33). In the case of calcium phosphate precipitation, the inhibition effect of Mg2+ results from its incorporation into the Ca-P solid lattice, which prevents the formation of a well-crystalline Ca-P structure (34). Mineral precipitation does not necessarily cause scaling, because scaling is a surface process while precipitation can happen in the bulk water. An added antiscaling chemical can prevent scaling in various ways (35). A good inhibitor chemical should work through multiple mechanisms to prevent scaling. Typical doses of the antiscalants added in cooling systems using freshwater are less than 10 mg/L. Most antiscalants that have been proved effective in freshwater may not, however, be as effective in treated MWW when similar dosing ranges are used (36). Corrosion. Another challenge when using treated MWW in recirculating cooling water systems comes from corrosion of metal surfaces in heat exchangers or conveyance pipes. Corrosion may be exacerbated due to the degraded quality of the wastewater and the simultaneous occurrence of mineral scaling and biofouling. Knowledge of traditional corrosion control approaches built on the experience with freshwater as cooling system makeup water needs to be re-evaluated. Among the many water quality parameters influencing corrosion in cooling systems using MWW, phosphate and ammonia are of the greatest concern. Ammonia and phosphate concentrations in secondary-treated MWW can reach up to 70 and 50 mg/L, respectively (Table 1). Ammonia is a strong complexing agent toward many metals and metal alloys and can be very corrosive (39). It is recommended that the ammonia concentration in cooling systems should not exceed 2 mg/L (as NH3) (40). Although the ammonia in the makeup water (MWW) can be high, several studies reported that the ammonia in recirculating cooling water was low (41-43). Such 8
findings are most likely due to stripping and nitrification in cooling towers, where high water temperature (40-50°C), high pH (8-9), and active aeration are all in favor of evaporative loss of ammonia. The knowledge of ammonia stripping and nitrification in cooling towers is important for determining corrosion control strategies to overcome the aggressiveness of ammonia. However, the study of ammonia stripping and nitrification in cooling towers has not yet been conducted. Phosphate can adsorb onto the surface of metals and metal alloys to form a protective thin film against corrosion (44) and has been used as a corrosion inhibitor in cooling systems. Calcium phosphate has limited solubility and its precipitation and surface deposition help to mitigate corrosion. However, excessive phosphate precipitation can be a problem when the water contains high hardness. Preliminary results of the stability of phosphate and phosphate-based corrosion inhibitors show that these inhibitors tend to precipitate in treated MWW. Furthermore, acid addition to the cooling water is still commonly practiced to control mineral precipitation. Decreased pH after acid addition can raise the corrosivity of the cooling water. As such, pH control-based scaling mitigation can potentially compromise the advantages of using mineral precipitation for corrosion inhibition. Studies on the optimal acidification strategies that balance scaling and corrosion control would be of great interest. As discussed in the previous section, cooling systems using MWW can have high biofouling potential, which can cause microbiologically influenced corrosion (MIC). Traditional approaches of maintaining free chlorine residuals for biofouling control may cause higher corrosion rates due to the direct attack of free chlorine on metals and metal alloys. Free chlorine also degrades tolyltriazole (TTA), a commonly used copper corrosion inhibitor. Instead of free chlorine, monochloramine may be more appropriate for use in cooling systems using MWW, but its influence on the corrosion of metal alloys and the corrosion inhibitors needs to be further evaluated. Chemical addition for corrosion control is the most widely employed approach in recirculating cooling systems. When a cooling system switches its makeup water source from freshwater to impaired water like treated MWW, traditionally-proven corrosion control approaches may not work well and need to be re-evaluated by taking into account both the degraded water quality and the scaling/biofouling processes and associated control strategies.
9
Potential Challenges and Opportunities Increasing population and economic development will continue to drive the demand for electric power in the years ahead. Despite growth in renewable energy sources, most of the electricitygenerating capacity in the decades ahead will still be from coal-fired and nuclear thermoelectric power plants. In most thermoelectric power production, water is used for cooling. About half of the existing power plants in the US employ once-through cooling, which will not be an option for proposed new power plants and may not be available for re-permitted plants. Meeting the freshwater demand of new power generation capacity will be very difficult in parts of the country that already have limitations on available freshwater, most notably in the west and southwest regions. Waters of impaired quality can be used as alternative sources of makeup water for recirculating cooling systems in thermoelectric power plants. Treated municipal wastewater is the most widely available and easily accessible water source for power plant cooling. Statistical analysis showed that half of the existing coal-fired power plants in the US can meet their cooling water needs using MWW from only 1 or 2 fairly large POTWs that are located within a 10-mile radius. The complicated water chemistry of MWW and the varying operating conditions of open recirculating cooling systems incur great challenges for the traditional approaches of using chemicals to control scaling, corrosion, and biofouling. New research is needed to provide technical support for more suitable control strategies to mitigate these problems. The feasibility of using chemical inhibitors in combination with advanced treatment to manage the cooling water quality need to be examined to ensure successful use of treated MWW in cooing systems. Systematic investigations employing both bench-scale systems and pilot-scale cooling towers are needed to explore optimal chemical treatment strategies for controlling corrosion, scaling, and biofouling before the most promising strategies can be tested in the field. Tertiary treatment, such as nitrification and/or filtration, may be required to reduce levels of suspended solids, bioactivity, organic matter, and alkalinity. Studies to determine the most cost effective approaches for managing MWW in cooling systems are critical. Changing regulatory issues, e.g., more controls on inter-basin transfers, concerns for extra contaminants in drift, need to be considered.
10
The environmental impacts during delivery, treatment, use, and discharge of MWW for cooling need to be investigated. In this regard, a comparative Life Cycle Assessment (LCA) of alternative pre-treatment and operating strategies may be suitable to inform sustainable design and decision-making. Social and institutional barriers to reuse of this water for power plant cooling may exist and work should be undertaken to identify such barriers and develop approaches to overcome them to ensure further development of power generation capacity in the future.
Acknowledgements Support received from the U.S. Department of Energy, National Energy Technology Laboratory, Grant No. DE-FC26-06NT42722. The authors gratefully acknowledge the Franklin Township Municipal Sanitary Authority, and especially manager James Brucker, for allowing and supporting performance of the pilot-scale cooling tower tests at their facility.
References (1)
Kenny, J. F.; Barber, N. L.; Hutson, S. S.; Linsey, K. S.; Lovelace, J. K.; Maupinet, M. A. Estimated use of water in the United States in 2005. U.S. Geological Survey Circular 1344, 2009.
(2)
U.S. Energy Information Administration / International Energy Outlook 2010. Cited July 5, 2010; Available from: http://www.eia.doe.gov/oiaf/ieo/electricity.html.
(3)
Hinrichsen, D. and Robey, B. Solutions for a water-short world. Johns Hopkins School of Public Health, Population Information Program, Baltimore, MD, 1997.
(4)
Gerdes, K. and Nichols, C. Water requirements for existing and emerging thermoelectric plant technologies. DOE/NETL-402/080108, U.S. Department of Energy, National Energy Technology Laboratory, Pittsburgh, PA, 2008.
(5)
OECD/IEA, Key world energy statistics. 2007.
(6)
DOE-NETL, Use of nontraditional water for power plant applications. Oak Ridge, TN, 2009.
(7)
Feeley, T. J. and Ramezan, M. Electric utilities and water: emerging issues and R&D needs. In 9th Annual Industrial Wastes Technical and Regulatory Conference. San Antonio, TX, Water Environment Federation, 2003. 11
(8)
Dishneau, D. Frederick County denies cooling water to proposed power plant. The Baltimore Examiner. 2007.
(9)
EPRI, Use of alternate water sources for power plant cooling. Palo Alto, CA, 2008.
(10) Roy, S. B.; Summers, K. V.; Goldstein R. A. Water sustainability in the United States and cooling water requirements for power generation. Universities Council on Water Resources/Water Resources Update, 2003(126): 94-99. (11) Nemerow, N. L. and Agardy, F. J. Strategies of Industrial and Hazardous Waste Management. John Wiley & Sons, New York, NY, 1998. (12) DOE-NETL, Use of reclaimed water for power plant cooling. Oak Ridge, TN. ANL/EVS/R-07/3, 2007. (13) Vidic, R. D. and Dzombak, D. A. Reuse of Treated Internal or External Wastewaters in the Cooling Systems of Coal-based Thermoelectric Power Plants. Final Technical Report to US DOE/NETL, project number DE-FC26-06NT42722, 2009. (14) USEPA, Guidelines for water reuse. U.S. Agency for International Development, EPA/625/R-04/108, Washington, D.C. 2004. (15) USEPA, Clean Watersheds Needs Survey: CWNS 2000 Report to Congress. U.S. Environmental Protection Agency, Research Triangle Park, NC, EPA-832-R-03-001, 2003. (16) Chien, S.-H. Use of impaired waters in power plant cooling tower system: Review of regulations and feasibility analysis. MS thesis, University of Pittsburgh, 2009. (17) Frayne, C. Cooling water treatment: Principles and practices. Chemical Publishing Co., New York, NY, 1999. (18) Rossmoore, H. W. Handbook of biocide and preservative use. Blackie Academic & Pro Professional, Chapter 3, pp. 50-77. 1995. (19) Ludensky, M. Microbiological control in cooling water systems and directory of microbiocides for the protection of materials: A handbook. pp. 121-139. 2005. (20) USEPA, Alternative disinfectants and oxidants guidance manual. United States Environmental Protection Agency, EPA 815-R-99-014, 1999. (21) Rao, T. S.; Nancharaiah, Y. V.; Nair, K. V. K. Biocidal efficacy of monochloramine against biofilm bacteria. Biofouling, 1998. 12(4): 321-332. (22) Turetgen, I. Comparison of the efficacy of free residual chlorine and monochloramine against biofilms in model and full scale cooling towers. Biofouling, 2004. 20(2): 81-85. 12
(23) Jolley, R. L.; Brungs, W. A.; Cotruvo, J. A.; Cumming, R. B.; Mattice, J. S.; Jacobs, V. A. Water chlorination: environmental impact and health effects, Volume 4, Book 1. 1983, Chemistry and Water Treatment, Ann Arbor Science, pp. 33. (24) Wolfe, R. L.; Ward, N. R.; Olson, B. H. Inorganic chloramines as drinking water disinfectants: a review. J. AWWA, 1984. 76(5): 74-88. (25) Metcalf & Eddy (AECOM), Water reuse: issues, technologies, and applications, written by Asano, T, Burton, F., Leverenz, H, Tsuchihashi, R., and Tchobanoglous, G., McGraw-Hill: New York, NY, 2007. (26) Liu, X. Y. A new kinetic model for three-dimensional heterogeneous nucleation. J. Chem. Phys. 1999. 111(4): 1628-1635. (27) Liu, X. Y. and Lim, S. W. Templating and supersaturation-driven anti-templating: Principles of biomineral architecture. J. Am. Chem. Soc. 2003. 125(4): 888-895. (28) Morse, J. W. and Arvidson, R. S. The dissolution kinetics of major sedimentary carbonate minerals. Earth-Science Reviews, 2002. 58(1-2): 51-84. (29) Morse, J. W.; Arvidson, R. S.; Luttge, A. Calcium Carbonate Formation and Dissolution. Chemical Reviews, 2007. 107(2): 342-381. (30) Ferguson, J. F. and McCarty, P. L. Effects of carbonate and magnesium on calcium phosphate precipitation. Environ. Sci. & Tech., 1971. 5(6): 534-540. (31) Walter, L. M. Magnesian calcite stabilities: A reevaluation. Geochimica et Cosmochimica Acta, 1984. 48(5): 1059-1069. (32) Deleuze, M. and Brantley, S. L. Inhibition of calcite crystal growth by Mg2+ at 100 C and 100 bars: Influence of growth regime. Geochimica et Cosmochimica Acta, 1997. 61(7): 1475-1485. (33) Nishino, Y., Oaki, Y.; Imai, H. Magnesium-Mediated Nanocrystalline Mosaics of Calcite. Crystal Growth & Design, 2009. 9(1): 223-226. (34) Suchanek, W. L.; Byrappa, K.; Shuk, P.; Riman, R. E.; Janas, V. F.; TenHuisen, K. S. Mechanochemical-hydrothermal synthesis of calcium phosphate powders with coupled magnesium and carbonate substitution. J. Solid State Chemistry, 2004. 177(3): 793-799. (35) Li, H. Mineral precipitation and deposition in cooling systems using impaired waters: mechanisms, kinetics, and inhibition. PhD dissertation, University of Pittsburgh, 2010.
13
(36) Li, H.; Hsieh, M.-K.; Chien, S.-H.; Monnell, J.D.; Dzombak, D. A.; Vidic, R. D. Control of mineral deposition in industrial cooling systems using secondary-treated municipal wastewater. Accepted by Water Research, 2010. (37) Shih, W.Y.; Albrecht, K.; Glater, J.; Cohen, Y. A dual-probe approach for evaluation of gypsum crystallization in response to antiscalant treatment. Desalination, 2004. 169(3): 213-221. (38) Li, H.; Vidic, R. D.; Dzombak, D. A. Electrochemical Impedance Spectroscopy (EIS) based characterization of mineral colloid deposition from precipitation reactions. submitted, 2010. (39) Stumm, W. and Morgan, J. J. Aquatic chemistry : chemical equilibria and rates in natural waters (2nd ed.). Environmental science and technology. 1996, New York: Wiley. (40) EPRI, Use of degraded water sources as cooling water in power plants (1005359). 2003: Energy Commission, Public Interest Energy Research Program, Sacramento, CA. (41) Goldstein, D.; Casana, J.; Wei, I. Municipal wastewater reuse as makeup to cooling towers. In Proceedings of the Water Reuse Symposium II, AWWA Research Foundation, Denver, CO., 1981. (42) Schumerth, D. J. Gray and impaired water cooling in surface condensers and heat exchangers. In Proceedings of PWR2006, ASME Power, Atlanta, GA, 2006. (43) Rebhun, M. and Engel, G. Reuse of wastewater for industrial cooling systems. J. Water Pollution Control Federation, 1988. 60(2): 237-241. (44) Nowack, B. Environmental chemistry of phosphonates. Water Research, 2003. 37(11): 2533-2546.
14
TABLE 1. Typical quality of secondary-treated municipal wastewater Parameter
Range
pH
7-8
BOD5 (mg/L)
3-30
COD (mg/L)
40-80
TDS (mg/L)
130-1600
TSS (mg/L)
10-50
Alkalinity (mg/L as CaCO3)
100-500
Ca (mg/L)
28-185
Mg (mg/L)
23-150
NH3-N (mg/L)
3-73
NO3-N (mg/L)
10-30
SO4 (mg/L)
60-293
PO4 (mg/L)
0.6-51
Conductivity (mS/cm)
0.2-1.2
15
POTW (> 3.1 MGD) Proposed power plant
FIGURE 1. Location of the proposed thermoelectric power plants (red dots) relative to the publicly-owned treatment works (POTWs) that have flowrates greater than 3.1 MGD of treated MWW (blue dots) in the lower 48 states.
16
FIGURE 2. Fraction of the proposed thermoelectric power plants with sufficient amount of treated MWW available within the specified radius (10 miles and 25 miles). The power plants are categorized by NERC regions (NERC: North American Electric Reliability Corporation).
17
Leaching Characteristics of Waste from PF Utilities and Transitional Technologies using Australian Coal David French1, Ken Riley1, Colin R. Ward1,2, Leanne Stephenson2 and Lila Gurba2 1
CSIRO Energy Technology, PO Box 52, North Ryde 1670, Australia School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney NSW 2052, Australia 2
KEYWORDS: leach behaviour, trace elements, fly ash, coal, pf, transitional technologies. ABSTRACT Australia produces approximately 13 million tonnes per annum of pf-combustion wastes (ash) from coal-fired power stations. A small proportion is utilised but most is deposited in wet ash dams or dry repositories. The nature of this waste stream is expected to change with adoption of transitional power generation technologies such as fluidised bed combustion, oxy-fuel combustion and integrated gasification combined-cycle processes using slagging gasifiers. The leaching properties of the major solid waste streams from each of these (transitional) processes have been compared with those of typical fly ashes from conventional pf power stations burning Australian bituminous thermal coals. Batch leaching tests were done at two liquid:solid ratios (3.5:1 and 20:1) and the results compared to similar pf ash studies. It is concluded that most trace elements were leached from the fly ash and bottom ash of FBC systems at similar concentrations to those present in equivalent leachates from ash of conventional power stations (lower concentrations of Cd, Co, Ni and Zn were leached from the waste of two FBC power stations). Trace elements were also leached in similar proportions from fly ashes derived from the same coals fired under air-fired (conventional) and oxy-fuel combustion conditions. Some of the trace elements (e.g. B, Cr, Mn, V and Zn) in the gasifier slags were not leached as readily as those in fly ash from conventional power stations; others (e.g. As and Se) were leached at similar concentrations from both types of ash materials..
INTRODUCTION Currently, Australia produces approximately 13 million tonnes per annum of pfcombustion by-products of which only a small proportion is utilised, most being deposited in wet ash dams or dry repositories. With the adoption of transitional power generation technologies (oxy-fuel pf, IGCC and FBC) the nature and amount of coal combustion products produced in Australia (and elsewhere) is likely to change significantly over time, as are the environmental impacts. To identify the potential changes in environmental impact arising from the introduction of these new technologies, a study was undertaken, to establish if differences exist between the waste products produced by transitional power
1
generation technologies and the fly ash (and bottom ash) produced by conventional pulverised fuel (pf) coal-fired power stations. Any differences found may indicate advantages associated with the environmentally sound disposal of the waste streams from transitional technologies, or alternatively that transitional wastes might have deleterious properties that impact adversely on the options for their disposal or use.
EXPERIMENTAL Samples of the major waste streams were obtained from either operational full-scale plants or pilot plants. The chemical and mineralogical compositions of these samples were determined by X-ray fluorescence spectrometry (XRF), inductively coupled plasma atomic emission spectrometry (ICP-AES), inductively coupled plasma mass spectrometry (ICP-MS) and X-ray diffractometry (XRD), with the latter using procedures described by Ward and French 1 to determine the amorphous content. The samples were also analysed for a number of trace elements using a variety of different techniques. In order to evaluate the potential release of environmentally significant elements during storage or use, the samples were extracted with MilliQ water, providing data that could be compared with the results of previous leaching studies on fly ash from conventional pf coal fired power stations 2-3. The samples (ashes or slags) were leached for three days in distilled water at liquid:solid ratios of 3.5:1 and 20:1 following the procedure described by Jankowski et al 2 . The leachates were then filtered through 0.22 μm Millipore cellulose acetate membrane filters. An unacidified portion of each filtrate was analysed for pH and conductivity. Another portion of each filtrate was acidified to approximately 1% v/v AR grade concentrated nitric acid and the concentrations of trace elements determined using appropriate techniques; for most elements that was ICP-AES and ICP-MS. The following materials were investigated: Fluidised bed combustion residues – samples of silo ash (fly ash) and bottom ash (bed material) from FBC power stations in Australia and in Japan. Oxy-fuel combustion ashes – coal feeds and pairs of ash samples from air-fired and oxy-fuel combustion tests conducted in Japan on three Queensland coals. IGCC slags – samples obtained from entrained-flow gasification trials completed at the Seimens test facility in Germany on a range of Australian coals. The received slag samples were air dried before analysis and the leaching tests. As well as these samples, a gasifier slag from the Buggenum plant in The Netherlands was also included in the study.
RESULTS AND DISCUSSION The data obtained from the analyses of the transitional waste products, and also those obtained from the leaching tests, are considered, and also compared with the
2
data obtained on fly ash from the same or similar Australian coals combusted in conventional pf-fired power stations FBC Ashes The mineralogy of the FBC ash samples (both fly ash or silo ash and bottom ash) is characterised by the abundance of calcium-rich phases, e.g. calcite, calcium sulphates (anhydrite and gypsum) and anorthite, CaAl2Si2O8 (Table 1). This is particularly so in the ash from the Japanese PFBC plant. The high calcium content is also evident in the data on the chemical composition (Table 2). Although lower proportions of Ca are found in the ashes from the Australian FBC station, compared with the ash from the Japanese station, the Ca concentrations are still higher than seen in pf ashes from Australian bituminous coals. Of the trace elements in the samples, Mo and Ni tend to be lower (Table 3) than in pf ashes. The leachates are characterised by high pH and conductivity, together with elevated levels of Ca, Cr, and Pb. Si, As, Cd, and Cu values tend to be lower than in typical pf ash leachates. The presence of high levels of Ca (as well as Ba and Sr), arising as a consequence of the addition of limestone to the bed material in FBC systems, is the major difference between ashes from FBC systems and ashes from conventional coal-fired power stations. Obviously there is a greater mass of waste ash from FBC systems per tonne of coal due to the addition of limestone to the coal bed. It is not evident from the data in Tables 4 and 5 that ashes from FBC are more environmentally benign than ashes from conventional pf-fired power stations. However, the greater concentrations of Ca and of the associated mineral phases indicate that FBC ashes would have greater neutralisation capacity, and therefore would be suited for backfilling at mine sites where acid mine drainage (AMD) may be a problem. Of interest is the very low concentration (below detection limit) of Mg in the leachates of the ash products from the Japanese PFBC plant. Presumably this reflects the presence of relatively insoluble magnesium carbonate. Ashes from Oxy Fuel Systems As seen from data on relevant pairs of ash samples (Table 1), the ashes produced in the oxy-fuel tests are very similar mineralogically to the ashes from the same coals used in conventional pf power stations. Although there are no clear differences between the mineralogy of the air-fired and oxy-fuel ashes, both are distinguished from the FBC and IGCC ashes by their higher mullite contents. Iron oxide contents also tend to be slightly higher (Table 1). .Although there are no consistent differences between the air and oxy-fuel fired pairs, a higher glass content is found in the ash from the air-firing of coal C, along with associated lower concentrations of quartz and mullite. There are similar but less marked differences in the pairs of ashes from coal A. The presence of sylvite (KCl) in the ashes from coal C is intriguing and is possibly related to the higher chloride and lower mineral matter in the feed coal. An unusual feature is the occurrence of minor proportions of the clay minerals illite and kaolinite, which would not have been expected to survive the temperatures typically found in pf combustion systems. As expected, the elemental compositions of each of the paired ashes are also similar (Table 2), but with two significant differences. The concentrations of Fe and Si vary
3
in the ash pairs of coals A and C. This may in part be the explanation for the variation in the inorganic phases seen within these two ash pairs. The trace element concentrations are similar to those found in the FBC and IGCC ashes, apart from slightly higher levels of Pb, U, V, and Zn. The concentrations of the trace elements in each of the paired ashes are also remarkably similar (Table 3), with few exceptions, one being the high Cr found in the oxy-fuel ash of coal C and the other high Th in the oxy-fuel ash of coal B. There is no obvious explanation for these observations. The pH, conductivity and trace element concentrations of each of the leachates are shown in Tables 4 and 5. In comparison to the FBC and IGCC leachates, the leachates from the oxy-fuel ashes appear to have higher contents of K, Si (only in the 3.5:1 leaching tests), As, Co, Cu, Mn, Mo, Ni, Sb, Se, V and Zn with lower Pb. The concentrations of trace elements in the leachates are also generally similar in the ashes from each pair (Tables 4 & 5). However, some differences do exist: the leachates of the air- fired ashes from coal A have higher contents of K, As, Mo, V, and Zn (less pronounced or absent in the case of As, Mo, and V in the 20:1 leachate), high occurs V in the oxy-fuel leachate of coal B; high Al, Mn, Ni and Zn are noted in the oxy fuel leachate of coal C. Again, there is no obvious reason for the observed behaviour. With the above exceptions most trace elements, including the environmentally important ones such as Cd and Se, are leached in similar concentrations. The general observation is that there is little difference between the leaching behaviour of ashes obtained from conventionally fired coal and from the same coal fired in an oxy-fuel boiler. However, there were some differences observed in the analyses of individual leachates from each pair for a limited number of specific trace elements. IGCC Slags The IGCC slags show a considerable variation in size distribution and morphology, ranging from blocky clinker to fine beads and glass droplets. This variation is related to factors such as the fusibility of the coal ash, the composition of the mineral matter and the gasification temperature. All other factors being equal, it is likely that variation in the morphology (i.e. particle size and shape) will have an impact on the leaching characteristics of a particular slag. The mineralogy of the slags is distinguished by an abundance of amorphous material, with correspondingly lower proportions of the crystalline phases. The observed mineralogical variations (Table 1) provide some evidence for the propensity of each of the coals to produce slags under the conditions of gasification, as shown by the high percentages of amorphous material present. For example, the slags from Test 104 and from Buggenum are completely amorphous. The presence of crystalline mineral phases rather than an amorphous phase is indicative of incomplete melting and subsequently poor slagging. Note also that the slags from Tests 102 and 103 contain high concentrations of carbon, which is obviously another indication of poor gasification. This is further reflected in high LOI values (Table 2).
4
Table 1. Ash mineralogy by quantitative X-ray diffraction Fluidised Bed Combustion Aus FBC Phase (wt %) Quartz
Silo ash 25.8
Bottom ash 30.0
Fly ash 3.7
Bottom ash 9.8
Cristobalite (?) Mullite
0.9
2.3
Air and Oxy-fuel pf
Japan PFBC
6.1
Illite
Coal A air-fired
Anorthite
air-fired
oxy-fired
3.8
3.4
12.8
10.9 0.1
29.2
32.9
0.8
Kaolinite
32.1
3.7
11.8
4.4
Gehlenite
8.5
6.3
Merwinite
6.6
5.0
5.8
5.7
Melilite, sodic Hematite
1.0
1.9
Maghemite
0.7
0.6
Magnetite
0.3
0.2
1.0
32.8
air-fired
oxy-fired
7.2
8.5
0.1
0.2
17.2
22.1
T100 0.1
1.1
1.5
0.2
0.0
1.0
0.7
0.7
1.7
1.1
0.9
2.8
3.7
2.0
0.6
0.8
2.5
Hercynite
T103
0.8
0.1
0.9
0.2
T104
Bugg
0.6
0.2
0.3
0.5
Goethite
3.2
Periclase (?)
0.2
0.7
Lime Portlandite
3.0
16.7
4.2
0.6
2.6
Calcite
1.6
0.9
3.1
43.6
Anhydrite
0.4
1.2
7.5
5.8
Gypsum
3.0
Sylvite (KCl)
0.5 65.5
54.8
41.6
99.9
99.8
100.0
99.9
0.2
0.2
11.7
24.9
1.0
Carbon Amorphous
T102
1.3 5.4
1.0
T101
1.7 0.7
0.7
Coal C
oxy-fired
0.2
IGCC
Coal B
56.0
53.8
57.8
58.9
72.4
65.9
85.5
97.9
86.1
74.6
100.0
100.0
100.0
99.9
100.0
100.1
100.1
100.1
100.0
100.0
100.0
100.0
100.0
100.0
5
Table 2. Major and minor element chemistry Fluidised Bed Combustion Aus FBC Sample
Air and Oxy-fuel pf
Japan PFBC
Coal A
IGCC
Coal B
Coal C
Silo ash
Bottom ash
Fly ash
Bottom ash
air-fired
oxy-fired
Air-fired
oxy-fired
air-fired
oxy-fired
T100
T101
T102
T103
T104
Bugg
8.10
0.40
4.50
3.00
0.28
0.61
0.23
0.28
0.56
0.63
1.10
1.70
11.70
21.60
0.90
1.40
wt% C SiO2
51.8
58.8
18.8
27.6
64.1
60.9
49.5
48.8
50.6
52.5
41.5
48.2
41.6
34.8
51.1
54.5
TiO2
0.80
0.57
0.35
0.97
1.82
1.91
2.13
2.10
1.19
1.35
0.85
1.13
1.33
1.03
1.68
1.07
Al2O3
20.6
23.6
7.9
20.1
29.1
30.8
35.0
35.3
26.4
27.4
21.5
18.9
21.2
21.5
32.1
22.0
Fe2O3
4.29
4.05
1.71
3.50
2.72
4.56
11.3
11.3
11.5
9.35
21.5
23.2
13.2
3.16
5.84
6.73
CaO
3.74
6.37
46.90
32.20
0.63
0.83
1.67
1.66
2.81
2.28
9.94
4.86
7.21
12.80
7.31
14.90
MgO
0.61
0.77
1.34
2.02
0.46
0.61
1.13
1.12
1.72
1.47
2.80
2.18
0.77
0.93
2.08
1.12
MnO
0.05
0.04
0.04
0.05
0.04
0.08
0.19
0.18
0.06
0.07
0.12
0.09
0.08
0.03
0.03
0.05
K2O
1.20
1.66
0.53
0.43
0.55
0.41
0.44
0.45
1.49
1.71
0.48
0.45
0.22
0.47
0.60
0.78
Na2O
0.23
0.09
0.72
0.95
0.17
0.16
0.14
0.10
0.79
0.55
1.51
0.58
0.06
0.72
0.96
0.22
P2O5
0.41
0.11
0.13
0.29
0.16
0.18
0.25
0.29
0.96
0.78
0.64
0.85
1.52
1.21
0.14
0.16
SO3
1.70
0.70
4.12
6.28
0.28
0.27
0.22
0.22
1.13
1.06
0.60
0.45
0.43
0.32
0.05
0.22
11.7
24.9
100.0
100.7
101.9
101.6
98.6
98.5
101.4
100.9
99.3
101.9
101.9
101.8
LOI
14.0
2.50
14.8
6.00
Total
99.4
99.3
97.3
100.4
(nd: not determined
6
Table3. Trace element chemistry Fluidised Bed Combustion Aus FBC Sample mg/kg As B Ba Bi Cd Co Cr Cu F Hg Mo Ni Pb Sb Se Sr Th U V W Zn
Silo ash 3.02 320 470 0.2 0.11 25 121 40 970 0.14 0.7 83 43 0.37 1.98 680 16 9 148 1.1 100
Bottom ash 0.63 37 260 0.2 0.06 9 84 20 110 <0.02 0.5 5 32 0.26 <0.1 320 24 6 63 0.5 86
Air and Oxy-fuel pf
Japan PFBC Fly ash 3.9 120 390 0.2 0.5 32 35 21 244 0.01 2.3 28 28 0.6 3 540 7.8 1.7 49 <0.1 12
Bottom ash 7.6 403 1500 0.9 0.9 32 67 94 371 0.07 12 52 45 1.9 48 1600 27 6.3 160 2.2 46
Coal A air-fired 1.4 30 550 0.48 0.34 66 91 120 210 0.077 2.7 61 80 0.79 1.04 440 32 13 371 4 240
oxy-fired 1.8 54 460 0.47 0.26 56 93 110 170 0.061 2.6 69 77 0.66 1.01 480 24 12 313 5 190
IGCC
Coal B air-fired 1.6 110 280 0.34 0.22 29 137 100 120 0.086 2.1 118 84 0.32 0.6 400 <5 8 178 4 170
Coal C
oxy-fired 1.7 100 300 0.37 0.26 34 148 100 110 0.058 2.1 131 85 0.35 0.62 470 22 9 223 4 130
7
air-fired 7.6 1160 350 0.48 0.26 88 83 110 660 0.197 5.3 174 69 0.87 1.91 710 21 11 231 10 220
oxy-fired 3.7 400 370 0.37 0.26 72 151 240 630 0.269 3.6 175 63 0.68 1.77 650 19 10 228 6 270
T100 9.3 200 2900 0.79 0.2 61 1000 100 200 0.09 200 240 35 1.9 1.7 2400 28 9 150 17 180
T101 1.4 64 3600 nd nd 100 780 71 90 0.02 27 240 8 0.4 0.5 2200 40 11 140 13 75
T102 4.8 160 2200 nd 0.18 41 200 91 220 0.23 14 320 12 0.5 0.9 1400 15 4 110 5 62
T103 2.2 40 480 nd 0.16 56 190 53 380 0.3 6 98 17 0.6 0.9 390 13 5 87 4 130
T104 0.8 91 1300 nd nd 38 200 56 90 nd 14 79 4 nd 0.2 1400 17 6 240 4 17
Bugg 2.7 120 840 nd nd 51 93 29 60 nd 4 76 nd 0.2 nd 530 17 6 170 5 10
Table 4. Trace element chemistry of the 3.5:1 leachates Fluidised Bed Combustion Leach Ratio 3.5:1 pH Condy* mg/L B Ca Fe K Mg Na SO4 Si µg/L Al As Ba Cd Co Cr Cu Li Mn Mo Ni Pb Sb Se Sr Th U V W Zn
(* mS/cm;
Aus FBC Bottom Silo ash ash 8.3 12.4 2.38 8.64
Air and Oxy-fuel pf
Japan PFBC Bottom Fly ash ash 12.3 12.3 7.37 9.25
Coal A air-fired
Coal B
oxy-fired
air-fired
IGCC Coal C
oxy-fired
air-fired
Oxy-fired
T100
T101
T102
T103
T104
BUG
6.4 3.3
6.1 2.3
7.5 2.6
7.5 2.4
6.5 11
6.2 12.5
7.46 0.53
7.85 0.16
7.82 0.6
9.09 0.79
8.68 0.08
8.05 0.21
2 624 <0.1 12 8.6 5.4 1523 <5
<0.2 1010 <0.1 60.6 <0.1 12.9 684 <5
0.2 610 <0.1 39.6 <0.1 55.7 19 <5
0.4 1386 <0.1 3.9 <0.1 4.4 1578 <5
2.61 157 nd 541 1.3 28 787 8.3
2.19 183 nd 274 1.2 27 775 4
2.89 192 nd 382 1.5 26 758 3.2
3 153 nd 346 2 22 664 4. 8
36.8 693 nd 2261 2.5 104 3258 12.4
28.7 708 Nd 2668 3 102 3378 12.4
0.33 81 0.16 1.73 8.51 7.53 201 2.38
0.06 22 0.05 0.62 2.58 4.53 46 0.71
0.14 104 0.02 1.86 4.65 5.56 255 1.18
0.06 142 0.07 2.31 2.11 16.47 435 0.18
0.01 8 0.3 0.3 0.85 5.63 18 1.59
0.09 28 0.15 0.61 0.58 9.19 52 1.46
265 2.5 77 0.65 1.74 2.49 0.41 15 10 295 0.7 <0.1 5.52 55.5 7469 <0.1 2.6 190 22 10
2 0.7 345 0.44 1.44 22.91 0.88 44 <0.1 184 1.5 1 0.01 1.3 4500 <0.1 <0.1 <1 0.5 8.5
973 0.2 37048 0.13 0.8 9 2.16 9218 <0.1 66 1.4 8 <0.01 42 >> 0 <0.1 <1 0.2 16
50 1.6 105 0.11 1.6 92 0.86 402 1649 52 2.7 2 <0.01 61 16109 0.02 0.1 <1 0.15 11
50 33 60 3.5 49 2.7 14 1483 1916 561 149 0.01 24 110 2670 0.03 0.05 1286 35 670
148 9 56 2.1 40 1.4 4 1480 2597 245 129 0.02 12 45 4597 0.02 0.02 300 9 349
77 48 88 1.7 3 36 9 1903 502 515 31 <0.01 5.3 72 3216 0.05 0.22 147 26 48
48 73 45 1.8 2 45 6.8 3273 502 637 31 <0.01 6.8 70 2201 0.03 0.14 326 52 19
119 27 48 3.5 73 16 10 3155 2097 581 279 0.53 10 269 7255 <0.02 0.01 134 16 291
382 19 69 6.6 100 17 26 3852 4631 290 466 <0.01 6.1 344 9202 <0.02 0.03 74 6 1517
129 0.48 104 2.96 1.33 1.03 1.44 57 240 198 30 0.15 1.4 11 2962 0.02 0.3 -0.52 0.44 11.6
958 1.78 169 2.8 0.64 0.9 1.13 31 12 76 10 0.21 3.62 6 872 0 0.75 0.54 2.4 5.9
552 8.02 350 3.04 1.27 0.87 0.81 42 29 79 21 0.12 3.55 57 2074 0.05 0.74 0.56 1.74 9.1
5864 3.47 68 3.14 1.48 1 1.8 85 0.39 239 1 0.12 10.29 15 522 0.03 0.13 5.17 8.45 8.8
1924 8.9 53 3 0.41 1.75 2.42 9 8 19 14 1.42 1.39 22 271 0.08 0.11 7.49 0.64 9.2
126 1.06 4 2.75 2.08 0.8 0.9 15 0.24 49 11 0.16 1.41 9 74 0 0.18 0.39 1.87 2.9
>>: over range)
8
Table 5. Trace element chemistry of the 20:1 leachates
Leach Ratio 20:1 pH Condy* mg/L B Ca Fe K Mg Na SO4 Si µg/L Al As Ba Cd Co Cr Cu Li Mn Mo Ni Pb Sb Se Sr Th U V W Zn
(* mS/cm;
Fluidised Bed Combustion Aus FBC Japan PFBC Bottom Bottom Silo ash Fly ash ash ash 8.6 12.4 12.2 12.3 1.38 8.24 6.83 8.95
Air and Oxy-fuel pf Coal B
Coal A air-fired
oxy-fired
air-fired
oxy-fired
IGCC Coal C air-fired
oxy-fired
T100
T101
T102
T103
T104
BUG
6.48 0.71
6.39 0.49
7.34 0.62
7.68 0.48
6.78 2.39
6.32 2.71
7.33 0.11
7.72 0.06
7.94 0.16
9.83 0.24
8.64 0.02
7.99 0.04
0.4 322 <0.1 4 1.9 1.4 779 <5
<0.2 872 <0.1 12.2 <0.1 2.7 265 <5
0.3 621 <0.1 6 <0.1 9.4 20 <5
0.8 979 <0.1 0.8 <0.1 1.1 365 <5
0.58 30 0.08 104 0.2 5.2 150 <1
0.46 31 0.03 52 0.3 4.9 131 <1
0.75 37 0.08 79 0.3 5.2 150 <1
0.56 27 0.08 63 0.3 3.8 106 <1
7.28 134 0.08 400 0.4 19 602 2.5
6.83 136 0.08 472 0.6 18 616 1.8
<0.04 16.2 0.05 0.5 1.15 1.36 24 0.91
<.0.01 9.6 0.04 0.37 0.8 1.43 12 0.72
<0.01 24 0.02 0.65 0.89 1.33 46 0.86
<0.01 39.7 0.02 0.7 0.63 3.38 85 1.16
<0.01 1.9 0.06 0.25 0.15 1.21 0.9 0.18
<.0.01 5.3 0.42 0.38 0.12 1.59 8.4 0.44
499 2.1 103 0.12 0.58 1.83 0.22 8.3 4 52 0.4 <0.1 2.22 12 3464 -0.02 0.8 125 6.9 8.5
1.7 0.6 268 0.13 1.13 8.63 0.56 16 <0.1 50 1 1 0.01 1 1535 -0.03 <0.1 1 0.7 7.5
1392 0.3 12067 0.12 0.8 7 1.41 1322 <0.1 48 1.1 8 <0.01 37 15830 0.01 <0.1 <1 0.25 11
5 0.9 270 0.05 1.2 37 0.68 82 2118 22 2.8 2 <0.01 13 4005 <0.01 <0.1 <1 0.11 8.7
54 10 26 0.8 14 1.3 9.3 278 611 92 44 0.06 6 22 588 <0.02 0.04 283 7.5 290
39 7.6 25 0.5 5.7 3.4 1.1 303 418 105 26 0.07 5.3 12 796 <0.02 <0.02 282 7.5 47
104 33 28 0.4 0.6 11 2.6 378 87 130 8 0 2 18 668 <0.02 0.02 96 13 16
196 60 12 0.5 0.37 13 2.1 453 47 162 5 0.19 2.5 18 460 <0.02 0.02 285 25 10
36 21 9.2 0.8 13 2.4 2.6 588 37 224 52 <0.01 4.6 59 1324 <0.02 <0.02 155 14 46
119 14 17 1.1 21 1.8 7.5 772 795 104 99 <0.01 2.9 72 1813 0.02 0.02 56 4.8 288
337 0.22 65 2.8 0.25 0.64 0.92 9.51 24 49.83 4.93 0.22 0.5 2.8 783 0.02 0.03 0.73 0.43 3.4
1029 1.55 112 2.78 0.21 0.71 1.01 6.93 9.9 14.31 2.51 0.3 0.7 2.4 362 <0.01 0.07 0.85 0.63 5.5
1428 24.8 171 2.99 0.2 0.67 1.85 9.52 4.9 17.34 3.51 0.22 1.2 37 594 <0.01 0.05 2.17 0.85 4
3843 4.65 23 2.67 0.36 0.64 1.39 18.61 0.1 37.82 0.46 0.29 4.17 11 134 <0.01 0.04 8.43 2.95 8.1
279 0.79 10 2.69 0.26 0.53 1.6 1.24 17 1.58 11.47 0.65 0.2 1.2 28 0.01 0.01 0.21 0.23 12
162 0.71 2.1 2.74 1.37 0.61 1.73 2.4 0.5 6 8.4 0.15 0.31 2.2 16 <0.01 0 0.45 0.78 4.5
>>: over range)
9
Generally, the compositions of the slags, when expressed as oxides, appear to be similar to those of ashes produced by other transitional technologies (Table 2) This may in part be a result of expressing the compositions determined by XRF in the traditional manner, as it is expected that more reduced species of some elements would in fact be present (e.g. FeO rather than Fe2O3). The trace element concentrations in the slags are also similar (Table 3). However, the concentrations of some of the more volatile elements, such as Bi, Cd, F and Pb, tend to be lower. Cr, Mo, Ni, and Th contents tend to be higher, although the Cr may possibly represent contamination from Cr-based refractories. The overall leachability (or solubility) of elements from the slags is lower than that for typical ashes from Australian pf-fired power stations tested in a similar way (e.g. Jankowski et al., 2006; Riley, 2007), or from the FBC or oxy-fuel fired systems in the present test program. In particular, B, Ca, K, Al, Co, Cr, Se and V concentrations are lower .This is also evident from a comparison of the conductivities of the leachates (Tables 4 & 5). The conductivities of the leachates (l:s = 20:1) from the FBC ashes range from 1.38 to 8.24 mS/cm; the conductivities of the ashes from the conventional pf and oxy-fuel tests range from 0.48 to 2.71 mS/cm, whereas the conductivities of the leachates from the IGCC slags range from 0.02 to 0.24 mS/cm. However some trace elements appear from the study to be leached from the IGCC slags at similar concentrations to those leached from conventional pf-fired power station ashes. Table 6. Comparison of leachates (liquid:solid = 20:1) from ashes/slags produced by conventional and transitional technologies
pH µg/L As B Cd Co Cr Cu Mn Mo Ni Pb Sb Se V Zn
Conventional pf #
FBC
Oxy-fuel
IGCC
4.4 - 12.0
8.6 – 12.3
6.3 - 7.7
7.3 - 9.8
<2 - 20 60 - 945 <0.2 - 5.5 <0.2 - 145 <1 - 48 <1 - 275 <0.2 - 365 5 - 275 1 - 193 <0.2 - 1.9 <0.1 - 9 3 - 60 5.6 - 70 < 1 - 1900
0.3 - 2.1 <200 - 800 0.05 - 0.13 0.6 - 1.2 1.8 - 37 0.2 - 1.4 <0.2 - 2118 22 – 52 0.4 – 3 <0.2 - 8 <0.1 - 2 1 – 37 <5 - 125 8 – 11
7 – 60 460 – 7280 0.4 - 1.1 0.6 – 21 1.3 – 13 1 - 9.3 37 – 795 92 – 224 5 – 99 <0.2 - 0.2 2.0 - 6.0 12 - 72 56 - 285 10 - 290
0.2 - 25 <10 - 40 2.8 - 3.0 0.2 - 1.4 0.5 - 0.7 0.9 - 1.9 0.1 - 24 1.5 - 50 0.5 - 12 <0.2 - 0.7 0.2 - 4.2 1 - 37 <5 - 9 3 - 12
(# data extracted from Riley 3 )
10
CONCLUDING REMARKS The aim of this study was provide a comparison of the leachability of trace elements from the fly ash of conventional pf coal-fired power stations and major waste streams (ash or slags) of transitional technologies such as FBC, coal-fired oxy-fuel stations and IGCC systems. The results of this study for a number of the environmentally significant trace elements can be compared to those of an earlier study 3 in which results were given for trace elements in leachates (l:s ratios of 20:1) of fly ashes from conventional pffired power stations in Australia. In Table 6, the ranges from that study are listed and compared with the ranges of a number of trace elements of the ashes or slags (from the transitional technologies) studied in this work. It is apparent from the present study that generally: •
Although there are possibly some exceptions, most trace elements in leachates from the fly ash and bottom ashes of FBC systems were found to occur at similar concentrations to those in leachates of ashes from conventional power stations (i.e. using coal fired as pulverised fuel in air).
•
Lower concentrations of Cd, Co, Ni and Zn were found in the leachates of the ash from the Australian and Japanese FBC power stations.
•
Trace elements are leached in similar concentrations from fly ash from coal fired in conventional and oxy-fuel combustion conditions; indeed there does not appear to be any significant differences in the leaching behaviour of trace elements in the ashes from the different systems.
•
Some of the trace elements in the gasifier slags do not leach as readily as those in the fly ash from conventional power stations; these trace elements are B, Cr, Mn, V and Zn and possibly Co and Cu. However, there does not appear to be any change in the leachability of the other trace elements, including As and Se. IGCC slag obviously has a larger particle size than pf ash, and the smaller surface area of the IGCC slags impacts upon element leachability. Although it is sometimes claimed that slag from IGCC plants “is environmentally benign and can be safely landfilled”, care must obviously be taken with its disposal as there are still trace elements that can be leached.
ACKNOWLEDGEMENTS Financial support for this project was provided by the Co-operative Research Centre for Coal in Sustainable Development (CCSD), which was sponsored in part by the Co-operative Research Centres Program of the Australian Government. Irene Wainwright and Dorothy Yu from UNSW and Owen Farrell, Rosemary Wood and David Jacyna from CSIRO are acknowledged and thanked for their assistance with the analytical work associated with the study.
11
REFERENCES [1] Ward, C.R. and French, D. (2006). Determination of glass content and estimation of glass composition in fly ash using quantitative X-ray diffractometry. Fuel 85, 2268– 2277. [2] Jankowski, J., Ward, C.R., French, D. and Groves, S. (2006). Mobility of trace elements from selected Australian fly ashes and its potential impact on aquatic ecosystems. Fuel 85, 243-256. [3] Riley, K.W., French, D., and Ward, C.R. (2007). Benchmarking of environmental credentials of Australian fly ash. Proceedings of World of Coal Ash Symposium, Covington, Kentucky, May 8-10 2007, Center for Applied Energy Research, University of Kentucky, 17 pp. (CD publication).
12
1
BOROVAC COAL CLEANING PROCESS Branislav Grbović, Director, Borovac International Pty Ltd P.O. Box 494, Dianella WA 6059, Australia, E-mail: [email protected] Jelenko Mičić, Assistant Director, Mining Basin Kolubara d.o.o. Svetog Save 1, 11550 Lazarevac, Serbia; E-mail: [email protected] Miroslav Spasojević, Chief of Sustainable Development Section, “Nikola Tesla" Power Plants B Urosevica 44, 11500 Obrenovac, Serbia, E-mail: [email protected] Miloljub Grbović, Technical Consultant, Borovac International Pty Ltd Bilecka 28a, 11040 Belgrade, Serbia, E-mail: [email protected]
Abstract The western parts of Serbian Kolubara coalfields contain significant quantities of interlayer waste that is causing considerable operational difficulties and inefficiencies at the mine and power plants. Future development of open pits will further worsen these issues as pits are expanding into the areas of lower quality coal, while at the same time the coal supply demand will increase due to construction of new power plant units. Testwork and investigations focused on Kolubara’s Tamnava coal have been conducted over the years aimed at developing a process capable of removing interlayer waste from complex coal seams typically present in most of Serbian lignite deposits. The results have confirmed that „Borovac“ coal cleaning process is capable of breaking the bonds between coal and waste and subsequently separating the waste from ROM coal. The implementation of the process will enable more efficient operation of existing plants and better utilisation of available coal reserves. The process has a potential to contribute significantly to reduction in overall environmental pollution caused by the products of combustion. This new Technical–Ecological–Commercial-Ethical concept is aimed to resolve the actual cause of the problem and prevent or minimize its chance of reoccurring in future, rather than attempting to heal the consequences once the problem has already been created. Key terms: Low quality coal (LQC), Scrubbing, Coal cleaning, Environment, Utilisation
Technical Problem The mining and power complex Kolubara-Obrenovac includes the mines and power plants originally designed and constructed on the assumption that a high quality coal (HQC) from eastern coalfields of the Kolubara basin will be available to feed both Kolubara and Nikola Tesla Power Plants (TENT).
2 However, the deep homogenous layers from Kolubara’s eastern coalfields have been exhausted during previous decades and mining activities are being shifted towards western coalfields that contain a low quality coal (Figure 1). Figure 1 Plan of Kolubara Coalfields Field Тamnava West Radljevo Zvizdar Shopic
кЈ/kg 6,589 6,483 5,639 5,550
Field kJ/kg T- D 8,032 T-Е 7,750 T-G 7,320 T-F 7,164 V.Crljeni 7,793 Т.East 7,319
Note: LHE values shown assume that mining includes waste interlayers up to 0.5m in thickness
Western coalfields contain interlayer waste (Figure 2), that is causing significant operational difficulties and inefficiencies at the mine and power plants. The low quality of coal will undoubtedly remain the single most important reason for insufficient power supply capacity that is facing the Serbian Power industry in a long-term. Existing process of mining and handling of the low quality coal at Tamnava West pit (Figure 3) has not been adapted yet to the natural physical characteristics of coal from remaining unexcavated parts of the deposit. Mining activities on complex coal-waste multilayer seams are performed by three excavators, which aim to selectively mine the coal and interlayer waste. Each excavator has its own dedicated discharge conveyor named M1, M2 and U6 respectively. When an excavator works on the coal layer, its respective conveyor is discharging coal to a collective transfer conveyor, which delivers the coal to crushing facility and further to power plants.
3 Figure 2 Western Tamnava Mine - Longitudinal sections
When excavator works on a complex coal seams and reached the interlayer waste and/or LQC (out-of-balance coal), it is diverted onto a collective waste and dumped into an internal landfill waste deposit (Figures 4 and 5). Figure 3 Existing handling of coal and waste at the western Tamnava deposit
This method of operating the high-capacity equipment at Tamnava mine is not cost-efficient and has two major disadvantages:
4
Figure 4 Western Tamnava - Appearance of dumped LQC coal within the internal landfill (June 2009) (Waste - lighter colour; Disposed coal - dark colour)
• •
Inadequate utilisation of the large mining equipment due to the excavators and conveyors not being able to work at their full capacity; and Given that the excavators are not designed for efficient selective mining, much of the coal from thinner layers is currently being excavated and disposed together with waste. This results in significant losses to recovery of available coal reserves from the deposit. Figure 5 Internal mine waste deposit detail
5 The existing mining process at Tamnava mine is not adapted to the physical characteristics of the coal from remaining unexcavated areas of the deposit. It creates numerous difficulties not just for the mining activities, but also for the power plants that use this coal for production of electricity. Impact of Coal Quality Inorganic waste mineral matter most frequently presents itself in the forms of pure clay, sand or sand-clay matter. Long-term statistics have fairly accurately defined the relationship between the Lower Heat Effect value (LHE) of Kolubara lignite and its content of waste matters - moisture M (%) and ash A (%): LHE = 25,053 - 250.53 A - 273.56 M
(kJ/kg)
In actual practice, it is established that the mine is required to supply payable units of coal that are strictly in excess of 5,280 kJ/kg. Moisture content of the Kolubara coal is normally in the range from 45% to 55%. This effectively means that i.e. open-pit coal with 45% moisture should contain no more than 28% of ash to be economically used. TENT has over the last few years published very detailed recordings on the impact of lower heat effect value of coal on the production levels and costs of electricity production. The impact of LHE on the efficiency of installed units at TENT is shown in Table 1. Even with the partial use of burners to enhance the fire in boilers, the utilisation of power plant’s unit capacities is low. Table 1 Effect of coal heat value on heat recovery
Power Plant Unit 6305 LHE kJ/kg 1 Burner Nos. 2,4,5 Coal *) structure 90 Recovery % Power Plant Unit LHE kJ/kg Burner Nos. Coal structure Recovery % *)
1
2
4608 6085 1
6653
7563
0
1
2,4,5
1,4,5
88
64
4 5386 8365 0
6508 2
1,4,5 1,2,3,4,5 2,4,5 90
88
91
4906 5872 2
5549 6477 1
5428
7187
6974
0
0
0
1
1
2,4,5
1,5
2,4,5
80
94
78
78
78
5 4470 6100 2
4894 8844 2
6907
1,2,5 91
6 4697 5596 1
6059
5733
6104
2
4908 5678 1
0
0
0
4330 5758 2
1,2,4,5
2,4,5
1,2,3
2,3,4,5
2,4,5
1,5
2,4,5
2,4,5
85
91
85
85
77
77
77
92
Coal structure: 1 – Xhilite; 2 – Sand; 3 – Young „lika“ wood; 4 – Soil, mud, clay; 5 – Non-homogenious
6
The wooden structure of multilayered sections is very profound. An increased wooden structure specifically negatively affects the block capacity utilisation (Table 2). Table 2 Effect of coal structure of coal on heat recovery
Power Plant Unit No.
1
2
LHE, kJ/kg
8512
8570
8215
7563
7837
8966
7475
8336
Burners, Nos.
0
0
0
1
0
0
0
0
Coal structure
1
1
1
1,4,5
1,5
1
1,5
1
Recovery, %
92
67
64.3
90
92
66
64
64.3
Apart from causing the reduction in power plant’s capacity utilisation, the decrease in the coal quality from western Kolubara coalfields also creates other difficulties: • • •
•
Total mass of coal transported from mine to power plants is increased Large consumption of crude oil is required to sustain fire in the boilers Because of a greater mass of silica sand in the coal from unexcavated parts of the deposit, there is significantly increased wear rate on equipment i.e. conveyors, mills, tube heat exchangers, filters and pumps for ash slurry. This affects the availability of the plant, reduces production levels and increases maintenanace costs. Increased content of waste inorganic matter in coal as a consequence has an increased environmental pollution and also a reduction in the lifetime of existing ash landfills.
The Borovac process offers possibile solution Numerous development programs have been conducted over the years aiming to find a coal cleaning process that is suited to natural characteristics of lignites from Balkan region. Differences in the physical properties between pure coal matter and waste inorganic matter from complex layers of Tamnava mine, served as the basis for development of hydraulic coal cleaning process named „Borovac“ (the process is patent protected). It should be noted that there are two types of waste substances that reduce the heat value of lignite, which are: •
Internal minerals that are formed from inorganic compounds at the time that the coal itself was being created. A portion of this matter is negligible, its connection to pure coal is stable and it is not likely to be removed from the excavated coal.
•
External mineral matter, which has emerged during the periods of interruption in formation of layers of clean coal. The share of this type of matter in the unexcavated parts of Kolubara deposit is large. It appeared in the deposit predominantly mechanically
7 in the forms of clay and/or sand. The connection of these minerals with pure coal is unstable as it has not been consolidated as a firm rock. A hydraulic method of scrubbing can break the bonds between the waste elements and pure coal and later separate them. The testwork and studies have demonstrated that from a low quality coal, the cleaning process can produce high quality coal with heat recovery of about 85%. It is to be noted that an extensive laboratory-scale testwork has been conducted on lignites from all major Serbian deposits (Table 3). The testwork results have demostrated the suitability of
the cleaning process and subsequent investigations have focused on Kolubara deposit. Table 3 Borovac Testwork Results
Coal deposit 1
Product 2 ROM coal
Kolubara Cleaned coal
Kostolac
Kosovo
Kovin
Ugljevik
Nova Manasija
Quality LHE, kJ/kg Ash, % 3 4 4,612 32.4
Mass 5 100
Distribution, % Heat 6 100
Ash 7 100
8,420
20
46.7
85.3
28.8
Waste
1,276
43.3
53.3
14.7
71.2
ROM coal
7,582
18.8
100
100
100
Cleaned coal
9,731
11.3
66.2
85
39.7
Waste
3,373
33.6
33.8
15
60.3
ROM coal
5,545
11
100
100
100
Cleaned coal
8,322
8.1
59.3
89
43.5
Waste
1,498
15.3
40.7
11
56.5
ROM coal
7,150
24.2
100
100
100
Cleaned coal
9,630
12.3
70
94.5
35.5
Waste
1,250
52
30
5.5
64.5
ROM coal
11,411
25.4
100
100
100
Cleaned coal
14,212
17.3
72.5
89.4
52
Waste
3,971
50
27.5
10.6
48
ROM coal
5,990
44.3
100
100
100
Cleaned coal
13,632
17.4
42.5
97.3
85
320
68.2
57.5
2.7
15
Waste
Semi-industrial scale testwork results on Kolubara coals have also confirmed that the coal can be cleaned and the required quality for supply to existing power plants succesfully achieved. An example of the achieved results is shown in Table 4 and Figure 6.
8
Table 4 Borovac Coal Cleaning Test on Kolubara
Product
Quality LHE, kJ/kg Ash, % 4,960 30.1
Run-of-Mine Coal
Distribution, % Mass Heat Ash 100.0 100.0 100.0
Cleaned Coal
8,200
16.4
52.0
85.9
30.0
Waste
1,460
43.8
48.0
14.1
70.0
Figure 6 Coal cleaning results on samples from Tamnava West coal supplied to power plants 100 90
ROM Coal Cleaned Coal
80 70
Mass, %
60 50 40 30 20 10 0 2.5
3.9
4.3
4.7
5.1
5.5
5.9
6.3
6.7
7.1
7.5
7.9
LHE, MJ/kg Coal cleaning implementation at Kolubara The implementation of the coal cleaning process into existing Kolubara flowsheet and facilities layout is relatively simple and without major technical obstractles i.e.: -
-
No significant changes are required to existing open pit at Tamnava West. The mining of coal layers with thicknesses in excess of 0.5m will continue with the same machinery as now (Figure 7). Overburden will continue to be diverted using the same conveyors and be dumped into an excavated East Tamnava area that is allocated for overburden and waste disposal. The process of coal cleaning will be conducted independently from mining of coal, so there will be no delay or disruptions to current production process. No technical changes are required at the TENT power plant side.
9
Figure 7 Integrated Plant Flowsheet with LQC Cleaning
Benefits of implementing coal cleaning process on LQC The Borovac process represents a new technical-ecological-economical-ethical concept, which resolves existing problems and decreases the probability of significant difficulities in future. Numerous benefits will be achieved by introducing the cleaning process on the low quality coals (LQC), including the following:
10 •
•
•
• •
•
Mines at Kolubara are offered an opportunity to produce and supply a good quality coal from their multilayered coalfields. This will enable power plants to operate closer to their original design duty, which they initially used to do when burning only good quality coals from homogenous eastern fields of the Kolubara basin. In summary, the existing plant units have the potential to increase utilisation of their capacity by 15%, which is equivalent to one additional 350MW unit! Commercially the coal cleaning will make future plant capacities (Kolubara B and extension of TENT B) much more attractive as an investment and much better justified on the assumption that supply of better quality coal is secured to feed them. Also it will make easier for the government and power industry of Serbia to secure a fnancial support for construction of these new units. Of particular importance is the fact that this process generates a clayish waste as a byproduct, which can be used as a protective cover/liner to protect soil from the drainage from fly ash disposals. This would allow the fly ash quantities that are not used for cement and roadwork, to be safely disposed within excavated spaces of Kolubara pits. This cleaning process also removes about 25% of sulfur in coal, which consequenlty improves the ratio of basic to acidic elements in the cleaned coal in favour of basic. This in turn assists in reducing SO2 gas emissions from flue gases at TENT. The Serbian Government has responsibility to actively contribute to global efforts in reduction of emission of pollutants into the environment. The ash disposals and stacks from TENT power plants are among the largest emitters in Serbia. The coal cleaning process would greatly assist Serbia to fulfil its duty. The Western coalfields of the Kolubara basin hold an estimated 350 million tonnes of unexcavated low quality coal, until now accounted for as an “off-balance coal”. This coal is either being dumped with overburden to an internal landfill, or is effectively wasted during combustion. By introducing the cleaning process to this coal it is possible to produce around 175 million tonnes of a good quality coal with LHE of approximately 8,000 kJ/kg and estimated value of 175,000,000 x 12 Euro/t = 2.1 billion Euro. This would represent a significant financial benefit to the company, country and overall region.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11–14, 2010 PROGRAM TOPIC: ENERGY SUSTAINABILITY – EFFICIENCY AND CONVERSION TO REDUCE GHG
REDUCING GREENHOUSE GAS EMISSIONS FROM COAL COMBUSTION BY ADDING MICRO-ALGAL BIOMASS Jacobus Brink, Sanette Marx School of Chemical and Minerals Engineering, North-West University, Potchefstroom 2520, South Africa
Abstract Each year millions of tons of greenhouse gases are released during combustion of coal in coal-fired power plants. Supplementing the coal feedstock to these power plants with micro-algal biomass, could maintain energy production, while decreasing the net contribution of greenhouse gases to the atmosphere, with resultant improvements to the environment.
The research project compared the calorific value and the elemental composition of low-grade Highveld coal, one of the major coal types mined in South Africa and one of the main types used for power generation in South Africa, against two sources of micro-algal biomass. One of the sources of micro-algal biomass was the dominant species in the Hartbeespoort Dam (37 km west of Pretoria, South Africa), Microcystis aeruginosa, which was harvested directly from dam. The second source of micro-algal biomass, Cyclotella meneghiniana, was cultivated in 4-liter Erlenmeyer bioreactors, using a modified growth medium. The low-grade Highveld coal had a calorific value of 21.65 MJ/kg, and a carbon, hydrogen, nitrogen and sulfur content of 56.06, 2.33, 1.51 and 0.78 mass %, respectively. In comparison, micro-algal biomass from the Hartbeespoort Dam had almost a similar calorific value of 21.02 MJ/kg compared to the low-grade coal’s; and a carbon, hydrogen, nitrogen and sulfur content of 42.00, 6.75, 8.87 and 0.72 mass %, respectively. The cultivated micro-algal biomass, Cyclotella meneghiniana, had a lower calorific value of 17.30 MJ/kg; and a carbon, hydrogen, nitrogen and sulfur content of 38.80, 6.49, 6.41, 0.33 mass %, respectively. Supplementing coal feedstock to coal-fired power plants with micro-algal biomass holds great potential for minimizing the net production of greenhouse gases.
Keywords: Highveld coal; Hartbeespoort Dam; Micro-algal biomass; Calorific value; Greenhouse gases
1. Introduction According to the Energy Information Administration’s International Energy Outlook 2010 (IEO2010), global energy demand is projected to increase by 49 percent in the next 28 years. Total world energy use is projected to rise from 495 quadrillion British thermal units (Btu) in 2007 to 590 quadrillion Btu in 2020 and 739 quadrillion Btu in 2035. In the absence of national policies and binding international agreements that would limit or reduce greenhouse gas emissions, world coal consumption is projected to increase by 56 percent from 132 quadrillion Btu in 2007 to 206 quadrillion Btu in 2035, at an average annual rate of 1.6 percent. Coal’s share of world energy consumption is expected to grow from 27 percent in 2007 to 28 percent in 2035. World energy-related carbon dioxide emissions are expected to grow from 29.7 billion metric tons in 2007 to 33.8 billion metric tons in 2020 and 42.4 billion metric tons in 2035. Coal’s share of world carbon dioxide emissions, which grew from 39 percent in 1990 to 42 percent in 2007, is projected to increase to almost 46 percent by 2035. Coal is the most carbonintensive of the fossil fuels, and it is the fastest growing carbon-emitting energy source.1
Growing concern over the impact of global warming on the environment, and the increased usage of fossil fuels worldwide, which is regarded as one of its prime contributors, is forcing governments and corporations to look into alternatives for fossil fuels like coal. Of these alternatives, biomass is one of the fastest growing renewable energy sources. Among the many sources of biomass, like wood and terrestrial crops, microalgae promise the highest yield per kilogram of biomass and also have much faster growth-rates than terrestrial crops,2 making them one of the most promising supplements for fossil fuels like coal. One of the largest reservoirs of micro-algal biomass in South Africa is the Hartbeespoort Dam, located 37 km west of Pretoria, and covering a surface area of approximately 20 km2. The Hartbeespoort Dam has great potential for micro-algal biomass production and beneficiation due to its size and close proximity to urban and industrial centres.
Figure 1: Hartbeespoort Dam
2
2. Objective of study The research project compared the calorific value and elemental compositions of low-grade Highveld coal against two sources of micro-algal biomass, in order to evaluate their potential for supplementing coal feedstock to coal-fired power plants. 3. Theory Coal currently provides 73% of South Africa’s primary energy and 95% of the country’s electricity is derived from coal-fired thermal generation. Coal in South Africa is found in 19 coalfields, located mainly in KwaZulu-Natal, Mpumalanga, Limpopo, and the Free State provinces. The Minerals Bureau estimates the total coal reserves in South Africa at 33.8 billion tons, and projects it to last until around 2050. The Highveld Coalfield is the second most productive coalfield in South Africa with ten operating collieries. In 2001, it accounted for about 73.65 million tons (24.92 %) of the total run-ofmine production. Low-grade Highveld coal consist of bituminous coal with an ash content of 20–35% and a calorific value of 18–25 MJ/kg.3
Microalgae, like all other biomass sources, are renewable energy sources, and do not contribute to the net emission of greenhouse gases, as they absorb the constituents of their biomass directly from the environment. The dominant micro-algal species in the Hartbeespoort Dam is the cyanobacterium Microcystis aeruginosa, which comprises over 90% of the total algal biomass in the dam. Excessive growth of these algae is stimulated by the influx of large quantities of nitrates and phosphates into the dam, from the run-off of fertilized agricultural land and effluent from sewage plants from the northern suburbs of Johannesburg.4
3
4. Experimental 4.1. Hartbeespoort Dam 4.1.1. Culture organisms Samples of Microcystis aeruginosa micro-algal biomass were collected from the Hartbeespoort Dam in 500 mL liter plastic bottles (see Figure 2).
Fig. 2: Micro-algal biomass on the banks of the Hartbeespoort Dam
4.1.2. Harvesting procedure The contents of the 500 mL plastic bottles were allowed to settle for one day (see Fig. 3). The bottom water layer was drained with a tube and the remaining top layer of concentrated micro-algal biomass pulp (wet weight micro-algal biomass) was poured onto metal palettes with a layer of approximately 500 g of building sand with a voidage of 40%. The palettes were then placed on top of the laboratory roof of the North-West University’s School of Chemical and Minerals Engineering, and allowed to dry for 24 hours.
Figure 3: Concentrated micro-algal biomass pulp separating to the top after 24 hours
4
Figure 4: Micro-algal biomass from the Hartbeespoort Dam drying in the sun on top of a base of building sand, located on top of the roof of the laboratory at the North-West University at Potchefstroom
Figure 5: Sun-dried palette of micro-algal biomass from the Hartbeespoort Dam on top of a base of building sand, located on top of the roof of the laboratory at the North-West University at Potchefstroom
Figure 6: Sand-filtered, sun-dried Hartbeespoort Dam micro-algal biomass flakes
5
Small samples of dried micro-algal biomass flakes were sent to the SEM Laboratory at the School of Chemical and Minerals Engineering, for electron-microscopy. Figures 7 and 8 show sun-dried microalgal biomass from the Hartbeespoort Dam on scales of 20 and 2 microns, respectively.
Figure 7: Shows a microphotograph of sun-dried Hartbeespoort Dam micro-algal biomass at 20 microns (SEM Laboratory, 2009)
Figure 8: Shows a microphotograph of sun-dried Hartbeespoort Dam micro-algal biomass at 2 microns (SEM Laboratory, 2009)
6
4.2. Cyclotella meneghiniana 4.2.1. Culture organisms Cultures of Cyclotella meneghiniana were obtained from the Botany Department of the North-West University. 4.2.2. Culture medium For the cultivation of C. meneghiniana, a growth medium was prepared according to the modified recipe set out by Krüger & Eloff.5
4.2.3. Culturing procedure Erlenmeyer flasks were used to cultivate C. meneghiniana algae. The flasks were filled to 4000 mL with growth medium prepared according to the above-mentioned recipe. A pipette was used to inoculate each 4-liter growth medium with 10 mL of C. meneghiniana. Continuous aeration was provided with air pumps. The cultures were kept indoors, but exposed to a 12 hour light and 12 hour night cycle.
Figure 9: Cultivation of Cyclotella meneghiniana in 4-L Erlenmeyer flasks
4.2.4. Harvesting procedure After six weeks of cultivation, the entire content of each Erlenmeyer flask were filtered through 0.7 μm glass microfibre filter paper and the biomass placed in an oven to dry for an hour at 110 oC. 7
4.3. Harvesting micro-algal biomass from the Hartbeespoort Dam The practicality of harvesting micro-algal biomass from the Hartbeespoort Dam was demonstrated with a small scale experiment (figures 10 and 11). Fourty liters of micro-algal biomass were collected from the dam and poured onto a patch of 4 m2 of cleared soil and allowed to dry for seven days. The dried micro-algal flakes were then collected manually.
Figure 10: A 4 m2 patch of cleared soil containing a layer of wet Hartbeespoort Dam micro-algal biomass
Figure 11: Sun-dried patch of micro-algal biomass from the Hartbeespoort Dam, on its bank at Kosmos
4.4. Calorific value measurement The calorific value of the coal and the two micro-algal biomass samples were measured with a CAL2k bomb calorimeter.
8
4.5. Elementary analysis The elementary analysis of the coal and the two micro-algal biomass samples were performed with a Leco 2000 CHN analyzer for the carbon, hydrogen and nitrogen analysis and a Leco SC-432 analyzer for the sulfur.
5. Results and discussion Harvesting of micro-algal biomass from the Hartbeespoort Dam could be performed relatively easily, utilizing abundant, renewable, solar energy for drying. A combination of sand filtration and solar drying yielded 4.7 g/L of dry weight micro-algal biomass per liter of dam water or 97.8 g/m2 of sundrying area, respectively. The small-scale experiment demonstrated the potential of harvesting microalgal biomass from the Hartbeespoort Dam energy-efficiently.
Table 1 compares the analysis of low-grade Highveld coal to cultivated Cyclotella meneghiniana and harvested micro-algal biomass from the Hartbeespoort Dam.
Table 1: Elementary and calorific analysis of coal and two micro-algal biomass samples
Content
Units
Coal
CM
HMB
C
% (wt.)
56.06
38.80
42.00
H
% (wt.)
2.33
6.49
6.75
N
% (wt.)
1.51
6.41
8.87
S
% (wt.)
0.78
0.33
0.72
CV
MJ/kg
21.65
17.30
21.02
The carbon content of the two biomass samples are more than 25% lower than the coal sample’s, while the hydrogen and nitrogen contents of the biomass samples, are thrice and four to six times that of the coal sample’s, respectively. The sulfur content of the cultivated C. meneghiniana is half the coal sample’s, while the biomass harvested from the Hartbeespoort Dam had almost a similar sulfur content than the coal sample. The calorific value of the cultivated C. meneghiniana was 20% lower than the coal sample’s, while the calorific value of the Hartbeespoort Dam micro-algal biomass was almost the same as the coal sample’s, making the latter a suitable supplement for low-grade coal feedstock for coal-fired power generation.
9
6. Conclusions Analysis of two micro-algal biomass samples revealed that harvested micro-algal biomass from the Hartbeespoort Dam had a relatively similar calorific value than low-grade Highveld coal. Supplementing this low-grade coal feedstock in coal-fired power plants with harvested micro-algal biomass from the Hartbeespoort Dam could maintain energy production, while reducing the net emission of greenhouse gases. In 2009 the South African Department of Water Affairs removed 4460 m3 or 21 metric tonnes of micro-algal biomass from the Hartbeespoort Dam.6 This is equivalent to 20 metric tonnes of low-grade Highveld coal or 441 gigajoules of energy. Utilizing the total surface area of the Hartbeespoort Dam has the potential of producing 9400 metric tonnes of micro-algal biomass, which is equivalent to 9100 metric tonnes of low-grade Highveld coal or almost 200,000 gigajoules of energy, enough to supplement a medium-scale coal-fired power plant with 1% of its total annual fuel requirements. Nomenclature CM
Cyclotella meneghiniana
CV
Calorific value
HMB
Hartbeespoort Dam micro-algal biomass
References 1. Energy Information Administration (EIA). International Energy Outlook 2010. Available from: www.eia.gov/oiaf/ieo/index.html. 2. Christi Y. Biodiesel from microalgae beats bioethanol: Trends in biotechnology. 2008;26(3):126131. 3. Jeffrey LS. Characterization of the coal resources of South Africa. The Journal of the South African Institute of Mining and Metallurgy. Feb. 2005. 4. Owuor K, Okonkwo J, Van Ginkel C, Scott W. Environmental factors affecting the persistence of toxic phytoplankton in the Hartbeespoort dam: Report of the Water Research Commision of South Africa, WRC Report:1401/3/07, p.1-89; 2007. 5. Krüger GHJ, Eloff JN. The influence of light intensity on the growth of different Microcystis isolates. J. Limnol. Soc. sth. Afr. 1977;3:21-25. 6. Department of Water Affairs. Harties metsi a me programme’s positive impact on Hartbeespoort Dam. 23 Sep. 2009.
10
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
PROGRAM TOPIC: 6. COAL SCIENCE THE CREATION OF GEOREACTOR GLOBAL SCIENTIFIC NETWORK Jan Rogut*, GIG, Central Mining Institute, Katowice, Poland Hans Bruining, Geoengineering, Delft University of Technology, The Netherlands Elizabeth Burton, Lawrence Livermore National Laboratory, California Jozef Dubinski, GIG, Central Mining Institute, Katowice, Poland Thomas Kempka, German Research Centre for Geosciences (GFZ) Sohei Shimada, University of Tokyo, Japan Subhas, K. Sikdar, US EPA, National Risk Management Research Laboratory, Cincinnati, OH Hema J. Siriwardane, West Virginia University, Morgantown, WV, USA Marc Steen, Institute for Energy, Joint Research Centre, European Commission Petten, NL Aleksandra Tokarz, GIG, Central Mining Institute, Katowice, Poland Tomasz Wiltowski, Southern Illinois University, Carbondale, IL, USA
Abstract: Sustainable, responsible and eco-friendly exploitation of geological space and resources requires gathering, ordering and integrating an enormous amount of scientific knowledge that has been generated in separate, different fields of geosciences. Critical, target-oriented evaluation, unification and utilization of data, models and approaches of geology, geophysics, geochemistry and related scientific geo-disciplines are required for setting up a trustworthy geo-technology platform aimed at facilitating low-emission production of energy and raw materials in-situ. Underground coal gasification, geothermy, enhanced oil and coal-bed methane recovery, long term safe storage of nuclear and hazardous wastes, CO 2 storage in deep geological formations and most recently methane extraction from gas shale are examples of key industrial technologies and processes that should benefit from such an initiative. Overcoming the present lack of sound, science-based, reliable and timely response by engineers and decision makers to growing societal concerns about environmental, technical and economic risks which could delay or fully stop the development of geotechnologies is imperative to realise their considerable potential to provide energy and materials for the growing human population in a clean, low-emission way. Because of their multi-disciplinary character and of the very limited number of tools available for controlling in-situ operations, geotechnologies rank among high risk engineering activities. Access to their supporting science and engineering base is limited because the major part of the knowledge has been generated in the frame of proprietary research by oil or mining technology giants or in highly classified nuclear weapons testing programs. Current increasing attention to climate change and to tightened environmental regulations opens new and additional opportunities and challenges for geo-research. The Interdisciplinary nature of geo-technologies generates a set of important questions but at the same time provides driving forces for research in a variety of emerging fields such as nano-geo-technology, biotechnology, chemical reaction engineering, dynamics of complex systems, reactive transport in multiphase systems, process intensification and integration. Last, but not least, the future of geo-technologies critically depends on a rational, objective socio-economic and environmental analysis of their strengths and weaknesses. The presentation, prepared by a voluntary team of authors is the first public presentation of the initiative to establish a global scientific network focused on excellence and effective exchange of knowledge of geo-sciences and geo-engineering that is presently dispersed across various fields of geo-technologies. An up-to-date curriculum on clean technologies for the production of energy and materials in situ for the coming generation of mining and energy professionals will be the key educational target of the network. Financial and managerial support of the network will be sought from the People programme of 7th EU Framework Programme for RTD and in the context of collaboration agreements with countries such as US, China, Japan, India, Australia, South Africa.
Converting brown coal into chemicals and hydrogen by steam cracking and gasification Nozomu Sonoyamaa,∗, Kazunari Nobutab , Tokuji Kimurab , Sou Hosokaic , Jun-ichiro Hayashid , Teruoki Tagob , Takao Masudab Coal & Environment Research Laboratory, Coal Business Office, Petroleum & Coal Marketing Department, Idemitsu Kosan Co., Ltd., 3-1 Nakasode, Sodegaura, Chiba 299-0267 b Division of Chemical Process Engineering, Graduate School of Engineering, Hokkaido University, N13 W18 Kita-ku, Sapporo, Hokkaido 060-8628 c Center for Advanced Research of Energy Conversion Materials, Hokkaido University, N13 W8 Kita-ku, Sapporo, Hokkaido 060-8628 d Division of Advanced Device Material, Institute for Materials Chemistry and Engineering, Kyushu University, 6-1, Kasuga Koen, Kasuga, Fukuoka 816-8580 a
1. Introduction Chemical products are derived primarily from petroleum, while coal and natural gas remain minor feedstocks. Recently, projects have been planned for plants producing chemical products from coal and natural gas. When chemical products are made from a different feedstock, we can select a lower-priced feedstock in the area where a plant is located, resulting in a reduction in the cost of the chemical products. Low-rank coals, such as brown coal and lignite, are not traded extensively on the world market, compared with higher-rank coal, because of their low calorific values, high moisture and oxygen contents, and propensity for spontaneous ignition. In the current study, we focused on the low-rank coals as a feedstock for chemical production. Many studies have been conducted for syngas production by gasification. Coals are decomposed into char and volatiles by pyrolysis. The tar inhibits char gasification, because it prevents steam and/or oxide from contacting the char and promotes the volatilisation of catalytic metals on coal by the reduction of the catalytic metals. We consider that the production of syngas by gasification and catalytic cracking are favourable for pyrolytic char and tar, respectively. An investigation of the integration of the process of char gasification and tar cracking for the chemical production of Loy Yang coal was performed. The problem with developing our process was that the pyrolytic char requires a trace of tar and this pyrolytic tar increases the atomic H/C ratio. In a previous study, we investigated the cracking of the pyrolytic oil of Loy Yang coal with iron oxide catalysts under a steam atmosphere, and the generation of hydrogen on the iron oxide catalysts from steam was indicated [1]. In the present study, we propose a process for converting Loy Yang coal into hydrogen and chemicals such as mono-aromatic hydrocarbons and oxygenated compounds, using a steam as hydrogen source. The potential and the optimum conditions of the process were investigated on ∗
Corresponding author: Tel: +81 438 62 9511; Fax: +81 438 60 1177 Email address: [email protected] (Nozomu Sonoyama )
the basis of experimental results for the pyrolysis of Loy Yang coal and cracking of the pyrolytic tar of Loy Yang coal. 2. Experimental 2.1. Catalyst preparation and characterization Three different iron-based catalysts were prepared by co-precipitation from aqueous solutions of nitrates of iron, aluminium, zirconium, and cerium: FeOx, Zr-Al-FeOx, and Ce-Zr-Al-FeOx. The catalysts were palletised with no binders, crushed, and sieved to yield particles of 300-850 µm in diameter. The structures of the catalysts were analysed using powder X-ray diffractometry. 2.2. Pyrolysis of Loy Yang coal A Victorian brown coal, for which the analytical data is shown in Table 1, from the Latrobe Valley, Victoria, Australia, was pyrolysed in a continuous reactor at a peak temperature in the range of 723-873 K with a mean particle residence time of 47 s. The volatile products contained non-condensable gases, water, water-soluble organics, oil, and pitch. The water-soluble organics, oil and pitch were referred to collectively as tar. The pitch was separated using a silica fibre filter located between the reactor and water/tar condenser, and heated at 423 K. Table 1: Analysis of Loy Yang coal and the char produced by the pyrolysis of Loy Yang coal at 873 K.
Moisture Ash [wt.%-ar] [wt.%-db] Loy Yang coal Char
58.4 3.6
3.3 6.9
C 68.5 82.0
H N S [wt.%-daf] 5.3 3.4
0.5 0.9
0.3 0.1
O
Heating value [MJ/kg-daf]
25.3 13.6
27.16 30.44
2.3. Catalytic cracking with iron oxide catalysts A benzene solution containing 10 wt.% oil was used as the feedstock for catalytic cracking. The inertness of benzene has been confirmed previously [2]. Catalytic cracking of the benzene solution under a steam atmosphere was performed at 773 K and ambient atmospheric pressure in a fixed-bed reactor that was loaded with about 1 g of one of the iron oxide catalysts. The feeding rates of the feedstock and steam were 1 and 1.5 g/h, respectively. The feedstock was supplied to the fixed-bed reactor through a 1/16-inch stainless steel tube by a syringe and heated at 473 K around an inlet of the fixed-bed reactor. A mixture of steam and nitrogen with partial pressures of 87 and 14 kPa, respectively, was introduced into the fixed-bed reactor. A train consisting of a condenser and a gas bag was installed downstream. The liquid products were collected via the condenser and analysed by gas chromatography. Non-condensable gases were allowed to pass through the condenser, collected in the gas bag, and analysed by gas chromatography.
100
Yield [ wt.%-db ]
Gas H2O
80
Tar
60 40
Char
20 0
723
823
873
Pyrolysis Temperature [ K ] Figure 1: Yield of pyrolytic products of Loy Yang coal.
3. Results and Discussion 3.1. Determination of pyrolysis conditions We experimentally determined the pyrolysis conditions for the integration of char gasification and tar cracking. Figure 1 illustrates the product distributions for the pyrolysis of Loy Yang coal. The production of volatiles (gas, H2 O, tar) was completed at a temperature between 823 and 873 K. The maximum yield of the tar occurred at 873 K. When the pyrolysis temperature exceeds a given temperature, the tar is converted into a gas. Accordingly, the yields of chemicals or the precursor compounds of the chemicals in tar may decrease with an increase in pyrolysis temperature beyond 873 K. The increase of the yields of phenols and ketones as chemical products requires the suppression of the decomposition of the oxygenated function, such as phenolic, ketone, and carboxylic groups, during pyrolysis. A small amount of phenolic function appeared to be decomposed even at 873 K. H2 O was generated at about 12 wt.% by the pyrolysis of Loy Yang coal. We expected that the generation of H2 O would be used for cracking with the iron oxide catalysts in steam. Figure 2 presents the yields of various components in tar. Pitch accounted for 80% of the tar at 723 K, and about 60% above 823 K. Obtaining both char containing traces of volatiles and the maximum yield of tar was important. The temperatures that were advantageous for gasification and catalytic cracking ranged from 823 to 873 K. We performed catalytic cracking of the oil produced at 873 K. 3.2. Investigation of catalytic cracking with iron oxide catalysts We investigated the potential and performance of iron oxide catalysts for the cracking of tar. Figure 3 illustrates the distillation curves of the oil and oil products after cracking with Zr-Al-FeOx for 2 h. The oil became lighter by cracking. Many components contained in the oil product had a
Yield [ wt.%-db ]
15 Others WS Oil Pitch
10
5
0
723
823
873
Pyrolysis Temperature [ K ] Figure 2: Yield of components containing the tar of Loy Yang coal (WS: water-soluble organics).
boiling point below 473 K. The cause of the decrease in heavier tar was not clear quantitatively, whether it was due to deposition on the catalyst or the decomposition of the heavier tar by the catalyst. Figure 4 shows the cumulative yields of mono-aromatic hydrocarbons, as demonstrated in our previous study [1]. FeOx yielded a greater amounts of xylenes, compared with the oil, but no cresol was detected in the product after cracking with FeOx. Product gases for FeOx were almost all CO2 and the structure of FeOx changed from a hematite structure to a magnetite structure. Accordingly, tar was predominantly oxidized by FeOx, so FeOx has too high an activity to degrade the oil for chemical production. Although Zr-Al-FeOx yielded a decrease in aromatic compounds, the product gases contained hydrogen and the structure of the Zr-Al-FeOx was maintained in the hematite structure. Compared with the other catalysts, Ce-Zr-Al-FeOx yielded more aromatic compounds. The iron oxide catalysts added to Ce and Zr were evaluated with promising results for chemical production by the cracking of tar with a steam atmosphere. We have proposed a mechanism for cracking with iron oxide catalysts on the basis of our previous study (Figure 5) [1]. Cracking with the iron oxide catalysts used the redox cycle, Fe2 O3 ↔ Fe2 O3−a + a2 O2 (0 < a < 1/3), for the catalytic event. Ce and/or Zr had the ability to decompose H2 O, generating active hydrogen and oxygen species. The active oxygen species were the agent of the oxidation of Fe2 O3−a and the active hydrogen species the agent of the hydrogenation of the oil. H2 O as a hydrogen source was important for the catalytic cracking with iron oxide catalysts containing Ce and/or Zr. Although the catalytic cracking of pitch was not investigated, we believe that the pitch can be decomposed by iron oxide catalysts with further development of the catalyst.
100
Fraction [ mol-C% ]
98 80
Oil product after cracking
96 60 Pyorlytic oil
94 40 92 20
熱分解油 接触分解生成物
0 90
0 273
200 400 473 673 Boiling Point Boiling Point[ ℃ [ K] ]
600 873
Figure 3: Distillation curves of the oil produced by the pyrolysis of Loy Yang coal and the product oil after cracking the oil with Zr-Al-FeOx for 2 h.
3.3. Evaluation of an integration process using steam as a hydrogen carrier Figure 6 shows a process that integrates the steam gasification of char and the steam cracking of tar using H2 O as a hydrogen carrier. On the basis of the above investigation, the pyrolysis and cracking temperatures of the process were set at 873 and 773 K, respectively. The enthalpy and mass balances of the process were calculated with changing char conversion for gasifire held at 1073 K. Steam for catalytic cracking was supplied from a pyrolyser that also dried the feed coal. Steam was set at 7.8 mol-H2 O to the atomic carbon of the tar. This value of steam/tar corresponded to the ratio of the total amounts of the moisture of Loy Yang coal and H2 O generated by the pyrolysis of Loy Yang coal to the tar. The consumption rate of H2 O for the catalytic cracker was assumed to be 20% of the input amount of steam in the catalytic cracker, although we obtained no data for the consumption of H2 O by cracking. The variation of enthalpy balance with char conversion is presented in Figure 7. The reaction heat values of char gasification and combustion were set at 131.3 and 434.6 kJ-LHV/mol-C, respectively. The heating value of H2 O from 298 to 1073 K was set at 70.3 kJ/mol-H2 O. The reaction heat value of pyrolysis and cracking was negligible. On the basis of the ultimate analysis of char (Table 1), the char gasification was described as C8 H4 O+7H2 O→8CO+9H2 O. The cold gas efficiency, η, of the gasifire was defined according to the following equation: η=
nCO LHVCO + nH2 LHVH2 × 100 nChar LHVChar
(1)
where ni represents the product or feed amounts of substance i, and LHVi is the lower heating value of substance i. The autothermal condition of the process is an enthalpy balance of 0%. Char conversion for the autothermal condition was about 50 mol-C% corresponding to a cold gas efficiency of about 65%.
Cresols Phenol Xylenes Toluene
Pyrolytic oil
FeOx
Zr-Al-FeOx
Ce-Zr-Al-FeOx 0
10
20
Fraction [ mol-C% on pyrolytic oil basis ] Figure 4: Yield of mono-aromatic compounds after cracking for 2 h.
Fe2O3 H* O*
Oil
H2O Fe2O3-Y 0 < Y < 1 3
CO2 + Oil-fragments
Figure 5: Catalytic reaction of oil decomposition on iron oxide catalysts under a steam atmosphere.
4. Conclusions We proposed a process for converting Loy Yang coal into hydrogen and chemicals such as mono-aromatic hydrocarbons and oxygenated compounds using steam as a hydrogen source. On the basis of the experimental results for the pyrolysis of Loy Yang coal and the cracking of tar with iron oxide catalysts under a steam atmosphere, we investigated the potential and the optimal conditions for the integration of the gasification and cracking processes. The autothermal condition for the process was determined by calculating the enthalpy and mass balances of the process. The energy efficiency of the process was affected not only by the amount of steam, but also by the rate of char conversion. The cold gas efficiency of the gasifire for the autothermal condition of the process was about 65%. Acknowledgments This work was funded in part by the Japan Coal Energy Center (JCOAL).
Exhaust gas Catalytic cracker Chemicals (BTX, Phenols, etc.) 773 K Loy Yang coal Dry 100mol-C Moisture 58wt%
Pyrolyzer 873 K
Steam
Volatiles Tar 17.9mol-C + Gas 8.1mol-C + H2O 150mol-H2O
Syngas (CO, H2) Steam Exhaust gas
Combustor
Gasifier
Char (67.5mol-C)
Unreacted Char
1073 K
Ash
Steam Steam 2 mol-H2O/mol-CinChar
Air
Figure 6: Schematic diagram of the production process of chemicals and hydrogen derived from Loy Yang coal.
Enthalpy Balance [ %-Feed LHV Basis ]
Cold Gas Efficiency of Gasifier [ % ] 50
0
32
65
97
129
25
50
75
100
25
0
-25
-50
0
Char Conversion [ mol-C% ]
Figure 7: Variation of the enthalpy balance of the proposed process with char conversion.
References [1] N. Sonoyama, K. Nobuta, T. Kimura, S. Hosokai, J. Hayashi, T. Tago, T. Masuda, Production of chemicals by cracking pyrolytic tar from Loy Yang coal over iron oxide catalysts in a steam atmosphere, International Conference on Coal Science & Technology, 2009. [2] E. Fumoto, T. Tago, T. Tsuji, T. Masuda, Recovery of useful hydrocarbons from petroleum, Energy Fuels 18 (2004), pp. 1770-1774.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COAL-DERIVED PRODUCTS PILOT SCALE PRODUCTION OF HUMIC SUBSTANCES FROM TURKISH LEONARDITES Bekir Zühtü Uysal, Department of Chem. Eng., Faculty of Engineering and Clean Energy Research and Application Center, Gazi University, Ankara, Turkey [email protected] Ufuk Gündüz Zafer, Department of Chem. Eng., Faculty of Engineering, Gazi University, Ankara, Turkey Ö. Murat Dogan, Department of Chem. Eng., Faculty of Engineering, Gazi University, Ankara, Turkey Duygu Öztan, Department of Chem. Eng., Faculty of Engineering, Gazi University, Ankara, Turkey Zeki Olgun, Turkish Coal Enterprises (TKI), Ankara, Turkey Mustafa Ozdingis, Turkish Coal Enterprises (TKI), Ankara, Turkey Selahaddin Anac, Turkish Coal Enterprises (TKI), Ankara, Turkey
Abstract: A large number of organic humic substances are increasingly being applied worldwide especially in agricultural applications. Turkish Coal Enterprises (TKI) has been producing humic acid and other humic containing substances from Turkish lignites and leonardites since 2008. A number of methods for the extraction of humic substances using sodium hydroxide solution in laboratory scale have been published. Potassium hydroxide and different acids (hydrochloric acid, sulfuric acid, nitric acid and phosphoric acid) were used for the production of humic acid in our laboratories before pilot plant was designed and pilot scale production was started. Humic substances were first extracted to the alkaline medium at 85˚C for 6 hours and then these extracted substances were precipitated with the acid. Experimental results showed that phosphoric acid resulted in the highest extraction efficiency of humic acid. The extraction efficiencies of Göynük leonardite were found as 61.4, 66.9, 79.0 and 91.2 % for studies with hydrochloric acid, sulfuric acid, nitric acid and phosphoric acid, respectively.
Heavy Metal Removal by Using Chemically Crosslinked Turkish Coal Based Humic Acid Tulay Inan1, Hacer Dogan1, Murat Koral1, Selahattin Anaç2, Zeki Olgun2 1
TUBITAK Marmara Research Center, Chemistry Institute, 41470 Gebze, Kocaeli TURKEY 2 TKI(Turkish Coal Enterprises)Hipodrom Cad. No:12 Yenimahalle 06330 Ankara
Abstract: Chemically crosslinked humic acid was synthesized by using epichlorohydrin as cross-linking reagent. Humic acid was obtained from TKI (Turkish Coal Enterprises). The reaction was optimized and the products were analysed by the Fourier transform infrared (FTIR) spectroscopy, thermal gravimetric analyser (TGA) and scanning electron microscopy (SEM). The efficiency of these samples as adsorbent has been studied as a function of amount, contact time and initial heavy metal concentration by a series of batch experiments. The study showed that crosslinked humic acid could be used as an effective sorbent for the removal of heavy metals.
1. Introduction Humic acids (HAs) are natural organic macromolecules, arise from the decay of plant and animal residues mainly through biochemistry and geochemistry, and widely exist in peat, lignite, stream river, wetland, lake and ground waters. HAs, which contain ~50-90% in total organic compounds, ~70-80% in soil, ~40-60% in organic carbon dissolved in water, are the main organic carbon dissolved in water, are the main organic carbon on the surface of the earth. The structures of HAs contain a carboxyl group (-COOH), an amino group (-NH2), a quinonyl group (C6H3O2), all of which can form functional HAs, by chemical methods. HAs and functional humic acids were mainly used in agriculture, forestry, biopharmaceutical, petroleum additive industry and so on [1-5]. Metal humates can be prepared in a relatively simple way by the precipitation of HAs with suitable metal compounds. Recent years, low cost and highly effective adsorban materials have been developed for heavy metal adsorption. Some of the reported low-cost sorbents include bark/tannin-rich materials, lignin, peat moss, iron-oxide-coated sand, leaf mould, seaweed/algae/alginate, dead biomass etc. Humic acid based materials are also shown as one of the best examples for low cost sorbents [6-7]. Humic acid extracted from Turkish coal in Ilgın Konya by TKI was used without any purification in the cross-linking humic acids (CL-HAs) and the reaction was optimized. The efficiency of these samples as adsorbent of the heavy metal mixture containing As, Pb, Cd and Ni in the amount of each 10 ppm and heavy metal mixture containing As, Cd, Co, Cr, Cu, Hg, Ni and Pb in the amount of each 50 ppm has been studied as
a function of amount, contact time and initial heavy metal concentration by a series of batch experiments. 2. Experimental 2.1. Equipments and reagents HA obtained from TKI. Epichlorohydrine (ECH), NaOH, and other reagents were obtained from commercial sources and used as received. The IR spectra were performed on a Perkin Elmer Spectrum One FTIR Spectrometer. Thermogravimetric data were obtained between 30 and 900 °C, under nitrogen, with a heating rate of 10 °C/min using a Thermogravimetric Analyzer Perkin Elmer Pyris. For heavy metal analysis Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) was used. 2.2. Humic acid characterization Humic acid sent by TKI was analyzed and the results are given in Table 1. Carboxyl groups, total acidity and the coefficient E4/E6 were then determined. The amount of carboxylic groups and the total acidity were determined using the calcium acetate and the barium hydroxide methods, respectively. The quantity E4/E6 is the ratio of absorbances at 465 and 665nm of humic acid solutions. Elemental analyses of carbon, hydrogen, nitrogen and sulfur were obtained with a Thermo Finnigan Flash 1112 Series EA elemental analysis instrument. E4/E6 ration was found as 3,26 of humic acid obtained from coal. This value shows that humic acid has high aromatic groups. 2.3. Adsorbent preparation Ten different procedures [1,8,9] were applied to optimize the process of CL-HAs as given below with the used amount of reagents as given in Table 2. HA, ECH and NaOH granules or NaOH solutions (5-20% in water) were used and the obtained mixtures were stirred at desired temperatures for desired times, resulting suspensions of CL-HAs were separated by centrifugation and rinsed with distilled water. Finally the solid complexes were dried at desired temperatures. The solubility of adsorbent was studied as a function of pH. HA-EP-1 [9] HA converted to Na-Humate by the addition of NaOH and obtaining a solution wih pH=12. EPC added dropwise to the system. Resulting suspensions of CL-HAs were separated by centrifugation and rinsed with distilled water. Finally the solid complexes were dried at 70 -105 °C HA-EP-2 [8] HA converted to Na-Humate by the addition of NaOH solution. EPC added dropwise to the system. The mixture was stirred and heated to 50 °C for 5 h. Resulting suspensions of CL-HAs were separated by centrifugation and rinsed with distilled water. Finally the solid complexes were dried at 70 -105 °C
HA-EP-3 Same procedure with HA-EP-2 was applied but the amount of 20% NaOH solution decreased from 6,01 gr to 0,33 gr by taking theoric hydroxyl value. HA-EP-4 Same procedure with HA-EP-2 was applied but the amount of 20% NaOH solution increased from 6,01 gr to 15 gr. HA-EP-5 HA converted to Na-Humate by the addition of NaOH solution. EPC added dropwise to the system. The mixture was stirred and heated to 50 °C for 1 h. Then mixture was stirred at 90-100 °C for 5 h. Resulting suspensions of CL-HAs were separated by centrifugation and rinsed with distilled water. Finally the solid complexes were dried at 70 -105 °C HA-EP-6 Same procedure with HA-EP-2 was applied but the reaction temperature was selected to be room temperature. HA-EP-7 Same procedure with HA-EP-2 was applied but the reaction temperature was selected to be 50 °C. HA-EP-8 HA and EPC added to the reaction vessel and stirred and heated to converted to NaHumate by the addition of NaOH solution. EPC added dropwise to the system. The mixture was stirred and heated to 50-70 °C’de for 1 h. Then NaOH granules added to the reaction medium mixture was stirred at 90-100 °C for 5 h. Resulting suspensions of CL-HAs were separated by centrifugation and rinsed with distilled water. Finally the solid complexes were dried at 70 -105 °C HA-EP-9 Same procedure with HA-EP-2 was applied but the reaction temperature was selected to be 100 °C. HA-EP-10 HA, EPC and NaOH all added to the reaction vessel. The mixture was stirred and heated to 50 °C for 24 h. Resulting suspensions of CL-HAs were separated by centrifugation and rinsed with distilled water. Finally the solid complexes were dried at 70 -105 °C
Table 1. Properties of humic acid TKI Humic Reference Acid Humic Acid* Content (Wt. %) Moisture Ash (at 950°C) Organic matter Inorganic matter Humic acid Hexane extraction residue Methanol extraction residue Petroleum ether extraction residue Elemental Analysis (wt. %) C H N S O Elemental ratio N/C H/C O/C Functional Groups (meq/g) COOH Phe-OH Total acidity E4/E6
11.8 38.2 56.6 43.4 31.3 0.12 3.63 0.018 55.4 6.4 1.4 0.8 36
54.72 4.04 1.47 0.36 38.54
0.022 1.38 0.49
0.023 0.88 0.53
2.16 6.84 9.0 3,26
5.0 1.16 6.16
* International Humic Substances- IHSS- (http://ihss.gatech.edu/ihss2/index.html) Pahokee Peat
Table 2. Amounts of reagents used in the preparation of CL-HAs HA†/EPC HA†/NaOH Ratio TKI NaOH (gr) Ratio HA† (gr) HA-EP-1 10/1 (wt/wt) 6/1(wt/wt) 36,06 6,01 (granules) HA-EP-2 1/10 (wt/wt) 1/3 (wt/wt) 2 6,01 (%20 solution) HA-EP-3 1/10 (wt/wt) 6/1 (wt/wt) 2 0.33 (granules) HA-EP-4 1/10 (wt/wt) 1/3 (wt/wt) 2 15 (%20 solution) HA-EP-5 1/40 (mol/mol) 6/1 (wt/wt) 2 0,34 (%5 solution) 10ml HA-EP-6 1/10 (wt/wt) 1/3 (wt/wt) 2 6,01 (%20 solution) HA-EP-7 1/10 (wt/wt) 1/3 (wt/wt) 2 6,01 (%20 solution) HA-EP-8 1/10 (wt/wt) 1/3 (wt/wt) 2 0,34 (granules) HA-EP-9 1/10 (wt/wt) 1/3 (wt/wt) 2 6,01 (%20 solution) HA-EP-10 1/10 (wt/wt) 1/3 (wt/wt) 2 6,01 (%20 solution) †HA contains inorganics and moisture in its structure (TKI HA contains %31 pure HA), EPC purity is %57 *
ECH* (gr) 2,1 11,4 11,4 11,4 10 11,4 11,4 10 11,4 11,4
2.4. Batch adsorption experiments Batch adsorption experiments were carried out by shaking 1 g of the adsorbent with 25 ml of solution of heavy metals with desired concentration in PE bottles. The metal solution together with the adsorbent was agitated magnetically in the closed PE bottle for different mixing times. Then the solid phase was separated by centrifugation, the concentration of the metal in the solution was determined immediately by ICP-AES. The adsorption capacity was calculated using the following equations:
q(mg / g ) =
(Ci − Ct ) V m
where Ci and Ct are the concentrations of the metal ion in initial and final solutions respectively. V is the volume of the aqueous phase (mL), m is the weight of Fe-HA (g). The adsorption degree, AD, as a function of time was also determined from the experimental data using the following relationship: AD % = 1 − C(t)/C(i) *100
3. Results and Discussion 3.1 FTIR Spectrums
Overlayed FTIR spectrums of pristine HA and CL-HAs are given in Fig 1. (a) and (b). The broad and strong band ranging from 3200 to 3600 cm-1 indicated the presence of –OH and NH2 groups, which is consistent with the peak 1033 and 1100 cm-1 assigned to alcoholic C-O and C-N stretching vibration. The peaks at 2925 and 2870 cm-1 can be assigned to asymmetric and symmetric –CH2 groups. The stong absorbtion peak at 1722 cm-1 is the characteristic of stretching main peak indicating the epoxy HA.
752.24 695.61
1402.93
3696.66
912.55
470.82 538.27
1621.72
3409.24
%T
1033.01
750.77 695.15
1453.82
1268.13
470.64
3696.92 2925.73
912.45
538.40
1722.13 1627.16
3620.22 3430.73
4000.0
3600
3200
2800
2400
2000
1800 cm-1
1600
1400
1200
1000
800
600
450.0
(a) HA 752.24 695.61 1402.93
3696.66
912.55
470.82 538.27
1621.72 3409.24
1033.01
HA -EP4
%T HA -EP5 750.77 695.15 1453.82 3696.92
2925.73
3430.73
3600
912.45
2800
2400
2000
538.40
1034.14
HA -EP7
3200
470.64
1722.13 1627.16
3620.22
4000.0
1268.13
1800 cm-1
1600
1400
1200
1000
800
600
450.0
(b) Figure 1. FTIR spectra of HA/ECH based CL-HAs
CL-HAs also post-croslinked by using an amine compound of 3-aminophenyl sulfone. The FTIR of the obtained products are given in Figure 2 (a) and (b). Overlayed FTIR spectrums of pristine HA and amine crosslinked -CL-HAs are given in Fig 2. (a) and (b). For the product of HA-EP5 after crosslinked with amine is given in Figure 2a and there are additional peaks coming from the crosslinking agent of amine compound in the FTIR spectrums. The spectrum of HA-EP5 is the combination of the amine compound and HA prepolymer and also as could be seen the main peak of C-O-C at 1722 cm-1 disappeared showing the post crosslinking reaction with amine compound.
752.24 695.61
HA
3696.66
1402.93
912.55
538.27 470.82
1621.72
3409.24
1033.01
HA -EP5
750.77 695.15 1453.82
3696.92
2925.73
1722.13 1627.16
3620.22
%T
1268.13
470.64 538.40
912.45
3430.73
1034.14
3-Aminophenylsulphone
3458.85 3390.03 3695.93 3620.05
4000.0
3600
HA -EP5 (Aminle pisirme sonrasi)
1623.19 1600.15 1455.46 1484.71
3200
2800
2400
2000
1800 cm-1
1286.71 1318.77
1089.99 1147.24
783.94 687.28 470.46 615.61 528.24
910.79
716.63
1600
1400
1200
1000
800
600
450.0
1600
1400
1200
1000
800
600
450.0
(a)
HA
HA-EP4
%T
HA -EP3
4000.0
3600
3200
2800
2400
2000
1800 cm-1
(b)
Figure 2. FTIR spectra of HA/ECH based CL-HAs crosslinked with amine compound of 3-aminophenyl sulfone.
3.2. Thermal Characterization
CL-HAs and amine croslinked CL-HAs thermograms are given in Figure 3 (a) and (b). The weight loss of the CL-HAs decreased after the reaction of epichlorohydrine but crosslinking with amine compound increased thermooxidative stability of the end products enormously. There are no decomposition up to 300 °C for the amine crosslinked products as seen in Fig3 (a) and (b).
(a)
(b) Figure 3. TGA Thernograms for CL-HAs and amine crosslinked-CL-HAs.
3.3. Solubility Tests Solubility tests performed using (0,3 g HA-EP and CL-HAs)/ 5gr water with the medium having pH=10. Amin crosslinked CL-HAs products are less soluble than HAEP products as given in Table 3.
3.4. Batch Adsorbtion Tests All adsorbtion tests were conducted in duplicates and results are reported as average. In general the studies on optimum pH, agitation rate and contact time were performed at room temperature (25±1 °C) in 50 ml vials for 24 hours exposure by stirring including 10 ppm of Pb, Cd, Ni, As. The adsorbed heavy metal content were determined with the help of ICP-AES. These tests were also repeated with the solutions including 50 ppm of heavy metals after 2 hours and 24 hours exposure for the sample coded HA-EP7 which gave the best results of all the adsorbents as given in Table 3.
Table.3. CL-HA Based Adsorbants: Their Solubility and Adsorbtion Values
D. No HA HA-EP1 HA-EP1 (amine crosslinked) HA-EP2 HA-EP2 (amine crosslinked) HA-EP3 HA-EP3 (amine crosslinked) HA-EP4 HA-EP4 (amine crosslinked) HA-EP5 HA-EP6 HA-EP7 HA-EP8
Solubility 0,3gr/5grof water(pH=12)
Adsorbtion Tests (10 ppm of mother solution) Pb <0,5
Cd 5,2
Ni 3,1
As 7,8
7,8
<0,5
10,0
9,4
12,7
0
<0,5 <0,5 <0,5
0,5 0,7 <0,5 <0,5 6,7 4,6
3,6 1,9 3,4
7,3
23,3 0 5,3 24,4
2 hr results HA-EP7
As 13,4
24 hr results HA-EP7
As 8,5
Adsorbtion Tests (50 ppm mother solution) Cd Co Cr Cu Hg Ni Pb 2 3,6 2,6 14,5 36,5 2,7 33,4 Adsorbtion Tests (50 ppm mother solution) Cd Co Cr Cu Hg Ni Pb 0,0 0,0 1,7 14,2 40,6 0,1 32,4
4. Conclusion We synthesized different HA/epoxide based adsorbents (CL-HAs). These adsorbants (CL-HAs) were also post-crosslinked with the help of amine compound of 3aminophenyl sulfone. The best adsorbant results were obtained for the sample coded HA-EP7 which was preapared at 50 °C. Crosslinking with amine compound increased thermo-oxidative stability of the adsorbants but decreased heavy metal adsorbance characteristics of the adsorbants.
References 1. Qijie Xu, Lei Zhang, Wenzhong Shi, Yuanchen Cui, “ catalytic performance of cross-linking humic acids supported Pd/Ni bimetallic catalyst for heck reaction”, Polish Journal of Chemical Technology, 11, 3, 22-26, 2009 2. F.J. Stevenson, Humus Chemistry. Genesis, Composition, Reactions (2nd Edition ed.), Wiley, New York (1994). 3. J. Novák, J. Kozler, P. Jano , J. Cezíková, V.Tokarová and L. Madronová, Humic acids from coals of the North-Bohemian coal field: I. Preparation and characterisation Reactive and Functional Polymers. 47 (2001) 101-109. 4. S. Erdogan, A. Baysal, O. Akba, C. Hamamci, Interaction of Metals with Humic Acid Isolated from Oxidized Coal, Polish J. of Environ. Stud. 16 (2007) 671675. 5. Humates and Humic Acids. How do they work?http://www.simplicitea.com/humates_how_they_work.doc 6. P. Janos, J. Fedorovic; P. Stanková; S. Grötschelová; J. Rejnek; P. Stopka, Iron Humate as a Low-Cost Sorbent for Metal Ions, Environmental Technology, 27 (2006)169-181. 7. P. Janoš, Iron humate as a multifunctional sorbent for inorganic and organic pollutants, Environ. Chem. 2 (2005) 31-34. 8. T. Malathan, R. Nicu and Valentin I. Popa“ Lignin modification by epoxidation”, Bioresources 3(4), 1371-1376 9. US 3 984 363
Analysis of the Effect of Internal Defect on Coke Fracture Behavior by Rigid Bodies-Spring Model Kenichi HIRAKI*, Hideyuki HAYASHIZAKI*, Yoshiaki YAMAZAKI*, Tetsuya KANAI*, Xiaoqing ZHANG*, Masakazu SHOJI*, Hideyuki AOKI* and Takatoshi MIURA* *Department of Chemical Engineering, Graduate School of Engineering, Tohoku University, 6-6-07 Aoba, Aramaki, Aoba-ku, Sendai, 980-8579, Japan Abstract In this study, to investigate the effect of internal defect in coke, fracture analyses using RBSM (Rigid Bodies-Spring Model) which is applied to fracture analysis of brittle materials such as concrete are carried out for coke. First, in case of taking no account of internal defect, the fracture analyses with RBSM assuming 4-point bending tests are carried out. As a result, analytical results reproduce the fracture behavior well. Next, it is known that internal defect such as non-adhesion region boundaries in coke influences coke strength, so in case of taking account of non-adhesion region boundaries, the fracture analyses with RBSM assuming 4-point bending tests are carried out. To investigate the location of non-adhesion region boundaries in coke, 3 percent non-adhesion region boundaries are randomly located in the coke and fracture analysis of two cases are carried out. Analytical results show that fracture load decreases when non-adhesion boundaries are located under region of high stress and microcracks happen at around non-adhesion region boundaries. It is supposed that these microcracks cause surface breakage. Analytical results also show that the fracture starts at non-adhesion region boundaries when non-adhesion region boundaries are located under region of high stress. In particular, fracture progresses at the location of continuous non-adhesion region boundaries. To investigate the fraction of non-adhesion region boundaries in coke, 3, 5 and 10 percent non-adhesion region boundaries are randomly located in coke. As a result, fracture loads decrease with an increase of non-adhesion region boundaries in the coke. 1. Introduction In the blast furnace operation, coke has four roles, which are heat source, reducing, carburizing and sustaining the flow passes. Above all, sustaining the flow passes of liquid metal and high temperature reducing gas in blast furnaces is the most important for stable operation of blast furnaces. Coke with high strength is required to satisfy this role. In previous studies of coke fracture1,2), internal defects such as non-adhesion region boundaries and connected pores influence coke fracture. Coke strength is generally estimated by drum strength tests of coke such as drum test and tumbler test. However, these tests do not indicate strength factors since they cannot estimate quantitatively the fracture behavior. The effect of internal defects such as non-adhesion region boundaries and connected pores in coke is not estimated enough. Above all, there are few quantitative data on non-adhesion region boundaries since it is difficult to observe non-adhesion region boundaries by microscope. Fracture behavior of material is generally complicated since various fracture morphologies such as delamination, sliding and crack propagation coexist. In particular, since coke is porous material and internal defects such as pore, crack and non-adhesion region boundaries in coke interfere with each other, fracture behavior of coke is very complicated. So, it is difficult to apply analytical method such as finite element method based on continuum mechanics3-6) to discontinuity behavior such as crack formation and propagation. Rigid bodies-spring model (RBSM) recently developed by Kawai7) is one of discrete approaches. Analytical object is divided into triangle elements interconnected along their boundaries by springs. Each element has two translational and one rotational degrees of freedom defined at the element center of gravity. The interface between two elements consists of two springs. Normal and shear springs are placed at the boundary of element. Energy which is given by surface force between elements instead of the internal work of each element is estimated in RBSM. This model is applied to concrete8) and brittle polycrystalline solid9) which behave as a discontinuity when fracture happens and progresses. In this study, to investigate the effect of non-adhesion region boundaries in coke, fracture analyses using RBSM are carried out for coke. First, in case of taking no account of non-adhesion region boundaries, the fracture analyses with RBSM assuming 4-point bending tests are carried out. Next, in case of taking account of non-adhesion region boundaries, the fracture analyses with RBSM assuming 4-point bending tests are carried out to investigate the location and the fraction of non-adhesion region boundaries in coke. 2. Rigid Bodies-Spring Model In this study RBSM is applied to coke. Fig. 1 shows element position between before and after deformation. Analytical object is divided into elements interconnected along their boundaries by springs. The interface between two elements consists of two springs. Normal and shear springs which resist relative displacement are placed at the boundary of element. Energy stored at the interface between two elements is estimated. Deformation of material is represented by translation and rotation of element. In RBSM, the relationship between relative displacement = (n, s)t
Fig. 1 Element position between before and after deformation and surface force per unit area = (n, s)t is shown as:
ζ D δ ,
(1)
where D is spring constant matrix. In elastic area, D is shown as:
kn D 0
0 , k s
(2)
where kn and ks are the elastic modulus of normal and shear spring, respectively. The following spring constants are obtained from elastic modulus E, Poisson's ratio and length of perpendicular h in plane stress as:
kn
E
(1 2 )h E ks . (1 )h
, (3)
Global stiffness matrix is constructed by these equations and is based on principal of virtual work. In RBSM, when fracture happens at the interface between two elements, embrittlement of material is represented by varying spring constant matrix D of the interface. Spring constant matrix in plastic state is shown as:
d11 d12 . D( p ) sym. d 22
(4)
When shear sliding happens, the following equation of stress increment is obtained from flow rule based on plasticity theory:
f f D(e ) t D(e) ζ ζ ε . ζ D(e) f ( e ) f D ζ ζ t
(5)
The following spring constant matrix elements after plastic deformation are obtained from substituting yield function based on Mohr-Coulomb condition into Eq. (5)10):
d11 k n
k n2 c n tan tan 2
k s s2 k n c n tan tan 2
,
(6)
Table 1 Material properties Tensile strength Elastic modulus Poisson's ratio Compressive strength Internal friction angle Shear strength
[MPa] [GPa] {-] [MPa] [°] [MPa]
5.4 4.0 0.2 61.6 32.0 9.4
Fig. 2 Analytical object
s Compression Fc fracture
0.5Fc
c
Shear fracture Ft
n
Tensile fracture
Fig. 3 Fracture criterion d12
k n k s s c n tan tan
k s s2 k n c n tan tan 2
d 22 k s
k s2 s2
,
k s s2 k n c n tan tan 2
(7)
.
(8)
When shear strain reaches constant value after shear sliding happens, it is regard as shear fracture and the following spring constant matrix is applied as: 0.25k n D( p ) 0
0 . 0
(9)
When crack happens by tensile stress, normal spring constant is set to zero and shear spring constant is set to 12.5 % of initial shear spring constant which is indicated by Yuzugullu and Schnobrich11). When normal strain reaches ultimate tensile strain, tensile fracture happens and all spring constants are set to zero. When normal stress reaches first compressive yield, normal spring constant is set to half normal spring constant. If normal stress exceeds first compressive yield while shear sliding happens, shear spring constant is set to zero. In this case, shear stress is kept and only normal stress increases. If shear fracture happens in these cases, Eq. (9) is applied. When normal stress reaches second compressive yield, normal spring constant is set to zero. 2. 1. Analytical object and conditions Fig. 2 shows analytical object. The fracture analyses with RBSM assuming 4-point bending tests are carried out and compared with experimental results12). Table 1 shows material properties. In this study, coke is assumed to be homogeneous material which consists of matrix and pore. Since shear strength c of coke is not clarified, the following approximate equation13) of the shear strength for brittle materials like glasses and rocks is applied as: c 30.5 Ft ,
(10)
where Ft is tensile strength. Fig. 3 shows fracture criterion. In this study, triangular elements are applied and the number of meshes is 1956. 2.2. Results and discussion Fig. 4 shows strain-load curve of calculation and experiment. Strain is evaluated from displacement of center bottom of analytical object. In the experiment, strain increases after load reaches the maximum load. This is because strain
800 700
Load [N]
600
Calc. Exp.
500 400 300 200 100 0 0.000
0.001
0.002
0.003
0.004
0.005
Strain [-]
Fig. 4 Strain-load curve of calculation and experiment
Fig. 5 Fracture appearances
Red : Non-adhesion region boundaries
(a) Case A (b) Case B Fig. 6 Arrangement of finite element mesh and 3percent non-adhesion region boundaries gauge broke since fracture happened at the point where strain gauge was set when load reached the maximum load. In this analysis, strain after yield is evaluated like experimental results. As shown in Fig. 4, fracture rapidly happens after plastic deformation happens, so analytical results reproduce brittle fracture behavior well. Fig. 5 shows fracture appearances which are obtained from analysis assuming 4-point bending test. The more the number of Fig. 5, the more load. As shown in Fig. 5, tensile crack happens at the center bottom of analytical object and tensile fracture happens. After fracture happens, crack propagates vertically upwards and coke collapses finally. Analytical results of fracture appearances agree with experimental results12). 3. Analysis of taking no account of non-adhesion region boundaries In low strength coke, non-adhesion region boundaries in coke are observed. It is thought that these defects in coke influence coke strength, so in case of taking account of non-adhesion region boundaries, the fracture analyses with RBSM assuming 4-point bending tests are carried out to investigate the effect of fracture load and propagation of non-adhesion region boundaries. 3. 1. Analytical object and conditions Fig. 6 shows analytical object in case of taking account of non-adhesion region boundaries. Red parts show non-adhesion region boundaries in Fig. 6. Non-adhesion region boundaries are taken into account by lowering fracture strength of coke assuming non-adhesion region boundaries. First, to investigate the effect of location of non-adhesion region boundaries, 3 percent non-adhesion region boundaries are randomly located in the coke and fracture analysis of case A and B (Fig. 6) which are different location are carried out respectively. Fracture strengths of non-adhesion
Stress [MPa]
Red : Non-adhesion region boundaries Black : Tensile fracture 0.4
1)
2)
3)
4)
0.0
Fig. 7 The stress distribution
Fig. 8 Fracture appearance in case A
2)
1)
Red : Non-adhesion region boundaries Black : Tensile facture
4)
3)
Fig. 9 Fracture appearance in case B region boundaries are set to one-tenth of fracture strength of coke. Next, to investigate the fraction of non-adhesion region boundaries in coke, 3, 5 and 10 percent non-adhesion region boundaries are randomly located in coke and 5 analyses in each case are carried out. Fracture strengths of non-adhesion region boundaries are also set to one-tenth of fracture strength of coke. 3.2. Results and discussion 3.2.1. The effect of location of non-adhesion region boundaries in coke Fig. 7 shows stress distribution in case of taking no account of non-adhesion region boundaries. Minus and plus represent compressive and tensile stress, respectively. As shown in Fig. 7, supporting part, load part, and center bottom of analytical object indicate higher stress. This tendency is similar to case of taking account of non-adhesion region boundaries. Fig. 8 shows fracture appearances in case A. Crack propagates vertically upwards from center bottom of analytical object and coke collapses finally. Microcracks happen around non-adhesion region boundaries. It is supposed that these microcracks cause surface breakage of coke. Fig. 9 shows fracture appearances in case B. Fracture starts at non-adhesion region boundaries, especially the continuous ones when non-adhesion region boundaries are located under region of high stress. Fig. 10 shows strain-load curve in base, case A and case B. In Fig. 10, base shows case of taking no account of non-adhesion region boundaries. As shown in Fig. 10, when non-adhesion region boundaries exist in coke, fracture load decreases. Fracture load in case B is lower than one in case A. This is because non-adhesion region boundaries are loca800 700
Base Case A
Load[N]
600
Case B
500 400 300 200
100 0 0.000
0.001
0.002
0.003
0.004
0.005
Strain[-]
Fig. 10 Strain-load curve during 4-point bending test in base, case A and case B
Fracture load [N]
600 500 400 300 200 100 0 0
5 10 Non-adhesion region boundaries fraction [%]
Fig. 11 Relationship between non-adhesion region boundaries fraction and fracture load. ted under region of high stress and they yield earlier than another coke part and stress concentration occurs around them. The slope of case B is lower than the others. It is thought that fracture starts from non-adhesion region boundaries and sliding and fracture happen at these parts and fracture progresses with stress release. 3.2.2. The effect of the fraction of non-adhesion boundaries in coke Fig. 11 shows relationship between non-adhesion region boundaries fraction and fracture load. Fracture loads decrease with an increase of non-adhesion region boundaries in the coke. This is because probability of non-adhesion region boundaries existing at high stress part increases with an increase of non-adhesion region boundaries in coke. It is thought that non-adhesion region boundaries fraction influences coke strength. 4. Conclusions In this study, to investigate the effect of non-adhesion region boundaries in coke, fracture analyses using RBSM are carried out for coke. First, in case of taking no account of non-adhesion region boundaries, the fracture analyses with RBSM assuming 4-point bending tests are carried out. Next, in case of taking account of non-adhesion region boundaries, the fracture analyses with RBSM assuming 4-point bending tests are carried out to investigate the location and the fraction of non-adhesion region boundaries in coke. The conclusions are summarized as follows. (1) When non-adhesion region boundaries exist in coke, fracture load decreases. Fracture starts at non-adhesion region boundaries, especially the continuous ones when non-adhesion region boundaries are located under region of high stress. Microcracks happen around non-adhesion region boundaries. It is supposed that these microcracks cause surface-breakage of coke. (2) Fracture loads decrease with an increase of non-adhesion region boundaries in the coke. It is thought that non-adhesion region boundaries influence fracture propagation and load, so they are important strength factor of coke. Nomenclature c D D(e) D(p) d11, d12, d22 E Fc Ft h k u, v
: : : : : : : : : : : : : : : :
shear strength spring constant matrix spring constant matrix in elastic state spring constant matrix in plastic state spring constant matrix element after plastic deformation elastic modulus compressive strength tensile strength length of perpendicular spring constant displacement component at element’s center of gravity relative displacement between two elements internal friction angle surface force per unit area for normal direction surface force per unit area for shear direction angle of rotation
Subscripts n
:
normal direction
[MPa] [MPa/m] [MPa/m] [MPa/m] [MPa/m] [GPa] [MPa] [MPa] [m] [MPa/m] [m] [m] [] [MPa] [MPa] []
s p
: :
shear direction plastic
References 1) T. Arima: Tetsu-to-Hagané, 87(2001), 274. 2) T. Arima: Tetsu-to-Hagané, 92(2006), 106. 3) K. Ueoka, T. Ogata, Y. Morozumi, H. Aoki, T. Miura, K. Uebo and K. Fukuda : Tetsu-to-Hagané, 92(2006), 184. 4) Y. Asakuma, M. Soejima, T. Yamamoto, H. Aoki, T. Miura and S. Itagaki : Tetsu-to-Hagané, 87(2001), 685. 5) Y. Asakuma, M. Soejima, T. Yamamoto, H. Aoki and T. Miura: ISIJ Int., 43(2003), 1151. 6) M. Soejima, Y. Asakuma, T. Mori, T. Yamamoto, H. Aoki, T. Miura, S. Tanioka and S. Itagaki : Tetsu-to-Hagané, 87(2001), 245. 7) T. Kawai: Journal of the Society of Naval Architects of Japan, 141(1977), 187. 8) K. Nagai, Y. Sato and T. Ueda: Journal of Advanced Concrete Technology, 2(2004), 359 9) Y. Toi and J. CHE: Transactions of the Japan Society of Mechanical Engineers A, 59(1993), 240. 10) N. Takeuchi, M. Ueda, A. Kamibayashi, H. Kitoh, S. Saito M. Tomida, H. Higuchi: Discrete Limit Analysis of Reinforced Concrete Structure, Maruzen, Tokyo, (2005) 11) O. Yuzugullu and W. C. Schnobrich: ACI Journal, 7(1976), 474. 12) T. Ogata, K. Ueoka, H. Hayashizaki, Y. Matsushita, H. Aoki, T. Miura, K. Fukuda and K. Matsudaira: Tetsu-to-Hagané, 93(2007), 736. 13) The Society of Chemical Engineers: Handbook of Chemical Engineers 6th Edition, Maruzen, Tokyo (1999)270.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 PROGRAM TOPIC: COMBUSTION EXACT TITLE OF PAPER: Ilmenite as an Oxygen Carrier in a Chemical Looping Combustion System: Reaction Kinetics and Fluidized Bed Performance Christopher K. Clayton, Graduate Research Assistant The University of Utah Department of Chemical Engineering 50 S. Central Campus Drive, Room 3290 Salt Lake City, Utah 84112 USA Tel. + 1-435-668-0043 Fax. +1-801-585-9291 [email protected] Gabor Konya, Associate Instructor The University of Utah Department of Chemistry 315 S. 1400 East, Room 2020 Salt Lake City, Utah 84112 USA Tel. +1-801-581-5976 Fax. +1-801-581-8433 [email protected] Edward M. Eyring, Professor The University of Utah Department of Chemistry 315 S. 1400 East, Room 2020 Salt Lake City, Utah 84112 USA Tel. +1-801-581-8658 Fax. +1-801-581-8433 [email protected] Kevin J. Whitty, Associate Professor The University of Utah Department of Chemical Engineering 50 S. Central Campus Drive, Room 3290 Salt Lake City, Utah 84112 USA Tel. +1-801-585-9388 Fax. +1-801-585-9291 [email protected]
ABSTRACT In order to better classify potential oxygen carriers for a Chemical-Looping Combustion system, a lab-scale bubbling fluidized bed reactor has been designed and built at the University of Utah. In order to simulate a dualfluidized bed system, the reacting gas is cycled between the oxidizing gas (Air) and the fuel gas (CH4/N2 or H2/CO/N2 mixture). Volume percentages of CH4, CO, CO2 and O2 are analyzed in real time by a CAI NDIR/O2 gas analyzer. Ilmenite, a naturally-occurring and relatively low cost titanium-iron oxide mineral (FeTiO3) is an interesting potential oxygen carrier for chemical looping combustion. Tests to characterize performance of ilmenite as an oxygen carrier and to measure oxidation and reduction kinetics have been conducted at temperatures ranging from 750°C to 950°C. One test was conducted for 24 hours in order to determine the long term performance of ilmenite over several cycles. Such data is important when considering applicability in industrial systems where attrition resistance and long-term reactivity are important. Results compare favorably with thermogravimetric (TGA) studies of reaction kinetics conducted at the University of Utah. This paper describes the experimental approach, data analysis and assessment of the potential for ilmenite as a carrier in full-scale systems. INTRODUCTION Chemical-looping combustion is quickly becoming an attractive alternative for cleaner combustion techniques. Utilizing the oxidation/reduction characteristics of a metal, hereafter known as the oxygen carrier, the process of chemical-looping combustion (CLC) scavenges oxygen from the oxidizing gas and indirectly delivers it to the fuel in a separate reactor. Separation of oxygen from nitrogen by that method inherently creates a nearly pure CO2 stream. Upon moisture removal the CO2 may then be sequestered. Chemical looping combustion (CLC) typically consists of 2 interconnected fluidized beds. The oxygen carrier is transported cyclically between these two beds. The first reactor is called the air reactor (AR) and is responsible for oxidizing the carrier. The second is called the fuel reactor and is responsible for reducing the carrier to its base form. Due to the nature of CLC an oxygen carrier material requires certain chemical and physical characteristics in order to be considered suitable as an oxygen carrier in a CLC system. Important characteristics include cost and availability, oxygen loading capacity, reactivity, resistance to deactivation, good attrition resistance, as well as desirable fluidization characteristics. Ilmenite (FeTiO3) is a naturally occurring mineral that is used primarily in the production of titanium dioxide. Ilmenite is the most abundant form of all titanium materials and is therefore mined in large quantities (1). Due to its abundance and availability ilmenite is an economically feasible option as an oxygen carrier in a CLC system. A number of studies have reported CLC testing using ilmenite as an oxygen carrier (1; 2). A study reporting the oxidation/reduction kinetics of ilmenite has not been found. This may be due to the fact that the term ilmenite may also refer to various ores containing varying amounts of the mineral
ilmenite. The variability that exists in the term ilmenite gives rise to difficulties in defining kinetic variables. This difficulty also arises from the different reaction of ilmenite at temperatures above and below 800oC. In its most reduced state the mineral ilmenite is FeTiO3. Below 800oC the ilmenite does not completely oxidize and follows:
Above 800oC two reactions occur:
where TiO2 is rutile and Fe2TiO5 (pseudobrookite) is the most stable phase (3). EXPERIMENTAL Apparatus A lab-scale CLC reaction system has been constructed at the University of Utah. Typically, a CLC system consists of dual fluidized beds. The reactor at the University of Utah is a single bubbling bed reactor. The dual fluidized beds are simulated in this single bed by switching between the oxidizing and reducing gases. Figure 1is a schematic representation of the CLC system at the University of Utah.
Figure 2: Schematic representation of the CLC reaction system constructed at the University of Utah. Effluent gases are analyzed in real time using a CAI 4 component NDIR/fuel cell analyzer. Three of the components (CH4, CO, and CO2) are measured by the infrared while the oxygen is measured using a fuel cell method. The reactor temperature is controlled by a Carbolite VFT-12-450 vertical tube clamshell furnace and has a 1200oC ceiling. Temperatures are measured both in the particle bed and below the bed using k-type thermocouples. Gas flow is controlled by rotameters and solenoid valves. OPTO-22 controls the solenoid valves and acquires all the data provided by the analyzer and the thermocouples. The ilmenite was purchased from Atlantic American Engineers whom provided the following composition: CHEMICAL ANALYSIS SPECIES Wt. % TiO2 63.90 Fe2O3 29.20 SiO2 0.56 Carbon 0.19 Phosphorus 0.04 Sulfur <0.001 Table 1: Chemical analysis conducted by Atlantic American Engineers of the ilmenite used for CLC testing at the University of Utah. During initial testing it was observed that the ilmenite contained an additional roughly 5 mass% water. Procedure Reacting gases flow from gas cylinders. Rotameters and solenoid valves control the flow of the reacting gases through the system. The flow rates
for both the oxidizing gas (Air) and the reducing gas (5.1% CH4 in N2) are maintained at 5 SLPM. Tests for ilmenite were conducted at 50oC increments between 650oC and 900oC. The operating temperature was controlled according to the measurement produced by the upper thermocouple. Cuadrat, et. al. determined that when calcined for at least 2 hours the ilmenite exhibited a substantial increase in reactivity (2). Therefore, the ilmenite was dried for 2 hours at 300oC and then calcined in air at 800oC for 2 hours. After calcination the ilmenite was then reduced to its base form by switching the reacting gas to the reduction mixture for 30 minutes. After the prepared ilmenite was cooled it was weighed and mixed with a diluent. Ilmenite was mixed with ceramic beads with similar fluidization properties in order to dilute the ilmenite. The dilution allowed for the reacting gases to exist in excess quantities within the reactor even at high temperatures. If the gases were not in excess and all gaseous reactants were consumed then there would be no indication as to the rate of reaction. Ilmenite was diluted to 19.7 mass% in an inert ceramic mixture. After dilution of the ilmenite a 100 g sample of the mixture was placed within the reactor and heated under nitrogen to the desired operating temperature. The experiment began once the center of the particle bed reached the desired operating temperature. When the operating temperature was obtained the cycling of reacting gases began starting at the oxidation stage and moving to the reduction stage. There was a 60 second interval between each oxidation and reduction cycle for gas purging (N2) in order to distinguish the two cycles. Upon completion of the testing the reactor was allowed to cool while the particle bed was continually fluidized under N2. Once the reactor had cooled the sample was reweighed in order to determine the mass increase of the ilmenite during oxidation. Analysis Data was sampled and recorded every second from the gas analyzer. It was determined that the reaction system has a certain gas residence time distribution that must be accounted for in order to properly analyze the data. With an empty bed the reactor was heated to 3 different temperatures. At each of these temperatures air was fed through the reactor at 5 SLPM. The analyzer recorded the O2 breakthrough time as well as the concentration of O2 versus residence time. Three tests were conducted at each of the 3 different temperatures. Averages were then
25 Vol% O2
20 15
100 C
10
600 C
5
750 C
%O2 out / % O2 in
calculated at each of the temperatures and plotted against time.
1 0.8
% Oxygen in Effluent
0.4
% Oxygen Consumed
0.2 0 0
0
0
50 100 Time (Seconds)
25 20 100 C
10
600 C
5
150
Figure 5: The actual oxygen uptake (bottom line) is calculated by taking the data output from the analyzer during oxidation (middle line) and subtracting that from the reactor residence time (top line). This test was carried out at 750 C. Once the percent of oxygen being taken up by the reaction was calculated (figure 5) the number of moles and the mass of oxygen consumed by the ilmenite was determined using:
The number of moles into the reactor was calculated using the ideal gas law at STP and 5 SLPM of air (21% O2). The total mass of oxygen consumed was then plotted at each temperature against time as displayed in figure 6.
0
50 100 150 Time (Seconds) Figure 4: Residence time of Oxygen with the initial oxygen breakthrough time subtracted. It can be determined from figure 4 that the curves generated at each of the different temperatures differ only by the breakthrough time, therefore, as long as the breakthrough time is known for a given temperature the average of the curves shown in figure 4 may be used as the general residence time of the reacting gases. Once the residence time data was gathered the test data was then subtracted from it. Figure 5 displays the residence time at 750oC for oxygen, the data gathered from the analyzer, and the difference between the 2 (actual percentage of oxygen consumed).
Mass of Oxygen Consumed (grams)
750 C
0
100
Time (Sec)
As can be seen from figure 3 each of the temperature exhibits a very similar curve. The major difference between the curves is the breakthrough time. Once the breakthrough times for each of the temperatures was removed the the curves collapsed on top of each other as can be seen in figure 4.
15
50
150
Figure 3: Plot of Residence time of Oxygen from the solenoid valves to the analyzer. Each temperature was tested 3 times and the average of the 3 is shown.
Vol% Oxygen
RTD 750 C
0.6
5.0 4.0 3.0 2.0 1.0 0.0 0
200 400 Time (Seconds) Figure 6: Total oxygen consumed oxidation cycle at 850oC.
600 during
The reaction rate at each temperature was then calculated at the point of 50% conversion and was based on the initial ilmenite loading by:
Where and represent the mass of oxygen and the initial loading of ilmenite in grams respectively.
RESULTS The reaction rate of ilmenite at varying temperatures determined using the fluidized bed is compared against the reaction rates determined using TGA analysis in figure 7.
Reaction Rate gO2/gFeTiO3_initial
TGA Fluid Bed 0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002 0
600
800
1000
Temp (C) Figure 7: Oxidation rates of ilmenite at various temperatures determined by TGA analysis and by Fluidized bed testing. Although differences were observed between the TGA analysis and the fluidized bed analysis these differences may be in part due to differences in the mixing of reactants due to the natures of the systems. Ilmenite was tested over several cycles in order to determine long term performance. As can be seen from figure XXX ilmenite suffered no decrease in reactivity over 24 hours of testing.
CONCLUSIONS Ilmenite was tested for its suitability as an oxygen carrier in a chemical-looping combustion system. Tests were conducted in a laboratory-scale fluidized bed reactor as well as within a TGA unit. Ilmenite demonstrated good attrition resistance as very little was lost due to attrition over the course of several hours and several cycles. Ilmenite also demonstrated good reactivity toward oxygen and did not lose reactivity over several cycles. Calcination of the ilmenite above 800oC prior to testing actually improved its performance. Due to the impurities often found within ilmenite along with its range of compositions it is difficult to derive specific kinetic constants. Overall, ilmenite performs well as an oxygen carrier in a CLC system. Especially when considering the environment of industrial-size systems ilmenite, being an inexpensive material, would be suitable as an oxygen carrier. REFERENCES
1. The use of ilmenite as an oxygen carrier in chemical-looping combustion. Leion, Henrik, et al. 2008, Chemical Engineering Research and Design, pp. 1017-1026. 2. Behaviour of Ilmenite as Oxygen Carrier in Chemical-Looping Combustion. Cuadrat, Ana, et al. Dresden : s.n., 2009. Clean Coal Technology Conference. 3. Low-temperature oxidation of ilmenite (FeTiO3) induced by high energy ball milling at room temperature. Chen, Ying. 1997, Journal of Alloys and Compounds, pp. 156-160.
Figure 8: Ilmenite testing at 850oC for 24 hours. The plot shows no loss in reactivity of ilmenite over the length of the tests. As a check for attrition resistance the sample material was weighed fully reduced prior to testing and then again upon test completion. After 40 hours of testing and 32 complete cycles a 19.99g sample of ilmenite had lost only 0.01g.
4. Evolution of Nitrogen and Other Species During Controlled Pyrolysis of Coal. Blair, D. W., Wendt, J. O. L. and Bartok, W. s.l. : Elsevier, 1977. Symposium (International) on Combustion. pp. 475-489. 5. Chemical-Looping Combustion of Petroleum Coke Using Ilmenite in a 10 kWth. Bergurand, Nicolas and Lyngfelt, Anders. 2009, Energy & Fuels, pp. 5257-5268.
6. Gasification of carbon: metal oxides in a fluidized powder bed. Lewis, W. K., Gilliland, E. R. and Sweeney, W. P. 1951, Chemical Engineering Progress, pp. 251-256. 7. Mineral Chemistry of Fe-Ti oxides from Xinjie PGE-bearing layered mafic ultra-mafic intrusion
in Sichuan, SW China. Wang, Christina Yan and Zhou, Mei-Fu. Beijing : Springer, 2004. Mineral Deposit Research, meeting the global challenge. pp. 481-485.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission COMBUSTION TECHNOLOGIES APPLICATION OF INORGANIC REMAINS ORIGINATING FROM WATER PURIFICATION AND SEWAGE SLUDGE ASHES IN CHEMICAL LOOPING COMBUSTION PROCESS Ewelina Ksepko, PhD, Institute for Chemical Processing of Coal, 1 Zamkowa, 41-803 Zabrze, POLAND, E-mail: [email protected], Tel: 00 48 32 271-00-41 int. 149 Grzegorz Łabojko, Ph.D., Institute for Chemical Processing of Coal, 1 Zamkowa, 41-803 Zabrze, POLAND, [email protected], Tel: 00 48 32 271-00-41 int. 133
Marek Sciazko PhD, Institute for Chemical Processing of Coal, 1 Zamkowa, 41-803 Zabrze, POLAND, E-mail: [email protected], Tel: 00 48 32 271-00-41 int. 240
Abstract: The presented paper contains results of research on chemical looping combustion (CLC). The objective of paper was to prepare new low cost materials [1] used as a oxygen carriers and investigate their reactivity in terms of applicability to energy systems. Multi cycle CLC tests in atmospheric TGA with oxygen carriers utilizing H 2 were conducted. Stability and oxygen transport capacity were described. Chemical phase composition was investigated by XRD technique. Three cycle test at 600 °C, 700 °C and 800 °C showed that proposed materials as oxygen carriers showed stable performance during the 3-cycle test at each temperature. The oxygen transport capacity has varied from 4 to 13.9 wt %, depending on temperature. The fractional reduction, fractional oxidation and global rates (dX/dt) of reactions were calculated. It was found, that thermal treatment of wastes did not affect reduction/oxidation reaction rate. Oxidation reaction was much faster than reduction for all oxygen carriers within time of 1.5 minute. High melting temperature in reducing atmosphere of 1180 ˚C was observed. Small particle size below 59 µm was detected. XRD pattern after multi-cycle test showed stable crystalline phase and confirmed that complete regeneration after multi-cycle tests was achieved. Acknowledgement: Work supported by the Polish Ministry of Education and Science, Grant № 11.10.009 The authors are gratefully to Upper Silesian Enterprise of Water Supply Joint Stock to make samples available. [1] Application for the patent No P390127.
“CHEMICAL LOOPING WITH OXYGEN UNCOUPLING: DESIGN CALCULATIONS AND PROCESS ENGINEERING SIMULATIONS USING KINETIC DATA” JoAnn S. Lighty Department of Chemical Engineering, Institute for Clean and Secure Energy University of Utah, 50 S. Central Campus Drive, Room 3290 Salt Lake City, UT 84112-9203 Email: [email protected] Adel F. Sarofim1, Asad H. Sahir1, Edward Eyring2, Gabor Konya2 1 Department of Chemical Engineering, Institute for Clean and Secure Energy, University of Utah 2 Department of Chemistry, University of Utah Abstract: Chemical Looping with Oxygen Uncoupling (CLOU) is a process which is being researched as a potential candidate for combustion of solid fuels facilitating effective CO2 capture. An engineering analysis has been developed using data from in-house experiments on copper oxide as an oxygen carrier and utilizing global coal char oxidation kinetics reported in literature. Insights from the experimental study of Lewis and Gilliland on carbon gasification with metal oxide along with the developed engineering analysis have been incorporated to develop scenarios to study the impact of the residence time of air and fuel reactors and the recirculation rate of oxygen carrier on the process. The analysis has been extended by employing an ASPEN PLUS simulation of stoichiometric reactor models to develop relevant material and energy balances in the formulated scenarios. In-house experiments for gathering the kinetic data on unsupported and supported copper oxide have been conducted using a Thermo-gravimetric analyzer (TGA). From these experiments conducted on CLOU, supported copper oxides have emerged as promising candidates and they are being explored further. The experimental data are currently being evaluated for kinetic parameters to be incorporated into the simulations. The results of the order-of magnitude engineering analysis based on gas-solid reaction kinetics, and global coal char oxidation kinetics are presented in this paper. Design aspects of importance in identifying suitable conversions for the fuel and air reactors have been investigated as well as the role and impact of oxygen carrier recirculation rate on the energy balance of the process. 1.
Introduction
Chemical Looping Combustion (CLC) is an emerging combustion technology which has the potential to assist in the economic capture of CO2. Lewis, Gilliland and Sweeney in 1951[1], reported the use of CuO as an oxygen carrier, which would be instrumental in supplying the requisite gas phase oxygen for combustion of a solid fuel. Independent investigations were carried out by Mattison et al. [2-4] in the latter part of the past decade for combustion processes
where oxygen dissociated from a metal oxide could be utilized in combusting solid fuels. They named these classes of CLC processes Chemical Looping with Oxygen Uncoupling (CLOU). CLOU facilitates the reaction of solid carbonaceous fuels with metal oxides, by proceeding via a gaseous intermediate. In contrast, CLC of a solid carbonaceous fuel involves intermediate steps of gasifying solid carbon into CO and H2, followed by oxidation of the products of the gasification reaction with a metal oxide supplying the requisite oxygen for the process. CLC of a solid fuel involves two reactions; release of oxygen from the metal oxide and the gasification of the solid fuel, which is rate limiting in nature [5]. CLOU has the potential to avoid the gasification step thus reducing the volumes of the fuel reactor. 2. Insights from Studies on Combustion of Solid Fuels using Copper Oxide Lewis, Gilliland and Sweeney in 1951[1], reported the augmentation of the C gasification reaction by employing a metal oxide. Their experimental setup consisted of an isothermal and isobaric batch fluidized reactor where 100-140 mesh metallurgical coke was gasified by the CuO supported on 20-200 mesh silica gel. Analysis of their experimental data reveals the sixty-fold enhancement of gasification rates by C+CO2 and C+O2 reactions when the O/Cu goes from 1 to 0.5, i.e. from CuO to Cu2O. The details of the analysis are described elsewhere [6]. A fifty fold increase in reaction rate has been reported by Mattisson et al. [2] for CuO oxygen carrier in CLOU experiments conducted on Mexican Petroleum Coke in the temperature range of 950-985°C, as compared to CLC with an iron based oxygen carrier. CLC using copper to cuprous oxide in a batch fluidized bed reactor has been used to study the combustion of lignite (Hambach, Germany) and two bituminous coals (Taldinskaya, Russia and US Illinois No. 5). The complete combustion of lignite was reported but the bituminous coals reacted incompletely [7]. In a recent study reported on copper oxide, experiments at temperatures of 500-700°C in a TGA have shown that solid-solid reactions mechanism greatly accelerate the reaction C + 2CuO → 2Cu + CO2 [8]. 3. Oxidation and Reduction Kinetic Studies The kinetics of the oxidation and reduction reactions with the unsupported copper oxide was studied using two TGA’s – a TA Q500 and TA Q600. First order reaction rate constants for the decomposition of CuO to Cu2O by the reaction (2CuO → Cu2O + ½ O2) and pseudo-first order oxidation rate constants (Cu2O + ½ O2 → 2CuO) were evaluated from the TGA data. The details of the experiment and method to elucidate the chemical kinetics are described elsewhere [6]. Figure 1 shows a plot of the logarithm of the rate constants versus the reciprocal of absolute temperature. The rate of decomposition increased with increasing temperature. On the contrary, the rate of oxidation decreased with increasing temperature. The decrease in oxidation rate is
attributed to the decrease in grain-boundary diffusion of oxygen through the CuO product surrounding the Cu2O due to the increased size of the CuO grains with increase in temperature [9, 10]. If a supported CuO is used, then the reduced depth of CuO minimizes such diffusion limitations. Hence, supported carriers are likely to yield better kinetics, as has been demonstrated by the experiments reported in literature for CLC of gaseous fuels [11, 12].
Figure 1 Plot of Kinetic Rate Constants obtained from TGA Experiments
4. Order of Magnitude Estimate Analysis for Design of CLOU systems for solid fuels and development of an ASPEN PLUS Simulation A preliminary order of magnitude design analysis has been formulated for CLOU for solid fuels which encompass the following aspects. – (The details of the underlying mathematical relationships and derivations used to formulate the criteria have been discussed elsewhere [6]): 1. The air and fuel reactors have to be operated at an optimum circulation rate of metal being introduced in the system. The design aspect has been discussed extensively for CLC of gaseous fuels [13]. The circulation rate aids in establishing the appropriate conversions at which the air and fuel reactors should be operated. An optimum recirculation rate of 135 kg CuO/MWt was determined based on in-house kinetic studies. From the results on the circulation rates, it was decided that a 69 % molar conversion of Cu2O and a 56 % molar conversion of CuO were close to optimum.
2. In the present work, the in-house kinetic studies have been utilized for establishing the temperature of 850°C for the air reactor and 950°C for the fuel reactor with the former providing the maximum rate of oxidation and the latter the highest rate of reduction of copper oxide before agglomeration becomes an issue. 3. The shrinking sphere model in conjunction with studies on pulverized coal char combustion kinetics [14] has been utilized to determine the burnout times for coal particles, which have been compared with the time required for decomposition of CuO required for the release of oxygen by the use of an idealized plug flow reactor model using the in-house kinetic studies. The analysis revealed that for an order of magnitude estimate, the residence time for CuO decomposition could be used as a design variable. An ASPEN PLUS simulation was conceptualized using the stoichiometric reactor RSTOIC model. For simplicity the following assumptions have been utilized: a) CuO releases gas phase oxygen by the reaction 2CuO → Cu2O + ½ O2 in the fuel reactor with a 56% molar conversion of CuO. b) Coal char is assumed to be a pure carbon as an engineering simplification for the simulation. The carbon completely reacts in the fuel reactor with the released gas phase oxygen from the CuO by the reaction C+O2 → CO2. c) Cu2O is re-oxidized to Cu2O by the reaction 2CuO → Cu 2O + ½ O2 in the fuel reactor with a 69% molar conversion of Cu2O. The results of an ASPEN PLUS simulation clearly indicate the possibility of energy recovery from the air and fuel reactors as oxidation of Cu2O is an exothermic reaction and the overall reaction in the fuel reactor (i.e. CuO decomposition and C combustion) are exothermic in nature. 5. Chemical Kinetics Effect in Solid Fuel Combustion in CLOU Combustion of solid fuels may either be governed by chemical kinetics or mass transfer. To investigate these regimes in CLOU, mathematical relationships have been developed to investigate the regimes with reference to solid fuels [6]. In the fuel reactor, the supply of O2 is governed by the decomposition of CuO by the reaction 4CuO → 2Cu2O + O2. The rate of CuO decomposition can be expressed as
𝑟𝑟𝑂𝑂2 ,𝐶𝐶𝐶𝐶𝐶𝐶 =
𝑘𝑘 𝐶𝐶𝐶𝐶𝐶𝐶 [𝐶𝐶𝐶𝐶𝐶𝐶 ] 4
−
1/2 2
𝑘𝑘 𝐶𝐶𝐶𝐶 2 𝑂𝑂 [𝐶𝐶𝐶𝐶 2 𝑂𝑂]𝑝𝑝 𝑂𝑂 2
(1)
At equilibrium when the rate of oxygen release equates to zero, the ratio of rate constants results in Equation (2) 𝑘𝑘 𝐶𝐶𝐶𝐶 2 𝑂𝑂 𝑘𝑘 𝐶𝐶𝐶𝐶𝐶𝐶
=
[𝐶𝐶𝐶𝐶𝐶𝐶 ]𝑒𝑒𝑒𝑒
1/2 2,𝑒𝑒𝑒𝑒
2[𝐶𝐶𝐶𝐶 2 𝑂𝑂]𝑒𝑒𝑒𝑒 𝑝𝑝 𝑂𝑂
(2)
Equation (2) can be substituted to formulate a relationship, Equation (3) in terms of fractional approach to equilibrium, fe
𝑟𝑟𝑂𝑂2 ,𝐶𝐶𝐶𝐶𝐶𝐶 =
𝑘𝑘 𝐶𝐶𝐶𝐶𝐶𝐶 [𝐶𝐶𝐶𝐶𝐶𝐶 ] 4
[1 − 𝑓𝑓𝑒𝑒 ]
(3)
The fractional approach to equilibrium has been defined as fe
𝑓𝑓𝑒𝑒 =
�
�
1/2 𝑂𝑂 2 � [𝐶𝐶𝐶𝐶𝐶𝐶 ]
[𝐶𝐶𝐶𝐶 2 𝑂𝑂]𝑝𝑝
[𝐶𝐶𝐶𝐶 2 𝑂𝑂 ]𝑒𝑒𝑒𝑒 𝑝𝑝
1/2 𝑂𝑂 2,𝑒𝑒𝑒𝑒
[𝐶𝐶𝐶𝐶𝐶𝐶 ]𝑒𝑒𝑒𝑒
(4) �
At equilibrium or fe = 1 , CuO decomposition is governed by mass transfer. In conjunction with the assumption of pulverized coal particles obeying the regime Sh = 2 and utilizing the models for pulverized coal char global kinetics, mathematical relationships were derived to investigate the limit of equilibrium. The details of the derivations have been discussed elsewhere [6]. To study the effect of Chemical Kinetics in solid fuel combustion for the limit fe = 0, a mathematical model for CLOU was developed with the following considerations: 1. The fuel reactor is an isothermal and isobaric ideal plug-flow reactor. 2. A pseudo-steady-state oxygen balance was formulated taking into account release of O2 by CuO particles and O2 consumption by pulverized coal char particles. 3. The char combustion was modeled using a shrinking sphere combustion model where the kinetic constants were obtained from literature studies on global models of pulverized coal combustion [14]. 𝑑𝑑𝑑𝑑 𝑘𝑘 1/2 = − 𝜌𝜌 𝐶𝐶 𝑝𝑝𝑂𝑂2 (5) 𝑑𝑑𝑑𝑑 𝐶𝐶
4. The decomposition of CuO obeys a first- order rate law as indicated by the in-house kinetic studies [6]. 𝑑𝑑𝜒𝜒 𝐶𝐶𝐶𝐶𝐶𝐶 = 𝑘𝑘𝐶𝐶𝐶𝐶𝐶𝐶 [1 − 𝜒𝜒𝐶𝐶𝐶𝐶𝐶𝐶 ] (6) 𝑑𝑑𝑑𝑑 The resulting set of differential and an algebraic equation was solved using ode15s solver in MATLAB. 6. Comparison of CLOU Experimental Data with developed analysis
To validate the approach, data from results of CLOU experiments conducted on a German Lignite and Mexican Petcoke using a 40% CuO supported on ZrO2 as an oxygen carrier were utilized [3,4]. For simplicity global, pulverized coal char combustion correlations of equivalent U.S. Coals were chosen for simulating German Lignite and Mexican Petcoke, in this case a Lower Wilcox coal char and Pocahontas coal char respectively as per Table-1[14, 15].
Table -1 Ultimate Analysis for Different Coals Coal German Lignite Lower Wilcox Mexican Petcoke Pocahontas
C(wt% d.a.f) 69.9
H (wt% O(wt% d.a.f) d.a.f) 5.4 23.1
N(wt% d.a.f) 0.6
S(wt% d.a.f) 1.0
Cl(wt % d.a.f) -
Heating Value (MJ/kg)-as recd. 20.9
72.34
5.21
20.11
1.35
0.94
0.07
16.4
88.8
3.1
0.5
1.0
6.6
-
30.9
91.48
4.38
2.30
1.10
0.69
0.06
33.4
The experimental results for 95% burnout for a devolatilized German Lignite were compared with the model simulations for the above referenced Lower Wilcox coal char. The results of the comparison are shown in Figure-2. As seen in this figure, the trend of the data and the simulation are comparable indicating that the assumptions of the model appear valid. To further understand the oxygen partial pressure concentration, an analysis was completed as stated in the previous section. Figure-3 illustrates the ratio of partial pressure of O2 in the reactor with respect to the equilibrium O2 (due to CuO decomposition). The ratio of partial pressure at the reactor outlet to that of equilibrium O2 is less than 0.30. A visual observation of the experimental data reported by Leion et al. [3] indicates that the O2 concentration was negligible at the exit of the fluidized bed reactor. Hence the simulation results indicating a low partial pressure of oxygen are in agreement with experiments. The low oxygen partial pressure is a result of the highly reactive lower ranked lignite which consumes oxygen at a higher rate.
Figure 2 Plot of comparisons of experimental data for CLOU Combustion of German Lignite with simulations for a Lower Wilcox Coal Char
Figure 3 Plot of ratio of partial pressure of oxygen at reactor outlet and equilibrium partial pressure of oxygen with time for the case of a Devolatilized German Lignite simulated as a Lower Wilcox Coal Char
The results of the simulations for 95% burnout Mexican Petcoke have been compared with the global pulverized coal char kinetics obtained for U.S. Pocahontas coal. Figure-4 represents the comparison of the experimental data obtained for 95% burnout of a Mexican Petcoke compared with the simulation results for a Pocahontas coal char. The trend of the experimental data and simulations are similar.
Figure-5 represents the plot of ratio of partial pressure of oxygen with the equilibrium partial pressure of oxygen with time. The ratio approaches unity for all the cases which corresponds with a visual observation of the data of Mattisson et al. [4]. The oxygen concentration at the outlet of the reactor remained near the equilibrium concentration for combustion of Mexican petcoke by CLOU suggesting the lower reactivity of the petcoke.
A comparison of Figures 2 and 4 show that the simulation more closely predicted the results for the Mexican Petcoke. Leion et al. [3] report a relatively high amount of CO release from the Devolatilized German Lignite, but no CO was evolved in the combustion of Mexican Petcoke, indicating a complete conversion of coke to carbon dioxide. Hence, this could be an explanation for the poorer fit of the simulation with the Lignite data, although other experimental factors could also be important.
Figure 4 Plot of comparisons of experimental data for CLOU Combustion of Mexican Petcoke simulated as a Pocahontas Coal Char
Figure 5 Plot of ratio of partial pressure of oxygen at reactor outlet and equilibrium partial pressure of oxygen with time for the case of a Mexican Petcoke simulated as a Pocahontas Coal Char
7. CONCLUSIONS In this study, mathematical models have been formulated which utilize data from in-house kinetic studies from TGA and global pulverized coal char kinetics in conjunction with material and energy balances from ASPEN PLUS. The importance of the kinetic and equilibrium regimes of combustion of solid fuels by CLOU is an important design consideration. Preliminary mathematical models which could analyze experimental studies on CLOU on solid fuels [3, 4] have been developed. Advanced analysis of these insights in future studies may be instrumental in the design of reactors for solid fuel combustion for CLOU process. ACKNOWLEDGEMENTS This material is based upon work supported by the U.S. Department of Energy under Award Number DE-NT0005015.
REFERENCES: 1. Lewis, W. K., Gilliland, E. R., and Sweeney, W. P. (1951) Gasification of Carbon Metal Oxides in a Fluidized Powder Bed, Chemical Engineering Progress,47 ,5, 251-256. 2. Mattisson, T., Lyngfelt, A., and Leion, H. (2009) Chemical-looping with oxygen uncoupling for combustion of solid fuels, Journal of Greenhouse Control, 3, 11-19. 3. Leion, H., Mattisson, T. and Lyngfelt, A. (2008) Combustion of a German lignite using chemical-looping with oxygen uncoupling (CLOU), The Clearwater Coal Conference - The 33rd International Technical Conference on Coal Utilization & Fuel Systems, Clearwater, Florida. 4. Mattisson, T., Leion, H. and Lyngfelt, A. (2009) Chemical-looping with oxygen uncoupling using CuO/ZrO2 with petroleum coke, Fuel, 88, 683-690. 5. Leion, H., Mattisson, T. and Lyngfelt, A. (2008) CO2 Capture from direct combustion of solid fuels with Chemical Looping Combustion, The Clearwater Coal Conference - The 33rd International Technical Conference on Coal Utilization & Fuel Systems, Clearwater, Florida. 6. Eyring, E.M., Konya, G., Lighty,J.S., Sahir, A.H., Sarofim, A.F., Whitty, K.J., Chemical Looping with Copper Oxide as Carrier and Coal as Fuel, in review, Oil & Gas Science and Technology, 2010. 7. Dennis, J. S., Müller, C. R., Scott, S. A. (2010) In situ gasification and CO2 separation using chemical looping with a Cu-based oxygen carrier: Performance with bituminous coals, Fuel, 89, 2353 – 2364. 8. Siriwardane, R., Tian, H., Miller, D., Richards, G., Simonyi, T., Poston, J. (2010), Evaluation of reaction mechanism of coal-metal oxide interactions in chemical-looping combustion, Combustion and Flame, doi:10.1016/j.combustflame.2010.06.008 9. Prisedsky, V.V., Vinogradov V.M. (2004) Fragmentation of diffusion zone in hightemperature oxidation of copper Journal of Solid State Chemistry, 177 , 4258–4268. 10. Zhu,Y. , Mimura,K. and Isshiki, M. (2004) Oxidation Mechanism of Cu2O to CuO at 600– 1050°C, Oxidation of Metals, 62, 3-4. 11. Garcia-Labiano, F., de Diego, L. F., Adanez, J., Abad, A., Gayan, P. (2004), Reduction and Oxidation Kinetics of a Copper-Based OxygenCarrier Prepared by Impregnation for Chemical-Looping Combustion, Industrial Engineering and Chemistry Research, 43, 81688177. 12. de Diego, L. F., Gayan, P., Garcia-Labiano, F., Celaya, J. Abad, A., Adanez, J. (2005), Impregnated CuO/Al2O3 Oxygen Carriers for Chemical-Looping Combustion: Avoiding Fluidized Bed Agglomeration, Energy and Fuels, 19, 1850-1856.
13. Abad, A., Adanez, J., Garcia-Labiano, F., de Diego, L. F., Gayan, P., Celaya, J. (2007) Mapping of the range of operational conditions for Cu-, Fe-, and Ni-based oxygen carriers in chemical-looping combustion”, Chemical Engineering Science, 62, 533 – 549. 14. Hurt, R.F., Mitchell, R.E. (1992), Unified High-Temperature Char Combustion Kinetics for a suite of coals of various rank, 24th International Symposium on Combustion, 1243-1250. 15. Smith, K.L., Smoot, L.D., Fletcher, T.H., Pugmire, R.J.(1994) , The Structure and Reaction Processes of Coal, Plenum Press.
NOMENCLATURE: [𝐶𝐶𝐶𝐶𝐶𝐶]
Concentration of CuO
(mol/m3)
[𝐶𝐶𝐶𝐶2 𝑂𝑂]
Concentration of Cu2O
(mol/m3)
𝑓𝑓𝑒𝑒
Fractional approach to equilibrium
(--------)
𝑘𝑘𝐶𝐶
Rate constant for Global Char Oxidation
𝑘𝑘𝐶𝐶𝐶𝐶𝐶𝐶
First order rate constant of CuO decomposition
𝑘𝑘𝐶𝐶𝐶𝐶 2 𝑂𝑂 𝑀𝑀𝑀𝑀𝑡𝑡
Megawatts of thermal energy
r
Radius of particle
𝑟𝑟𝑂𝑂2 ,𝐶𝐶𝐶𝐶𝐶𝐶
Rate of oxygen release by CuO
Rate constant of Cu2O oxidation
𝑝𝑝𝑂𝑂2
Sherwood Number
t
Time
𝑆𝑆ℎ
Partial pressure of O2
(g/cm2atm0.5 s) (s-1) (s-1atm-0.5) (MJ/s) (cm.) (mol/m3s) (atm) (--------) (s)
Greek Symbols 𝜒𝜒
𝜌𝜌𝐶𝐶
Molar Conversion of CuO Density of char
Subscripts eq
Equilibrium
(--------) (g/cm3)
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 PROGRAM TOPIC COMBUSTION – NOVEL TECHNOLOGIES (OXYFUEL, CHEMICAL LOOPING, ETC) CHEMICAL LOOPING COMBUSTION AND GASIFICATION – A NOVEL TECHNIQUE TO PRODUCE CONCENTRATED STREAM OF HYDROGEN AND CARBON DIOXIDE FROM VICTORIAN LIGNITES Presenting Author Chiranjib Saha PhD Student Department of Chemical Engineering, Monash University, Victoria – 3800, Australia Telephone: +61 3 9905 1347, Fax: +61 3 9905 5686 E-mail: [email protected] Co-Author Dr. Ali Akhavan Postdoctoral Research Fellow Department of Chemical Engineering, Monash University, Victoria – 3800, Australia Telephone: +61 3 9905 4915, Fax: +61 3 9905 5686 E-mail: [email protected] Co-Author Dr. Sankar Bhattacharya Associate Professor, Department of Chemical Engineering, Monash University, Victoria – 3800, Australia Telephone: +61 3 9905 9623, Fax: +61 3 9905 5686 E-mail: [email protected]
Abstract: Chemical-Looping combustion and gasification is a novel technique where an oxygen carrier is used to transfer oxygen from the combustion air to the fuel, thus avoiding direct contact between air and fuel. This technology is an alternative to conventional combustion and gasification where oxygen needs to be supplied from air or air separation plants which are energy intensive. This new methodology prevents CO2 from being mixed with combustion gases. Chemical looping concept has been widely studied for combustion of natural gas; however its application to solid fuel, such as coal, is being studied only recently. From literature it is evident that limited number of studies has been performed on chemical looping using lignites. Victoria has large resources of lignites (>500 years at current consumption rate) and therefore there is a strong incentive for development of efficient technologies, such as chemical looping, for power generation from lignites. Oxide of metals such as Nickel, Copper, Cobalt, and Manganese are good oxygen carrier candidates and have been studied extensively to be used in chemical looping process. Iron oxide is an inexpensive mineral in Australia and when used in chemical looping, it is expected to generate concentrated stream of CO2 and H2. However, much is unknown about the yield of products such as H2, CO2, CO, Char etc. from this process as a function of time, temperature, particle size and type of lignites. Present literature also lack information regarding the fate of the externally added Fe2O3 particles through their interaction with the constituents of lignites and their prospects for regeneration. This paper explores the possibility of using lignite in chemical looping with iron oxide as oxygen carrier. The reduction and reoxidation properties of Fe2O3 are investigated using thermogravimetric analyzer (TGA). Scanning electron microscopy (SEM) along with EDX of fresh solid reactants is compared with used reactants to understand the changes in surface morphology and mineral composition. Surface elemental information of agglomerates as a function of temperature and time are also being investigated. This paper presents preliminary results from this ongoing study.
1. Introduction Conventional coal based power generation involving combustion and gasification requires direct contact between air and coal. This dilutes the resulting gases due to the presence of nitrogen (N2) in air. So the concentration of CO2 is reduced in flue gases which make CO2 separation and capture energy intensive. Chemical looping combustion and gasification (CLC/G) is one of the promising technologies to achieve potentially easier CO2 capture. In chemical looping combustion, an oxygen carrier; preferably metal oxide, is used to transfer oxygen from combustion air to fuel. In this way direct contact between air and fuel is avoided which essentially prevents CO2 from being mixed with combustion gases. The CLC system is composed of two reactors: an air reactor and a fuel reactor. In the fuel reactor, the fuel particles are reduced by oxygen from metal oxide and reduced metal oxide is reoxidized in the air reactor to be used in the next cycle. The main advantage of this system is a concentrated stream of CO2, free from dilution with nitrogen and its oxides can be obtained from fuel reactor after condensing the water vapour. A generalized description of the overall system is given in Figure 1. Reaction stoichiometry is given by the equations below:
(2n+m) MexOy + CnH2m → (2n+m) MexOy-1 + mH2O + nCO2
(1)
MexOy-1 + ½ O2 → MexOy
(2)
where Me stands for metal and MeO as metal oxide. N 2/O 2
CO 2/H 2O
MeO /Me
Air Reactor
Fuel Reactor
Me/ MeO
Air
Coal
Fig. 1. Schematic of Chemical Looping Combustion.
Chemical looping concept has been widely studied and important progress has been made for combustion of natural gas (Adanez et al., 2010).The Chemical looping methodology has been investigated previously with synthesis gas, derived from coal, as feedstock (Diego et al., 2004; Lyngfelt et al., 2008). However, only a few studies have been conducted directly using combustion of solid fuels, such as coal (Siriwardane et al., 2009). From literature it is evident that limited numbers of investigations have been performed on chemical looping using lignites. In situ chemical looping gasification of Hambach German Lignite has been performed with Copper oxide as oxygen carrier (Dennis & Scott, 2009). However chemical looping has never been tested with Victorian lignites. Victoria has large reserves of lignites (>500 years at current consumption rate) and therefore there is a strong incentive for development of efficient technologies, such as chemical looping, for power generation from lignites. Metal oxides of Nickel, Copper, Cobalt, Manganese and Cadmium have been discussed in literature for Chemical looping combustion of gaseous fuels such as natural gas and CH4 (Rubel et al., 2009). Pure Hematite (Fe2O3) and externally impregnated support metal oxides with this has been investigated for application in chemical looping mostly with gaseous fuel (Lyngfelt et al., 2008). Coal gasification experiments in a fluidized bed proved the feasibility of coal CLC with Fe2O3 as the oxygen carrier. It was observed the amount of CO and CO2 produced is consistent with the amount of coal added (Siriwardane et al., 2009; Lyon et al., 2000). Natural iron titanium oxide or Ilmenite has been investigated as oxygen carrier with South African Coal and Petroleum coke by Berguerand & Lyngfelt (Berguerand et al., 2008 & 2009). Synthetic particle of 60% active Fe2O3 and 40% MgAl2O4 has also been investigated with solid fuel by Leion et al. (Leion et al., 2008). Dennis & Scott reported gasification of lignite char under steam and CO2 atmosphere with Fe2O3 as oxygen carrier (Dennis & Scott 2009). It is evident that application of Fe2O3 as oxygen carrier for CLC of lignites has not been investigated widely. Moreover iron oxide is inexpensive in Australia and when used in CLC it is expected to generate concentrated stream of CO2 and H2. There are issues which need to be solved for CLC application of Victorian lignite with Fe2O3. Much is unknown about the rate and yield of products (such as H2, CO2, CO, Char etc.) from this process as a function of time, temperature, particle size and type of lignites. The fate of externally added Fe2O3 particles through the reaction process is unknown. The durability of oxides to sustain repeated cycles through oxidation and reduction in presence of solid residues is another area of concern. Issues related to agglomeration of solid particles over the period of time with variation in temperature during reduction in real reactor situation are also unknown. In the present work some of the issues identified above are explored. The reduction and reoxidation properties of Fe2O3 have been investigated in a Thermo gravimetric Analyser (TGA). Scanning electron microscopy (SEM) along with EDS of fresh solid reactants is compared with used reactants to understand the changes in surface morphology and mineral composition. Surface elemental information of agglomerates as a function of temperature and time are also being investigated. This paper presents preliminary results from this ongoing study.
2. Experimental section Coal samples Chemical looping experiments have been performed primarily with a Victorian brown coal - Loy Yang. The raw coal samples have been first sieved to particle size of 100-150 µm. Similar coal particle size have been used by Siriwardane et al. and Cao et al. in their Chemical looping combustion experiments with coal (Siriwardane et al., 2009; Cao et al., 2006). The sieved coal is then dried in an oven for 6 hours at 105°C. Experiments have also been conducted using a Brazilian lignite to compare the weight loss and reactivity characteristics between a high (Brazilian lignite) and low (Loy Yang) ash content coal. The Brazilian lignite used is from Santa Catarina state and prepared similar to the Loy Yang lignite for experimental purposes. Oxygen carriers The oxygen carrier used in this study is particles of Fe2O3. Fe2O3 (Ferric oxide calcined) is directly from the manufacturer - BDH Chemicals Ltd. with a purity of more than 99%. It has a maroon colour. Initially the particles were in larger lumps. It was broken and particles were sieved in size fraction of 100-150 µm (same as coal particle size) to provide a good physical mixing with coal samples. The procedure was repeated until a sufficient quantity of particles in the size range was obtained. 120 mg of Fe2O3 particles were mixed with coal particles to maintain the desired mass ratio. Samples preparation Coal was physically mixed with metal oxides for Chemical looping combustion experiments. The mass ratio used (metal oxide mass/coal mass) was 6:1which correspond to stoichiometric oxygen supply. 20 mg coal samples have been used for all the experiments. In case of Redox experiments, same amount of coal has been mixed with reacted metal oxide/ash in the TGA crucible. Ash could not be separated after each cycle and hence was accumulated after each cycle in TGA crucible. Instruments Tests were conducted using Thermo gravimetric analyser (TGA). Scanning electron microscope (SEM) was used for solid characterization. The solid residue samples were collected carefully after each test for SEM analysis.
Thermogravimetric Analyser (TGA)
The effect of consecutive reduction-oxidation cycles on the reactivity of metal oxides using various experimental conditions and gas compositions was assessed in a TG-DTA/DSC apparatus with steam injection capability. Experiments were carried out in an alumina crucible of 18 mm ID and 1 mm wall thickness with a tolerance value of ±0.5. The coal and metal oxide mixture was heated from ambient temperature to 950°C at a heating rate of 10°C/min under
different experimental environment and gas compositions. Different test runs were performed to optimize the heating rate and based on that 10°C/min heating rate has been finalized. All the experiments included a 2 hour isothermal hold at the end of the heating segment. Several test runs proved that at 950°C, after a 2 hour isothermal, the loss in weight of samples are almost constant. In case of Redox experiments, the sample was kept isothermal at 950°C for 1 h and afterward air was introduced to oxidize the reduced particles for 1 more hour. The reducing environments used were pure air and 20%CO2 + 80% N2. The metal oxide particles have been re-oxidized using air. Scanning Electron Microscopy (SEM)
The reaction residues were carefully placed on carbon tape for secondary electron imaging with energy dispersive spectra (SEM-EDX). The analysis was performed using a JOEL 840A SEM instrument.
3. Results and Discussions The composition of the dried Loy Yang brown coal and the Brazilian lignite used in this study are given in Table 1. Table 1: Component analysis of Loy Yang Brown Coal Loy Yang Coal Properties (% db) Ash
Weight (%)
1.5-1.7
Volatile Matter
50.5-51.3
Carbon
68.3-69.2
Hydrogen
4.8-4.9
Sulphur
0.4
Nitrogen
0.5
Oxygen
25
TGA Baseline test with Coal Baseline tests were conducted with Loy Yang (without oxygen carrier) by heating up to 950°C in air. Fig. 2 show the weight loss and reaction rate profile at different reaction temperature of Loy Yang coal.
0.0035 Weight Loss (%) Reactivity (S-1)
100
0.0030
0.0025
0.0020
60
0.0015 40
Reactivity (S-1)
Weight Loss (%)
80
0.0010 20 0.0005
0
0.0000 0
200
400
600
800
Reaction Temperature (0C)
Fig. 2. TGA test of Loy Yang in air. A peak was observed at 830°C for Loy Yang coal in case of combustion in air. Ash content was about 2% which means 98% of the weight loss was due to combustion in air. It is evident from Fig. 2 that coal volatilization and combustion is initiated at 250°C and proceeds until 850°C. After 860°C the weight loss is constant for Loy Yang coal. A relatively smaller peak is observed for at around 100°C which indicates the moisture release from coal. Multi cycle redox experiments using metal oxides and coal A five cycle TGA test was conducted with Fe2O3-Loy Yang coal mixtures in N2-CO2 environment to evaluate the coal combustion and metal re-oxidation process. After each reoxidation process at 950°C, reacted metal oxide/ash was mixed with same amount of coal used in previous cycles. Ash could not be separated after each cycle and hence accumulated TGA crucible. The result of redox cycles in CO2 gas are shown in Fig. 3 with Fe2O3. The extent of coal combustion and re-oxidation of metal oxides, as shown by the weight loss and weight gain respectively, during TGA tests show a small decrease for the Fe2O3-coal mixture. However, NiO shows progressively less mass loss and gain over multiple cycle operation. More over, during experiments, no mass loss of Fe2O3 particles was observed from initial value with each cycle. However, even with ash accumulation in TGA crucible, the percentage of combustion at fifth cycle was 89% for Fe2O3 oxygen carrier.
105 Reduction and Oxidation of Fe2O3 and Loy Yang
Weight Loss (%)
100
95
90
85
80
75 0
200
400
600
800
1000
1200
Time (Min)
Fig. 3. Five (5) cycles TGA redox test of Loy Yang coal with Fe2O3 in N2-CO2 environment. The maximum reactivity during each reduction and re-oxidation of oxygen carrier is plotted in Fig. 4 (A)-(B) for Fe2O3. The reduction reactivity is maximum at second cycle for the oxygen carrier tested and shows a decrease in trend in consecutive cycles. This can be attributed to the fact that the oxygen carrier particles give off oxygen more easily for combustion at second cycle due to the structural changes during operation in first cycle. But during progressive cycles the continuous ash accumulation in the TGA crucible may restrict the contact between fresh coal and the re-oxidized metal oxide. One of the major focuses of this study will be how to avoid the ash interference to fresh coal-oxygen carrier interaction in a practical system. Detailed investigation pertaining to this will be addressed in future work. The maximum reactivity in each re-oxidation cycle for Fe2O3 shows maximum value at second cycle similar to maximum reduction reactivity curve. But after that reactivity is almost constant as ash accumulation does not play any role during reaction of oxygen carrier and fresh air. 0.0060
0.0112
0.0056
-1
Reactivity (S )
Reactivity (S-1)
0.0108
0.0104
0.0052
0.0048
0.0044
0.0100
0.0040
0.0036
0.0096 0
1
2
3
4
5
6
0
1
2
(A)
3
4
5
6
Cycle
Cycle
(B)
Fig. 4. Maximum Reactivity Comparison of Fe2O3 with Loy Yang coal during Redox in N2-CO2 environment (A) Reduction (B) Oxidation.
SEM Images of fresh and used oxygen carriers Figure 5 (A) - (D) shows the SEM images of the fresh and reacted Fe2O3 oxygen carrier particles in CO2 gas composition. The images A and C are at lower magnification (1500X) for fresh and used Fe2O3 respectively whereas B and D are at higher magnification of 6000X for the same. The edge of the fresh particle is rougher as compared to the used particle that can be seen from figures A and C. The crystalline structure of fresh particle is absent from the used one rather used particle seems to be highly agglomerated and increased in size as compared to the fresh one. The fresh particle is comprised of loose smaller particles and with less porosity as can be seen from figure (B). However, the porosity increased in case of the used particle (figure D) due to oxygen transfer in multicycle operation. The melting point of Fe2O3 is about 1500°C and the temperature range selected for experiment was 950°C. So severe sintering of oxygen carrier is not expected and absent from these images. But only possibility is that at this temperature range ash particles can melt and stick with metal oxides. However Loy Yang coal has only 2% ash content (Figure 2). There is no evidence of agglomeration with ash after 5 cycle operation from images. This is further explained through EDX semi quantitative analysis.
(A)
(B)
(C)
(D)
Fig. 5. SEM Images of fresh & used Fe2O3 (A) & (B) – Fresh Fe2O3, (C) & (D) – Used Fe2O3.
EDX and Semi-quantitative component analysis of the fresh and used oxygen carrier The EDX analysis provided semi-quantitative weight composition of the major elements detected on the surface of the fresh and used oxygen carrier particles. For the fresh particles of Fe2O3, Fe and O are detected and the weight percentages are 18.18 and 81.82 respectively as shown in figure 6 (C). For used Fe2O3 particles (after five redox cycles) these weight percentages are 17.76 and 82.24 as can be seen from Figure 6 (D). This indicates that there is little ash deposition on the used particles of Fe2O3 under the conditions of the experiments. Thus there is no appreciable mass loss of Fe2O3 during the five redox cycles, a fact backed up by experimental observations (Figure 3).
(A)
(B)
(C)
(D) Fig. 6. SEM – EDX analysis of the oxygen carrier fresh and used particles
4. Conclusions The in-situ CLC/G with a low-ash Victorian brown coal was investigated in a TGA using Fe2O3 as oxygen carriers. CO2 was used as gasification agent. Fe2O3 showed a high coal combustion percentage of more than 90% under CO2 gas in a single cycle operation. No agglomeration between ash and Fe2O3 particles was observed. No mass loss of Fe2O3 was observed for the duration of the experiments. However, at 5th cycle the percentage of combustion achieved by Fe2O3 was 89% in CO2 environment. Cycle 2 showed maximum reactivity (during reduction) with a decreasing trend during the subsequent cycles. Though these initial experiments did not reveal much agglomeration between ash and oxygen carrier, longer duration experiments are required to explore this issue further. REFERENCES Adanez, J., Cuadrat, A., Abad, A., Gayan, P., Diego, L. F. de., & Garcia-Labiano, F., “Ilmenite Activation during Consecutive Redox Cycles in Chemical-Looping Combustion”, Energy & Fuels, 24, 1402-1413, 2010. Berguerand, N., & Lyngfelt, A., “Chemical Looping Combustion of Petroleum Coke Using Ilmenite in a 10 kWth Unit-High Temperature Operation”, Energy & Fuels, 23, 5257-5268, 2009. Berguerand, N., & Lyngfelt, A., “Design and Operation of a 10 kWth chemical-looping combustor for solid fuels – Testing with South African Coal”, Fuel, 87, 2713-2726, 2008. Cao, Y., Casenas, B., & Pan, W-P., “Investigation of Chemical Looping Combustion by Solid Fuels. 2. Redox Reaction Kinetics and Product Characterization with Coal, Biomass, and Solid Waste as Solid Fuels and CuO as an Oxygen Carrier”, Energy & Fuels, 20, 1845-1854, 2006. Dennis, J. S., & Scott, S. A., “In Situ Gasification of a Lignite Coal and CO2 Separation using Chemical Looping with a Cu-based Oxygen Carrier”, Fuel, Article in Press, 2009.
Diego, L. F. de., Garcia-Labiano, F., Adanez, J., Gayan, P., Abad, A., Corbella, B. M., & Palacios, J. M., “Development of Cu-based Oxygen Carriers for ChemicalLooping Combustion”, Fuel, 83, 1749-1757, 2004. Leion, H., Mattisson, T., & Lyngfelt, A., “Solid Fuels in Chemical Looping Combustion”, International Journal of Green House Gas Control, 2, 180-193, 2008.
Lyngfelt, A., Johansson, M., & Mattisson, T., “Chemical-Looping Combustion – Status Development”, 9th International Conference on Circulating Fluidized Beds (CFB-9), May 13-16, 2008, Hamburg, Germany. Lyon, R. K., & Cole, J. A., “Unmixed Combustion – An Alternative to Fire”, Combustion and Flame, 121, 249-261, 2000.
Rubel, A., Liu, K., Neathery, J., & Taulbee, D., “Oxygen Carriers for Chemical Looping Combustion of Solid Fuels”, Fuel, 88, 876-884, 2009. Siriwardane, R., Tian, H., Richards, G., Simonyi, T., & Poston, J., “Chemical-Looping Combustion of Coal with Metal Oxide Oxygen Carriers”, Energy & Fuels, 23, 3885-3892, 2009.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11-14, 2010 Program Topic: Underground Coal Gasification (UCG) BLOODWOOD CREEK UCG PILOT 2008 – 2010 Presenting author Dr Cliff Mallett Technical Director Carbon Energy Limited Level 12, 301 Coronation Drive, Milton QLD 4064/ PO Box 2118 Toowong DC, QLD 4066 AUSTRALIA +61419753719 [email protected]
Co-author Burl E. Davis UCG Process Consultant Carbon Energy Limited 245 Lynn Ann Drive New Kensington, PA 15068 412-793-6058 [email protected] EXTENDED ABSTRACT BLOODWOOD CREEK UCG PILOT 2008 - 2010 Carbon Energy Limited is a listed Australian public company focused on underground coal gasification. It uses technology developed by CSIRO, the Australian Government scientific and industrial research organisation. Carbon Energy has been producing syngas from its pilot UCG plant at Bloodwood Creek in southeast Queensland since October 2008. It is completing the construction of a 5MW power station to be fueled with UCG syngas. It is developing a second UCG project near Valdivia in southern Chile. Since its formation in 2006, Carbon Energy has worked to bring a small demonstration project into production. In its first year it explored its coal permits and selected a site for a UCG trial. After another year of site investigations and design development, a UCG panel and associated surface plant was constructed, and ignited in October 2008. During a 100 day experimental period the UCG reactor was operated under a variety of conditions and injection gas types including air, oxygen and steam. Following this the reactor was operated in low flow mode on air injection, through 2009 and 2010. Decommissioning of panel 1 and clean up were commenced in August 2010.
The Bloodwood Creek site is located approximately 200km west of Brisbane in the Surat Basin. The region has limited rural industry developments, mainly pastoral, native forest or feedlots with soils unsuitable for intensive agriculture. Groundwater is unsuitable for human or animal consumption or for irrigation. The UCG pilot panel and surface facilities are constructed in a cleared paddock beside Bloodwood Creek. Underground operations are surrounded by three series of monitoring wells. The first set of operational monitoring wells are located within 10m of the UCG reactor and panel. The second set of sentinel wells are located approximately 50m from the UCG reactor. The third set of monitoring wells are environmental compliance wells located on the boundary of the permitted UCG pilot area which is approximately 1km2in extent. Carbon Energy has developed a UCG panel design that can produce 1PJ of syngas per year, and will operate for 4-5 years. The panel is defined by two horizontal boreholes 30m apart, in the base of the coal seam, that travel in the seam for 5-600m. These horizontal boreholes turn in and intersect a vertical ignition well. The coal is ignited at the base of the ignition well, and after the burn is established, this well is sealed and air/oxygen is injected down one of the horizontal wells, and the product gas is extracted from the other horizontal well. Hot gases move from the injection well to the product well across a face of coal, that is progressively gasified, causing the face to retreat as in a conventional longwall mine. At Bloodwood Creek, the coal is 200m deep and the seam is around 10m thick. The pilot was commenced on the 8th October 2008, initially with air injection. In January 2009 the injection gases were changed to steam and oxygen, and an experimental program carried out with these gases throughout January. Following these experiments the panel continued to operate with low flow air injection. On Day 140 of the trial the UCG cavity grew to engulf monitoring well No. 6, which had been placed in the centre of the panel. This confirmed the position of the reactor and provided data on ground conditions. Temperature sensors had been installed in Well 6 at points 60m above the coal, two at 12m above the coal, and within the coal. These showed that sensors 12m above the coal started heating 10 days before burn through, and continued to heat until they both failed at around 100oC, 10 - 12 days after the burn through. The sensor 60m above the coal has shown no heating. The sensor in the coal was very slightly affected the day before burn through, but showed no significant rise in temperature until a few hours before it was consumed by the face advance. The rate of face advance appears to be faster than the propagation of heat in front of the face, so there is no significant heating in the coal in advance of the cavity. Temperature from a cross-section through the reactor cavity shows that heat migration is mainly in the roof above the reactor, but heating does not reach to 60m above the coal. Temperature sensors in the coal at 10m from the reactor cavity showed no heating for almost a year. The reactor grew to very close to the 1 series of monitoring wells placed to the west of the reactor, and sensor 1M gradually heated from ambient of 26 oC to 70 oC between July 2009 and August 2010. Well 1L which was cased through the coal seam showed heat effects that indicated that the margin of the reactor cavity had affected the
casing and monitoring equipment. Well 2M, 10m to the east of the reactor cavity showed a temperature increase of less than 2 oC over the 22 months of operation of Panel 1. A 3D seismic survey was conducted to locate faults in the coal and allow the design of UCG panel layouts. The coal seam is relatively flat with faults of 20 to 40m throw spaced at approximately 1km intervals. Experimental processing of the seismic data was carried out by Peters et al 2010 to see if the UCG cavity could be identified in the seismic records. This found that the RMS amplitude showed an anomaly centered on the UCG cavity. While the anomaly was centered over the cavity, neither its outline or extent precisely coincided with the position of the cavity based on borehole location and the Well 6 burn through data. The results are very encouraging however, and represent the first example of geophysics pinpointing a UCG reactor, and it is hoped that further analyses will refine the predictive capacity of this method. The key operating requirement to safely operate a UCG reactor is the control of the reactor pressure to less than surrounding hydrostatic pressure. This allows the surrounding water to constrain reaction byproducts and prevents the escape of potential pollutants. Water pressures are continually monitored in the wells around the reactor to ensure that these conditions are maintained. If reactor pressure is allowed to fall too low, excess water runs into the reactor, and it is possible to create a draw down cone of depression around the reactor, that cannot be recharged quickly enough by regional recharge. A balance is required that allows sufficient water to flow into the reactor at all levels to prevent escape from the reactor, but to let only enough water flow in that will not depress the hydrostatic pressure surrounding the reactor. UCG is new as a commercial operation and is yet to achieve acceptance from the public or regulators. A recent event in Queensland has illustrated the potential for problems when other stakeholders are not fully informed on UCG activities. A UCG company, Cougar Energy, had commenced a pilot burn near Kingaroy in SE Queensland, but discontinued the trial for operational reasons after a short period. A sample from their monitoring wells provided a high contaminant level result which was subsequently established as a laboratory error. The original result became known and used by local activists to create public concern and the State Government shut down the UCG operation and constrained other UCG operators far from Cougars operations. After a month of extensive testing around all three UCG operations, the Government announced that it could find no indications of anything associated with the UCG operations that were a threat to people or stock. However the damage to the industry has been done as it is now associated with poisoning of farm water supplies, despite all the Government evidence that it is not. Carbon Energy is aiming to demonstrate consistent syngas production that can fuel a commercial 5MW power station. This can be done with one Carbon Energy UCG module, and a power station is nearing completion at the Bloodwood Creek site. This power station is using reciprocating engines fuelled with syngas, and will feed the electricity into the local electricity network.
Future plans are to increase power generation in stages, next with a 25MW unit and subsequently up to 300MW. The highest value can be extracted from UCG syngas through using it as a feedstock for chemicals, and it is planned to ultimately construct an ammonia and synthetic natural gas plant at Bloodwood Creek. UCG is an attractive development option for Queensland’s abundant coal deposits as it extracts most energy from the in situ coal resource, and creates rural industries in the on site conversion of syngas into high products such as fertilizers and explosives. Coal seam gas competes for access to the coal resource, but can extract only one twentieth of the energy compared with UCG operations. Although UCG is still in its infancy as a commercial venture in western economies, it holds the promise of much higher recovery of in situ resources, it avoids the safety and environmental problems of conventional coal mines, and provides the energy in a conveniently utilized form for a range of applications such as power, chemicals and liquid fuels, and allows for economic capture of carbon dioxide, an essential for a future clean coal future.
Manuscript Not AVAILABLE
Studies on Gasification of Turkish Lignite via Underground Coal Gasification Project Director Şahika Yürek , M.Sc.Mining Engineer Kıvanç Het, Directorate of Turkish Coal Enterprises (TKİ), Yenimahalle, Ankara, Turkey [email protected], [email protected]
ABSTRACT It is known that with nearly 12.3 billion tons of reserve coal, Turkey produces nearly 22% of its electricity needs via the burning of these reserve fuels. TKI, Turkey’s largest mining organization and the sixteenth largest national company, owns a nearly 2.5 billion ton share of this reserve. TKI, engaged in extracting coal from mines for heating and thermal purposes, is also engaged in studies directed towards utilizing coal for alternative purposes, following the latest technological advances in the world. One of the spearheading studies is geared towards obtaining synthetic gas via the underground gasification of coal. The process of underground coal gasification is an operation in which steam, composed of air and water, is directed with the assistance of injection wells into underground coal deposits and the resulting gas from this reaction is directed to the surface by conduction channels. This technique aims to render feasible what has been a difficult and costly means of production for accessing reserves in the sea and at great depths. Furthermore, this technique would render possible the above-ground processing of obtained gas into H2, CH4, NH3, CO, and CO2 for a variety of applications. However, these are not the only advantages of underground gasification of coal; this process also aims to contain CO2, one of the key factors in the level of greenhouse gases, within the areas from which coal is extracted from beneath the surface of the earth. It has been widely accepted that the process of underground coal gasification is a means of clean coal production which minimizes environmental impacts with low costs and high productivity. Within the scope of these studies TKI has carried on contacts with both local and foreign firms, and the most rigorous study in this area has been conducted under the guidance of the leading expert in the field, the American firm Lawrence Livermore Laboratories. If a suitable site replete with the necessary conditions for an experimental initial run involving the underground gasification of coal in Turkey is identified by the advising firm, then plans will be laid for the undertaking of larger scale underground gas production domestically. In conclusion, without a doubt this technique could once again bring to the fore those lignite coal resources of ours which are currently unexploited, which entails the financial
inactivity of those sites. Within the scope of Clean Coal Technology, this technology would be an important step in ensuring a prudential means of meeting of our future energy needs.
Underground Coal Gasification and Applicability to Thrace Basin Lignite in Turkey
Ayşe Yıldırım, Advisor, Turkish Petroleum Company [email protected] Serdar Doğan, Chemical Engineer, Turkish Petroleum Company sedoğ[email protected]
Key Words: UCG, IGCC, Gasification
Abstract As it is known that, hydrocarbon reserves have been declined rapidly and with other problems for energy demand, there have been expanded and renewed interests in new alternative technologies in worldwide. Underground Coal Gasification (UCG) is the one of those technologies for new energy sources. UCG is a gasification process carried on low calorific value, non mined or no minable coal seams due to the geological conditions (high fracture frequencies, volcanic, complex storage/tectonic structures). UCG process converts coal in situ into product gas (syngas) by using oxygen/ steam mixture (air, enriched air, oxygen/water and carbon dioxide/oxygen). Here the coal beds react as a chemical reactor, thus gasification process is maintained underground rather than conventional gasification methods. In this process; coal, steam and oxygen are brought together to the combustion temperature for coal by adjusting the amount of oxygen carefully, the coal is not completely burned but decomposed chemically. The process is a partial oxidation rather than combustion. The resulting mixture (carbon monoxide, hydrogen, carbon dioxide, methane) is UCG gas (syngas) can be used Integrated Gasification Combined Cycle (IGCC) configuration as a supplement and substitute fuel for electricity generation and chemical synthesis resulting in manufacturing of synthetic liquid fuel or chemicals by FisherTropsch process. Turkish Petroleum Company (TPAO), has a vision being an energy company and taking in the commission to assess the low calorific value lignite reserves in Turkey by clean coal technologies under the roof of Ministry of Energy and Natural Resources (ETKB) conjunction with General Directorate of Mineral Research and Exploration (MTA), Turkish Coal Enterprises (TKI), Electric Power Resources Survey and Development Administration (EIEI), Electricity Generation CO.INC (EUAS), ETI MINE Works General Management (ETI Maden), Turkish Hard Coal Enterprises (TTK) and General Directorate of Mining Affairs (MİGM) . In this review, geological, geophysical and chemical studies have been done on the lignite beds, which were determined during natural gas drilling in Thrace Basin in Turkey and their applicability is being discussed to UCG processes.
Introduction The defining feature of global energy markets remains high and volatile prices, reflecting a tight balance of supply and demand. This has put issues such as energy security, energy trade and alternative energies at the forefront of the political agenda worldwide. Energy consumption growth slowed in 2007 compared with 2006, it was still above 10-year average for the fifth consecutive year. World primary energy consumption increased by 2.4% in 2007-down from 2.7% in 2006, but still the fifth consecutive year of above-average growth. Globally, oil consumption rose, but at the weakest rate of all the fossil fuels, reflecting the pressure from high prices. The oil price has been on an upward path for more than six years now. Global oil consumption grew by 1.1% in 2007. Natural gas consumption rose by3.1 in 2007, slightly above the 10-year average. Coal, seen as affordable and locally produced in many parts of the world, was the fastest growing fossil fuel for the fifth year in a row. Renewable energy remains a small share of total global energy use, but most renewable sources experienced rapid growth in 2007. Ethanol output rose by 27.8%. Global capacity for wind and solar electricity generation
grew broadly in line with average of 28.5% and 37%, respectively. The ongoing growth in fossil fuel consumption suggests that global carbon dioxide emissions are still rising. Turkey primary energy consumption increased by 2.9% in 2009. Turkey has limited reserves of oil and natural gas, but proven reserves of lignite in the order of 8.4 billion tones. Combustible renewable, especially wood and the country’s water sources are the important indigenous energy sources. Turkey should use its domestic lignite sources in industry and power generation. In this paper, it is recommended that UCG will be a new technology for new energy and power generation source for Turkey.
Underground Coal Gasification Process Description Underground coal gasification converts coal in-situ into a gaseous product, commonly known as synthesis gas or syngas through the same chemical reactions that occur in surface gasifies. Gasification converts hydrocarbons in to a synthesis gas at elevated pressures and temperatures and can be used to create many products (electric power, chemical feedstock, liquid fuels, hydrogen). Gasification provides numerous opportunities for pollution control, especially with respect to emission of sulfur, nitrous oxides and mercury. UCG could increase the coal resource available for utilization enormously by gasifying otherwise non minable deep or thin coal reserves. In the process gas is produced and extracted through wells drilled down into the coal seam, to inject air or oxygen to combust the coal in-situ, and to produce the coal gas to the surface for further processing, transport or utilization. The process relies on the natural permeability of the coal seam to transmit gases to and from the combustion zone, or on enhanced permeability created through reversed combustion, an in-seam channel or hydro-fracturing.
Underground Coal Gasification
The overall chemistry underlying coal gasification processes is well understood. There are two main chemical reactions. 1.Oxidation Almost all the oxygen from the blast is depleted upon contact with the coal Oxidation zone should be kept as short as possible Oxidation zone should spread out laterally from the injection point C + O2 → CO2
- 393.5 kJ mole-1
C + ½O2 → CO
- 110.5 kJ mole-1
CO + ½O2 → CO2
- 283
kJ mole-1
Both of these reactions are exothermic (generates heat) 2. Reduction The hot gas leaves oxidation zone and passes towards the reduction zone. into contact with hot char. C + CO2 ↔ 2CO
CO2 is reduced to CO as it comes
+172.3 kJ mole-1
With the introduction of steam, H2 is also formed C + H2O ↔ CO +H2
+131.4kJ mole-1
Both of these reactions are endothermic (consumes heat) Too much water (groundwater ingress) can quench the system
Key Considerations for UCG 1.
Thickness of coal seam >1m – any thickness (in theory)
2.
Depth of coal seam >300m
3.
Rank of the coal
4.
Proximity to fault structures >200-500m
5.
Proximity to built up areas >500m (In UK)
6.
Proximity to mines both active and historic >500m
Why Consider Underground Coal Gasification UCG has numerous advantages over conventional underground or strip mining and surface gasification:
Reducing operating costs, surface damage and eliminating mine safety issues such as mine collapse,
Coals that are un minable (too deep, low grade, thin seams) are exploitable by UCG, thereby greatly increasing domestic resource availability,
No surface gasification systems are needed, hence capital costs are substantially reduced,
No coal is transported at the surface, reducing cost, emissions, local foot print associated with coal shipping,
Most of the ash in the coal stays underground, thereby avoiding the need for excessive gas clean-up, and the environmental issues associated with fly ash waste stored at the surface,
There is no production of some criteria pollutants (SOx, NOx) and many other pollutants (mercury, particulates, sulfur species) are greatly reduced in volume and easier to handle.
UCG eliminates much of the energy waste associated with moving waste as well as usable product from the ground to the surface UCG compared with conventional mining combined with surface combustion, produces less greenhouse gas and has advantages for geologic carbon storage.
Potential Limitations and Concerns for UCG Even though UCG has a number of advantages, the technology is not perfect and has several limitations.
UCG can have significant environmental consequences: aquifer contamination, and ground subsidence,
While UCG may be technically feasible for many coal resources, the number of deposits that are suitable may be much more limited because some may have geologic and hydrologic features that increase environmental risks to unacceptable levels,
UCG operations cannot be controlled to the same extent as surface gasifies. Rate of water influx, the distribution of reactants in the gasification zone, growth rate of cavity are the important process variables.
The economics of UCG has major uncertainties,
UCG is inherently an unsteady-state process, flow rate and the value of the product gas will vary over time.
Comparatıve Costs and Fınance for UCG
Large-scale UCG with power generation (300MWe) undertaken remote from the gasification site has a generation cost comparable to, and possibly less than, Integrated Gasification Combined Cycle (IGCC) technologies.
However, small scale developments (~50MWe) are not likely to be economically viable as standalone project
UCG with CCS has a power generation and capture cost which lies between two estimates for IGCC with CCS and is comparable with gas turbine combined cycle (GTCC) with CCS (assuming postgasification capture).
Turkey Lignites Almost half of Turkey’s energy production is represented by coal (43% lignite). The lignite deposits are widespread with reserve estimates of more than 8 million metric tons. Turkish Coal/Lignite reserve is shown in the following graph.
Turkish Coal/Lignite Reserves
Potential Issues of Shallow Turkish Lignite for UCG
Almost of all the Turkish lignite is Tertiary age.
Shallow lignite is likely to cause surface subsidence which may lead to gas escape.
Lignite with a moisture content >40% may be considered unsuitable due to quenching of oxidation zone.
Shallow lignite may be situated close to valuable groundwater bodies and UCG may be considered too marginal due to groundwater contamination issues.
Careful selection of lignite reserve necessary to avoid coal with low calorific value.
However- these issues may be reduced or mitigated by careful site selection.
Applicability of UCG to Thrace Basin Lignite There are seven onshore and four major offshore basins in Turkey. The onshore ones are called SE Turkey (Anatolia) Basin, Thrace Basin, Adana Basin, Tuz Gölü Basin, East Anatolia Basin (including several subbasins) and the onshore Black Sea Basin (Zonguldak and Sinop). The offshore basins can be named as; Black Sea, Marmara Sea, Aegean and Mediterranean Sea. The most active onshore basins are, as far as the exploration and production concern; firstly SE Turkey (Anatolian) and Thrace, secondly Adana and Tuz Gölü Basins, whereas SE Turkey (Anatolian) Basin is known to be the oil prone and the Thrace is the gas prone ones. All the oil and gas fields are located in these two basins where the total production reaches 50M bbl for oil and 60MM SCFD for gas. The European side of Turkey is called the Thrace Basin. It is the largest and thickest Tertiary sedimentary basin in Turkey. The basin is triangular shaped, trends WNW-ESE and was formed by extension in late Middle Eocene to latest Oligocene times. This gas prone basin is located northwest of Istanbul, Turkey. In basin About 9,000 meters of Eocene-Miocene marine clastics and continental sediments were deposited. Several basement faults were reactivated and underwent strike-slip motion in late Miocene times where the famous North Anatolian Fault bounds the southern part of the basin. The Tertiary sedimentary succession, overlying Paleozoic to Mesozoic metamorphic basement, comprises interbeded fine to coarse grained clastics from a variety of depositional environments, turbidities, muddy carbonates with local reef developments, river channels and tuffs. Thrace Basin is one of the most important basins because of coal and hydrocarbon potential. The basin is surrounded on the north by the Istranca, on the west by the Rodop, and on the south by the Menderes Massifs. Coal explorations have been conducted by General Directorate of Mineral Research and Exploration (MTA). important coal formations in the basin arewithin Danişmen Formation of Oligocene age. Danişmen Formation has a lithology including gray-green claystone, sandstone, conglomerate, tuff and lignite, and was named as lignitic sandstones in the region by fırst studies. The lignite occurences have mostly been developed at lagoonal, deltaic, flood plain and lacustrine marshes but economical coal occurrences take place at delta plain, lagoonal and lacustrine marshes. Investigated samples show that the huminite group, the coals are seen to be abundant of huminite maceral group as well as gelinite maceral. From the vitrinite reflection values, the coals seem to be classified as sub bituminous in rank and the coals were determined to have deposited in limnic-paralic environments. There are lignite occurrences on foothills of Istranca Mountains nort of the Thrace Basin, and these are mentioned as Saray (Edirköy, Safaalan, Küçükyoncalı), Vize (Topçuköy), Kırklareli and Demirhanlı fıelds in general, and Keşan, Malkara, Uzunköprü and Meriç fields in the south. Lignites exposed at N and S of the basin gradually deepen toward center of the basin, and reach to 11 000 meters of sedimentary sequence in central parts of the basin and extend to depths below 600m.
Geological Map of Thrace Basin is shown in the following graph.
GEOLOGİCAL MAP of THRACE BASIN
VAKIFLAR
Chemical and Geological Properties of the lignite reserves in Thrace Basin are:
The lignites are in Tertiary age,
The coal thickness is between 30-50m,
The average coal seam depth is 200-400m,
Lignite are almost with a moisture content >40 %.
Lignite reserves are in low calorific value.
In Thrace Basin with these properties the UCG process might be done under following conditions:
A geological/hydro geological model should be build for an existing resource,
The size of the deposit should be in excess of 100 million tonnes
If all parameters are successful, pilot site selection and plant should be considered.
References: 1.
Aiman, W.R., R.J. Cena, R.W. Hill, C.B. Thorsness, 1980, Highlights of the LLL Hoe Creek No.3 Underground Coal Gasification Experiment., Lawrence Livermore National Laboratory, Livermore, CA. UCRL-83768.
2.
Ambrozic, T, and Turk G., 2003, Prediction of subsidence due to underground mining by artificial neural networks Computers an&Geosciences 29,627-737
3.
Blinderman, M.S., and Jones, R.M., 2002, The Chincchilla IGCC Project to date: UCG and Environment, 2002 Gasification TECHNOLOGİES Conference, San Francisco, USA, October 27-30, 2002.
4.
Şengüler, İ. (2003) Öz Kaynaklarımız İçinde Linyitin Yeri ve Önemi. Türkiye 9. Enerji Kongresi Bildiriler Kitabı, 59-67, İstanbul.
5.
Uysal, B.Z., (2008) Temiz Kömür Teknolojileri. Türkiye 16. Kömür Kongresi, Bildiriler Kitabı, 335340, Zonguldak.
CO2-Reduction through Biomass co-firing in Coal Fired Power Plants Dr. Roland Jeschke*, Dr. Klaus-Dieter Tigges*, Dr. Alfred Gwosdz*, Alfons Leisse* *
Hitachi Power Europe GmbH, Schifferstraße 80, 47059, Duisburg, Germany
1
Introduction
Facing the threatening background of global warming the impetus for more and more activities to reduce CO2 emissions is given. Biomass co-firing, being regarded as CO2 neutral, has a potential to reduce CO2 emissions generated by mankind. This particularly applies for power generation based on fossil fuels. Besides that operators of utility boilers are always interested in reducing production cost with a special focus on fuels. Biomass is deployed increasingly in power stations as wood pellets, chips and others. Moreover most of the new power stations being under construction right now envisage biomass combustion in the short or mid term. Hitachi Power Europe has been involved in co-firing for a couple of years starting with grinding of wood pellets. In consequence the development of biomass firing technology was extended to cover the entire value chain from storing and grinding up to firing a wide range of biomass products. This technology can be applied for both lignite and hard coal fired steam generators. It is based on HPE’s well proven firing technology for hard coal and lignite and was refined under full scale conditions in a 35 MW test facility. The combustion tests provided a comprehensive set of operating data to evaluate the technology and validate the combustion models especially adapted to biomass firing conditions. This paper outlines the biomass technology for co-firing woody biomass in hard coal or lignite fired steam generators up to an amount of 100% and shows possible applications taking different biomass features into account. Results of combustion tests are described and calculated and measured data are shown.
2
Overview
Considering co-firing solid biomass in coal-fired steam generators an overview containing the relevant aspects of secondary fuel storage, grinding, feeding, injection or admixture as well as the burner technology is depicted in Figure 1. Depending on the chosen co-firing design the feeding or injection type may vary. This is indicated by the multiple arrows starting from the storage icon and pointing to various positions between the grinding and the burner. The options for co-firing design are depending on the main fuel used, because this determines the boiler design of the respective power plant. The boiler design itself determines which biomass fuel is most appropriate for being co-fired, because it affects the residence time of the particles in the burner chamber. Furthermore the change of burner and/or mill configuration from design coal case to biomass configuration may be complicated. The choice depends on the design and the dimensions of the existing boiler on the one hand and the physical and chemical characteristics of the biomass fuel on the other hand.
Page 1 of 12
Figure 1
Overview of relevant aspects for co-firing biomass in coal fired boilers
In the following Figure 2 some of the options for feeding biomass into a coal-fired power plant are shown. Option 1. (BL) stands for direct injection via burner for a complete burner level. Option 2. shows the implementation of separate burners. In option 3. (B) the biomass is admixed to the pulverized fuel piping close to the burner. Option 4. (S) is realised through the admixture on to the conveyer before the mill and option 5. (S) accounts for pre-conditioning e.g. pre-crushing, carbonisation,…
Figure 2
Feeding/Injection types for biomass co-firing
The bracketed (S) after options 4. and 5. stands for ‘small’, indicating that the selection type may also be determined by the amount of biomass to be co-fired. Examples: for co-firing less than 10% of full thermal load the options 4. and eventually 5. can be applied. If one needs to get up to 100% of full thermal load on a burner level option 1. has to be chosen. With this option the possible biomass heat input depends e.g. on burner geometry, because the used burner type may not be capable / suitable of firing biomass fuel with considered amounts. An amount of 10-100% of full thermal load on a single burner needs option 3. (B). As mentioned before the options applicable depend on the fuel characteristics of the chosen biomass fuel as well. A high moisture content (e.g. forest residue), a high ash content (e.g. sewage sludge), a high content of corrosive elements (e.g. straw), the existence of elements that deactivate the DENOX catalyst (e.g. phosphorus) or the tendency to increase slagging and fouling (e.g. via ash fusion temperature) are important restricting factors. Page 2 of 12
This paper deals with wood pellets (WP), wood chips and sawdust. Concerning the restricting factors depending on the fuel properties these three avoid some of the limiting properties discussed above and are therefore suitable for co-firing in a large coal-fired steam generator.
3
DS®-Burner technology
The DS®-Burner technology from Hitachi Power Europe GmbH, used for hard-coal fired steam generators, is able to burn up to 100% biomass. The burner configuration as well as the mill configuration of the MPS mill is easily adaptable to a biomass mode. The DS®-Burner is of swirl staged type and has a concentrically design. A basic set up is shown in Figure 3. DS®-Burners mainly consist of the following items: • Ignition equipment containing light oil gun and high energy sparking rod. • Core air tube. • Primary air tube with integrated fuel nozzle, adjustable swirl vanes and impact elbow at the inlet. • Secondary air tube with adjustable swirl vanes and air deflection throat. • Tertiary air nozzle with adjustable swirl vanes. • SA / TA wind box with connection to the air supply.
Figure 3
DS®-Burner basic burner set-up
DS® burners are focussed on the local definition of the ignition point considering pyrolysis and oxidation processes. Oil gun and sparking rod are guided inside the core air tube and the core air tube is centrically arranged in the pulverized fuel tube. The p.f. tube is connected with the coal line through an impact elbow at the inlet. The impact elbow and the impact table located on the core air tube take care for highly uniform distribution of the pulverized coal in the primary air tube. Page 3 of 12
Function of DS®-Burner Adjustable swirl vanes generate a rotary primary air/coal flow resulting in particle concentration at the outer circumference of the pulverised fuel tube. Before leaving the burner the fuel particles are slowed down by collision against the inner fuel nozzle ring. As a result the flow velocity of the particles is reduced below the fuel’s characteristic flashback velocity by means of the impact on that ring. For inducing and maintaining the ignition process a sufficient amount of oxygen is available in the primary air stream only. The impact of the pulverised fuel particles on the fuel nozzle ring results in reflection and acceleration in flow direction again by diversion into the primary air flow, which discharges into the furnace. Devolatilisation of the pulverised fuel occurs in a controlled way under low-oxygen conditions, so that generation of nitrogen oxide is already severely restricted in the flame and thus the total emission is noticeably reduced. The stable swirl flow formed by the secondary- and tertiary- air which surrounds the core flame causes oxygen enrichment in the peripheral burner zone and thereby ensures an oxygen rich flue gas atmosphere near the furnace walls. Firing 100% Biomass Based on this DS® Burner experience, a special version of this burner was developed that is capable of dealing with fuels being fed in a dense phase i.e. 4-10 kg pulverized, dried fuel per kg carrier gas. This burner type deals with pulverized dried lignite and biomass and is the suitable burner for high fuel concentration in the centre. A further special feature of this burner is its extended control range. A basic set-up is shown in Figure 4. The core air tube is shortened to allow the core air being mixed with the primary air having a high dust load. This happens just before the inner fuel nozzle ring, where the particles entrained in the primary air flow are slowed down by collision with this ring. Helical baffles on the outer diameter of the core air tube support the rotation of the primary air/coal flow. Secondary and tertiary air handling is as for the DS® burner.
Figure 4
DS® T Burner basic set-up
Page 4 of 12
3.1
Testing and refinement in 35 MW test facility
A series of tests with different fuels using the DS® as well as for the DS®-T-burner were carried out at the 35 MWth test facility of the Centro Combustione Ambiente (CCA, an ANSALDO Caldaie company) in Gioa del Colle, Italy. Figure 5 shows a photo of the burning chamber. The natural circulation type boiler has a horizontally shaped combustion chamber. The burner is installed at the front wall. The combustion chamber is partially refractory lined in order to obtain similar heat transfer and flame temperatures as in a real boiler. The fuel supply can be controlled in different ways depending on whether the firing is direct or indirect. The direct system includes a coal silo, a ball mill, a rotating classifier and the PA fan. In case of indirect firing, grinded fuel is transported by a pneumatic conveyer. The flue gas is cleaned from dust by a bag house filter before leaving the chimney. Observation ports are located on the left side wall of the combustion chamber at various distances from the burner. Water-cooled probes can be introduced here into the flame perpendicular to the flow axis. In this way it is possible to measure the temperature distribution at different distances to the burner and at different distances from the centre of the flame. In the test series the concentrations of flue gas species as O2, CO and NOx as well as the release of hydrocarbons were measured in the flame. All air flow rates are measured separately and adjusted by means of control dampers. Figure 5
35 MWth combustion chamber at the Centro Combustione Ambiente (Ansaldo Caldaie company). Length app. 12 m, width app. 4.3 m, height app. 6 m (w/o hopper), ash hopper 2.5 m
The following Figure 6 clearly shows a stable flame and an early ignition very close to the burner even at a sawdust firing rate of 86.6 % for the DS®Burner. The sawdust is characterized by a Residue of 2.6 % on the 1000 micron sieve and 90.7 % on the 90 micron sieve. The stability of the biomass flame over a wide load range assures good conditions for low NOx emission figures as well as low loss on ignition (LOI) figures.
Figure 6
View of the flame of a DS® Burner burning 86.6% sawdust Page 5 of 12
3.2
Calculated data
Figure 7 and Figure 8 show the temperature field (dimension Kelvin) of a DS® – Burner operating with 50% hardcoal and 50% WP or 100% coal as fuel, respectively. The figures show similar temperature field contours so that the well proven fuel range for the DS®-Burner is extended for co-firing biomass that fulfils the fuel characteristics of WP.
Figure 7 50% Coal and 50 % Wood Pellets [K]
Figure 8
100% Coal [K]
Figure 9 and Figure 10 display the 3D temperature field comparison for 50% WP co-firing and for pure hardcoal combustion (dimension degree Celsius). As depicted, the temperature fields look similar for both cases.
Figure 9 50% Coal and 50 % Wood Pellets [°C]
Figure 10
100% Coal [°C]
As shown in the figures, the different behaviour of combustion of coal and biomass and the different fineness of both fuels on the other hand are well compensated to a large extent by the DS®-Burner characteristics. Due to the stable ignition of the fuels by the DS®-burners and the controlled combustion in the furnace, the simulations show that co-firing of biomass is admissible in a wide range. The predictions meet reality for the test facility very well. Co-combustion in utility boilers show the same trends as predicted.
Page 6 of 12
4 4.1
Grinding technology MPS mill
The MPS coal mill, usually deployed for grinding hard coal, can easily be modified into a MPS wood pellets mill, including explosion suppression and safety system for bio-fuel. Reference for this type of modification is AMER 9 in the Netherlands. The MPS wood pellets mill can be re-modified into a MPS coal mill in a couple of days, if desired. For AMER 9 PS, the quantity of WP co-firing is up to 15% of full thermal load with one of six mills, the WP net calorific value was 16.9 MJ/kg, the water content 9.6%ar and the ash content 2.12%ar. The biomass is completely fed into one burner level which means that this burner level runs on 100% biomass combustion. The modification of the MPS mill includes installation of explosion detection and suppression systems, employing relatively cold primary air, carbon monoxide detection and quick-acting isolation systems as shown in Figure 11. As an example explosion detectors mounted within the mill body are capable of triggering a set of extinguishing systems located throughout the mill. Furthermore a double wall is necessary to let the biomass particles be carried out in an annular gap providing sufficient effective lift velocities to prevent unnecessary recirculation of biomass back into the mill.
Figure 11
Sketch of MPS-mill modified for wood pellet operation
Page 7 of 12
Function of MPS mill A motor drives a grinding bowl with exchangeable grinding bowl segments via a gear. The material centrically introduced from the metering hopper onto the rotating grinding bowl through a discharge chute is carried to the grinding track by centrifugal forces and rolled over by the three fixed grinding rollers. The grinding force is a function of the mill load and is set using the hydro pneumatic grinding force system. The process of combined drying and pulverizing follows the airflow principle. The hot primary air enters the grinding chamber around the grinding bowl through the nozzle ring to take over drying and transportation of the pulverized material to the electrically driven lamella classifier (static classifier shown in Figure 11) arranged above the mill housing. Separation of the coarse grain from the fine grain takes place at this point. The dust-air flow rising from the pulverizer is deflected once it has passed the gravity screening/cleaning zone. Using a dynamic classifier a radial acting force is transmitted through the rotating lamella cage against the particles in the bearing air stream which deflects the coarse dust particles from their present flow direction and repels them outward. There, these coarse particles get into a downward directed flow field that transports them back centered into the grinding bowl for post-pulverization via the funnel-shaped oversize material return flow. The separative capacity of the dynamic lamella classifier is determined by the variable rotational frequency of the rotor. For static classifiers it is adjusted by positioning the classifier vanes. The finished dust leaves the classifier via several single dust pipes leading to the burners.
4.2
DGS mill
The DGS mill is usually deployed for grinding of lignite. Depending on the Amount of Biomass to be (co-) fired the feeding/injection can be carried out by direct feeding into the mill by adding the crushed/grinded biomass to the coal on the conveyer for the DGS mill. Up to about 5% of full thermal load biomass share may be added. DGS technology involves the raw coal and flue gases being initially put through a beater section for precrushing. There will be no further pulverisation of the wood fibres in the mill. This ensures excellent air and coal dust allocation into the beater wheel. As a result, there is often no need for any classifier something both raising the pressure balance and cutting back on energy requirements. Figure 12
DGS Mill
Page 8 of 12
4.3
Edge mill The Edge Mill is a vertical spindle mill, see Figure 13, and belongs to the group of 'Applied Force Mills'. It may be used for pre-crushing wood chips to a size, which can be fed to e.g. a DGS mill. The mill is equipped with a lower die, which, depending on the mill size, is connected to one or two drives. Above the lower die grinding rollers are arranged which are stationary and are pressed to the lower die hydraulically. When the die is set to rotation it forces the grinding rollers start crawling on top and overrunning the wood chips, which have been fed in-between. The wood chips are pressed through the stencil and are reduced to small pieces. This procedure allows the wood chips not only to be cut parallel but also transverse to the wood fibres. That is not possible with a beater mill. Figure 13
4.4
Edge Mill by Fa. Kahl
Comparison of fineness measurements
For comparison the fineness achievable for wood chips pre-crushed in an edge mill and of WP after treatment in a MPS mill is shown in the next table. These Residues taken from measurements show the need for utilizing a burn off grate for the co-firing of wood chips. [micron] 6300 4000 2000 1000 500 200 125 90 63
Chips after edge mill [%] 2.5 25.2 51.1 76.4 94.9 97.2 97.7 97.9
Pellets after MPS mill [%] 0.9 14.9 54.8 85 92 95.5 97.4
Table 1 Fineness of pre-crushed wood chips and wood pellets milled by MPS mill
5 5.1
Possible Applications for co-firing woody biomass Co-firing with hard coal
For a hard coal fired boiler, usually using roller mills and swirl burners, the biomass fineness requirement is challenging, because there is no grate in the hopper. For this case, an appropriate biomass is wood pellets. These pellets consist of compressed sawdust particles, which can be disaggregated in roller mills. Page 9 of 12
Figure 14
Wood pellets
As shown in Table 2 the ash content of the WP is rather low, the content of VM is very high and the sulphur content is low. This influences UBC, NOx and SO2 emissions directly. The slagging and fouling tendency of the WP shown beneath are different; the WP2 have low melting temperatures, so that the risk of slagging is much higher than for WP1 pellets. NCV Water Ash Volatile matter S Cl N IDT Softening Temp. Hemisph. Temp. Fluid Temp.
[MJ/kg] [%] [%ar] [% ar] [% ar] [% ar] [% ar] [°C] [°C] [°C] [°C]
WP 1 18,07 4,85 0,30 80,14 0,05 0,04 0,17 1430 1465 1470 1490
WP 2 17,46 9,02 0,86 75,51 0,05 0,03 0,13 1167 1170 1172 1177
Table 2: Exemplary properties of wood pellets
Applying option 1. (BL) shown in Figure 2 ‘direct injection via burners of a complete burner level’ to a hard coal fired boiler, i.e. 100% biomass on one burner level means running a mill totally with WP. The conversion of such a mill to enable pellet combustion involves measures and solutions for dosing to the milling system, storage, optimisation of the firing system, heat and mass balances as well as CFD calculations. This option was realized at AMER 9 PS.
Storage concept for admixture to p.f. piping For firing from 10% up to 100% biomass in a hard coal fired boiler option 3 (B) according to Figure 2 is most suitable. Admixture to pulverized fuel piping can be done e.g. by 3 way valves or by direct injection with non-return valve installed before inlet. Here is the challenging question how to store and transport the pulverized fuel to the burners. In the following example a dense phase transport to each burner is considered. The biomass is injected in a straight-lined pipe section near to the DS®-Burners. Page 10 of 12
The storage concept for this example may look as shown in Figure 15: • Pneumatic conveying of the wood dust to the burners at a high dust load (blue lines). • The carrier-air mass flow for the biomass remains constant for all loads but the dust loading changes with the variation of the quantity of wood. • Wood dust feeding to the burners is made - using rotary piston blowers, - from silos, - via dosing rotary seals, - and , - via conveying lines to the burners. • A silo and a blower supply one burner level. • The feeding into the combustion chamber is made via the DS® Burners.
Silo
Rotary seal
Piston blower
Figure 15
5.2
Mixing device
Storage concept according to Sil & Atex for dense phase transport
Co-firing with lignite
Considering a lignite fired boiler the biomass fineness requirements are not as challenging as for the hard coal case. This is mainly because of the burn off grate. Coarse particles which are not lifted by the upstream gas flow fall onto the grate and incinerate. For this case, especially concerning the less challenging fineness requirements, wood chips instead of WP can be chosen for co-firing. Despite of the afore mentioned grate the chips have to be crushed to smaller peaces before they can be burned. The resulting fineness of the comminuted chips is considerably coarser than for WP. That means the altered combustion behaviour, especially regarding burnout time, has to be taken in consideration. Wood chips, according to Figure 16, have different shape than WP. The water content is also higher and ranges usually between 15% and 50%, so the NCV can vary between 5 and 18 MJ/kg. This influences the appropriate type of mills. A suitable feedstock size for co-firing with lignite is G30 according to Ö-Norm 7133, where the main fraction is 2.8 mm to 16 mm with 60-100 % wt. The coarse fraction has dimensions of more Page 11 of 12
than 16 mm to a maximum of 85mm with maximum 20 % wt. share. The fine fraction consists of particles smaller than 2.8 mm with a maximum of 20% wt. share. Less than 4 % wt. share have dimensions smaller 1 mm. This nearly corresponds to the definition of class P16 in Prenorm CEN/TS 14961. The size of the crushed wood chips will typically be in the range of some millimetres. Therefore a sufficient residence time in the furnace must be ensured or a burn off grate has to be available, which is true for lignite fired units. Applying option 5. (S) ‘Pre-conditioning’, i.e. pre-crushing in this case, and option 4. (S) ‘Admixture before mill on conveyer’ to a lignite fired boiler allows an amount of 35% of full thermal load to be co-fired. Figure 16
Wood chips
For biomass co-firing with lignite the HPE-DGS mill is used with upstream Edge Mill for biomass pre-grinding. A defibration plant, consisting of edge mill and silos, is used for crushing wet wood chips which are stored in silos before feeding on the coal feeder that supports the DGS mill. The finished material is transported mechanically to the boiler house and is supplied to the coal feeders. Laying on top of the coal both fuels will be conveyed together to the mills. The feeder is equipped with a belt weigher to determine the coal mass introduced into the pulverizer. The mill supplies coal and biomass to the boiler. This process is well suited for wet raw material. A drying process is not necessary. On the contrary, the higher the water content, the lower the power required for grinding. As a result the energy consumption of the total process is relatively low. To ensure the wet condition of raw material the minimum water content is limited to 20%. For wood chips with water content between 20 - 30% a humidification of the woodchips is intended. Another advantage of the high water content of the finished property is that there are no problems with ATEX.
6
Outlook
The technology for grinding and co-firing biomass with coal is available. Different types of biomass can be co-fired. The amount of thermal input determines the options for grinding and co-firing that should be chosen. 100% biomass firing is possible for a complete burner level. Several options and applications were shown. It obviously makes sense to use this experience and design and built Power plants capable of firing up to 100% biomass. It could be expected that new plants to be build are able to account for a broad spectrum of biomass. Limitations arise mainly from the properties of the biomass chosen which lead to corrosion, slagging or fouling problems.
Page 12 of 12
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
VENTILATION AIR METHANE ABATEMENT AT CONSOL ENERGY’S ENLOW FORK MINE Richard A. Winschel, Director Research Services, CONSOL Energy Inc. 4000 Brownsville Rd., South Park, PA, 15129 USA [email protected], 412-854-6683 Deborah A. Kosmack and William P. Fertall, CONSOL Energy Inc. Jerry Gureghian, Green Holdings Corp.
Abstract: CONSOL Energy Inc. and Green Holdings Enlow, Inc. (a subsidiary of Green Holdings Corp.), are developing one of the largest coal mine ventilation air methane (VAM) emission abatement project in the United States at CONSOL's Enlow Fork Mine in southwestern Pennsylvania, USA. The status of the project is discussed herein.
Background and Introduction: Coal mining, and particularly coal mine ventilation air, is a major source of anthropogenic methane emissions, a greenhouse gas that is 21 times more potent than carbon dioxide (CO2). Globally, VAM emissions from coal mines amount to approximately 300 million tCO2e (metric tonnes of CO2 equivalent) each year. Oxidation of methane to carbon dioxide and water reduces its global warming potential by about 87%. The VAM abatement equipment to be installed at the mine, based on regenerative thermal oxidation (RTO) technology, will capture and destroy methane released during the mining process that would otherwise escape to the atmosphere through the mine's ventilation system. The project is designed to reduce the mine's VAM emissions by the equivalent of 190,000 tCO2e each year. It is expected that the system will be operational in early 2011. Green Holdings will supply the capital, will operate the unit and will be responsible for selling the emissions reduction credits. CONSOL will provide the ventilation air, the site and technical support.
Description of Mine and Bleeder Fan: Enlow Fork Mine is an active underground coal mine located in Washington County, Pennsylvania, USA, approximately thirty-five miles southwest of Pittsburgh. The topography is rolling hills typical of the Allegheny Plateau region of North America. Enlow Fork Mine produces approximately 11 million short tons (10 million metric tonnes) of clean coal per annum from the great Pittsburgh Seam primarily via the longwall mining method (two separate longwall systems are employed), and it is one of the largest underground coal mines in the United States. The ventilation systems in CONSOL’s longwall coal mines use “bleeder” fans to ventilate each longwall section. The bleeder fans draw and exhaust air from the mine through vertical shafts that connect the surface and the underground workings. The bleeder shafts at the mine are
drilled from the surface with a finished diameter of 6 to 8 feet (1.8 – 2.4m), and each one provides ventilation for one of the two longwall mining systems at the mine. At Enlow Fork Mine, typical bleeder fans will exhaust about 300,000 standard cubic feet per minute, scfm (510,000 standard m3 per hour) of air. This ventilation air will contain methane at a typical concentration of 0.6% to 1.0% by volume. Bleeder fans will have a typical life of six to eight years before the mining district they were designed to ventilate is permanently sealed. Figure 1 shows a typical bleeder fan layout at the Enlow Fork Mine with a proposed methane emission abatement system installed. The methane abatement system will be installed on the area where the materials removed during the drilling of the bleeder shaft have been placed and covered with earth; this area is known as the cuttings pit, and it provides a large, flat and level space for the installation in a region where flat, level space is rare.
Description of Regenerative Thermal Oxidation (RTO) System: The RTO plant is designed to safely destroy at least 96.5% of the very low concentrations of methane gas (1.4% or less) contained in high volumes (140,000 to 300,000 scfm or 240,000 to 500,000 m3 per hour) of ventilation air exhausted by underground coal mines, commonly termed ventilation air methane, or VAM. The plant design, based on proven emission control technology, is composed of three individual RTO units, each capable of processing 47,000 scfm or 80,000 m3 per hour of VAM with methane concentrations between 0.5% and 1.2%. Each RTO unit is equipped with: -
two heat-exchange ceramic media beds an oxidation chamber (with propane burners) an inlet/exhaust valve housing direct drive supply fan exhaust stack temperature and methane sensors
The RTO units are arranged in an array connected to the mine evase by a 100-foot-long (30 m) duct, which captures no more than 80% of the mine fan’s output, see Figure 1. At one end, a collection box, designed not to restrict the exhaust of the ventilation air from the mine, connects the duct to the mine evase. At the other end of the duct a damper shuts off all VAM supply to the RTO array in the event methane concentrations exceed the operational limits (0.3% to 1.4%) of the RTO units. The damper is activated by methane sensors located within the collection box. When within operational limits, the VAM is safely carried through the duct to each RTO unit with the assistance of the RTO supply fans to overcome the pressure drop caused by the heatexchange beds. The VAM enters the RTO through the first media bed to the oxidation chamber, which is preheated to 1800 °F (1000 °C) using the propane burners, which are shut off after preheating. The methane contained in the VAM oxidizes in an exothermic reaction, which, above 0.3%, generates sufficient heat to maintain the chamber at 1800 °F (1000 °C). The VAM
exits the oxidation chamber via the second media bed where the air is cooled to 480 °F (250 °C) as heat is transferred to the ceramic bed. The directional flow of VAM is alternated every two minutes to prevent overheating of the media beds and to maintain the temperature of the oxidation chamber within operational limits. Data from methane sensors located at the inlet of the RTO and in the RTO exhaust stack are used to determine the actual quantity of methane which has been destroyed and serves as the basis for claiming the carbon credits which will be issued to the project. Once operational, the RTO plant is expected to function with minimal human intervention although a modem programmed to call or page a remote service center or personnel to broadcast any upset condition.
Permitting: In order to construct, install, and operate the equipment, several permits are required from various government agencies. The authorizing agencies include local, state, and federal groups. There are five organizations that must either approve or inspect the system before it is permitted to be constructed or operated. The status of each of the permits is discussed below. The permit required at the federal level is an addendum to the Mine Ventilation Plan. It is submitted and approved through the U.S. Mine Safety and Health Administration (MSHA). A modification to the existing plan which details the safety features of the abatement system is required. This plan is also reviewed by the Pennsylvania Department of Mine Safety, which is responsible for inspecting the equipment before it is operational. The addendum was formally submittal to MSHA District 2 on August 9, 2010. The existing Mining Activities Permit, issued by the state of Pennsylvania, also required a modification because the project changes the activity on the mine site. The installation of the equipment will remain within the existing permitted area and there will be no impact to road access and water drainage, therefore the modifications to the permit were minor. On July 14, 2010, the Bituminous Coal Mining Activity Permit including the installation of a ventilation air methane regenerative thermal oxidizer on the exiting bleeder shaft was officially approved by the Pennsylvania Department of Environmental Protection (PA DEP) Bureau of Mining and Reclamation. A Plan Approval and Operating Permit from the PA DEP Bureau of Air Quality is required before the system can be operated. The focus of this permit is to evaluate the impact of the operations on air quality. Although the abatement system is an air-cleaning device that will reduce the emissions of methane entering the atmosphere, there will be some minor emissions of nitrogen oxides and carbon monoxide. The application was submitted on June 23, 2010. A Conditional Use Permit is required from the local municipality, Morris Township, where the bleeder fan is located. The focus of this permit is to assure that the project will comply with local ordinances and minimize disturbances such as noise levels and increased traffic. On July 6, 2010, the Supervisors of Morris Township granted the Conditional Use Application for a
ventilation air methane abatement system to be operated as an adjunct facility in connection with the Enlow Fork Mine bleeder shaft.
Status and Plans: Green Holdings issued a request for price to a pre-screened selection of RTO manufacturers in the first quarter of 2010. After all the permits for the project have been issued, Green Holdings will notify the selected equipment supplier and they expect the RTO Plant to be operating at the Enlow Fork site within 22 weeks of such notification. The project is expected to begin generating credits in the second quarter of 2011.
Figure 1. Conceptual layout of bleeder fan site with VAM abatement system installed.
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
NOVEL METHODS OF COAL SEAM GAS CONTENT DETERMINATION FOR ESTIMATION OF GREENHOUSE GAS EMISSIONS FROM MINING Abouna Saghafi CSIRO Energy Technology, P.O. Box 52, North Ryde, NSW 1670, Australia Email: [email protected]
Coal seam naturally contains greenhouse gases, dominantly methane but also carbon dioxide and to a lesser extent higher hydrocarbons. With coal mining most gas volumes trapped in coal seams and strata are liberated which mostly end up in atmosphere. In order to assess fugitive emissions from mining new methods are devised. In situ gas content of coal seams is a primary parameter required in these methods. Traditionally the purpose of gas content determination in coal mines has been the safe operation of mining and workers. Over the years various methods of measurement have been developed for ungrounded mining. However, these methods are not always adequate for requirement of emissions assessment for greenhouse gas inventory and new definition of gas content and more accurate measurement methods are required. The liability of the coal producer or coal user or both in relation to the emitted and remaining gas in coal can be a major factor for the way gas content is defined and measured. In order to accommodate the future emission trade scheme (ETS), one of the first steps is to debate the definition of gas content and developing more adequate methods of measurement. In this paper the authors looks at possible definitions of gas content in relation to the purpose of its use and suggests novel methods of its determination. Keywords: Coal seam gas, gas content, greenhouse gas, fugitive emissions, methane, carbon dioxide
INTRODUCTION Coal seams are high capacity gas reservoirs and to some degree, most coals contain some gas. The mine gas consists generally of methane (CH4) and carbon dioxide (CO2) which are partly or totally released during mining. Mining leads to large disturbance of coal seam reservoir as the fracturing develops both in coal and rocks. Gas escapes to the atmosphere via fractures and exposed coal surfaces. In 2009 the annual greenhouse gas (GhG) emissions from anthropogenic sources in Australia were about 539 Mt CO2-e. About 40 Mt CO2-e of these were fugitive emissions. The coal mining is the main source and represents more than 70% of the total fugitive emissions. The emissions are mostly due to release of the
1
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
trapped coal seam gas which is liberated during and after the completion of mining. The intensity of emissions depends on flow properties of strata and diffusivity and matrix permeability of coal. The method of mining affects the extent and density of induced fractures. Permeability could increase by several orders of magnitude. Moreover, fracturing of strata causes the discharge of water from mining area leading to further increase of permeability and acceleration of gas desorption from coal. Though the rate and intensity of gas release from mining at a given time is primarily a function of temporal gas content and flow properties of coal and strata, the total volume of gas liberated over the life of mine is a function of the virgin, pre-mining magnitude of gas content of coal seams and gas trapping strata. Novel methods are currently being developed to estimate GhG emissions from coal mining. For the case of open cut mining, where the emission is diffuse, direct measurement is difficult and in many cases virtually impossible. In a novel method, developed by author for Australian mines (Saghafi, 2010), the emissions estimate is based on a number of parameters related to the lithology of the strata and coal seam properties. Among these parameters the in situ gas content of coal is of primary importance. Accurate measurement of gas content is therefore primary for reducing the uncertainty of estimation. The appropriate definition and corresponding measurement method of gas content can influence, sometime largely, the magnitude and accuracy of emissions estimation. While measurement of low gas content may not be of any importance to safety issues in underground mining it is quite important for accurate estimation of GhG emissions from both open cut and underground mines. In the next sections after discussing various modes of gas storage in coal various definitions of gas content are provided; this is followed by description of the standard method of the gas content determination and proposed method for low gas content testing. Error associated with the new method will be discussed.
MECHANISMS OF GAS STORAGE IN COAL AND GAS CONTENT Coal is a porous rock where gas can be stored in the large space available on the pores’ internal surface. Gas is stored both in free and adsorbed phases. The adsorbed phase is held on the pore surfaces and constitutes the greatest portion of the stored gas for shallow to medium depths (<600 m). At higher depths due to increase in the density of gas larger volumes of free gas can be stored in pore volumes. The main and largest part of the pore surface area is located in the micro pore system (size <2 nm) where the pores are only a few times larger than the coal seam gases molecular sizes. The adsorbed phase storage in the micro pore system follows a pore filling mechanism and therefore reaches its maximum value (adsorption capacity) when the pore system is fully filled. The stored gas is then retained due to a combination of adsorption and capillary forces. Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining
2
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
Definition of gas content Gas content can be generically defined as the volume of gas contained in unit mass of coal. In Australia the standard conditions are a temperature of 20°C and a absolute pressure of 101.325 kPa (Standards Australia, 1999). Based on the physics of gas storage in coal, total gas contained in coal is the sum of free gas content and adsorbed gas content. However, the traditional method of measurement does not allow the determination of the free gas component of gas in coal. Using the direct method, the volume of gas which is measured consists only of adsorbed gas. This gas in turn can be divided into the ‘desorbable’ and ‘residual’ components of gas content. Either one or both components could be required depending on the purpose of the emissions studies. Fresh core drill samples are generally collected for gas content testing. When the drill rig hits the coal seam coal starts releasing its gas rapidly and almost all free gas is gone before coal reaches the ground surface. At the surface (atmospheric pressure), the adsorbed gas would be released slowly. Desorption processes could continue for days or weeks until there would be no measureable volume of gas. The release of gas is due to forces of gas pressure and gas concentration gradients. In Figure 1 the processes of gas desorption from coal is depicted. Total amount of gas naturally desorbed from coal is called ‘desorbable’ gas content (Qd) while the remaining gas in coal is called residual gas content (Qr). The sum of desorbable and residual gas content constitutes the total gas content (Qt).
Qt = Qd + Qr
(1)
Gas Content (m3/t)
Qd
S de low so rp tio
Figure 1 Desorbable (Qd) and residual (Qr) gas contents
n
Qr Time (day)
Note that this gas volume is only the adsorbed gas which is released slowly and naturally. The current ‘direct method’ of gas content measurement does not allow the determination of the free gas. Besides the generic definition above the definition of gas content can be a Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining
3
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
function of its use as well and other definition for gas content may be required to address the need of various users. For example, from the viewpoint of gas reserves of in coal (i.e. in CBM industry), gas content may include free gas volumes. This definition can also be extended to GhG emissions inventory where it is required to have an estimation of free gas in coal as well.
MEASUREMENT OF GAS CONTENT OF COAL Gas content of coal is determined by direct measurement of the volume of gas desorbed from coal. In Australia two methods are routinely used, namely the ‘slow desorption’ and ‘fast desorption’ methods (Williams et al., 1992; Saghafi et al, 1998). Variants of both methods have been also used in various forms in other coal mining countries. In France a variant of the fast desorption was being used since the early 1960s (Bertard et al, 1970). In USA various variants of the slow desorption method have been used (Kissell et al, 1973; Diamond and Levine, 1981; Diamond et Schatzel, 1998). The Australian slow desorption method was a development of the USBM method after some enhancement (Australian Standard, 1991). The fast desorption method otherwise known as quick crash method was developed in early 1990s (Williams et al., 1992; Saghafi et al, 1998) and is currently the method of choice for assessment of gas emissions and outburst risk in underground coal mines (Australian Standard, 1999). Though both methods consist of similar steps to determine the gas content coal, the length of the procedure is significantly longer in the slow desorption method. In the fast desorption method coal is crushed after a short period of natural desorption so that all its gas is forced to release rapidly (minutes up to an hour). An advantage of fast desorption method besides its rapidity is reduction in the risk of dissolution of CO2 in measuring water for mixed gas conditions (when both CH4 and CO2 are present in seam gas). The slow desorption method and fast desorption methods are both based on measurement or estimation of the volume of gas desorbed from coal in several stages. Both methods start with estimation of ‘lost’ gas during drilling and retrieval of coal sample to the surface. In the slow desorption method desorbs its gas ‘naturally’ until to further desorption is detected. In fast desorption method, however, after a short time allowed for gas to be released naturally during coal transport to the laboratory and in the lab, coal is crushed. Note that in the slow desorption method, the crushing stage may be also included to determine the ‘residual’ gas content of coal (Qr). The three components of gas contents from three stages of gas content testing in fast desorption method are commonly represented by Q1, Q2 and Q3 parameters. The ‘measured gas content’, Qm, is the sum of the 3 components (Australian Standard, 1999), The Q1 or the lost gas stage is identical for the two methods. The Q2, gas desorbed during transport and in the lab is also called desorbed gas in the slow desorption method is the main component of the gas content in this method. In this method this stage is allowed to continue until no further measurable gas desorption is observed. In the fast desorption method the Q2 step is generally short as coal is crushed any time after reaching lab depending on the availability of a measuring system and proper conditions. The last Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining
4
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
component of gas content is Q3 which is the gas desorbed from crushed coal in fast desorption method is also usually the largest volume of desorbed gas in this method. For the slow desorption method this stage is often of no importance as residual gas is expected to be small. As discussed the measurement of the volume of gas released (in all three stages) is done using a measuring cylinder. The released gas is admitted into a water filled inverted cylinder. The displacement of water provides the measure of the volume. This system has worked well over the years and is used routinely in Australia. There are, however, some problems with this way of measuring the volume of gas including gas partial pressure effect and dissolution of gas in water. These have been addressed over the years and improvements have been suggested and applied (Saghafi et al, 1998). Reporting error of gas content testing Note that in each step of measurement of gas content there would be some level of error ‘Measured’ gas content inherits all these errors. The total error of gas content (εm ) could be calculated as follow,
ε m = ε1
Q Q1 Q + ε 2 2 + ε1 3 Qm Qm Qm
(2)
where ε1, ε2 and ε3 are the errors produced in measuring Q1, Q2 and Q3 components. As an example in one case we had, Q1 = 0.5 m3/t, ε1= 20% Q2 = 3.0 m3/t, ε2= 15% Q3 = 2.5 m3/t, ε2= 5% The measured gas content is then presented as Qm = 6.0 ± 0.67 m3/t.
GAS CONTENT MEASUREMENT FOR GREENHOUSE GAS ESTIMATION Fugitive gas emissions from coal mining are considered to be an important source of total anthropogenic emissions. Gassy mines are generally underground mines and estimation of emissions is straightforward as the total gas released is either captured through drainage pipeworks or through ventilation shafts. These quantities can be accurately measured and hence there are no major issues with quantification of such emissions. However, emissions from surface mining are subject to significant uncertainties due to diffuse character of emissions. One parameter influencing largely the emissions is the gas content of in-situ coal. However, the gas content of coals from a surface mine coal can be very low. Note that this is also true for ‘non gassy’ underground coal mines. In these cases the conventional method of measurement of volume by using water displacement may be quite impossible and prone to large relative errors. Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining
5
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
To illustrate the limit of measurability using the standard method it should be noted that the smallest volume that can be confidently measured by using the water displacement in a 250 cm3 measuring cylinder, is about 5 cm3. Therefore for a sample of 100 g size the method can measure low gas content of about 0.1 m3/t if some appropriate level of confidence is to be respected. The accuracy and limit of detectability of measuring gas content could have important impact on coal mining economics as the carbon credit/tax is based on amount of emissions from mine and given the large volume of production even small change in gas content can affect the economy of mining. Measurement of low gas content For very low gas content coals (Qm<0.1 m3/t) the water displacement method of measuring the volume is not adequate. For this case there would be no measureable Q1 and often no measurable Q2. For measurement of gas content of these coals (‘non-gassy coals’), the author suggests that the best practice would be to seal the fresh sample in a gas tight canister in the field, and then dispatch it to the lab for crushing. Ideally coal could be sealed in a canister which can be directly mounted on the crusher so that there would be no need to open the coal canister before crushing. The total desorbed gas can then be indirectly evaluated using a method based on measurement of gas composition rather than measuring gas volume. One method to measure the volume of desorbed gas and determine the gas content is to keep the crushed coal in the crusher container for sometime after the completion of crushing and measure gas composition in the crusher canister after a period of time (a few hours). Gas content 3
(m /t) 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
Q1
Q2
Figure 2 Three main components of gas content testing and post crushing component Q3’
Q3
The volume of desorbed gas is then determined Q 3' from the knowledge of void volume in the 0 100 200 300 400 500 600 700 800 900 1000 crusher canister and gas Absolute gas pressure (kPa) Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining 6
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
composition values. This method was used by the author at CSIRO in the course of a number of national mining projects on determination of gas content of coal. The method was applied after the completion of the three stages of gas content testing (Q1, Q2 and Q3). This new component of gas content was called Q3’ (Q3 prime). In Figure 2, relative quantity of Q3’ component of gas content for a coal is given. For routine measurement of low gas content of coal, a similar approach is suggested. The set up is conceptually illustrated in Figure 3. The crushing canister is initially flushed with helium (He) and then coal is placed in the crusher and crushed. After the completion of the crushing and allowing time for temperature equilibrium, the canister is opened to a closed circuit with an in-line pump. A gas sample is collected from the system after a sufficient period of time and gas composition is measured using gas chromatography. Knowing the volume of the total void space in the system the gas content is determined.
To GC
Crushed coal
Micro pump (vacuum)
Figure 3 Schematic diagram of measurement gas content for low gas content coals
He
The lower limit of gas content which can be determined using this method can be evaluated from the knowledge of void volume in the system (crusher and piping) and the lowest or optimal lower limit of gas chromatography in use. For instance if the volume of void is about 500 cm3 (typical void in the CSIRO quick crush canister for a 100g coal sample), and a GC which can measure accurately a concentration of 100 ppm of methane is used, then the gas content of about 0.001 m3/t can be theoretically determined. This method, therefore, can measures gas content values of at least 100 times smaller than the standard method.
CONCLUSIONS Coal mining is the main source of fugitive greenhouse gas emissions in Australia. It presents more than 70% of the total fugitive emissions. The emissions are mostly due to release of the trapped coal seam gas and strata liberated during mining and after the completion of mining. Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining
7
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
As the inventory of emissions is becoming part of the economics new methods are being developed to estimate GhG emissions. In the new methods the in-situ gas content of coal is a primary parameter for evaluation of emissions. However for low gas content conditions the methodologies in use are not accurate and in some cases can not detect the gas. Because of the large volume of coal produced in mining any small error of gas content measurement is magnified and reflected in the emission inventory. A novel method of measurement for low gas content coals is suggested based on measurement of gas composition. The method should largely increase the limit of detectablity and accuracy of gas content measurement. The analysis of the method indicates that theoretically gas content of two orders of magnitude smaller can be measured by using the new method instead of current methods.
ACKNOWLEDGEMENT The author wishes to thank the CSIRO Energy Technology and the Australian Coal Association Research Program for funding this work.
REFERENCES Australian Standard, 1991, Guide to the determination of gas content of coal – Direct desorption method. AS 3980-1991, Standards Association of Australia. Australian Standard, 1999, Guide to the determination of gas content of coal – Direct desorption method. AS 3980-1999, Standards Association of Australia. Bertard, C., Bruyet, B. and Gunther, J., 1970, Determination of desorbable gas concentration of coal (Direct Method), International Journal of Rock Mechanics and Mining Sciences., Vol. 7, pp 43-65. Diamond, W.P, Levine, J.R., 1981. Direct method determination of the gas content of coal: Procedure and results. Report of Investigations 8815, US Bureau of Mines. Diamond, W. P. and Schatzel, S. J., 1998. Measuring the gas content of coal: A review, International Journal of Coal Geology Volume: 35, pp. 311-331. Kissell, F.N., McCulloch, C.M., Elder, C.H., 1973. The direct method of determining methane content of coalbeds for ventilation design. US Bureau of Mines, Report Investigation 7767, 17 pp. Saghafi, A., 2010. A tier 3 method for estimating fugitive emissions from open cut coal mining: application to hunter coalfield, In the Proceedings of 37th Symposium on the Geology of Sydney Basin, 6-7 May, Pokolbin, Hunter Valley, Australia. Saghafi A, Williams D.J. and Battino S., 1998. Accuracy of measurement of gas content of coal using rapid crushing techniques, In: Proceedings of the 1st Australian Coal Operators Conference COAL’98, Wollongong, 18-20 February 1998, Australia, pp. 551Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining
8
Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2
559, also available from http://outburst.uow.edu.au/presentations_publications/coal_1998/Saghafi-1998.pdf Williams, D. J., Saghafi, A., Drummond, D. B. and Roberts, D. B., 1992. Development of a new equipment for rapid determination of coal gas content. Symposium on Coalbed Methane Research and Development in Australia, Townsville, Australia, Vol. 4, pp 2130.
Saghafi: Gas content determination for estimation of greenhouse gas emissions from mining
9
Swelling of Moist Coal in Carbon Dioxide and Methane Richard Sakurovs, Robyn Fry and Stuart Day CSIRO Energy Technology PO Box 52 North Ryde NSW 1670, Australia Phone: 61 2 9490 5309 Email: [email protected]
Abstract The possibility of injecting CO2 into coal seams for enhanced coalbed methane (ECBM) recovery while simultaneously providing long-term carbon sequestration is an active area of research. It is now well known that coal swells in the presence of water and gases, which in turn may affect the permeability of the coal. If the swelling of the coal matrix by each component can be quantified, it may be possible to make better predictions about the suitability of particular seams for ECBM and carbon sequestration. Although there have been numerous studies where coal swelling has been measured in gases or water, there is relatively little information relating to how swelling of coals by gases is affected by water. In this paper we report on the gas-induced swelling behaviour of four moist Australian coals. Blocks of coal, nominally 30 10 10 mm, were cut parallel and perpendicular to the bedding plane from larger lumps. Samples were moisture-equilibrated at 97 % relative humidity before being exposed to CO2 or CH4 at pressures up to 16 MPa and a temperature of 55 °C. Swelling of each sample was measured directly using an apparatus where digital cameras monitored the change in length of the block as a function of pressure. Results show that swelling was greater in CO2 than CH4, with lower rank coals swelling more than high rank material. The presence of moisture significantly reduced the amount of additional swelling by the gas compared to dry coals, however, the degree to which the swelling of the coals was affected by moisture depended on the rank of the coal. It was also found that, proportionally, CH 4-induced swelling was more affected by the presence of moisture than CO2-induced swelling. Although moist coals swelled less in CO2 than dry coals, if the swelling due to moisture is included, the total swelling is more than that induced by CO2 in the dry coal.
1
Introduction
Interest in CO2 sequestration into coal seams and enhanced coalbed methane (ECBM) recovery has driven worldwide research aimed at better understanding the gas-coal processes involved. One of the fundamental properties of concern is the amount of CO2 that can be stored in a particular seam. However, gas adsorption is accompanied by some degree of volumetric swelling (Levine, 1996; Cui et al., 2007; Kelemen and Kwaitek, 2009; Day et al., 2010), which in turn may affect the gas flow characteristics of the target coal seams (Pan and Connell, 2007).
1
In some cases the amount of swelling may be quite significant; Day et al. (2010) found that in CO2 some coals swelled by around 5 % at 16 MPa. Although these measurements were made on unconfined coal, the implication is that under in situ conditions, the swelling will result in closure of the cleat and pore system thus restricting the free flow of gas throughout the seam. A reduction of the CO2 injectivity has already been observed in some ECBM field trials such as the RECOPOL project in Poland, Alison Unit San Juan Basin USA and the Yubari pilot in Japan (van Bergen et al., 2006; Shi et al., 2008; Durcan and Shi, 2009; Fujioka et al., 2010), and this effect is usually attributed to coal swelling. Most laboratory studies of gas-induced swelling in coal have been with dry material (Reucroft and Patel, 1986; Kelemen and Kwaitiek, 2009; Day et al., 2008a). While these investigations have provided valuable insight into the swelling behaviour of a range of coals in various gases, coal seams with the potential for CO2 sequestration and ECBM production will most likely be saturated with water. It is therefore necessary to understand the swelling characteristics of coal under water-saturated conditions. However, at present there is very little quantitative information on gasinduced swelling of wet coal. In a recent paper van Bergen et al. (2009) reported measurements of swelling of two European coals in CO2, CH4 and Ar at pressures up to about 8 MPa. Experiments were made on both air-dried material and samples that had been moistened by exposing to an atmosphere at 97 % relative humidity. The results suggested that moisture tends to compete for adsorption sites on the coal, implying that gas-induced swelling will be suppressed in the presence of moisture. These results are consistent with the results of gas sorption measurements on moist coals. Day et al. (2008b) for example, found that the amount of gas adsorbed by coal was reduced in the presence of moisture, however, the extent of the reduction depended on the rank of the coal and the type of gas. Since the amount of gas adsorbed is closely related to swelling (Day et al., 2010), it is probable that the effect of moisture on swelling is not straightforward. In view of the relative lack of knowledge regarding swelling in moist coals, we undertook a study to examine the swelling behaviour of some Australian coals in detail. Swelling was measured in four coals, of various ranks, in CO2 and CH4 at pressures up to 16 MPa under both dry and moist conditions. We present here the results of these measurements and discuss some implications for sequestration and ECBM applications.
2
Experimental
2.1
Samples
Four coals were used for these experiments, details of which are summarised in Table 1. The coals ranged in rank for sub-bituminous to bituminous and were from the Bowen Basin in Queensland (Coals A, I and N) and the Illawarra Coalfield in New South Wales (Coal O). These samples were used in a previous study where moisture-induced swelling was examined (Fry et al., 2009) so the sample identification codes have been retained here to allow easy comparison if required.
2
Table 1. Properties of the four coals used in this study. Coal A
Coal I
Coal N
Coal O
Moisture (% ad)
5.7
1.7
11.5
1.1
Ash (% db)
4.7
22.5
10.8
16.9
Volatiles (% daf)
29.6
38.3
36.6
21.7
Carbon (% daf)
83.4
82.9
79.3
88.9
Hydrogen (%)
4.1
5.5
4.1
4.5
Vitrinite Reflectance (%)
0.69
0.95
0.62
1.40
Coal blocks with nominal dimensions of 30 10 10 mm were cut from lumps of coal using a diamond saw. Two blocks were made from each sample; one with the long axis perpendicular to the bedding plane and the other with the long axis parallel to the bedding plane (Figure 1)
Figure 1. Coal sample blocks were cut from larger coal pieces. The blocks were then laid inside the sample cell. Only the length measurements were used in the experiments due to their greater accuracy.
For the dry measurements, the sample blocks were dried at 60 °C under vacuum for at least 48 hours before commencing the swelling measurements. Moist samples were prepared by placing the blocks in desiccators containing a saturated solution of K2SO4 to maintain a relative humidity of 97 %. Samples were left in the desiccators at room temperature (21 to 23 °C) for 3 to 4 weeks, which we have found to be sufficient time to allow for complete equilibration.
2.2
Apparatus
Gas-induced coal swelling was measured using an optical technique which has been described in detail previously (Day et al., 2008a). In this system, the prepared sample 3
blocks are contained in a high-pressure cell with transparent windows. Gas is supplied to the cell via a syringe pump to provide pressures up to 16 MPa. The cell is housed in a temperature-controlled oven to maintain isothermal conditions throughout each experiment. Length changes of the blocks are recorded with digital cameras positioned directly above the cell window. A steel block (which does not swell) with the same dimensions as the coal blocks is used as a reference to avoid any artefacts caused by the changing refractive index of the gas as the pressure is increased within the cell.
2.3
Procedure
The dimensions of the perpendicular and parallel blocks were first measured with a micrometer prior to being placed in the swelling apparatus. After sealing the sample cell, the apparatus was heated and maintained at a temperature of 55 °C. When examining dry samples, our usual procedure is to evacuate the cell overnight to fully degas the sample block before starting the measurements. However, this procedure would remove a significant amount of water from the moist blocks. Hence for the moist samples, the cell was only briefly evacuated (for less than one minute) to remove air in the system. Experiments were run in CO2 or CH4 by increasing the pressure in a series of steps up to a maximum of 16 MPa. The system was held at each pressure step for at least several hours after swelling had ceased; typically 24 hours or more, depending on the gas and the coal. The change in volume of the blocks at each pressure step, V + V, was calculated using Equation 1:
V V l per l per l par l par
2
(1)
where lper and lpar are the lengths of the perpendicular and parallel samples, respectively.
3
Results and Discussion
Swelling of these coals was anisotropic, as found previously. In both CO2 and CH4, linear expansion in the direction perpendicular to the bedding plane was always greater than in the direction parallel to bedding. There did not appear to be any difference in the degree of anisotropy between the wet and dry samples, although there was substantial variability within the sample set which would tend to mask small differences, if present. Taken as a whole, the ratio of perpendicular to parallel strain in all samples was about 1.35. Although the ratio of the perpendicular to parallel samples was broadly similar for both wet and dry samples, there was an interesting difference in the way the samples behaved when initially exposed to gas. With dry samples, addition of gas to the cell resulted in a step change in the length of the block followed by a slower increase as it approached equilibrium. Wet samples also increased in length rapidly but tended to “overshoot” then slowly contract to the equilibrium length. This was observed in both directions, parallel and perpendicular to the bedding plane. An example of the difference between wet and dry behaviour is shown in Figure 2 where the linear strain of the perpendicular block for Coal A in CO2 is plotted as a function of time. In
4
this example the pressure was increased from 2 to 4 MPa over a period of about 30 minutes.
1.0
Linear Strain (%)
0.8
0.6
0.4
0.2 Dry Wet 0.0 0
10
20
30
40
Time (hours) Figure 2. Linear strain of wet and dry Coal A perpendicular sample blocks as a function of time. The samples were in CO2 and the pressure varied from 2 to 4 MPa.
It is probable that the apparent contraction after the initial increase in length of the wet blocks was caused by moisture evaporating as the fresh dry gas was admitted to the cell. This is argument supported by the observation that contraction was much more pronounced in the two lower rank coals (A and N), presumably due to their higher moisture contents. As well, the amount of contraction decreased at the higher pressure step measurements, so that towards the end of the run, the moist block behaviour was similar to the dry blocks. Moisture loss is a potential problem with this experimental procedure because gasinduced swelling occurs simultaneously with some shrinkage due to possible partial drying. This can complicate interpretation of the data, but it may be that the effect on the gas-induced swelling is actually quite small. We have previously shown with gas sorption experiments (Day et al., 2008b) that there is a certain moisture content in coal above which additional water has no effect on the sorption capacity of the coal. For CO2, the limiting value was found to be the equilibrium moisture content that would be achieved by exposing the coal to 40-50 % relative humidity, while for methane the relative humidity would be in the range of 60-80 %. Since adsorption and swelling are very closely related (Kelemen and Kwaitek, 2009; Day et al., 2010), it seems reasonable that this limiting moisture content also applies to swelling. The moisture contents of the blocks measured at the end of each experiment indicate that the remaining moisture is at least as high as the limiting moisture content, hence we have assumed that the effect of evaporation on gas-induced swelling during these measurements can be ignored. There is, of course, some shrinkage of the block due
5
to the loss of this moisture, but because the amount of water lost is relatively small, this is minor compared to the gas swelling component. Volumetric swelling of each sample was determined using Equation 1. Previous studies have shown that volumetric swelling can be accurately represented for many gases by a modified Dubinin-Radushkevich (DR) adsorption model (Day et al., 2010). In this model, the density of the gas is used instead of pressure according to Equation 2:
Q Qmax e
D ln L / g
2
(2)
where Qmax is the maximum swelling of the coal, g is the density of the gas at the temperature and pressure, L is the density of the condensed gas, and D is a constant. This relationship was confirmed for all four samples in both CO2 and CH4 as illustrated in Figure 3 where the volumetric swelling is plotted as a function of gas density for Coals A and I. The lines in each plot correspond to the model and show excellent agreement with the measured data. Although Figure 3 shows only dry coals, the DR model fitted the wet samples equally well (see Figure 5 for examples). Using the DR model, the maximum swelling, Qmax, in CO2 and CH4 was calculated for each sample, which are listed in Table 2 along with the corresponding D value. For comparison, the moisture-induced swelling, which had been previously measured for these samples (Fry et al., 2009) is also shown. Note that for Coal N, the moistureinduced swelling was almost as high as that for CO2 in the dry material. CH4 Density (kg m -3) 5
0
20
40
60
CH4 Density (kg m -3)
80
100
120
5
0
Coal A
40
60
80
100
120
Coal I
4
4
Volumetric Swelling (%)
Volumetric Swelling (%)
20
3
2
1
3
2
1
CO2
CO2
CH4
CH4
0
0 0
100
200
300
400
500 -3
CO2 Density (kg m )
600
700
0
100
200
300
400
500
600
700
CO2 Density (kg m -3)
Figure 3. Volumetric swelling of dry samples of Coals A and I in CO2 and CH4 as a function of gas density. The line plots represent the DR model results.
6
Table 2. Dubinin-Radushkevich parameters for the four coals (both dry and moist) in CO 2 and CH4. Maximum moisture-induced swelling in these coals is also shown (moisture results from Fry et al., 2009). Coal A
Coal I
Coal N
Coal O
CO2 Dry
Wet
Dry
Wet
Dry
Wet
Dry
Wet
Qmax (%)
4.40
2.62
2.39
1.97
5.08
2.21
2.01
1.98
D
0.0702
0.0811
0.0667
0.0765
0.0731
0.0685
0.0349
0.0607
CH4 Dry
Wet
Dry
Wet
Dry
Wet
Dry
Wet
Qmax (%)
2.60
0.97
1.45
0.96
2.85
0.68
1.24
0.99
D
0.1050
0.1210
0.1072
0.1252
0.1029
0.1010
0.0858
0.0952
4.90
-
0.43
-
Water Qmax (%)
2.64
-
1.32
-
In CO2, Qmax of the dry coals varied from about 2 % in Coal O to a little more than 5 % in Coal N. Swelling in CH4 was substantially lower than that produced in CO2, ranging from 1.2 % in Coal O to 2.6 % in Coal N. However, the ratio of the CO2 to CH4 swelling was relatively constant, averaging at about 1.7 for all four coals. In common with other studies, the results presented here show a strong rank dependence with higher swelling associated with lower rank coals (Figure 4).
7
6
5
Qmax (%)
4
3
2
1
CO2 CH4
0 0.4
0.6
0.8
1.0
1.2
1.4
1.6
Vitrinite Reflectance (%) Figure 4. Maximum volumetric swelling (Qmax) in dry coals as a function of vitrinite reflectance.
It is clear from the Qmax values in Table 2 that moisture significantly reduced the amount of swelling, analogous to results of gas adsorption measurements in moist coals. This is illustrated more clearly in Figure 5 where the volumetric swelling measured in both CO2 and CH4 for Coals A and I is plotted as a function of gas density.
8
5
5
Coal A
Coal I 4
Volumetric Swelling (%)
Volumetric Swelling (%)
4
3
2
1
3
2
1
CO2 Dry
CO2 Dry
CO2 Wet
CO2 Wet
0 0
100
200
300
400
500
600
0
700
0
100
-3
CO2 Density (kg m )
300
400
500
600
700
CO2 Density (kg m-3)
2.5
2.5
Coal A
Coal I 2.0
Volumetric Swelling (%)
2.0
Volumetric Swelling (%)
200
1.5
1.0
0.5
1.5
1.0
0.5 CH4 Dry
CH4 Dry
CH4 Wet
CH4 Wet 0.0
0.0 0
20
40
60
80 -3
CH4 Density (kg m )
100
120
0
20
40
60
80
100
120
-3
CH4 Density (kg m )
Figure 5. Volumetric swelling of dry and wet samples of Coals A and I in CO2 (top) and CH4 (bottom) as a function of gas density. The line plots represent the DR model results.
In all cases, swelling in the moist samples was substantially less than in the dry coal, however, the effect was not constant across the samples or between the two gases. Proportionally, the effect of moisture was much greater in the lower rank coals and was also more pronounced in CH4 compared to CO2. Figure 6 compares the percentage reduction in Qmax (relative to the dry coal) for CO2 and CH4 as a function of vitrinite reflectance.
9
Swelling Reduction (%)
80
60
40
20 CO2 CH4 0 0.4
0.6
0.8
1.0
1.2
1.4
1.6
Vitrinite Reflectance (%) Figure 6. Reduction in swelling due to moisture as a function of vitrinite reflectance.
For Coal N, the lowest rank coal, moisture reduced the maximum swelling in CO2 by about 55 %, whereas for Coal O (the highest rank sample), moisture had very little effect on the CO2 swelling. But as shown in Figure 6, in CH4 moisture had a substantially higher proportional effect. In Coal N, methane swelling was reduced by nearly 80 % in the moist block and even in Coal O, which was virtually unaffected in CO2, showed about a 20 % reduction in methane swelling. The reduction in swelling in the presence of moisture is due to moisture preferentially occupying polar sites on the coal surface. In dry samples, it appears that CO2 and CH4 can readily adsorb on these sites, but when water is present, they become unavailable to other sorbates. Thus, in moist low rank coals, with a high proportion of polar sites, much of the coal’s internal surface is already covered by water so there is relatively little surface available for gas sorption and therefore little swelling. The proportionally higher effect of moisture on CH4-induced swelling compared to CO2 can be explained in terms of their differing affinity to water. It has been suggested that when water adsorbs on polar sites, hydrogen bonding between water molecules tends to form clusters of water around each site. Although these clusters effectively exclude other sorbates, some gases such as CO2 seem to be able to form a layer over the top of each water cluster. Methane on the other hand, is unlikely to attach to water and is therefore restricted to unoccupied hydrophobic adsorption sites and thus the effect of water is somewhat greater than seen with CO2. When considering the moist coal samples it should be remembered that they are actually already swollen due to the presence of moisture (see the water Qmax values in Table 2). Hence the total volumetric swelling of each coal, relative to the dry material, is the sum of the gas-induced and moisture-induced components. The total swelling can be estimated by adding the gas-induced moist coal Qmax values to the 10
water Qmax or alternatively, by drying the sample after the gas-induced swelling of the moist coal has been measured. This is shown in Figure 7 where the volumetric swelling of Coal A is plotted as a function of time. At the end of the high pressure run, the sample cell was evacuated and the sample allowed to dry. In this example, the sample shrank below its starting point by about 2.7 %, which is essentially the same amount for the Qmax in moisture for this coal. The shrinkage measured upon drying of the moist coal samples after the high pressure gas runs are listed in Table 3 and in all cases are reasonably close to the water Qmax values, also shown in Table 3.
3
Volumetric Swelling (%)
2
1
0
-1
-2
-3 0
5
10
15
20
25
Time (days) Figure 7. Volumetric swelling in moist Coal A (in CO2) as a function of time, showing the effect of drying the sample.
The total swelling in Coal A shown in Figure 7 was 5.3 %, somewhat higher that the maximum swelling measured in any single component in dry coal. Total swelling calculated for the other moist coals were also higher than the corresponding Qmax values for single gases on dry coal (Table 2).
11
Table 3. Measured volumetric shrinkage of moist coal samples after drying. The total swelling (i.e. gas-induced swelling plus swelling due to the presence of moisture) in the moist samples. Coal A
Coal I
Coal N
Coal O
After CO2 run
2.68
0.70
4.76
0.50
After CH4 run
-
0.73
5.66
0.49
2.68
0.72
5.21
0.50
Water Qmax (%)
2.64
1.32
4.90
0.43
Wet CH4 Qmax (%)
0.97
0.96
0.68
0.99
Combined Water and Wet CH4 Qmax (%)
3.61
2.28
5.58
1.42
Dry CO2 Qmax (%)
4.40
2.39
5.08
2.01
Shrinkage (%)
Average Shrinkage (%)
It is clear from these results that injecting CO2 into wet coal seams will cause some degree of swelling. However, countering this will be some shrinkage as water and CH4 are displaced by the CO2. The net effect of the swelling and shrinkage can be estimated from the Qmax values determined above. Adding the moisture swelling and wet methane Qmax components yields an estimate of the total shrinkage as a result of the removal of water and methane, assuming their total removal. These results are shown in Table 3 (as Combined Water and Wet CH4 Qmax) and it can be seen that they are roughly equal to the dry CO2 Qmax for each coal. If this does turn out to be the case in practice, it may be that the swelling induced by CO2 injection, and the concomitant reduction in permeability, may be largely offset over time as the seam is flushed of water and methane.
4
Conclusions
The swelling characteristics of four Australian coals were measured in CO2 and CH4 at 55 °C and at pressures up to 16 MPa. Measurements were made on dry coals and samples that had been moisturised by equilibrating with moisture at 97 % relative humidity. The results of these measurements showed:
Swelling in dry coal is greater in CO2 than CH4 by a factor of about 1.7, regardless of the rank of the coal.
Swelling in dry coal is inversely related to the rank of the coal in both CO2 and CH4.
All samples, both dry and moist showed significant anisotropy with, on average, about 35 % greater strain induced in the direction perpendicular to the bedding plane. The presence of moisture did not appear to have any effect on this property.
Moisture significantly reduces the degree of gas-induced swelling in both CO2 and CH4. However, the amount to which the swelling is reduced is a function of the rank of the coal. In lower rank coals, the effect of moisture on the amount of swelling is high; in the lowest rank coal examined here, swelling was depressed by about 55 % whereas the highest rank coal was barely 12
affected. In CH4, however, the effect of moisture is more pronounced, with up to about an 80 % reduction in swelling in low rank coal compared to the dry material.
The reduction in swelling in the presence of moisture is probably due to water molecules occupying polar sites on the coal’s internal surface, thus preventing adsorption of gas. The higher proportion of polar sites in low rank coals would account for the greater effect in these coals.
Although moist coals swell less than dry samples, they are in fact already partly swelled due to the presence of the water. If this pre-swelling is included, the total swelling of these coals (i.e. water-induced and gas-induced swelling) is higher than that induced by CO2 in dry coal.
Swelling of coal seams during CO2 injection may be significant, but this may be offset by the combined shrinkage induced as the in situ methane and water are displaced from the seam.
Acknowledgements The authors gratefully acknowledge the financial support provided to this work through the CSIRO Coal Technology Portfolio program.
References Cui, X., Bustin, R.M., Chikatamarla, L., 2007. Adsorption-induced coal swelling stress: Implications for methane production and acid gas sequestration into coal seams. Journal of Geophysical Research 112, No. B10202. Day, S., Fry, R., Sakurovs, R., 2008a. Swelling of Australian coals in supercritical CO2. International Journal of Coal Geology 74, 41-52. Day, S., Sakurovs, R., Weir, S., 2008b. Supercritical gas sorption on moist coals. International Journal of Coal Geology 74, 203-214. Day, S., Fry, R., Sakurovs, R. Weir, S., 2010. Swelling of coals by supercritical gases and its relationship to sorption. Energy and Fuels 24, 2777-2783. Durucan, S., Shi, J., 2009. Improving CO2 well injectivity and enhanced coalbed methane production performance in coal seams. International Journal of Coal Geology 77, 214-221. Fry, R., Day, S., Sakurovs, R., 2009. Moisture-induced swelling of coal. International Journal of Coal Preparation and Utilization 29, 298-316. Fujioka, M., Yamaguchi, S., Nako, M., 2010. CO2-ECBM field tests in the Ishikari Coal Basin of Japan. International Journal of Coal Geology 82, 287–298. Kelemen, S.R., Kwaitek, L.M., 2009. Physical properties of selected block Argonne Premium bituminous coal related to CO2, CH4 and N2 adsorption. International Journal of Coal Geology 77, 2-9. Levine, J.R., 1996. Model study of influence of matrix shrinkage on absolute permeability of coal bed reservoirs. Coalbed Methane and Coal Geology, Geological Society Special Publication No 109, pp 197-212 Pan, Z., Connell, L.D., 2007. A theoretical model for gas adsorption-induced coal swelling. International Journal of Coal Geology 69, 243-252. 13
Reucroft, P.J., Patel, H., 1987. Gas-induced swelling in coal. Fuel 65, 816-820. Shi, J., Durucan, S., Fujioka, M., 2008. A reservoir simulation study of CO 2 injection and N2 flooding at the Ishikari coalfield CO2 storage pilot project, Japan. International Journal of Greenhouse Gas Control 2, 47–57. van Bergen, F., Pagnier, H., Krzystolik, P., 2006. Field experiment of enhanced coalbed methane – CO2 in the upper Silesian basin of Poland. Environmental Geoscience 13, 41-52. van Bergen, F., Spiers, C., Floor, G., Bots, P., 2009. Strain development in unconfined coals exposed to CO2, CH4 and Ar: Effect of moisture. International Journal of Coal Geology 77, 43-53.
14
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission PROGRAM TOPIC: 4. CARBON MANAGEMENT - CO2 SEQUESTRATION EVALUATION OF TOTAL POROSITY AND THE AMOUNT OF INACCESSIBLE PORES IN COAL USING SMALLANGLE NEUTRON SCATTERING Yuri B. Melnichenko, Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6393, USA Email: [email protected] L. He, Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6393, USA Email: [email protected] R. Sakurovs, CSIRO Energy Technology, North Ryde 2113 Australia Email: [email protected] M. Mastalerz Indiana Geological Survey, Indiana University, 47405-2208 USA [email protected] T. Blach Nanoscale Science and Technology Center, Griffith University, Brisbane, Queensland 4111, Australia Abstract: Carbon sequestration in geological formations such as deep unminable coal seems is one of the proposed measures for arresting the rising concentration of atmospheric carbon dioxide. The efficiency of CO2 sequestration and enhanced coalbed recovery depends crucially on the ability to predict sorption capacity of green house gases in coal, which may vary significantly depending on the total porosity of the coal as well as on the amount of pores that are actually accessible to a specific fluid. Due to the high penetration power and relatively short wavelength of neutrons, small-angle neutron scattering (SANS) as well as ultra small-angle scattering (USANS) techniques are ideally suited for assessing the phase behavior of various fluids in engineered and natural porous systems including coal. SANS and USANS offer a number of advantages for evaluating the total coal porosity, as all pores both accessible and inaccessible to the fluid contribute to the scattering at ambient conditions. One of the most important advantages is that scattering experiments can be performed with samples saturated by non- or weakly adsorbed supercritical fluids at pressures corresponding to the zero average contrast (ZAC) condition. At this condition the neutron contrast between solid matrix of coal and all pores accessible to a fluid is annulled and any residual scattering may be attributed to the scattering from inaccessible pores which do not belong to the interconnected porous channels and therefore cannot be filled with the fluid. In this talk we will discuss the results of the USANS and SANS studies of four bituminous coals from the Illinois Basin (USA) and and Bowen Basin (Australia) saturated with CO2 and methane at pressures up to 8000psi (53MPa) including pressures corresponding to the ZAC condition. The scattering patterns at different pressures are used to demonstrate that the scattering length density of different coals, the parameter crucial for calculating ZAC condition can be accurately predicted from evaluated based of the chemical composition of a particular coal obtained by ultimate analysis. Scattering patterns at the ZAC are compared with SANS and USANS measured at ambient conditions and all studied coals reveal non-zero residual scattering, which shows that there are pores in coal inaccessible to methane and CO2 on the timescale of the experiment (days). Scattering patterns from pores with sizes greater than ~ 100 Angstrom are analysed using a newly developed approach which is used to determine the amount of pores accessible to CO2 and
methane in each coal as a function of pore sizes. We demonstrate that the way the volume of accessible pores changes as a function of pore sizes is unique for each of the examined coals and the volume fraction of accessible pores may vary between 90 % (macropore region) to 30 % (mesopore region). We also analyse scattering from micropores, which reveals strong condensation effects of CO 2 and methane in all for coals. The developed methodology may be applied for evaluating the volume of accessible pores in other natural porous materials of interest for CO2 sequestration, such as saline aquifers, shales, and sandstones.
EARLY STAGE DETECTION OF COAL SPONTANEOUS COMBUSTION IN VIEW OF PRETREATMENT OF THE COAL Boleslav Taraba*, Zdenek Pavelek**, Jiří Janek Ostrava University, Dept. of Chemistry, 30.dubna 22, Ostrava 1, Czech Republic * E-mail: [email protected] ** OKD, HBZS, a.s., Ostrava-Radvanice, Czech Republic Abstract The main aim of the investigations was to elucidate possible changes in oxidation behaviour of coal from the mine district, that had to be sealed off because of fire incident. With this respect, two samples of bituminous coal were studied to describe changes in evolution of the gases indicating early stage of spontaneous combustion process (namely carbon monoxide, ethylene, propylene). Three types of coal pre-teatment procedures were used to simulate in situ conditions of the coal lying near the „spon-com“ site in the sealed off area: i) pre-oxidation of coal (200, 250°C), ii) pre-heating of the coal under inert gas (200, 250°C), and iii) extinguishing of pre-heated coal (200, 250°C) by liquid water. After the pretreatments, amount and composition of gases evolved from the coal during aerial oxidation at temperatures 40 – 200°C were measured at continuous flow reactor. Irrespective of type of the pre-treatment, oxidation of the pre-treated coal was found to be accompanied by increased CO evolution in comparison with not-treated coal (up to temperature 100°C). Opposite to it, evolution of ethylene and propylene was ascertained to be somewhat lowered by the coal pre-treatment procedures. However, experiments proved that threshold temperature for the unsaturated hydrocarbons (i.e. minimal temperature when evolution of the indication gas is obviously detected) is not depended on the coal pre-treatment procedure. Obtained results thus indicate that carbon monoxide and unsaturated hydrocarbons have their relevance as spontaneous combustion indicators even for mine areas that had been sealed off because of coal fire incident, and that are to be re-opened. KEYWORDS: spontaneous heating; oxidation; coal; gas evolution
1. Introduction The process of spontaneous heating of coal continues to be serious problem, particularly for underground extraction of the coal. A key to early warning system of the spontaneous heating of coal insists in the fire gases detection. In this respect, carbon monoxide and/or unsaturated hydrocarbons (ethylene, propylene) are generally accepted as the main indication gases /1/. When coal spontaneous combustion occurs, the mine area is usually sealed and, later (when fire is extinguished), it is re-opened.
However, lack of information is felt
regarding the possible changes in oxyreactivity of the coal from the re-opened areas. Coal in the vicinity of spontaneous combustion site could be exposed to increased temperature,
both in the presence and deficience of oxygen. Thus, change in oxidation behaviour of the exposed coal can be anticipated. The paper is aimed at elucidating possible changes in evolution of the indication gases emitted from coal taken from mine district that had to be sealed off because of fire incident.
2. Experimental 2.1 Coal sample Two samples of the bituminous coal from Ostrava-Karviná Coal Basin (denoted as “Darkov07” and “TX05”) were studied. Basic analytical characteristics of the coals are given in table 1. Table 1: Proximate and ultimate analyses of the studied coal Proximate analysis Ultimate analysis d daf daf C (%) Hdaf (%) Ash, A (%) VM (%) Darkov07 5.5 24.1 89.5 4.9 TX 05 6.2 30.5 87.8 4.8 d – dry basis; daf – dry, ash- free basis; VM - volatile matter; Sample
2.2 Coal pre-treatment procedures A lump of coal was milled and sieved for required grain size, i.e. 0.4 - 2.0 mm. Subsequently, three types of coal pre-treatment procedures were applied to simulate in situ conditions of the coal lying near the „spon-com“ site in the sealed off area: a) pre-oxidation of the coal by aerial oxygen at temperatures 200 and 250°C respectively; b) pre-heating of the sample under inert gas (nitrogen) at temperatures 200 and 250°C respectively. c) immersion of pre-heated coal (250°C) by liquid water (simulation of extinguishing process of the “burning” coal). Each pre-treatment procedure took five hours.
2.3 Experimental approaches The samples were examined with respect to amount and composition of the gases evolved from the pre-treated coal during its (subsequent) oxidation. Attention was paid
especially to gases indicating the spontaneous heating stage, namely, carbon monoxide, ethylene and propylene. For the measurements, continuous flow reactor operating under air at temperatures 40 – 200°C was used /2/. As comparison basis for the experimental results, data obtained for original, not-treated coal samples were used.
3. Results and discussion Figure 1 shows evolution of carbon monoxide (CO) during oxidation of the coal sample with different pre-treatment. The dependence is expressed in terms of rate of the CO evolution (ml/(kg.min)) versus temperature.
450
Original coal N2 200°C
400
N2 250°C
CO evolution (10
-3
ml/kg.min)
500
Pre-ox 200°C
350
Pre-ox 250°C + H2O (liq) at 250°C
300 250 200 150 100 50 0 30
50
70
90
110
Temperature (°C)
Figure 1: Evolution of the carbon monoxide during oxidation of the pre-treated coal, sample DARKOV07
From Figure 1, it can be seen that, irrespective of the type of pre-treatment, oxidation of pre-treated coal is accompanied by increased CO evolution in comparison with nottreated coal (up to 100°C). All pre-treatment procedures thus evidently resulted in rise of concentration of structural precursors (functional groups) on coal surface that subsequently oxidative decompose evoluving carbon monoxide. For unsaturated hydrocarbons, however, the situation is a bit different. Irrespective of noteworthy scatter of the experimental data, certain decrease in ethylene and propylene evolution for pre-treated coals was observed. Such a trend was obvious especially for coal sample TX05. This is demonstrated in the Figure 2 for ethylene. From the practical point of view, the finding underlines the importance of even trace amounts of unsaturated
8 Original coal 7
N2 200°C "N2 250°C"
Ethylene evolution (10
-3
ml/t.min)
hydrocarbons in the air of re-opened localities - as spontaneous heating indicators.
Pre-ox 200°C
6
Pre-ox 250°C + H2O (liq) at 250°C
5
4
3
2
1
0 50
70
90
110
130
150
170
190
Temperature (°C) Figure 2: Evolution of the ethylene during oxidation of pre-treated coal, sample TX05
In addition to rate evolution measurements, unsaturated hydrocarbons were also tested with respect to their threshold temperature.
The value of threshold temperature
represents minimal temperature when given hydrocarbon is “primarily” evolved from coal and it can be thus obviously detected in the air. Experimentally determined values of the threshold temperatures as a dependence of the type of coal pre-treatment are summarised in the Table 2.
Tab. 2: Threshold temperatures of ethylene and propylene evolved from studied coals as a dependence on the type of coal pre-treatment ORIGINAL Pre-oxid. Pre-oxid. N2 200°C N2 250°C + H2O (liq) 200°C 250°C at 250°C
SAMPLE
T X 05
Ethylene
100- 120°C
120°C
100°C
140°C
120°C
60- 80°C
Propylene
100- 120°C
100°C
100°C
100°C
120°C
80°C
SAMPLE
D A R K O V 07
Ethylene
100°C
120°C
120°C
100°C
100°C
80°C
Propylene
120°C
100°C
120°C
120°C
100°C
80°C
From the Table 2, it is obvious that both oxidative and inert pre-treatment of the samples practically do not change the values of threshold temperatures for ethylene and propylene (experimental error of the determination is about 20°C). However, pre-treatment procedure simulating extinguishing of the heated coal (Table 2, last column) was found to result in some decrease of threshold temperature, by approx. 30 + 10°C. Thus, evidently, liquid water when applied on “hot” coal causes chemical changes in its structure as reported already earlier /3/. 4. Conclusion On the basis of obtained experimental results one can conclude that carbon monoxide and unsaturated hydrocarbons principially keep their relevance as spontaneous combustion indicators even for mine areas, that had been sealed off because of coal fire incident, and that are to be re-opened.
5. Acknowledgement Financial support of CBU Project 55-07 is gratefully appreciated. 5. References /1/ Mackenzie-Wood, P. and Strang, J.: Fire Gases and their Interpretation, The Mining Engineer 149 (1990) pp. 470-478. /2/ Taraba, B. and Peter, R.: Initial Stage of Coal Oxidation in View of Threshold Temperature, in Proceedings of the International Conference on Coal Science & Technology (CD), October 9-14, 2005 Naha, Okinawa, Paper No. 4B03. /3/ Taraba, B.: Low temperature oxidation and spontaneous combustion of coal, Ostrava University, 2003, pp. 112, ISBN 80-7042-832-5, (in Czech).
SEM study of some Indian natural cokes (jhama) #
* Ashok K. Singh, **Mamta Sharma, ***Mahendra P. Singh, *Nandita Choudhury
*Central Institute of Mining & Fuel Research, CSIR, Govt. of India Dhanabd-828109, India **National Metallurgical Laboratory, CSIR, Govt. of India, Jamshedpur-831007, India ***Dept. of Geology, Banaras Hindu University, Varanasi-221005, India (#corresponding author: [email protected])
Abstract Scanning Electron Microscopy (SEM) is an indispensable tool for studying the mineral and organic matrix, particularly in the baked coals or natural cokes, which form due to magmatic intrusion (dykes and sills) in coal seams. In this paper this tool has been used to demonstrate the relationship between morphology and genesis of natural coke or jhama samples derived from Damodar Valley Coalfields of India. The SEM studies were carried out using JEOL-840A JSM under magnifications ranging from 200X to 50000X on selected heat altered coal/natural coke, CPC and graphite samples to decipher the microstructures of these samples and comparative studies. The chips of coal lithotypes, heat altered coal/natural cokes, CPC and graphite, of about 1cm size were isolated from the coal blocks using hammer, chisel and forceps and were washed with alcohol to remove the superficial dust and mounted on brass stubs using silver paste as glue and coated with gold (thickness about 400A°). SEM photographs reveal some textural features of natural coke which would not be detectable using optical microscopy. The similarity between images of natural and artificial coke and also the mineral assemblages present in the natural coke matrix is proof that temperatures over 350°C were experienced during contact metamorphism. Due to its greater depth of focus compared to optical microscopy, SEM provides information on textures in unpolished samples complementary to microscopic data. The main findings include identification of carbonized matrix, mosaics and flow structures of various dimensions, nature and dimensions of the micro, meso, macropores and cracks formed due to escape of volatiles and relationship and association of mineral matter or their altered products (glassy matrix) with the generated pores. The system thus permits a combined maceral-mineral analysis, which can be used to advantage in coal geology studies.
Key words: SEM, natural cokes (jhama), carbonised matrix, microstructures. International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 1
Introduction
In Indian coalmines huge reserves of heat affected coals are available, but have remained unused due to lack of systematic approach to their physico-chemical and petrographic characterization. Intrusions of igneous rocks in coal seams are common in Indian coalfields. In general, the coals are altered to “natural coke” or “jhama” in the vicinity of intrusions, irrespective of their mode as dyke or sill, and the magnitude of heat effect shows gradual decrease if you take a traverse away from the coal intrusive contact.
The igneous intrusions, mostly of mica peridotite and dolerite types, in most of Indian Gondwana coalfields, were placed during Early Cretaceous period (Blanford, 1861; Cooper et al., 2006; Fox, 1930 and 1934; Ghose, 1967; Chatterjee and Ghosh, 1970; Ghosh et al., 1977; Chandra, 1992, Chandra and Srivastava, 1975). They occur in discordant (dykes) and concordant (sills) forms at variable depths and mostly along the weak planes, particularly into the rocks of lowest resistance at the boundary between variable lithologies. The dykes are shown to have punctured the coal seams and have actuated severe baking effects. In addition, this has caused physico-mechanical and chemical destruction of the coal seams and has increased the ash content of affected coals due to varying degree of distillation (Pascoe, 1959; Merritt, 1990; Querol et al., 1997; Golab, 2003; Kwiecinska and Petersen, 2004; Stewart et al., 2005; Cooper et al., 2006; Golab et al., 2007). Further, Singh (1998) has shown that the impact of heat is generally of varying extent but near the intrusion-coal contacts, the coals have given rise to natural coke and the inter-seam sediments (mostly shales and sandstones) have produced paralava. Furthermore, he has shown that the para-lava is characterized by the occurrence of minerals such as biotite, olivine, tridymite, cristobalite, mullite, calcite, siderite, ankerite, International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 2
and pyrite. Further detailed work on Indian natural cokes was carried out by Singh et al., 2003; 2007; 2008a & b; Sharma, 2009 a&b, etc.
In this paper it has been tried to show how Scanning Electron Microscopy (SEM) is a useful and indispensable method for the study of coals/altered products, minerals and rocks. It can be used as a rapid tool to demonstrate relationships between morphology and genesis of natural coke or ‘jhama’ samples. Here, SEM has been applied in a study of coals altered to natural coke by a magmatic intrusion. SEM photographs reveal some textural features of natural coke, which can not be detectable using optical microscopy. The similarity between images of natural and artificial coke is proof that temperatures over 350°C were experienced during contact metamorphosis. Due to its greater depth of focus compared to optical microscopy, SEM has provided information on texture in unpolished samples complementary to microscopic data.
Geological Features
The Jharia Coalfield is the store house of coking coals in India. Unfortunately, thick piles of Lower Gondwana sequences of these coalfields are intruded by magma. This has caused enormous damage to these valuable coal deposits in various ways. The major concern of the coal industry has been to make an assessment of the impact of intrusive activity on coal and to evolve effective technologies for their mining and utilization. Moreover, natural coke or jhama, produced by carbonization through magmatic heat need to be properly understood, as international and national standards, including ISO and BIS, are yet to be precise about their physical, chemical and technological properties.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 3
Typical ‘jhama’ or ‘natural coke’ is produced from coking coals and ‘natural char’ from non-coking coals under suitable carbonization conditions (Singh et al., 2003, 2007, 2008a&b, 2009, Sharma, 2009). With the help of cartoon (Singh et.al, 2008, Sharma, 2009) the sequence of thermo-contact alterations of coal due to concordant and discordant intrusives is shown, particularly in the context of Damodar Valley coalfields. It has been observed through field investigations that the dykes, being perpendicular, have limited impact, whereas the sills being parallel to the seams have made extensive damage (Singh et al., 2008). This coal magma intrusion has resulted in evolution of CO2 and CO gases near the contact at very high temperatures, reaching sometimes even up 1500°C, which due to lack of escape routes, re-deposited and give rise to the genesis of carbonate minerals such as siderite, calcite and ankerite (Singh et al., 2008 a, b). These carbonate minerals have actually inflicted higher volatile matter content in the higher specific gravity fractions with higher ash content at different sizes.
It has been observed that in the zones with relatively low temperatures, around 3003500C, initiation of metaplast phase of coal alteration took place, which commenced the process of coke formation. This temperature zone has been marked as the outer most shell of reactive domain and the core of it has been projected as a zone having thermal activity with temperatures reaching up to 15000C. This has been corroborated by the presence of minerals such as mullites, cristobalites and glassy matrix in the high temperature zones, the zone close to igneous intrusion (temperature rising up to 1500°C), where coal has been transformed into natural coke (Singh, 1998, Singh et al., 2007, 2008a&b). This observation has also been supported by physical, chemical and petrographic characteristics of studied samples (Singh et al., 2007, Sharma, 2009) and has finally led to the demarcation of zones such as zone of severely affected coal, moderately
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 4
affected coal and unaffected coals with respect to distance from the intrusion (Singh, et al., 2007, 2008a&b, Sharma, 2009a&b).
Materials and Methods
For optical microscopy the procedures defined by Bureau of Indian Standards (IS-9127, Part 2and 3) were followed. The study was performed under reflected light; in case of anisotropic materials such as natural coke, CPC and graphite, crossed polarized light mode was applied. For SEM studies standard manuals of the machine were used for studying these samples with utmost care.
The chips of coal lithotypes, heat altered coal/natural cokes, CPC and graphite, of about 1cm size were isolated from the coal blocks using hammer, chisel and forceps and were washed with alcohol to remove the superficial dust and mounted on brass stubs using silver paste as glue and coated with gold (thickness about 400A°). To decipher the microstructures at higher magnifications, Scanning Electron Microscopy (SEM) on selected heat altered coal/natural coke, Calcined Petroleum Coke (CPC) and graphite samples was carried out. For this purpose, JEOL-840A JSM was used. The samples were scanned under magnifications ranging from 200X to 50000X.
Greater depth of focus obtained in SEM images makes it a promising tool and also makes it possible to use unpolished samples. It facilitates quicker study and different view including 3-D viewing as compared to optical microscopy. Although many researchers are using optical microscopy of coal, natural coke, coke and other minerals, which is a tedious method, here, if SEM is also used the better information can be deciphered. International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 5
With SEM, the features, such as microtextures and microstructures in natural cokes can be studied in greater detailed. As shown in this paper, SEM gives quicker and even better results as compared to optical methods. This technique is also helpful in studying the association and relationship between different rock constituents including coal-magma interaction/alteration giving rise to different types of zones i.e. severely, moderately and slightly altered zones of natural cokes.
Experimental
The parent coals were drawn from seam XIV and XV of Jharia coalfield. Typically, the coals of these seams contain the vitrinite (~50 vol. %), (inertinites ~30 vol. %), liptinites (~1-2 vol. %), mineral matter (15-25 vol. %) and vitrinite reflectance (VRO ~0.95%). However, the heat altered natural coke (jhama) samples derived from these seams contain the unaltered coals 5-10 vol.%, fine mosaics 10-15 vol.%, medium mosaics 10-20 vol.%, coarse mosaics 10-20 vol.%, very coarse mosaics 2-5 vol.%, fine flow 2-5%, medium flow 5-10 vol.%, coarse flow 5-7 vol.%, very coarse flow 3-5 vol.%, isotropic components 15-25 vol.% and mineral matter 20-30 vol.%. Seams XIV and XV are dominated by anisotropic components mostly in the baked portions. These are mostly altered reactive macerals i.e. vitrinites, liptinite, semivitrinites; whereas the inertinite macerals (semifusinites, fusinites, secretinites, macrinites and micrinites) give rise to isotropic domains, sometimes with traces of alterations and basic anisotropy as viewed through polarized light petrological microscope.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 6
Results and Discussion
Application of detailed physical, chemical and petrographic techniques along with XRD, SEM, etc were done on large number of samples derived from intrusion baked zones of different seams in Damodar Valley coalfields (Singh, 1998, Singh et al., 2008, Sharma, 2009). However, in this paper only comparative aspects of features obtained through optical microscopy and SEM have been dealt. Both the techniques here act as complementary to each other in deciphering the details from natural coke, CPC and graphite samples. The features studied are on nature and shape of cracks, fissures and micropores, unaltered and altered coal fragments/remnants, infilling of minerals and their associations in the pores, anisotropic and isotropic components. Association, size and shape of mesophase spheres, mosaics, flow structures and deposited carbon could be studied in greater details using both above techniques. In the figures 1a-16b (attached as photoplates 1-5), different photographs under optical/crossed polarized light (left) and SEM (right) for comparative studies have been shown. The figures 1a-3b in the left, show the images under optical/crossed polarized light microscopy and in the right comparative SEM images. The microstructural features are very clear in the SEM images i.e. the unaltered grains of vitrains, mosaics, flow structures and cracks are more prominent, whereas anisotropy is prominent in optical images.
Mesophase spheres (liquid crystals) in Figure 4a (optical)-4b (SEM) have been better depicted. As can be appreciated, mesophase spheres were identified through anisotropy through optical view, whereas the 3-D view through SEM gave spherical shape to confirm it. In figure 5a, medium to coarse grained mosaics and devolatilization pores have been shown under crossed polarized light, whereas in figure 5b, compact natural International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 7
coke ground mass with prominent differentiation in sizes of pores have been shown through SEM.
In the figures 6a-7a, the natural coke groundmass (left), and mosaics and unaltered inertinite (fusinite) grain under crossed polarized light have been shown, whereas in right side figure 6b&7b, same groundmass with prominent pores and fibrous inerts (fusain) has been identified through SEM. Figure 8a (under crossed polarized light) shows mosaics (heat altered products) at different domain sizes (very fine to coarse) depicting very strong anisotropy along with different sizes of pores, while in figure 8b under SEM, the same features have been highlighted with 3-D view and identity of individual pores becomes more distinct.
In figure 9a, mesopores in natural coke groundmass showing mineral matter infilling under Crossed Polarized light, while the in-filled minerals could be better differentiated using SEM images as shown in Figure 9b. Similarly, in Figure 10a, natural coke groundmass showing mosaics and big size devolatilization pores having infilling of mineral matter has been seen under crossed polarized light have been reported. This was better differentiated as distinct crystal of quartz in the devolatiloization pores at higher magnification using SEM.
Anisotropy in the mosaics developed through heat alteration is generally very prominent in the optical (crossed polarized light) images (figure 11a and figure 12a), whereas the 3D images (Figures 11b & 12b) act as complement showing prominet flow and lamellar structures and their association with pores and other textures and structures. Coarse to very coarse flow and lamellar textures have been shown in figures 12a and 12b.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 8
Anisotropy is prominent in optical image, whil 3-D view under SEM gave better picture with proper orinetation. Different types of folded structures are also prominet in these iamges.
In figures 13a to 14b, association of minerals and their mode of occurrence has been shown under optical as well as SEM images. In figure 13a and 14a, natural coke groundmass showing devolatilization pores with blurred appearance of something included in the pores, while in figure 13b, and 14b as studied in SEM, acts as very helping tool in identifying the minerals, indicate kaolinite (13b) and carbonate and silicate minerals (14b) respectively. Figures 15a and 15b have shown surface features of calcined petroleum cokes (CPC) including mosaics, flow and lamellar textures and pores under SEM, while different sizes and shapes of natural graphite have been shown in Fig. 16a-b.
Conclusion
This study has been helpful in establishing the usefulness of optical microscopy (crossed polarized light) as well as Scanning Electron Microscopy (SEM) for study of heat altered coals giving rise to natural cokes. Different types of microtextures and microstructures and different contact metamorphic associations available in natural cokes, in addition to better identification of minerals have been possible using both the techniques together. Anisotropy/isotropy studies using petrological microscopes under crossed polarized light act as supplementary tool to SEM, which gives better opportunity in identification of minerals and association of minerals and baked coal/natural coke under 3-D view. Through this study it has been observed that combined maceral-mineral analysis can be International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 9
used to advantage in coal geological studies. SEM is more useful in studying the contact metamorphic features and assessing relationships between shape and formation of organic rocks. Effect of magmatic intrusion on coal structure and mineral composition has become easier using SEM and optical microscopy together.
Acknowledgements
The authors are thankful to the Director, Central Institute of Mining and Fuel Research, CSIR, Dhanbad for permitting to publish this paper and to SSRC, Ministry of Coal, Govt. of India, BCCL, Tata Steel, Dhanbad for their financial support in the form of project and also permitting to collect the samples from their leasehold.
References
1. Blanford, W.T., On the geological structure and relations of the Raniganj coalfield, Bengal. Mem. Geol. Surv. Ind., 3(1), 1-196, 1861. 2. Chandra, D., Mineral Resources of India, 5 (Jharia Coalfield), Geological Society of India, Bangalore, P. 149, 1992. 3. Chandra, D. and Srivastava, G.P., Reflectance of burnt coals. Fuel, 57,251-252, 1975. 4. Chatterjee, G.C., and Ghosh, P.K., Tectonic Framework work of the Peninsular Gondwanas of India., Records of the Geological Survey of India, 98 (2), 1-15, 1970. 5. Cooper, J.R., Crelling, J.C., Rimmer, S.M., Whittington, A.G., Coal Metamorphism by Igneous Intrusion in the Raton Basin, CO and NM: Implications for Generation of Volatiles, Int. J. of Coal Geol, 71, 15-27, 2006. 6. Fox, C.S., The Jharia coalfield. Mem. Geol. Surv. Ind. Vol. 56, 1930. International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 10
7. Fox, C.S., The Lower Gondwana coalfields of India. Mem. Geol. Surv. Ind. Vol. 59, 1934. 8. Ghosh, S., Chatterjee, C.N., Mallik, P.C., Regional metamorphism of coal in major coalfields of Damodar valley region, India. In proceeding of IV International Gondwana symposium, 1977, Calcutta, India. Hindustan Publishing Corporation (India), Delhi-110007, 286-298, 1977. 9. Ghose, T.K., A study of temperature condition at igneous contact with certain Permian coal of India. Economic Geology, 62, 109-117, 1967. 10. Golab, A., The impact of igneous intrusions on coal, cleat carbonate, and groundwater composition. Ph.D. thesis, School of Geosciences, University of Wollongong, Australia, 2003. 11. Golab, A. N., Hutton A.C., French D., Petrology, carbonate mineralogy and geochemistry of thermally altered coal in Permian coal measures, Hunter Valley, Australia. International Journal of Coal, 70,150-165, 2007. 12. Kwiecinska, B., Petersen, H.I., Graphite, semi-graphite, natural coke, and natural char classification-ICCP system. Int. J. Coal Geol. 57, 99–116, 2004. 13. Merritt, R. D., Thermal alteration and rank variation of coals in the Matanuska field, south- central Alaska. International Journal of Coal Geology 14, 255-256, 1990. 14. Pascoe, E.H., A manual of the geology and Burma. Govt. of Ind. Publication, 9091016, 1959. 15. Querol, X., Alastuey, A., Lopez-Soler, A., Lana, F., Fernandez-Turiel, J.L., Zeng, R., Xu, W., Zhuang, X., Spiro, B., Geological controls on the mineral matter and trace elements of coals from the Fuxin Basin, Liaoning Province northeast China. International Journal of Coal Geology 57, 99-116, 1997.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 11
16. Sharma, Mamta, characterisation and industrial importance of coal carbonised due to igneous activity in the coalfields of Damodar valley, India, Ph.D. thesis (submitted), Banaras Hindu University, Varanasi, India, 2009. 17. Sharma, Mamta, Innovations in natural coke studies, Indian Science Congress Association (ISCA, Young Scientist Award), Shillong, India, 2009. 18. Singh, Ashok K., Petrology of coals from the fire affected areas of Jharia coalfield, Damodar Valley Bihar (now Jharkhand). Ph. D. thesis (unpublished), Department of Geology, Banaras Hindu University, Varanasi, U.P., India, 1998. 19. Singh, A.K., Sharma, M. and Singh M. P., Genesis of Natural Cokes: Some Indian Examples, International Journal of Coal Geology 75, 40-48, 2008a. 20. Singh, A.K., Singh, Alpana, Singh, B.D., Sharma, M., Kumar, S., Shukla, N.K. and Choudhury, N., Coal Bed Methane genesis and its status in Indian context, Memoir Geological Society of India, 71,149-171,2008b. 21. Singh, A.K., Singh, M.P., Sen, K., New Classification Scheme for Microtextures of Natural Cokes: A Case Study from Heat Affected Coking Coals of Jharia Coalfield, India. 20th International Pittsburgh Coal Conference, USA. 2003. 22. Singh, A.K., Singh, M.P., Sharma, M., Srivastava, S.K., Microstructures and microtextures of natural cokes: A case study of heat-altered coking coals from the Jharia coalfield, India. International Journal of Coal Geology 71, 153-175, 2007. 23. Stewart, A.K., Massey, M., Padgett, P.L., Rimmer, S.M., Hower, J.C., Influence of a basic intrusion on the vitrinite reflectance and Chemistry of the Springfield (No. 5) coal, Harrisburg, Illinois. International Journal of Coal Geology 63, 58– 67. 2005.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Page 12
PHOTOPLATE-1: Optical (left-a) & SEM (right-b) photographs showing details of studied samples at different magnifications.
Fig.1a: Vitrain showing normal grain under reflected light.
Fig.1b: Under SEM vitrain shows ground mass, crack and conchoidal fracture.
Fig.2a: Unaltered vitrain grain and altered mosaics, flow structures and pores under crossed polarized light.
Fig.2b: Natural coke showing pores of different sizes, compact ground mass and unaltered vitrain grain under SEM.
Fig.3a: Unaltered vitrain grain alongwith altered natural coke groundmass under crossed polarized light.
Fig.3b: Unaltered vitrain grain and altered mosaics, flow structures and pores as seen using SEM.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
PHOTOPLATE-2: Optical (left-a) & SEM (right-b) photographs showing details of studied samples at different magnifications.
Fig.4a: Mesophase spheres (liquid crystals) developed in the heat altered coking coal under crossed polarized light.
Fig.4b: Mesophase spheres at higher magnification in natural coke sample studied through SEM.
Fig.5a: Medium to Coarse mosaics and devolatilisation pores under crossed polarized light.
Fig.5b: Compact natural coke groundmass showing different sizes of pores studied through SEM.
Fig.6a: Photomicrograph showing porous natural coke groundmass and unaltered fusain under crossed polarized light.
Fig.6b: Fibrous nature of fusain in natural coke groundmass showing no alteration under SEM.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
PHOTOPLATE-3: Optical (left-a) & SEM (right-b) photographs showing details of studied samples at different magnifications.
Fig.7a: Isotropic fusain (inert) in natural coke groundmass as studied under crossed polarized light.
Fig.7b: Fusain in natural coke groundmass with better details as studied using SEM.
Fig.8a: Natural coke groundmass showing very fine to coarse mosaics and devolatilization pores under crossed polarized light.
Fig.8b: Very fine to very coarse mosaics and devolatilization pores as seen using SEM.
Fig.9a: Mesopores in natural coke groundmass showing mineral matter infilling under Crossed Polarized light.
Fig.9b: Natural coke groundmass showing minerals infilling in the devolatilization pores as studied using SEM.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
PHOTOPLATE-4: Optical (left-a) & SEM (right-b) photographs showing details of studied samples at different magnifications.
Fig.10a: Natural coke groundmass showing mosaics and big size devolatilization pores having infilling of mineral matter as seen under crossed polarized light.
Fig.10b: Infilled mineral matter in the devolatiloization pores at higher magnification using SEM.
Fig.11a:Natural coke groundmass showing fine to coarse flow with prominent anisotropy under crossed polarized light.
Fig.11b:Fine to coarse flow structures with prominent pores as seen using SEM.
Fig.12a: Lameller flow texture in the natural coke groundmass showing very prominent anisotropy under crossed polarized light.
Fig.12b: Lameller flow texture in similar natural coke sample in 3-D view as studied under SEM.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
PHOTOPLATE-5: Optical (left-a) & SEM (right-b) Figs. 13a-14b & only SEM Figs 15a-16b photographs showing details of studied samples at different magnifications.
Fig.13a: Natural coke groundmass showing devolatilization pores with hazy appearance of something sitting in the pores.
Fig.13b: Kaolinite minerals as studied using SEM on natural coke samples at high magnification.
Fig.14a: Mesopores in natural coke groundmass showing hazy appearance of mineral matter infilling under crossed polarized light.
Fig.14b: Natural coke groundmass showing carbonate and silica minerals at higher magnifications using SEM.
Fig.15a:Calcined Petroleum Coke (CPC) showing altered mosaics, flow structures and pores under SEM.
Fig.15b:Same sample at different view under SEM.
Fig.16a: Natural graphite flakes under SEM.
Fig.16b: Same natural graphite at higher magnification using SEM.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Examination of Low Temperature Air Oxidation Mechanism of Brown Coal for Suppressing Self Ignition Tendency Kouichi Miura*, Ryota Okajima, Mitsunori Makino, and Ryuichi Ashida Department of Chemical Engineering, Kyoto University Kyoto 615-8510, JAPAN E-mail: [email protected] Abstract Low rank coals including brown coal and lignite are valuable energy resources in this century. For the storage and transportation of the low rank coals, suppression of self ignition tendency as well as dewatering is essential. The self ignition tendency is closely related to the rate of oxidation of the coals. Therefore it is very important to examine the rate and mechanism of the oxidation of the coals at low temperatures. In this paper the mechanism of air oxidation of an Australian brown coal, Loy Yang, was investigated at less than 300°C. The coal was heated up to an oxidation temperature of 260, 280, or 300 °C at the rate of 10 K/min in an helium stream containing oxygen by either 11 % or 22 %, and kept for o, 30, or 60 min at the oxidation temperature. Weight change, gas formation rate, and fate of surface functional groups during the oxidation were investigated using thermogravimetry, gas chromatography, and in-situ FTIR spectrometry. The amount of gas phase oxygen consumed, ξO2, was estimated from the weight change and the gas formation rate. The oxidized coal was cooled to 65°C in the oxygen atmosphere and then heated up to 700°C at the rate of 10 K/min in the helium stream containing oxygen by 22 % to estimate the gasification rate of the oxidized sample. The self ignition tendency was characterized by the temperature, Tg, at which 10% of the oxidized coal is gasified. The higher heating value (HHV) of the oxidized coal was estimated by the Motto-Spooner equation with the ultimate analysis of the oxidized sample.
1. Introduction Low-rank coals, subbituminous coal and brown coal/lignite, account for about a half of world recoverable coal resource. Low rank is in general associated with low heating value because it contains a large amount of water. Australian brown coal, for example, contains a large amount of water amounting to ~ 60 %. This means that dewatering or drying is essential when brown coals are transported and stored to be utilized. Dewatered brown coals, however, unfortunately tend to have high self ignition tendency as compared to non-dewatered coals, which causes serious problems for transportation and storage of the dewatered brown coals. This is why the low rank coals are currently used just as fuels near coal mines. It is therefore essential to develop effective self ignition suppression methods for utilizing low rank coals worldwide. To do so, it is also
essential to clarify the mechanism of self ignition. It has been reported that various factors, such as chemical structure, mineral content, moisture content, particle size, or surface area of coal, affect the self ignition of low rank coals1). Several works have been performed to examine the mechanism of the self ignition1-3). However, the mechanism has not been completely elucidated and hence effective methods to suppress the self ignition tendency have not been developed yet. In this study air oxidation of an Australian brown coal, Loy Yang, was investigated at less than 300°C to examine the mechanism of self ignition and to find an effective self ignition suppression method. 2. Experimental 2.1 Coal used. An Australian brown coal, Loy Yang (abbreviated to LY), was used as a coal sample. Its ultimate analysis (d.a.f. basis) and proximate analysis (dry basis) are as follows: C: 64.0 %, H: 4.4 %, N: 1.2 %, O: 30.5 % (by difference) and VM: 51.5, FC: 47.0, ash, 1.5 %) was used in this study. The coal was ground into fine particles of less than 53 μm in diameter. Demineralized coal sample (LYdem) was also prepared by treating the ground LY coal by 2N- HCl for 12 h, and it was used for comparison purpose. 2.2 Experimental procedure. A thermogravimetric analyzer (Shimadzu; TGA-50) connected to a micro gas chromatograph (Varian; micro GC CP-4900) was used for continuous analyses of weight change and produced gas during the treatment. Figure 1 shows a schematic of typical experimental run. About 10 mg of coal were placed on a pan made of platinum (6 mm I.D and 2 mm high) in the TG, and the coal sample in the pan was heated up to 260, 280, or 300 °C at the rate of 10 K/min in a helium stream containing either 22 % or 11 % oxygen and held for 0-120 min at the temperature. The oxidized coal was cooled to 65°C in the oxygen atmosphere and then heated up to 700°C at the rate of 10 K/min in the helium stream containing oxygen by 22 % to estimate the gasification rate of the oxidized sample. The self ignition tendency was characterized by the temperature, TG, at which 10% of the oxidized coal is gasified. Weight change and gas formation rates were simultaneously measured by use of the thermobalnace and the gas chromatograph. The amount of gas phase oxygen consumed during the oxidation treatment, ξO2, was estimated from the weight change and the gas formation rates. The oxidized coal at the stage ④ was subjected to the elemental analysis. The fate of surface functional groups during the oxidation was investigated using an in-situ FTIR spectrometry (JEOL; JIR-WINSPEC50) at four stages given as ①, ②, ③, and ④ in Fig. 1. The higher heating value (HHV) of the oxidized coal was estimated by the Motto-Spooner equation with the ultimate analysis of the oxidized sample. The treated coals are referred to by their treatment condition; the oxidation temperature (°C), isothermal period (min) of oxidation treatment, oxygen concentration (%). LY(300, 30, 22), for example, indicates the treaded coal oxidized at 300 °C for 30 min in a helium stream containing 22 % oxygen. In total 10 treated coals were prepared from LY under different treatment conditions. Only LYdem (300, 30, 20) was prepared from the demineralized LY for comparison purpose.
Gasification initiation temperature, TG [℃]
O222 % / He or O211 % / He
600 400
② 200
Drying 110℃
0
Oxidation
③
O222 % / He Gasification
④
300℃, 30min
①
65℃ [mg/min] Gasガス生成速度 formation rate [kg/(kg min)]
Relative weight [wt%-d.a.f.] [wt% (d.a.f)] 相対質量
Temperature 温度 [℃] [ ℃ ]
800
0.4 0.04
100
Relative weight
75 50
CO2 H2O
25
CO
-10 %
0.3 0.03 0.02 0.2
Oxygen consumed, ξO2
0
0.01 0.1 00
0
50
100 時間[min] [min] Time
150
200
Figure 1. Changes in weight and gas formation rates during a typical oxidation treatment and gasification. 3. Results and Discussion 3.1 Self ignition tendency in relation to yields and heating values. The lower panel of Fig. 1 shows the weight change, gas formation rates, and the amount of oxygen consumed during the preparation of LY(300, 30, 22) and its gasification. The weight started to decrease accompanying the formation of CO2 and H2O at around 200°C, and its decreasing rate reached a maximum when the temperature reached the oxidation temperature and then gradually decreased with increasing oxidation time at the oxidation temperature. The net change during the oxidation was the consumption of carbon and hydrogen of the coal. The yields and elemental composition of the oxidized coal were estimated from the measurements and the ultimate analysis of the oxidized coal. Figure 2 shows the yields and elemental compositions for all the oxidized coals prepared. The yields of oxidized coals tended to decrease with the increase of severity of oxidation conditions, namely with the increase of oxidation temperature, holding time during the oxidation, and the oxygen concentration of the oxidizing agent. It is clearly shown that carbon and hydrogen are lost through the oxidation treatment, but the amount of oxygen was kept almost constant during the oxidation. Figure 3 compares the TG values between the treated coals and LY coal. The TG values of the treated coals were all higher than the TG value of LY coal, indicating that the low temperature oxidation is effective to suppress the self ignition tendency of LY coal. The TG values also tended to
C LY
H N
50.1
LY(280, 30, 22)
51.7
LY(280, 60, 22)
1.8
59.1
LY(300, 15, 22)
1.9
51.6
LY(300, 30, 22)
1.4
47.6
LY(300, 15, 11)
52.9
LY(300, 30, 11)
51.1
LYdem(300, 30, 22)
1.3
48.9 0
20
29.0
1.9
27.7
2.6
1.5
1.6 1.0
2.2 1.6
29.0
1.5
1.6
48.2
LY(300, 0, 22)
29.0
1.3
1.9
30.5
29.2
1.6
2.3
53.6
LY(260, 120, 22)
1.2
2.6
57.4
LY(260, 60, 22)
1.2
4.4
64.0
LY(260, 30, 22)
O
27.6
26.2 27.1
1.4 1.4 0.7
40 60 Yield [wt%-d.a.f.]
27.8 26.0 30.3 80
100
相対質量 Relative weight[wt%-d.a.f.] [wt% (d.a.f)]
Figure 2. Yields and elemental compositions of the treated coals.
LY(280, 30, 22) LY(300, 30, 11)
100
LY(300, 30, 22)
LY 90
LYdem(300, 30, 22) LY(300, 15, 22) LY(260, 30, 22)
80
LY(300, 0, 22) 70℃
70 150
200
250
T
300 G 350 温度[min] [℃] Time
400
450
500
Figure 3. Gasification profiles and the gasification initiation temperatures, TG, of the treated coals. increase with the severity of oxidation conditions.
The TG value of LY(300, 30, 22) was 384°C which was 60°C higher than the TG value of LY and was the highest among the treated coals prepared from LY. The TG value of LYdem(300,30, 22) was 391°C and was 7°C higher than the TG value of LY(300, 30, 22). This means that some minerals in LY may enhance the self ignition tendency. The severity of oxidation is expected to be represented by the amount of oxygen consumed during the oxidation, ξO2. Then the TG values of the treated coals were plotted against the ξO2 values in Figure 4. The TG values increased almost directly proportional to the ξO2 values for the treated coals prepared from LY, indicating that ξO2 well can be an index of the degree of
Gasification initiation temperature, TG [℃]
420 LYdem(300, 30, 22)
400
LY(300, 30, 22) LY(300, 30, 11) LY(300, 15, 22) LY(300, 15, 11)
380 360
LY(280, 60, 22) LY(260, 120, 22) LY(280, 30, 22) LY(260, 60, 22) LY(260, 30, 22)
LY(300, 0, 22)
340
LYdem LY
320 0
20
40
60
80
Amount of oxygen consumed, ξO [wt%] 2
Figure 4. Relationship between the gasification initiation temperatures, TG, and the amounts of oxygen consumed during the oxidation, ξO2 of the treated coals.
Heating value [MJ/kg-coal (d.a.f.)]
25 Demineralization
LY(300, 15, 11) LYdem(300, 30, 22) LY(300, 15, 22)
20 LY
LYdem
15 Demineralization
LY(300, 0, 22)
10
LY(260, 60, 22) LY(260, 30, 22) LY(280, 30, 22)
5
LY(300, 30, 22) LY(280, 60, 22) LY(300, 30, 11)
LY(260, 120, 22)
0 320
340 360 380 T [℃] ガス化開始温度 G Gasification initiation temperature, TG [℃]
400
Figure 5. Relationship between the heating value and the gasification initiation temperatures, TG, of the treated coals. suppression of self ignition tendency. The relationship of TG and ξO2 for LYdem(300,30, 22) deviated from the linear relationship. This shows that removing minerals which may enhance the self ignition tendency increases TG value at the same oxidation degree. The low temperature oxidation was found to be effective to suppress the self ignition tendency of LY coal as clearly shown in Figs. 3 and 4. However, the yields and the amounts of carbon and hydrogen decrease with the increase of the severity of oxidation as shown in Fig.2, which means the heating values decrease with the increase of the severity of oxidation. It is desirable to minimize
the decrease of heating value through the oxidation treatment. Then the relationship between the higher heating value of the treated coals based on the raw coal basis and the self ignition
Temperature 温度 [℃][ ℃ ]
temperature of the treated coals was examined in Figure 5. The heating values almost linearly decreased with the increase of TG for the treated coals prepared from LY. The heating value of LY(300, 30, 22) was 15 MJ/kg-coal and was only 63% of the heating value of raw LY coal. This indicates that there exists a trade-off between the suppression of self ignition and the loss of heating value through the oxidation treatment. If we see Fig. 5 more closely, the heating values of the treated coals prepared at higher temperatures tend to have slightly larger heating values when compared at a same TG value. This indicates that the oxidation treatment at higher temperatures is advantageous from the viewpoint of heating value.
800 600
②
400 200
300℃, 30min
110℃
③ ④
①
0 0
65℃
50
100
150
200
2500
2000
K-M function function[a.u.] [a.u.] K-M
Time [min] 時間 [min] aliphatic C-H -OH
① ②
aromatic C-H
④ ③
4000
3500
3000 -1 Wavenumber [cm-1] ] Wavenumber [cm
(a) =
=
ーCーOー ーCーOH
O
=
K-M function [a.u.]
1900
ーCーO-C-
O
O
=
O
=
O
ーCーH
④ ③ ② ①
aromatic C=C
highly conjugated carbonyl anion -COO-
1800
1700
1600
1500
-1]-1] Wavenumber 波数 [cm[cm
(b) Figure 6. In-situ FTIR spectra measured at the four stages during the preparation of LY(300, 30, 22).
3.2 Examination of mechanism of low temperature oxidation. The in-situ FTIR spectra measured at the four stages during the preparation of LY(300, 30, 22) are shown in Figures 6a and 6b. Fig. 6a clearly shows that the amounts of aliphatic C-H and OH groups significantly decrease with the progress of oxidation. On the other hand, Fig. 6b clearly shows that the amounts of anhydrides and esters significantly increase with the progress of oxidation. Quantitative changes of these moieties with the progress of oxidation were examined by deconvoluting the FTIR spectra shown in Figs. 6a and 6b. Figure 7 shows the changes in the amounts of these moieties against the treatment time. It is clearly shown that the amounts of aliphatic C-H and OH groups decrease , whereas the amounts of anhydrides and esters increase during the oxidation . The amounts of carboxylic groups and aldehydes decreased slightly with progress of oxidation. The changes in the amounts of the moieties with the progress of oxidation are consistent with the oxidation mechanism proposed by Miura et al.4) and Worasuwannarak5) (Figure 8). The first step of the oxidation is the formation of peroxides by the insertion of oxygen between C and H of aliphatic C-H groups. The peroxide is then converted to aldehyde, carboxyl, anhydride, and finally to ketone with the progress of oxidation by one mechanism. It is converted to ester and finally to CO2 by another mechanism. Esters and anhydrides are much more resistant to oxidation. This is why the oxidation treatment is effective to suppress the self ignition tendency of LY coal.
Area Area[a.u.] [a.u.]
Drying
① ②
Oxidation aliphatic C-H
aromatic C-H
④ -OH
anhydride
ester
Area Area[a.u.] [a.u.]
Cooling
③
-COOH -CHO 0
20
40 60 [min] 時間[min] Time
80
100
Figure 7. Changes in the amounts of several moieties of coal against the treatment time.
Figure 8. The oxidation mechanism proposed by Miura et al.4) and Worasuwannarak5)
4. Conclusion Air oxidation of an Australian brown coal, Loy Yang, was investigated at less than 300°C to examine the mechanism of self ignition and to find an effective self ignition suppression method. The oxygen content remained interestingly almost unchanged during the oxidation. The yield of the oxidized coal decreased from 92 to 77 % on d.a.f. basis depending on the severity of the oxidation conditions. The net HHV values also decreased from 23. 8 MJ/kg-coal to 22.0 - 17.9 MJ/kg-coal depending on the severity of the oxidation conditions. The self ignition index temperature, TG, on the other hand, increased from 323°C of the raw coal to 345 – 392 °C of the oxidized coal with the increase of the severity of oxidation, indicating that the oxidation suppressed the self ignition tendency significantly. The TG value increased with the increase of the amount of oxygen consumed for the oxidation, ξO2, which confirmed that the degree of the suppression of self ignition tendency is well correlated with the degree of oxidation as expected. The in-situ FTIR measurement clarified that the aliphatic C-H groups of the coal was selectively oxidized to form either anhydrides or CO2 via aldehydes, esters, and carboxyls accompanying H2O formation. These results showed that the oxidation converted reactive C-H groups to stable anhydrides, suppressing the self ignition tendency. This oxidation mechanism was well supported by the above findings. Thus it was clarified that the air oxidation suppresses the self ignition tendency of brown coal by converting reactive C-H groups to rather stable anhydrides. This suggests that the pre-oxidation at less than 300 °C would be an effective method to suppress the self ignition tendency of brown coal. References 1) V. Calemma, R. Rausa, R.Margarit, and E. Girardi, FUEL, 67 (1988), 764-770. 2) A.H. Clemens, T.W. Matheson, and D. E. Rogers, FUEL, 70 (1991), 215-221. 3) H. Wang, B.Z. Dlugogorski, and E. M. Kennedy, Prog. Energy Combust. Sci., 29 (2003) 487-513. 4) K. Miura, H. Nakagawa, and K. Hashimoto, Carbon, 33 (1995), 275-282. 5) N. Worasuwannarak, PhD Dissertation, Kyoto University, Chapter 2 (2003).
Beneficiation prospects of baked coking coals from seam XV, Jharia Coalfield, Damodar Valley, India #
Ashok K Singh, N. K. Shukla, *Mamta Sharma, N. Choudhury
Central Institute of Mining and Fuel Research, CSIR, Govt. of India, Dhanbad-828108, India *National Metallurgical Laboratory, CSIR, Govt. of India, Jamshedpur-831007, India (#corresponding author: [email protected])
Abstract Baked coals in the Jharia Coalfield, due to its irregular physical and chemical characteristics such as hardness, loss in caking property and slow in burning due to loss of volatile matter, have got very poor response from utilization point of view in industry. In this attempt the baked coals from seam XV, lying mostly unused in Jharia coalfield, have been beneficiated through F & S in the laboratories at CIMFR, Dhanbad, India. The washability investigations were carried out at the size fractions 50-25, 25-13, 13-06, 06-03 and 03-0.5 mm and at the specific gravity range of 1.4 to 1.8 at a difference of 0.05 and 0.1. The washability data developed on the size 50-0.5 mm has been presented through conventional washability curve and Mayer’s curve and it is observed that at 11.9% ash level 25.8% cleans may be used for carbon artifact industry and the rejects (74.2%) at 23.7% ash level may be used in Indian Power Plants after judicious blending with high VM power coals.
Key words: Baked coals, beneficiation, washability curve, Mayer’s curve, carbon artifact.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
Introduction
Coal is a heterogeneous mixture of organic (macerals) and inorganic (minerals) matter. The inorganic components form potential impurities of coal. It becomes extremely crucial to understand the quality and quantity of inorganic matter for the potential use of organic part. There exists a large difference in density between coal macerals and mineral matters. While density of major macerals lies between 1.23 and 1.72, the densities of major minerals are above 2.3. Therefore, density based separation processes are most common in coal preparation, which is referred to as ‘Washability’ of coal. Washability is carried out through float and sink test of coal at different sizes and different densities. Petrographic and chemical analyses of each fraction obtained through Float-sink provide characteristics of such fractions and their relevance vis-à-vis downstream processing and utilization. Scanty work on beneficiation and end use of natural coke has been carried out, mention may be made of Sanyal, (1965); Singh, (1998); Singh, and Singh, (1989); Singh, et.al, (2006a); Singh, et.al, (2006b, 2006c); Singh, et. al. al., 2008, 2009, Roy et.al, 2008.
Location and Geological Setting
The sample was collected from seam XV of Digwadih Colliery, which is situated in the southeastern part of Jharia coalfield (latitude 230 41’40” to 230 43’ 0” and longitude 860 23’ 30” to 860 24’30). The thickness of seam was approx 3.0 meter and the depth of seam was around 150 meter from the surface. In this mine, underground mining was
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
2
being done through Board and Pillar system. The dip of the seam was about 1 in 7. These heat altered coals/natural cokes are of Barakar Formation under Damuda Group of Lower Gondwana Super group (Sanyal, 1965).
Methodology Sampling
The sample was collected from above colliery following the BIS procedures (IS-436). Due to lack of market, the production of natural coke was very low and on irregular basis from the seam XV in this colliery and it was about 850 t/month. Approx. 3.0 tonnes of sample of the size mostly <50 mm was collected from stored coal near pithead following standard procedures and was transported to CIMFR for detailed washability investigation.
Experimental Sample preparation For bulk sample
The sample received at CIMFR, Dhanbad was initially air dried and crushed to 50 mm in a continuously operated single roll crusher. The crushed sample was thoroughly mixed and after proper sampling it was divided into four parts. Out of these four parts, one part was considered as head sample and characterized for proximate analysis following BIS procedures (IS-1350, Pt-I). The second part was subjected to size International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
3
analysis in 50, 25, 13, 6, 3 and 0.5 mm standard screens. The third part was used for float and sink studies as per BIS-13810 of 1993. Head sample as well as individual size fractions were further characterized for proximate, ultimate, GCV, CO2 and petrography following respective BIS procedures. The fourth part was kept as reserve. The work plan is shown in figure 1.
Size analysis
The crushed coal/jhama was subjected to screen analysis at 50, 25, 13, 6, 3 and 0.5 mm. The results are reported in Figure 2.
Float and Sink test
The sample (crushed to -50 mm) was subjected to Float-sink studies on the sizes 5025, 25-13, 13-06, 06-03, 03-0.5. Each of the individual size fractions were subjected to sequential float and sink tests with the relative density range from 1.30 to 1.80 at interval of 0.05 and 0.1. Weight and ash content of all gravity fractions were noted and the same is shown in table 1.The washability data of the fractions between 50-0.5 mm has been shown in table 1 and figures 3 and 4. The data obtained were used for plotting washability curves for the sample to establish the washability characteristics. Since all good coal is expected to be floated at 1.4 to 1.80 specific gravity, a comparison is made in all cases for gross yield and ash level up to 1.6 specific gravity. This would indicate preferential response of some heat altered coal to float and sink. International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
4
The minus 0.5mm coal fraction was not subjected to float and sink tests, keeping in view the difficulties encountered for separation of ultra fines.
Result and Discussion
Chemical and Petrographic characteristics Head Samples
Head sample was characterized through proximate (air dried basis), ultimate analyses, petrographic analyses, HGI, GCV, ash analysis, real density etc. On air dried basis the natural coke of study area contained moisture 2.4%, VM 6.5% (8.4% daf), FC 71.1% (91.6% daf), GCV 6050 kcal/kg and real density ~1.77. It is indicative of almost complete conversion of coaly material to natural coke after baking. This is also depicted through very low hydrogen 1.45% (1.87% daf), high elemental carbon 69.92% (90.10% daf), very less reactives macerals 2.9% (4.3% vmmf) and high heat altered maceral constituents 49.3% (72.1% vmmf).
Among oxides in the ash, the SiO2
(34.78%), Al2O3 (31.7%) and Fe2O3 (9.38%) are the main contributors.
Screen Analysis and Washability ·
Screen analysis depicts that -0.5 mm size fractions have high ash content (24.2% on dry basis) with respect to higher fraction because of segregation of mineral matter at finer sizes (Figure 2).
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
5
·
Characterization of different specific gravity fractions (1.40 to 1.80) of 50-25, 25-13, 13-06, 06-03 and 03-0.5 mm sample was done. Through washability studies of 50-25 mm size fraction, it has been observed that about 64% of the wt. is confined to 1.45 to 1.80 sp. gr. The higher moisture contents (3.8 to 5.1%) in the lower specific gravity ranges 1.40 to 1.60 was probably due to higher porosity providing locales for water retention and hygroscopic nature of natural coke. At 12.9% ash the cumulative float was 23.7% and at 16.3% ash the cumulative float was 76.5% (Table 1). In the size range 25-13mm, 68.6% wt was confined to the sp. gr. between 1.45 and 1.80. Initial gravity fractions (1.40-1.60) show higher moisture contents (3.4 to 4.3%). At 12.3% ash the cumulative float was 26.8% and at 15.2% ash the cumulative float was 82.6%. In the size range 13-06mm, 68.8% wt was confined to 1.45 to 1.80 sp. gr. Higher moisture was confined to sp. gr. 1.40 to 1.60. At 11.1% ash the cumulative float was 23.9% and at 13.9% ash the cumulative floats was 81.8%. For the size range 06-03 mm, 66.1% wt was confined between 1.45 and 1.80 sp. gr. Moisture showed decreasing trend probably due to lesser porosity. At 10.7% ash the cumulative float was 27.4% and at 13.0% ash the cumulative float was 80.1%. In the size range 03-0.5 mm at the sp. gr. 1.4 the wt% was slightly higher (11.9%) as compared to other size fractions. 62.8% wt was confined between 1.45 and 1.80 sp. gr. The moisture again showed lower values. At 10.6% ash the cumulative float was 55.5% and at 12.0% ash the cumulative float was 80.9% (Table 1).
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
6
·
At 50-0.5 mm size, sample depicted wt% 5.4 at the ash 10.1%. At cumulative wt 49.7% the ash content was 13% and at 79.4 wt% the ash was 14.9% only. (Table 1, Fig. 2 and 3).
·
For the size fraction 50-25 mm, the VM showed initially decreasing trend from 7.7% (9.2% daf) to 5.3% (6.6% daf) but finally increased up to 10.6% (19.6% daf) due to presence of CO2 up to 7.92%. Fixed carbon in this fraction was high up to 76.6% (92.0% daf), while elemental carbon around 74% (90.0% daf). In the sp. gr. range 1.41.8 the GCV was 6860-5780 kcal/kg. Low VM 7.7% (9.2% daf) to 5.3% (6.6% daf) and hydrogen (2.61%-1.12%) and higher real density (1.6-1.9) are indicative of significant alteration of coking coal into natural coke.
·
The fraction 25-13mm showed anomalous VM due to considerable CO2 gas (1.30% to 7.15%). Fixed carbon was slightly higher than previous fraction and ranged from 73.2% (93.6% daf) to 79.2% (93.6% daf), while elemental carbon between 74% (94.5% daf) and 77% (91.5% daf). GCV was also higher than previous fraction and ranged between 5950 and 6930 kcal/kg.
·
The size fraction 13-06 mm also showed erratic VM with VM contributing CO2 ranging from 1.23 to 8.18%. FC ranged from 75.3% (94.0% daf) to 80.9% (92.7% daf) and elemental carbon 75.85% (94.7% daf) to 81.09% (92.9% daf). Hydrogen ranged from 2.84 to 1.42%). The GCV also increased as compared to other fractions and varied between 6120 kcal/kg (sp.gr. 1.8) and 7170 kcal/kg (sp.gr. 1.4).
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
7
·
06-03 mm size fraction depicted VM 7.2% (8.3% daf) decreased up to 3.3% (3.9% daf) again increased up to 11% (18.2% daf) at sp.gr. >1.8. The reason being higher carbonate minerals contributing CO2 (1.24 to 6.47%). In the sp.gr. range of 1.8 to 1.4 the FC 77.6% (94.6% daf) to 83.1% (93.5% daf) and elemental carbon 78.03% (95.2% daf) to 83.48% (93.9% daf) showed increase as compared to other higher size fractions.
·
Similarly, for the size fraction 03-0.5mm, VM showed anomalous result and was corrected through determined CO2, ranging from 0.87 to 8.39%. With decreasing sp.gr. from 1.8 to 1.4 the FC 77.8% (93.8% daf) to 83.6% (93.4% daf) and elemental carbon 73.84% (89.1% daf) to 82.34% (92.0% daf) showed increasing trend.
·
Petrographically, the size fraction 03-0.5 mm depicted the distribution of unaltered macerals (reactives) in meager quantity 0.7% -3.4% (1.1% -5.2% on vmmf basis), inertinites 24.1%
to 6.4% (33.7%-8.8% on vmmf basis), mineral matter 25.1% to
35.6% and the heat altered maceral components showing strong anisotropy and different structures (very fine to very coarse and flow structures) varied between 41.1% and 65.0% (62.2 and 89.6% on vmmf basis).
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
8
Conclusion and recommendations
The results of this exercise suggest that the unaltered coaly parts could be separated at specific gravity <1.40, the main anisotropic constituents (carbonaceous material) could be enriched within specific gravity range of 1.50 to 1.70 with better yields in finer fractions. Moreover, these experiments have ruled out the earlier projections that heat altered coal / jhama is not washable through normal float and sink tests, instead the proposed procedure is competent to liberate the mineral matter and provides enough scope to enrich the natural coke fraction. In addition, the enriched natural coke fraction could be used for graphitization and other industrial applications, as can be inferred based on various chemical and petrographic studies. In terms of quantitative data, the jhama of XIV seam can be beneficiated up to 19.8% cumulative float at 10% ash. With the help of washability study of the heat altered coal samples, it has been shown that these coals can be beneficiated with optimum yield and natural coke rich fractions could be made suitable for industrial uses, particularly in carbon artifacts, power plants and cement industries. Further, natural coke at very low ash can be considered for making artificial graphite (alone or even in blends with Calcined Petroleum Coke and Low Ash Metallurgical Coke). At the desired level of ash content (10% and 15%), the yield of >20% and >80% respectively was obtained and most of the anisotropic carbon was confined to specific gravity 1.40 to 1.70 (around 70%) in almost all size ranges.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
9
Acknowledgment
The authors are thankful to the Director CIMFR, CSIR, Government of India; Dhanbad for providing the laboratory facility to carry out this research work and permitting to publish this work. The financial support received from SSRC, Ministry of Coal, Govt. of India is thankfully acknowledged. The cooperation from the officers and staff members of concerned colliery of Jharia Coalfield, during fieldwork is duly acknowledged. The authors wish to acknowledge the S&T support obtained from different divisions especially Coal Preparation Division of CIMFR.
Reference
1. Roy B. N., Singh A. K., Prasad N., Shukla N. K., Sharma Mamta, Prospect of Jhama coal of Damodar Valley Coalfield of India for its Industrial application, Proceedings, national Seminar on Energy Resources at Birsa Institute of Technology (BIT), Dhanbad; February 2008. 2. Sanyal, S. P. (1965). Nature of thin vein of solidified tarry matter formed during natural carbonization of coal from the Victoria West Colliery, Raniganj Coalfield, India, Fuel, 44(5): 333-335. 3. Singh, A.K. (1998): Petrology of coals from the fire affected areas of Jharia coalfield, Damodar valley Bihar, unpublished Ph. D. thesis, Department of Geology, Banaras Hindu University, Varanasi, U.P., India.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
10
4. Singh, A. K., Shukla, N. K., Mukherjee, A., Sharma, Mamta, Choudhury, N., Gauricharan, T., Haldar, D.D., (2006), Cleaning Potential of natural coke (jhama) through washability investigation and suitability for different end uses. Proceedings, International Pittsburgh Coal Conference, Pittsburgh, USA. 5. Singh, A. K., Shukla, N. K., Sharma, M., Singh, V., Gauricharan, T., Haldar, D. D., Basu, C. K., Srivastava, S. K., (2006), Studies on exploring new dimensions for utilizing heat affected coal as a supplementary resource in Indian industries. Proceedings, G.T.F.M. 24 Nov. 2006 held at ISM, Dhanbad. 6. Singh, A. K., Shukla, N. K., Srivastava. K., Haldar, D. D., Roy, B. N., and Sharma Mamta, 2009, Precautionary measures in determining Volatile matter in natural coke Washability fractions, International Journal of Coal Preparation and Utilization, Taylor and Francis, 29: 34–47, 7. Singh Ashok K., Shukla N. K., Roy B. N., Sharma Mamta and Bhattacharya K. K., 2008, State of the art in graphitization of carbonaceous materials (Invited Paper) Indian Carbon Society, National workshop on Modern Carbon Products and Processing held at NPL, New Delhi on 3rd March 2008. 8. Singh Ashok K., Shukla N.K, Sharma Mamta, Choudhury S., Roy B.N. and Srivastava S.K, Genesis of natural coke and its industrial utilization, Proceedings 23rd International Pittsburgh Coal Conference (25-28 September 2006) held at Pittsburgh, USA. 9. Singh, T. N., and Singh, R. (1989): Scope of wining and utilisation of jhama-The burnt coal. Int. Symp. On Coal utilisation-trends and challenges, sec 14-19, CFRI, Dhanbad.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
11
Raw Jhama Screen at 50 mm
+50 mm
Head Sample
(Proximate Analysis, Ultimate Analysis, Calorific Value, Petrographic Analysis)
Crushed to -50 mm
Size Analysis 50, 25, 13, 6, 3 and 0.5 in mm
-50 mm
Float and Sink Test at Sp. Gr. 1.40 to 1.80 for each size fraction
Reserve
Ash, Moisture (Proximate Analysis, Ultimate Analysis, Calorific Value, Petrographic Analysis)
Figure 1: Work plan for detailed study of Jhama
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
12
Sp. Gr.
Float
Cum. Float
Cum. Sink
Ch. Wt%
Mayer's
Wt.%
Ash%
Wt.%
Ash%
Wt.%
Ash%
pt. value
<1.40
5.4
10.1
5.4
10.1
94.6
21.3
3.7
0.5
1.40-1.45
8.2
11.9
13.6
11.2
86.4
22.2
9.5
1.5
1.45-1.50
12.2
12.6
25.8
11.9
74.2
23.7
19.7
3.1
1.50-1.60
23.9
14.2
49.7
13.0
50.3
28.3
37.8
6.5
1.60-1.70
16.4
16.4
66.1
13.8
33.9
34.0
57.9
9.1
1.70-1.80
13.3
20.1
79.4
14.9
20.6
43.0
72.8
11.8
>1.80
20.6
43.0
100.0
20.7
89.7
20.7
100
20.7
Table 1. Washability data of Seam XV, JCF (ROM, natural coke crushed to 50mm) size tested 50-0.5mm
40 35
Ash % of diff. fraction
Ash(dry) %
30 25 20 15 10 5 0 50-25(39.6) 25-13(20.8) 13---6(14.3) 6--3(8.0)
3-.5(9.2)
_0.5(8.1)
Size (wt %) Figure 2: Result of screen analysis (mm)
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
13
Conventional washability curves 2.00 0.0
1.90
1.80
Sp.Gravity 1.70 1.60
II
III
1.50
1.40
1.30
10.0 20.0
IV
I
Cleans-Yield%
30.0 40.0 IV.YldGrav I.float
50.0 60.0
II.char-wt
70.0
III.sink
80.0 90.0 100.0 0
10
20
Ash% 30
40
50
Figure 3: Conventional Washability Curve (ROM, natural coke crushed to 50mm) size tested 50-0.5mm
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
14
Mayer's washability curve 2.00 0.0
Sp. Gravity 1.60
1.80
1.40
1.20
10.0 20.0 30.0
%Yield
40.0 Yld-grav
50.0 60.0 70.0 80.0 90.0
100.0 0
5
10
%Ash
15
20
25
Figure 4: Mayer’s Curve (ROM jhama crushed to 50mm), size tested 50-0.5mm.
International Pittsburgh Coal Conference, October, 2010, Istanbul, Turkey
15
Emissions from cofiring coal with renewable materials such as biomass and sewage
Dr Lesley Sloss FRSC FIEnvSci Principal Environmental Consultant IEA Clean Coal Centre Gemini House, 10-18 Putney Hill, London SW15 6AA UK Phone: 01592 581132 Fax: 01592 713100 [email protected]
Abstract: The practice of co-firing biomass with coal in full-scale coal utility plants is increasing due to the benefits with respect to reduced fossil fuel-based CO2 emissions. Biomass also tends to have a lower sulphur content than coal and therefore emissions of SO2 are reduced. The same is true for NOx emissions from lower fuel nitrogen content. Further, the lower flame temperatures and different combustion stoichiometry of biomass can result in lower thermal NOx production. The exception to this is the combustion of sewage sludge which may contain significantly more nitrogen than coal. However, co-firing of coal with sewage sludge helps to reduce overall NOx emissions. A reduction in ash, especially when co-firing wood, is another advantage of co-firing. Increased chlorine and/or changes in ash particle adsorbency can help reduce trace element emissions such as mercury and arsenic. However, some biomass materials, such as straws and grasses, can have higher potassium and chlorine than coal which may lead to problems such as slagging and fouling. There are also potential issues with respect to changes in the operation of pollution control technologies. For example, in some situations, the formation of dioxins can result from co-firing biomass in less than optimal conditions. Phosphorus in sewage sludge can react with lime to reduce sulphur capture in fluidised bed systems. Sewage sludge tends to have higher concentrations of several trace elements such as Cr, Cu, Ni, Pb, Zn and Fe than coal. It has been shown that these elements tend to end up in the fly ash and it is important to ensure that this has no detrimental effects on the intended use of such ash. In most cases, the balance between the characteristics of the coal and biomass and the plant operation can control any detrimental plant effects. In practice, full-scale coal-fired plants such as Drax in the UK and Fiume Santo in Sardinia note little or no detrimental change in trace element emissions following the introduction of biomass as a co-fuel. It would appear that, for the most part, the benefits of co-firing biomass far outweigh any negative effects. In fact, it would seem that the majority of environmental impact assessments regard the production, transport and preparation of the biomass fuels more important than changes in the stack emissions from the plant as a result of the co-firing. Detrimental effects, however, can be an issue for ash management.
Introduction In some areas or situations, biomass may be considered as carbon-neutral and is therefore growing in popularity as an alternative fuel in situations where carbon reduction is a priority. According to Koppejan and Baxter (2005), replacing 5% of coal with biomass on an energy basis worldwide would produce 40 GWe of power and an emission reduction of 300 Mt/y. In 2004, cofiring of biomass with coal was employed at 135 plants worldwide, with most of these being situated in the USA, Germany and Finland (Leckner, 2007). The use of biomass is largely dependant on the availability of the fuel as well as any legislative or economic incentive in place. Biomass materials are generally cellulose based and therefore physically and chemically quite different to coal. Since these materials are also highly variable (from straw materials and agricultural wastes, to sewage sludge, plastics and other refuse derived fuels) each poses a different challenge with respect to efficient and clean combustion. The majority of the challenges are with fuel delivery to the combustor, and potential slagging, fouling and corrosion issues at the plant. There seem to be few, if any, issues being raised with respect to increased emission levels of major elements such as particulates, SO2 or NOx. In fact, the increased efficiency of combustion of biomass with coal and the presence of emissions control systems on coal-fired plants mean that detrimental emissions from waste/biomass combustion alone can be reduced or avoided. However, there are still issues with changes in ash characteristics to be dealt with. Further, the inclusion of waste materials in full-scale coal-fired units may, in some situations mean that new and more stringent emission standards must be met (Sloss, 2010). SO2 emissions It has previously been noted (Fernando, 2005), SO2 emissions “invariably” decrease during co-combustion, often in proportion to the amount of biomass used. This is largely due to the fact that biomass material generally contains less sulphur than coal. One exception is sewage sludge, which may contain significantly more sulphur than coal due to the flocculate materials added as a part of sewage sludge processing. However, this higher sulphur content does not necessarily result in higher emissions as the alkali and alkaline earth products also present in sewage sludge can help capture SO2 in solid and liquid waste streams. It is therefore important that systems cofiring sewage sludge measure and evaluate the SO 2 in the flue gases and, if necessary, take measures to control any potential increase in emissions. Some waste tyres may also contain “high” levels of sulphur (0.9% on a dry basis) which may be higher than in some coals. However, any increase in emissions as a result of cofiring would be relatively minimal (<10%) and could be handled by existing flue gas desulphurisation (FGD) units (Sloss, 2010). NOx emission NOx emissions arise from both fuel nitrogen and the nitrogen in the combustion air and so the nitrogen content of any fuel can not generally be used to predict NOx emissions. Most biomass materials contain lower nitrogen contents than coal but give rise to greater NOx emissions when fired alone due to the formation of greater amounts of thermal NOx. However, although some biomass fuels can lead to an increase in NOx emissions when cofired with coal, the majority of published literature would suggest that emissions commonly decrease. This is due to the more efficient combustion conditions provided in larger coal-fired units. The increased volatile matter in the biomass matter, compared to coal, can cause
the rapid reduction of O2 which in turn reduces the rate of formation of NOx from fuel nitrogen. The larger particle sizes which may be present in biomass fuels may also help to reduce the excess air which also reduces NOx formation The NOx reduction effect is often regarded as beneficial for coal-fired power plants and cofiring of biomass is considered to have “intrinsic NOx reduction potential” (Di Nola and others, 2009). The high calorific value and low nitrogen content of waste tyres make them ideal as a reburn/cofiring fuel to reduce emissions from coal combustion. Tests at an 80 kWth plant in the UK showed that tyres were more efficient as reburn fuel than coal (33% better) or gas (53% better) (Singh and others, 2009). The cofiring of biomass with coal can lead to potential problems with SCR (selective catalytic reduction) catalysts being poisoned with alkali or alkaline-earth metals (such as potassium chlorides and sulphates) (Fernando, 2007). Halogen emissions The major issue with halogen in biomass materials is increased plant corrosion. The range of chlorine content of biomass materials is far larger than that in coal. The chlorine contents of RDF (refuse derived waste) and MSW (municipal solid waste) can be up to 1% Cl which is ten times greater than in typical bituminous coals (Leckner, 2007). Sewage sludge has a low Cl content than coal. Despite some biomass materials containing problematic levels of chlorine (with respect to plant corrosion) there have been no reports published of increases in halogen emissions from co-combustion units. However, although the halogens may not themselves be an emission issue, they can significantly affect the behaviour of trace elements during coal combustion. Higher chlorine contents can be used as a control measure for mercury as they lead to greater proportions of the oxidised mercury form which can be captured in existing pollution control devices. Trace elements The behaviour of trace elements in coal combustion systems can be complex (Sloss and Smith, 2000) and emissions do not necessarily correlate directly with fuel trace element content, especially for elements such as mercury. The majority of elements remain in the ash from the plant but more volatile species such as Se, As and Hg, can remain in the flue gas in trace quantities. The trace element contents of different biomass materials vary significantly. For example, waste tyres may result in reduced emissions of Hg, Cd and Tl but higher emissions of Cr, Mn, Fe and Zn. Sewage sludge can lead to an increase in Hg emissions whereas straw can actually be considered, in some situations, a means of lowering Hg emissions from coal combustion (Sloss, 2007). The presence of calcium and other alkali materials in the fuel mix can have an important effect in determining whether trace elements are trapped in the waste streams or released in the flue gas. In the UK, Drax power station has been testing various biomass materials to cofire with coal. Although some materials led to an increase in trace elements in the flue gas (such as increased Hg from milled pal nut), all materials considered resulted in emissions which were below any legislated limit.
Organic compounds Organic emissions arise from combustion as a result of inefficient and incomplete combustion. Since full-scale coal-fired plants are generally significantly more efficient than biomass combustion systems, the emissions of organic compounds are normally considered insignificant. The addition of coal to MSW combustion systems is actually an effective method of suppressing dioxin and furan formation (Davidson, 2005). De novo synthesis of dioxins and furans in the ESP of co-combustion units could theoretically be an issue in some situations but there do not seem to have been any reported cases of this occurring in the literature (Sloss, 2010). Fine particulates Biomass combustion alone is less efficient than coal combustion and leads to significantly greater particulate emissions. The physical characteristics of biomass fuels also lead to changes in combustion behaviour and affect the peformance of particulate control devices. For example, dried sewage sludge may contain higher concentrations of CaO which can neutralise the surface acidity of the particles and reduce ESP efficiency (Fernando, 2007). However, despite shifts in particle size ranges due to differences in nucleation rates and gas-to-particle conversion processes, the particulates from co-combustion can be handled effectively by existing particulate control systems. Fly ash In countries such as the Netherlands, coal fly ash is a valuable commodity. Any changes to fly ash characteristics as a result of cofiring biomass materials are therefore important as they may affect the usability of the ash and create a loss in income whilst creating a new waste problem. Biomass materials tend to have reactive ashes due to the higher quantities of alkali metals. Unburnt carbon may increase due to the low ash content contribution from the biomass. Since many trace elements remain in the ash, any increased contribution of trace elements from biomass materials could lead to an increase in the trace element content in the ash. For trace elements such as Cr, this can potentially pose a problem with respect to leaching. Petcokes can lead to slightly increased ash concentrations of Mo and Ni; sewage sludge can increase ash concentrations of Cd, Pb and Zn; and paper sludge can increase Cd, Hg and Zn concentrations (Lamers and others, 2000). However, there has been nothing published to suggest that any of these changes in ash are considered a problem with respect to the safety of ash use or disposal. From published studies, it would seem that any increase in the trace element content of the ash as a result of cofiring is not a barrier to cofiring as such. However, legislated fly ash standards in Europe (EN450) and the USA (ASTM618) specify that ash to be used in concrete must be derived directly from coal combustion. Therefore any plant choosing to start cofiring biomass will automatically lose revenue from ash sales. Extensive testing in the Netherlands has shown that this “blanket ban” on ash from cofiring may be too strict. Koppejan and Baxter (2005) showed that fly ash from cofiring systems did require increased aerating agent to produce concrete which would meet specified freeze-thaw conditions and that setting times could be longer but that, if the aerating agent was provided at an increased rate, the final material was as good, if not better, than concrete from coal fly ash alone. It is therefore suggested that the current legislation on the use of fly ash from cofiring is too stringent and could be an unnecessary barrier to the cofiring of biomass at plants that obtain revenue from the sale of fly ash. Legislation and incentives for biomass cofiring
The cofiring of biomass at full scale coal-fired plants is being encouraged in several countries in Europe as part of the move towards greener energy. However, there is no single EU policy on cofiring. Some individual member states are therefore making their own legislation and action plans, most of which are based on financial incentives such as feed-in-tariffs and premiums. Coal-fired plants in the EU must meet the current LCPD (Large Combustion Plant Directive) emission limits. Waste combustion plants must meet more stringent limits under the WID (Waste Incineration Directive). Therefore plant cofiring waste must meet new emission limits which are calculated as a combination of the LCPD and WID limits under what is termed the “mixing rule”. In many cases, the resulting limit is particularly stringent and is the deciding factor on whether a plan is prepared to consider co-combustion. In most cases, plants cofiring waste and biomass with coal at relatively low concentrations (such as below 20 or 30% by weight) the standards are easy to meet. However, this is not the case when firing higher concentrations of more challenging wastes such as sewage sludge or RDF/MSW. Variations in national legislation result in different attitudes to cofiring. For example, Scotland had to abandon cofiring of sewage sludge with coal due to the tightening of UK limits whereas Germany cofires sewage sludge at around 20 plants. The tighter legislation on landfill in Germany also makes co-combustion more economically sensible. As mentioned above, the existing legislation in the EU and USA disallows the use of coal ash from cofired systems which means that many plants will simply no consider cofiring as a viable option. However, this legislation is often regarded as too stringent and Denmark has already reduced the stringency to accept ash from co-combustion systems. The EU is currently reconsidering the requirements set out in EN450. Conclusions In the majority of situations, the co-combustion of biomass with coal in full-scale coal-fired systems leads to a reduction in emissions of particulates, SO2, NOx, trace elements and organic compounds. In most situations, biomass cofiring can actually be considered a valid technique to reduce NOx emissions from coal combustion. Co-combustion can also be regarded as a method of harnessing the energy present in the biomass whilst reducing emissions that would have been far more significant had the biomass material been burned alone. There are exceptions, such as sewage sludge and RDF/MSW materials, which can prove to be more of a challenge. Although biomass and waste materials are known to be highly variable in their chemistry, it would seem that the majority of these materials can be a valuable source of clean energy to replace coal in full scale combustion systems. Fuels such as these could be tested on a case-by-case basis to determine whether they can in fact be used as a carbon-neutral fuel without any detrimental effects on emissions and ash sales. References Davidson RM (2005) Chlorine in coal combustion and cofiring. CCC/105. London, UK, IEA Clean Coal Centre, 32 pp (Nov 2005) Di Nola G, de Jong W, Spliethoff H (2009) TG-FTIR characterisation of coal and biomass single fuels and blends under slow heating rate conditions: partitioning of the fuel-bound nitrogen. Preprint from: Fuel Processing Technology (2009) Fernando R (2005) Fuels for biomass cofiring. CCC/102. London, UK, IEA Clean Coal Centre, 37 pp (Oct 2005) Fernando R (2007) Cofiring of coal with waste fuels. CCC/126. London, UK, IEA Clean Coal Centre, 37 pp (Oct 2005)
Koppejan J, Baxter L (2005) Global operational status of cofiring biomass and waste with coal - experience with different cofiring concepts and fuels. In: IEA Bioenergy co-utilisation of biomass with fossil fuels - Exco55, Copenhagen, Denmark, 25 Mat 2005, 14 pp (2005) Leckner B (2007) Co-combustion - a summary of technology. Thermal Science: 11 (4); 5-40 (2007) Singh S, Nimmo W, Gibbs B M, Williams P T (2009) Waste tyre rubber as a secondary fuel for power plants. Fuel; 88; 2473-2480 (2009) Sloss LL (2007) Trace elements and fly ash utilisation. CCC/122. London, UK, IEA Clean Coal Centre, 54 pp (Mar 2007) Sloss LL (2010) Emissions from cofiring coal, biomass and sewage sludge. London, UK, IEA Clean Coal Centre. 57 pp (Aug 2010) IN DRAFT Sloss LL, Smith IM (2000) Trace element emissions. CCC/34. London, UK, IEA Clean Coal Centre, 29 pp (Jun 2000)
Element Leaching from Coal Stockpiles – Case Studies from the Sydney and Collie Basins, Australia Colin R. Ward1,2, David French2, Ken Riley2, Leanne Stephenson1, Owen Farrell2, Zhongsheng Li1 1 – School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney 2052, Australia 2 – CSIRO Energy Technology, PO Box 52, North Ryde 1670, Australia
ABSTRACT Laboratory testing has been carried out to evaluate the potential for release of elements from coal in exposed stockpiles, drawing on sample sets from the Sydney Basin in New South Wales and the Collie Basin in Western Australia. Testing in the Collie Basin was further extended to evaluate element leaching from the coastal sand deposits on which the stockpile was constructed, to ascertain whether the stockpile might represent a significant source of contaminants to the associated groundwater system. The mobility of the elements in the coal and sand was evaluated by batch leaching tests; samples of each material were shaken in sealed polyurethane bottles with water or a controlled pH (acid) solution for 18-24 hours, after which the concentrations of major and trace elements in the resulting leachates were determined. For the Collie Basin complementary tests were also carried out in which coal and coal underlain by sand were separately placed in Perspex columns and demineralised water or an acid solution was allowed to permeate through. The natural pH of the Collie Basin coal in water was shown by the batch tests to be 4.75 to 5.0, while that of the Sydney Basin coals, and also of the Collie Basin sand, was 8.0 to 9.3. Trace elements of environmental concern were mostly released at low concentrations from the coals of both basins. For the Sydney Basin some elements, such as Mn, Co, Ni and Zn, were released at slightly higher levels when the coals were tested with acid rather than water, suggesting a slight influence of pH on element mobility from the coal samples. Other elements, such as Mo and Se, were less mobile from the Sydney Basin coals under acid conditions than at natural (alkaline) pH levels. Element release from the Collie coals, which had a naturally acid pH when tested with water, showed a similar pattern to that displayed by the Sydney Basin coals under acid test conditions. Trace elements in the Collie Basin study, especially As, Cr, Cu, Mo, Pb, Sb, Th, U, V and Zn, were mostly released in higher concentration from the sand than from the coal. A few elements, such as Ba, Mn and Ni, were released in higher concentrations from the coal (although values were still low), while others, such as Cd and Co, were released in low but approximately equal concentrations from both coal and sand samples. Column tests on the Collie coal gave similar results to the batch leaching studies, although with slightly higher pH levels. The coal/sand column produced a leachate with a pH of 8.2 to 8.6, consistent with neutralisation of the acidic coal leachate by the carbonate minerals in the sand sample. Flow through the coal/sand column became blocked in the course of the test, apparently due to solution and re-precipitation of calcium carbonate in the sand by the acidic coal leachate. Higher concentrations of most trace elements, including As, B, Ba, Cr, Cu, Mo, Sb, Se, U and V, were released from this column before the blockage compared to the
1
columns with coal alone. Comparison to the batch test results suggests that these elements were mainly derived from leaching of the sand component. INTRODUCTION Effective environmental management of coal mining, preparation and utilisation facilities needs to take into account the nature of any potential leachates that might be derived from exposed stockpiles and refuse emplacements, including the possible interactions of those leachates with the underlying strata and groundwater systems. Many trace elements of environmental significance, such as As, Se, Pb and Zn, are typically associated with sulphide minerals such as pyrite (Finkelman, 1994; Ward et al., 1999). Community concerns associated with coal stockpiles are commonly based on experience with pyritic coals (e.g. Anderson et al., 1991, Pimentel and da Luz, 2009), where acidic mine waters may be generated from oxidation of the pyritic material in the coal, coal overburden or preparation refuse materials. Australian coals, especially those from most mines in the Sydney and Bowen Basins, tend to be low in pyrite (Ward and Swaine, 1995; Ward et al., 1999; Ruan and Ward, 2002), and stockpiles of such coals are therefore less likely to be associated with adverse environmental impacts. Community concerns nevertheless exist, in Australia and elsewhere, about the potential for release of toxic elements from exposed coal stockpiles. More specific data are therefore required to evaluate the extent to which such concerns may be justified, and to identify any specific cases where remedial measures may be required. A related issue is the potential for interaction between the leachates from coal stockpiles and the substrate or surrounding materials on which the stockpile has been built. Even though coal stockpiles are commonly underlain by impermeable barriers or have associated leachate collection systems, assessment of such interactions may be of value in assessing possible impacts if the barrier is breached, overflow occurs across the barrier, or material escapes from the leachate collection facility. While extensive studies have been reported on the leaching of elements from coal ash under different conditions (e.g. Jones, 1995; Hassett et al., 2005; Ward et al., 2009a), and also on the mobility of different elements from coal using a range of selective leaching techniques (e.g. Palmer et al., 1993; Finkelman, 1995), only limited research has been published on the nature of leachates from exposed coal stockpiles (e.g. Wang et al., 2008; Pimentel and da Luz, 2009), and even less on the interaction of stockpile leachates with associated substrate materials (Anderson et al., 1991), to provide a reference base for such environmental evaluations. This paper summarises the results of two separate studies of element leaching from coal stockpiles. One of these is based on a stockpile containing somewhat pyritic sub-bituminous coal from the Collie Basin of Western Australia (Ward et al., 2009b; 2010); the other is based on low-pyrite bituminous coals taken from stockpiles at four different mines in the Hunter Coalfield of the Sydney Basin, New South Wales (Riley, 2003). The Collie Basin study was extended to investigate the interaction of potential stockpile leachates with coastal sands making up the substrate material on which the stockpile is built, partly to compare the leaching behaviour of the coal to that of other materials in the area and partly to assess the possible impact of leachates from the coal on the underlying strata and groundwater system. SAMPLES AND TEST PROCEDURES A 20 kg sample of coal from the Collie Basin stockpile, air-dried and further crushed to <5 mm, was provided for the study by the plant operating company. The coal in the stockpile 2
has a nominal top size of around 40 mm; the further crushing was intended to provide material more suited to laboratory testing, increasing the available surface area and scaling down the permeability characteristics. A 10 kg air-dried sample of the underlying sand was also provided by the company, taken from a point near the stockpile site. Samples of 10 kg, crushed to < 10 mm, were taken from product stockpiles at several different mines in the Hunter Coalfield for the Sydney Basin study. Stockpiles of two different products were sampled at one site, to allow comparison of coals from the same mining operation with high and low ash yields. A program of laboratory tests was conducted to assess the levels of trace elements in leachates that might be derived from watering (for dust suppression) and extended rainfall on the stockpiles, including exposure of the coal to demineralised water and to any acid waters that might associated with regional precipitation. Evaluation of the Sydney Basin samples was based on batch leaching tests, while that for the Collie Basin stockpile included both batch tests and column leaching tests. As well as tests on the respective coals, the program for the Collie Basin study also involved testing of the coastal sand on which the stockpile had been built. Batch Leaching Tests The mobility of the elements from these samples was tested by a series of batch leaching tests, in which the samples were shaken in sealed polyurethane bottles with water or a specific pH solution for a fixed time period, after which the concentrations of major and trace elements in the resulting leachates were determined. The sand was tested in a dried but otherwise natural state; the coals were air-dried and then crushed to <10 mm for the Sydney Basin study and <5 mm for the Collie Basin study. Two different batch leaching protocols were used to evaluate the behaviour of the coals and, for the Collie Basin, the associated sand sample: Shake (Water Washing) Test: Based on ASTM Method D-3987 (ASTM, 1995), this was intended to assess of the mobility of the inorganic constituents in the samples at their natural pH levels. The samples were agitated in de-ionised water for 18-24 hours, using solid to liquid (s:l) ratios of 1:20 for the Sydney Basin samples and 1:20 and 1:3.5 for the Collie Basin materials. Simulated Precipitation Leaching Protocol (SPLP) Test: Based on US EPA Method 1312 (US-EPA, 1992), this involved treatment of the coal and sand samples with a solution made up to simulate acid rain water at pH 3.0 (Sydney Basin) to 4.2 (Collie Basin). Extraction was carried out with mechanical agitation for 18-24 hours at a solid to liquid ratio of 1:20. The leachates were separated from the respective solid phases at the conclusion of each test by centrifugation, followed by filtration through 0.22 m Millipore cellulose acetate membrane filters. Approximately 30 ml of each sample was immediately acidified using approximately 0.5 ml of analytical grade concentrated nitric acid, and the acidified leachates were analysed by ICP-AES and ICP-MS techniques. The remaining portions of the nonacidified leachates were used for pH measurement.
3
Column Leaching Tests Laboratory column leaching tests were carried out on the Collie Basin coal sample using two columns 100 mm in diameter (to minimise channelling of liquids down the wall) and approximately 1 metre long (Ward et al., 2010). One column (Column 1) used demineralised water and the other (Column 2) used a synthetic acid rain water (pH = 4.2) as the leaching solution. A third column (Column 3) was also constructed, with a layer around 300 mm thick of the sand sample under a 700 mm coal column, in which the synthetic rain water (pH 4.2) was also used as the leaching solution. A filtration membrane was placed at the bottom of each column to prevent passage of any suspended solid particles. Each test was run to provide leaching under unsaturated conditions for up to 7 days, with sampling of the leachate in each case after approximately two (first flush), 24, 48, 72, 96 and 192 hours. The columns were then left exposed to the air for approximately two weeks, to allow any changes (e.g. oxidation of pyrite) to develop, after which a second series of tests was carried out on the exposed coal or coal-sand columns. The second series was intended to investigate the effects of exposure under alternating wet and dry conditions. The leachates from the column tests were separated through 0.22 m cellulose acetate filters, and analysed in the same way as those from the batch leaching tests. RESULTS FROM BATCH LEACHING TESTS The pH and concentrations of selected elements in the leachates from the batch tests are given in Tables 1 and 2. Table 1: Element concentrations in leachates from batch tests with demineralised water Sydney Basin
pH As B Ba Be Cd Co Cr Cu Mn Mo Na Ni Pb Sb Se Th U V Zn
g/L g/L g/L g/L g/L g/L g/L g/L g/L g/L mg/L g/L g/L g/L g/L g/L g/L g/L g/L
Collie Basin
Coal A 8% ash
Coal A 24% ash
Coal B 8% ash
Coal C 12% ash
Coal D 12% ash
8.14 0.1 <10
8.59 <0.1 <10
9.18 0.2 <10
9.30 0.4 <10
9.30 1.1 <10
<0.01 0.02 0.02 0.14 0.1 6.7 1.5 4.3 0.11 <0.1 0.04 0.27 0.03 0.01 0.09 <1
<0.01 0.02 <0.01 0.1 0.14 0.28 4 6.3 0.07 0.1 0.05 0.35 0.14 0.05 0.21 <1
<0.01 0.01 0.05 0.3 0.13 0.57 1.7 5.6 0.18 <0.1 0.23 0.58 <0.01 0.01 0.43 <1
<0.01 0.02 0.06 0.26 0.28 11 2.1 6.2 0.19 <0.1 0.14 0.57 0.08 0.03 1.0 <1
<0.01 0.02 0.14 0.36 0.23 0.34 2 6.9 0.39 <0.1 0.15 0.91 0.01 0.05 1.3 <1
Coal (7% ash) 1:3.5 1:20
Sand 1:3.5 1:20
4.76 0.25 20 40 0.08 0.34 2.1 0.38 <0.01 549 0.01 24.8 8.26 0.38 0.01 0.21 <0.01 0.01 0.04 4.73
8.1 2.99 50 10.6 0.02 0.63 0.34 6.69 3.62 28.6 1.79 10.2 3.35 2.68 0.36 0.99 0.5 0.21 6.4 24.6
5.07 0.2 <10 88 <0.01 0.66 0.44 0.07 <0.01 102 0.01 3.47 2.1 <0.1 0.01 0.07 <0.01 <0.01 <0.07 <1
8.33 1.26 <10 5.03 <0.01 0.37 0.12 3.03 1.87 10.7 0.46 <0.06 0.88 1.12 0.11 0.05 0.24 0.11 3.73 8.03
The natural pH of the Sydney Basin coal samples in water was shown by the batch tests to be mildly alkaline (8.1 to 9.3). This is consistent with a predominance of clay and in some cases carbonate minerals, as are typically reported in coals from this region (e.g. Ward and Swaine, 1995; Ward et al., 1999; Ruan and Ward, 2002). Lower pH values (4.4 to 8.3) were obtained 4
from the tests with synthetic acid rainwater, especially for the low-ash coals from sites A and B. As well as the relatively neutral behaviour of the minerals present (low buffering capacity), this may reflect the low overall proportion of mineral matter in these particular coal samples which are export coals that had been processed to obtain low ash yields. Table 2: Element concentrations in leachates from batch tests with acid rain water Sydney Basin
pH As B Ba Be Cd Co Cr Cu Mn Mo Na Ni Pb Sb Se Th U V Zn
g/L g/L g/L g/L g/L g/L g/L g/L g/L g/L mg/L g/L g/L g/L g/L g/L g/L g/L g/L
Collie Basin
Coal A 8% ash
Coal A 24% ash
Coal B 8% ash
Coal C 12% ash
Coal D 12% ash
Coal 1:20
Sand 1:20
5.10 0.1 <10
6.62 <0.1 <10
4.41 0.2 <10
7.43 0.2 <10
8.28 0.7 <10
0.14 0.04 0.73 0.03 0.3 78 <0.01 4.2 1.7 0.2 0.02 0.18 <0.01 0.01 0.02 4
0.05 0.04 0.27 0.05 0.21 33 0.19 6.5 0.56 <0.1 0.02 0.13 <0.01 0.02 0.01 4
0.38 0.08 8.2 0.38 5.3 110 0.07 5.9 8.5 0.8 0.36 0.31 <0.01 0.02 0.19 14
0.02 0.02 2 0.02 0.36 95 1.2 5.8 1.5 <0.1 0.17 0.32 0.01 0.01 0.87 <1
0.01 0.01 2.3 0.04 0.2 53 1.4 6.8 2.2 <0.1 0.12 0.64 <0.01 0.15 0.73 <1
5.04 0.17 <10 85 0.04 <0.01 0.51 <0.17 <0.01 118 <0.04 6.31 1.15 <0.1 <0.001 0.05 <0.01 <0.01 <0.07 <1
8.49 1.24 40 4.71 0.03 0.09 0.12 <0.17 <0.01 9.71 0.49 3.29 0.16 0.3 0.1 0.04 0.22 0.08 <0.07 <1
The pH of the Collie Basin coal at the two different s:l ratios tested in water was 4.7 and 5.0, while that of the associated coastal sand tested under the same conditions was 8.1 and 8.3. This particular coal has a 7% ash yield (air dried) and around 10% mineral matter, with around 10% of the mineral matter being represented by pyrite (Ward et al., 2009b; 2010), and hence pyrite oxidation is probably responsible for generation of the acid pH levels. The sand, by contrast, consists mainly (around 80%) of carbonate minerals (calcite and aragonite), derived from an abundance of shell fragments; this explains the alkaline nature of the leachate derived from that material. Leaching with synthetic acid rainwater does not significantly alter these pH values (Table 2), due to the abundance of the pyrite and carbonate minerals in the respective samples. Low concentrations of most trace elements were released from the Sydney Basin coals in the water-based batch leaching tests (Table 1). The only element of environmental significance with a concentration above 1 g/L was Mo (1.5-4.0 g/L). Acid leaching of the same coals (Table 2) produced higher concentrations of Ni (up to 8.5 g/L, Mn (up to 100 g/L) and Zn (up to 14 g/L) than the water-based tests, but lower concentrations of Mo (<1.5 g/L). Plots showing the pH and concentration of selected elements for each coal leached with water and acid are given in Figure 1. The higher concentration of Mo in the leachates from the Sydney Basin coals when treated with water (Table 1), compared to the same element in the leachates after acid treatment (Table 2; Figure 1), probably reflects the formation in solution of an oxy-anion (cf. Eary et al., 1990; Ward et al., 2009a). In contrast to metals in solution as cations, which are more soluble 5
in acid conditions, such oxy-anions tend to be more mobile at neutral to alkaline pH levels. The higher concentrations of elements such as Co, Ni and Zn from the acid-treated Sydney Basin coals (Figure 1), by contrast, are consistent with behaviour as cations in the mobilisation process. Testing of the Collie coal with water at similar solid:liquid ratios (1:20) released slightly higher concentrations of Cd, Co, Mn, Ni and Zn than were obtained from the Sydney Basin coals, but much lower Mo concentrations. Testing of the Collie coal with synthetic acid rainwater (pH = 4.2), however (Table 2), produced leachates with very similar characteristics to those obtained from the Sydney Basin coals under similar (acid-leaching) conditions.
D‐Acid
C‐Acid
B‐Acid
A2‐Acid
A1‐Acid
100 80 60 40 20
A1‐Acid
A2‐Acid
B‐Acid
C‐Acid
D‐Acid
A2‐Acid
B‐Acid
C‐Acid
D‐Acid
Nickel
10
Concentration (ppb)
A1‐Acid
A1‐Water
Molybdenum
B‐Water
0
D ‐ Acid
C ‐ Acid
B ‐ Acid
A2 ‐ Acid
A1 ‐ Acid
D ‐ Water
C ‐ Water
0.0
B ‐ Water
Manganese
A2‐Water
0.1
D‐Water
A1‐Water
Concentration (ppb)
0.2
A2 ‐ Water
0
D‐Acid
C‐Acid
B‐Acid
A2‐Acid
A1‐Acid
D‐Water
C‐Water
B‐Water
0.3
4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0
2
120
0.4
8 6 4 2
D‐Water
C‐Water
B‐Water
A2‐Water
A1‐Water
D‐Acid
C‐Acid
B‐Acid
A2‐Acid
A1‐Acid
D‐Water
C‐Water
B‐Water
A2‐Water
0
A1‐Water
Concentration (ppb)
A2‐Water
Chromium
0.5
A1 ‐ Water
Concentration (ppb)
A1‐Water
2
4
D‐Water
4
6
C‐Water
6
C‐Water
8
8
B‐Water
Concentration (ppb)
Leachate pH
10
Cobalt
10
A2‐Water
Leachate pH
12
Figure 1: Plots showing pH and concentration of selected elements in leachates from Sydney Basin coal samples from water (left) and acid (right) testing.
Most of the trace elements in the Collie Basin study were released in higher concentration from the coastal sand than from the coal, especially As, B, Cr, Cu, Mo, Pb, Sb, V and Zn. A few, such as Ba, Mn and possibly Ni, were released in greater concentrations from the coal (although values were still low), while others, such as Cd and Co, were released in low but approximately equal concentrations from both the Collie coal and the associated sand sample.
6
RESULTS FROM COLUMN LEACHING TESTS Data on the pH and selected element concentrations in the first flush from each of the column tests, before and after standing, are given in Table 3. The column tests on the coal (Columns 1 and 2) gave similar results to the batch leaching studies (Tables 1 and 2), although with slightly higher pH levels. The coal/sand column (Column 3) produced a leachate with a pH of 8.2 to 8.6, consistent with neutralisation of the acidic coal leachate by the carbonate minerals in the sand sample and with the alkaline leachate obtained from the sand in the batch tests.
Sand SPLP
Sand 1:20
Coal SPLP
Sand 1:3.5
Sand SPLP Sand SPLP
Sand SPLP
0.0
Coal SPLP
0.0
Sand 1:20
0.5
Coal 1:20
0.5
Sand 1:20
1.0
Sand 1:20
1.0
Sand 1:3.5
Concentration (ppb)
1.5
Sand 1:3.5
1.5
Sand 1:3.5
Molybdenum
Coal SPLP
2.0
Sand SPLP
Sand 1:20
Sand 1:3.5
Coal SPLP
Coal 1:20
0
Cobalt
2.5
Nickel
10 8 6 4 2 0
Coal SPLP
2
Coal 1:20
0
Coal 1:3.5
Concentration (ppb)
4
2.0
20
Sand SPLP
Sand 1:20
Sand 1:3.5
Coal SPLP
6
Coal 1:3.5
Concentration (ppb)
Coal 1:20
Chromium
8
Coal 1:3.5
Concentration (ppb)
Coal 1:3.5
0
40
Coal 1:20
1
60
Coal 1:3.5
2
80
Coal 1:3.5
Concentration (ppb)
Concentration (ppb)
3
Barium
100
Coal 1:20
Arsenic
4
Figure 2: Concentration of selected elements in leachates from Collie coal (left) and sand samples (right) tested in water at s:l = 1:3.5 and 1:20 and in acid (SPLP) at s:l = 1:20.
Flow ceased in the coal/sand column between observations at 24 and 48 hours, apparently due to build-up of solids near the coal-sand contact. Pumping of liquid into this column was therefore stopped. It was restarted 24 hrs later, but stopped again after 2¼ hrs as the column was almost completely clogged. The column was left without further throughput until a spontaneous breakthrough occurred some 21 days later, when a lot of fluid held up in the column came through. Pumping into the column was resumed; initial output was only about 1 drop every 4 minutes, but this increased to normal flow over a six hour period. X-ray diffraction analysis of samples taken from the sand layer in Column 3, after the leaching tests were completed, showed an increase in the proportion of carbonate minerals (mainly magnesian calcite) in the middle and bottom of the sand bed, with a corresponding 7
reduction in the other phases such as quartz and microcline. This suggests that carbonate was partly dissolved from the top of the sand layer during the leaching process and redeposited towards the base of the column to cause the blockage that developed. Higher concentrations of most elements, including As, B, Ba, Cr, Cu, Mo, Sb, Se and V, were released from the coal/sand column before the blockage developed, compared to the columns tested with coal alone. Comparison to the batch test results suggests that the elements released from this column were mainly derived from leaching of the sand component. A few elements, notably Mn and Zn, were released in lower concentrations from the coal/sand column, suggesting possibly adsorption or precipitation when the coal leachate came into contact with the sand layer. Table 3: Composition of leachates from first flush of column tests Column 1 Coal Demineralised water pH As B Ba Be Cd Co Cr Cu Mn Mo Na Ni Pb Sb Se Th U V Zn
µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L mg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L
Initial Test 4.39 0.48 64.8 24.1 0.56 0.042 9.8 0.12 2.3 3091 0.23 185 7.7 0.09 0.007 2 0.006 <0.003 0.57 523
After standing 4.84 0.15 55.6 30.2 0.29 0.02 5 0.14 1.4 1963 0.2 74 4.8 0.06 <0.003 0.3 0.005 <0.003 0.1 1262
Column 2 Coal Synthetic rainwater Initial test 4.54 0.45 65.8 21.1 0.32 0.027 6.6 0.08 1.8 2411 0.21 163 5.3 0.06 0.006 0.9 0.005 <0.003 0.52 755
After standing 5.08 0.15 50.2 35.1 0.27 0.019 4.3 0.13 1.4 1753 0.21 78 3.9 0.07 0.004 0.5 0.005 <0.003 0.1 2047
Column 3 Coal + sand Synthetic rainwater Initial test 8.32 4.5 196.8 86.3 0.04 0.066 0.9 0.81 4.8 3.37 9.8 344 14 0.12 1.1 3.2 0.031 2.0 5.6 311
After standing 8.07 8.2 58.8 38.1 <0.01 0.014 0.61 1.06 0.66 160.7 5.1 33 5.2 0.06 6.4 0.9 0.02 1.1 1.4 40.8
COMPARISON TO WATER QUALITY GUIDELINES Of the few elements released at significant concentrations from the coals in the batch and column tests, only Zn and possibly Co appear in some cases to exceed the current ANZECC/ARMCASNZ Water Quality Guidelines (ANZECC/ARMCANZ, 2000). However, if the common dilution attenuation factor of 100 is applied, the concentrations expected under field conditions would still be below the “trigger values” set by those Guidelines for marine environments, especially in cases, such as in the Collie Basin study, where the coal leachate is attenuated by contact with a carbonate-bearing sand beneath.
8
ACKNOWLEDGEMENTS Thanks are expressed to Verve Energy, and especially to Peter Christian, Environmental Manager, for sponsorship of the Collie Basin study, as well as on-site assistance and permission to publish this paper. Funding for the Sydney Basin study was provided by the Australian Coal Association Research Program (ACARP). Thanks are expressed to Dorothy Yu and Irene Wainwright, of the UNSW Analytical Centre, and to AMDEL Pty Limited, for provision of the chemical analysis data on the Collie Basin samples. REFERENCES Anderson, M.A., Bertsch, P.M., Feldman, S.B., Zelazny, L.W., 1991. Interactions of acidic metal-rich coal pile runoff with a subsoil. Environmental Science and Technology 25(12), 2038-2048. ANZECC/ARMCANZ, 2000. Water Quality Guidelines. http://www.ea.gov.au/water/quality/nwqms/ ASTM, 1995. Standard test method for shake extraction of solid waste with water – ASTM D 3987. Annual Book of ASTM Standards 11.04, 14-17. Eary, L.E., Rai, D., Mattigod, S.V., Ainsworth, C.C., 1990. Geochemical factors controlling the mobilization of inorganic constituents from fossil fuel combustion residues: II. Review of minor elements. Journal of Environmental Quality 19, 202-214. Finkelman, R.B., 1994. Modes of occurrence of potentially hazardous elements in coal: levels of confidence. Fuel Processing Technology 39, 21-34. Finkelman, R.B., 1995. Modes of occurrence of environmentally-sensitive trace elements in coal. In: Swaine, D.J., Goodarzi, F. (editors), Environmental Aspects of Trace Elements in Coal. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 24-50. Hassett, D.J., Pflughoeft-Hassett, D.F., Heebink, L.V., 2005. Leaching of CCBs: Observations from over 25 years of research. Fuel 84, 1378–1383. Jones, D.J., 1995. The leaching of major and trace elements from coal ash. In: Swaine D.J., Goodarzi, F. (eds), Environmental Aspects of Trace Elements in Coal, Kluwer, Dordrecht, The Netherlands, pp 221–262. Palmer, C.A., Krasnow, M.R., Finkelman, R.B., d’Angelo, W.M., 1993. An evaluation of leaching to determine modes of occurrence of trace elements in coal. In: Chiang, S.H. (editor), Proceedings of 10th International Pittsburgh Coal Conference, School of Engineering, University of Pittsburgh, pp. 1062-1067. Pimentel, D.A., da Luz, J.A.M., 2009. Pluvial percolation in pyrite-bearing coal stockpiles. Revista Escola de Minas 62(1), 99-105. Riley, K.W., 2003. Leachability of Trace Elements from Coal Stockpiles. Report on ACARP Project C11021, Australian Coal Association Research Program, Brisbane, 18pp.
9
Ruan, C.-D., Ward, C.R., 2002. Quantitative X-ray powder diffraction analysis of clay minerals in Australian coals using Rietveld methods. Applied Clay Science 21, 227-240. US-EPA 1992, Synthetic Precipitation Leaching Procedure. US Environmental Protection Agency Method 1312. http://www.epa.gov/epaoswer/hazwaste/test1_series.htm Wang, W., Qin, Y., Song, D., Wang, K., 2008. Column leaching of coal and its combustion residues, Shiuishan, China. International Journal of Coal Geology 75, 81-87. Ward, C.R., French, D., Jankowski, J., Dubikova, M., Li, Z., Riley, K., 2009a. Element mobility from fresh and long-stored acidic fly ashes associated with an Australian power station. International Journal of Coal Geology 80, 224-236. Ward, C.R., French, D., Riley, K., Stephenson, L., Farrell, O., Li, Z., 2009b. Element leachability from a coal stockpile and associated coastal sand deposits. Proceedings of International Conference on Coal Science and Technology, Cape Town, South Africa, 26-29 September, 2009, Paper 12-2, 9 pp. (CD publication). Ward, C.R., French, D., Riley, K., Stephenson, L., Farrell, O., Li, Z., 2010. Element leachability from a coal stockpile and associated coastal sand deposits. Fuel Processing Technology (in press). Ward, C.R., Spears, D.A., Booth, C.A., Staton, I., Gurba. L.W., 1999. Mineral matter and trace elements in coals of the Gunnedah Basin, New South Wales, Australia. International Journal of Coal Geology 40, 281-308. Ward, C.R., Swaine, D.J., 1995. Minerals and inorganic constituents. In “Geology of Australian Coal Basins” (ed. C.R. Ward, H.J. Harrington, C.W. Mallett and J.W. Beeston), Geological Society of Australia Coal Geology Group, Special Publication 1, 93-109.
10
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission COAL RESOURCE ESTIMATION IN ISIKLAR-KISRAKDERE (SOMA, MANISA, TURKEY) A. Erhan Tercan, Prof.Dr. Hacettepe University, Mining Engineering Department, Ankara, Turkey [email protected], tel: 0312.297 76 00, fax: 0312.299 21 55 Bahtiyar Ünver, Prof.Dr., [email protected], tel: 0312.297 76 00, fax: 0312.299 21 55 Mehmet Ali Hindistan, Assist.Prof.Dr. [email protected], tel: 0312.297 76 00, fax: 0312.299 21 55 Perihan Çorbacı, Technical Expert, General Directorate of Turkish Coal Enterprises, Ankara, Turkey [email protected], tel: 0312 384 17 20, fax: 0312 384 62 34 Kıvanç Het, M.Sc. Mining Engineer, General Directorate of Turkish Coal Enterprises, Ankara, Turkey [email protected], tel: 0312 384 17 20, fax: 0312 384 62 34
Abstract: Işıklar-Kısrakdere is a site that is located in Soma Neocene region and contains lignite coal suitable for underground mining extraction. This paper presents a case study in which coal resource estimation is addressed. For this purpose a data base is constructed using drill hole information such as collar, survey, sample and geology tables. Then solids for coal seams are produced by considering faults and geological layers. The solid models are divided into equal size and shaped blocks. The mean calorific values of these blocks are estimated by geostatistical techniques. Finally quality-tonnage curves are produced.
Manuscript Not AVAILABLE
Coal Explorations in Turkey: New Projects and New Reserves Dr. İlker ŞENGÜLER MTA Genel Müdürlüğü Enerji Dairesi 06800 Ankara [email protected]
Of the primary energy sources, coal, which is regarded as a solid fossil fuel, is very crucial with respect to other primary energy sources since it has an undoubted longer time of reserve and a wider distribution than the other sources. World recoverable coal reserves are 915 billion tonnes, and our country with own lignite reserves is among 10 leading countries in the world. Tertiary deposits cover a total area of 110.000 km2, and they are distinguished as about 2% Eocene, 14% Oligocene, 52% Miocene and 32% Pliocene. Total prospecting area performed by the General Directorate of Mineral Research and Exploration is 205.000 km². 42.000 km² areas of Tertiary, which could be prospective for coalfields, are studied with detail and the coalification is determined in an area of about 1500 km². Coal exploration studies were about to cease at the beginnings of 1990’s years due to the fact that our country exploited more increasingly natural gas. These studies augmented again by the projects, which commenced and co-ordinated by the General Directorate of Mineral Research and Exploration in 2005. The priority in these studies is given to reinvestigate the areas suitable for coal deposition in our country, and so to determine new prospective fields. In this context, new coal deposits are discovered in Thrace, Soma (Manisa), Karapinar (Konya), Dinar (Afyonkarahisar), Alpu (Eskisehir) and Afsin-Elbistan (Kahramanmaras) basins, there are also reserve increases in the known coal fields. Our lignite reserves, known as 8.3 billion tonnes for a long time, reached to 12.6 billion tonnes due to the exploration and research studies, commenced in 2005 and conducted by General Directorate of Mineral Research and Exploration. However, taking into consideration the exploited amount in the lignite fields is 1.1 billion tonnes so far, it is concluded that lignite reserves in our country is 11.5 billion tonnes. After 2005 an increasing amount of 4.3 billion tonnes of proven+probable+possible reserves is a very crucial source of energy because it is domestic. The utilization for fuels of coal-powered plants by our coals, which often have been ranked as low-level lignites, is inevitable for a sustainable energy and therefore a sustainable development. With newly discovered oil and coal fields, the General Directorate of Mineral Research and Exploration pioneered the establishment of TPAO and TKİ, both of which are the most important institutions, and has been constituting a leadership in energy. Today, our institution again took this mission via proposed and conducted projects.
Activated Carbon from Brown Coal by Chemical Activation
Luguang Chen* and Sankar Bhattacharya Department of Chemical Engineering, Monash University GPO Box 36, Clayton, Victoria 3800, Australia
*: Corresponding author, Tel:03-99054915, Fax:03-99055686, Email: [email protected]
Abstract In Victoria, brown coal is a cheap resource, which at current consumption rate has over 500 years of reserves. A possible utilization of these materials is converting them into activated carbon, which has potential for use as adsorbent and catalysts. Activated carbons can be produced by two methods - physical activation and chemical activation. Chemical activation is preferred over physical activation owing to the lower temperatures and shorter time needed for activation. The activation property of activated carbon strongly depends on its surface area. In this study, we prepared activated carbons from Loy Yang brown coal using chemical activation. The materials were dipped into the solution for 24 h at room temperature, thoroughly washed by distilled water in order to remove the exact chemicals, and then filtered. Then the impregnated materials were pyrolyzed in a furnace at temperatures of 400ºC-600ºC for 30 mins under flowing N2 at 20 ml/min. The type of activating reagent (KOH and K2CO3) and the effect of carbonization temperature on the texture of activated carbons have been investigated.
1
The morphologies of activated carbons were observed by SEM. The surface areas of activated carbons were measured by volumetric adsorption analyzer, using N2 as the adsorption gas at -196ºC. This paper summarises the different chemical activation reagent and the carbonization temperature effects on activated carbons’ porous structures and identifies the optimum preparation conditions for activated carbon from Loy Yang coal. Keywords: Activated carbon; Brown coal; Chemical activation; Surface characterisation 1. Introduction Activated carbon is predominantly an amorphous solid with a large internal surface area and pore volume[1]. Activated carbon is one of important adsorbents for purification of gaseous and liquid phase mixtures[2]. There are two basic ways to prepare activated carbons[3, 4], physical and chemical activation. Physical activation involves carbonization under an inert atmosphere, or a partial and controlled gasification at high temperature with steam[5], CO2 [6, 7], air[8, 9], or a mixture of these[10]. The chemical activation is to prepare activated carbons using chemical agents which are normally alkali and alkaline earth metal containing substances such as KOH, K2CO3, NaOH, ZnCl2, and some acid such as H3PO4 and H2SO4[2, 11-13]. Chemical activation is the process that normally takes place at a lower temperature and a shorter time than that used in physical activation. Moreover, the carbon yields of chemical activation are higher than physical one. The challenge in the preparations of activated carbons is to produce the carbon with high specific surface area and pore volumes from low cost materials at low temperature. Victoria has extensive brown coal resources, estimated to be over 200 thousand million tones, of which some 33 thousand million tones are regarded as readily recoverable reserves using present day open-cut mining technology[14]. The vast supplies low cost of Victorian brown coals make them particularly attractive feedstock for production of activated carbons. Victorian born coals are also very low in ash content, typically 2% on a dry basis. 2
The aims of this research are to investigate the influence of (1) activation agents, KOH or K2CO3 (2) the impregnation ratio (3) carbonization temperature on the pore structure of the prepared activated carbons from Victorian Loy Yang brown coal. 2. Experimental 2.1 Chemical activation Loy Yang brown coal was selected as starting material for preparation of the activated carbons. The properties of Loy Yang raw coal are listed in Table 1. The oven dried material was ground and sieved, and a particle size no less than 600 μm was used in the experiments. Chemical activation of the coal was performed using KOH and K2CO3. Different impregnation ratios and carbonization temperatures were studied in order to establish the optimal conditions for producing high surface area activated carbons from the coal. For the chemical activation, 10 g Loy Yang coal was mixed with KOH or K2CO3 solution with different impregnation ratio for about 3 h at 50ºC. The concentration of the solution was adjusted to give a mass ratio of chemical agent to the material between 1:2 and 5:2 (mass basis). The homogeneous samples were dried at 115 ºC for over 24 h, followed by carbonization in a muffle furnace at desired temperatures (400ºC, 500ºC and 600ºC) for 30 min before cooling down. The carbon products were washed with distilled water until the pH of solution reached 7. After that, the activated carbons were dried at 115ºC for over 24 h. The activated carbons produced under different conditions are listed in Table 2.
3
Table 1 Proximate and ultimate analysis of Loy Yang raw coals
Loy Yang Coal 2.6 3.7
Moisture % (oven dried basis) Ash Ultimate analysis wt % Carbon Hydrogen Oxygen Nitrogen Sulphur Chlorine
65.0 4.6 25.5 0.72 0.5 0.11 Ash analysis (% db)
SiO2 Al2O3 Fe2O3 TiO K2O MgO Na2O CaO SO3 P2O5
56.5 20.5 4.6 1.5 1.3 3.6 4.7 1.6 5.0 0.2
Table 2 Conditions for the preparation of the activated carbons.
Type of activated Starting carbons material Loy Yang LH1-400,500,600 coal Loy Yang LH3-400,500,600 coal Loy Yang LH5-400,500,600 coal Loy Yang LC1-400,500,600 coal Loy Yang LC3-400,500,600 coal Loy Yang LC5-400,500,600 coal
Activation Agent
Impregnation ratio
Carbonization temperature (ºC)
KOH
1:2
400, 500, 600
KOH
3:2
400, 500, 600
KOH
5:2
400, 500, 600
K2CO3
1:2
400, 500, 600
K2CO3
3:2
400, 500, 600
K2CO3
5:2
400, 500, 600
2.2 Pore structure characterization The pore structures of the activated carbons were determined by N2 adsorption at -196ºC using Micromeritics ASAP-2020 volumetric sorption analyser. The activated carbons 4
were degassed at 250ºC in a vacuum condition for over 24 h. The volume of micropores was deduced from the application of the Dubinin-Radushkevich (DR) equation[15]. Additionally, secondary electron images of the activated carbons were recorded by using JEOL 840A scanning electron microscopy. 3. Results and discussion 3.1 Yield of activated carbon The yield of activated carbon is estimated as follows
Yield AC ( wt.%) [
W AC ] 100% , Wcoal
(1)
where WAC and Wcoal are the weights of the activated carbon product and the raw coal, respectively. Fig. 1 shows the effects of carbonization temperature on the yield of activated carbon with different impregnation ratios. As for the activation with KOH or K2CO3, the carbon yield was found to be a decreasing function of the carbonization temperature and the impregnation ratio. This can be attributed to the fact that volatile evolution, carbon oxidation to form carbonations, and carbon gasification are enhanced by the increase in temperature, leading to the increase in weight loss. The untreated raw Loy Yang coal showed the highest carbon yield, 68% at a carbonization temperature of 400ºC and the KOH treated Loy Yang coal showed the lowest carbon yield, 21% at a carbonization temperature of 600ºC with the impregnation ratio of 5:2. The carbon yield of the activated carbons by the chemical treatment is lower than that of the untreated coal. The reason can be addressed that alkaline surface salt complexes can be formed through the interactions between potassium composite and carbonaceous materials at low temperatures[16] and these complexes are the active sites for gasification. Thus, the carbon gasification by KOH and K2CO3 over these sites contributed to the loss in carbon yield at high temperatures and the high porous structure.
5
50 45
Yield /%
40 35 30 25 20 15 300
LH1 LH3 LH5 LC1 LC3 LC5 350
400
450 500 550 Temperature /ºC
600
650
700
Figure 1 Effect of carbonization temperature on the yield of activated carbon from Loy Yang coal.
The experimental results indicate that carbonization temperature plays an important role on the yield of activated carbon. Moreover, the impregnation ratio affects significantly the activated carbon yield. K2CO3 activation produced a higher carbon yield compared to KOH activation. 3.2 Effect of KOH activation on pore structure The effects of impregnation ratio and carbonization temperature on surface area and pore volume of activated carbons activated by KOH are listed in Table 3. The last column in Table 3 indicates the variation in results among the multiple repeat experiments. It can be seen that KOH enhanced the developments of porous structures of the activated carbon by gasification through reaction shown in Equation (2).
6KOH C 2K 3H 2 2K 2 CO3
(2)
6
The presence of metallic potassium will attach to the carbon matrix. The phenomenon results in widening of the spaces between carbon atomic layers and increase of the pore volume[17]. The BET surface area increased with the increase of carbonization temperatures and impregnation ratios; furthermore, the carbonization temperature played more important role to increase the porous structure. When carbonization temperature increased from 400ºC to 600ºC, the micro pore volume increased significantly from 0.144 to 0.450 cm3/g and micro surface area rose from 259.37 to 395.62 m2/g, impregnated in the same concentration of KOH. The surface areas of the activated carbons produced from the experiment are located in the range of 500-2000 m2/g, which is the typical surface area range of commercial activated carbons[18]. The maximum surface area was obtained from the activated carbon produced at the carbonization temperature of 600ºC with the impregnation ratio of 5:2, which is 712.00 m2/g. Table 3 Pore structure characterization of activated carbons activated by KOH with impregnation ratio from 1:2 to 5:2 at carbonization temperature of 400 ºC - 600 ºC.
Types of Activated Carbons LH1-600 LH3-600 LH5-600 LH5-400 LH5-500
Surface Area (m2/g) BET Surface Area 649.56 668.62 712.00 514.17 543.23
DubininRadushkevich Micropore 330.56 342.95 395.62 259.37 324.78
Pore Volume (cm3/g)
Variation*
Total Pore Volume
Micro Pore Volume
(%)
0.438 0.470 0.552 0.352 0.402
0.324 0.401 0.450 0.144 0.227
0.2 1.0 0.5 0.4 0.2
* variation is defined as the maximum value of variation = ABS(Experimental value-Average value )/Average value ×100%.
The pore size distribution of the activated carbon produced by KOH activation from Loy Yang coal at variable impregnation ratio at carbonization temperature of 600ºC is shown in Fig. 2. Higher impregnation ratio enhanced the pore structures. When the impregnation ratio increased from 1:2 to 5:2, the more surface metal complexes forming on the surface are responsible for the increase of the pore development, and creation new pores.
7
0.07
Pore Size Distribution (cm³/g·Å)
0.06 LH1-600 LH3-600 LH5-600
0.05 0.04 0.03 0.02 0.01 0 0
2
4
6
8 10 Pore Width (Å)
12
14
16
18
Figure 2 Pore size distribution of the activated carbon by KOH with variable impregnation ratios at carbonization temperature of 600 ºC.
The effect of carbonization temperature on the pore size distribution of the activated carbon by KOH with impregnation ratio of 5:2 is shown in Fig. 3. These figures confirm that the activated carbons produced at carbonization temperatures 400-600ºC contained heterogeneity of porous structure while 600ºC enhanced the porous structure.
8
Pore Size Distribution (cm³/g·Å)
0.07 0.06 LH5-400 LH5-500 LH5-600
0.05 0.04 0.03 0.02 0.01 0 0
2
4
6
8
10
12
14
16
18
Pore Width (Å)
Figure 3 Pore size distribution of activated carbons by KOH with impregnation ratio of 5:2 at variable carbonized temperatures.
3.3 The effect of K2CO3 activation on pore structure The pore structure characterization of activated carbons by K2CO3 activation from Loy Yang brown coal with impregnation ratio from 1:2 to 5:2 at carbonization temperature from 400 to 600 ºC is summarized in Table 4. As mentioned before, the last column in Table 3 indicates the variation in results among the multiple repeat experiments. The reaction between potassium carbonate and carbon occurs according to the following reaction K 2 CO3 2C 2K 3CO
(3)
The surface area (both BET and micropore surface area) and pore volume (both total and micropore volumes) both increased continuously with carbonization temperature and
9
impregnation ratio. The pores were resulted from the evolution of gaseous carbonization products and catalytic oxidation of carbon surface by potassium metallic salt. BET surface area was almost tri-fold from 514.28 m2/g and total pore volume increased from 0.286 cm3/g to 0.738 cm3/g when the carbonization temperature rose from 400 to 600ºC with the impregnation ratio of 5:2. The impregnation ratio also has significant enhanced the porous structure of the activated carbon at the same carbonization temperature of 600ºC. Compared to the results from KOH activation activated carbon shown in Table 3, K2CO3 activation produced activated carbons with higher surface area. The results shown in Table 4 also illustrate that K2CO3 can be used as a chemical activation agent to produce commercial activated carbon with surface area higher than 500 m2/g. Table 4 Pore structure characterization of activated carbons activated by K2CO3 with impregnation ratio from 1:2 to 5:2, and at carbonization temperature from 400 to 600 ºC.
Types of Activated Carbons LC1-600 LC3-600 LC5-600 LC5-400 LC5-500
Surface Area (m2/g) BET Surface Area 653.77 785.37 1408.70 514.28 593.86
DubininRadushkevich Micropore 554.64 601.45 862.92 360.38 431.86
Pore Volume (cm3/g)
Variation*
Total Pore Volume
Micro Pore Volume
(%)
0.352 0.370 0.738 0.286 0.324
0.290 0.293 0.513 0.126 0.143
0.3 0.5 0.3 0.7 0.8
* variation is defined as the maximum value of variation = ABS(Experimental value-Average value )/Average value ×100%.
The effect of carbonization temperature on the pore size distribution of the activated carbon by K2CO3 activation with impregnation ratio of 5:2 is given in Fig. 4. From this figure, it is obvious that the carbonization temperature gave significant effect on the pore structure of activated carbon produced. At low carbonization temperature of 400 ºC, the pore structure of activated carbon produced mainly consisted of mesopore, however, with
10
the increase of carbonization temperature, the creation of the mesopore structure and widening of mesopores to macropores also increased.
Pore Size Distribution (cm³/g·Å)
0.12
0.10
LC5-400 LC5-500 LC5-600
0.08
0.06 0.04
0.02 0.00 0
5
10 Pore Width (Å)
15
20
Figure 4 Pore size distribution of activated carbons by K2CO3 with impregnation ratio of 5:2 at variable carbonized temperatures.
The pore size distribution of activated carbons by K2CO3 activation with variable impregnation ratio at a carbonization temperature of 600 ºC is shown in Fig. 5. The effect of impregnation ratio in the pore development is very significant as seen in Fig. 5. Increasing the impregnation ratio from 1:2 to 5:2 lead to increase of the pore development and creation of new pores.
11
Pore Size Distribution (cm³/g·Å)
0.12 0.10 LC1-600 LC3-600 LC5-600
0.08 0.06 0.04 0.02 0.00 0
2
4
6
8
10
12
14
16
18
Pore Width (Å) Figure 5 Pore size distribution of the activated carbon by K 2CO3 with variable impregnation ratios at carbonization temperature of 600 ºC.
3.4 Comparison of the surface morphology of the carbons
Figure 6 Morphologies of (a) char from the untreated Loy Yang coal, (b) LH5-600, and (c) LC5-600.
The morphologies of carbon produced from the experiment by SEM are shown in Fig. 6. Comparison of the images confirms that both KOH and K2CO3 enhanced the porous structures. Furthermore, K2CO3 activation appeared to result in more meso and
12
macroporosity compared to KOH activation at the same impregnation ratio (5:2) and carbonization temperature of 600ºC. 4. Conclusion Activated carbons with high surface area were produced from Victorian Loy Yang brown coal with chemical activation using potassium hydroxide and potassium carbonate. The effects of impregnation ratio and carbonization temperature on pore structure were investigated. Carbonization temperature and impregnation ratio has a significant effect in the pore characterisation of activated carbons. Compared to potassium hydroxide, potassium carbonate activation produced activated carbon with higher carbon yield and surface area. The surface areas of activated carbons produced in this study are all in the specific surface area range from 500 to 2000 m2/g of the commercial activated carbons. Acknowledgement We acknowledge the support from the Victorian government (Energy Technology Innovation Strategy Programme).
References 1.
Yorgun, S., N. Vural, and H. Demiral, Preparation of high-surface area activated carbons from Paulownia wood by ZnCl2 activation. Microporous and Mesoporous Materials, 2009. 122(1-3): p. 189-194.
2.
Hayashi, J.i., et al., Preparing activated carbon from various nutshells by chemical activation with K2CO3. Carbon, 2002. 40(13): p. 2381-2386.
3.
Rodríguez-Reinoso, F. and M. Molina-Sabio, Activated carbons from lignocellulosic materials by chemical and/or physical activation: an overview. Carbon, 1992. 30(7): p. 1111-1118.
4.
Bagheri, N. and J. Abedi, Preparation of high surface area activated carbon from corn by chemical activation using potassium hydroxide. Chemical Engineering Research and Design, 2009. 87(8): p. 1059-1064.
13
5.
Bouchelta, C., et al., Preparation and characterization of activated carbon from date stones by physical activation with steam. Journal of Analytical and Applied Pyrolysis, 2008. 82(1): p. 70-77.
6.
Plaza, M.G., et al., Development of low-cost biomass-based adsorbents for postcombustion CO2 capture. Fuel, 2009. 88(12): p. 2442-2447.
7.
Nabais, J.M.V., et al., Production of activated carbons from coffee endocarp by CO2 and steam activation. Fuel Process. Technol., 2008. 89(3): p. 262-268.
8.
Dawson, E.A., et al., An investigation of the porosity of carbons prepared by constant rate activation in air. Carbon, 2003. 41(3): p. 571-578.
9.
Puziy, A.M., et al., Synthetic carbons activated with phosphoric acid III. Carbons prepared in air. Carbon, 2003. 41(6): p. 1181-1191.
10.
Aworn, A., P. Thiravetyan, and W. Nakbanpote, Preparation of CO2 activated carbon from corncob for monoethylene glycol adsorption. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2009. 333(1-3): p. 19-25.
11.
Prahas, D., et al., Activated carbon from jackfruit peel waste by H3PO4 chemical activation: Pore structure and surface chemistry characterization. Chemical Engineering Journal, 2008. 140(1-3): p. 32-42.
12.
Basta, A.H., et al., 2-Steps KOH activation of rice straw: An efficient method for preparing high-performance activated carbons. Bioresource Technology, 2009. 100(17): p. 3941-3947.
13.
Uçar, S., et al., Preparation and characterization of activated carbon produced from pomegranate seeds by ZnCl2 activation. Applied Surface Science, 2009. 255(21): p. 8890-8896.
14.
Guy, P.J. and G.J. Perry, Victorian brown coal as a source of industrial carbons: a review. Fuel, 1992. 71(10): p. 1083-1086.
15.
Dubinin, M.M., In Progress in Surface and Membrane Science ed. D.A. Cadenhead. Vol. 9. 1975, New York: Academic Press.
16.
Ehrburger, P., et al., Carbonization of coals in the presence of alkaline hydroxides and carbonates: Formation of activated carbons. Fuel, 1986. 65(10): p. 14471449.
17.
Ahmadpour, A. and D.D. Do, The preparation of activated carbon from macadamia nutshell by chemical activation. Carbon, 1997. 35(12): p. 1723-1732.
18.
Rouquerol, F., J. Rouquerol, and K.S. Sing, Adsorption by Powders and Porous Solids; Principles, Methodology and Applications. 1999, London: Academic Press. 14
15
Low-temperature catalytic graphitization of carbon material Ch. N. Barnakov1, A. P. Kozlov1, V. Yu. Malysheva1, S.K. Seit-Ablaeva2, M. A. Kerzhentsev3 Z. R. Ismagilov3 (1) Institute of Coal and Coal Chemistry SB RAS, Pr. Sovetskiy 18, Kemerovo, 650099, Russia (2) Kemerovo Technological Institute of Food Industry, Kemerovo, 650010, Russia (3) Boreskov Institute of Catalysis SB RAS, Pr. Ak. Lavrentieva 5, Novosibirsk, 630090, Russia Graphitization is a physical process (crystallization), which is classified by some authors as homogeneous or heterogeneous. Graphitization of “non-graphitizable carbon materials” or difficultly graphitizable materials is considered as a heterogeneous process. As a matter of fact, “non-graphitizable carbon materials” do not exist; any carbon material can be graphitized. Difficultly graphitizable materials should be heated to 3000-3200C (evaporation of carbon is followed by its condensation), and initial substance (carbon black, charcoal) should be strongly compressed before the process. Easily graphitizable materials are graphitized in the temperature range of 2200-2800С [1]. Graphitization temperature can be reduced, e.g., to 1900C using vanadium (1-3 wt.%) as a catalyst [2]. The use of various metal nanoparticles provides a further decrease in the crystallization temperature of carbon material. The patent application [3] discloses a method for synthesis of nanostructural carbon material from carbon precursor (an aromatic substance comprising one or two rings and having the COOH, C=O, OH, C=C, SO3, NH2, SOH or N=C=O functional groups) with the use of catalytic template nanoparticles. Carbon nanostructures are formed around a set of nanoparticles that serve as a base. Metal nanoparticles (iron, cobalt, nickel, etc.) are used for the purpose. Catalytic nanoparticles operate both as the nucleation sites of carbon nanostructure and as catalysts during carbonization. The carbonization proceeds in a nitrogen flow at 1150C for 3 hours or at 850C for 4 hours. The resulting nanostructural carbon material has high specific surface area, high porosity and high degree of graphitization. The patent application [4] outlines a methods for synthesis of nanostructural crystalline carbon material with high specific surface area. At the first step, carbon precursor is mixed with salts of transition metals and different oxides. The precursor can be represented by resorcinformaldehyde, phenol-formaldehyde, phenolic resin, melamine-formaldehyde, polyfurfuryl alcohol, polyacrylonitrile, saccharose, polypyrrole, polyvinylbenzene or petroleum pitch. Compounds that consist of metal cations (Fe, Co, Ni, Mo, V, Y, Zr, Nb, Li, Mg, Al, Si, K, Ca, Ti, Cr, Mn, Cu, Zn, Ga, Ge, As, In, Sn, Sb, La, Hf, Ta, W) and anions (CH3COO–, CH3COCH=CO–CH3, F–, Cl–, Br–, NO3–, SO42–, PO43–, COO–, ClO4–, RO–) can be used as the salts. SiO2, Al2O3, TiO2, CeO2, ZrO2, SnO2 and Y2O3 can serve as oxides of inorganic materials.
Amounts of components are taken so as to provide the precursor to salt ratio of 20:1 – 1:2 and the precursor to oxide ratio of 20:1 – 1:2. The method includes four steps. It is exemplified by the synthesis where a mixture of water, cobalt salt, nickel salt, resorcin, formaldehyde and quartz is stirred intensely, kept at 85C for 3 hours, and subjected to graphitization in a flow of argon at 900C for 3 hours. Nanostructural carbon material is obtained upon mixing the composite with 3 M NaOH for 3 hours to remove quartz particles, which is followed by washing with 2.5 М HNO3 for 1 hour to remove metal particles. From the indicated patents, it is evident that the use of salts and oxides of various transition metals, including nanoparticles of these metals, makes it possible to decrease significantly the graphitization temperature of carbon material and obtain graphite nanoparticles with high specific surface area. The presence of graphite nanoparticles is revealed by XRD spectra and electron micrographs, which demonstrate the packing of graphite stacks. The patent application [5] considers a low-temperature catalytic graphitization of carbon material into graphite. Nanoparticles of 3d metals (iron, cobalt, nickel) or their mixture stabilized by nanostructural carbon material [6], which we called Kemerit [7], in the amount of 0.2-8 wt.% referred to metal, were used as a catalyst to convert coal-tar pitch into foam graphite at a temperature of 600-1000С. The XRD studies were performed on a URD-6 diffractometer with CuК radiation. XRD spectra of the synthesized foam graphite samples are quite similar: they have a graphite component and an admixture of turbostratic carbon (see Fig. 1). Temperature dependence of the samples electrical conduction was determined by the four-contact method in the temperature range of 4.2-300 K. The samples under consideration were inserted into an ampoule. Contacts to the ampoule were made of silver wire 0.1 mm in diameter. For some samples, conduction of the obtained foam graphite at 300 K (see Fig. 2, example 1) corresponds to that of crystal graphite (see Fig. 3). Note that Raman spectroscopy reveals a much higher defectness of the foam graphite obtained from pitch as compared to the foam graphite produced via intercalation of graphite compounds [8] followed by their thermal expansion. Both the coal-tar and petroleum pitches always contain benzpyrene, which is a condensed penta-nuclear aromatic cycle with pronounced carcinogenic activity. Benzpyrene content varies from 1.3 [9] to 7 – 9 wt.% [10], thus making the problem of benzpyrene remediation quite topical. For the first time the possibility of benzpyrene fixing was shown in [11] for catalytic synthesis of anodic material from a mixture of needle coke and medium-temperature coal-tar pitch. Main results are discussed in [12] and partially reported below.
Fig. 1. XRD pattern of a foam graphite sample. 1.8
120
1.6
Examle31 Пример Пример Examle12
1.4
80
1 60 0.8 0.6
r(mWcm)
r(mWcm)
1.2
100
40
0.4 20 0.2 0 0
50
100
150
200
250
0 300
T(K)
Fig. 2. Temperature dependence of resistivity of the foam graphite synthesized at 600-1000С and different concentrations of catalyst (0.4% Ni + 5.8% FeSi – example 1); (0.4% Ni – example 3).
The temperature analysis of coal-tar pitch pyrolysis in the presence of catalyst was made using a synchronous thermoanalyzer STA 409 with quadrupole mass spectrometer QMS and STA 449 with quadrupole mass spectrometer QMS.
Figure 3 shows the curve obtained on a quadrupole mass spectrometer QMS for distribution of 113,126 mass fragments, the most typical fragments of benzpyrene, which release depends on the temperature of coal-tar pitch pyrolysis. Note that the obtained results support data of [9, 10] indicating the presence of benzpyrene in the initial coal-tar pitch, which is released at pyrolysis in the temperature range of 125 – 350С. Besides, benzpyrene is formed during pyrolysis of coal-tar pitch at 400 – 550С. Thus, due to pyrolysis of coal-tar pitch, amount of benzpyrene releasing to the atmosphere strongly exceeds its amount in the initial pitch. Figure 4 shows a characteristic curve obtained on a quadrupole mass spectrometer QMS for distribution of 113,126 mass fragments, the most typical fragments of benzpyrene, versus the pitch pyrolysis temperature in the presence of catalyst (cobalt nanoparticles). Note that in Fig. 4, the intrinsic pitch benzpyrene is released at 125 – 250С, whereas benzpyrene forming at further pyrolysis of the pitch is absent. A similar pattern is observed with nickel and iron nanoparticles. This allows a conclusion that during pyrolysis of coal-tar pitch, nanoparticles of 3d metals hinder the formation of benzpyrene at catalytic synthesis of foam graphite from coal-tar pitch, but do not fix benzpyrene that was present in the pitch.
TG, % 1 00
Ion c urrent, A TG
-1 2
90 80
9 1 0
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
[126]
70 60 [113]
50 40 30 20 0
200
400
600
O
800
1000
Tem peratur e, C
Fig. 3. Temperature dependences of the weight loss (TG) and mass spectra of the fragments of medium-temperature coal-tar pitch thermal decomposition.
T G, %
100
Ion cur rent, A -12
1 .0 0
TG
0 .9 [126]
0 .8 90
0 .7 0 .6 0 .5
[113]
80
0 .4 0 .3 70 0
200
400
600
Tem perature,
O
800
0 .2 1000
C
Fig. 4. Mass spectra of the fragments of medium-temperature coal-tar pitch thermal decomposition in the presence of cobalt nanoparticles.
Figure 5 shows a curve obtained on a quadrupole mass spectrometer QMS for distribution of 113,126 mass fragments, the most typical fragments of benzpyrene, versus the pitch pyrolysis temperature in the presence of ferrosilicium, which is used as a catalyst. In this case, benzpyrene traces are absent in the temperature range of 125 – 350С, but appear at 400 – 550С, when major part of the pitch is pyrolyzed. T G, % 100
I o n cu r re n t, A TG
8 1 0
-12
7
90
6 80
5
70
4 3
60
[113]
2 50
[126]
1
40 0
200
400
600
Te m p e ra tu re ,
O
800
1000
C
Fig. 5. Mass spectra of the fragments of thermal decomposition of needle coke and medium-temperature coal-tar pitch mixture in the presence of ferrosilicium.
As it was noted above, nanoparticles of 3d metals completely suppress the formation of benzpyrene during the coal-tar pitch pyrolysis, and ferrosilicium fixes benzpyrene occurring in the pitch at the initial stage of pyrolysis. A mixture of 3d metal nanoparticles and ferrosilicium completely eliminates the release of benzpyrene upon pyrolysis of coal-tar pitch. The TG, DSC curve of STA 449 thermoanalyzer in Fig. 6 shows the pitch pyrolysis that proceeds, in distinction to Figs. 3, 4 and 5, without growth of the curves of 113, 126 mass fragments corresponding to benzpyrene.
Ion current, A DSC, -12 10 mW/mg 0,1 0,50
TG, %
TG
100
DSC
[126]
0,0
0,45
-0,1
0,40
-0,2
0,35
-0,3
0,30
-0,4
0,25
-0,5 800
0,20
90 [113]
80
0
200
400
Temperature, OC
600
Fig. 6. Temperature dependences of the weight loss (TG), specific heat flow found from DSC and mass spectra of the fragments of medium-temperature coal-tar pitch thermal decomposition in the presence of nickel and ferrosilicium nanoparticles.
The presented data give grounds to state that the use of 3d metal nanoparticles in a mixture with ferrosilicium completely suppresses the release of benzpyrene to the reactor atmosphere and increases the yield of carbon residue due to benzpyrene fixing.
The work was supported by the RAS Presidium and SB RAS (Complex Integrated Project RAS Presidium No. 27.58; Integrated Project SB RAS No. 88).
REFERENCES 1. V.N. Krylov, Yu.N. Vil’k. Coal-Graphite Materials and their Application in Chemical Industry. Khimiya. 1965. 148 p. 2. E.V. Novikov, E.A. Belenkov, E.M. Baitinger. The effect of vanadia addition on graphitization of coke-pitch carbon materials. Bulletin of Chelyabinsk Scientific Center, vol. 3 (16), 2002.p.30-35. 3. US Patent application 2007/0265162, Nov. 15, 2007 4. US Patent application 2005/0008562, Jan. 13, 2005 5. RF Patent application 2009132561/15(04684) 6. RF Patent 2206394. A method for synthesis of nanostructural carbon material / Barnakov Ch.N., Seit-Ablaeva S.K., Kozlov A.P., Rokosov Yu.V., Fenelonov V.B., Parmon V.N. Issued 20.06.2003. Bull. 17. 7. Certificate for KEMERIT trademark No. 245674. Registered at the RF State Register of trademarks and service marks on 12.05.2003. 8. N.E. Sorokina, O.N. Shornikova, V.V. Avdeev. Formation regions of intercalation graphite compounds in the systems graphite-HNO3 (H2SO4)-H2KMO4. Inorganic Materials, 2007, vol. 43, No. 8, p. 924-928. 9. Afanasov I.M., Kepman A.V., Morozov V.A., Seleznev A.N., Avdeev V.V. Determination of the content of polyaromatic hydrocarbons in coal-tar pitch. J. Analyt. Chem. 2009. Vol. 64. No. 4. p. 376. 10. Privalov V.A., Stepanenko M.A. Сoal-Tar Pitch. Moscow, Metallurgy. 1981. 208 p. 11. RF Patent 2370437. 2009. 12. Barnakov Ch.N., Kozlov A.P., Malysheva V.Yu., Seit-Ablaeva S.K., Romanenko A.I., Anufrienko V.F., Ismagilov Z.R., Parmon V.N. A method of benzpyrene removal at the production of anodic carbon material, in particular at the production of graphite. Izvestia Vysshykh Uchebnykh Zavedeniy, Khimia i Khimicheskaya Tekhnologia (Bulletin of Universities, Chemistry and Chem. Engineering) 2010. Vol. 53. Issue 10.
Research on the Preparation of Highthermalconductivity Carbon Block by the Ordered Growth of Self-assembled Mesophase Jin Minglin1,*,Liu Ronghua1,2, ,Bao Zonghong2, Cheng Qingzhong1 ,Hu Jingxia1 (1 Department of Material Engineering; Shanghai Institute of Technology; Shanghai 200235;2 College of Chemistry and Chemical Engineering; Nanjing University of Technology; Nanjing 210009; China) Abstract Abstract: Highthermalconductivity carbon material is a focus in the new carbon material research field. With few QIs coal tar pitch as raw material in this work, the thermoplastic mesophase pitch was prepared by the thermal polymerization processes. The effect of molding pressure on the development of ordered mesophase was studied. In our experiments, the mesophase pitch containing the same TI was molded respectively at 25℃ to 600℃ under 6.18MPa,9.24 MPa,13.84 MPa,20.05 MPa, carbonized at 1000℃ and graphitized at 2800℃. The results showed that the graphitization degree of the carbon blocks ranges from 48.8% to72.1% by graphitized at 2800℃. With the increase of pressure, the apparent density increased from 1.47 g/cm3 to 1.58g/cm3 and the porosity decreased. It is shown by SEM that the regional mesophase displayed a sort of highly ordered structure with long fiber and the mesophase liquid crystal molecules were aligned perpendicular to the compressing in the molding process. The axial and radial direction of thermal conductivity range of the carbon blocks was 140.5-156W/m·K , 169.2-219.4W/ W/m·K, respectively. The anisotropic ratio increased as the pressure increased and reached 1.41.It was proved that one-step hot press molding not only increased the density and decreased the porosity of the carbon blocks, but also promoted the growth of mesophase in order, formed the long-range ordered, fibrous structure which was similar to carbon fibers. Consequently, the phonons conducted along the long-range ordered structure, the hindrance of the conduction of phonons was reduced. Key words: highthermalconductivity; pitch; mesophase; carbon material
1 INTRODUCTION Compared with metallic materials ,carbon materials has the advantage of low density, low heat expansion coefficient and good mechanical properties at high temperature. It is now widely used in the fields of aeronautics and integrated circuit(IC),such as the cooling devices of IC, aircraft nose cones and high-speed jet aircraft brake[1-4]. At home and abroad, there are mainly two ways to study the high thermal conductivity carbon materials: Firstly, carbon fibers as matrix material were waved, densificated and graphitized, the high thermal Supported by the Shanghai Committee of Science (No.09520500700) *Corresponding author: Fax:+86-21-64085150 E-mail addredd:[email protected](M.L.Jin)
and
Technology,
China
conductivity carbon material was obtained, such as Ma et.al have prepared the carbon material of which thermal conductivity(TC) was at room temperature(RT) higher than 600W/(m·K)[5]. Because of high cost, its application was limited. Secondly with the needle coke, staple fiber as aggregate ,the high-thermal-conductivity carbon materials were prepared by adding the binder, isocratic pressing, densification and graphitization processes, for example Qiu et.al [6]have studied that the TC of these carbon material was 165W/(m·K).Its TC was lower than the metal, so these carbon material have not been used in practice. With small amount of QIs coal tar pitch as raw material in calcination experiment, the thermoplastic mesophase pitch was prepared by the thermal polymerization process. The effects of by in-situ growth under pressure on the orderly growth of mesophase were studied. The highthermalconductivity carbon block has the long-range ordered, fibrous structure which was similar to carbon fibers.
2 Experiment 2.1 Raw material The properties of coal tar pitch were shown in Table 1 Table 1 Properties of Softening point /℃
Vf
coal tar pitch
TI /wt%
TS /w%
64.5
35.5
/% 282.5
28.83
2.2 Preparation of the thermoplastic mesophase pitch Mesophase pitch was synthesized from coal tar pitch in 12 tubes ( φ 20 × 200mm) which was placed one-time to a multi-tube, well-crucible stove with programmable temperature control. The temperature program comprised two steps: RT to 410 ℃, heating rate of 5 ℃ / min, and constant temperature for 18h in 410℃; up to 420 ℃ at 5 ℃ / min, then heated for 18h. 2.3 Hot press molding for the ordered growth The thermoplastic mesophase pitch was compacted in a mould by a hot-pressing apparatus(RT~600℃,0~40MPa)The process of hot press molding: the mesophase pitch containing the same TI were molded respectively at 25℃ to 600℃ under 6.18MPa,9.24 MPa,13.84 MPa,20.05 MPa. 2.4 Carbonization The samples which were placed into a porcelain boat were carbonized in the tube furnace. The temperature program is as follows: first the room temperature was raised to 650 ℃ at a rate of 2℃/min ,then programmed to rise to 800℃ at 1℃/min, finally raised to 1000℃
at 2℃/min and held for 1 hours. The nitrogen atmosphere was used in the whole process. 2.5 Graphitization The samples were graphitized in the graphitizing furnace. The temperature program is as follows: first the room temperature was raised to 1000 ℃ at a rate of 15℃/min ,then programmed to rise to 2700℃ at 26℃/min, finally raised to 2700℃ and held for 1 hours. The atmosphere of high purity argon gas was used in the whole process. 2.6 Testing of properties 2.6.1 Testing of physical properties TS content was measured according to GB/T2292-1997.The microstuture of pitch samples were observed by 500 times under Leica DMR optical microscope. The density of samples was measured according to GB/T 3810.3-2006.Measurements of the thermal diffusivity (a) and specific heat capacity (Cp) of the specimen (φ10×3mm) were carried out by using the laser-flash diffusivity instrument (Nano-Flash-Apparatus, LFA 447, NETZSCH). 2.6.2 Microstructure analysis X-ray diffraction (XRD) with a Cu Ka radiation (k = 0.15406 nm) was used for determining the primary structure and parameter of the graphite crystal: d002, where d002 was the graphite interlayer distance. The degree of graphitization (g) was calculated by the expression g = (0.3440-d002)/(0.3440-0.3354).Fractured micrographs of graphite blocks were observed through a field emission scanning electron microanalyser (FESEM, JEOL JSM-6360F, Japan).
3. RESULTS AND DISCUSSIONS 3.1 Analysis of Density and porosity Through the methods of macroanalysis , the density of carbon blocks is an important factor to determine their thermal performance [7]. The reduction of the pores inside the carbon materials with the increase of the density leads to the decrease of thermal resistance during the heat transfer process, so the carbon materials will have high thermal conductivity. Table3 The apparent density and porosity of the samples apparent density Pressure
porosity
hot-pressing
Carbonized
Graphitized
hot-pressing
Carbonized
Graphitized
samples
samples
samples
samples
samples
samples
6.18mpa
1.36
1.46
1.47
18.99%
24.46%
31.63%
9.24mpa
1.36
1.47
1.53
18.45%
23.28%
28.84%
13.79mpa
1.37
1.48
1.57
17.41%
23.47%
26.98%
20.05mpa
1.38
1.49
1.58
17.26%
22.86%
26.51%
The true density of hot-pressing sample is 1.67 g/cm3.After
graphitized at 2800℃,the sample true density increased roughly by 28.74% and reached to 2.15g/cm3.By carbonized and graphitized, the polycondensation and decomposition reaction of the polymer aromatic hydrocarbon in the samples had happened, so the grains moved closer to each other, the crystallization degree of the samples increased and the stacking between the basal plane which have aromatic structure became close.Table3 showed that the apparent density of the samples increased and the porosity decreased with increasing pressure from 6.18mpa to 20.05mpa. It showed that the molding pressure improved the tightness of grains in the samples and decreased the cracks and pores inside the carbon blocks. 3.2 Analysis of XRD Turbostratic Carbon was changed to graphitizing carbon by heat treatment. The degree of graphitization is quantitative characterization of the degree of closeness of the ideal crystal structure of graphite[8]. In our experiments, the micro-structural changes of the samples were investigated from two aspects which were heat treatment temperature and molding pressure,and then to explore its effect on thermal conductivity. The thermal conductivity of carbon material is determined by the concentration of charged particles (carrier) and the charged particles (carrier) mean free path of the basal plane which has aromatic structure.It showed that graphitization not only caused chemical bond fracture,increased non-matching electron and the concentration of carrier but also led to the increase of the basal plane which has aromatic structure ,the increase of the carrier mean free path and the decrease of grain boundaries and dislocations[9]. The decrease of the hindrance of the conduction of phonons improved the thermal conductivity of carbon materials.
Fig1 The X-ray diffraction spectrum of specimens from the samples heated at several temperature.( a:600℃;b:1000℃;c:2700℃) Fig1 was the X-ray diffraction pattern of mesophase pitch which was heat-treated at 600℃ under 20.05mpa,carbonized at 1000℃ and graphitized at 2700℃. The diffraction intensity of the heat-treated samples at 600 ℃ and 1000℃ respectively was low but not sharp, thus the degree of graphitization was difficult to calculate. When the heat-treated temperature reached 2700℃,the diffraction intensity (002) became much more high and sharper, the degree of graphitization reached 72.1%.It showed that the thermoplastic mesophase pitch converted to the crystalline structure of graphite by rising the heat-treated temperature in the heat treatment process, the basal plane which has aromatic structure and the carrier mean free path was increased.
Fig2 The influence of pressure on the graphitization degree Fig2 was the relationship of the degree of graphitization of the samples in different sections with the molding pressure. As can be seen from Fig2,the degree of graphitization with the molding pressure was increased, and the graphitization degree of cross-section was higher than longitudinal-section. The molding pressure was not only conducive to the ordered arrangement of the liquid crystal molecules in the thermoplastic mesophase pitch, but also promoted and controlled the orderly growth of liquid crystal molecules along perpendicular to the compressing direction which made the degree of order in cross-section higher than that of the longitudinal-section. 3.3 Analysis of SEM Fig3 were the SEM morphology of the samples under different molding pressure. As can be seen from Fig3,the longitudinal section of the samples showed wide axis layered graphite sheet showing layers of stacked form, and the surface has rupture and nodes, but no bump-shaped pores. During press molding process, the wide axis, ordered graphite layers in the mold rotated because of the pressure and a large cross-section aligned perpendicular to the compressing direction, when the graphite layers were the most stable[10]. From Fig a, b, c, d in Fig3 we can see that as the pressure increased, the graphite layer spacing had been reduced, then the thermal resistance between the graphite layers would also be reduced.
(a)
(b)
(c)
(d)
Fig3 SEM photos of the samples in the longitudinal section (a:20.05MPa;b:13.84MPa;c:9.24MPa;d:6.18MPa)
Fig3 and Fig4 showed the difference between the longitudinal-section and cross-section of samples. The surface of graphite sheet had the ordered, fibrous structure which was similar to carbon fibers. Further analysis could discover that the fibrous structure were clustered distribution and the surface of graphite sheet curl when the samples were forming under 6.18mpa.With the pressure growing, the cross-section samples were highly ordered and streamlined fibrous structure. This showed that the pressure promoted the further development of ordered structure in the thermoplastic mesophase pitch. Consequently, the phonons conducted along the ordered structure which was parallel to the graphite sheets and the hindrance of the conduction of phonons was reduced.
(a)
(b)
(c)
(d)
Fig3 SEM photos of the samples in the cross section (a:20.05MPa;b:13.84MPa;c:9.24MPa;d:6.18MPa)
3.4 Analysis of thermal conductivity The thermal conductivity of the samples is listed in table2.After graphitizing at 27000℃, The axial and radial thermal conductivity range of the carbon blocks was 140.5-156W/m·K , 169.2-219.4W/m· K, respectively. Either axial or radial of thermal conductivity increased with the increase of molding pressure and the degree of graphitization. It was proved that the increase of the density and the degree of graphitization promoted while the hindrance of the conduction of phonons was reduced. Table4 The thermal conductivity of the samples molded under the several pressure
The radial directi on The
(25℃)
Pressure/MPa
Density g/cm3
The degree of graphitization / %
thermal conductivity /W/m·K
6.18 9.24 13.84 20.05 6.18
1.447 1.526 1.566 1.576 1.447
65.1% 65.1% 70.9% 72.1% 48.8%
169.1 174.1 196.7 219.3 140.4
axial directi on
9.24 13.84 20.05
1.526 1.566 1.576
63.9% 66.3% 66.3%
156.0 154.9 155.8
Fig5 Variety curve of the thermal conductivity of the samples (the radial and axial direction)
Combined with SEM observation, we found that the wide axis layered graphite sheets showing layers of stacked form were aligned parallel to the radial direction. Theoretically, radial heat conductivity is carried out in the graphite layers, while the axial heat conduction is carried out between the layers[10]. As shown in Fig5, the radial thermal conductivity is higher than the axial thermal conductivity. The anisotropic ratio reached 1.41. The anisotropy ratio tended to increase with the growing pressure molding. The heat conductivity in the graphite layers is higher than that between the graphite layers, so the samples show a clear anisotropy.
4 CONCLUSION (1)With few QIs coal tar pitch as raw material, the thermoplastic mesophase pitch which was prepared by the thermal polymerization processes and contained same TI were molded respectively at 25℃ to 600℃ under 6.18MPa,9.24 MPa,13.84 MPa,20.05 MPa, carbonized at 1000℃ and graphitized at 2800℃. The reduction of the pores inside the carbon materials and the increase of the density the increase pressure lead to the decrease of thermal resistance during the heat transfer process, so the carbon materials will have high thermal conductivity. (2)By the analysis of XRD, the graphitization degree range of the carbon blocks was 48.8%-72.1% by graphitized at 2800℃,and the
graphitization
degree
of
cross-section
was
higher
than
longitudinal-section; In the hot-pressing progress, the wide axis, ordered graphite layers in the mold rotated by the pressure and a large cross-section aligned perpendicular to the compressing direction. With the pressure growing, the cross-section samples were highly ordered and streamlined fibrous structure. Consequently, the phonons conducted along the ordered structure which was parallel to the graphite sheets and the hindrance of the conduction of phonons was reduced. (3)The axial radial thermal conductivity range of the carbon blocks were 140.5-156W/m·K , 169.2-219.4W/ W/m·K respectively when the samples were graphitized. The radial thermal conductivity is higher than the axial thermal conductivity. The anisotropic ratio reached 1.41. The anisotropy ratio tended to increase with the growing pressure molding. The heat conductivity in the graphite layers is higher than that between the graphite layers, so the samples show a clear anisotropy.
REFERENCES [1] Lu Ruitao,Huang Zhenghong,Kang Feiyu.Research Progress in Carbon-based Funetional Materials with High Thermal Conductivity. Materials Review 2005;19(11): 69-72 [2]
Fitzer
E.
The
future
of
carbon–carbon
composites
[J].
Carbon,
1987;25(2):163–190. [3] Liu Zhanjun, Guo Quangui, Shi Jingli, Zhai Gengtai, Liu Lang. Graphite blocks with high thermal conductivity derived from natural graphite flake. Carbon 2008;46:414-421 [4] Blazewicz S, Blocki J, Chlopek J, Godlewski J, MickalowskiJ, Pakonski K, Piekarczyk T et al. Thin C/C composite shells for high energy physics: manufacture and properties [J]. Carbon, 1996; 34(11) :1393–1399. [5] Ma Zhaokun, Shi Jingli, Song Yan, Guo Quangui, Zhai Gengtai, Liu Lang. Carbon with
high
thermal
conductivity,
prepared
from
ribbon-shaped
mesosphase
pitch-based fibers. Carbon 2006;44: 1298-1352 [6] QIU Haipeng,GUO Quangui,SONG Yongzhong,ZHAI Gengtai,SONG Jinren,LIU Lang. Study of
the relationship between thermal conductivity and microcrystaline
parameters of bulk graphite.New Carbon Material
2002;17(1):36-40
[7] Gao Xiaoqing, Guo Quangui, Liu Lang, Song Jinren.The study progress on carbon materials
with
high
thermal
conductivity.
Journal
of
Functional
materials
2006;3(37):173-177 [8] Chen Jie, Xiong Xiang, Xiao Peng.Research and Development of High-thermal Conductivity Carbon/Carbon Composites. Materials Review 2006;11(20):431-434 [9]L
M
Manocha,
Ashish
Banerjee.Thermophysical
Warrier,
properties
of
S
Manocha,
densified
pitch
D
Sathiyamoorthy, based
S
carbon/carbon
materials—I. Unidirectional composites [J].Carbon 2006;44:480–487 [10] Michio Inagaki,Kang Feiyu.Carbon Materials Science and Engineering-From Fundamentals to Application. Beijing: Tsing University Press, 2006:23-40
UNEP Coal Combustion Partnership Area Activities Prior to 2013 Global Mercury Treaty Lesley Sloss, Principal Environmental Consultant, IEA Clean Coal Centre, London, UK [email protected] Wojciech Jozewicz, Department Manager, ARCADIS, Research Triangle Park, North Carolina, USA [email protected] Gunnar Futsaeter, Programme Officer, United Nations Environment Programme, Geneva, Switzerland [email protected] For presentation at the 2010 International Pittsburgh Coal Conference, Istanbul, Turkey, October 11-14, 2010 Abstract To address the risk to human health and the environment from anthropogenic mercury releases, the Governing Council (GC) of the United Nations Environment Programme (UNEP) decided in 2003 to establish the UNEP Mercury Programme. The UNEP Mercury Programme currently consists of two main complimentary activities: fulfilling the aims of the Global Mercury Partnership and the negotiation of a legally binding instrument. Negotiations of a legally binding instrument on mercury began in June 2010 and are to be finalized by February 2013. The Global Mercury Partnership is the main mechanism for the delivery of immediate actions on mercury during the negotiation process. The largest single anthropogenic emission source of mercury is the combustion of coal in power plants and industrial boilers. Practices capable of providing reductions in mercury emissions from coal-fired power plants have been summarized in UNEP’s POG (Process Optimisation Guidance) document. These practices include plant efficiency improvement, precombustion control/prevention measures, maximization of plant performance, co-benefit mercury removal by other pollutant emission control technologies and mercury-specific control processes. This document was presented to government agencies in four selected countries: South Africa, Russia, India, and China. When finalized, the document will serve as a tool for individual coal-fired, electricity-generating plants and for individual countries to: evaluating the opportunities to achieve multi-pollutant emission reductions at coal-fired power plants, assessing opportunities to improve energy efficiency of power plants and thus achieving reduction of greenhouse gas (GHG) emissions together with the reduction of mercury emissions, and providing guidance on how to optimize co-benefits of multi-pollutant control systems to reduce mercury emissions. This paper presents the contents of the POG document, including information on coal usage and typical examples of air pollutant control equipment deployed at coal-fired plants in each of the four countries of interest. Up-to-date progress and future plans for dissemination of the POG document in the four selected countries are also given. Finally, the paper presents the progress achieved in the four focused countries toward the completion of the UNEP background study.
1.
BACKGROUND
In 2009 the Governing Council (GC) of the United Nations Environment Programme (UNEP) agreed to elaborate a globally legally binding instrument on mercury to reduce risks to human health and the environment. UNEP was requested to convene an intergovernmental negotiating committee with the mandate to prepare the instrument with the goal of completing it prior to the GC session in 2013 (UNEP, 2010a). UNEP held the first International Negotiating Committee (INC1) meeting in Stockholm, Sweden, in June 2010 to begin finalizing an international, legally-binding instrument for the reduction of mercury emissions and use worldwide. Currently, it is unclear what form this instrument will take, but it has been decided that the instrument should be “comprehensive” and include provisions on a number of areas, e.g.: -
“To reduce the supply of mercury and enhance the capacity for its environmentally sound storage; To reduce the demand for mercury in products and processes; To reduce international trade in mercury; To reduce atmospheric emissions of mercury; To address mercury-containing waste and remediation of contaminated sites; To increase knowledge through awareness-raising and scientific information exchange.”
Coal combustion has been identified as the largest source of mercury emissions from human activities, contributing an estimated 26 percent of global emissions in 2005. It is therefore one of the priority target areas for work under the UNEP Global Mercury Programme. The Coal Partnership has been formed under the auspices of UNEP and comprises interested representatives from government, industry, academia, and commercial institutions worldwide. The partners meet at least once a year to discuss the current work and to determine how further progress may be promoted. Significant funding has been obtained from the European Commission to initiate a project in four target countries: China, India, Russia, and South Africa. Further funding has been provided by UNEP, and in-kind assistance has been forthcoming from groups such as the International Energy Agency’s (IEA) Clean Coal Centre (CCC), the U.S. Environmental Protection Agency (USEPA) and the U.S. Geological Society (USGS). Activities has presently only been initiated in China, Russia, and South Africa. The work is divided into three main areas: 1. Workshops to discuss, promote and disseminate the information provided in the Process Optimization Guidance document (POG) (UNEP, 2010b). Based on a “decision-tree” (see Figure 1), the POG - currently in final draft stage - has been produced as a simple tool to provide assistance to plant operators and governing agencies on the options available for mercury reduction at large-scale coal-fired power plants. These options range from efficiency improvements and fuel treatment, to optimizing existing control technologies for acid gases and particulates, and even bolt-on mercury-specific control processes.
The POG aims to simplify the decision-making process for determining the most appropriate and cost-effective mercury control options on a plant-by-plant basis. The POG has been reviewed by experts and interested parties in the three target countries. Workshops providing more specific information and guidance on mercury control are ongoing. 2. Assistance with inventory preparation. Collect information to improve accuracy of future emissions inventories for the sector (information on coal used (statistics, analysis, prognoses), characterization of power plants with regard to technologies used and develop an emissions inventory for the sector mostly based on using developed emission factors. Since mercury emission factors can vary by an order of magnitude or more for coal plants around the world, it is difficult to determine which emission factor is the most appropriate in each case. Measurements at full-scale plants firing typical coals can provide the most appropriate data for updating country- or plant-specific emission factors. A few mercury measurements campaigns at selected power plants in Russia and in South Africa have been conducted. 3. Mercury-reduction demonstration projects. Implement studies to demonstrate efficiency of co-beneficial techniques through and build local/national capacity on these issues, also with the aim of transferring information and lessons learnt to facilities and governments in other countries through full-scale projects at plants in the target countries.
2.
PERFORMANCE OPTIMIZATION GUIDANCE DOCUMENT (POG)
The POG is a document, currently in final draft form, summarizing the various ways of reducing mercury emissions from existing coal-fired power plants. Mercury reduction in coalfired plants is not always simple as the complex chemistry of mercury in coal combustion can mean that emissions from two relatively similar plants can be very different; mercury control at one plant may be much simpler and less expensive than at another plant. There is no single option for mercury control, understanding and comparing the options is a significant challenge for any plant operator (Sloss, 2008; 2009). For some plants, improving plant operating efficiency would not only reduce running costs but could reduce mercury emissions by up to 7 percent and, at the same time, reduce emissions of other pollutants. Coal cleaning/switching/blending, if feasible with available coal supplies, can maximize the proportion of mercury in the “sticky” oxidized form, which is easier to capture in existing pollution control systems for particulates, NOX, and SOX. Additionally, these pollution control systems can be fine-tuned or enhanced with additives to increase the concomitant capture of mercury without significant plant disruption or retrofit. The POG decision tree (Figure1) assists the user in analyzing which sections of the POG report are most relevant in each case. Where possible, the POG includes comparisons of the relative costs of various mercury reduction options, but it is emphasized that these options may vary significantly on a plant-by-plant basis.
NOX Control
SNCR / No NOX Control
NOX SCR
PM Control
ESP
PM
FF
ESP
Other
SO2 Control
Yes Wet
SO2
Wet FGD Additives
Yes
Yes
Dry
Wet
• ESP Tuning • Treated/Enhanced ACI • Coal Treatment
Figure 1. POG Decision Tree
SO2
Yes
Yes
Dry
Wet
No
Treated / Enhanced ACI
Wet FGD Additives
Oxidants
FF
Other
No
Hg Control Options
PM
SO2
Yes
Yes
Dry
Wet
Treated / Untreated Sorbents
• Co-benefit Oxidation/Capture • Oxidation Additives
Dry
No
No
• Lime • Treated/Enhanced ACI
Yes
SO2
• Co-benefit Oxidation/Capture • Oxidation Additives
Treated / Enhanced ACI
• ESP Tuning • Treated/Enhanced ACI • Coal Treatment
• Coal Blending • Oxidants
• Lime • Treated / Untreated Sorbents
Treated / Untreated Sorbents
3.
INTERNATIONAL ACTIVITIES
This section summarizes the work achieved to date as well as planned work in the selected target countries. 3.1
China
Consuming more coal than any other nation, China is in a prime position to demonstrate significant mercury reductions. Past, current, and proposed pollution control requirements in China - which focus on efficiency, particulate control, and acid gas control - are likely to result in a co-benefit reduction in mercury emissions. The UNEP coal partnership project work to date in China has concentrated on improving the data in the national inventory on emissions from large-scale coal-fired power plants. The work has been carried out by Tsinghua University, Beijng under subcontract to the Ministry of Environmental Protection (MEP). This work has involved the collection of the following information:
National and provincial installed power plant capacity and electricity generation via coal combustion in 2008, as available (see Figure 2 for capacity/generation trend)
Installed configuration of any air pollution control systems (particulate matter, SOX, NOX)
Any published emissions measurement data for mercury
Although this work is still underway, some initial data are already available. In 2008, there were over 600 GW of coal-fired capacity providing over 2803×109 kWh of electricity. More than 95 percent of these plants have electrostatic precipitators (ESP) and over 60 percent have flue gas desulfurization (FGD). The installation rate of low NOX burners and selective catalytic reduction (SCR) units is also increasing, with all plants installed after 2003 having to have some form of NOX control. Mercury emission reduction rates at Chinese plants ranged from under 10 percent to over 80 percent.
Figure 2. Installed Coal-fired Capacity and Electricity Generation in China (Wang and Wu, 2010) To improve the data available and to enhance the production of appropriate mercury emission factors, the UNEP coal partnership project in China also involves the collection of 150 coal samples. Analysis of the mercury content of these samples is being combined with data from previous studies to provide a more comprehensive database of coal mercury content from different mines and regions in China. Based on currently available studies, the average mercury content in Chinese raw coal ranges from 0.15 to 0.33 mg/kg (Wang and Wu, 2010), as shown in Figure 3. All these data, along with information on expected growth through 2020 of electricity production, coal use, and control technologies, are being submitted to UNEP’s paragraph 29 study.
Figure 3. Average Mercury Content of Chinese Coals based on Different Studies (Wang and Wu, 2010).
3.2
Russia
The UNEP coal partnership project work in Russia has concentrated on improving the data in the national inventory on emissions from coal-fired plants. The work has been carried out by Russia’s Scientific Research Institute for Atmospheric Air Protection, St. Petersburg. The work has involved the collection of the following information:
National installed power plant capacity, as available
Any installed air pollution control systems
Any published emissions measurement data for mercury
Although this work is still underway, some preliminary data are already available. Currently, there are over 40 GW of coal-fired capacity in Russia. There are no SOX or NOX controls installed. The three main types of particulate matter (PM) emission controls are wet PM scrubbers, ESPs, and cyclones. In-kind assistance and support from the USEPA in the form of a team of stack testing professionals equipped with the USEPA mercury measurement tool-kit (Siperstein et al., 2010) allowed monitoring of mercury emissions from four different coal-fired units at two separate coal-fired plants (Romanov, 2010). At the same time, mercury in coal and ash were analyzed in paired studies by Russian experts and the USGS. The major types of coal used for power generation in Russia are shown in Figure 4. This work was only begun in May 2010 and, following further assessment, it is hoped that an updated emission factor appropriate to Russian coals and utilities will be provided to update and enhance the accuracy of the paragraph 29 study with respect to emissions from Russia.
Figure 4. Types of coal used for power generation in Russia (Romanov, 2010)
Two projects have been proposed to determine the efficiency of mercury removal under conditions representative of utilities in Russia. The first project will test additives (chemical reagents) to a wet PM scrubber. These additives will be tested at different concentrations in an attempt to find the optimum addition rate for maximization of mercury oxidation. Wet PM scrubbers for fly ash collection are commonly used gas cleaning equipment at Russia’s coal–
fired power plants. These scrubbers address over 30 percent of the total coal-fired capacity of the Russian power sector. The second project will involve the injection of activated carbon into the flue gas, where activated carbon captures mercury through physical and chemical reactions on its surface. The carbon is then captured by the plant’s PM control device. Testing will be conducted at a pilot plant facility located at a power plant outside of Moscow. The plant currently fires a low-rank coal that is used widely in Russia. Thus, this project will provide mercury control data on a low-rank coal similar to that used in other countries of the world. Sources of activated carbon to be used in the project will be explored in Russia, Europe, and the United States. It is anticipated that the work on both of these projects will commence late in 2010. 3.3
South Africa
Work has been initiated through South Africa’s Department of Environmental Affairs (DEA), in Pretoria, Johannesburg, along with support from the Council for Scientific and Industrial Research (CSIR), Eskom (the state owned utility), Sasol, and other South African organizations. In-kind assistance and support from the USEPA in the form of a team of stack testing professionals equipped with the USEPA mercury measurement tool-kit (Siperstein et al., 2010) allowed monitoring of mercury emissions from six different coal-fired units at two separate coal-fired plants. At the same time, mercury in coal and ash were analyzed in paired studies by South African experts and the USGS. Although this work was only begun in May 2010, initial results indicate extremely interesting results with respect to the greater capture of mercury in baghouse PM control systems than in ESPs (Scott 2010). Following further assessment, it is hoped that an updated emission factor appropriate to South African coals and utilities will be provided to update and enhance the accuracy of the paragraph 29 study with respect to emissions from South Africa. A project has been proposed to determine the efficacy of a coal cleaning process (designed for sulfur control) in the removal of mercury. The coal treatment system proposed has a low water-requirement and would therefore be ideal for coal treatment in water-restricted areas of South Africa. It is proposed that coal will be analyzed before and after coal treatment along with emissions and ash from the clean and unclean coals fired in a pilot combustor. It is hoped that work will commence on this project in the autumn of 2010.
4. FUTURE WORK As mentioned above, some of the project work in Russia and South Africa is proposed to commence within the next year. UNEP will continue to report on the results of these projects through the coal partnership on the UNEP website. While India has agreed to take part in all activities of the project, no contract has been signed. Possible mercury demonstration projects are being discussed in the hope that India may move quickly once an agreement is signed. UNEP and the coal partnership are keen to help establish projects in other economies in transition and would be happy to discuss potential collaborative agreements with any interested parties. The coal partnership and other UNEP partnerships are always keen to welcome new members. If you are interested, please log on for more information: http://www.unep.org/hazardoussubstances/Mercury/GlobalMercuryPartnership/tabid/1253/lan guage/en-US/Default.aspx
REFERENCES Romanov, A. (2010) Update on UNEP Inventory Work in Russia, Presented at the MEC-7 Meeting, Glasgow, Scotland, June 2010. Scott, G. (2010) Update on the UNEP Inventory Work in South Africa, Presented at the MEC7 Meeting, Glasgow, Scotland, June 2010. Siperstein, J., He, Q., Mashianov, N., Ryan, J., Senior, C., and Van Otten, B. (2010) Experience with Hg Field Measurements Using USEPA Reference Methods, Presented at the MEC-7 Meeting, Glasgow, Scotland, June 2010. Sloss L L (2008) Economics of mercury control. Report CCC/134. London, UK, IEA CCC, 60 pp, ISBN: 978-92-9029-453-5, January 2008. Sloss L L (2009) Impact of emissions legislation on existing coal-fired plants. Report CCC/145. London, UK, IEA CCC, 51 pp. ISBN: 978-92-90290-464-1, February 2009. UNEP (2010a) http://unep.org/hazardoussubstances/Mercury/Negotiations/Mandates/ (UNEP Report of the Governing Council, 25th session 16-20 February, 2009; GC25/5, III, para 25&26). UNEP (2010b) Progress optimization guidance document for reducing mercury emissions from coal combustion in power plants. Geneva, Switzerland, UNEP Chemicals Branch, DTIE, 94 pp (IN DRAFT). Wang, S. and Wu, Y. (2010) Reducing Mercury Emissions from Coal Combustion in the Energy Sector, Presented at the MEC-7 Meeting, Glasgow, Scotland, June 2010.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COAL SCIENCE: TRACE ELEMENTS/EMISSION DIRECT MEASUREMENT OF MERCURY IN SIMULATED FLUE GAS Bihter Padak Department of Energy Resources Engineering, Stanford University 367 Panama Street, Green Earth Sciences 065, Stanford, CA 94305, USA [email protected], 1-650-725-0752. Jennifer Wilcox Department of Energy Resources Engineering, Stanford University 367 Panama Street, Green Earth Sciences 065, Stanford, CA 94305, USA [email protected], 1-650-724-9449.
Abstract: Homogeneous oxidation of mercury in the flue gas of coal combustion utility boilers has been studied for many years to understand the speciation of mercury. In spite of a vast amount of experimental studies, supported by modeling efforts, there are still many questions to be answered and the speciation of mercury is not fully understood yet. One needs to be able to make precise mercury measurements to understand its speciation and accurately predict the extents of mercury oxidation. Traditionally “difference” techniques are used which involves the direct measurement of elemental mercury. These techniques do not allow for distinguishing between the two different oxidized forms, Hg+ and Hg+2 which makes it difficult to understand mercury speciation. It has recently been shown that mercury measurements performed with wet chemical conditioning systems are biased, resulting in inaccurate partitioning between oxidized and elemental mercury species. Given the shortcomings of the difference techniques, it is essential to measure oxidized and elemental mercury directly and hence separately to have a complete understanding of mercury speciation. In this study a custom-built electron ionization quadrupole mass spectrometer (EI-QMS) will be used to directly measure mercury species in combustion flue gas. To accurately measure the low concentrations of different mercury species present in coal combustion flue gases, the EI-QMS must be sensitive to concentrations in the ppb range, which can pose a challenge. To increase the sensitivity, the system has been upgraded to have a supersonic beam and a skimmer is placed after the first orifice that is heated. In such a setup, scattering of the molecular beam is avoided and the amount of gas which reaches the ionization region, and subsequently the ion detector, is maximized, thereby improving the sensitivity of the instrument. Also a tuning fork chopper has been implemented in the system along with a lock-in amplifier to enhance the signal-to-noise ratio. A benefit of employing a mass spectrometer is, unlike traditional impinge-based methods, the oxidized forms can be isolated and individually identified because it separates the products based on their mass-to-charge ratio. With this custom-built instrument, mercury species will be directly measured for the first time for high temperature combustion applications. By directly measuring mercury species accurately, one can determine the actual extent of mercury oxidation in the flue gas, which will aid in developing mercury control technologies. In addition, not only mercury but also other trace metals in flue gas such as arsenic and selenium can be measured and speciated in flue gas.
2010 International Pittsburgh Coal Conference Istanbul, Turkey, October 11 – 14, 2010 Manuscript Submission PROGRAM TOPIC: COMBUSTION
Modeling Trace Element Release from Included and Excluded Pyrite during Pulverized Coal Combustion Wayne S. Seames, Ph.D. Professor Department of Chemical Engineering Harrington Hall Room 323 241 Centennial Drive Stop 7101 Grand Forks, ND 58202-7101 Phone: (701)777-2958 Fax: (701)777-3773 Email: [email protected] Esam I. Jassim, Ph.D. Post Doctoral Fellow Department of Chemical Engineering Harrington Hall Room 323 241 Centennial Drive Stop 7101 Grand Forks, ND 58202-7101 Phone: (701)777-4244 Fax: (701)777-3773 Email: [email protected] Steven A. Benson, Ph.D. Professor Department of Chemical Engineering Harrington Hall Room 323 241 Centennial Drive Stop 7101 Grand Forks, ND 58202-7101 Phone: (701)777-5177 Fax: (701)777-3773 Email: [email protected]
Abstract: A mathematical model was developed to study the vaporization of trace elements during the combustion of individual coal and pyrite particle. The initial modeling has focused on the release of As, Se and Sb associated with pyrite since it is the host site for these elements. The model considers the release of the trace elements from pyrite included in coal particles and excluded from coal particles. The model predicts the amount of elements vaporized as a function of coal particle size, mineral grain size, mineral grain association, flame temperature, and residence time. The model considers the combustion of the coal char and the decomposition, fragmentation, melting, and coalescence processes of the pyrite particles.
Introduction: The ability to model the transformations of trace elements during the combustion process and emissions control equipment is very important due to their emissions and impacts on system performance. considered in the investigation include arsenic, selenium and antimony.
The elements
These elements vaporize upon
combustion and will react and condense to varying degrees depending upon the characteristics of the element and other constituents in the gas stream. The mineral pyrite in coal is a host site for many trace elements (Finkelman, 1994). For antimony, some work suggests that Sb is in the form of sulfides in the coal and dispersed with the organic matrix. Arsenic association with pyrite is well known (Minkin and other , 1984). Selenium behaves like sulfur and can be incorporated into sulfide minerals and be organically associated in the coal. Selenium has been found to be associated with pyrite and galena. In coal pulverized to -200 mesh pyrite grains can be either included in coal particles or excluded from coal particles.
The transformational environments associated with these two forms of pyrite are significantly
different. Pyrite associated with a coal particle will be exposed to devolatilization followed by the combustion of the char particle. The physical and chemical characteristics of the char structure during the course of the combustion process have an impact on the particle size distribution on the release of the trace elements and other inorganic ash forming components. Oxidation of the char increases the temperature of the char particle as well as the decomposing pyrite and will impact the degree of vaporization of the trace elements. When char particles burn at high temperatures (above 1800 K), the carbon oxidation rate is sufficiently fast enough to cause oxygen partial pressures within the porous char extremely low. Under these conditions, the local reducing environment occurring within the burning char particles may facilitate the chemical reduction of refractory metal oxides to highly volatile suboxides (SiO, Al2O3) or metals (Mg, Ca, and Fe). As suboxides or metal vapors diffuse out of the char and through the char external boundary layer, they encounter increasing oxygen potential of the bulk gas. The vaporized trace element can react and condense during gas cooling and be concentrated in the submicromet ash size fractions or may remain in the vapor phase. Several size modes for ash materials have been identified for ash materials produced during coal combustion.
Seames (2003) found that for coal combustion, ash
particulates having the diameter ranging from 0.03 to 100 μm several modes. The finest mode at about 0.1µm is due to the homogeneous condensation of flame volatilized species, intermediate mode of fine fragment mode is due to multiple mechanisms that include condensation, fragmentation, and disintegration of ash particles; and the large mode at a diameter of about 10 to 15 µm is due to the minerals and organically associated components
interacting upon combustion.
This paper involves modeling the evolution of the trace elements from pyrite
particles during combustion. Modeling Approach: As the coal is heated during combustion devolatilization occurs allowing the volatile matter and possibly organically associated inorganic elements to be released. The emitted volatile matter will combusts, where the associated organically bound elements can be vaporized, oxidized, nucleate and eventually form the submicron particulates. The porous char containing the pyrite particle, remaining after devolatilization, will combust and the associated pyrite particle will undergo chemical and physical transformation that result in the release of trace elements including As, Se, and Sb. The model considers devolatilization, char combustion, and trace element vaporization. Devolatilization Model: The devolatilization model can be expressed as simple reaction rate formulations. The volatile releasing rate may be written as: N
m vo m p i Ri
1
i 1
Ri Ai exp E i / RT p
i = 1,2,--N
2
Where: N is the number of volatile elements Tp the particle temperature Ai the rate constant of devolatilization for element i βi the mass fraction of element i in the coal particle mp mass of the coal particle
Char Combustion Model: The mechanism of char burning includes homogenous and heterogeneous reactions, mass transfer effects inside the pore of the char, and diffusion inside the boundary layer surrounding the particle. The characteristics, properties and sizes of the char particle may significantly vary depending on the type of the coal and devolatilization conditions. Char particles are burned at the exterior and interior surface areas with shrinking in size and/or density. First, the oxygen diffuses from the bulk flow to the exterior surface of char particle and into char pore which eventually react at both surfaces. The reacted gas products in the pores as well as the products that formed at particle exterior surface diffuse to the bulk flow. Since different chars have different pores structures, the rate of gas diffusions in the pore are different.
For single char particles, it is reasonable assuming that the temperature inside a particle is uniformly distributed. The heat released from carbon combustion and the heat released/absorbed by other inclusion reaction at the particle surface are balanced with the convection, radiation and heat absorbed by the char particle, i.e. m pC p
dT p dt
d p2 hTg T p Tw4 T p4 H R
3
or m p C p( c ) dT p 4rs2
dt
T p4 TW4 hT p T
1 4rs2
k
MW Q H i 1
i
i
i
4
where ΔHi exothermic/ endothermic heat of reaction; C p(c ) specific heat for carbon, σ Boltzmann constant, MW molecular weight, Q mole released per sec., ε emissivity. The convection heat transfer coefficient (h) is calculated as: Nu
2hrs 2 ; where Kg stands for thermal conductivity of the surrounding gas. Kg
The radiative transfer term is determined by assuming that a given particle faces only the distance wall temperature (TW), and is not affected by other particles. Evolution of Trace Elements: Zeng et al (2001) have proposed an approach to describing the vaporization process of trace elements. A three consecutive sub-processes, depict in Figure 1, govern the releasing of TEs from pyrite. Those are: i) transport of molecules or atoms through the bulk pyrite or other mineral liquid (melt) to the melt/gas interface, ii) vaporization of elements at the surface of the melt, and iii) transport of molecules/atoms through the pore of the char to the atmosphere. Vaporization Diffusion
pore
Figure 1. Vaporization sub-processes of Trace elements.
The correlation used to predict the rate of vaporization can be written as follow: k T Ae E / RT
5
Where A: the overall mass transfer coefficient (mole/s), E: activation Energy (J/mole K).
Results and Discussion:
Included Pyrite The initial associate of the pyrite particle is in the coal particle centered in the core as shown in Figure 2. The pyrite is concentrated at the centre of a sphere particle and exposed to hot environment once all carbonaceous element burns out. The particle surface temperature increases during the devolatilization process of volatile matter and combustion of char as shown in Figure 3. However, once all organic matter consumed and the pyrite is exposed to high temperature, sulfur starts to evolve and exothermically reacts with oxygen increase the surface temperature. As the sulfur reaction and evolution gradually decays, the surface of the pyrite melts.
Figure 2. Coal particle with pyrite inclusion.
Figure 3. History of Particle surface temperature during combustion The modeling results show that the rate of trace element evolution from the pyrite as a function of time. Figure 4 presents the vaporization rate of As, Se, Sb. The evolution process of the trace elements is assumed to occur after carbon is completely combusted and pyrite surface is exposed to the high temperature oxidizing environment.
Figure 4. Vaporization rate of trace elements for included pyrite.
Excluded Pyrite Equation 3 can be applied to the process of trace elements’ evolution from excluded pyrite as well by defining ∆H as the exothermic reaction of sulfur released as a result of pyrite chemical transformation during the course of combustion. Since the vaporization rate of trace element changes directly as a function of the pyrite surface temperature during the course of combustion, the instantaneous rate of vaporization is not constant. Hence, a parameter described as the average rate of vaporization is introduced. The average rate of vaporization is simply determined by integrating the instantaneous rate for entire residence time as illustrated in Eq. (6) and shown in Figure 5. t
2 R k k (t )dt tR 0
6
Figure 5. Instantaneous and average rate of vaporization for excluded pyrite. Figure 6 illustrates the average rate of vaporization as a function of excluded pyrite particle size. It is clearly presented that submicron sizes release higher rate of trace elements than large particles. This returns to the fact that small particles remains at high temperature during course of combustion for relatively long period. Also, the figure demonstrates that the vaporization rate of Antimony is higher than Selenium and Arsenic.
Figure 6. Vaporization rate of trace element as a function of excluded pyrite size.
Conclusion
The evolution of three trace elements during coal combustion has been modeled using the criteria of three consecutive transportation processes of Zeng et al (2001). Since the rate of vaporization varies with the temperature of particle, the model has introduced a parameter named average rate of vaporization to study the impact of residence time and flame temperature on the vaporization process. The results have shown that antimony has highest rate of vaporization followed by selenium then arsenic. In term of particle size, the study demonstrates that large amount of elements is predicted to be released from small particles.
Acknowledgment
Funding for this study was provided by the U.S. Department of Energy (Grant DE-FG02-6ER46292). Neither the United States Government nor the agency thereof, nor any of the employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information disclosed. Reference herein does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or the agency thereof or by the State of North Dakota.
References:
Desrosiers R.S., Riehl J.W., Ulrich G.D., Chiu A.S. In Seventeenth Symposium (International) on Combustion; the Combustion Institute: Pittsburgh, PA, 1979; pp 1395-1406. Finkelman R.B. Modes of Occurrence of Potentially Hazardous Elements in Coal: Levels of Confidence. Trace Element Transformations in Coal-Fired Power Systems; Fuel Processing Technology 1994, 39 (1–3); Benson, S.A.; Steadman, E.N.; Mehta, A.K.; Schmidt, C.E., Eds.; pp 21–34. He R., Suda T., Fujimori T., Sato J. Effects of Particle Sizes on Transport Phenomena in Single Char Combustion. International Journal of Heat and Mass Transfer, 2003, 46, 3619-3627. Lockwood F.C., Yousif S.A. Model for the Particulate Matter Enrichment with Toxic Metals in Solid Fuel Flames. Fuel Process. Technol. 2000, 65-66, 439-457. Minkin J.A., Finkelman R.B., Thompson C.L., Chao E.C.T., Ruppert L.F., Blank H., Cecil C.B. Microcharacterization of Arsenic- and Selenium-Bearing Pyrite in Upper Freeport Coal. Indiana County, Pennsylvania. Scanning Electron Microscopy 1984, IV, 1515–11524. Mitchell R.E., Ma L., Kim B. On the Behaviour of Pulverized Coal Chars. Combustion and Flame, 2007, 151, 426-436. Quann R. J., Neville M., Janghorbant M., Mims C.A., Sarofim A.F. Mineral Matter and Trace-Element Vaporization in a Laboratory Pulverized Coal Combustion System. Environ. Sci. Technol. 1982, 16 (16), 776781. Seams W.S. An initial study of the fine fragmentation fly ash particle mode generated during pulverized coal combustion. Fuel process Technology, 2003, 81, 109-125. Williams A., Backreedy R., Habib R., Jones J.M., Pourkashanian M. Modelling Coal Combustion: The Current Position. Fuel 8, 2002, 605-618. Zeng T., Sarofim A.F., Senior C.L. Vaporization of Arsenic, Selenium and Antimony during Coal Combustion. Combustion and Flame, 2001,126(3), 1714-1724.
ONLINE MONITORING OF BORON IN FLUE GAS DESULFURIZATION EFFLUENTS BY FULLY AUTOMATED MEASURING EQUIPMENT Seiichi Ohyama a), Keiko Abe a), Hitoshi Ohsumi a), Hirokazu Kobayashi b), Naotsugu Miyazaki c), Koji Miyadera c), and Kin-ichi Akasaka d) a)
Central Research Institute of Electric Power Industry, 1646 Abiko, Abiko-shi, Chiba 270 1194, Japan Central Research Institute of Electric Power Industry, 1-6-1 Ohtemachi, Chiyoda-ku, Tokyo 100 8126, Japan c) DKK-TOA Corporation, 2-214 Sakuragaoka, Higashiyamato-shi, Tokyo 207 0022, Japan d) DKK-TOA Corporation, 1-29-10 Takadanobaba, Shinjuku-ku, Tokyo 169 8648, Japan b)
ABSTRACT We developed a fully automated measuring equipment for aqueous boron (referred to as the online boron monitor) on the basis of a rapid potentiometric determination method using a commercial BF4- ion-selective electrode (ISE). The equipment can measure boron compounds with concentration ranging from a few to several hundred mg/L, and the measurement is completed in less than 20 min without any pretreatment of the sample. In the monitor, a series of operations for the measurement, i.e., sampling and dispensing of the sample, addition of the chemicals, acquisition and processing of potentiometric data, rinsing of the measurement cell, and calibration of the BF4- ISE, is automated. To demonstrate the performance, we installed the monitor in two full-scale coal-fired power plants and measured the effluent from a flue gas desulfurization (FGD) unit. The boron concentration in the effluents varied significantly depending on the type of coal and the load of power generation. An excellent correlation (R2 = 0.987) was obtained in the measurements between the online boron monitor and the conventional measurement (ICP-AES). The online boron monitor revealed long term behavior of boron in the FGD effluent for the first time and thus it can serve as a useful tool for managing boron emission in process effluents. INTRODUCTION Boron is an essential micronutrient for plants and animals, but also is toxic in large doses. For plants, different species require different levels of boron for optimum growth and its allowable range is extremely narrow. For humans, excessive ingestion over a long period may result in eczema, stomachache and nausea.1 Thus, the World Health Organization (WHO) reports the drinking-water guideline value of boron to be 0.5 mg/L.1 In Japan, the Environmental Quality Standard for water pollution, which is a target level to be achieved in public water to protect human health, is set at 1 mg/L for boron.2 Reflecting this establishment, the National Effluent Standard for boron is set at 10 mg/L and 230 mg/L for effluents discharged to terrestrial water bodies and to coastal water bodies, respectively.3 Boron is one of the toxic trace elements contained in coal.4, 5 Since boron is relatively volatile, the boron in coal is released to the flue gas owing to its combustion. Then the gaseous boron species cool and condense on solid surfaces or in aqueous solutions by adsorption or absorption. In pulverized coal-fired power plants with a flue gas desulfurization (FGD) unit, the boron in coal is released into the flue gas during combustion and then transferred downstream to the flue gas cleaning system. The boron in coal is mainly partitioned into clinker (bottom ash of the boiler), fly ash and FGD wastewaters.6-9 The main source of aqueous boron in power plants is effluent from the FGD unit. We must achieve better control of boron emission to meet the severe regulations. In order to manage and control boron emission effectively, we must monitor the concentration of boron in the effluent. This study is aimed at developing fully automated measuring equipment for monitoring process effluent and at demonstrating its performance and effectiveness for the management of boron emission. THERMODYNAMICS To build a process monitor for aqueous boron, a simple and rapid determination method that can be applied to on-site measurement in a compact size is required. The typical measurement methods of boron, i.e., inductively coupled plasma atomic emission spectrometry (ICP-AES) and colorimetric analysis (spectrophotometry), require bulky and costly analytical equipment or a time-consuming complicated pretreatment, both of which are difficult to apply to onsite analysis. We focus on potentiometric measurement using a commercial tetrafluoroborate (BF4-) ion-selective electrode (ISE).
In the conventional potentiometric measurement,10-16 aqueous boron species react with fluoride species (HF, F-, H2F2, HF2-) to form BF4-, which is measured with a BF4- ISE. In the case of H3BO3 with NaF, the overall reaction is (1) H3BO3 + 4NaF → 4Na+ + BF4- + 3OH-. If we convert aqueous boron species completely to BF4 , we can measure the total concentration of aqueous boron as BF4- with a BF4- ISE. We calculated the chemical equilibrium of the reaction according to the procedure reported by Mizoguchi et al.,17 considering the following fluoroborate-forming reactions, (2) H3BO3 + F- → BF(OH)3- , (3) BF(OH)3- + F- → BF2(OH)2- + OH- , (4) BF2(OH)2- + F- → BF3(OH)- + OH- , (5) BF3(OH) + F → BF4 + OH , and fluoride-related reactions, (6) HF → H+ + F- , (7) 2HF → H2F2 , (8) HF + F- → HF2- . Figure 1 shows the equilibrium fraction of BF4- at a stoichiometric ratio for the reaction (1) (F/B = 4, molar). The BF4fraction depends on the pH and the concentration of borate. Higher boron concentrations and lower pH values favour BF4- formation, but at F/B = 4 the complete (100%) conversion to BF4- is not available at any pH values and boron concentrations.
Figure 1 Equilibrium fraction of BF4- at different boron concentrations (0.1-10 mmol/L) with stoichiometric F- ion (F/B =4). The inset concentration indicates the initial H3BO3 concentration or the atomic concentration of boron.
Figure 2 Equilibrium fraction of BF4- at boron concentrations of (a) 1 mmol/L and (b) 10 mmol/L with different F/B ratios.
Figure 3 Equilibrium fraction of BF4- for the reaction of 0.1-10 mmol/L borate and 100 mmol/L fluoride. Thermodynamically, an excess amount of fluoride (F-) should be added to aqueous boron in order to achieve the complete conversion of borate to BF4-. The effect of excessive F- on the conversion to BF4- is illustrated in Figure 2, which shows a molar fraction of BF4- at the boron concentrations of 1 and 10 mmol/L with different F/B ratios. At both concentrations, upon adding 100 mmol/L fluoride, almost full conversions (over 99%) can be obtained at below pH 3. Figure 3 summarize equilibrium fraction of BF4- for the reaction of 0.1-10 mmol/L borate and 100 mmol/L fluoride. Consequently, 0.1-10 mmol/L borate with an excess amount of fluoride (100 mmol/L) can be completely converted to BF4- at below pH 3. EXPERIMENTAL 18, 19, 20 Rapid determination in potentiometric measurement. When an excess amount of F- is added to boron in acidic solution, the full conversion to BF4- requires roughly 1 h at ambient temperature. Figure 4 (a) shows an example of the reaction profile, in which 10 mg/L (as boron) of H3BO3 is converted to BF4- at pH 1. To accelerate the reaction, higher temperatures (333-353 K) and low pH are usually employed to accelerate the reaction.10, 15, 16 Under the conditions that 100 mmol/L of NaF and 0.1-10 mmol/L of H3BO3 are present in solution, thermodynamic calculation indicates that the equilibrium conversion of H3BO3 to BF4is 99% at below pH 3. Thermodynamically, pH 3 is sufficient for the complete conversion, but kinetically, pH 1 is necessary for the rapid conversion of H3BO3. We devised a rapid determination method without the need to heat the sample.18, 19 By analyzing the reaction kinetics in the case of a large excess of F- over H3BO3 in acidic solution, we found that the reaction obeys the first-order kinetics with regard to H3BO3, where the BF4- concentration CB is given by (9) CB = CA0 (1 - exp (-kt)) + CB0, where initial concentration of H3BO3, CA0: initial concentration of BF4-, CB0: concentration of BF4- at arbitrary time, C B: k: rate constant. By applying equation (9) to a BF4- profile, we can obtain three unknown parameters, CA0, CB0 and k, from the curve fitting, where the sum of CA0 and CB0 is the final BF4- concentration that we must obtain. If we can precisely estimate the parameters from a BF4- profile in a short period, a reduction in the measurement time can be achieved without the need to heat the solution. When F- is added to a borate solution in large excess of the stoichiometric ratio for reaction (1) and the solution is acidified with H2SO4 at pH 1, the measurement of BF4- in the initial 10 to 15 min is sufficient for ± 10% estimation. Thus, the final BF4- concentration can be kinetically obtained from the initial BF4- profile (Figure 4 (b)), leading to a large reduction in the measurement time. The details of the determination method are described elsewhere.18, 19
Figure 4 BF4- profiles in the conversion of H3BO3 with NaF. NaF was added to a 1 mmol/L H3BO3 solution (as boron) acidified with H2SO4 at pH 1. Added NaF in solution amounts to be 100 mmol/L in solution after addition. (a) The complete conversion requires 0.5-1 h. (b) The final concentration of BF4- can be kinetically obtained on the basis of the profile in the initial 10 min.
Figure 5 Online boron monitor (dimensions: 500 mm width × 500 mm depth × 1500 mm height). (a) Inside view. (b) Close-up of the measurement cell. Instrumentation.20 On the basis of our method, we constructed the fully automated measuring equipment for aqueous boron, referred to as the online boron monitor. Its internal views are shown in Figure 5. The online boron monitor automates a series of operations required for the measurement: the sampling and dispensing of the sample, addition of chemicals, acquisition and processing of potentiometric data, rinsing out of the measurement cell, and calibration of the BF4- ISE. The monitor was equipped with a measurement cell made of acrylic resin with an internal volume of 50 mL, which is maintained at roughly 313 K by circulating heated water in a water jacket around the cell. The cell housed a BF4- ISE and a reference electrode, both of which were obtained from DKK-TOA Corporation. The BF4- ISE (EL7464L) is a liquid membrane electrode, which disperses tetrafluoroborate trioctylammonium as an ionophore, in the polyvinyl chloride (PVC) membrane. The ISE can measure the concentration of BF4- anion in the range of 0.1-1000 mg/L as boron, but in practice, 1-100 mg/L in the case of high matrix samples containing ions interfering with the electrode. The severely interfering ions with the BF4- ISE are perchloride (ClO4-), iodide (I-) and cyanide (CN-) ions, while nitrate (NO3-), sulfide (S2-) and bromide (Br-) ions weakly interfere with the BF4measurement.21 The sample is collected through a sampling loop, the volume of which is adjustable between 1 mL and 10 mL by changing its length. After a certain amount of sample is injected into the cell, 5 mL of 9 M H2SO4 solution is added, and then the solution is diluted with deionized water by a factor of 5 to 50 depending on the loop volume, which is controlled by a liquid-level gauge in the cell. The monitor waits for the solution temperature to become stable,
and then 10 mL of 0.5 M NaF solution is added to the solution, and the measurement starts. The absolute calibration is automatically carried out between two BF4- concentrations of one order difference, e.g., 50 mg/L and 500 mg/L in the case of a 50-fold dilution of the sample, in the same manner as the measurement. Thus, the online boron monitor can measure the concentration of boron compounds in the range of 1-500 mg/L, which covers the Japanese effluent standards, in less than 20 min without the need for any pretreatment. In this study, we dilute an effluent from power plants and measure a higher boron concentration in range of several hundreds mg/L. With and without the sample dilution, the online boron monitor measures the boron concentration in range of 1-10 mg/L. Thus, if interference by coexisting ions in the sample is negligible, it can also be applied to a municipal wastewater treatment process and a desalination process, in both of which boron should be controlled at around 1 mg/L. Field demonstration test.20 The monitor was installed in two full-scale coal-fired power plants (plant A with a 250 MW unit and plant B with a 700 MW unit) and the boron concentration in the FGD wastewater was monitored for 1 month and 4 months, respectively. The monitor automatically measured the effluent at 3-hour intervals and was calibrated daily. For the verification of the measurements, the FGD wastewater was periodically collected and analyzed with ICP-AES. In both power plants, the boron compounds in the effluent was treated in a downstream wastewater treatment facility and discharged to public water bodies with the boron concentration below the effluent standard. Plant A was operated with nonblended coal combustion (one kind of coal at a time), whereas plant B was operated with blended combustion, in which a mixture of different kinds of coal is burned at different compositions. RESULTS AND DISCUSSION 20 Plant A. The online boron monitor was successfully operated at plant A for one month. Figure 6 shows the boron behavior in the FGD effluent, as revealed by the monitor. The values measured by ICP-AES are also depicted as blue triangles in Figure 6 (a). The boron concentration varied significantly in the range of 40-210 mg/L. Probable factors governing the boron behavior in the FGD effluent, i.e., the boron content in coal and the power generation, are also shown in Figures 6 (b) and (c). Three kinds of coal with different boron contents were used during the test period. It is apparent that the kind of coal is a governing factor of the boron level in the effluent, but no proportional relationship was recognized between the boron concentration in the effluent and the boron content in coal. Two reasons are envisaged for this. The total amount of boron released during the combustion is governed by the boron content, but its partitioning would be governed by other factors, i.e., the properties of coal, such as calcium content or alkalinity.6, 7 If the partitioning of boron largely differs owing to the chemical properties of coal, the boron concentration would not be proportional to the boron content in coal. In addition, if the FGD operation differs greatly depending on the properties of coal, i.e., sulfur content, this also should affect the boron behavior. The power generation was constant at 250 MW in full load operation during the test period, except from the 13th to 15th days, in which the plant was shut down and then restarted (Figure 6 (c)). In response to this, a V-shaped profile was recognized in the boron concentration at around the 16th day. This would suggest the constant operation of the FGD unit even during the shutdown period. Thus, the boron concentration in the effluent reflects changes in the kind of coal and the power generation with a delay of 2-3 days. Such details of the boron behavior in the FGD effluent have never been revealed before, because the measurement of boron concentration at a power plant is usually performed a few times a week at most. The boron concentration showed a relatively sluggish response with regard to time probably due to the large buffer tank for the effluent. The daily calibration values suggest a change in the ISE performance. The performance is estimated from the response slope and the standard potential E0, which are shown in Figure 6 (d). The response slope is the difference in the electrode potential when measuring two concentrations of one order difference, and E0 is the potential of 1 mg/L BF4solution. The response slope increased linearly from -58 mV to -55 mV, whereas E0 decreased from 313 mV to 293 mV, both of which brought about a slight shift during the test period. Thus, no significant deterioration of the ISE membrane was observed. Figure 7 shows the values measured with the monitor and the ICP-AES. An excellent correlation (R2 = 0.987) was obtained between the two methods, which demonstrates the effectiveness of the online boron monitor as a process monitoring tool.
Figure 6 Boron concentration in FGD effluent at plant A. (a) Boron concentration in FGD effluent. (b) Boron content in coal. (c) Load of power generation. (d) Response slope and standard potential E0 for calibration of ISE.
Figure 7 Relationship between values measured with the ICP-AES and the online boron monitor (at plant A). Plant B. The boron profile over a period of 4 months at plant B is shown in Figure 8. The boron concentration varied drastically in the range of 100-500 mg/L, which roughly reflects the boron content in coal. For the initial 2.5 months, a good correlation was obtained between the monitor and ICP-AES measurements. Thereafter, large discrepancies were
observed for 2 weeks, which is depicted as A in Figure 8 (a). The values measured with the monitor were twofold those measured with the ICP-AES. During this period, we observed no significant change in the sample composition, i.e., no ionic species such as ClO4-, I- and CN-, which are seriously interfering with the BF4- electrode. 21 However, some foaming phenomena in the FGD unit were reported during this period. Thus we suspect that the discrepancies might be caused by bubbles becoming attached onto the ISE membrane. The air bubbles can come from the sample (effluent) and also from the utility water in sample dilution. Since we did not see any perturbation in the calibration (Figure 8 (d)), we suppose that the air bubbles would come from the effluent. We have experimentally confirmed that air bubbles on the membrane reduce the electrode potential, leading to higher boron concentrations. As a measure to the foaming, we modified the cell configuration by tilting the BF4- ISE that had been installed perpendicularly in the cell (Figure 5 (b)), so that air bubbles will have difficulty in attaching on the ISE membrane and also can be easily removed from there. This configuration has already been put into practice in the monitor that is installed at a different plant and has no problem so far. Additionally, the use of anti-foaming agents would be a solution for dissipating the foam. In this case, we must examine prior to its use that the agent results in no or little interference with the ISE. For the initial 3 months, the response slope of BF4- ISE increased linearly, but remained below -50 mV, as shown in Figure 8 (d), which is a criterion for the exchange of the ISE membrane. Then, the slope fluctuated greatly and exceeded -50 mV. A similar behavior was also observed in E0. This suggests that the life of the ISE membrane is 3 months under the test conditions. The fluctuation of values measured with the monitor during period B, shown in Figure 8 (a), is affected by the life of the ISE membrane.
Figure 8 Boron concentration in FGD effluent at plant B. (a) Boron concentration in FGD effluent. (b) Boron content in coal. (c) Load of power generation. (d) Response slope and standard potential E0 for calibration of ISE.
CONCLUSIONS The online boron monitor, a fully automated measurement device, was developed on the basis of the rapid potentiometric determination using a BF4- ISE, which measures boron compounds in the range of 1-500 mg/L in less than 20 min without the need for any pretreatment of the sample. In the demonstration tests at two power plants, the online boron monitor was successfully operated, and the detailed behavior of boron in the FGD effluent, which reflects the kind of coal and the power generation, was revealed. The life of the ISE membrane was estimated to be 3 months from the test results. The online boron monitor was proved to be a useful tool in managing the boron emission in process effluent. REFERENCES (1) World Health Organization, Guidelines for Drinking-Water Quality, 3rd Ed.; Vol. 1. Recomendations; Geneva, 2004. (2) http://www.env.go.jp/en/water/wq/wp.pdf (3) http://www.env.go.jp/en/water/wq/nes.html (4) Clarke, L. B.; Sloss, L. L., Trace elements – emissions from coal combustion and gasification. IEACR/49; IEA Coal Research; London; 1992. (5) Ito, S.; Yokoyama, T.; Asakura, K. Emissions of mercury and other trace elements from coal-fired power plants in Japan. Science Tot. Env., 2006, 368, 397-402. (6) Clemens, A. H.; Damiano, L. F.; Gong, D.; Matheson, T. W. Partitioning behaviour of some toxic volatile elements during stoker and fluidized bed combustion of alkaline sub-bituminous coal. Fuel, 1999, 78, 1379-1385. (7) Clemens, A. H.; Deely, J. M.; Gong, D.; Moore, T. A.; Shearer, J. C. Partitioning behaviour of some toxic trace elements during coal combustion—the influence of events occurring during the deposition stage. Fuel, 2000, 79, 1781-1784. (8) Iwashita, A.; Sakaguchi, Y.; Nakajima, T.; Takanashi, H.; Ohki, A.; Kambara, S. Leaching characteristics of boron and selenium for various coal fly ashes. Fuel, 2005, 84, 479-485. (9) Noda, N.; Ito, S. The release and behavior of mercury, selenium, and boron in coal combustion. Powder Tech., 2008, 180, 227-231. (10) Carlson, R. M.; Paul, J. L. Potentiometric determination of boron as tetrafluoroborate. Anal. Chem., 1968, 40, 1292-1295. (11) Gulens, J.; Leeson, P. K. Direct ion-selective electrode determination of micromolar boron as tetrafluoroborate. Anal. Chem., 1980, 52, 2235-2237. (12) Lanza, P.; Mortera G. Potentiometric determination of boron in silicon with an ion-selective electrode. Anal. Chim. Acta, 1975, 75, 149-384. (13) Lanza, P.; Mortera G. The determination of boron in waste waters with an ion-selective electrode. Annali di Chimica, 1983, 73, 371-384. (14) Johnson, M. L.; Ward, W. G. Analysis of boric acid by tetrafluoroborate specific ion electrode. Metal Finishing, 1988, 86, 49-51. (15) Imato, T.; Yoshizuka, T.; Ishibashi, N. Potentiometric flow-injection determination of boron by using a flowthrough tetrafluoroborate ion-selective poly(vinyl chloride) membrane electrode. Anal. Chim. Acta, 1990, 233, 139-141. (16) Imato, T.; Yoshizuka, T.; Ishibashi, N. Flow-injection determination of boron by using poly(vinyl chloride) membrane-based tetrafluoroborate ion-selective electrode. Bunseki Kagaku, 1993, 42, 91-98 (in Japanese). (17) Katagiri, J.; Yoshioka, T.; Mizoguchi, T. Basic study on the determination of total boron by conversion to tetrafluoroborate ion (BF 4-) followed by ion chromatography, Anal. Chim. Acta, 2006, 570, 65-72. (18) Ohsumi, H.; Ohyama, S.; Kudo, S.; Sakata, M. A simple and rapid determination method of boron in wastewater using ion-selective electrode. J. Environ. Chem., 2004, 14, 81-89 (in Japanese). (19) Ohsumi, H.; Ohyama, S.; Kudo, S.; Sakata, M. Japanese Patent, 2008, 4210146. (20) Ohyama, S.; Abe, K.; Ohsumi, H.; Kobayashi, H.; Miyazaki, N.; Miyadera, K.; Akasaka, K.; Fully automated Measuring equipment for aqueous boron and its application to online monitoring of industrial process effluents, Environ. Sci. Technol., 2009, 43, 4119-4123. (21) Instruction Manual, Tetrafluoroborate Ion Electrode, ION-LB08102, DKK-TOA Corp., Tokyo, 2007 (in Japanese).
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COMBUSTION- FLUE GAS CLEAN UP
MERCURY SORPTION ON BROMINATED ACTIVATED CARBON Erdem Sasmaz PhD Candidate, Department of Energy Resources Engineering, Stanford University, Stanford, CA 94305 Tel: 650-725-0752, Fax: 650-725-2099, [email protected] Jennifer Wilcox Assistant Professor, Department of Energy Resources Engineering, Stanford University, Stanford, CA 94305 Tel: 650-724-9449, Fax: 650-725-2099, [email protected]
EPA estimated that approximately 75 tons of mercury were found in the coal delivered to power plants each year and about two thirds of this mercury was emitted to the air, resulting in about 50 tons being emitted annually. This 25-ton reduction was achieved through existing pollution controls such as fabric filters (for particulate matter), scrubbers (for SO2) and selective catalytic reduction units (for NOx). To predict the levels of mercury emissions from coal-fired power plants and determine the best applicable control technologies, it is important to understand heterogeneous mercury reaction mechanisms on activated carbon surfaces. Experimental studies are conducted using a bench-scale packed-bed system to further understand heterogeneous mercury reaction mechanisms. Methane is combusted in a tubular burner and the effect of combustion products on the sorption of mercury are investigated. Additionally, other flue gas species will be introduced in a controlled manner to the reactor system to understand their effect on mercury adsorption. Activated carbon-based sorbent materials before and after simulated flue gas exposure are analyzed using X-ray photoelectron spectroscopy (XPS). Characterization experiments are conducted to determine the chemical bonds, surface coverage and speciation of mercury. Our XPS results for mercury sorption tests on brominated powder activated carbon surfaces conducted in an air environment suggest that mercury is adsorbed on carbon surfaces in the form of oxidized mercury; however, the identification of the oxidation state of mercury on our samples was limited due to presence of silicon on the surfaces and its subsequent interference with mercury spectral lines.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
COMBUSTION
EFFECT OF COAL VOLATILE MATTER ON EMISSIONS OF BOILER COMBUSTION Hasancan Okutan, Prof.Dr. İstanbul Technical University, Chemical-Metallurgical Faculty, Chemical Engineering Department 34469, Maslak-İstanbul - TURKEY Nalan Erdöl Aydın, Dr. İstanbul Technical University, Chemical-Metallurgical Faculty, Chemical Engineering Department 34469, Maslak-İstanbul – TURKEY Erhan Böke, Assist.Prof. Dr. İstanbul Technical University, Mechanical Faculty, Mechanical Engineering Department 34437, Gümüşsuyu-İstanbul - TURKEY
Abstract : Turkey’ s most important bituminous coal deposits are lacated especially in Zonguldak basin on the shares of the Black Sea in the Northen Turkey. Coal obtained from Zonguldak basin by means of underground mining is transferred to various washery plants in districts such as Zonguldak, Çatalağzı, Armutçuk and Amasra, where its sulfur amd mineral matter contents are reduced and it is classified as coking, semicoking and noncoking coal. Coking coal is used as feed stock in the coke production, while noncoking coal is consumed as industrial and domestic fuel throught the country. Bituminous coals from Zonguldak basin hava a considerable amount of volatile matter contents, some of which are easy to desorp from the coal surface during heating and they leave the coal structure without burning. In this study, effect of coal volatile matter on the combustion efficiency and CO, SO2, NOx and particulate emissions of the bituminous coals and lignite from Turkey and Sibirya are investigated in a 174 kW, half cylindirical hot water boiler. Experiments were carried out for 27,0% volatile matter Zonguldak washery, 27,5% volatile matter Çatalağzı washery, 34,0% volatile matter Armutçuk washery, 35% volatile matter Amasra washery, 0,8% volatile matter Zonguldak coke, 38% volatile matter Soma and 18,9% volatile matter Sibirya coals. Combustion tests were performed according to Turkish Standarts (TS) 4040 and 4041, which are equivalent to recpective ISO standarts. Individual combustion efficiency, particulate and gas emissions profiles and cumulative emissions are determined and the experimental results are discussed.
THE IMPROVEMENT OF UCG PROCESSES Karol KOSTÚR Faculty of Mining, Ecology, Process Control and Geotechnology Technical University of Košice, Košice, Slovak Republic, [email protected] Abstract: The idea of the transformation of coal in underground into synthetic gas so called syngas has interesting in world more centuries. UCG (Underground Coal Gasification) is an in situ technique to recover the fuel or feedstock value of coal that is not economically available through conventional recovery technologies. Today, less than one sixth of the world’s coal is economically accessible. Today, similarly to all other countries in the world also in Slovakia there is an interest in the revival and perfection of the UCG technology. From the viewpoint of content the research is directed toward to increasing heating capacity of syngas. From the standpoint of the methods used one can characterized the research by 2 approaches: experiments in UCG laboratory and mathematical modeling, including of simulation studies. Both approaches have helped to discover complicated relationships during UCG and they will be subject this lecture. Most important are topologies/methods, humidity of coal, heat losses, temperatures in relevant zones, composition of oxidation agents and permeability. Calorific value of syngas is 0.55 – 4.45 MJ.Nm-3 on average with maximum 25.51 MJ.m-3 if oxidation agent was used air only. In case using mixture air + oxygen have been obtained calorific value from range 0.43- 6.38 MJ.m-3 on average with maximum 27.53 MJ.m-3. The analysis has been done for these big differences by aim to improve UCG. Therefore in paper is described: - Structure of thermodynamics model including simulation cases, - Experimental plant, including measured data, - Relevant results (the influence of reaction’s area). Key words: mathematical models, UCG laboratory, analysis important parameters on calorific value of syngas 1. INTRODUCTION In present time, the solving of Global climate belongs to the most important activities in world. On this place, we must to emphasis the fact, that UCG belongs to so called clean technologies. From professional standpoint this problem is solved more ways: reducing consumption of energy, the renewable sources of energy (a wind, water, and the sun), the improving of combustion of fusil fuels, the increasing of efficiency electrical and thermal equipment, etc. Of course, there are other possibilities for example the utilizing nuclear energy. In general are well-known problems with nuclear power stations and with nuclear waste. Probably, renewable sources of energy will not belong to main group of energetically sources in short future. Coal is our most abundant fossil fuel. While oil and natural gas account for 64 percent of the world’s energy consumption, they total only 31 percent of the world’s known fossil fuel reserves. Therefore, the coal appears as considerable source for energetistic and chemical industry in future. But today in Europe a coal have to be extracted from a bigger depth. It increases costs to coal extraction. It main reason of the reducing
1
extraction coal in Europe on app.2,5 % from world coal production. The reason of high costs is in a great consumption of specific energy on unit exploited coal. UCG is an in situ technique to recover the fuel or feedstock value of coal that is not economically available through conventional recovery technologies. Today, less than one sixth of the world’s coal is economically accessible. In present time most coal is extracted by classical mining technology. In generally are well known disadvantages classical mining (a big cost, a danger work, problems by ecology,…). First tentative works directed toward utilization of the in situ method of gasification under conditions of Northern Bohemian brown-coal basin [1] date back to 1956 in formerly Czechoslovakia. Six experiments of UCG in situ were realized late during sixty years of last century. The obtained results did not confirm the technological assumptions of gasification especially regarding the quality of the gas recovered. The average heating capacity reached the value from 1,59 – 3,44 MJ.m-3 . Today, similarly to all other countries in the world also in Slovakia there is interest in the revival and perfection of the UCG technology. The mainly advantage is fact that UCG can use otherwise inaccessible coals, can produce syngas for power and for fuels (i.e., liquid fuels, synthetically natural gas, or hydrogen), various chemical products and has attractive economics. Based on published and new cost estimates, engineering analyses, and new commercial pilots it appears that UCG can produce syngas for ½ to ¼ of the cost compared to surfaces gasifies [2]. Importantly, UCG may have special promise in combination with carbon capture and sequestration. The U.S. government has declared “clean technology” a critical coal for the near term, and the state of California and other government entities have mandate the reduction of CO2 emissions [3]. Field test in Western Europe demonstrated the technical feasibility of underground coal gasification at the great depth with different geological conditions (deposition seam, tectonics) and different coal type. U.S.A joined on European program of gasification and they targeted to drill technique, research gasification cavity, modeling process and environmental implication this technology. There has been UCG activity in China carried out in abandoned mines to recover some of the coal/energy left behind. From the viewpoint of content the research is directed toward to increasing heating capacity of syngas. In the period of 2006 – 2010, the research was focused on improving complex mutually influenced processes of the underground coal gasification [4] and on their optimal control. The basic research methods have used mathematical modeling of processes and experiments in laboratory conditions. Both approaches have helped to discover complicated relationships during UCG and they will be subject this lecture. Most important are topologies/methods, humidity of coal, heat losses, temperatures in relevant zones, composition of oxidation agents and permeability. Currently, with the growing global energy consumption is UCG technology in competitive fight with alternatives, especially renewable resources, which return is also relatively long. Therefore, the aim of our research was to find out which main factors during the UCG process would enable to streamline, improve this technology. Relevant results views of point calorific value were obtained from following the analysis of UCG for lignite: -the influence of temperature, -the influence of ratio oxidizing agents, -the influence of coal moisture, -the influence of velocity streaming gases, -the influence of reaction’s area. The composition of one lignite sample is shown in tab 1.
2
Table 1 The analysis of coal Moisture Wt(r) Volatile V(r) Ash A(r) Heat value Qid (r)
Proximate analysis (%) 38.2 28.0 5.81 16.0 MJ.kg-1
Elementary analysis (%) C (daf) 42,80 H (daf) 2.21 N (daf) 0.83
2. MATHEMATICAL MODELING There were numerical simulations performed before every experiment for particular coal composition by thermodynamic and dynamic thermal model. Thermodynamic model Simulations by the thermodynamic model were based on a solution of the following optimization task [6]: To minimize Gibson overall energy system in equilibrium state
G = ∑ nj µ j N
j =1
µ
j
= µo j
(1) + RT ln
n
j
N
∑n j =1
j
p . po
(2)
In keeping with the established system of restrictions based on balance equations for the relevant chemical reactions in the underground gasification An=b
(3)
Where µ j - is a standard Gibson energy [kJ], R – is an universal gas constant, T – is a
thermodynamic temperature [K], p – is a pressure in a gasifying system [Pa], p0 – is a standard pressure [Pa], N is the total number of mol in system, b is a vector of an input number of gram atoms in a gasifying system, nj – is a number of gram atoms of the component j-th in a in equilibrium state of the mentioned system, A – is a coefficients´ matrix of individual elements (C,O,…) in system’s components in a steady state.
3
The outputs of these simulations were dependences of the arisen gas composition on temperature and theoretical amount of oxidative media needed for the coal gasification-see Fig.1. The qualitative conclusions from these simulations can be formulated as follows. At temperatures above 500 ° C with increasing temperature the percentage of CO in syngas significantly increases and at temperatures above 1200 ° C its stabilization is reached. On the contrary methane, which has the highest calorific value, reaches maximal concentrations at lower temperatures (400 - 500 ) oC. Hydrogen , as well as CO, with increasing temperature increases its own concentration from temperature starting at 250 oC and at temperature 850 oC it reaches maximal values (18,6 %). Its concentration begins to decrease slightly with further raise of temperature. The actual temperature during gasification in this system will be set by the energetic balance including heat loss, but the thermodynamic model does not handle this. Similarly, it does not reflect the kinetics of processes (heterogeneous reactions are significantly slower than the homogeneous reactions).
Fig. 1 Syngas composition dependence on temperature obtained by using the thermodynamic model simulations
4
Dynamic thermal model A mathematical model was created in order to simulate the process of coal gasification, which allowed studying the impact of a cavity shape, the size of a reaction surface, the heat loss, a moisture of the coal on the thermal fields, syngas calorific value. The simulation model was created by elementary balances method. There is a coal layer in the model divided into elements through which the gas medium gets through. The processes in the element take place in the mutual interaction of the material and gas medium (Fig.2).
Elements
E1
E2
E3
E4
E5
E6
EN
Stream of gasses Interactions Coal
Fig.2 Dividing of coal gasification unit to the elements The structure of a simulation model is based on solving the basic energy and balance equations (4) and (5). A thermal balance for element during the time step ∆τ : Qmaterial In + Qoxidation gas In + Qreactionheat + Qconde n = Q materialOut + Q gas Out + Qpyrolisis + Qgasify + Qloss + Qevaporation
[J]
(4)
where Qmaterial In -a thermal energy of input coal into gasification, Q materialOut - a thermal energy of coal after ∆τ time of gasification, Q reactionheat - a thermal energy released by oxidative reactions with a carbon, Qconde n – a thermal energy released by a condensation of a vapor, Q oxidation gas In - a thermal energy carried to element by a flowing gas oxidizing medium, Q pyrolisis – a thermal energy utilized on pyrolysis, Q gasify – a thermal energy utilized on reduction reactions, Qloss – a thermal energy conducted away from the gasified element into the surrounding ( heat losses), Qevaporation – a thermal energy utilized on an evaporation. Mass balance for the element during the time step ∆τ :
5
mmat_in + mgas_in = mmat_out + mgas_out
[ kg ]
(5)
where mmat_in is a weight of an initial coal in element, mgas_in is a weight of gas medium on an element´s input, mmat_out is a weight of coal in element after ∆τ, mgas_out is a weight of a gas medium on an element´s output for∆ τ time. The simulations There were made simulations of chosen ways of gasification by a mathematical model. The following methods of coal transformation and flowing gas media were simulated: • piston gasification (alternative 1) •
gasification in cavity (alternative 2)
•
reverse gasification in cavity (alternative 3).
The three shapes of moving gasification front and relevant UCG processes by individual colors are shown on Fig. 3. Piston gasification Oxidizing zone Gasification Distillation Evaporation Air/Oxygen Syngas
Gasification in cavity
Reverse gasification in cavity
Fig.3 A scheme to simulate the three means of gasification queue/line in a coal movement
6
Piston direction is based on the analogy of piston internal - combustion engines and represents the same movement of gasification front along the thickness of the gasified coal bed. This mean of gasification front movement can be regarded as ideal, but in real conditions hardly achieved. The second method, so - called "Gasification in cavity" is by its shape closest to the real conditions of underground coal gasification. The third mentioned method is characterized by an inversion to the second case and its weak point is a partially burned syngas, particularly its volatile and pyrolysis utility components. In practice, this method may occur in the reverse flow, when it is necessary to restore the process of UCG. At the same time in examining the impact of the cavity on the syngas calorific value, we studied also the impact of coal moisture. There is the process of a theoretical syngas calorific value by the shape of the cavity depending on the coal moisture illustrated in the Fig.4. The Alternative No. 1 represents a piston gasification, the Alternative No. 2 reflects the gasification in cavern and the Alternative No. 3 illustrates the calorific value process in the reverse cavity. 8
3
Calorific value[MJ/m ]
6
4
Alternative 1 Alternative 2 Alternative 3
2
0 0
5
10
15
20
Humidity[%] 25
30
Fig.4 The impact of the gasification shape front and moisture on syngas calorific value 3. EXPERIMENTAL RESEARCH UCG processes (physical and chemical) are very complicated-see Fig.3. There are more unknown geological parameters. A big problems is fact that we "we do not see" the processes running under the ground and the possibility of measuring the temperature, continuous analysis of gases are impossible in practice. The improving of UCG requires the studying these processes not only by help simulation cases but there are necessary verified knowledge from an experimental research. The experimental research is very expensive in situ and it does not enable to achieve required knowledge. Therefore, we have chosen experimental research in laboratory conditions.
7
Coal seam
Model of UCG
Over burden
..................................................................... .................isolation....................
q
q
Coal
Over burden
Coal q Under burden
Under burden
........................................................ .............isolation.............................. .......
q
a - coal seam
b – coal model
Fig.5 The application of similarity theory for coal model The substratum, coal and overburden in the unaltered form are put into described generator. Maintaining the original characteristics of selected samples from the underground is ensured by their protection against contact with air during transport to the laboratory. For the specific purposes, a coal experimental model depended on the chosen topology, respectively gasification method is created. Experimental model [4] is built on the basis of respecting the fundamental knowledge of the similarity theory. The principle of the application of similarity theory for experimental coal model is shown on Fig.5. In particular, it is a geometric similarity of coal model to potential coal seam for gasification. We can express it by the criteria (6) and (7)
cons tan t =
1=
xmod el
(6)
xcoalseam
β mod el
(7)
β coalseam
where x are the dimensions (length, width) and β is the relevant slope of a coal bed with a horizontal plane. Coal is cut into blocks before insertion into the generator, they are stuck together to adhere similarity criteria (6) and (7). However, it is impossible to apply geometric criteria for the thickness of the coal bed in the laboratory because of practical reasons. Therefore, instead of this criterion, we apply criterion to test equality of heat flows in coal-rock interface (8) in the model and real experiments in situ. q = idem
(8)
This criterion has been realized by help mineral isolation which is inserted between coal (overburden, underburden) and internal generator surrounding. The isolation thickness is
8
computed by solving of Fourier’s system partial differential equations for first type of boundary conditions. Last criterion of equivalence between coal seam properties and coal model is realized by storing of unaltered coal in form blocks in direction underground historical created layers. 3.1 Experimental devices and analysis of experiments The experimental devices built in 2007 - 2008, which currently represent the only laboratory of underground coal gasification in Slovakia, were used to study following impacts on the efficiency of underground gasification: - methods, the topography of preparing the coal bed for gasification to achieve maximal calorific value, - the impact of temperature, - the impact of the oxidizing gas mixture composition, - the impact of coal moisture, - the impact of the gas flow velocity in gasification channel, - the impact of the reaction area, - the control and controlling systems for UCG process. Our ambition was to create an experimental facility to allow studying the underground gasification processes experimentally by various methods. The base of this plant is a steel generator (G) - see Fig. 6. The tilting generator enables to simulate an inclination of coal seam. The compressor (K) forces air into a pressure vessel (TN). The pressure in vessel is controlled by programming logic controller (PLC) through solenoid (SV). The air can be concentrated by oxygen, water steam or carbon dioxide by connection on input tube. The air or gas mixture is blown into generator by valves and tubes. This system of valves (Vi) and tubes will enable to simulate various arrangements of inlet holes and outlet holes. The pressure flow direction of gas can be regulated in generator by valves and flap valves. But this possibility is very limited. The syngas is possible exhausted by fan and it is combusted by a burner. Experimental plant has others measurement equipments, e.g. the flue gas analyzer, flow meters, the calorimeter, etc. Sampling data are processing by monitoring and control system [7]. The coal samplings were stored in generator G in original state in form various coal models. Later, UCG laboratory was spread about further equipments e.g. special UCG generator G2, compresors, mixing station for preparing oxidation mixture, etc.
9
Fig. 6 The scheme of experimental plant.
3.2 Methods, the topography of preparing the coal bed for gasification to achieve maximal calorific value In this analysis and on the basis of the results of simulations and experimental results from the previous period, we verified the various methods of preparing the coal bed. We carried out this by construction of the following so-called coal models from two various coal seams (Novaky, Cigel). Date of experiment
Generator
Name of coal model
1/2009 (February)
G1
The system composed of
blocks of coal with central rectangular cross-section of gasification channel-Novaky
2/2009 (May)
G2
Monolithic system of coal
with drilled diameter 20mm of gasification channel-Novaky
3/2009 (June)
G2
Monolithic system of coal
with drilled diameter 40mm of gasification channel-Novaky
4/2009 (September)
G1
Multi – channel system of
coal gasification – Cigeľ
10
1/2010 (February)
G1
Multi – channel system of
G2
Multi – channel system of
G1
Coal model with split coal
coal gasification – Nováky 2/2010 (June) coal gasification – Cigeľ 3/2010 (June / July)
Fig.7a Monolithic system of coal with
Fig.7b Multi – channel system of
drilled diameter 20mm of gasification
coal gasification (4/2009)
channel (2/2009)
For illustration, monolithic model with drilled gasification channel is shown on Fig. 7a and principle of multichannel system is shown on Fig. 7b. Multichannel system was constructed from exactly cutting coal samplings for purpose of measurement and following evaluation reaction area, velocity of streaming gases, etc. The cut of coal models is shown after end of gasification on Fig.8. UCG generator was filled by nitrogen in moment of stop UCG for purpose of evaluation coal models (the mass of ash, the rest of coal) after gasification. We can see the rests of coal (black pieces) on coal model with drilled gasification channel see Fig.8. It means a little small effectiveness of transformation coal to syngas in case of monolithic system of coal with drilled gasification channel. An arithmetic weighted means, from the measured calorific values and standard deviations of SYNGAS for the mentioned coal models, are shown in Table 12.1. Characteristics are given for both criteria, that means, for the calorific value”H” higher than 3 MJ/m3 as well as for the calorific value during the whole experiment (H> 0). Except for this criteria, in the table, there
11
a
b
Fig.8 Coal models after gasification (8a Monolithic system of coal with gasification channel, 8b Multi – channel system) are listed statistical characteristics of calorific value in terms of the input type of oxidizer. Calculated characteristics are for two types of oxidizer. The first one is only an air without the addition of technical oxygen and the second one is a mixture of an air and oxygen. The mixture of an air and oxygen means also the case when oxygen was not added to the mixture (it was equal zero), which was actually during the whole experiment. The highest average calorific value was obtained during the 1 / 2009 experiment, which value, during the whole experiment, was 6.38 MJ / m3. On the contrary, the lowest average calorific value was during the 3 / 2009 experiment. Table 2 Statistical characteristics of the SYNGAS calorific value exper.
oxidant
H > 3 MJ/m3
H > 0 MJ/m3
average standard deviation MAX
MIN
average standard deviation
MAX MIN
only air 1/2009 mixture air+O2
6.75
2.59
14.53 3.00
4.45
3.28
14.53 0.00
8.12
3.11
14.59 3.00
6.38
3.93
14.59 0.00
only air 2/2009 mixture air+O2
5.43
3.82
19.86 3.01
1.64
1.51
19.86 0.13
5.27
3.23
19.86 3.01
1.28
1.31
19.86 0.00
only air 3/2009 mixture air+O2 only air 4/2009 mixture air+O2
3.63
0.94
8.01
3.00
0.55
0.84
8.01 0.00
3.68 5.45
0.73 2.53
8.01 3.00 25.51 3.00
0.43 3.94
0.77 2.69
8.01 0.00 25.51 0.00
6.11
2.81
27.53 3.00
4.58
3.17
27.53 0.00
only air 1/2010 mixture air+O2
4.27
1.61
10.14 3.00
1.02
1.18
10.14 0.00
4.46
1.80
12.31 3.00
1.04
1.24
12.31 0.00
only air 2/2010 mixture air+O2
5.78
2.86
24.56 3.00
3.21
2.90
24.56 0.00
5.83
2.75
24.56 3.00
3.32
2.89
24.56 0.00
12
3.3 The impact of a temperature The impact of temperature is clear depending on desired aims of a gasification product. If the aim of the gasification is a maximal concentration CO/H2 in syngas, then it will guarantee the achievement of maximal temperatures in the oxidation and reduction zone. This was confirmed by simulations and experiments at temperatures over 1200oC. If the aim is maximal calorific value / maximal concentration of CH 4, then this can be achieved at temperatures between 800 to 1000 oC. 3.4 The analysis of the impact of oxidizer proportion on SYNGAS quality The oxidizer proportion on SYNGAS quality was analyzed in the experiment 1/2009. Gradually, was the proportion of the volume oxygen flow to the air flow – VO2/Vvzd increased from 10% to 70% by ten percent during the experiment. There is the process of SYNGAS calorific value and the proportion of VO2 /Vvzd shown in Fig.9. VO2 is the volume flow of oxygen and Vvzd is the volume flow of air. Calorific value is the highest at proportion of 75% on average which means 14 MJ/m3. 100 14
90 80 70
10
60 8
50 40
6
ratio VO2/Vvzd
3
calorific value [MJ/m ]
12
30
4
20 2
10
24.02.2009_10:48:00
24.02.2009_06:15:00
24.02.2009_01:42:00
23.02.2009_21:09:00
23.02.2009_16:36:00
23.02.2009_12:03:00
23.02.2009_07:30:01
23.02.2009_02:57:01
22.02.2009_22:24:00
22.02.2009_17:51:00
22.02.2009_13:18:00
22.02.2009_08:45:00
22.02.2009_04:12:00
21.02.2009_23:39:00
21.02.2009_19:06:00
21.02.2009_14:33:01
21.02.2009_09:59:00
21.02.2009_05:26:00
21.02.2009_00:53:00
20.02.2009_20:20:00
20.02.2009_15:47:00
20.02.2009_11:14:00
20.02.2009_06:38:00
20.02.2009_02:05:00
19.02.2009_21:32:00
19.02.2009_16:59:00
19.02.2009_12:26:00
19.02.2009_07:53:00
19.02.2009_03:20:00
18.02.2009_22:33:00
18.02.2009_18:00:00
18.02.2009_13:27:00
18.02.2009_08:54:00
18.02.2009_04:21:00
17.02.2009_23:48:00
17.02.2009_19:15:00
17.02.2009_14:42:00
17.02.2009_10:09:00
0
17.02.2009_05:34:00
0
Date and time calorific value
ratio VO2/Vvzd
Fig. 9 The course of SYNGAS calorific value and the proportion of input oxidizer- an experiment 1/2009 3.5 The impact of coal moisture The impact of moisture from the simulation has already been shown in Fig. 4. At the piston flow, the impact is clear. By other means, respectively shapes of gasification line to 8 resp. 12 percent of water content in coal, is its effect on the calorific value positive. We can explain it by the reactions of a steam with coal, where the final product of these reactions, determining the syngas calorific value, is methane. At these lower concentrations of moisture, this positive effect prevails over the negative and that is a higher consumption of a thermal energy in
13
nearly the oxidation and reduction zone due to the transformation of water into vapor, resulting in decrease of temperature in this two relevant areas. At the moisture over 12 percent, there prevails a negative moisture effect in a coal over calorific value in all three shapes of gasification front. As the moisture of studied lignite ranged from 23% to 40%, a negative factor of increased coal moisture on syngas calorific value clearly prevailed. This fact was also confirmed by an experimental research. There is an average measured syngas calorific value for two different coal moistures compared in the table 3. We would like to emphasize that this experiments are comparable, also in terms of channels´ types (multichannel system) and the method of gasification using air only in the role of oxidizer. Table 3 The comparison of an average calorific value for two different coal moistures No. of experiment Calorific value of syngas [MJ.m-3] Moisture of coal [%] 1/2010 2/2010
1,02 3,21
38,3 24,3
3.6 The impact of the gas flow velocity in gasification channel The aim was to determine an interval of an optimal velocity of flowing gas media in gasifying channels. It was impossible to measure the flowing velocity in the oxidation zone directly because of extreme conditions (aggressive environment, high temperatures, the changing shape of the cavity, the change of a free cross-section of the probe body). Therefore, the following methodology was chosen. The velocity was calculated from the oxidizer volume flow, measured before entering the generator and the free cross - section channel, through which the gases flew. We evaluated the so-called limited syngas calorific value on flow velocity, only if this condition (9) was fulfilled: 800 oC ≤ Toxid.zone ≤ 1100oC AND % O2syngas ≤ 0.1 % (9) The process of evaluated dependence for 4/2009 experiment is shown in Figure 10. 30
Calorific value of syngas MJ/m3
25
20
15
10
5
0 0.000
0.200
0.400
0.600
0.800
1.000
1.200
1.400
1.600
1.800
Velocity in channel [m/s]
Fig. 10 A limited calorific value dependence on the flow of the oxidizing media without oxygen
14
2.000
In the following table, according to described method, there are processed results of the gasification with oxygen and with the mixture of an air and oxygen for the first four experiments. The pressure of oxidation agent was app. 4 kPa. Table 4 Parameters of gas velocity and Reynolds’s criterion for calorific value with fulfilling conditions (9) Experiment Air speed without Values of Re Air speed with Values of Re O2
(m/s)
criteria
O2 (m/s)
criteria
1/2009 G 1
1.5 – 2.0
400 - 500
3-6
1000-1800
2/2009 G2
16
5000
0.2-2.0
50 - 800
3/2009 G2
0.5 – 1.0
500 - 1000
0.2
85
4/2009 G1
0.18 –1.05 (0.5)
50 – 240 (120) 0.02 – 0.5 (1,2)
20 – 288 (120)
Considering the density of points in the diagrams of the graph type 10, from all experiments we consider the optimal range of "air" flow speed in the simply gasification channel from 1.5 - 2m.s-1 (the highest calorific value of syngas). Optimal air speed for more times increasing reaction area (30 X-multichannel system) is 0.5 m.s-1 on average, exactly the optimal range from 0.18 – 1.05 m.s-1. It is lower in comparing with simply gasification channel and calorific value has been from 3 – 20 MJ.m-3 (4/2009).With an increasing oxygen content in the oxidation mixture we can regard as the optimal the lower order values. 3.7 The impact of reaction area The significant factor for the gasification is also the coal moisture (experiments No. 1/2009, 2/2009, 3/2009, 4/2009, 1/2010), but it was necessary to answer the question, why experiment 1/2009 with moist coal proved excellent calorific value. It is clear from the analysis done, that an important criteria for a quality of UCG is a reaction area, on which run the chemical reactions, pyrolysis and physical processes such as evaporation, heat energy transfers, because these processes limit the UCG on the boundary area of an oxidative media and a coal. The processes inside the coal massif are set by diffusion and this is characterized in the gas-solid environment by very low speeds. Processes, on the limit contact of a gas and solid phase - reaction area, are determining. However, the reaction area does not reflect a density of the gas flows through the unit weight of the coal gasification model. Therefore, we introduced the term of a specific reaction area, which, set by the following relation, more accurately characterizes the suitability of the reaction area geometry - the conditions for the course of basic processes in the UCG.
Ra =
O.1 Gc 15
Where O- a circuit of a free cross - section, through which gas flows [m], l is the length unit [m], Gc is the weight of a gasified coal [kg], Ra is the specific reaction area of the coal model before the gasification [m2 /m.kg]. In the Table 5, there are compared, at almost the same coal moisture, syngas calorific values at the gasification with air only. Both coal models varied by size of the specific reaction area (the simple gasification channel, multi - channel). Tab.5 The comparison of a specific reaction area No. of experiment 3/2009 4/2009
Specific reaction area 2,4.10-4 7.10-3
Calorific value of syngas [MJ.m-3 ] 0,55 3,94
Moisture of coal [%] 39,01 38,30
By comparing the values of a specific reaction area in the above – mentioned tables, we come to the conclusion that 8-times higher average calorific value was reached, if the reaction area was approximately 30 - times higher. The discovery of this parameter (specific reaction area) is significant to regulate the development of methods for accessing the coal field for the purposes of the underground coal gasification. We demonstrated this knowledge and clearly proved it by a mathematical modeling and set of experiments. 3.8 The control and controlling systems for the process of the underground coal gasification Because the process of underground gasification is slow, long-term, it is inevitable to automatically process the relevant quantities, which was ensured by a developing monitoring system on the basis of the measured values. The results of the gasification depend also on the quality of operators (subjective factor), which provide the process control. Therefore a controlling system was simultaneously developed and improved. At first, at the stabilization level, this ensured the stabilization of relevant control variables by help of the basic regulative circuits. We researched and, at the same time, we experimentally verified several ways of control, e.g. feed forward control (with the model) and feedback control methods (PI controllers, control with a dead zone,...). Based on a comparison of the controlling quality at the level of stabilization, the most suitable is the feedback control principle, implemented by PI discrete regulators at the control level of: - a volume air flow (by pulses of a servo – valve), -
a volume oxygen flow,
-
an air pressure,
-
rev of syngas ventilator ( a frequency converter), which was used to control the oxygen content in syngas.
In the Figure 11, there is showed a fundamental scheme of a control system, which task is to optimally control the UCG process. Therefore, immediately after dealing with the
16
stabilization level, we continued to deal with the superior optimization level. The result is a two – leveled controlling system consisted of: - stabilization level - optimization level.
Optimisation criterion
Optimisation
w
Stabilization Feedforward with automatic calibration
Feedback
z u У Process of UCG
Fig. 11 The structure of an optimal control
Where the task of the optimization level is a continuous real-time setting of a required values on stabilization level. If these values are kept then the whole process has an optimal course. Like at works on the stabilization level, also when dealing with optimization level, we researched various optimization methods, such as extremal control, optimization using gradient methods with restrictive conditions [7]. We verified them and continuously evaluated. The result is that the extremal regulation was able to fully replace the stabilization level at controlling only one input variable (volume air flow). Dealing with the multi parametric optimization (volume air flow, volume oxygen flow, rev of ventilator) two-leveled controlling structure has proven where the level of optimization was carried out by gradient method. The aim of a real – time optimization controlling was to maximize values of functional, which can mean: - the maximization of syngas calorific value, - the maximization of some required item (CO,H2, CH4,...) in syngas, - the maximization of a temperature in oxidation zone, - the maximization of CO / (CO + CO2 ) proportion. The optimal control method of multi – parametric system was based on the application of the gradient method with the restrictions on input quantities. We state a brief analysis of its work at one of the experiments (september/2009). By way of illustration, Fig.12 shows the optimal control process for maximizing the calorific value (criteria J-marked in blue) when the optimization algorithm optimized the flow of air, oxygen, and rev of ventilator. The algorithm run a total of 45 hours and increased the calorific value from an initial 4.4 to 8 MJ/m3. This result was achieved by gradually reducing the air flow, increasing oxygen flow and reducing the frequency of a converter. The opening of the servo - valve was reduced from the initial 16 to 5 pulses, the required oxygen flow increased from an initial 0.85 to 1.39 m3/ h and the frequency of a converter reduced from the initial 25 to 5 Hz. The algorithm had also set restrictions that did not allow to reduce the valve opening below three pulses and the converter frequency below 5 Hz. This restriction ensured a minimal air supply (4 m3 /h). The
17
function of an auto - restart algorithm was not activated during the test. Each optimized control variable u1 (air flow), u 2 (oxygen flow), u 3 (rev of ventilator for exhaust of a gasification product-syngas) had its own parameter h , ( h1 , h2 , h3 - optimization steps). The course of the optimal control in Figure 12 shows that the controlling system after each new calculated control action „ u „ gradually increased measured calorific value of syngas. With the lapse of time, it is necessary to add that the development of optimal control, can not be considered as completed. It has been shown even in the case, that the Gradient method has a feature that provides a finding a local and not a global extreme. Therefore, in another dealing, it would be desirable to choose a different method or eliminate the mentioned imperfection of the gradient method.
Fig. 12 The course of optimal control of UCG process
18
4. CONCLUSIONS The project researched underground coal gasification processes for Slovak lignite by thermal dissociation. The product of underground coal gasification (UCG) is a synthesis gas -syngas. In project, suitable coal supplies were analyzed and for the research was chosen the coal seem Cigel. The counter maps constructed on the lines of equal thickness of coal seam (focus on the seam h1 in the Cigeľ mine) showed it in 0 - 4 meters. This thickness of coal is appropriate for UCG method instead of conventional underground Coal-Mining Methods. The upper sealing component, above this coal seam h1, is built of roof clay seams of min. thickness 100 meters. The calorific value of coal from the seam h1 is up 8 to 16 MJ.kg-1, and it is suitable for planned purpose. The solving of project was based on the mathematical modeling and an experimental research. Thermodynamically and dynamical mathematical model was created. Thermodynamically model enables to simulate the composition of syngas on temperatures. Dynamical model serves for the study of the moving gasification front and temperature fields in UCG. Results from simulation were verified during 15 experiments in special laboratory for UCG. The laboratory consists from 2 generators. The original coal including undeburden, overburden was stored in agree criteria of similarity and it was gasified in this suitable chosen a generator. During solving of project was defined a new important parameter – specific reaction area. This parameter has a big influence on the effectiveness of UCG. This new knowledge is the call for a research and praxes. Properties of syngas were researched from follows view of points: types of coal models, the composition of oxidation gas mixture, temperatures of oxidation zones, pressures of gases, chemical composition of coal samplings and moisture of coal. In frame of project, the control of complicated UCG processes was solved also. Architecture of control system consists of two levels: -basic (stabilization level) -optimization level. This control system should arrange the optimal control. Criteria of effectiveness were defined. On the basis of these criteria, UCG has a higher yield in comparison with conventional underground Coal-Mining Methods. The syngas had the calorific value on average 3 – 6 MJ.m-3 for air and for mixture (air + oxygen) till 13 MJ.m-3. In conclusion, we can state that using the gained knowledge, in particular its accurate implementation in practice, it is possible to improve UCG technology and obtain syngas with
19
a higher quality, especially with a higher calorific value. That will lead to the attractiveness of this technology because the payback period of investments will shorten.
Acknowledgment This work was supported by the Slovak Research and Development Agency under the contract No. APVV-0582-06 and HBP a.s. Prievidza (Upper Coal Mines Nitra,a.s. Prievidza). References [1] Valeš , J. and Šafářová, M.,2006 . Zplynovaní uhlí (Coal Gasification), Report VV-123/06,Most: VUHU (Research Institute of Brown Coal, Most). [2] Friedmann, S.J., Upadhye, R., Kong, F.M. Energy Procedia 1 (2009), pp. 4551 – 4557 http://www.sciencedirect.com [3] Walter, K. Fire in the hole, S & TR LLNL (April 2007), pp. 12 -18, www.llnl.gov/str/April07/Friedmann.html [4] Kostúr, K., et. al. Underground Coal Gasification by thermal dissociation. (Research report in Slovak langue), Technical University Kosice, (2008), p.359 [5] Kostúr, K. Importance of pre-project research and evolution of UCG process before ITS realisation, UCG Summit, London, UK, (October 2009), http://www.iqpc.com/presentations/login.aspx [6] Kostúr, K., 2007. The mathematical modeling of relevant processes for underground coal gasification. In Takashi,L. et al. (Eds.). Proceedings of 5th International symposium on Earth Science and Technology, Kyushu, Japan 29-4 December 2007. Fukuoka: University Fukuoka, 475-480. [7] Kostúr K., Kačúr,J.: The monitoring and control of underground coal gasification in laboratory conditions, Acta Montanistica Slovaca ,Vol. 13 (2008), No. 1, 111-117, http://www.actamont.tuke.sk, ISSN 1335-1788 [8] Kostúr, K.: THE PROPOSAL OF OPTIMAL PROCES CONTROL FOR UNDERGROUND COAL GASIFICATION, Transfer inovácií, Volume 14 (2009),pp.13-17, http://www.sjf.tuke.sk/transferinovacii
20
AN INTEGRATED 3-D UCG MODEL FOR PREDICTING CAVITY GROWTH, PRODUCT GAS, AND INTERACTIONS WITH THE HOST ENVIRONMENT John J. Nitao, Lawrence Livermore National Laboratory, P.O. Box 808, L-052, Livermore, CA, 94550, U.S.A., [email protected], 925 423 0297 Thomas A. Buscheck, Lawrence Livermore National Laboratory, P.O. Box 808, L-223, Livermore, CA, 94550, U.S.A., [email protected], 925 423 9390 Souheil M. Ezzedine, Lawrence Livermore National Laboratory, P.O. Box 808, L-188, Livermore, CA, 94550, U.S.A., [email protected], 925 422 0565 S. Julio Friedmann, Lawrence Livermore National Laboratory, P.O. Box 808, L-175, Livermore, CA, 94550, U.S.A., [email protected], 925 423 0585 David W. Camp*, Lawrence Livermore National Laboratory, P.O. Box 808, L-223, Livermore, CA, 94550, U.S.A., [email protected], 925 423 9228 * Corresponding Author 27th Annual International Pittsburgh Coal Conference, October 11-14, 2010, Istanbul, Turkey LLNL document number LLNL-CONF-459611
Abstract Underground coal gasification (UCG) is an alternative to surface gasification for recovering the energy of coal in the form of an energy rich product gas. While having surface gasification’s advantages of low pollutant emissions and efficient CO2 management, UCG avoids the environmental and safety issues of coal mining and processing by using process wells to gasify the coal deep underground in its native seam. We are developing a comprehensive scientific and engineering simulation tool to support UCG site selection, design, environmental assessment, operations, monitoring, and closure. Simulation of UCG is computationally challenging because it involves highly coupled multi-physical/chemical processes over wide ranges of time and space scales. We apply a “divide and conquer” computational strategy in our UCG simulator architecture. Domain models simulate the physical and chemical processes within five distinct domains. Model-integration components interface the domain models and determine the evolving boundaries between those domains. The physical domain models of the simulator are the Wall Zone Model (WZM), Rubble Zone Model (RZM), Cavity Gas and Heat Model (CGM), Thermal-Hydrological Model (THM), and Geomechanical Model (GMM). The integration components of the simulator are the Boundary Evolution Model (BEM) and the Simulator Manager (SM). The thermal-hydrological and geomechanical models are based on mature pre-existing codes that have been validated. Progress on new models includes a nearly-completed wall zone model and a partially-completed boundary evolution model. Plans are in place to complete all model components, integrate them into a full-scope UCG simulator, and test the simulator and its domain models against lab- and field-scale data. We expect this simulator to be a useful tool for growing our understanding of UCG and informing engineering decisions during the life cycle of a UCG project.
1 1.1
Introduction
Overview of Underground Coal Gasification (UCG) Coal is the most widely distributed energy source in the world, supplying 27% of worldwide demand in 2007, second only to oil. As oil reserves decline, reliance on coal is expected to continue to rise, supplying 29% of worldwide energy demand by 2030, when it will nearly overtake oil (IEA, 2009). Because the combustion of coal emits 1.3 times more CO2 than oil and 1.7 times more than gas, the increasing reliance on coal poses a major challenge to the need to reduce greenhouse gas emissions. Gasification technologies have been developed to recover the energy content of coal in an environmentally clean way. Typically, coal is mined, transported, crushed or pulverized, cleaned, and fed into a specially-designed gasification reactor operating at high temperatures and often high pressures. Oxygen (or air) and sometimes steam are injected and react with the coal to produce an energy-rich product gas containing primarily H2, CO, CH4, and CO2. This product gas is cleaned of particulate and criteria pollutant species (and CO 2 if desired), and
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Nitao et al., LLNL
then can be burned for heat or power generation, or used as the feedstock for production of synthetic liquid fuels, natural gas, or chemicals such as methanol or ammonia. Large investments have been made worldwide in coal gasification because of its ability to greatly reduce emissions of conventional criteria pollutants. Gasification is now receiving increased attention and investment because of its efficient compatibility with carbon capture and sequestration (CCS) schemes. Simply, it is easier and cheaper to remove CO 2 and other pollutants from gasifier product gas than from post-combustion flue gas. Underground coal gasification (UCG) is a method of gasification that is done in situ, deep underground below the water table. It uses the coal seam itself as the gasification reactor, and drilled wells for the injection of oxidants and removal of product gas (Figure 1). UCG has been demonstrated to be technically feasible in pilot, demonstration, and commercial operations from the 1960’s through the present in the Former Soviet Union, United States, China, South Africa, Australia, and others. In addition to the environmental advantages UCG shares with surface gasification, UCG avoids the many environmental and worker-safety impacts of coal mining, transportation, processing, and waste disposal, and may reduce the demands on local potable water supplies. UCG has its own unique environmental issues to manage. It produces pollutants, and a fraction of these may be left underground. The consumption of coal creates a cavity, and the rock above it may fracture and/or collapse into the cavity. Surface subsidence to some extent is possible. UCG can and has been done with minimal, acceptable levels of environmental and safety impacts. This has been demonstrated in tests such as Rocky Mountain 1 (Hanna Basin, Wyoming USA) and Chinchilla (Queensland, Australia). The environmental interactions must be understood well and managed properly through proper site selection, design, monitoring, operation, and closure protocols. UCG’s economics appear attractive enough that all or most of the projects that have been started and are being proposed worldwide since the late 1990’s have been done with private financing with little or no government subsidy. UCG’s favorable economic estimates are because the costs of mining, coal transport and processing, and the gasification reactor are replaced with a few wells completed into the subsurface, and a surface piping array. UCG is also generally well suited to coal seams that have low value for conventional mining-based uses because of some combination of large depth, low quality, or small thickness. A typical UCG module involves an oxidant injection well and a product gas production well. During module preparation, a channel or permeable link is created between the injection and production wells. Types of linkages include: (1) horizontal borehole, which is used in Controlled Retraction Injection Point (CRIP) modules (Hill, 1986), (2) pneumatically- or hydraulically-stimulated fractures, used in the former Soviet Union (FSU), (3) Typical UCG Injection Well reverse combustion, which creates a small-diameter yet Well Layout sometimes wandering channel (c.f. Stephens, 1981), and (4) Subsidence ?? galleries left after mining operations, as done in China (Khadse et al. 2007). After the link is established, combustion is initiated at or near the outlet of the injection Water table well, some combination of air, oxygen, and/or steam is Production pumped into the injection well, the coal is gasified in a Well 200 to growing cavity within the seam, and the product gas is ~1500 m recovered through the production well. The simulator described in this paper assumes in its initial configuration that the process wells, link, and initial ignition stage have already been completed.
~ 30 m
~ 100 m
Figure 1. UCG gasifies coal in situ, using wells to feed air and/or oxygen and (sometimes) steam and to recover an energy-rich product gas. While this example depicts a linear Controlled Retraction Injection Point (CRIP) module (Hill 1986), many other module configurations are possible.
Lawrence Livermore National Laboratory
2
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Wall Zone (Rock) Wet rock
25 to
Cavity
Cavity
H2, CH4, CO, H2O, & CO2 Wall Zone (Coal)
Rock
Dry rock
Rubble Zone
Produce gas (100-300 Btu/scf)
Inject air or O2 & H2O (steam, optional)
40oC
Wet coal Dry coal
Coal
1100oC
Char
Cavity
Cavity
Rock, coal, char, ash pieces
300oC
Char Dry coal Wet coal Wet rock
Wall Zone (Coal) Wet coal
Rock
Dry coal Lawrence Livermore National Laboratory
Coal (~C H0.8 O0.2) + O2 + H2Ol
Nitao et al., LLNL
Cavity
CO + H2 + CH4 + HC’s + CO2 + H2Og
Figure 2. A representative cavity configuration is shown at a mid-to-late stage of a linked vertical well module. UCG involves several distinct near-field domains, including the cavity, the wall zone containing coal, the wall zone containing rock, and the rubble zone. The illustration qualitatively shows the higher flux of water into the cavity expected at lower elevations due to the greater pressure difference between the cavity and the hydrostatic surroundings. The hydrological and geomechanical domains extend far beyond the scale of this illustration to the scale shown in Figure 1.
Figure 2 illustrates one of the simplest module concepts, a pair of vertical wells that have been linked by a process such as reverse combustion. It also illustrates some of the many of the phenomena that occur within the cavity gas, wall, and rubble zones. These phenomena are not limited to linked vertical well modules. The oxidant gas is introduced into the hot oxidative region. The regions where oxidation, gasification, pyrolysis, and drying occur propagate outward from the injection well and forward towards the production well. In coal, the wall zone transitions from hot char and ash adjacent to the cavity, to hot dry coal, to wet coal. Important pyrolysis, gasification, and oxidation reactions occur in the coal and char regions within the wall zone. In rock, the wall zone transitions from hot dry rock adjacent to the cavity, to wet rock. The drying and reaction fronts within the wall zone typically propagate a few centimeters to a meter or more per day. The open cavity volume governs the transport of species and heat and is host to gas phase reactions. The rubble zone contains ash from gasified coal, small pieces of coal, char, and rock that have spalled from the ceiling or side walls, and larger pieces of coal or rock that have fractured and fallen in due to mechanical phenomena. Coal, char, and gas reactions in the rubble zone are the same as elsewhere, but gas flow and mass and heat transport more closely resemble those of a packed bed than an open volume or porous medium. Outside of the near-cavity region important phenomena include transport of water and heat, and geomechanical phenomena of structural roof collapse, fracturing, stress changes, and subsidence. Cavity growth and process performance depend on many factors. Some of these can be controlled, such as the oxidant composition and injection rate and cavity pressure. Some are natural-system factors, including the thermal-hydrological-chemical-mechanical (THCM) properties of the host coal/rock environment. Coal and char chemistry, oxygen and steam availability, and temperature (controlled by convective and radiative heat-transfer) affect the reactions and heat balance. Formation permeability, the surrounding hydrostatic pressure field (which would be affected by pumping), and the cavity pressure govern the water influx into the cavity as a function of 3
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Reactions
Heat
Mechanics
Flow
Nitao et al., LLNL
Figure 3. UCG involves tightly-coupled multi-physical/chemical processes, involving chemical reactions, heat and mass transfer, hydrological flow, and geomechanical behavior. These coupled processes are influenced by the thermal-hydrological-chemical-mechanical (THCM) properties of the heterogeneous coal/rock environment, and the UCG module design and operating parameters.
position and time. Water influx affects the gasification chemistry and the local and system heat balances (Buscheck et al. 2009). Geomechanical response of the coal and rock strata to the growing cavity affect the stability and fracturing of roof materials, and/or collapse of roof material into the cavity. Cavity roof collapse affects local and system heat balances, raises the cavity ceiling location, adds coal or rock rubble into the cavity, and influences surface subsidence. Geomechanical fractures and stress reduction will alter the permeability field which influences water influx. All of these effects interact within the heterogeneous coal and rock environment in a dynamic, non-linear manner and affect the UCG cavity growth and process performance. UCG involves a wide range of coupled physical and chemical phenomena that operate over a wide range of length and time scales (Figure 3). The evolving size of the cavity formed during a UCG operation impacts directly the economic and environmental factors crucial to its success (Britten and Thorsness, 1989). The maximum cavity dimensions that can be achieved and that are environmentally acceptable will strongly affect the overall economics of a UCG project. The maximum cavity size (especially the short dimension in the plane of the seam) that can be achieved will be limited by the ability of oxygen and heat to be transported to the coal walls, and the ratio of heat generated by coal consumption to heat lost to boiling of water permeation influx and heating and drying of exposed and/or collapsed roof rock. The maximum cavity size that is environmentally acceptable will be limited largely by the geomechanical responses (stress reduction, fracturing, collapse, and subsidence) and their effects on the regional hydrology and surface profile. 1.2
Purpose and Requirements for a Full-Scope UCG Simulator UCG projects require a wide range of engineering decisions on many topics during the UCG project life cycle. These include: Site selection o resource evaluation o site screening o evaluation of candidate sites in relation to module concepts Module design and construction details o evaluation of alternative module types (linked vertical wells, linear CRIP, parallel CRIP, etc.) o consideration of coal analyses and options for injectant composition o consideration of geotechnical factors such as seam thickness, dip, and strata strengths and permeabilities o type and location of process wells o evaluation of alternative linking technologies o location of water-management (sump/recharge) wells Analyses of UCG-environment interactions o analyses of the effects of the project on the environment o analyses of the effects of environmental factors on UCG performance o environmental assessments for permitting o environmental management plans 4
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Nitao et al., LLNL
Operating conditions as a function of time o air/oxygen/steam composition, injection rate, and pressure (cavity, injection, or production) o water well pumping or recharge rates o planned operating conditions o real-time operational decisions Process and environmental monitoring o design, operation, and interpretation o regulatory compliance o scientific and operational data for process understanding, model calibration, and scale-up o Interpretation of monitoring data to infer system performance Module and project closure protocol
A UCG simulator should be able to inform and guide decision makers over the full range of these decision topics. A UCG simulator will also be an important tool for gaining scientific insight into UCG phenomena and their interdependencies. To provide high quality analyses to aid the decision topics listed above, a simulator must accurately predict how choices being considered will influence coal consumption and cavity growth, product gas composition and rate, and environmental interactions. A full-scope UCG simulator should be fully threedimensional and include the following capabilities: Ability to input 3-D geology and rock and coal property fields Ability to input time-dependent operator-controlled parameters Arbitrary cavity shape and size evolution, with local rates governed by local conditions and processes Tracking of the spatial geometry and content of the rubble zone and its effect on cavity growth Rubble zone flow, heat and mass transport, and reactions of coal and char pieces Overburden stress changes, fracturing, cavity roof collapse, and surface subsidence Resolved cavity gas flow, turbulent mass and heat transfer, and radiative heat transfer Detailed coal and char reactions, including the influence of spatially-varying water influx Hydrological interactions between water influx to the cavity and actively-managed surroundings Table 1 summarizes the extent to which existing UCG simulators have the last seven of these capabilities. The model of Britten and Thorsness (1989) is one of the most comprehensive published UCG simulators to date and it has only three of the seven tabulated capabilities. We plan that the full-scope UCG simulator described in this paper will incorporate all nine of the capabilities listed above. Given the complex physics and chemistry involved and the large number of possible cases to consider, a UCG simulator should be computationally efficient and be able to perform high-fidelity runs on high-performancecomputing (HPC) platforms. Another important attribute of a UCG simulator is that it has a flexible modular design that facilitates revisions to code modules and flexible execution. While some analyses will require fullscope integration, others will be adequately and more efficiently executed by exercising only one specific model component, perhaps with initial and boundary conditions provided by the output of an integrated simulator run.
2
Simulator Architecture
2.1
Computational Approach UCG involves many physical and chemical processes, occurring in distinctly different domains. The nearfield domains include the gas-filled UCG cavity, the reacting, spalling porous wall zone (consisting of char and coal or rock), and the rubble zone (consisting of ash, char, coal, and rock pieces and/or slag). Extending beyond the cavity, thermal, hydrological, and geomechanical processes influence water influx, roof collapse (blocks of overburden falling into the cavity due to mechanical stresses), fracturing, and subsidence of the ground surface. As shown in Figure 4, we apply a “divide and conquer” computational strategy in our UCG simulator architecture. Rather than attempting to solve all the complex coupled multi-physical/chemical processes with a single monolithic simulation tool, our approach is to solve them with a set of models, each designed for the respective domains that they represent, and then to interface their results. Domain models simulate the physical 5
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Nitao et al., LLNL
Table 1. Summary of capabilities for existing UCG simulators and the one in development described here. Author (Year)
Dinsmoor et al. (1978) Eddy and Schwartz (1983)
Thorsness and Britten (1986)
Biezen et al. (1995) Kuyper et al. (1996) Yang (2006) Perkins Sahajwalla (2005, ‘08) Khadse et al. (2006) Daggupati et al. (2009) This effort
Arbitrary 3D cavity; local rates depend on local phenomena No – 1-D channel No – 2-D horizontal cylindrical to rectangular segments No – 2-D vertical cylindrical segments Yes – coarsely resolved in space & time No – 1-D channel No – 1-D channel No – 1-D channel No – 1-D packed-bed No – CFD & reactor compartments Yes
Rubble zone tracked and affects cavity growth No
Rubble zone transport and reactions No
Geomech. Collapse, fracturing, subsidence computed No
Resolved cavity gas flow; mass and heat transport No – 1-D flow No – 1-D flow
Detailed coal/char reactions with local water influx No
Hydrology coupled with active local-area pumping No
No
No
No
No
No
Yes
Yes
No
No – wellmixed
Yes – semianalytic
No – porous medium flow equations No
No
No
Yes – ad hoc rule based No
No – hydro. model but no pumping No
Yes – Highly idealized & coarse Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes – 3-D at lab- scale No – 1-D flow No – 1-D packed-bed Yes – CFD & reactor compartmnts Yes
No – simplified No
No
Yes
No
Yes
No
Yes
No
Yes
Yes
No
Thermal-Hydrological Model (THM)
Geomechanical Model (GMM) • Roof collapse • Subsidence • Stress changes & fracturing
Wall Zone Model (WZM) • • • •
Reactions & drying Heat & mass transport Spallation Gas & solid compositions
• Groundwater flow & influx • Pore pressure field • Thermal response
Cavity Gas Model (CGM)
Simulator Manager (SM)
• Execution control • Information exchange • Data conversions & merging
Boundary Evolution Model (BEM)
• Reactions & gas composition • Heat & mass transport • Turbulent mixing
Rubble Zone Model (RZM) • Reactions & drying • Heat & mass transport • Gas & solid compositions
• Cavity growth from wall reactions & structural failure • Rubble zone geometry • Roof movement due to structural failure
6
Figure 4. System architecture for the full-scope UCG simulator.
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Nitao et al., LLNL
and chemical processes within five distinct domains. Two model-integration components interface the domain models and determine the evolving boundaries between them. The component models of the simulator are Wall Zone Model (WZM) Rubble Zone Model (RZM) Cavity Gas and Heat Model (CGM) Thermal-Hydrological Model (THM) Geomechanical Model (GMM) Boundary Evolution Model (BEM) Simulator Manager (SM) The WZM, RZM, CGM, THM, and GMM are domain models that calculate multi-physical/chemical processes within their respective domains. The SM and BEM are model-integration components (rather than models of domain physics) that work to tie the physical domain models together and track the boundaries between them. Each component of the simulator is described below. The domain models will also be able to run as standalone models so they can be used for conducting validation, calibration, sensitivity or scoping analyses. 2.2
Simulator Manager (SM) The functions of the SM are Control all of the domain models (WZM, RZM, CGM, THM and GMM) and model-integration component (BEM). Communicate data and information needed for interfacing respective domain models. Convert between different data formats used by the respective models and the BEM model-integration component.
As part of its control function, the SM must determine the time scale over which each model is called in order to meet accuracy and numerical stability requirements, yet maintain computational efficiency. In many cases, a model may pass to the SM when it should be called again, or when it may give changes or rates of change in important physical/chemical variables. For example, the GMM may indicate to the SM when it must be called again because the roof is approaching the critical stress conditions. An additional software design goal is to the ability to substitute different models of the physical domains into the simulator. 2.3
Wall Zone Model (WZM) The WZM is responsible for all of the small-scale physical and chemical processes in the coal and rock that directly influence the conversion of coal and cavity growth, including drying, coal pyrolysis, char gasification and oxidation, and coal and rock spallation. The very steep temperature gradients in the wall zone result in the smallscale processes typically occurring within the first tens of centimeters into the cavity wall. Because of the steep gradients in temperature and composition existing there, it would be numerically overwhelming to have a finely resolved 3-dimensional computational mesh near the cavity wall, given a typical module scale of tens to hundreds of meters. Our approach is to take advantage of the fact that fluxes are approximately normal to the cavity wall, which allows us to use an array of 1-D WZM models, oriented orthogonal to, and distributed along the cavity wall surface (Figure 5). An array of 1-D wall zone models are used to (1) compute mass and energy fluxes that are exchanged with the cavity gas, (2) compute mass and energy fluxes where the wall is in contact with the rubble zone, (3) exchange fluid transport and heat fluxes with the THM, and (4) pass the coal/char conversion rate and the coal/char and rock spallation rates to the BEM and RZM for computing changes in the cavity and rubble zone boundaries. The WZM is coupled to the THM as shown in Figure 5. The ends of the 1-D wall models that point away from the wall are coupled to the THM, which is a 3-D porous-media model for groundwater flow and heat transport in the surrounding coal and rock.
7
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
~ 2"
Thermal-Hydrological Model (THM) Geomechanical Model (GMM)
Caprock 1-D WZM
Cavity cavity Gas Model (CGM)
Wall Zone Model (WZM)
oxygen + combustible and waste gases + heat
Coal
Figure 5. An array of 1-D Wall Zone Models (WZM’s) surround the surface of the cavity. These interface on the inside with either the Cavity Gas Model (CGM) or the Rubble Zone Model (RZM, not shown here), an on the outside with both the ThermalHydrological Model (THM), and the Geomechanical Model (GMM).
Bedrock
Not drawn to scale
Cavity Gas Model (CGM)
Nitao et al., LLNL
char breakoff
dry coal
THM & GMM
wet coal/cap/bedrock
15+ reactions included evaporation
2.4
Rubble Zone Model (RZM) The RZM simulates all processes within the rubble pile which may consist of coal, char, ash, and/or rock fallen from the cavity ceiling and side walls. These may be in particles of various size and shape, or molten or frozen slag. The reactions within the rubble are essentially the same as those in the WZM. However, fluid flow and mass and heat transfer are more complicated because the pile is a mixture of micropores, consisting of lowpermeability fragments of coal, ash, and rock, surrounded by macropores which give the rubble a high permeability. Two approaches are available to address this: chemical engineering packed bed reactor modeling approaches (c.f. Thorsness and Kang, 1986); and a dual porosity reactive transport work from the hydrological flow and transport field. Physical-chemical processes within coal fragments may be modeled using representative porous spheres with effective properties and equivalent surface area and volume. The reactive gas transport in the macropore space may be modeled as a porous continuum using a 3-D numerical model that is coupled to the CGM. 2.5
Cavity Gas Model (CGM) The CGM encompasses all gas-phase processes within the cavity: fluid flow and turbulent mixing, mass and heat transfer, thermal radiation, and reactions. This model solves the Navier-Stokes equations, coupled with species and energy balances. Cavity combustion is highly dependent on turbulent mechanisms by which the injected gas stream mixes with the fluxes emanating from the wall due to water influx and drying, coal pyrolysis, and char gasification and oxidation, as modeled by the WZM. Pre-existing commercial and non-commercial computational fluid dynamics (CFD) codes with combustion are currently being evaluated for the use in the CGM, with the currently favored code being STAR-CCM. Simplified, abstracted models using transfer coefficients are also being considered for performing fast, but approximate, calculations. However, the lack of established transfer coefficients under mixed convection (i.e., simultaneous free and forced convection) conditions, expected in the UCG cavity, limits the accuracy of this latter approach, unless calibration using CFD codes is performed.
8
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Nitao et al., LLNL
2.6
Geomechanical Model (GMM) The purpose of the GMM is to predict roof collapse, large-scale spallation, and large-scale subsidence due to changes in the geomechanical stress field. These events are induced by the stresses in response to changes in the cavity wall geometry and the temperature field. The following geomechanical codes, developed at LLNL, and which have complementary capabilities, are being adapted for use in the GMM. Geocentric is a finite element framework for geomechanical simulations. It provides the ability to model nonlinear mechanical behavior via continuum elastic and elastic-plastic models, and can also model tightly-coupled hydromechanical systems involving solid deformation and variably saturated fluid flow. Geocentric uses primarily implicit formulations, and achieves parallel scalability through the use of algebraic-multigrid, preconditioned, Newton-Krylov solvers (White and Borja, 2008, White, 2009; White and Borja, 2010). LDEC is an explicit discrete element code, in which a highly-fractured domain is decomposed into a series of polyhedral blocks that can contact and slide past one another. These blocks can either be rigid or deformable, and can be used to simulate the dynamic collapse of underground structures due to joint slip. (Morris et al. 2003, 2006). 2.7
Thermal-Hydrological Model (THM) The THM is responsible for 3-D thermal-hydrologic flow throughout the subsurface domain that extends from the outer cool end of the WZM, all the way out to the far-field boundaries of the ground surface and a radius at which UCG effects are negligible. Some modest distance away from the cavity, due to hydrogeologic heterogeneity (mainly strata-based properties), heat and water flux vectors are no longer normal to the cavity wall surface. This is one of the several reasons for requiring the THM to be 3-D. Moreover, large-scale geomechanical effects can alter fracture permeability. Therefore the THM needs to be interfaced with the GMM to determine whether significant alteration in fracture permeability has occurred. It is important that the THM have broad capability, including the ability to address the following: Calculation of local fluxes of water permeating into the process zone boundary (through the wall zone) as a function of time Water removal from the formation into the cavity (and production in the form of vapor) will influence the surrounding subsurface water balance. It will be important to assess the hydrological impact of UCG operations on the surroundings. In the case of very large scale UCG operations, it will be important to calculate the effect on the regional groundwater hydrology. At very large scale, it will likely be more appropriate to use only the THM model, uncoupled from the UCG simulator. Pumping from (or possibly recharge into) water sump wells outside of the cavity may be used in the preparation and operational management of a module to affect the local water pressure and flow fields. The location of these wells might be anywhere from a sump located in a low point in or very near the cavity, to water wells tens, hundreds, or maybe thousands of meters away. The ability to model this in 3D is required to inform operational decisions before, during, and after UCG operations. Industrial-scale UCG operations may generate enough heat in situ to influence groundwater temperatures up to 100 m away from the UCG modules, with a persistence of years. It will be important to address this effect in conducting environmental risk assessments. The post-operation cavity washing protocol will need to be modeled. For example, ramping the cavity gas pressure down to atmospheric will cause water to permeate faster into the hot rock, coal, and char wall and rubble zones which will produce steam and drive gasification reactions until lower temperatures slow the kinetics, and then continue to produce steam and finally warm water that will approach ambient. Modeling this will be a good use of the simulator and requires a good THM. Once cooled below boiling, it is likely that the THM will be initialized through the simulator, then run uncoupled to assess long term thermal hydrology and possible transport of any contaminants produced by the UCG operation away the UCG cavity in order to manage the process and assess potential environmental risks. An existing code, NUFT, developed at LLNL (Nitao, 1998), will be used for the THM.
9
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Nitao et al., LLNL
2.8
Boundary Evolution Model (BEM) The BEM advances (in three dimensions) the surfaces that bound the three major domain models: the cavity volume, the wall (coal/rock) zone, and the rubble zone. Note that the term “cavity volume” used here refers to the open gas-filled portion that is not occupied by rubble. The evolution of the boundary between the cavity and the coal/char or rock wall zone is the result of the following effects. Solid char is converted to gas products by oxidation and gasification reactions, calculated by the WZM Small-scale spalling of ash, char, coal, and rock at the cavity wall, calculated by the WZM Structural roof collapse can occur, moving the ceiling higher, calculated by the GMM. The evolution of the boundary between the cavity and the rubble zone is the result of the following effects. Ash and unreacted char falls from cavity wall, calculated by WZM Spalled pieces of ash/char/coal or rock fall from cavity wall or ceiling, calculated by WZM Structural roof collapse can occur, adding blocks of roof material to the rubble, calculated by the GMM Pyrolysis, gasification, and oxidation of coal and char pieces, calculated by the RZM Rubble surface recedance, due to particles shrinking, compacting, or slagging, calculated by the RZM Solid particles sliding on the rubble surface when the angle of repose is exceeded, calculated by the RZM The evolution of the boundary between the rubble zone and the coal/char or rock wall zone is the result of the following effects. The wall zone and rubble zone mechanisms described above, calculated by the WZM and RZM
3
Progress to Date
3.1
Wall Zone Model A working 1-D model for the WZM has been completed. It is an extension of the comprehensive mathematical model of Perkins and Sahajwalla (2005, 2008). The governing partial differential equations are 1 balance equation for gas phase momentum, 1 balance equation for the gas phase, Ng balance equations for the Ng gas species, Ns balance equations for the Ns solid species, 1 balance equation for energy, 1 balance equation for the macropore porosity, 1 balance equation for the micropore porosity. Reactions between gas species are: water shift: CO + H2O ↔ CO2 + H2 , methanation: CH4 + H2O ↔ CO + 3H2. The combustion of char is given by the reactions: C + H2O ↔ CO + H2, C + CO2 ↔ 2CO, C + 2H2 ↔ CH4. The pyrolysis reactions are of the form: VMi(s) ↔ VMi(g), where VMi = CO, CO2, H2, H2O, CH4, C2H6, H2S, TAR, N2. The TAR is a pseudo-species that encompasses all of the non-listed volatiles. Other reactions, such as char oxidation, can be added since the code is completely flexible in this regards. The coal is considered to be a porous medium with bi-modal pore size distribution: volatilization increasing macropore size and char combustion, increasing micropore size. The specific char reactive surface area is a function of micropore porosity. The balance equation for gas phase momentum used for volume-averaged flow in the pore space is the Brinkmann equation: 10
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
V t
VW V p
Nitao et al., LLNL
W g h 0 ,
K
where gas density, V gas velocity, W specific discharge (= V ), gas volume fraction, viscosity, p gas pressure, K permeability, g gravitational acceleration,
h elevation as a function of position.
The gas species balance equations includes transport of multiple gas species by advection and molecular diffusion. The energy balance equation includes the following effects: Thermal radiation occurs in the cavity. Thermal conduction occurs in the bulk medium. Thermal advection occurs due to gas flow. Heat of reaction is included, including that resulting from the evaporation of water. Energy transport from species molecular diffusion is included. Note that this effect is not included by Perkins and Sahajwalla (2008). Specific heat depends on porosity and liquid water content. The 1-D wall zone for coal is comprised of two zones: a dry zone that extends from the hot char-cavity interface to the drying “front”, and a wet zone of coal which is water-saturated. (The dry “coal” zone includes the continuum from dry coal at the steam saturation temperature to char to residual ash which falls when the ash/char ratio exceeds a limit.) Evaporation of water occurs at the interface between the two zones. A boundary-layer transfer coefficient is used to model mass and energy and transfer between the coal face and the cavity gas. Our current version of the WZM uses the computationally simple representation of Perkins and Sahajwalla (2008) for spalling that occurs continuously as the non-ash content at the cavity end decreases below some critical value. The discontinuous spallation approach of Camp (1980) and Thorsness and Britten (1986) will be implemented in later revisions. Because of the high gradients involved and the moving cavity wall boundary and wet-dry zone interface, a moving computational grid has been implemented. Validation of the WZM is in progress, including comparison with data from experiments reported by Westmoreland and Forrester (1977). Because not all of the data is available in their reports, we used data from Tsang (1980), who may have obtained it through personal communication. In the experiment a 15-cm-diameter cylindrical block of sub-bituminous Wyodak coal was heated in an argon-filled oven to provide an inert atmosphere. Figure 6a shows the progress of the drying front as inferred from the temperature. Also shown is predicted drying front using our WZM. The predicted curve is very close to that reported by Perkins and Sahajwalla (2005). Figure 6b shows a snapshot of the temperature versus the normalized radius. Comparisons with other experiments will be made in the future. 3.2
Cavity Gas Model (CGM) The application of the STAR-CCM code to the CGM is in the early phase of development.
3.3
Geomechanical Model (GMM) The GMM is expected to utilize the Geocentric code (White and Borja, 2008, White, 2009; White and Borja, 2010) in the near term. The LDEC code (Morris et al. 2003, 2006) may be used in later versions of the GMM for the near- to mid-field representation to better-model block fracturing. These are both mature codes; however, there has been only limited experience in applying them to UCG. Some analyses have been done geomechanical processes in UCG (Oleg et al. 2008; Morris et al. 2009). 11
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
Nitao et al., LLNL
Figure 6. Wall Zone Model predictions (solid curve) and experimental data (points) for a 15-cm-diameter cylinder of sub-bituminous Wyodak coal (Westmoreland and Forrester, 1977). The comparisons include (a) temperature versus dimensionless distance, and (b) drying front propagation.
3.4
Thermal-Hydrological Model (THM) The THM, based on the NUFT code (Nitao, 1998), is one of the more mature domain models of our evolving UCG simulator. Some work has been done with NUFT to analyze the influence of thermal-hydrological processes on UCG behavior (Buscheck et al. 2009). 3.5
Boundary Evolution Model (BEM) Work is proceeding on the BEM which governs movement of cavity and rubble zone boundaries. Our implemented algorithm has the following advantages over other surface tracking methods: It conserves volume exactly. For example, the volume increase in the cavity growth due to small-scale wall spallation must equal the total time-integrated volume flux from the WZM. It correctly distributes volume fluxes computed by the 1-D wall-zone models to points along the entire wall surface. It accounts for volume flux differences resulting from spatial variability in coal properties. It is spatially consistent in that the wall surface absolutely never enters the region where the coal no longer exists. Figure 7 is a demonstration of the BEM algorithm for an idealized test case with the following simplifications: 30% of the coal volume that is removed from the wall falls onto the rubble pile. No reactions can occur on the wall where it is in contact with the rubble. No reactions can occur in the rubble zone (i.e., pile is purely ash). Ash will slide when the angle of repose (assumed to be 30º) is exceeded. Spatially uniform and time-constant cavity gas composition and temperature is used. We initially start at (t = 0 day) with a cylindrical cavity (Figure 7a). The cavity expands as wall material is consumed or detaches according to a rate computed from a single 1-D wall model. (In the final simulator, multiple 1-D wall models will be distributed along the cavity wall.) Note that the insulating ash rubble causes the cavity to expand upwards, as observed in field tests through back-excavation (Thorsness and Britten, 1986). The narrow white region between the cavity wall and the rubble is an artifact due to the color mapping used by the contouring program.
12
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
A Simple Preliminary Test Example
Nitao et al., LLNL
Figure 7. This simple demonstration of the Boundary Evolution Model (BEM) starts with a 1-m radius cavity. A single 1-D WZM calculation predicts the advance rate for all exposed coal surfaces. 30% of the affected coal falls as rubble. No wall reaction occurs when covered by rubble. Angle of repose distributes rubble.
Assume an initial cylinder (r = 1 m)
4
Future Work
Planned work includes completion and testing of all model components, and validation of the domain models and the full simulator using available lab- and field-scale experimental data. Sensitivity studies will be run to gain scientific insight into UCG system behavior. We expect to use the simulator and its domain models to guide decisions and interpret results for current and future UCG projects. We plan for iteration between model development, validation, scientific improvement, and application to projects.
5
Summary
Underground coal gasification is an economically Rubble and environmentally promising option for converting coal into an energy-rich gas without having to mine the Zone coal. The coal is gasified in its native seam deep underground by injecting air or oxygen and sometimes steam, and recovering the product gas through wells drilled into the seam. Successful UCG requires a series of engineering decisions on site selection, module design and construction, environmental analyses, operating conditions, process and environmental monitoring, and closure protocol. A UCG simulator can aid these decisions by predicting what would happen for different scenarios. We are developing a three-dimensional simulator with the following capabilities: Ability to input 3-D geology and rock and coal property fields Ability to input time-dependent operator-controlled parameters Arbitrary cavity shape and size evolution, with local rates governed by local conditions and processes Tracking of the spatial geometry and content of the rubble zone and its effect on cavity growth Rubble zone flow, heat and mass transport, and reactions of coal and char pieces Overburden stress changes, fracturing, cavity roof collapse, and surface subsidence Resolved cavity gas flow, turbulent mass and heat transfer, and radiative heat transfer Detailed coal and char reactions, including the influence of spatially-varying water influx Hydrological interactions between water influx to the cavity and actively-managed surroundings We apply a “divide and conquer” computational strategy in our UCG simulator architecture. Five domain models simulate the physical and chemical processes within the wall zone, rubble zone, cavity gas zone, thermalhydrological, and geomechanical domains. Two model-integration components interface the domain models and determine the evolving boundaries between those domains. The thermal-hydrological and geomechanical models are based on mature pre-existing codes that have been validated. Progress on new models includes a nearly completed wall zone model (of coal and of rock) and a partially completed boundary evolution model. Plans are in place to complete the rubble zone and cavity gas zone models, integrate them into a full-scope UCG simulator, test the simulator and its domain models against lab- and field-scale data, and apply it to current UCG projects. 13
An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment
6
Nitao et al., LLNL
Acknowledgements
This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under contract DE-AC52-07NA27344.
7
References
Biezen, E.N.J, Bruining, J, and Molenaar, J. An integrated 3D model for underground coal gasification, SPE Ann.Tech. Conf. and Exhibition, SPE Paper No. 30790-MS 1995. Britten, J.A. and Thorsness, C.B. “A model for cavity growth and resource recovery during underground coal gasification”, In Situ, 13 (1 & 2), 1-53, 1989. Buscheck, T.A, Hao, Y, Morris, J.P, and Burton, E.A. “Thermal-hydrological sensitivity analysis of underground coal gasification”, Proceedings of the 2009 International Pittsburgh Coal Conference, Pittsburgh PA, September 20-23, 2009. Camp, D.W, Krantz, W.B, and Gunn, R.D. A water-influx model for UCG with spalling-enhanced drying, Proc. 15th Intersociety Energy Conversion Engineering Conference, Seattle, WA August 18-22, 1980. Dinsmoor, B, Galland, J.M, Edgar, T.F. The modeling of cavity formation during underground coal, J. Petrol. Technol., 30, 695-704, 1978. Eddy, T.L, and Schwartz, S.H. A side wall burn model for cavity growth in underground coal gasification, J. Energy Resour. Technol., 105, 145-155 1983. Kuyper, R.A, van der Meer, Th.H, and Bruining, J. Simulation of underground gasification of thin coal seams, In Situ, 20, no. 3, 311-346 1996. Magnani, C.F. and Farouq Ali, S.M. A two-dimensional mathematical model of the underground coal gasification process, Fall Meeting of the SPE, SPE Paper No. 5653-MS, 1975. Morris, J.P., Rubin, M.B, Blair, S.C, Glenn, L.A, and Heuze, F.E. Simulations of underground structures subjected to dynamic loading using the Distinct Element Method, Engineering Computations, Vol. 21 (2/3/4), 384-408. 2003. Morris, J.P., Rubin, M.B, Block, G.I, and Bonner, B.P. Simulations of fracture and fragmentation of geologic materials using combined FEM/DEM analysis, International Journal of Impact Engineering, 33, 463-473. 2006. Morris, J.P., Buscheck, T.A, and Hao, Y. “Coupled geomechanical simulations of UCG cavity evolution”, Proceedings of the 2009 International Pittsburgh Coal Conference, Pittsburgh PA, September 20-23, 2009. Nitao, J.J. Reference manual for the NUFT flow and transport code, version 3.0, Lawrence Livermore National Laboratory, UCRL-MA-130651-REV-1, 1998. Perkins, G. and Sahajwalla, V. A mathematical model for the chemical reaction of a semi-infinite block of coal in underground coal gasification, Energy Fuels, 19, 1679–1692, 2005. Perkins, G. and Sahajwalla, V. Steady-state Model for estimating gas production from underground coal gasification, Energy & Fuels, 22, 3902-3914, 2008. Stephens, D.R. “The Hoe Creek experiments: LLNL’s underground coal gasification project in Wyoming”, Lawrence Livermore National Laboratory, Livermore, CA, UCRL-53211, 1981. Thorsness, C.B. and Britten, J.A. A mechanistic model for axisymmetric UCG cavity growth, Lauwrence Livermore National Laboratory, UCRL-94419 (NTIS no. DE87011570), 1986. Tsang, T.H.T. Modeling of Heat and Mass Transfer during Coal Block Gasification, Ph.D. Thesis, Univ. of Texas, Austin, Tx, 1980. Yang, L. Three-dimensional unstable non-linear numerical analysis of the underground coal gasification with free channel, Energy Sources, Part A, 28, 1519-1531, 2006. Vorobiev, O, Morris, J.P, Antoun, T, and Friedmann, S.J. “Geomechanical simulations related to UCG activities”, Proceedings of the 2008 International Pittsburgh Coal Conference, Pittsburgh PA, September 29-October 2, 2008. Westmoreland, P.R. and Forrester, R.C. III. “Pyrolysis of large coal blocks: implications of heat and mass transport effects for in situ gasification”, A.C.S. Division of Fuel Chemistry”, vol. 22, no. 1, p. 93, 1977. White, J.A. and Borja, R.I. “Stabilized low-order finite elements for coupled solid-deformation/fluid-diffusion and their application to fault zone transients”, Computer Methods in Applied Mechanics and Engineering, 197(49-50):4353–4366, 2008. White, J.A. “Stabilized finite element methods for coupled flow and geomechanics”, PhD thesis, Stanford University, Stanford, CA, 2009. White, J.A. and Borja, R.I. “Block-preconditioned Newton-Krylov solvers for fully-coupled flow and Geomechanics”, Computational Geosciences, in review, 2010. 14
Quantification of The Effects of Various Thermal Boundary Conditions in The Underground Coal Gasification Cavities using a Compartment Model Sateesh Daggupati1, Ramesh Naidu Mandapati1, Sanjay Mahajani1, Anuradda Ganesh2, Sapru R.K3, Sharma R.K.3, and Preeti Aghalayam1* 1
Department of Chemical Engineering, IIT Bombay, Powai, Mumbai-400076, India 2
Energy Systems Engineering, IIT Bombay, Powai, Mumbai-400076, India
3
UCG Group, IRS, ONGC, Chandkheda, Ahmedabad -380005, Gujarat-, India
*Corresponding Author: [email protected] Ph.No: +91(22)2576 7295, Fax: +91(22) 2572 6895
Abstract The UCG product gas can be used for electricity generation, or as a chemical feed stock, and gas turbine power generation combined with UCG is one of the promising ways of accomplishing clean coal utilization. In the UCG process, a cavity consisting of ash, char rubble and void space is formed and its size increases three dimensionally in a non-linear fashion. Operational control of the UCG process is difficult because of the several phenomena that are occurring simultaneously such as the detachment of coal from the cavity roof (i.e. spalling), water intrusion, chemical reactions, heat and mass transfer, and so on. These phenomena also lead to a complex flow distribution in the cavity. The characterization and quantification of this non-ideal flow field is necessary as it influences the performance of the UCG process. It is affected by several parameters such as the temperature gradients, inlet nozzle position and orientation, and coal properties such as thermal conductivity. The primary objective of this work is to study the effect of temperature gradients and various thermal boundary conditions on the reactant gas flow patterns in an underground cavity, through mathematical simulations. CFD simulations are performed for each case in order to get the flow pattern and residence time distribution curves. The effects of various thermal boundary conditions in the underground coal gasification cavities are quantified by performing the compartment modeling simulations independently. The results presented here may provide good insight of the UCG cavity under different scenarios of the UCG process.
Keywords: UCG, Cavity, RTD, CFD
1. Introduction It is important to characterize the flow patterns inside the cavity under two different situations i.e. thermo mechanical failure or spalling case and non-spalling case. Thermo mechanical failure is an important cavity phenomena and whether spalling would take place or not depends on the properties of coal and other geological aspects. There is a basic difference in two different situations as regards to the locale of reaction and temperature gradients inside the cavity. The non-spalling case the combustion reaction occurs at the cavity roof and the temperature at the roof is sufficiently higher than the floor whereas, in case of spalling, cavity floor is at sufficiently higher temperature than the roof. This is because of combustion of spalled particles occurs on the floor of the cavity. Underground coal gasification (UCG) involves the conversion of coal from the coal seam directly into a
1
combustible gas thus avoiding the cost of mining and transportation. The effect of radiation on flow pattern also needs to be considered while simulating the flow field. The flow pattern is strongly influenced by many operating and structural parameters such as inlet feed flow rate, temperature, spalled particles, radiation, positions and orientation of the inlet nozzle. CFD is the best tool to visualize, characterize and quantify the non-ideal flow behavior especially when the flow patterns are non-ideal as in the case of UCG. In our earlier work, for different sizes of UCG cavity, the development of a CFD based „compartment model‟ (i.e., a network of ideal flow reactors) were presented. In the first step, CFD simulations were performed for isothermal case on four different cavities of chosen size and shape, in order to determine flow patterns and corresponding residence time distribution (RTD). Based on the CFD results and considering RTD as the main criteria, a compartment model was proposed. In the next step, the parameters of the compartment model were estimated so that the RTD of the model is in good agreement with that obtained by CFD under similar conditions. Finally, the developed compartment model is further validated by comparing the homogeneous reaction- enabled steady state CFD simulation results with that obtained from the steady state flow sheet simulator (i.e. ASPEN PLUS). The proposed model is generic enough and is suitable for all the cavity sizes studied, with model parameters such as residence time, bypass ratio etc. changing with time.
The main objective of this work was to propose and validate the compartment modeling approach considering few specific cases. In the present work, we take one step ahead consider different possible scenarios and examine the effect of other operating conditions on the flow pattern and hence the compartment model parameters. The work is organized as follows: we first study the flow pattern in the case of non-isothermal conditions under various conditions such as different inlet feed temperatures, feed flow rates, positions and orientations of inlet nozzle. We further study the effect of spalling on flow pattern in two different approaches. The first one is by assuming the entire cavity floor is at sufficiently high temperature than roof and another approach demonstrates how the flow pattern is affected by the spalled particles (at high temperature) which lie on the floor of the cavity. The effect of all these parameters on the flow pattern is quantified with the help of validated compartment model. The following procedure is used to characterize the flow patterns under different specified conditions:
CFD simulations are performed on a selected cavity studied earlier, in order to determine the steady state flow patterns of the reactant gases under different boundary and physical conditions such as inlet temperatures, flow rates, spalling and non-spalling cases, position and orientations of inlet nozzle.
In the second step, steady state flow field obtained in the first step is frozen and a tracer is introduced as a pulse at the inlet and unsteady state model for the tracer balance is solved in order to obtain the RTD.
In order to quantify the effect of various parameters on the flow pattern, validated compartment model is taken and Matlab simulations are performed to estimate the model parameters.
2. CFD Simulations Geometry
2
First, the cavity of size 1 with vertical injection studied earlier is chosen to study the flow pattern under different situations. This geometry was chosen based on the data available in the literature (Dagguapti et al. 2010; Harloff 1983). Geometry is created for the cavity of vertical injection and converted into the discrete control volumes (i.e. meshed) by using commercial CFD meshing preprocessor, GAMBIT. The mesh used is unstructured with tetrahedral grid cells in the whole volume of the reactor. Finer meshing is used in the vicinity of inlet and outlet because the velocity gradients are stiff in these regions due to expansion and contraction. The grid used for this analysis is shown in figure 1. This is the optimized grid after the grid independency studies which has been detailed out in Daggupati et al. (2010). The dimensions of cavity used for the simulations are given in Table 1. It also indicates total number of cells of meshed geometry. Separate geometry was created for the spalling case by considering the spalled particles on the floor of the cavity.
Figure 1: Grid of the UCG cavity Table 1: Dimensions of the UCG cavities Distance between the injection and production wells (m)
9.1
Length of the cavity behind the inlet (backward direction) (m)
1
Length of the cavity dome from inlet (forward direction) (m)
2.5
Height of the cavity (m)
1.32
Width of the cavity (m) at the injection well
0.9
Width of the cavity (m) at the outflow channel
0.3
Width of the cavity (m) at the back side of injection well
0.25
Width of the cavity (m) at the outlet
0.2
Inlet & Outlet diameter (m)
0.1
3
Total volume, m
Grid cells (tetrahedral and hexagonal), Lakhs
4.0961 9,83,334
3
Input Specifications and Convergence criteria At the inlet a „velocity_inlet‟ boundary condition with a velocity of 4 m/s was specified, such that the feed coming in has a volumetric flow rate of 113.14 m3/hr in the injection well. A „pressure outlet‟ boundary condition at the outlet with the gauge pressure of 5 atm., which is taken from the field trail data base (Cena and Thorsness, 1981). For the non spalling case, presence of reaction was indirectly considered by the high temperature (a constant wall temperature, 1273 K) boundary condition at the roof and rest of the cavity walls were assumed to be at a constant heat flux of 25 W/(m2-K). In case of spalling the cavity bottom is considered at a constant wall temperature (i.e.1273 K) condition and other walls are assumed to be at a constant heat flux of 25 W/(m2-K). The realizable κ-ε model with standard wall functions is used in these computations to model the turbulence in the cavity. This scheme is suitable for flow involving recirculations (Shih et al. 1995; Fluent, 2007). The boundary conditions used for this study are given in Table 2. In order to study the effect of radiation on the flow patterns, P-1 radiation model is considered while simulating the flow field. P-1 model works reasonably well for combustion applications where the optical thickness is large. The major advantage of this model is it can be easily applied to complicated geometries with curvilinear coordinates like UCG cavities. It is easy to solve the model equation with a little CPU demand and with reasonable accuracy (Sazhin et al., 1995; Fluent, 2005). Based on the work of Kuyper et al., (1993) and Perkins and Sahajwala, (2007), the extinction coefficient for the gas is assumed to be constant 0.12 m-1 (absorption coefficient = 0.12 m-1 and also assumed that scatters all directions equally). Simulations using other radiation models (such as grey gas model or discrete ordinate model) for the absorption coefficient did not show any significant difference in the solution obtained by P-1 radiation model and a constant absorption coefficient.
The equations for the conservation of mass, momentum, energy and radiation (see equation 1, 2,3 and 4) in the turbulent regime for these meshed geometries were solved to obtain the flow patterns under the specified boundary conditions by using the computational software, FLUENT v6.3.26. The SIMPLE (Semi-Implicit Method for Pressure-Linked Equations) algorithm is used to resolve the pressure-velocity coupling and a steady state second order upwind accurate discretization model is used to solve the above equations. In case of steady state simulations, the solution was judged to be converged when the scalar residuals of each transport equation goes below 1×10-7. The mass flow difference between in and out was also verified, and which was of the order of 10-8. In the unsteady state simulations temporal accuracy for species balance equation is solved till the scalar residue of this equation reduces to less than 1×10-7. Four independent simulations are performed to study the effect of feed temperature, and another three independent simulations for the effect of flow rate. Further several simulations are also performed to examine the effect of spalling, position and orientation of the inlet nozzle. Table 2: Boundary conditions and other specifications used for CFD studies Inlet velocity Carrier fluid Outlet pressure, atm Viscous model
4 m/s oxygen 5 Realizable κ-ε
4
Continuity Equation
.v 0 t Momentum Equation
v .vv p .( ) g F t
...1
...2
Energy Balance Equation
E .vE p . h j J j t j
...3
P-1 Radiation Model
qr aG 4aT 4
...4
RTD Simulations Once steady state solution is obtained for the cavity under specified conditions, the flow field is frozen. In order to perform the virtual tracer experiments, a tracer, which has similar properties as the carrier fluid, is injected to the cavity in the pulse mode, i.e. at time =1 sec the feed consists of tracer mass fraction of 1.0. In order to minimize the diffusion effects, diffusivity of the tracer is taken as a sufficiently low value i.e. 1 x 10 -15 m2/s. The unsteady state model is solved for the governing species balance equations (Eq. 5) to obtain the concentration of the tracer. The RTD is calculated from the area weighted average of tracer concentration data against time, obtained at the outlet. Exit age distribution (E curve) data is generated for all the cases of four different feed temperatures, four different flow rates, spalling and radiation situations.
Yi .vYi .J i t
...5
2. Compartment Modeling In addition to CFD simulations, this work explores the utilization of the traditional compartment modeling technique to quantify the non-ideality of the flow patterns that exists in the UCG cavities. A validated compartment model has been proposed by Dagguapti et al. (2010) for the cavities at different times of the UCG process. The proposed compartment model (see Figure 2) consists of a series combination of five perfectly mixed reactors and the parallel stream consists of a series combination of five small perfectly mixed reactors followed by a plug flow reactor.
Formulation of the Compartment Model
In formulating the compartment model, unsteady state tracer balance equations for each reactor in the proposed combination of reactors may be written as follows:
5
Main stream reactors
dCm j ,i
(Cm j 1,i 1 Cm j ,i 1 )
... 6
m, j
Cm j ,i Cm j ,i 1 (h dCm j ) Parallel stream
... 7
reactors
dCbk
Cbk 1,i 1 Cbk ,i 1
... 8
bk
Cb j ,i Cb j ,i 1 (h dCb j )
... 9
for time = 1 sec, Cm0,1 Cb0,1 1
and
for time >1 sec, Cm0,i Cb0,i 0
j= 1 to 5 (no. of rectors in series)
and
i= 1 to operation time
Finally
Cout (1 b)Cm5 (b Cb5 )
... 10
Where, b is the ratio in which the inlet stream splits as indicated in Figure 2.
Figure 2: Proposed compartment model for UCG cavity The above set of unsteady state differential equations are solved using ode15s solver in the MATLAB to obtain the exit age distribution data. E curves of both CFD and compartment model are fitted by the NLINFIT optimization tool in Matlab, in order to estimate the model parameters. The optimizer used to find the parameters requires two inputs, firstly, initial guesses for all the parameters to be estimated, secondly, E curve data of the system to be dealt with over a certain period of time. These are then used to find an optimum parameter set representative of the system. These parameters are obtained for all the cases given in Table 3. The goodness of the fit was measured in terms of the average percentage error ( e ) which is defined as follows (Singh et al., 2007):
e(%)
1 n | Emod el ECFD | 100 E n i 1 CFD
... 11
The errors obtained in all the cases are given in Table 3 and Table 4. The percentage error is less than 3% in all the cases. The parity plot comparing the E (t) values predicted from both CFD and compartment model is shown in Figure 3, which indicates that more than 99 % of data lies within ±3 % of the ideal case.
Table 3: Compartment model parameters for effect of feed temperature and spalling
6
Effect of feed temperature
Model Parameters
473K
B VM1 (m3) VM2 to VM5 (m3) VB1 (m3) VB2 to VB5 (m3) VP m3 Total Volume % dead volume % error
0.189 1.399 0.174 0.007 0.005 0.280 2.40 41.36 1.258
673K
873K
0.047 2.035 0.217 0.007 0.007 0.350 3.29 19.73 2.423
Effect of spalling
1073K
0.143 2.238 0.248 0.072 0.021 0.380 3.77 8.06 1.85
0.153 2.378 0.251 0.075 0.025 0.400 3.96 3.40 2.67
RHT
BHT
SPHT
0.047 2.12 0.221 0.071 0.05 0.36 3.64 11.26 2.423
0.277 1.840 0.289 0.078 0.078 0.390 3.78 7.81 2.89
0.142 2.24 0.248 0.070 0.020 0.400 3.78 7.67 1.95
Table 4: Compartment model parameters for effect of feed flow rate and nozzle orientation Effect of feed flow rate, m/s
Model Parameters
2
B VM1 (m3) VM2 to VM5 (m3) VB1 (m3) VB2 to VB5 (m3) VP m3 Total Volume % dead volume % error
4
0.037 1.923 0.171 0.010 0.005 0.400 3.035 25.986 2.12
Effect of nozzle orientation
8
0.143 2.238 0.248 0.072 0.021 0.380 3.77 8.06 1.85
Vertical
0.272 2.147 0.297 0.177 0.059 0.250 4.000 2.438 1.12
Horizontal
0.143 2.238 0.248 0.072 0.021 0.380 3.77 8.06 1.85
0.163 2.525139 0.094577 0.349089 0.019488 0.17101 3.50 14.52 1.2
0.016
CFD
0.012
NO-HI
SP-RHT
FT-473K
SP-SPHT
FT-673K
SP-BHT
FT-873K
FT-1073K
FR-2 m_s
FR-8m_s
0.008
0.004
0 0
0.004
0.008
0.012
0.016
Compartment Model
Figure 3: Parity plot for E (t) 3. Results and Discussions Effect of Feed Temperature Inlet feed temperature makes significant impact on the efficiency of gasification in the UCG process. It is necessary to understand how the flow pattern is affected by changing the feed temperature. For this purpose, four temperatures are chosen, i.e. 473K, 673 K, 873 K and 1073 K to study flow pattern in a non-spalling case. By keeping other
7
boundary conditions constant, four independent simulations are performed to obtain the steady state flow pattern and the E curve. In order to quantify this effect, the parameters of compartment model are also estimated. Figure 4 shows the path lines of the injected tracer in the cavity for a chosen time step which clearly shows that at low temperatures the active volume is less, further it is increases with increasing the feed temperature. This is due to the fact that at higher feed temperature the overall cavity temperature increases, which results in a higher volumetric flow rate inside the cavity. Figure 5 shows the velocity contours on various vertical planes in the axial direction, which indicates that by increasing the feed temperature, velocity magnitude increases. Figure 6 shows the comparison of E curves corresponding to various feed temperatures. By increasing the feed temperature, the magnitude of the peak in the E curve decreases, which indicates that there is an increment in active volume. It is also observed that the initial time delay i.e. PFR-like behavior, increases with an increase in the feed temperature. The compartment model parameters are given in Table 3. It can be seen that volumes of all back mixed reactors in the main stream as well as in the parallel stream and the plug flow reactor volume increase with feed temperature which leads to reduction in the dead volume. It also shows that the bypass ratio is high in the case of feed temperature 473 K and decreases further with increase in feed temperature.
Figure 4: Comparison of various feed temperature path lines (colored by velocity magnitude, m/s) for a chosen time step
Figure 5: Comparison of various feed temperature velocity (magnitude m/s) contours
8
0.0125 Feed Temperature=473K Feed Temperature=673K 0.01
Feed Temperature=873K Feed Temperature=1073K
E(t)
0.0075
0.005
0.0025
0 0
200
400
600
Time (sec)
Figure 6: Comparison of RTD for various feed temperatures
Effect of Spalling Three independent CFD simulations were performed to study the effect of spalling in the UCG cavity. Two different situations are considered while simulating the spalling behavior. In the first case, cavity floor is assumed to be at high temperature than the roof. In this case, flow pattern is driven by the temperature gradient. In the second case, spalled particles (at high temperature) are considered at the cavity bottom, which cause obstruction to the flow and facilitate back mixing. In this case, flow pattern is affected by both the spalled particles and temperature gradient. In order to compare these results with non-spalling case, CFD simulations are also performed by assuming the cavity roof at high temperature (non-spalling case). All these simulations are performed under otherwise similar operating conditions such as feed flow rate, temperature, pressure etc. Figure 7 shows the comparison of temperature contours on different vertical planes along with the axial direction. It indicates that the temperature is well distributed in the cavity for non-spalling case compared to the spalling case. Figure 8 shows the velocity contours in the cavity for these three cases. It can be seen that the velocity magnitude is more in the case of spalling (i.e. bottom at high temperature) because the temperature gradient is high in this case (see figure 7). Figure 9 shows the comparison of path lines (colored by velocity magnitude, m/s) for both spalling and non-spalling cases for a chosen time step. It indicates that more mixing occurs for the spalling case 2, which is due to the obstruction to flow by the spalled particles which lies on the floor of the cavity. It also indicates that the active volume is more in the case of spalling and rigorous mixing can be seen in figure 9. The comparison of RTDs of spalling case and non-spalling case is shown in figure 10. The initial sharp peak is observed in the case of nonspalling case indicating that bypass ratio is more for non-spalling case. The magnitude of peak in the E curve is less for the spalling case which also suggests that active volume is more for spalling case. Table 3 gives the compartment model parameters, which indicates that the volume of mixed reactor in the main stream and plug flow reactor
9
volume are more for spalling case compared to that of the non-spalling case. It is also observed that the bypass ratio is high and % of dead volumes is more in the case of non-spalling.
Figure 7: Comparison of both spalling and non-spalling temperature (K) contours
Figure 8: Comparison of spalling and non-spalling velocity (magnitude m/s) contours 0.01
Non-Spalling- (Roof at High Temperature) 0.0075
Spalling (Bottom at High Temperature)
E(t)
Spalling Spalled particles at High Temperature
0.005
0.0025
0 0
200
400
600
Time (sec)
Figure 9: Comparison of both spalling and non-spalling path lines (colored by velocity magnitude, m/s) for a chosen step size
Figure 10: Comparison of RTD for both spalling and non-spalling case
Effect of Flow Rate Three different flow rates (i.e. 2 m/s, 4 m/s and 8 m/s) are chosen to study the effect of flow rate on the flow pattern. These simulations are performed for non-spalling case with feed temperature of 873 K. As expected, by increasing the flow rate, mixing increases in the cavity which leads to increase in the active volume as can be seen in figure 11. Table 4 gives the compartment model parameters, which suggests that the bypass ratio increases with flow rate which is expected results, at low flow rate most of the gas remains in the cavity for a sufficient long time which can
10
be seen in figure 11 and further bypass ratio increases with an increase in the flow rate. The volumes of mixed reactors in the main stream as well as parallel stream increase with an increase in the flow rate. Percentage of dead volume decreases with increasing flow rate (see Table 4). The comparison of RTDs of different feed flow rates are shown in figure 12. As expected, by increasing the flow rate, the lag becomes less because of less residence time in the case of high flow rates.
0.016
Flow rate = 8m_s Flow rate = 4m_s
E (t)
0.012
Flow rate = 2m_s
0.008
0.004
0 0
200
400
600
800
Time, sec
Figure 11: Comparison of various feed flow rate path lines (colored by velocity magnitude, m/s) for a chosen step size
Figure 12: Comparison of RTD for various feed flow rates
Effect of Nozzle Orientation Two different injections (i.e. vertical and horizontal) are considered to study the effect of inlet nozzle orientation on the flow pattern. Two independent simulations are performed under similar operating conditions (i.e. feed flow rate= 4 m/s and Feed temperature=873K, etc.) in the non-spalling case. Temperature contours on different vertical planes along the axial direction are shown in figure 13. This figure indicates that the temperature gradient is less in the case of horizontal injection. Figure 14 shows the path lines comparison of (colored by velocity magnitude, m/s) both the injections for a chosen step size. It indicates that by virtue of vertical injection, flow distribution takes places around the injection well. In case of horizontal injection, more mixing occurs in the vicinity of injection well in the bottom area, which is responsible for the less temperature gradient as most of the upper area is a stagnant region. Figure 15 shows the comparison of RTDs for both vertical and horizontal injections, which indicates that the magnitude of peak in the E curve is less for the vertical injection case. This means that the active volume is more for vertical injection case as compared to the one with horizontal injection. It is also observed that initial sharp peak is observed in the case of horizontal injection, indicating that bypass ratio is more for horizontal injection case. Table 4 gives the compartment model parameters, indicating that the bypass ratio is more for the horizontal injection and the active volume is more for the vertical injection.
11
Figure 13: Comparison of temperature (K) contours injection for both vertical and horizontal injections
Figure 14: Comparison of vertical and horizontal path lines for a chosen step size
0.01
0.0075
Hrizontal injection
E (t)
Vertical Injection 0.005
0.0025
0 0
200
400
600
Time (sec)
Figure 15: Comparison of RTD for both vertical and horizontal injections
4. Conclusions We consider different possible scenarios and examine the effect of various operating and structural parameters such as feed flow rate, feed temperature, spalling, orientation of inlet nozzle on the flow patterns. The average velocity magnitude in the cavity increases either with increase in the feed temperature or increase in the feed flow rate. This and the presence of spalled particles lead to a decrease in the temperature gradient inside the cavity thereby bringing uniformity in temperature. All the above said effects are quantified by performing the compartment modeling in each case. The sensitivity of compartment model parameters with respect to the changes in various conditions was studied. These results provide a good insight into the cavity under different conditions of a UCG process. Nomenclature
v p
Velocity tensor
Stress tensor
Static pressure
12
g
Gravitational body force
F
External body force
k eff
Effective thermal conductivity
Jj
Diffusion flux of species j
Yi
Mole fraction of species „i‟
σ
Stefan-Boltzmann constant
a
Absorption coefficient m-1
G
Incident radiation
qr
Radiation flux
V0
Volumetric flow rate, m3/s
B
Bypass ratio
VMj
Volume of jth CSTR in main stream of the compartment model, m3
VBj
Volume of jth CSTR in side stream of the compartment model, m3
VP
Volume of plug flow reactor in the compartment model, m3
τmj
Residence time of jth CSTR in main stream, sec
τbj
Residence time of jth CSTR in bypass stream, sec
τe
Residence time of PFR at the exit, sec
dCm j ,i
Derivative of tracer concentration of 'j'th CSTR in the main stream for 'i'th iteration
dCb ji
Derivative of tracer concentration of 'j'th CSTR in the bypass stream for 'i'th iteration
Cm j ,i
Tracer concentration in the 'j'th CSTR in the main stream for 'i'th iteration
Cb j ,i
Tracer concentration in the 'j'th CSTR in the bypass stream for 'i'th iteration
h
Time step size
e
Average percentage error
References 1.
Daggupati, S.; Mandapati, R.N.; Mahajani, S.M.; Ganesh, A.; Mathur, D.K.; Sharma, R.K.; Aghalayam, P.; Laboratory studies on combustion cavity growth in lignite coal blocks in the context of underground coal gasification. Energy 2010, 35, 6, 2374.
2.
Harloff G.J., Underground coal gasification cavity growth model. J Energy 1983;7:410.
3.
Daggupati S, Mandapati RN, Mahjani SM, Ganesh A, Aghalayam P. Compartment model for underground coal gasification. In: Proceedings of the 25th Annual International Pittsburg Coal Conference. Pittsburg,.USA, 2008.
13
4.
Daggupati, S., Mandapati, R., Mahajani, S.M., Ganesh, A., Sapru, R.K., Sharma, R.K. and Aghalayam, P., “Prediction of UCG performance by using Compartment model”, 26 th Annual International Pittsburgh Coal Conference, 2009.
5.
Cena RJ, Thorsness CB. Underground coal gasification data base, Report No.UCID-19169. U.S. DOE, Livermore, CA: Lawrence Livermore National Laboratory; 1981.
6.
Sazhin S.S., Sazhina E.M., Saravelou O.F. and Wild P. The P-1 model for the thermal radiation transfer: advantages and limitations, Fuel, 1996, 75(3), 289.
7.
Shih, Tsan-Hsing, Liou, W. W., Shabbir, A., Yang, Z., and Zhu, J. A new k−ε eddy viscosity model for high Reynolds number turbulent flows. Computers and Fluids, 1995, 24(3), 227.
8.
FLUENT 6.3. Users Guide. Fluent Inc. Lebanon NH; December 2005.
9.
Singh, K.K.; Mahajani, S.M.; Shenoy, K.T.; Ghosh, S.K.; Computational Fluid Dynamics Modeling of a bench scale pump-mixer: head, power and residence time distribution, Ind Eng Chem Res 2007, 46, 2180.
10. Kuyper, R.A., Van Deer Meer, H. and Hoogendoorn, C. J. “Turbulent Natural
Convection Flow Due to
Combined Buoyancy Forces during Underground Gasification of Thin Coal Layers”. Chemical Engineering Science. 1994, 49(6) 851. 11. Perkins, G.; Sahajwalla, V. Modeling of heat and mass transport phenomena and chemical reaction in underground coal gasification. Chemical Engineering Research Design 2007, 85, A3, 329.
14
Estimation of Chemical Reaction Occurred in Underground Coal Gasification(2) Mamoru Kaiho 1, Osamu Yamada 1*, Sohei Shimada 2, Shouji Fujioka 3, and Jie Liang 4 1
National Institute of Advanced Industrial Science and Technology 2
The University of Tokyo, 3Japan Coal Energy Center, and 4China University of Mining and Technology 1
16-1 Onogawa, Tsukuba, Ibaraki 305-8569 JAPAN
*Tel +81-29-861-8423
Fax +81-29-861-8417
E-mail [email protected]
Abstract: We have proposed a method to derive the gasification reaction formula, from the result of ultimate analysis and composition CHmOn + αO2 + βH2O -> γH2 + δCO + εCO2 + ηCH4, of gas produced.
Since the numerical expressions to determine the value of α to η in above reaction
formula were derived without any arbitrary assumption and approximation, we consider that the equations are generally used for the analysis of data obtained by any kind of gasification process. The characteristics of chemical reaction process of underground coal gasification (UCG) in Fushin coal mine in China was elucidated by our method as presented previously.
We have successfully
prepared the method to estimate the temperature of gas in UCG reactor based on the heat of reaction hr calculated from reaction formula by following equation ; hr = γhH2 + δhCO + ηhCH4 - hcoal where hH2 , hCO ,
hCH4 , and hcoal show molar heating value of H2, CO, CH4, and coal.
Since the wall of the reactor of UCG is made of coal, hr seems to be utilized effectively in pyrolysis and drying of coal.
The UCG reactor was regarded as adiabatic, therefore, hr agrees with
a total of sensible heat of gas exist in the reactor at that temperature. We compared the reaction processes and gasification temperature of UCG carried out at Xinghe mine, Liuzhuang mine, Ezhaung mine, Xiyang mine, and Fuxin mine. was 81.67, 82.66, 82.87, 92.34, and 79.70% respectively.
Carbon content of each coal
Average amount of α in each coal was
estimated as 0.333, 0.376, 0.477, 0.171, and 0.347 mol/mol respectively, and that of β was also estimated as 0.688, 0.399, 0.397, 0.717, and 0.247 mol/mol respectively. α and β seemed to correlate to the value of hcoal.
1. Introduction In recent years, drastic changes in development of underground coal gasification (UCG) have occurred by application of oil drilling technique. Several commercial processes have been scheduled. The chemical reaction process of UCG, however, has not been fully understood because of the lack of information on chemical reactions in the UCG reacting site such as temperature, pressure, gas composition, and so forth. We have found a theoretical method to derive chemical reaction formula from composition of gas produced. We analyzed chemical process, heat of gasification reaction, and gasification temperature of UCG carried out in Fuxin mine, Xinghe mine, Liuzhuang mine, Ezhuang mine, and Xiyang mine in China. Features of chemical process of UCG are discussed based on the theoretical method. 2.
Method for analysis of gasification reactions
2.1
Derivation of reaction formula
It is generally accepted that the chemical process of coal gasification should be interpreted based on reaction rate of each elemental reaction such as C+0.5O2→CO、C+CO2→2CO、C+ H2O→CO+H2, and so on.
It will be difficult, however, to determine contribution of more than ten elemental
reactions, even if we simplified the gasification reaction as in formula (1). CHmOn+ αO2 + βH2O → γH2 + δCO + εCO2 + ηCH4
(1)
Therefore, we tried to establish a theoretical method that determined the coefficients from α to η in eq.(1) based on stoichiometry of eq. (1) shown below. For C :
1 = δ +ε + η
For H :
m + 2β = 2γ + 4η
For O :
n + 2α + β = δ + 2ε
We proved that α~η in eq.(1) were expressed by using equations from (2) - (7), when we define p, q, r, and s as concentrations of H2, CO, CO2, and CH4 respectively. q +2r —p —2s 0.5m —n α = ──────── + ───── 2(q + r + s) 2 P + 2s β = ───── — 0.5m (q + r + s) p γ = ────── (q + r + s) q δ = ────── (q + r + s) r ε = ────── (q + r + s)
(2)
(3) (4) (5) (6)
s η = ────── (q + r + s)
(7)
2.2 Derivation of equations considering gasification mechanism In order to analyze the chemical process of eq. (1) obtained from composition of gas produced, we defined eq. (8) as a standard gasification formula, since it is completely free from arbitrary part because it is determined only by m and n in coal molecule. CHmOn + 0.5( 1— n )O2→0.5mH2 + CO
(8)
The difference between amounts of O2 in eqs. (1) and (8) is defined as Oex, as shown in eq. (9). Oex =α – 0.5( 1 – n )
(9)
In the case of Oex ≥ 0, amount of O2 in eq. (1) is greater than that in eq. (8), 2Oex mol of H2 and CO are consequently burned and same mol of H2O and CO2 are produced. If the amount of burned H2 is expressed by x, then the amount of CO burned will be descrived as (2Oex - x), and that of H2O and CO2 produced will be x and (2Oex - x), respectively.
Combustion caused by increase of O2
alters eq. (8) as a standard reaction formula and eq. (10) is obtained. CHmOn+{0.5( 1 – n )+ Oex }O2 → (0.5m – x )H2+ (2Oex - x)CO + xH2O+ ( 2Oex - x)CO2
(10)
In addition, considering y mol of shift reaction (CO+H2O→H2+CO2) andzmol of methane formation reaction (CO+3H2→CH4+H2O) convert eq. (10) into eq. (11). CHmOn +{0.5( 1 – n ) +Oex}O2 +(y – x –z )H2O → (0.5m + y – x – 3z)H2 +(1– 2Oex – y + x – z)CO+(2Oex + y – x )CO2 + zCH4
(11)
In the case of Oex < 0, shortage of O2 makes eq. (12) based on eq. (8). CHmOn+{0.5( 1 – n ) + Oex }O2 → 0.5mH2 +(1+2Oex)CO –2OexC
(12)
Here –2OexC shows the amount of residual carbon. -2Oex mol of H2O cause water gas reaction with –2OexC to avoid the generation of residual carbon, and eq. (12) is replaced by eq. (13). CHmOn+{0.5( 1 – n ) + Oex }O2 –2OexH2O → ( 0.5m –2Oex )H2 + CO
(13)
In addition, y and z mol of CO caused shift reaction and methane formation reaction respectively. Eq. (13) can be changed to eq. (14). CHmOn+{0.5( 1 – n ) + Oex}O2 + (– 2Oex +y – z )H2O → (0.5m – 2Oex + y – 3z)H2 +(1 – y – z )CO+yCO2 + zCH4 We can calculate Oex by eq. (9).
(14)
x is estimated as the product of volume of gas burned with Oex and
the proportion of yield of H2 to total gas, 2Oex×0.5m/(1+0.5m) = mOex/( 1+0.5m). by the comparison of eq. (1) with mathematical formula, eqs. (11) and (14).
y is determined
In the case of Oex < 0,
y = ε, is set up by comparison of eq. (1) with eq. (14).
In the case of Oex ≥ 0, y is estimated by eq.
(15) which is derived by putting the amount of CO2 in eq. (1). This is equal to that of (11). y = ε –2Oex +{ mOex/( 1+0.5m)}
(15)
z is determined as z = η, regardless of the signs of Oex. Contributions of combustion of product gas in eq. (1), water gas reaction, shift reaction, and methane formation proceeded quantitatively based on standard reaction formula as in eq. (8) and the values of Oex, x, y, and z.
2. 3
Estimation of heat of reaction and temperature of gasification
Heat of gasification reaction, hr (kcal/mol-coal) is calculated by eq. (16). (16)
hr=γhH2+δhCO+ηhCH4−hcoal
Here, hcoal, hH2, hCO, and hCH4 are molar combustion heats (kcal/mol) of coal, H2, CO, and CH4 respectively.
The values of hH2, hCO, and hCH4 are -68.32, -67.64, and -212.80 kcal/mol
respectively. In the case of UCG, hr is considered to be utilized for preheating, drying, and pyrolysis of coal effectively to allow the reactor regarded as adiabatic. Temperature of gas produced by eq. (1) may be estimated by assuming that most of hr is converted to the sensible heat of gas under adiabatic condition. Gas in the UCG reactor consists of products of gasification and H2O.
Amount of N2 per 1 mol of
carbon in coal, ζ, is calculated by following equation; ζ = (concentration of N2 in product gas)/(sum of concentration of CO, CO2, and CH4) The amount of H2O in product gas cannot be measured precisely by analytical instrument.
It is
difficult to predict the amount of H2O as result of the material balance of the gasification process, because unknown amount of groundwater was certainly mixed in the product gas.
We assumed,
therefore, that shift reaction attains equilibrium at gasification temperature and calculated the amount of H2O, θ, using the equilibrium constant of shift reaction, K, and concentrations of H2, CO, and CO2. Sensible heat of product gas, Qsh was calculated by the integration of thermal capacity of each component gas, which is given by following expression, Cp = a + bT +cT2 + dT3, from 298K to gasification temperature T. Firstly, we assumed gasification temperature T and determined the value of θ using γ、δ、and ε in eq. (1), and equilibrium constant of shift reaction, K, which is calculated by the thermodynamic
equation.
The Qsh at T in the reactor, which was calculated for γ ~ η in eq. (1), ζ, and θ, was
compared with hr of eq. (1).
Gasification temperature that generates Qsh equal to hr was obtained
by series of calculations performed for temperatures changed in system.
Estimation of gasification
temperature was carried out for each result of gas analysis obtained for five coals. 3.
Analysis of reactions in UCG Reaction formula of coal gasification, heat of reaction, and gasification temperature of UCG
carried out at Fuxin mine, Xinghe mine, Liuzhuang mine, Ezhuang mine, and Xiyang mine in China were estimated based on the data obtained from China University of Mining and Technology. The results of ultimate analysis of five coals were shown in Table 1 with their heating values.
Fuxin
coal、Xinghe coal、Liuzhuang coal、and Ezhuang coal were classified as bituminous coal and Xiyang coal was classified as anthracite.
Table 1
Results of ultimate analysis and data of heating values of coals Ultimate analysis
Coal
C
H
O
Heating N
S
(%-daf)
Value (MJ/kg)
Fuxin coal
79.70
5.35
12.91
1.36
0.68
23.85
Xinghe coal
81.67
5.57
9.08
1.39
1.70
27.59
Liuzhuang coal
82.66
5.63
8.85
1.47
1.09
27.82
Ezhuang coal
82.87
5.68
9.22
1.40
0.83
25.01
Xiyang coal
92.34
3.22
2.48
1.17
0.79
30.92
An example of analysis of chemical process of UCG, results for Fuxin mine are summarized in Tables 2 ~ 4. Composition of gases produced by UCG are listed in Table 2. The values of α ~ η in eq. (1) estimated from gas composition in Table 2 were shown in Table 3. also summarized in Table 3.
The values of Oex and y were
Table 2
Composition of gas produced by UCG in Fuxin mine Gas composition(%)
No.
O2
H2
CO
CO2
CH4
N2
1
1.01
3.8
11.6
6.8
6.5
59.60
2
0.80
5.6
14.8
8.2
6.2
60.30
3
0.60
10.2
16.8
6.7
4.5
48.70
4
0.70
10.5
17.5
6.8
5.5
62.60
5
0.90
10.3
19.6
4.3
4.8
52.60
6
0.60
11.0
17.5
7.4
5.0
57.47
7
0.60
8.0
17.5
7.8
4.2
63.80
8
0.50
6.4
16.7
8.4
4.7
56.40
Table 3
Coefficients in reaction formula of gasification
Coefficients of reaction formula and indicators of gasification (mol/mol) No.
α
Β
γ
δ
ε
η
Oex
y
1
0.311
0.271
0.152
0.466
0.273
0.261
-0.129
0.273
2
0.367
0.213
0.192
0.507
0.281
0.212
-0.073
0.281
3
0.337
0.283
0.364
0.600
0.239
0.161
-0.103
0.239
4
0.303
0.318
0.353
0.587
0.229
0.184
-0.137
0.229
5
0.286
0.290
0.359
0.683
0.149
0.167
-0.154
0.149
6
0.322
0.306
0.372
0.591
0.240
0.169
-0.118
0.240
7
0.424
0.152
0.271
0.593
0.264
0.142
-0.016
0.264
8
0.437
0.128
0.215
0.560
0.282
0.158
-0.003
0.282
Oex : Excess amount of O2 calculated by following equation Oex= α – 0.5( 1 – n ) y : mol of CO that caused shift reaction with H2O gasification agent
Gasification temperature estimated were shown in Table 4 together with the amount of N2, equilibrium constant K, the amount of H2O equilibrated with concentration of H2, CO, and CO2 at that temperature, heat of reaction, and value of H2O, θ+β which is considered to be the amount of ground water penetrated in the UCG reactor.
Table 4 No.
N2
K
Heat of reaction and gasification temperature Equil. H2O
(mol/mol)
4.
hr
Total H2O
Gasification
θ+β
Temperature
(mol/mol)
(kcal/mol)
(mol/mol)
(℃)
1
1.156
0.7721
0.1153
-18.45
0.386
901
2
1.364
0.5633
0.1889
-23.37
0.402
1013
3
1.253
1.2684
0.1143
-16.19
0.397
759
4
1.126
1.9875
0.0693
-12.92
0.387
657
5
1.063
3.9375
0.0199
-9.63
0.310
535
6
1.197
1.5967
0.0946
-14.54
0.401
704
7
1.576
0.4914
0.2455
-27.06
0.398
1069
8
1.624
0.4133
0.2620
-29.71
0.390
1147
Reaction mechanism of UCG
4.1 Analysis of reactions of UCG Chemical reaction of UCG were analyzed using results of Fuxin, Xinghe, Liuzhuang, Ezhuang, and Xiyang mines. Average compositions of gases produced by five UCG processes were shown in Table 5.
Table 5
Average compositions of gases produced by UCG Gas composition
(%)
Coal
O2
H2
CO
CO2
CH4
N2
Fuxin
0.71
8.28
16.50
7.01
5.18
57.68
Xinghe
-
21.72
8.82
13.59
2.64
53.36
Liuzhuang
1.49
14.77
13.98
10.73
3.69
55.31
Ezhuang
0.00
16.42
5.83
28.47
9.48
39.80
Xiyang
0.20
9.30
9.20
9.20
7.20
64.70
The mean value of each coefficient of reaction formula obtained for UCG operation in five coal mines were shown in Table 6.
Table 6
The average coefficients of reaction formula of gasification
The average coefficients of reaction formula and indicators of gasification (mol/mol) Coal
Α
Β
γ
δ
ε
η
Oex
y
Fuxin
0.347
0.247
0.287
0.573
0.245
0.182
-0.093
0.347
Xinghe
0.333
0.688
0.882
0.348
0.545
0.108
-0.126
0.333
Liuzhuang
0.376
0.399
0.537
0.499
0.365
0.135
-0.084
0.376
Ezhuang
0.477
0.397
0.375
0.133
0.650
0.217
0.019
0.477
Xiyang
0.171
0.717
0.363
0.359
0.359
0.281
-0.319
0.171
4. 2
Relationship between heat of reaction and quantity of O2 reacted Chemical process of UCG seems to be practically governed by the value of Oex.
relationship between heat of reaction, hr, and the amount of O2, α, is shown in Fig. 1.
The
Reaction
state of UCG seems to be practically dominated by quantity of O2 reacted. In agreement with conventional way in thermodynamics, we employed the minus sign to present exothermic reactions.
30 Xinghe Xiyan Fuxin Ezhuang Liuzhuang
20 Xinghe
10 hr (kcal/mol)
Fuxin(red)
0 Xiyan
-10 Ezhuang(purple)
-20 -30 -40
Liuzhuang
-50 0.0
0.1
0.2
0.3 0.4 α(mol/mol)
0.5
Fig. 1
Relationship between α and hr
0.6
0.7
Theoretical amount of O2 required for combustion is defined by ζ = 1+ 0.25m - 0.5n. Both sides of equations for C, H, and O appeared in section 2.1, which is presenting the elemental balance for eq. (1), are summed according to 2(C) + 0.5(H) – (O), then we have eq. (17). 2 (C)
:
0.5(H)
:
+ ) - (O) ――――――――― 2(C) + 0.5(H) – (O)
:
2 (1 = δ + ε + η ) 0.5( m + 2β = 2γ + 4η ) +) - (n + 2α +β = δ + 2ε) ――――――――――――――――― ( 2 + 0.5m – n ) – 2α = (γ + δ) + 4η
2ζ– 2α = (γ + δ) + 4η
(17)
Assuming that η is equal to 0, total amount of H2 and CO, (γ + δ), is equal to 2ζ– 2α.
On the other
hand, since the value of hH2 (-68.32kcal/mol) is very close to that of hco (-67.64kcal/mol), hr can be approximately expressed by the following equation, hr = hcoal +68.32(2ζ– 2α).
Therefore, the
relationship between α and hr in Fig. 1 is found to be linear. Changes of hr and β with elapse of time in the results obtained at five coal mine are shown in Figs. 2 ~ 6.
Fuxin 40
1.6
hr β
30
1.4
20
hr (kcal/mol)
1.0
0 hr
-10
0.8
-20
0.6
-30
0.4
β
-40
β(mol/mol)
1.2
10
0.2
-50 -60
0.0 1
3
5
7
9
11
13
15
17
19
21
23
day
Changes in hr and β with elapse of time at Fuxin mine
Fig. 2
Xinghe 40
1.6
hr β
30
1.4
20 hr (kcal/mol)
1.0
0 -10
0.8
β
-20
0.6
-30
β(mol/mol)
1.2
10
0.4
-40
hr
0.2
-50 -60
0.0 1
3
5
7
9
11
13
15
17
19
21
23
day
Fig. 3
Changes in hr and β with elapse of time at Xinghe mine
Liuzhuang 40 hr
20
1.6
hr β
30
1.4
hr (kcal/mol)
0
1.0
-10
0.8
-20
0.6
-30
β(mol/mol)
1.2
10
0.4
-40
β
0.2
-50 -60
0.0 1
3
5
7
9
11
13
15
17
19
21
23
day
Changes in hr and β with elapse of time at Liuzhuang mine
Fig. 4
Ezhuang 40
1.6
hr β
30
1.4
20
hr (kcal/mol)
0
1.0
-10
0.8
-20
0.6
-30
β
β(mol/mol)
1.2
10
0.4
-40 hr
-50
0.2
-60
0.0 1
3
5
7
9
11
13
15
17
19
21
23
day
Changes in hr and β with elapse of time at Ezhuang mine
Fig. 5
Xiyan 40
1.6
hr β
30
1.4
20
0
1.0
-10
0.8
-20
0.6
β
-30
β(mol/mol)
hr (kcal/mol)
1.2
hr
10
0.4
-40
0.2
-50 -60
0.0 1
4
7
10
13
16
19
22
25
28
day
Fig. 6
Changes in hr and β with elapse of time at Xiyang mine
In general, but not always, increasing hr gives increase in β.
In spite of constant rate of air supply,
hr repeatedly increase and decrease and it occasionally shows positive value, that means the endothermic reaction process is presumably proceeded.
We assume the fluctuation of hr is caused
by following mechanism. Air fed to the UCG reactor is considered to burn coal at some hot spot in the reactor and some hollows may be formed in coal seam. exothermic reaction of coal and O2.
The wall of hollow accumulates the heat produced by
When temperature of the wall become hot enough to cause
water gas reaction (C+H2O→H2+CO), consumption of carbon in the wall is proceeded not only by combustion but also by the endothermic reaction described above with the expansion of the hollow. It increase β and hr, occasionally, hr turn out to positive and endothermic gasification proceed. On the other hand, the amount of O2 fed per unit area of the wall decrease with expansion of the hollow.
It lowers supply of heat to the wall by the reduction of exothermic reactions between
carbon in the wall and O2 per unit wall area.
Since water gas reaction ceased by a drop in
temperature, it also decreases β and hr. Temperature of wall becomes cold enough to terminate gasification reactions in total, combustion begins at some hot spots remained on the wall, and new hollow is formed and growing. be caused by such way.
The periodic fluctuation of β and hr shown in Figs. 2 ∼ 6 seem to
This features consists an essential part of actual UCG reactions.
In the case of hr ≥ 0, gasification temperature will be estimated to be lower than room temperature. Such low temperature, however, gasification can never occurr.
It is considered that the practical
temperature may be kept at high enough to maintain proceed of water gas reaction at a sufficiently rapid rate by supply of heat accumulated in the wall. In the case of UCG, there is a considerable possibility of disagreement between Ts, estimated from hr, and actual gasification temperature. 4. 3
Variation of each component of gas
Relationship between α and quantity of each gas component was investigated using data of UCG carried out at Fuxin mine, Xinghe mine, Liuzhuang mine, Ezhuang mine, and Xiyang mine to elucidate the chemical reaction mechanism. Fig. 7 shows the relationships between α and δ (left hand), and and β, γ, and η based on the data acquired during the operation of UCG for 8 days. against α gave a straight line.
A plot of amount of each component gas, γ to η,
As can be seen from Table 3, Oex was lower than zero.
This means
that formation of CO2 could be responsible for shift reaction, since combustion of CO does not occur in the case of Oex < 0.
Fuxin
1.5
Fuxin
1.5
β(H2O) γ(H2) η(CH4)
1.1
0.7 δ(CO) ε(CO2)
0.3
1.1
yields of gas (mol/mol)
yields of gas (mol/mol)
δ(CO) ε(CO2)
0.7
γ(H2)
β(H2O)
0.3
η(CH4)
-0.1
-0.1 0.0
0.1
0.2
Fig. 7
0.3 0.4 α(mol/mol)
0.5
0.6
0.0
0.7
0.1
0.2
0.3 0.4 α(mol/mol)
0.5
0.7
Relationship between amount of gas component and α obtained at Fuxin mine
Xinghe
1.5
Xinghe
1.5
β(H2O) γ(H2) η(CH4)
δ(CO) ε(CO2)
1.1
yields of gas (mol/mol)
yields of gas (mol/mol)
0.6
0.7
0.3
1.1
0.7
γ(H2) β(H2O)
0.3
η(CH4)
-0.1 0.0
0.1
Fig. 8
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
-0.1
0.7
0.0
0.1
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
0.7
Relationship between amount of gas component and α obtained at Xinghe mine.
Fig. 8 shows the relationship between quantity of component gas and α obtained at Xinghe mine. All the plots except for three points are found to be 0.3≤α≤0.5 and the average Oex is -0.126. It is considered that shift reaction may proceed since quantity of CO2, ε, is more than that of CO, δ. Average value of y , molar amount of CO caused shift reaction, is calculated to be 0.541 mol/mol. Liuzhuang
1.1 δ(CO) ε(CO2) δ(CO)
0.7
0.3
Liuzhuang
1.5
yields of gas (mol/mol)
yieds of gas (mol/mol)
1.5
β(H2O) γ(H2) η(CH4)
1.1
γ(H2)
0.7
β(H2O) 0.3
ε(CO2)
η(CH4) -0.1
-0.1 0.0
0.1
Fig. 9
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
0.7
0.0
0.1
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
Relationship between amount of gas component and α obtained at Liuzhuang mine
0.7
Fig. 9 shows results in analysis of UCG carried out at Liuzhuang mine. Periodical change of hr with elapse of time is observed as shown in Fig. 4, and hr showed rise to positive value twice. Since the average value of Oex is -0.084, CO2 could be generated by shift reaction. In Fig. 10, results in analysis of UCG carried out at Ezhuang mine are shown. relatively high, and average value of Oex is calculated to be 0.019.
Values of α are
Since Oex is positive, amounts of
H2, and CO decreased as a result of combustion of these gases caused by Oex of O2. The production of CH4 is, however, held at 0.157 to 0.293 mol/mol-coal, which is considered to be relatively high yield.
Ezhuang
1.5
Ezhuang
1.5
yields of gas (mol/mol)
yields of gas (mol/mol)
δ(CO) ε(CO2)
1.1
0.7 ε(CO2)
0.3
β(H2O) γ(H2) η(CH4)
1.1
β(H2O)
0.7
γ(H2) η(CH4)
0.3
δ(CO)
-0.1
-0.1 0.0
0.1
Fig.10
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
0.0
0.7
0.1
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
0.7
Relationship between amount of gas component and α obtained at Ezhuang mine
Fig. 11 shows results in analysis of UCG carried out at Xiyang mine.
Values of α are
concentrated in the region of ca 0.1 to 0.3 mol/mol-coal, which is remarkably lower than the values observed in other four mines. only by shift reaction.
Since the mean value of Oex is -0.319, CO2 seems to be produced
Xiyang mine produces anthracite because, from the ultimate analysis, its
carbon content is 92.34%. Reactivity for gasification is nevertheless large enough to carry out UCG with relatively small amount of O2. Average η of Xiyang mine is found to be 0.281mol/mol, and it is much larger than those of Fuxin, Xinghe, Liuzhuang, and Ezhuang mines as shown in Table 3.
Large value of η may be correlated with lower yields of γ and δ. Xiyan
Xiyan
1.5
β(H2O) γ(H2) η(CH4)
δ(CO) ε(CO2)
1.1
yields of gas (mol/mol)
yields of gas (mol/mol)
1.5
0.7 ε(CO2) δ(CO)
0.3
1.1
0.7
β(H2O) γ(H2)
0.3
η(CH4)
-0.1
-0.1 0.0
0.1
Fig. 11
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
0.7
0.0
0.1
0.2
0.3 0.4 α(mol/mol)
0.5
0.6
Relationship between amount of gas component and α obtained at Xiyang mine
0.7
4. 3
Reaction mechanism of CH4 formation
Most of H atoms in Xiyang coal molecule are presumably combined with C atoms at the rim of aromatic rings considering low m (H/C) value. Very limited numbers of methyl group or hydrocarbon chains exist in the molecule.
This means that the resulting CH4 is produced by
methanation reaction (CO+3H2→CH4+H2O) or hydrogasification reaction (C+2H2→CH4), not by pyrolysis of coal. Fig. 12 shows the relationship between the yield of methane and gasification temperature. According to the principle of chemical equilibrium, yield of CH4 increases with lower gasification temperature, and the maximal yield of CH4 obtained at Xiyang coal mine.
Since the both methane
formation reactions are considered to be enhanced by surface catalytic activity of coal, char or ash, such tendency of CH4 yields against temperature obtained at five coal mines having varieties of coal rank might be influenced by chemical structure of char or properties of metal compounds in ash which differ from a coal to another.
0.4 Xnghe Liuzhuang Xiyan Fuxin Ezhuang
Xiyan
η(mol/mol)
0.3
Ezhuang
0
0.2 Fuxin Liuzhuang
0.1 Xinghe 0.0 0
500
1000
1500
Ts (℃)
Fig. 12
5.
Relationship between η and gasification temperature
Conclusions Actual operation data of UCG carried out in Fuxin mine, Xinghe mine, Liuzhuang mine,
Ezhuang mine, and Xiyang mine in China were investigated using an analytical method proposed by the authors at National Institute of Advanced Industrial Science and Technology and estimated the amounts of O2 reacted, H2O reacted, H2, CO, CO2, and CH4 produced, progress of shift reaction, heat of gasification reaction, and gasification temperature. Heat of gasification reaction showed a repetition of irregular rise and fall with elapse of time. It sometimes have positive value, which indicates endothermic gasification is proceeding with the aid
of excessive heat stored in the wall of UCG reactor during the period of exothermic combustion. The amount of reacted H2O also showed a similar repetition. When we plot yields of H2, CO and CO2 against the amount of O2 reacted a linear relationship is found.
CH4 is produced even for Xiyang coal that is classified as anthracite.
This indicates that
CH4 is produced not only by pyrolysis but also by methanation reaction or hydrogasification reaction.
COMPUTATIONAL FLOW MODELING OF UNDERGROUND COAL GASIFICATION (UCG) PROCESS Sateesh Daggupati1, Ramesh Naidu Mandapati1, Sanjay Mahajani1, Anuradda Ganesh2, Sapru R.K3, Sharma R.K.3, and Preeti Aghalayam1* 1
Department of Chemical Engineering, IIT Bombay, Powai, Mumbai-400076, India 2
Energy Systems Engineering, IIT Bombay, Powai, Mumbai-400076, India
3
UCG Group, IRS, ONGC, Chandkheda, Ahmedabad -380005, Gujarat-, India
*Corresponding Author: [email protected], Ph.No: +91(22)2576 7295, Fax: +91(22) 2572 6895
Abstract
Underground coal gasification (UCG) is a technique which permits access to coal which either lies too deep underground, or is otherwise too costly to exploit using conventional mining techniques. At the same time, it eliminates many of the health, safety and environmental problems of deep mining of coal. An irregular shape cavity is formed in the coal seam when coal is converted to gaseous products and its volume increases progressively as the coal is consumed. The complexity involved in modeling UCG process thus compels one to adopt a rigorous modeling approach that calls for use of computational fluid dynamics (CFD), which solves all balance equations simultaneously on a high speed computer. The simulation tool developed in the present work is capable of simultaneously predicting temperature distribution in the coal seam and profiles of velocity, temperature and species inside the cavity of a given size and shape. Further, with the help of this simulator we study the effect of various inlet conditions such as steam to oxygen ratio, feed temperature etc., on the product gas compositions. Ultimately, this work would help one to obtain the optimum conditions to produce product gas of high calorific value for a given cavity along with the specified inlet and boundary conditions. A broader objective of this simulation work is to track the growth of cavity and the associated changes in the UCG performance.
Keywords: UCG, Cavity, RTD, CFD
1. Introduction Underground coal gasification (UCG) is the process, in which the recovery of energetic and chemical content of coal taking place without mining (in-situ). UCG is the attractive method of producing syngas from coal. The syngas is a potential feedstock for the production of methanol and is a substitute for natural gas at prices competitive with existing energy sources. Moreover, UCG product gas can also be used as a fuel source for power generation 1. Developing an understanding of UCG process can be a formidable challenge as the process involves simultaneous interactions of several complex phenomena such as chemical reactions (both homogenous and heterogeneous), complex flow patterns of reactant gases, heat and mass transfer to the roof of the cavity and inside the coal seam, water influx from surrounding aquifers and thermo-mechanics of the coal and surrounding rock strata2.
1
The primary goal of the research on in-situ gasification is to find the near optimal operational parameters and identify significant safety issues before going into the actual operation / field test. The theoretical studies on UCG process are necessary as field trails are highly expensive, time consuming. Also these studies are unable to provide detailed information on the fundamental processes occurring in a burning coal seam underground. Mathematical modeling is the only tool which can address several issues such as process feasibility, design and prediction of performance for any field scale operation3. The design of a field test would greatly benefit from a simple dedicated simulator. The model must be able to describe the development from the early stages of gasification to a fully developed gasifier/cavity.
UCG models of differing levels of complexity have been reported in the literature. More complex models often give better prediction compared to the simple models; however, the computational time increases considerably. Generally, two main approaches have been considered for modeling: the packed bed model and channel model. The former assumes that coal gasification occurs in highly permeable porous media with a stationary coal bed 3. To obtain reliable results, it is necessary to use fine grids or adaptive grid refinement in the vicinity of the front, which causes some limitation for field applications. In addition, these methods cannot clearly model the cavity growth. The latter approach assumes that, during UCG, a permeable channel or a cylindrical cavity is expanding, in which the gasification occurred in the channel or cylinder wall. The process model developed by Britten and Thorsness 4, describes cavity growth and gas production during the UCG process. The basic assumption in this two-dimensional model is that the oxidant injection is maintained low in the coal seam and rubble detaches from the coal wall or roof and fall over the injection point, thus offering resistance to the flow of the injected stream. Other important assumptions made for the modeling are that the cavity is axisymmetric about the injection point and void space at the cavity top is well-mixed. The important factors responsible for cavity growth, such as water influx, cavity surface recession rate, flow dispersion of injected stream in the rubble bed, thermal degradation, gasification reactions and chemical attack on the coal wall, are incorporated through various sub-models. This model is applicable to flat-seam UCG of sub bituminous or lower rank coals in which the oxidant injection point remains low in the coal seam and might not be applicable for thinner seams. Perkins and Sahajwalla5 have developed a CFD based model capable of representing the UCG as a channel gasifier is presented. The model includes coupling of the heat and mass transfer and chemical reactions occurring at the boundary of the cavity. The performance characteristics of underground coal gasification in the channel mode are examined with reference to Newman Spinney P5 field trial. These model predictions are more sensitive to process changes, such as expanding cavity shape, than is observed during the field trials. They suggested that extension of the model to enable dynamic prediction of cavity growth and to account for an ash rubble pile is expected to improve the match between predicted and measured performance. Pirard et al.6 have developed a more comprehensive model, for studying the transport phenomena in UCG cavities based on a finite volume discretization of the Navier-Stokes equations and the k-ε turbulence model. Depending on the operating conditions, the fluid flow was dominated by either buoyancy due to temperature or concentration
2
gradients. The predicted composition and heating value of the product gas were in reasonably good agreement with the experimental data from the Price town field test. Biezen et al.7 have developed a three-dimensional model which incorporates natural convection driven coal surface gasification and permeable bed gasification together with temperature and stress induced failures of coal and rock. Nourozieh et al.8 developed a three dimensional model based on porous media approach by using CRIP configuration. This approach combines the effects of heat, mass transport, and chemical reactions to achieve this goal. CFD simulations were performed by using CMG software STARS by considering the averaged properties of all coal and clay stone layers. The model is capable of predicting the cavity formation, temperature profile, and gas composition for deep coal seams.
The under ground coal gasification cavity has an irregular three dimensional geometry. The shape and rate of growth of this cavity will strongly impact other important phenomena, such as reactant gas flow patterns, kinetics, temperature profiles, and so on. In our earlier work11 on compartment modeling for flow characterization in UCG cavity work, we had demonstrated that the flow patterns are markedly impacted by the injection method (vertical or horizontal) and ultimately affects the performance of UCG process. In all the earlier mathematical models, threedimensional cavity with vertical injection has not been considered as a input for predicting the performance of UCG process. At the same time it can be concluded that, computational fluid dynamics (CFD) is the only way to understand the complex UCG process in detail. There is also not much work evident in open literature on the three dimensional CFD modeling for the underground cavity reactor with vertical injection of the feed gases.
In the present work, a complete CFD based model has been developed, that can be used as a real time tool to assist in the operation and control of UCG process. The primary objective of the developed CFD model of UCG gasification is to study the impact of changing the parameters on the product gas composition and at the same time it is capable of simultaneously predicting temperature distribution in the coal seam and profiles of velocity, temperature and species inside the cavity of a given size and shape. Because of incorporation of many simultaneous processes it computationally expensive exercise. Nevertheless, the results obtained in this case would serves as a benchmark for tallying the outcomes of the computationally simplified models 11 being developed for UCG.
2. Model Description Assumptions The schematic of the geometry considered for simulation is shown in Figure 1. The key assumptions of the model are as follows: Coal seam is represented as a solid layer with stationary boundary and providing the reaction surface area for reacting the feed gas mixtures. For simplicity, coal is assumed to be pure carbon.
3
Heat transfer from cavity roof to the gas in cavity is governed by both convection and radiation and heat transfer in the coal seam is through conduction.. Both the heterogeneous reactions i.e. combustion with oxygen and gasification with steam of carbon are considered and the rate of each can be expressed by Arrhenius type equation. The major physical properties or thermodynamic parameters, such as heat conduction coefficient, specific heat, and heat exchange coefficient, are assumed to be the functions of temperature.
Input specification and boundary conditions Based on the data available in literature (Daggupati et al.,9 ; Harloff10 ), a cavity size along with the coal seam is chosen in which the dimensions are given in Table 1. Different views of the geometry used for the CFD studies are shown in figure 1. The quality of grid affects the rate of convergence, accuracy of solution and CPU time required. Mesh quality depends on the number of cells, aspect ratio and skewness factor. The mesh density should be high enough to capture all the process features. Optimum grid size is obtained by grid independency test which shows that the average cell size/ spacing of 3 cm is sufficient to obtain the grid independent solution. The basis of the selection of grid size is discussed elsewhere in Daggupati et al. 11. The grid consists of 1164487 unstructured tetrahedral grid cells in the whole volume. The gas flow inlet to the cavity was defined as 'mass flow_inlet' and the outlet was specified as 'pressure_outlet'. At the inlet, feed mixture contains oxygen and steam with different mole ratios at 473 K. The boundary conditions used for this study are given in Table 2.
Heterogeneous reactions (equations 1 and 2) are assumed to be taking place at the cavity roof i.e. interface between the coal seam and cavity. Rest of the cavity walls were assumed to be at a constant heat flux of 25 W/ (m 2-K). In order to study the effect of radiation on flow pattern, the emissivity = 0.8 was specified to the wall for accounting the radiative heat transfer (Kuyper et al.12 ). P-1 radiations model is activated as it is considered to be more suitable for combustion systems having thick optical thickness >1 (Sazhin et al., 13; Fluent,14). The realizable κ-ε model with standard wall functions is used in these computations to model the turbulence in the cavity. This scheme is suitable for flow involving re-circulations (Shih et al.,15; Fluent,14 ). In the UCG process, the combustion and gasification of coal with steam (see 1 and 2) are important and the rates of these reactions determine the economics of the UCG process. In addition to the continuity and momentum balance equations, the reaction enabled species balance is also invoked to get the product gas composition for the gasification with oxygen and steam reactions.
Under turbulent flow conditions the homogeneous reaction rates may be limited by the turbulent mixing of the reactants and products. In this case, the eddy-dissipation model is used in combination with the finite rate expression given by equation (7) to determine the overall reaction rate. The reaction kinetics given by equations 3 and 4 is incorporated (Perkins and Sahajwalla,5). The mass, momentum, energy and species balance equations (see equations 5 to 9) for the above specified cases of meshed geometries are solved. SIMPLE algorithm was used for pressure– velocity coupling and second order upwind scheme was used for discretization of all the transport equations for the computational domain. Simulations were considered converged when the scaled residuals for each transport
4
equation were below 1×10−7. Most simulations required about 15000–20000 iterations for convergence. This procedure required 400 to 500 CPU hrs of calculation on a Pentium Core 2 Duo 2.67GHz Work station. A total of twelve CFD Simulations are performed in a chosen cavity and the results are discussed in this section. The section is divided in two parts- the first part covers studies on only combustion whereas the second part deals with simultaneous combustion and gasification. k1
C O2 CO 2 k2
C H 2 O CO H 2
Where,
H 393.8MJ / kmol
...1
H 131.4MJ / kmol
....2
kJ 179.4 mol k1 2.503 10 T exp RT 17
kJ 231 mol k 2 8.593 T 0.5 exp RT
Continuity Equation Momentum Equation
.v 0 t v .vv p .( ) g F t
....3
....4
....5 ....6
Energy Balance Equation
E .vE p . h j J j t j
....7
P-1 Radiation Model
qr aG 4aT 4
....8
Species Balance Equation
Yi .vYi .J i Ri t
....9
Figure 1: Grid used for the CFD studies
5
Table 1: Dimensions of the geometry Distance between the injection and production wells (m)
9.1
Length of the cavity behind the inlet (backward direction) (m)
1.2
Length of the cavity dome from inlet (forward direction) (m)
2.5
Height of the cavity (m)
1.5
Width of the cavity (m) at the injection well
1.1
Width of the cavity (m) at the outflow channel
0.5
Width of the cavity (m) at the back side of injection well
0.45
Width of the cavity (m) at the outlet
0.4
Coal layer thickness, m
0.2
Inlet & Outlet diameter (m)
0.1
Grid cells (tetrahedral and hexagonal)
1164487
Table 2: Boundary conditions and other specifications used for CFD studies 31.66
Combustion (Oxygen)
63.32 1:0.5
Inlet mass flow rate, kg/hr Gasification (Oxygen : Steam)
40.56
1:1
49.47
1:2
67.28
1:3
85.1
Outlet pressure, atm
5
Inlet temperature, K
473
Viscous model Cavity roof Cavity other walls, Heat flux, W/m-K
Realizable κ-ε Wall surface reaction 25
Radiation model
P-1
Emissivity of coal wall
0.8
3. Studies on Combustion Reaction Effect of Feed flow rate For these simulations, pure oxygen is introduced to the cavity at the inlet at a temperature of 473 K. It is observed that, by increasing the flow rate, combustion reaction rate increases, indicating that the reaction is controlled by
6
mass transfer of oxygen from cavity to the roof wall. The temperature contours in the cavity can be seen in figure 2 and reaction rate contours are shown in figure 3. As expected the coal seam temperature increases with increase in feed flow rate due to more heat generation at the roof of the cavity. It is also observed that, by increasing the flow rate, more heat dissipation takes place to the coal seam (see figure 2). The contours of the concentration of CO 2 on different vertical planes in the axial directions are shown in figure 4. They are in line with the conversion rates realized at different flow rates.
Figure 2: Effect of flow rate on cavity temperature under combustion (Feed at 473K)
Figure 4:: Effect of flow rate on CO2 mole fraction (Feed at 473K)
Figure 3: Effect of flow rate on combustion reaction rate (Feed at 473K)
Figure 5: Effect of feed temperature on CO2 mole fraction (flow rate at 31.66 kg/hr)
Effect of Feed Temperature Two independent simulations are performed to study the effect of feed temperature on the gasification of coal with the oxygen. The feed flow rate is considered as 31.66 kg/ hr at two different temperatures i.e. 473 K and 673 K for these simulations. Figure 5 shows that CO2 mole fraction increases with an increase in the feed temperature. It can be understood that the active volume or circulations increase within the cavity as the feed gas velocity also increases with temperature. It is also observed that the combustion reaction rate increases with an increase in feed temperatures under the conditions studied.
7
4. Studies on Steam Gasification Reaction Effect of steam to oxygen ratio Steam to oxygen ratio in the feed is one of the important parameters to study in the UCG process. It pays a vital role in any gasification process, whether it is surface or in-situ, as it influences the quality of the product gas (i.e. CO to H2 ratio). In the present study, we consider four different steam to oxygen molar ratios (i.e. 0.5, 1, 2 and 3) at a constant feed temperature of 473 K. Optimum steam to oxygen ratio is obtained by performing four independent simulations. From figure 6, it is observed that the cavity roof temperature decreases drastically with the increasing presence of steam at a given feed flow rate and temperature. It is expected that by increasing the steam to oxygen ratio, combustion reaction rate decreases, which leads to decrease in cavity temperature. This is supported by the contours of CO2 mole fraction on various vertical planes along the axial direction in the cavity which are shown in
1250
0.3
1100
0.2
Roof Temperature CO or H2 mole % at outlet
950
Mole fraction
Temperature, K
figure7.
0.1
800
0 0
1
2
3
Steam to Oxygen ratio
Figure 6: Effect of steam to oxygen ratio on cavity roof temperature and CO or H2 mole fraction at the outlet (feed at 473 K, flow rate=31.66 kg/hr)
Figure 7: Contours of CO2 mole fractions for simultaneous combustion and gasification case (feed at 473 K and flow rate=31.66 kg/hr)
From figure 6, it is clear that the by increasing the steam to oxygen ratio, mole fraction of CO 2 decreases which indicates that there is a decrease in the combustion reaction rate. At the same time, steam gasification is favored by the presence of steam up to a certain level and further the gasification rate decreases as it is highly endothermic (see equation 2) as can be seen in figure 6. Figure 6 shows that the mole fraction of hydrogen or carbon monoxide increases with increase in steam to oxygen ratio up to 1 and then decreases with a further increase in this ratio. It is concluded form figure 8 that the optimal steam to oxygen ratio is 1 under the given conditions studied. It may be noted that we have not considered CO and H2 oxidation as well as water gas shift reaction. This may have some
8
influence on the optimum steam to O2 ratio determined here. Figure 8 shows the contours of mole fraction of H 2 or CO on the various vertical planes along the axial direction within the cavity, and it is clear that the maximum mole fraction is obtained for a cavity with steam to oxygen ratio as 1.
Figure 8: Contours of CO or H2 mole fractions for simultaneous combustion and gasification case (feed at 473 K, flow rate=31.66 kg/hr)
Figure 9: Contours of CO or H2 mole fraction for simultaneous combustion and gasification case (flow rate=31.66 kg/hr and steam: O2=1)
Effect of feed temperature In order to study the effect of feed temperature on gasification, three independent simulations are performed at a fixed flow rate (31.66 kg/hr) and optimal steam to oxygen ratio (i.e. 1). As expected that higher feed temperature enhances the endothermic gasification reaction rate, which can be seen in figure 9. It shows the contours of H2 or CO on various planes created in the cavity and reveals that the H 2 or CO mole fraction is higher at higher feed temperature i.e. 873 K.
4. Conclusions A pseudo steady state three-dimensional CFD model is developed to estimate the performance of underground coal gasification process. The model includes heat and mass transfer, and chemical reactions occurring at the boundary of the cavity roof. The effects of feed temperature and flow rate on gasification reactions are studied. The optimal steam to oxygen ratio is obtained for a chosen conditions. Future work may be aimed at performing a detailed
9
parametric study considering other parameters such as pressure, presence of other reactions on the product gas compositions, etc. Nomenclature ρ
Density, kg/m3
t
Time, sec
v p
ρg F
Velocity tensor Static pressure Stress tensor Gravitational body force External body force
k eff
Effective thermal conductivity
Jj
Diffusion flux of species j
Yi
Mole fraction of species „i‟
σ
Stefan-Boltzmann constant
a
Absorption coefficient m-1
G
Incident radiation
qr
Radiation flux
Ri
Rate of reaction species j
kf
Forward reaction rate constant
kb
Backward reaction rate constant
A
Frequency factor
Ef
Activation energy for forward reaction
Eb
Activation energy for backward reaction
References 1. Shafiro vich, E.; Var ma, A. Underground coal gasification: a brief review of current status. Ind. Eng. Chem. Res. 2009, 48, 7865-7875
10
2.
Khadse A, Qayyumi M, Mahajani SM, Aghalayam P. Underground coal clean coal utilization technique for India. Energy
3.
4.
5.
new
coal
Reactor Engineeri ng 2006;4:A37. during
1-53.
Perkins, G.; Sahajwalla, V. Modeling of heat and mass transport phenomena and in underground coal gasification. Chemical Engineering
6.
underground
Britten, J.A.; Thorsness, C.B. A Model for Cavity Growth and Resource Recovery Underground Coal Gasification”, In-situ 1989, 13,
a
2007;32:2061.
Khadse A, Qayyumi M, Mahajani SM, Aghalayam P. Reactor model for the gasificatio n (UCG) channel. I nt J of Chemical
gasificatio n:
chemical reaction
Research Design 2007, 85, A3, 329-343.
Pirard J.P., Brasseur A., Coeme A., Mostade. and Pirlot P. Results of tracer tests during the EI tremedal underground coal gasification at great depth. Fuel, 2000, 79, 471-478.
7.
Biezen, E. N. J. and Bruining, J. An Integrated 3D Model for Underground Coal Gasification. Society of Petroleum Engineers Annual Technical Conference, Dallas, U.S.A. 1995.
8.
Nourozieh, H., Kariznovi M., Chen Z.,, and Abedi J., , Simulation Study Coal Gasification in Alberta Reservoirs: Geological Structure and
of
Underground
Process Modeling. Energy Fuels,
2010 (in press). 9.
Daggupati, S.; Mandapati, R.N.; Mahajani, S.M.; Ganesh, A.; Mathur, D.K.; Aghalayam, P.; Laboratory studies on combustion the context of underground
coal gasification.
cavity growth
Sharma,
in lignite
R.K.; coal
blocks
in
Energy 2010, 35, 6, 2374-2386.
10. Harloff GJ. Underground coal gasification cavity growth model. J Energy 1983;7:410. 11. Daggupati, S.; Mandapati, R.N.; Mahajani, S.M.; Ganesh, A.; Mathur, D.K.;
Sharma,
R.K.;
Aghalayam, P.; Compartment modeling for flow characterization of underground coal gasification cavity. (submitted to IECR) 12. Kuyper, R.A., Van Deer Meer, H. and Hoogendoorn, C. J. “Turbulent Natural Combined Buoyancy Forces during Underground
Convection Flow Due to
Gasification of Thin Coal Layers”. Chemical
Engineering Science. 1994, 49(6). 851-861. 13. Sazhin S.S., Sazhina E.M., Saravelou O.F. and Wild P. The P-1 model for the thermal radiation transfer: advantages and limitations, Fuel, 1996, 75(3), 289. 14. FLUENT 6.3. Users Guide. Fluent Inc. Lebanon NH; December 2005. 15. Shih, Tsan-Hsing, Liou, W. W., Shabbir, A., Yang, Z., and Zhu, J. A new k−ε eddy viscosity model for high Reynolds number turbulent flows. Computers and Fluids, 1995, 24(3), 227.
11
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
CO2 SEQUESTRATION IN UNMINABLE COAL WITH ENHANCED COAL BED METHANE RECOVERY: THE MARSHALL COUNTY PROJECT Richard A. Winschel, Director Research Services, CONSOL Energy Inc. 4000 Brownsville Rd., South Park, PA, 15129 USA [email protected], 412-854-6683 James E. Locke and Ravi S. Srivastava, CONSOL Energy Inc. Richard A. Bajura, Thomas H. Wilson, Hema J. Siriwardane, Henry W. Rauch, Douglas G. Patchen, Brad D. Hega, and Raj K. Gondle, West Virginia University Arthur W. Wells, National Energy Technology Laboratory, U.S. Dept. of Energy
Abstract: A pilot test is being conducted in Marshall County, West Virginia, USA, to evaluate enhanced coal bed methane recovery and simultaneous carbon dioxide sequestration in an unmineable coal seam in the Northern Appalachian Basin. Injection of carbon dioxide (CO2) began in September 2009 and it continues at the time of this writing. This paper describes the project and its current status.
Introduction: Researchers from CONSOL Energy Inc., West Virginia University and the National Energy Technology Laboratory are collaborating in this effort. It is planned to inject up to 20,000 short tons of carbon dioxide at a pressure of up to 933 psig and at a rate of up to 27 short tons per day into an “unmineable” region of the Upper Freeport coal seam located at depths of 1200 to 1800 feet, depending on surface topography. Horizontal coal bed methane wells were drilled in a modified five-spot pattern into the Upper Freeport seam and separately into the overlying (ca. 600 ft. shallower) Pittsburgh coal seam, and these wells have been producing coal bed methane (CBM) since 2004. The peripheral horizontal wells enclose a 200-acre square. The center wells were converted to CO2 injection wells, and CO2 injection commenced in September 2009. A Class II underground injection control permit was obtained from the West Virginia Department of Environmental Protection Office of Oil and Gas. The impacts of CO2 injection into the center wells on the production and composition of the coal bed methane produced in the peripheral and overlying wells is being monitored. Injection will cease when either 20,000 short tons have been injected or when the coal bed methane from the peripheral or overlying wells becomes contaminated with CO2. The pilot test incorporates numerous site characterization and monitoring initiatives including 1) monitoring of the gas and water produced from seven active coal bed methane wells and abandoned deep gas wells in the area of review, and from two observation wells drilled for this project, 2) ground water and domestic well water monitoring, 3) stream water monitoring, 4) soil gas monitoring, 5) surface monitoring for perfluorocarbon tracers injected with the CO2, 6) tiltmeter observations, and 7) seismic observations. Monitoring will continue for two years after injection ceases.
Description of Wells: Project wells were directionally drilled at three locations, referring to Figures 1 and 5, termed “north”, “center”, and “south”. The north wells were drilled down-dip using the slant-hole technique, and they cannot be effectively dewatered from their own wellheads. In the past, gas production from the two Pittsburgh seam wells (MH-3 and MH-4) and one Freeport seam well (MH-5) had been sporadic and they have been shut in for periods of time. Since the start of unattended operation, production from these wells has become more consistent and they have been left open to production. An unrecoverable drill string was left in the Upper Freeport Seam well MH-6, which is sealed. The center well site consists of two horizontal wells drilled into the Upper Freeport Seam; each well has two legs. The wells presently serve as the injection wells for the project; however, the wells were initially drilled as production wells with the intention to convert them to injection wells once gas production diminished. Well MH-18 is the northern-most of the two, with northerly- and westerly-oriented horizontals, while well MH-20 is located to the south of the site and has horizontals that extend towards the south and east. In July 2007, the center wells were shut-in to begin their conversion to injection wells. South wells were drilled into both coal seams; each well was drilled with two horizontal legs. The Pittsburgh Seam well (MH-12) has consistently produced CBM at an average daily rate of approximately 200 thousand cubic feet per day (mcfd). The Upper Freeport Seam well (MH-11) has produced CBM at approximately 9.5 mcfd since production was initiated. . Injection Program: The injection system consists of: 50-ton liquid CO2 holding tank, CO2 injection pump that takes liquid CO2 from the tank and pumps it towards vaporizer, injection tubing, and two injection wells. The liquid CO2 is delivered to the site by truck and transferred into the holding tank. During injection, a twocylinder reciprocating pump transfers the liquid CO2 from the holding tank to the vaporizer. The pump is capable of generating a discharge pressure of up to 1500 psig and is driven by a variable frequency drive that allows the pump speed to be varied either by a 4-20 milliamp signal or by a potentiometer. The vaporizer consists of a natural gas-fired boiler that heats water, and a vaporizer section where the liquid CO2 is vaporized by heat transfer from the hot water. Upon exiting the vaporizer, the gaseous CO2 is split into two streams and is transferred to the two injection wells, designated MH-18 and MH-20. Each injection line is equipped with a flow meter and pressure transducer as well as a pneumatic control valve, which allows for accurate measurement and control of the gaseous CO2 injection. Following commissioning in September 2009, the system underwent a safety review to determine if it could be operated unattended. The findings of the review resulted in:
The installation of automated shut-off valves to effectively stop the flow of liquid and gaseous CO2, A relief valve installed to provide liquid CO2 trapped by a shutdown with a path to exit the pump discharge line without causing any damage to the fittings, and An air receiver tank was installed at the site to supply instrument air to all the valves to ensure that they remain in a desired state until the injection was brought to a systematic halt.
After these changes were implemented, CO2 injection was reinitiated on November 23, 2009. Supervised injection of CO2 continued through the remainder of November and December, with injection occurring for an 8 to 9 hour period each day for four days in November and six days in December. Injection at the site was halted on December 9, by a severe wind storm, which disrupted power in the area for several days. Power to the site was restored on December 14, and injection was restarted; however, the injection was temporarily suspended again because of high differential pressure measured across the CO2 filter caused by debris collected on the filter element surface. Filter elements were replaced and injection was reinitiated on December 15. This problem re-occurred several times through the first two quarters of 2010, but it has since subsided. Fully automated CO2 injection commenced on January 26, 2010. Although weather became a major challenge to continuous injection during the first quarter of 2010, we injected CO2 continuously throughout most of the quarter. Delivery issues encountered due to weather conditions allowed the CO2 storage tank to empty, forcing the injection system to shut down on January 30. The region was hit by a major winter storm on February 5 – 6 that blocked all access to the area and caused a power outage at the site. Access and power were finally restored and injection was resumed on February 24, and has continued with only minor interruptions since. By April, injection pressures achieved the maximum limit of 700 psig allowed by the Underground Injection Control (UIC) permit issued by the West Virginia Department of Environmental Protection (WVDEP). At this pressure the mass injection rate averaged only 6 to 10 tons of CO2 per day (TPD), which is well below the project goal of 27 TPD. CONSOL pursued and received a modification to our UIC permit to raise the maximum injection pressure to 933 psig. To safely and successfully achieve this pressure, various pieces of equipment were replaced, as they were only rated for continuous operation at 750 psig. On August 8, the system was removed from service to allow for the replacement of the components and upgrades to the system‟s programmable logic control. On August 19, the upgrades were completed and the injection system was returned to service. At the time of this writing, injection pressures were approaching 900 psig and the injection rate was averaging approximately 21 TPD. Figure 2 summarizes the injection pressure and cumulative tons of CO2 injected. To date, just over 1,000 tons of CO2 have been injected into the coal seam and there has been no indication of enhanced coal bed methane production. Figure 3 plots the monthly CBM production from January 2007 through June 2010. Missing periods resulting from well shut-ins are omitted.
Geophysical Characterization: Geophysical characterization of the site employs well-log-derived structure and isopach maps, integration of structure inferred from the horizontal well paths, and partial 2D and 3D seismic coverage. Density and sonic logs from one well help establish an approximate tie between the seismic response and subsurface stratigraphic intervals. The seismic survey over the site was designed as a swath 3D. Geophones and sources were positioned along a series of seven lines crossing the site (dark blue lines, Figure 1). Line locations were less than optimal relative to the injection laterals because one of the local landowners denied access. 3D data were obtained through the northern and western parts of the site through simultaneous recording on all geophones of seismic disturbances created by each source in the survey. The midpoint spacing within the 3D grid is 20 feet.
Subsurface Structure. CONSOL Energy provided well log data for all vertical and lateral wells drilled at the site. These wells often had gamma ray and density logs which were used to confirm picks on the Upper Freeport and overlying Pittsburgh coals. Structure of the Upper Freeport based on vertical well penetrations and lateral well path locations within the seam reveal local structural highs to the north and west. The area is gently deformed. Structural relief drops to the south and east with a dip of about 2 degrees or less (Figure 4). Although the laterals oscillate up and down in the Upper Freeport seam, the thickness of the Upper Freeport varies between 2 to 6 feet across the site so that the structure shown in Figure 4 is felt to have an accuracy of approximately 2 feet in the vicinity of lateral well control points. Structural variability encountered along the forked injection laterals across the center of the site suggests that local structures with relief of about 10 feet could be present elsewhere in the diamond. Seismic Data. 3D seismic data extracted from the swath survey provide a spatially continuous view of the subsurface reflection response over the northern part of the site (Figure 5). Reflection travel times to the Upper Freeport generally lie between 0.29 to 0.3 seconds. The bandwidth of the shallow data extends, on average, from approximately 30 to 100 Hz with a dominant frequency of about 60 to 65 Hz. Sonic from a nearby well indicates an Upper Freeport coal velocity of about 9300 feet/second. The velocity and dominant frequency limit resolution to 25 feet and greater. Detection of thinner seams is possible. The Upper Freeport has a thickness of at most 6 feet in the area and although not resolvable as a separate reflection event its presence has noticeable influence on the composite seismic response from the Upper Freeport coal and surrounding intervals. Time slice views through the 3D cube (Figure 6) reveal the presence of subtle structural and stratigraphic features in the vicinity of the Upper Freeport. Attribute Analysis. A variety of post-stack seismic attributes were computed and evaluated in order to gain additional insights into the possible existence of penetrative deformation that might facilitate leakage at the site. The absolute value of the derivative of trace amplitudes enhances the resolution of stratigraphic and structural features. The vertical seismic display (Figure 7) reveals patterns of disrupted reflection events between 0.24 to 0.34 seconds. This time interval includes the Upper Freeport coal. Considerable local stratigraphic and structural variability is observed above and below the Upper Freeport. The lowermost interval overlying the Upper Freeport is approximately 200 feet thick (orange circled area in Figure 7). Although there is no evidence of major penetrative deformation, subtle features observed in the derivative display suggest that small-scale faults and/or fracture zones may be present that could intersect the reservoir and extend into overlying intervals. Summary of Geophysical Characterization. Geophysical characterization of the site reveals that the area is mildly deformed. Bedding dip in the vicinity of the site is generally less than 2 degrees. Updip directions within the diamond lie to the west and northwest. 2D and 3D seismic suggest the presence of complex stratigraphic environments in intervals bounding the Upper Freeport coal. Although major penetrative faults and folds are not observed in the seismic data over the site, the presence of near-vertical alignments of subtle features in the attribute display do suggest that minor faults and/or fracture zones may be present. This is not surprising, since fractures are commonly observed in outcrop, core and FMI log. The seismic data over the site has provided valuable stratigraphic and structural information about the local geology that is difficult to infer
from well data alone. Acquisition of post-injection seismic survey is planned to provide a comparison with baseline 3D seismic. Future time lapse analysis is planned to evaluate possible 3D seismic uses for detecting and mapping the distribution of CO2 within the reservoir. Monitoring of Deep Wells: Prior to injection, fifteen wells within a quarter-mile of the area of review were sampled to determine baseline characteristics of the gas and water (if available) produced. Four samples were collected in consecutive weeks to establish the baseline. Gas samples were analyzed for oxygen, nitrogen, methane, ethane, higher hydrocarbons, and CO2. Water samples were analyzed for pH, dissolved oxygen, ORP and conductivity, dissolved solids, and numerous anions and cations. Temperature was also measured. During the injection period, we are performing monthly monitoring for CO2 breakthrough by sampling water and CBM gas and the annular space of several wells located within a quarter-mile of the area-of-review as defined in the UIC permit. The most recent gas monitoring results show no evidence of CO 2 breakthrough. Water monitoring results also show no evidence of CO2 breakthrough.
Shallow Hydrogeologic Monitoring: Environmental monitoring of stream water, shallow ground water, and shallow vadose zone gas has been ongoing, both before and during the injection of CO2 gas, at the CONSOL Energy Inc. test site. The primary purpose of this monitoring is to determine if any sign of significant land-surface leakage of injected CO2 gas is occurring at the test site, through making and comparing between geochemical water measurements for background pre-injection environmental conditions and during-injection environmental conditions. Another purpose of this monitoring is to test and develop better hydrogeologic monitoring procedures for detecting any ground surface seepage of injected CO2 gas at this or other CO2 geologic sequestration sites with similar hydrogeologic settings. A secondary research purpose is to develop a better general hydrogeologic understanding of coal and natural gas bearing sedimentary rock terrains within the Allegheny Plateau Geologic Province at which future CO2 sequestration in unmined coal beds is possible. Figure 1 illustrates the CONSOL Energy Inc. CO2 gas injection pilot test site along the blue-colored stream of Pennsylvania Fork of Fish Creek, located between the towns of Bellton to the northwest and Georgetown to the southeast. The stream flows northwest from Georgetown towards Belton. Houses near these towns are shown as black dots, and forested land is shown in green color. Stream monitoring locations are indicated with large red square dots; shallow ground-water monitoring wells are shown by large blue square dots, and sampled domestic water supplies (wells and one spring) are shown by small magenta square dots. Other sampled features by coauthor scientists are blue seismic geophysical survey lines, tiltmeter holes (small square black dots) for land surface monitoring, and capillary adsorption tubes (CATs – small square orange dots) for atmospheric gas monitoring of chemical tracer gas. The test site black and red lines indicate the CO 2 gas injection (center square lines) and coal bed methane production (peripheral wells) of CONSOL Energy Inc. The test site is located within Deep Valley, which has a gently sloping south side and a steeply sloping north side relative to Fish Creek, which bisects the site. The major meander bend of this stream dictates the test site topography, with the outside (northern portion) of the meander bend being associated with steep hillsides due to erosion over geologic time along the cut bank there. This site is within the soil covered sedimentary rock terrain of the Dunkard Group of Permian age; the stratigraphic sequence consists of alternating layers of clastic
sedimentary rocks (sandstone, siltstone, shale, mudstone, and claystone), limestone, and coal beds. Fish Creek, monitored water wells W-1, W-2, and W-3, and sampled domestic wells B-1, B-2, G-1, G-3, and G-4, and domestic spring G-2 are all located within the Washington Formation of the middle Dunkard Group. Procedures. Monitoring water wells W-1, W-2, and W-3 were each drilled in May 2008, to 6 inches (15 cm) in diameter from 0 to ~20 ft (0 - 6 m) in depth , and 5 inches (13 cm ) in diameter from ~20 to 105 ft (6 - 32 m) in depth; these wells have respectively 21.5 ft (6.5 m), 24 ft (7 m), and 16.5 ft (5 m) of PVC plastic casing pipe, of which 2.2 ft (67 cm), 2.5 ft (76 cm), and 1.4 ft (43 cm) respectively are the pipe stickup heights. The upper ~20 ft (6 m) of hole annulus zone outside the casing pipe of each well was grouted with bentonite clay to ground surface to lessen the danger of surface contaminants entering these wells. Also, 60 ft (18 m) of slotted PVC plastic liner was installed in each well within the 45 to 105 ft (14 - 32 m) depth zone to prevent hole collapse. Well W-1 is the highest yielder with a driller‟s initial yield of ~25 gpm (6.6 L/m), and a transmissivity of 874 ft2/d (0.944 m2/d) and a hydraulic conductivity of 12.5 ft/d (3.8 m/d) based on a hydraulic aquifer test; the main aquifer for this well is coarse grained sandstone. Well W-2 has a driller‟s initial yield of 2.4 gpm (0.63 L/m) and has a shale to fine grained sandstone for an aquifer. Well W-3 has a driller‟s yield of ~6 gpm (1.6 L/m) with a sandy shale aquifer. Formal field sampling and field chemical analysis was done starting in October 2008 (after trial field sampling in September 2008), and continuing through this month (August 2010) and to at least year‟s end. Stream discharge was measured and water sampling was done weekly to biweekly at spots S-1 (upstream), S-2, and S-3 (downstream). Monitoring wells W-1, W-2, and W-3 were measured for depth to water table and sampled weekly, as were the stream spots. The six domestic water supplies (G-1, G-2, G-3, G-4, B-1, and B-2) were sampled biweekly, to more broadly determine background ground water geochemistry well outside of any anticipated environmental zone of influence for injected CO2 gas; Belton wells B-1 and B-2 had sampling discontinued in May 2010 at the request of their property owners. All sampled waters were field tested for electrical conductivity, temperature, and pH with appropriate meters. Total dissolved sulfide was also field measured beginning October 2008, for the three monitoring wells, and for the domestic water wells exhibiting a hydrogen sulfide odor. Well head space gas for the three monitoring wells was measured weekly for CO 2 gas content beginning May 2009, using a portable Vaisala Inc. meter and probe. Water chemical analyses were performed on all sampled waters by CONSOL Energy Inc Research and Development. Laboratory chemical parameters analyzed for are pH, electrical conductivity, major cation elements (Ca+2, Mg+2, and Na+1), minor cation elements (K+1, Fe+2, Al+3, and Mn+2), and major anions (HCO3-1, Cl-1, and SO4-2). Following water geochemical checking and correcting for data errors, the computer program MINTEQ was run to calculate the saturation index for calcite (SIC) and theoretical equilibrium CO2 gas partial pressure for water. Also, speciation calculations were done via EXCEL for hydrogen sulfide species (H2S, HS1 ) in water, based on pH and total dissolved sulfide. Results and Interpretation. Table 1 illustrates the average (mean) chemical values for selected parameters for the sampled wells and stream spots. The stream waters show significantly higher field pH than lab pH, with the greatest difference being the stream spot S-2 (7.87 field pH versus 7.15 lab pH). Theories for this pH difference are being explored, but analytical error has been mostly ruled out. One theory is that stream water, either directly or by the influence of ground-water runoff, is undergoing chemical reactions to elevate the field pH, such as by reactions (1) and (2) below, whereby methane gas is reacting with sulfate ions to produce bicarbonate ions, hydrogen sulfide gas, and hydroxyl ions. HCO3- +HS- + H2O CH4 + SO42- (1)
H2S(g) + OHHS- + H2O
(2)
Evidence for this theory is the strong sulfurous odor that commonly is noted during sampling near stream spot S-2 and well W-2. These reactions could be followed by reactions such as (3) - (5) below within water sample bottles during transit to and storage within the lab prior to the lab analysis chemical analysis procedure, causing the lab pH to drop. SO42- + H+ HS- + 2O2 (3) 2+ SO4 + 2H H2S + 2O2 (4) +1 HCO3 +H + H2O CH4 + 2O2 (5) Another theory is that photosynthesis occurring during warm months and especially low flows raises the stream field pH by reactions (5) and (6) below, but this would not explain the lower lab pH observed. CH2O CO2 + H2O + O2 (6) + CO2 + H2O H + HCO3 (7) Lesser, but yet still significant, differences exist between the field pH (higher) and the lab pH for the three monitoring wells (W-1, W-2, and W-3) and two of the domestic wells (G-1, G-4). Reactions (1) to (5) could partially explain this effect, and so could reaction (8) below for wells W-1, G-1, and perhaps G-4; wells G-1 and G-4, and sometimes wells W-1 and W-2 exhibit a strong odor similar to that of hydrogen sulfide (or methyl mercaptan for well W-2) during field sampling. Also, wells W-1 and G-1 have strongly positive SIC values indicating super-saturation for CaCO3 within the lab tested water samples. Precipitation of some CaCO3 within sample bottles prior to analysis would have acted to lower the lab pH. CaCO3 + H+ Ca+2 + HCO3- (8) The equivalents concentration data (not shown here) were used to characterize the overall water chemistry (dominant cation and anion) for Table 1 sampled water locations. For the Fish Creek spots, the predominant ions are Ca+2 and HCO3-1, except for spot S-2 during early December 2009 through June 2010 when SO42- was the predominant anion. Wells W-1 and W-2 are predominantly Na+1 and Cl-1/ HCO3-1 while well W-3 is predominantly Na+1 and HCO3-1. Wells W-1 and W-2 have suffered some contamination effects, especially W1, from brine unrelated to CO2 gas injection. One known soil contamination episode with leaking fluids from a slop tank was discovered with brine occurred in 2006 ~300-400 ft (92 – 122 m) south of well W-1 and prior to CO2 injection initiation; this could have been caused by a brine spill or leak near the present holding tanks adjacent to the methane pump jacks for gas production wells MH-11 and MH-12. Well W-3 has an exotic chemistry with extremely low Ca+2 and Mg+2 and extremely high pH, HCO3-1, and Na+1 which is difficult to explain. This chemistry could possibly be caused by four types of geochemical reactions. The first type could be caused by a strong upward seepage of methane gas from a natural reservoir or contamination (such as from a leaky gas pipeline) source, which could cause reactions (1) and (5) above, resulting in extremely high HCO3-1 concentration; this situation has been witnessed before in Rauch‟s research of natural gas impacts on water wells in southern West Virginia and eastern Kentucky. Second, neutralization of the acid created by reaction (5) could be occurring due to chemical weathering of rock minerals by the reverse of reaction (8) for example, but this is unlikely to explain the unusually high pH. Third, hydrolysis reactions like reaction (2) could occur to create the very high pH. Fourth, a strong natural water softening effect associated with local shale and mudstone rock is likely occurring, whereby aqueous Ca+2 and Mg+2, created by chemical weathering reactions, are exchanged for Na+1 via cation exchange on clay mineral grain surfaces. Potential Geochemical Signals for Surface Seepage of Injected CO2 Gas. To date (August, 2010) no probable signals, involving geochemical indicators of surface seepage of injected CO2 gas, have been observed based on available shallow ground water chemistry and stream chemistry. Potential geochemical signals being monitored
include direct signs of elevated CO2 gas, and indirect signs of CO2 gas surface seepage. Direct signs could include significantly elevated concentrations of CO2 gas dissolved in stream water, shallow ground water of the monitoring wells, or shallow ground water of the sampled domestic supply wells or spring; and elevated concentrations of CO2 gas vapor within the vadose zone head space gas of the monitoring wells. Defining “significantly elevated” is somewhat arbitrary, but CO2 gas concentrations exceeding 2 to 10 times normal background levels would be suspicious. Background is defined by pre-injection monitored conditions for onfield-site waters (monitoring wells and stream of Figure 1), and both pre- and post-injection monitored data for the domestic water supplies, which are well outside of (beyond 1 km of) the injection test site square (the potential geographic zone of influence for injected CO2 gas seepage). Also, any apparent (from auger holes) seepage signals would be checked by follow-up carbon isotope analyses of vadose gas or well water through another geologic consultant and lab, and by sampling and testing of vadose gas for injected chemical tracer from soil (within auger holes) and shallow aquifers (sampling well head space gas). Lastly, any probable geochemical signals received would be checked for a trend against the historical record of injected CO2 gas volume and injection borehole pressure for deep wells MH18 and MH20. Indirect signs of probable CO2 gas seepage are also being investigated using the principles discussed above. Two or more simultaneous indirect signs would be required, involving shallow water dissolution products or vadose zone content of CO 2 gas. For example, signs could be excessive methane gas displaced by upward seepage of CO2 gas, or extreme values of chemical reaction byproducts of CO2 gas seepage, such as elevated pH, bicarbonate, hardness, and hydrogen sulfide or total dissolved sulfide.
Ground Response During CO2 Injection: The site consists of a lower unmineable coal seam and an upper mineable coal seam. CO2 is being injected into the lower coal seam through horizontal wells, while coalbed methane is produced from horizontal wells in both the lower and upper coal seams. The primary objective of the work described in this section is to study the ground response caused by injection of CO2 in the lower coal seam. In order to study the ground response, thirty six high precision tiltmeters with two GPS receivers were installed that could measure surface displacements in the sub-millimeter range, as shown in Figure 8. The use of tiltmeter technology is relatively new and has the potential for use in monitoring fluid flow and surface deformations in commercial CO2 sequestration projects. Such measurements can be useful in calibrating geomechanical models that can predict reservoir response. These instruments were installed in hilly ground terrain near the injection zone. The dense vegetation and steep slopes at the site made the construction of these tiltmeter stations extremely difficult. Some earthwork was done to place these instruments at the site. Figure 9 shows a typical tiltmeter station at the field site. A typical tiltmeter station consists of a battery, a solar panel, and radio telemetry box to communicate with the central processing unit located at the site. All 36 tiltmeters are functional and have been calibrated prior to injection of CO2. Real-time measurements are being collected during the injection of CO2 on a regular schedule. The CO2 injection began on September 08, 2009 and as of this writing, about 1000 tons of CO2 have been injected. Very small surface displacements have been observed in the tiltmeter data. Preliminary analysis of tiltmeter data shows small surface displacements, and a typical surface deformation pattern at the site is shown in Figure 10. Three-dimensional coupled flow-deformation analyses based on the finite element method were performed to determine fluid flow and surface displacements at the site. A model description and the idealized well configuration are shown in Figure 11 and preliminary results on computed surface displacements are shown in Figure 12. The data collection, data analysis, and modeling work are underway. Results obtained from the
tiltmeter measurements and numerical modeling of the geomechanical response of the reservoir will be presented in a subsequent paper.
Near Surface Monitoring: Perfluorocarbon (PFC) tracers are added directly to the injected CO2 at the well heads and serve as a surrogate for potential leakage of CO2 into the near surface. The PFC tracers can be analyzed at the level of femto-liters tracer per liter of soil-gas or atmosphere. The two injection wells will each have 500 mL of one of two PFC tracers injected from syringe pumps capable of matching the injection pressure. The injections are made at a rate of 12.5 mL tracer per ton of CO2 with tracer injection taking several days to complete. Tracer injections are made as soon as a steady state flow of CO2 has been achieved. The two PFC tracers to be injected at the Marshall county site are perfluoro-1,2-dimethylcyclohexane and perfluorotrimethylcyclohexane. Soil-gas tracer monitors consist of steel pipes driven 1 meter into the ground where sufficient soil is available, sealed at the top, and open at the bottom. A 3-inch long glass tracer absorption tube containing Ambersorb is lowered to near the bottom of the pipe and left in place to collect any PFC tracers present in the soil-gas. The diffusion rate of soil-gas into the end of the open glass tube corresponds to an exposure of 200 milliliters soilgas during a 24 hour period. The tubes remained underground for periods ranging from one week immediately after tracer injection to approximately two months after the first month of monitoring. The sorbent tubes are exchanged as sets from all monitoring locations. The pipes have atmospheric PFC tracer monitors clipped on at about 4 feet above the ground where sorbent tubes are exposed to the atmosphere inside a rain shield. Analysis is conducted at NETL‟s tracer laboratory using thermal desorption, and gas chromatography with chemical ionization, mass spectrometric detection. The near surface monitoring grid consists of 28 locations within 1,200 meters of the injection wells. Of these locations, 18 are within 200 meters of the center point between the two injection wells. The monitors were located near features identified as representing an above average leakage probability or potential migration pathways from the subsurface. These include locations adjacent to existing well bores, and along potential fault lines identified from lineament analysis of satellite images and surface conductivity surveys. Monitors are also placed at „hydrocarbon anomalies‟ where light hydrocarbon concentrations were found to increase with soil-gas depth or where atmospheric methane plumes were not associated with wells or surface contamination. A subset of the PFC tracer monitoring locations (12 locations) was used to conduct CO2 surface flux measurements, and to take samples of soil-gas for CO2 and light hydrocarbon analysis. Soil-gas fluxes are measured using a LiCor 6400 photosynthesis chamber with the soil CO2 flux chamber (6400-09). At each location, the chamber is pressed into the soil surface to a depth of 1 cm, and three flux measurements are made. For each measurement, the CO2 in the chamber is scrubbed to 10 ppm below the ambient atmospheric concentration and allowed to rise to at least 10 ppm above ambient. The flux rate is calculated from the rate of CO2 concentration change in each measurement. Soil-gas samples are taken by driving a probe (AMS Gas Vapor Probe 427.01) into the soil with an annular hammer. The probe is essentially a stainless steel pipe that has a closed point at the end to facilitate insertion into the soil. There are perforations just above the tip to allow soil gas to flow into the pipe. A solid steel rod is set inside the hollow probe during insertion and extraction to prevent the holes from clogging with soil. At each sampling depth, the hammer and soil insert are removed. A cap fitted with a septum seal is placed on the top to allow gas sampling. A gas-tight disposable syringe is used to extract samples through the septum. The first 50 mL of gas extracted from each depth is discarded in order to flush the volume of the probe. Each sample of 50 mL is injected into a pre-
evacuated 25 mL bottle through a septum. This results in a sample that is at approximately 2 atm of pressure. Soil gas samples are laboratory analyzed for CO2 concentration and C1-C3 hydrocarbon concentrations. Soilgas samples and CO2 flux measurements are taken at different seasons prior to injection and then repeated every four months following the start of injection. Monitoring for PFC tracers in fugitive atmospheric emissions at the production wells MH-3, -4, -11 and -12 will be conducted to monitor for tracer appearance in production gas. Samplings of production gas will also be taken when sorbent tube exchanges are made. Results from these measurements will provide information on the movement of CO2 through the sequestration reservoir. Headspace-gas in the monitoring water wells are also sampled for the presence of PFC tracers by pumping 3 liters of headspace-gas through sorbent tubes at the same time sorbent tube exchanges are made.
Summary: To date, over 1000 short tons of CO2 have been injected with no sign of either CO2 breakthrough or of enhanced production of coal bed methane. Injection will continue until CO2 breakthrough occurs or 20000 short tons of CO2 are injected, whichever is first. Monitoring by a broad array of methods will continue throughout the injection period and for two years following its completion.
Acknowledgment: This work was funded in part by the U. S. Department of Energy through its National Energy Technology Laboratory under cooperative agreement No. DE-FC26-01NT41148, and in part under Montana State University Grant No. G137-05-W0221.
Location W-1 W-2 W-3 S-1 S-2 S-3 B-1 B-2 G-1 G-2 G-3 G-4
From (date) 10/10/2008 10/10/2008 10/10/2008 10/10/2008 10/10/2008 10/10/2008 10/10/2008 10/10/2008 10/10/2008 10/10/2008 3/9/2009 4/6/2009
Average To Average Average Field EC (date) Field pH Lab pH (μS/cm) 7/21/2010 8.35 8.03 1058 7/21/2010 8.46 8.29 667 7/21/2010 9.28 9.14 820 6/16/2010 7.73 7.17 202 6/16/2010 7.87 7.15 231 6/16/2010 7.85 7.15 267 5/17/2010 7.09 7.17 783 5/17/2010 7.24 7.35 776 6/14/2010 8.61 8.31 485 6/14/2010 7.00 6.68 256 6/14/2010 5.58 5.83 227 6/14/2010 7.89 7.73 568
Average Lab EC Average (μS/cm) HCO3- (mg/L) 1024 293 684 175 862 442 197 80 224 80 263 80 825 256 833 361 449 278 247 103 223 20 611 196
Average Average Average Average Ca2+ Average Mg2+ Average Na+ Cl- (mg/L) (mg/L) (mg/L) K+ (mg/L) (mg/L) SO42- (mg/L) 161.0 38.5 5.8 1.2 161.7 6.3 106.0 10.1 1.7 1.1 125.9 8.6 25.1 0.9 0.3 0.6 198.6 5.4 7.3 25.6 4.3 1.8 7.0 19.5 16.1 26.4 4.5 1.8 10.8 19.6 28.6 27.9 4.9 1.9 15.5 19.3 104.3 73.5 13.2 2.0 67.4 20.6 67.0 61.7 8.6 2.2 98.6 1.6 6.2 9.3 1.6 1.1 98.5 1.8 9.4 32.4 6.2 1.4 7.8 21.5 16.7 15.8 4.3 12.7 9.5 32.2 59.1 23.4 3.3 1.7 98.8 5.3
Table 1: Mean chemical parameter values for sampled water wells. Wells W-1, W-2, and W-3 are monitoring wells constructed for this research; wells B-1 and B-2 are domestic well supplies located in Bellton; supplies G-1, G-2, and G-4 are domestic wells located in Georgetown; and supply G-2 is a domestic spring near Georgetown.
Figure 1. Site map showing production, injection, and monitoring wells, seismic survey lines, tiltmeter locations, stream monitoring locations, and NETL near-surface monitoring locations.
Figure 2. Injection summary.
Figure 3. Monthly Gas Production from project-related CBM wells.
Line 1
Line 7
Line 6
Figure 4. Structure of the Upper Freeport coal. Contour interval is 5 feet. Elevations are subsealevel. XY coordinates are metric UTM NAD83. Locations of seismic source and receiver lines are shown in light blue.
Figure 5. Seismic coverage at the site. The cross line and inline grid outlines 200 foot by200 foot areas over the site. Injection laterals (black lines) extend to the northwest and southeast.
Figure 6. A portion of the 3D seismic volume is shown in the area covering the northwestern laterals. CO2 is injected in the laterals to the northwest through the MH18 well.
Figure 7. Absolute value of derivative along 2D Line 7. Area of disrupted reflection events is circled in orange. The red diagonal line locates a seismic feature interpreted to be a minor fault. Log picks shown on the synthetic trace (blue trace at left) show the locations of the Pittsburgh coal (Pit), Bakerstown coal (BakrTwn), the Mahoning coal (Mh), the Upper Freeport coal (UFr) and an undifferentiated Allegheny coal (Alleg).
N
36 Tiltmeters
Production Laterals in Upper Freeport Coal Seam Production Laterals in Pittsburgh Coal Seam Figure 8. Configuration of horizontal wells and 36 tiltmeters
Figure 9. A typical tiltmeter station at the site
Figure 10. Preliminary surface deformations at the field site based on tiltmeter data
(a) Finite element model
(b) Assumed well configuration
Figure 11. Three-dimensional finite element model
Figure 12. Typical computed surface deformations due to fluid injection
A GIS-DSS FOR A CO2-ECBM PROJECT FEASIBILITY STUDY: CASE OF SULCIS COAL BASIN (SARDINIA, ITALY). Ciccu Raimondo1, Deiana Paolo 2, Mazzella Alessandro3, Tilocca Caterina4 1
Professor, Dept. of GeoEngineering and Environmental Technologies - University of Cagliari, Italy - [email protected]
2
Dept. of Energy - Section Implant and Processes - ENEA - Roma, Italy - [email protected]
3
PhD Student, Dept. of GeoEngineering and Environmental Technologies - University of Cagliari, Italy - [email protected]
4
Researcher, Dept. of GeoEngineering and Environmental Technologies - University of Cagliari, Italy - [email protected]
ABSTRACT The correct identification of the potential sites for geological storage of greenhouse gases (i.e. CO 2) is the most important operative step to assure the success of the project. It is obvious that, in the CO 2-ECBM technologies field, the success of a geological storage project can be reached only if, in addition to the technical and the economical accomplishment, the safety for the people and the ecosystems directly exposed are always guaranteed. In the CO2-ECBM technologies field, the geological reservoir evaluation must consider a great number of parameters, often very heterogeneous, mainly concerning: the coal seems features, the hosting formations properties, the environmental restrictions and the human/administrative constrains. The variety and variability of these parameters make each evaluation unique and, for this reason, it is impossible to give some “universally-valid” guidelines to help the Decision Maker in the planning procedure. For these reasons it is not advisable for the Decision Maker using only his own experience rather it would be advisable the use of a Decision Support System (DSS). In the Carbon Geological Storage (CGS) field, the combination of DSS and GIS technologies allows the realization of very powerful tools able to identify and solve technical problems (like the location of geological
reservoirs
able
to
contain
safely
large
quantities
of
gases)
and
to
evaluate
technically/economically different alternatives (like the environmental monitoring system planning) in order to assure the right balance between the technical/economical success of the project and its environmental sustainability. This research work concerns the first results accomplished by our research team in the development of a GIS-DSS for the pre-feasibility study of a CO 2-ECBM project within the Sulcis coal basin (Italy, SW Sardinia). The GIS-DSS has been completely implemented using the ArcGIS Model Builder 9.3. The available data concerning: the basin DTM, the coal bed 3D-model, the geological maps, the aerial photo, the land use data and the environmental constraints have been combined using the Weighted Overlay Process (WOP). The WOP applied (also known as Multi-Criteria Evaluation - MCE) produced an output grid by combining the input grids values and weighting them by using a relative importance scale. This operative step allowed to define, within the basin area, the most suitable zones where the next detailed studies should be concentrated (“Basin Suitability Map”). The current version of the GIS-DSS, fully functional although yet under development, can be modified to consider more input parameters and/or to produce different results varying the weighting relative scale.
INTRODUCTION According to the latest data published by International Energy Agency [1], fossil fuels fulfil more than 80% of world energy demand and about 98% of transport (Figure 1).
Coal
Hydro
Oil
Comustible renew ables & w aste
Gas
Other
Nuclear
Figure 1: World energy production by source [1].
Figure 2: CO2 emission in the atmosphere trend [2].
The widespread use of these non-renewable power sources produced, starting from the years of the Industrial Revolution, a progressive raise of the concentration of greenhouse gases in the atmosphere (Figure 2). This fact had led to the well-known environmental problems related to the release to the atmosphere of greenhouse gases produced during combustion (especially: CO 2, N2O, CH4 e O3). However, because the economic development of each country needs large quantities of energy, it is estimated that fossil fuels will remain a strategic energy source for at least another century ahead [2], or until alternative technologies will be able to fulfil the entire world's energy demand. For this reason, in recent years, the scientific community is studying all the techniques that can produce a reduction in emissions of greenhouse gases into the atmosphere. Among the possible alternatives, the injection of CO 2 into coal bodies has taken, in recent years, particular importance due to the fact that chemical and biological processes that develop between the injected gas and the coal matrix, determine the release of natural gas (predominantly formed by a variable mixture of methane, ethane, propane and butane) imprisoned into the organic matter [3]. This technique, known as Enhanced Coal Bed Metane (ECBM), allows a reduction of the concentration of CO 2 in the atmosphere with revenues from the exploitation of the released natural gas capable of counterbalancing the operating costs for the capture, compression, transport, injection and environmental monitoring. Currently, because of the energy and climate crisis, even the Sulcis orebody is at the centre of intensive studies about its alternative uses. In this frame, the University of Cagliari has been participating since many years in a study regarding the assessment of geological capture of CO 2 with ECBM extraction of CH4.
SULCIS COAL BASIN (SARDINIA, SW) Geographical Settings The Sulcis coal basin is located in the Sardinia SW and it is formally delimited by the Gulf of Fontanamare in the North, the Gulf of Palmas in the South, the Palaeozoic outcrops in the East and the island of San Pietro in the West (Figure 3).
Figure 3: Sulcis Coal Basin localization.
The coalfield covers an area of about 200 square kilometers with proven reserves for over 130 million tonnes of coal. Geological and Morphological Settings The morphology of the region (Figure 4) is characterized by the presence of a volcanic coverage (mainly ignimbritic) that shows a general inclination towards SW with average slope of 8÷10%. These formations have been strongly modified by weathering processes and endogenous phenomena; as result a pattern of reliefs with tectonic structures NS and EW have been created. The coast area, mainly formed by the accumulation of recent sediments (such as silt and sand), shows a general flat morphology. In some zones of this area are still present the features of marshy conditions. In the southern area, upon the tabular ignimbritic base, the presence of some andesitic bodies create a mountain-like morphology.
Figure 4: Morphological map.
Figure 5: Geological Map [4] [5].
From a tectonic point of view the entire coalfield is crossed by a series of regional fractures with prevalent direction NW-SE, NE-SW and, subsequently, E-W (Figure 5). These structures are strongly related with the complex tectonic events that affected the southern Sardinia from Miocene to Quaternary. Some of the studies made over the last years have identified 13 structural faults linked to a dense population of smaller fractures that accompany the main ones, showing the complex effect of alternating forces of compression and distension [6]. Because of the presence of this complex field of fracture, as result of the tectonic movements that affected the area, it is impossible to use a 3D model of the ore body that does not take into account the results of tectonic phenomena, which locally produced significant deformation and discards of the coal formation [7] [8]. Mining Settings The Sulcis coal deposit is hosted in a sedimentary sequence (dated between the upper Paleocene and the Oligocene) above a Mesozoic transgressive formation and closed on the top by a volcanic Miocene sequence (from rhyolitic to ignimbritic) [10]. The Produttivo system shows a thickness between 40 and 100 meters and it is formed by at least 16 seams with variable thickness, related to 16 different genetic events. Three of the seams have historically been mined for over 100 years to provide energy for the economic development of Sardinian and Italian industries [11]. The Produttivo system crops out in the NE sector and dips down to 800 meters below sea level in correspondence to the coastline. The influence of the genetic conditions variability together with the tectonic events produced a blockfeatured ore body with a "coal whose characteristics are different from seam to seam and from area to area, particularly going from North to South” [12].
Sulcis Coal Characteristics Sulcis coal is classified as sub-bituminous B/A (ASTM) or "glanzbraunkohle” (DIN) [9], at the limit of lignite. Macroscopically Sulcis coal looks shiny black with intercalated layers and lenses of varying brightness (Figure 6); often are present intense fracturing lines, perpendicular to each other, and along which thin beds of calcite are visible (Figure 7) [13].
Figure 6: Morphological map.
Figure 7: Calcite between Vitrinite beds [12].
For the mineralogical composition of the Sulcis coal, the bibliography [12] indicates (Table 1):
Table 1: Sulcis Coal Composition
•
Vitrinite Group
71.2 %
Exinite Group
11.6 %
Inertinite Group
5.6 %
Other Minerals
11.6 %
the constant presence of pyrite (widespread both as small and single crystals with dimensions between 2 and 5 microns or as clustered crystals with size up to 200 microns);
•
the abundant presence of a carbonate phase (i.e. calcite and dolomite) filling the contraction cracks in the Vitrinite beds;
•
the presence of a clay phase (mainly kaolin) as tiny beds.
•
the presence of other minerals like quartz, gypsum, iron oxides and sulphurs, calcium phosphate.
Environmental Constraints From an environmental point of view, the Sulcis Basin covers an area with an high social and cultural value: an high number of archaeological sites (such as temples and necropolis), historical mining sites and parks overlie within the Basin (Figure 8). The protection of these artistic, historical and natural heritages must represent a priority against any project that produces an impact on the environment.
Figure 8: Environmental constraints localisation.
Figure 9: Soil use map.
A GIS-DSS FOR A CO2-ECMB PROJECT FEASIBILITY STUDY Introduction The Sulcis Coal Basin characteristics represent, in different ways, a problematic situation for a rational approach to a CO2-ECBM project. Geological and tectonic conditions together with the existence of anthropogenic and environmental constraints, require a careful evaluation in order to obtain the optimum in terms of technical and economical goals but even for the safety of the people and the ecosystems directly exposed. In fact the faults and the fractures by which the Basin is crossed, represent areas of dangerous weakness, that affect the surface drainage patterns and could become hazardous way of escape for the injected gas. Sulcis Basin Terrain Model Starting from the available data, the digital terrain model of the Sulcis-Iglesiente has been made up (Figure 10). Elevation data provided by 3D points and contour lines have been interpolated in order to produce a morphological model of the area. Data processing has been carried out by the TIN-DTM construction algorithm. Available aerial imagery have been draped on the DTM producing a landscape model of the entire basin field (Figure 11) crucial for the subsequent phase of recognition of the tectonic structures.
Figure 10: Sulcis Coal Basin DTM.
Figure 11: Sulcis Coal Basin DTM.
Starting from the DTM of the basin, the slope (Figure 12) and exposure (Figure 13) of rock formations have been calculated.
Figure 12: Sulcis Coal Basin slope map.
Figure 13: Sulcis Coal Basin exposure map.
Sulcis Basin Structural Model Through a combined analysis of the available literature [6], geological maps [5] [14] [15], satellite data and aerial photos, a structural model for the basin area has been developed [7] [8]. This analysis allowed the identification of spatial sub-domains in which no tectonic structures were recognizable.
Figure 14: Sulcis Basin Structural model.
Figure 15: Sulcis Basin structural sub-domains identification.
Sulcis ore body 3D Model The 3D modelling procedure is based on the sample data carried out by Carbosulcis SpA during the last 30s [14]. For building the Sulcis deposit 3D model an innovative modelling technique based on a combination of TIN meshing, geometric extrapolation and linear interpolation [8] has been developed. The modelling results are shown in Figure 16 and Figure 17. Combining top and bed modelling results a thickness model has been created trough a “extrusion between” procedure (Figure 18).
Figure 16: Top model.
Figure 18: Thickness map.
Figure 17: Bed model.
Figure 19: 3D Model.
Data Processing Data processing represents the most important operative step in building a Decision Support System. For the purposes of this research work it was implemented using the ArcGIS Model Builder 9.3. The flow chart for the identification of potential CO2-ECBM site is shown in Figure 20.
3d points
contour lines
DTM
slope map
aspect map
geologic map
aerial photos
satellite data
point sample
DTM
DTM
structural model
structural model
bed 3D model
top 3d model
constraints
soil use map
thickness model
constraints
evaluation map
Figure 20: GIS-DSS flow chart.
All the available information and data converged in the ArcGIS 9.3 Model Builder. The Model Builder allowed to obtain a final evaluation through the Weighted Overlay Process (WOP). The WOP applied (also known as multi-criteria evaluation – MCE) created the output evaluation map by combining, through appropriate weights established during the construction of DSS, the suitability values of each different input theme. The suitability levels for different factors for identifying potential sites for CO 2-ECBM implemented in this GIS-DSS beta version are summarized in Table 2. Combining the information in Table 2, the suitability level (
S ) for each i cell is given as:
n
∑ wi∗V i S=
1
n
(1)
∑ wi 1
where:
– wi is the relative importance weight for each input layer (as explained in Table 2); –
Vi is the suitability value for each cell in each input layer.
Table 2: Suitability levels for different factors for identifying potential sites for CO2-ECBM influence
1
2
3
4
5
coal body depth
HIGH
0 ─ -25
-25 ─ -100
-100 ─ -200
-200 ─ -400
-400 ─ -800
coal body thickness
HIGH
0 ─ 25
25 ─ 50
50 ─ 100
100 ─ 150
150 ─ 250
main faults distance
HIGH
0 ─ 50
environmental constraints
HIGH
protected areas
-
-
-
free areas
geology
MEDIUM
sedimentary deposits
weathered sedimentary rocks
strong sedimentary rocks
metamorphic rocks
magmatic rocks
soil use
MEDIUM
urbanized areas, mining areas, wetlands,...
agricultural fields (permanent crops)
agricultural fields (not permanent crops)
elevation
MEDIUM
500 ─ 300
300 ─ 150
150 ─ 50
50 ─ 25
25 ─ 0
slope
LOW
90% ─ 45%
45% ─ 30%
30% ─ 15%
15% ─ 5%
5% ─ flat
50 ─ 100
100 ─ 200
mediterranean scrub
no vegetation
CONCLUSIONS AND FUTURE DEVELOPING In the CO2-ECBM technologies field, the geological reservoir evaluation must consider a large amount of parameters, often very heterogeneous. The variety and variability of these parameters make each evaluation unique and, for this reason, it is impossible to give some “universally-valid” guidelines to help the Decision Maker in the planning procedure. For these reasons it is not advisable for the Decision Maker using only his own experience rather it would be suggested the use of a Decision Support System (DSS). The proposed combination of DSS and GIS technologies allows the realization of very powerful tools able to identify and solve the technical problem of the location of geological reservoirs able to contain safely large quantities of gases. The preliminary evaluation of potential CO 2-ECBM sites carried out by the analysis with GIS-DSS enabled the delimitation, inside the central zone of the Sulcis Coal Basin, of the most suitable areas responding to the established requirements (Figure 21). The green-highlighted areas in the GIS-DSS output map, are those who have the best suitability values for the purposes of the CO 2-ECMB responding to the outlined evaluation criteria. The current GIS-DSS version offers an appropriate processing and evaluation tool geographic, geological, structural information and, especially, It is a very flexible tool for suitability analysis implementation. In fact, the possibility: •
to vary the number or the typology of the input layers to be considered for the final evaluation; and
•
to modify the weights of each input layer
•
to evaluate and to compare different scenarios. The proposed version of the GIS-DSS is still in development in order to consider a larger number of
input themes and parameters
Figure 21: Two different evaluations obtained considering different influence for the “main faults distance” parameter.
ACKNOWLEDGEMENTS This research work would not have been possible without the essential and gracious support of ENEA. REFERENCES [1] International Energy Agency, (2009). Key World Energy Statistics. France 2009, pp. 82 [2] International Energy Agency, (2009). CO2 emission from fossil fuels. France 2009, pp. 82 [3] White C.M., Smith D.H., Jones K.L., Goodman A.L., Jikich S.A., LaCount R.B., DuBose S.B., Odzemir E., Morsi B.I., Shroeder K.T., (2005). Sequestration of Carbon Dioxide in Coal with Enhanced Coalbed Methane Recovery – A Review, in: Energy & Fuels Volume 19, N.3, May-June 2005, pp. 599-724 [4] Fadda A., Ottelli L., Perna G., (1994). Il Bacino Carbonifero del Sulcis. Geologia, Idrogeologia, Miniere. Edisar – Cagliari, pp.152. [5] Assorgia A., Brotzu P., Callegari E., Fadda A., Lonis R., Ottelli L., Ruffini R., Abbrate T., (1993). Carta Geologica del Distretto Vulcanico Cenozoico del Sulcis (Scala 1:50.000). [6] Valera R., (1966). Iglesiente. Il campo di frattura: Relazione Conclusiva. Resoconti [7] Mazzella A., (2009) L'utilizzo di tecnologie GIS per la ricostruzione tridimensionale di corpi geologici profondi: il caso del Bacino Carbonifero del Sulcis (Sardegna SW). Resoconti dell'Associazione Mineraria Sarda. Iglesias, vol. CXII/CXIII, pp. 87-95, ISBN/ISSN: 0376-2130.
[8] Mazzella A., Mazzella A., (2010). A combination of TIN meshing, geometric extrapolation and linear interpolation for deep ore bodies modeling: case of Sulcis Coal Basin (Sardinia, Italy). Proceedings of IAMG 2010, Budapest 28 August-2 September 2010, pp. [9] Carta M., Del Fa C., Agus M., Carbini P., (1978). Studio di una campionatura del I fascio della miniera di Seruci del bacino carbonifero del Sulcis., Resoconti dell’Associazione Mineraria Sarda, n° 1, pp. 111-169 [10] Carta M., (1976). Note sulle miniere carbonifere del Sulcis., La Programmazione in Sardegna, Anno 10, Novembre-Dicembre, pp 3-18 [11] Agus M., Garbarino C., (1986). Solfo organico negli strati di carbone di un foro di sonda del bacino del Sulcis (Sardegna sud-occidentale)., Rendiconti Società Italiana di Mineralogia e Petrologia, Vol. 41, pp. 6974 [12] Agus M., Carta M., (1976). Delle caratteristiche minero-petrografiche, chimiche e tecnologiche del carbone della miniera Seruci-Nuraxi Figus., Atti della Facoltà di Ingegneria, Vol. 7, Anno IV, n° 2, pp. 73-107 [13] Agus M., Cara S., Garbarino C., Ottelli L., Tocco S., (1989). Distribuzione del solfo nei livelli di carbone della I vena del bacino carbonifero terziario del Sulcis (Sardegna SW)., Congresso Internazionale di Geoingegneria Torino, 27-30 Settembre 1989, Vol. 1, pp. 323-329 [14] Diana G. F., (1985). Contributo alla conoscenza del Bacino Carbonifero del Sulcis: ricostruzione GeoStrutturale dell’area di Seruci-Nuraxi Figus. Implicazioni e considerazioni sullo sviluppo delle future coltivazioni. Tesi di Laurea in Ingegneria Mineraria, Università degli Studi di Cagliari, relatori Prof. Tocco S., Dott. Ottelli L. [15] Tocco S., Marcello A., Mazzella A., Naitza S., Pretti S., Valera P., Valera R., (2008). Carta Metallogenica delle Georisorse della Sardegna (1:250.000)
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 CARBON MANAGEMENT
Effect of rock composition on mineralization in sequestration Prashanth Mandalaparty, Graduate Research Assistant, University of Utah. Chemical Engineering Dept, University of Utah, 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112 Email: [email protected] Phone: 3365099709 Dr.Milind Deo, Professor, Chemical Engineering, University of Utah. Chemical Engineering Dept, University of Utah, 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112 Email: [email protected] Phone: 8016520451 Dr.Joseph Moore, Research Assistant Professor, Energy and Geo-science Institute, University of Utah. Energy and Geo-Science Institute, University of Utah, 423 Wakara Way suite 300, Salt Lake City, UT 84108.
Email: [email protected] Phone: 8015856931
Abstract The reactive behavior of pure CO2 with limestone, sandstone, arkose and peridotite was examined in this study. The experimental apparatus consists of series of high-pressure reactors with pure CO2 as feed gas at 1000C to evaluate the dependence of kinetics and mineralogical changes on rock composition. Arkose was observed to exhibit the highest tendency of participating in these reactions, which can be attributed to the geochemical complexity of its initial mineral assemblage. Layers of calcite were seen growing on the surface of the arkose. Analcime deposits are almost omnipresent either occurring as large
connected aggregates or as deposits on surfaces of other minerals. Ankerite and calcite deposition were observed as amorphous mass intergrown with starting minerals. Continuous dissolution of limestone was seen with the release of CO2 gas, indicated by the increasing pressure in the reactor (formation of a gas chamber), which occurs due to the lack of any source of alkali to buffer the solution. With sandstone, there was slight increase in pH due to dissolution of feldspars. The rate of carbonation of feldspars is pretty slow compared to the dissolution. Hence the precipitation of carbonates at the measured time scale is not evident. In peridotite experiments, carbonation of peridotite forming calcium and magnesium carbonates along with serpentine was evident. Hence, arkose has the geochemical complexity for permanent sequestration of CO2 as carbonates and is an ideal choice. Geochemists workbench (GWB) was used for kinetic modeling of these reactions. A full factorial statistical analysis was carried out to identify the most sensitive parameters (kinetic rates and reactive surface areas) in the arkose. The results from simulations indicate much rapid pace of these reactions when compared to experiments. Hence, caution should be exercised when using the calculated rates of reaction for making long-term process predictions. The study provides useful laboratory data (with model comparisons) when considering CO2 sequestration in different geologic The brine chemistry results for the experiments with different initial rock compositions are presented in this report. These results complement the changes in the rock chemistry.
Introduction Carbon dioxide is the most abundant GHG (64%). The atmospheric concentration of CO2 has increased from 280 ppm during the preindustrial period to 385 ppm with about half this increase having occurred since the mid 1960’s. In the United states majority of the CO2 emissions are from power plants which account for about 40% of the total emissions (IPCC report 2007) and it was identified before by Holloway et al 1997. Carbon dioxide sequestration appears to be an important potential method by which emissions into the atmosphere can be reduced. In this method anthropogenic carbon dioxide is disposed into geologic formations including saline aquifers, depleted oil reservoirs (CO2 EOR) and unmineable coal seems (Enhanced coal bed methane recovery). These formations are widely available and also in close proximity to majority of the point emission sources
(Holloway 1997). Injection of CO2 deep underground is particularly promising because deep sedimentary formations have the potential to retain CO2 in the subsurface for thousands to millions of years (Bachu 2002). In the United states the capacity of deep saline aquifers is greater than any other geologic formation with very high estimates (Bergman and Winter 1995) and they are also found within close proximity to power plants (Bachu 1994). Aquifers with salinities exceeding 10,000mg/l total dissolved solids are excluded by the U.S. Environmental Protection Agency as underground sources of drinking water (Xu et al2005). Hence these form a primary target for the eventual disposal of CO2. The temperature in these aquifers (100C-200C) varies greatly depending on the depth and also the local geothermal gradients (Bachu 2002). Also CO2 is injected at depths greater than 700m to ensure that it stays in the super critical form because its critical point is approximately 304 K and 1073 psi (Span and Wegner 1996) and also is soluble in water. The notion of CO2 disposal in aquifers has been discussed in the literature with specific aquifers as target like The Netherlands (Lohuis 1993) and the Alberta Basin, Canada (Gunter 1993 Bachu 1994 Perkins and Gunter 1995, Gunter 1996 and 1997 Bachu 1996). Geologic formations, which are artificially charged with CO2, are called Carbon repositories (Kaszuba et al 2005). Once injected CO2 being more buoyant than water will rise until it hits an impermeable membrane or low permeability seal. This is called structural or stratigraphic trapping of CO2. Then depending on a rate controlled by several factors such as the rate of CO2 injection, the rate of CO2 dissolution into the pore water, the surface area available for the reaction and the rate of diffusion of the CO2 into the pore water away from the pore water- CO2 interface, CO2 dissolves in the formation water (Allen et al 2005). This process is termed solubility trapping. Then CO2 is trapped as a discontinuous phase (either in supercritical form or as dissolved CO2) in the pores of formation leading to capillary trapping. Then CO2 forms carbonic acid, which dissociates to form carbonate and bicarbonate ion leading to ionic trapping of CO2. Finally these anions react with the cations resulting from the dissolution of primary minerals, due to decreased pH, which in turn leads to precipitation of carbonates permanently sequestering
CO2 (Bateman et al 2005). Hence logically, mineral precipitation can be termed sequestration where as other mechanisms can be termed as storage (Kaszuba et al 2005). Less importance has been given to mechanisms by which CO2 is actually trapped as immobile mineral phases. Kaszuba et al (2003) carried out a study to determine the extent of fluid rock interactions in addition to carbonate mineral precipitation that may occur in an experimental system that simulates geologic storage and sequestration of CO2. They reported the precipitation of magnesite, siderite and analcime. Kaszuba et al (2005) then carried out a study to analyze the effect of CO2 injection on fluid rock reactions and also assess the effect of these geochemical interactions on the integrity of the cap-rock. Shale was used to model the aquitard. Such analyses for the integrity of cap-rock had been done previously by Lindeberg (1997) using simple numerical models (Darcy equation and Fick’s law). They attributed the leakage of CO2 to gravity migration with subsequent release through sub vertical fractures and faults. A few experimental studies examined geochemical reactions in a saline aquifer in response to CO2 injection. Gunter et al 2000 and Perkins and Gunter 1995 carried out numerical geochemical modeling studies that incorporated kinetic laws and some studies combining experiment and modeling, in which dissolution of silicate minerals in brine and precipitation of carbonate are reported. The dominant state of carbonic acid with respect to pH plays an important role in governing the factors that lead to formation of carbonates in aquifer (soong et al, Drever 1988, Drucken miller et al 2005). Drucken miller et al reported the formation of calcite at pressures of around 600psi and at 150C. Rosenbauer et al reported the affect of presence of So4- to the geochemical reactions and also the changes it brings about in the solubilities and the porosities of the repository. One of the main factors governing this complex geological environment is the initial mineral assemblage. It has been stated in the literature that the type of rock in the reservoir also governs the type of the trapping mechanism of CO2. For example in a limestone reservoir the principal method of trapping is ionic trapping and structural trapping. Inorder to verify this hypothesis the analysis of these sequestration reactions with both the rock and brine chemistry has been provided in this study.
Figure 1: Schematic diagram of the experimental setup.
Experimental setup The experimental apparatus (Figure 1) consists of series of stainless steel reactors made of 316-grade stainless steel, which were rated for 4000psi at 600C. The reactors were pressure sealed with high-pressure SWAGELOK fittings. The reactor has provision for retrieving the rock sample without disturbing it through a detachable cap (Swagelok fitting) at the bottom. The CO2 was pressurized into the reactor using a single cylinder high pressure reciprocating DB Robinson pump. The flow of CO2 into the reactor was controlled using high-pressure needle valves. Syphon CO2 (CO2 in supercritical form) was used for these reactions. High purity nitrogen was used to pressure test the entire setup at 3000psi. The temperature was controlled using a bench top temperature controller with SPECVIEW as the interface via K-type thermocouples. Brine was prepared from laboratory grade NaCl with its initial composition shown in Table2. The ions in addition to Na and Cl might be due to the impurities in the salt used. High purity nano-filtered de-ionized (DI) water was used to prepare the brine samples. The rock analysis was carried out using X-ray diffraction XRD, Scanning electron microscope and EDS analyses. The advantage of using such apparatus over the previous experimental
described in Seyfried et al which was adopted by Kaszuba et al (2003,2005) and Rosenbauer et al (2005), is that we can correlate the changes in mineral chemistry with the changes in brine chemistry at each stage the sample is collected, which gives us a comprehensive picture of the geochemical interactions taking place. The changes that the occur due to sampling ie., the shift in the equilibrium between CO2 and brine, the geochemical reactions occurring at the brine-mineral interface are also not disturbed. Since this is a closed system there is a one-one relationship between brine chemistry and mineral chemistry. The changes we see in brine chemistry should correspond to the changes we see in mineral chemistry. This experimental setup helps us in evaluating that.
Experiment conditions The different sets of experiments are tabulated in Table 1. All the rocks were crushed to 100µm and mixed to construct the arkose. All the minerals were angular to circular in shape and were ranging from 40-100µm in size. Expt
Rock used
Temperature
Pressure
Set A
Limestone
1000C
2000 psi
Set B
Sandstone
1000C
2000 psi
Set C
Peridotite
1000C
2000 psi
Set D
Arkose
1000C
2000 psi
Table 1: Experimental conditions
All the experiments were carried out with 3g of rock. For the Arkose, the initial sample was prepared by mixing equal proportions of rock (0.5 grams each).
After the
experiment, the sample was retrieved from the reactor by detaching the bottom (wherein
the sample rests in a cap) and carefully taking the sample out without disturbing the precipitants on the surface. The sample was dried overnight at 500C, temperature not high enough to alter the crystal structure of any minerals. The sample is then retrieved, cooled down and divided into equal parts for XRD and SEM/EDS analyses. The quantitative estimates of the minerals before and after the experiments were obtained by XRD. Since very small amounts of precipitates were formed, any attempt to quantify produced solids proved difficult. Hence the changes in the initial mineral assemblage and the mineral assembly after the reaction were calculated. These changes can be corroborated with the precipitated minerals seen in the SEM/EDS analyses and a reaction mechanism leading to the precipitation of new minerals can therefore be postulated. After the experiments the brine samples were diluted and filtered using Whatman 40 filter paper (retention capacity of 8µm) in a vacuum filtration setup. The sample was then divided into sub equal fractions for all cation and anion analyses. The sample was acidified a little by addition of sulfuric acid to prevent any further precipitation. All the dissolved cations were analyzed using ICPMS and dissolved anions by ion chromatography. Table 2 gives the initial composition of the brine Na
Mg
K
Ca
Al
Mn
Fe
Ba
Si
S
Cl
mg/L
mg/L
mg/L
mg/L
µg/L
µg/L
µg/L
µg/L
mg/L
mg/L
mg/L
23032
1
<6
<4
<8
<1
54
<2
0.4
<6
26542
Table 2: Initial composition of the brine
Results and discussion Set A: Experiments with pure Limestone Initial XRD and SEM analysis (Figure 2) of limestone revealed it to be 99.4 calcite and 0.6% quartz . Hence the sample was relatively pure limestone. During the experiment, which was carried out for a period of 42 days the pressure increased to 2060 psi after 17 days and 2090 psi after 42 days. This can be attributed to a four-step process as indicated in the order of the reactions below.
(Eq 1)CO2 (g)
CO2 (aq)
(Eq 2)H2O + CO2 H2CO3 (Eq 3)H2CO3 H+ + HCO3(Eq 4)CaCO3 +H+ Ca2+ + HCO3The carbon dioxide dissolves in water to form carbonic acid (Eq1 and Eq 2). This leads to a decrease in pH. pH decreased from a value of 5.4 to about 4.9 after 28 days and 4.7 after 42 days. Because of this increased acidity calcite, which is sensitive to changes in pH undergoes dissolution to release CO2 (Eq 3 and Eq 4 ), which causes the increase in pressure inside the reactor. The absence of any silicate minerals to buffer the brine solution in the reactor rules out the possibility of any precipitants and also mineral sequestration of CO2. The dissolution patterns are almost omnipresent (Figure 3) through out the sample when the SEM analysis was carried out.
Figure 2: SEM analysis of initial calcite
Figure 3: SEM analysis of the reacted calcite after 42 days
140
Concentration mg/l
120 100 80 Ca ion
60
Mg ion 40 20 0 0
17
28
42
Time(Days)
Figure 4: Principal changes in brine chemistry in the Limestone experiments
The primary ions tracked here are Ca ion and Mg ion (Figure 4). The Mg ion is derived from
dolomite, which was present as traces in the initial mineral assemblage. The
continuous increase in the concentrations of both the principal ions indicates dissolution of calcite and dolomite. The absence of any alumino silicate mineral to buffer the carbonic acid rules out the possibility of precipitation of any carbonates in a limestone aquifer. The primary modes of sequestration in these aquifers are structural trapping and ionic trapping. Set B experiments with Sandstone Initial XRD analysis of the sandstone revealed 44% calcium feldspar, 26.2 % Na feldspar and the rest potassium feldspar. During the experiments pressure was constant around 2000 psi after 17 days remained stable for the rest of the experiment. The pH increased from a value of 5.4 to about 5.9 after 28 days and 6.4 after 42 days. The principal reactions in this case can be formulated as follows. (Eq 5) 2H+ + CaAl2Si2O8 + H2O (Eq 6)CaAl2Si2O8 + H2CO3 + H2O (Eq 7) NaAlSi3O8 + 3 H2O
Ca2+ + Al2Si2O5(OH)4
CaCO3 + Al2Si2O5(OH)4 NaAlSi2O6.H2O + H4SiO4
(Eq 8) 2KAlSi3O8 + 9H2O + 2H+
Al2Si2O5(OH)4 + 2 K+ + 4H4SiO4
Dissolution of feldspar buffers the brine (Eq 5), which turned acidic because of the formation of the carbonic acid. This leads to the next sequence of reactions, which is the carbonation of calcium feldspars (Eq 6), which leads to the precipitation of calcite and kolinite. At these temperatures and pressures albite undergoes phase change (Eq 7) to precipitate analcime which was detected in the XRD analysis. Dissolution of potassium feldspars (Eq 8) also leads to precipitation of kaolinite (Figure 7). The primary product of these dissolution reactions is silica. This leads to the brine becoming saturated and in some cases supersatured with silica. Hence when the samples are retrieved for analysis this silica undergoes heterogeneous deposition as amorphous silica on the surface of other host minerals, which was evident in the EDS analysis on all reacted samples. There was also evident precipitation of halite on these samples (Figure 5 and Figure 6). Halite precipitates when the sample is prepared for analysis, ie., drying overnight.
Figure 5: EDS analysis of halite
Figure 6: Halite chunk
Figure 7: Deposition of kaolinite on plagioclase feldspar
100 90
concentration mg/l
80 70 60 "Ca "
50
"Al "
40
"K"
30
Si
20 10 0 0
17
28 Time (Days)
Figure 8: Principal changes in brine chemistry
42
Figure 8 shows the trend of Ca, K , Al and Si in the brine with plagioclase feldspar and microcline being the primary alumino silicates in the initial mineral matrix.
No
precipitation of carbonates was found in these experiments. The decrease in Al ion indicates its precipitation in solid phase, which is evident from the precipitation of kaolinite which is aluminosilicate hydroxide. Analcime was detected in XRD but its precipitation was not evident in the reacted samples.. Si concentration increases continuously in the samples because of the dissolution of the alumino silicate minerals. Set C: Experiments with Peridotite Peridotite is composed largely of the minerals Olivine [(Mg,Fe)2SiO4] and pyroxene [(Ca,Mg,Fe)2Si2O6] which react with CO2 and H2O near the earth surface to form hydrous silicates (Serpentine), Iron oxides (magnetite) and carbonates (calcite, magnesite and dolomite). These reactions can be formulated as: (Eq 9) 2Mg2SiO4
+
Mg2Si2O6 + 4 H2O
=
2Mg3Si2O5(OH)4
(Mg-olivine)
(Mg-pyroxene)
(serpentine)
(Eq 10) Mg2SiO4
+ 2CO2
=
(Mg-olivine) (Eq 11) Mg2SiO4
2MgCO3
(Magnesite)
+
SiO2
(quartz)
+ CaMg2Si2O6 + 2CO2 + 2H2O = Mg3Si2O5(OH)4 + CaCO3 +
MgCO3 (Mg-olivine)
(Ca-Mg-pyroxene)
(serpentine)
(calcite)
(magnesite) Natural carbonation of peridotite by weathering and low temperature alteration is common. Evidence for natural, low-temperature hydration and carbonation of peridotite ca be found in springs and associated travertines inn catchments composed of peridotite and in outcrops of altered peridotite with abundant carbonate veins. High alkalinity, stable isotope ratios and formation of travertine and carbonate cemented conglomerates in springs indicate ongoing serpentenization involving meteoric water at low temperature. Ground water reacting with peridotite in near-surface, open systems forms water rich in
Mg- and HCO3-, which when on continuous reaction with peridotite leads to precipitation of abundant magnesite and dolomite as veins. The resulting waters become progressively richer in Ca- and OH-. These waters emerge near surface to mix with Mg-HCO3- waters or react with the atmosphere; they precipitate abundant calcite and dolomite in near surface veins, and carbonate cement in unconsolidated sediment and travertine. Enhanced natural processes may provide an important alternative to mineral carbonation. Peridotite was obtained as big green crystalline granules from the deposits in the Samail Ophiolite, Sultanate of Oman. The distinctive green color was due to the presence of significant portions of Iron in the crystalline structure. Kinetics of these carbonation reactions is very slow unless olivine and serpentine reactants are ground to fine powder heated and hels at elevated pressures and temperature. Hence these granules were ground to very fine powder. A size distribution analysis revealed that major portion was within the range of 60-100 microns in size. 3 grams of peridotite was fed into a 40CC reactor along with 20CC of brine prepared by mixing 3grams of laboratory grade NaCl in 20CC of distulled water and brine was allowed to saturate the sample for about 2 days. CO2 was then fed into the reactor at a pressure of 2000 psi and the temperature was maintained at 100C. The reactor was isolated and the experiment was carried out for 47 days. The pressure in the reactor first decreased and then stabilized around 1980 psi for the rest of the experiment with minimal fluctuations, which can be attributed to changes in ambient conditions. Initial analyses of the sample revealed that Olivine was the major component in the sample. The sample was composed of 97.8 wt% olivine and 2.2% pyroxene (Figure 9 and Figure 10) . Hence we can conclude from the reactions mentioned above that carbonation of this sample will mainly yield magnesite and serpentine. Because of the presence of access CO2 in the reaction setup, carbonation of peridotite will dominate hydration and hence magnesite precipitation should be omnipresent and formation of serpentine should be in trace amounts.
Figure 9: EDS analyses of the initial sample.
Figure 10 : SEM analyses of the initial rock
Figure 11: EDS analyses of the reacted rock.
The EDS analysis of the reacted rock (FIgure 11) indicates the presence of MgCO3 in the sample. This occurs due to the carbonation of the peridotite to form magnesite and silica. This process is shown in the following reactions. The presence of silica is also revealed in the analysis. The SEM analysis of the product revealed the precipitation of magnesite as orthorhombic crystals (Figure 12) with pitted rough faces. Depending on the shape of the crystal it can be concluded that by the time the experiment the brine was under saturated with respect to magnesite. The dissolution of the alumino-silicate minerals produces silica. This silica precipitates out (Figure 13) mainly as amorphous silica on the products when the experiment is terminated and the temperature is decreased because the brine is supersaturated with respect to silica during the latter stages of the experiment.
Figure 12: Growth of magnesite crystal in the reacted sample
Figure 13: Trace amounts of silica mostly as fur on host peridotite
100 90
Concentration mg/l
80 70 60 Mg
50
Ca
40
Si
30
Fe
20 10 0 1
2
3
4
Time (Days)
Figure 14: Principal changes in brine chemistry in Peridotite experiments
The initial increase in the primary ions Mg, Ca and Si (Figure 14) indicates the dissolution of the aluminosilicates, which leads to a decrease in the pH of the solution. This leads to the carbonation reactions (favored at higher pH) dominating the hydration reactions. The carbonation of peridotite leads to the precipitation of magnesite and siderite. The steep increase in the concentration of Si is a result of the dissolution of the silicate minerals. The reacted samples were characterized with the deposition of silica, which precipitated from the brine when the reactor was degassed. SET D: Experiments with Arkose All the participating minerals (constituents of arkose or dirty sand) (Table 3) are seen to be participating in the reaction. The major change is seen in the composition of dolomite and feldspars (microcline and andesine) with a small change in calcite and chlorite. The change in quartz composition was not expected and dissolution of quartz usually doesn’t occur at pH 5. But as the pH increases and hits a pH of about 7 the H4SiO4 dissociates leading to the dissociation of quartz (Dreever 1988) . While dissolution patterns dominate the XRD analysis in the sample after 62 days precipitation patterns dominate after 134 days. Andesine, calcite and microcline are found precipitating while dolomite and quartz
composition decreased in the final experiment. These results are in good correlation with the brine chemistry analysis and also the SEM/EDS analysis for the rock explained in the sections to follow. Rock
Quartz
Andesine
Dolomite
Chlorite
Microcline
Calcite
Formula
SiO2
NaxCayAlSi2O8
CaMg(CO3)2
(Fe, Mg, Al) 6(Si, Al) 4O10(OH) 8
KAlSi3 O8
CaCO3
Class
Silicates
Silicates
Carbonates
Silicates
Silicates
Carbonates
Group
Quartz
Feldspar
Dolomite
Chlorite
Feldspars
Calcite
Specific gravity
2.65
2.68-2.71
2.86
2.6-3.4
2.5
< 2.5
Hardness
7
6-6.5
3.5-4
2-3
6-6.5
3
Table 3: Composition of the arkose
Figures 15 and 16 show the SEM analysis of the reacted sample after 62 days and 134 days. The calcite in the initial samples serves as a standard to examine the precipitation and the dissolution patterns in the samples at different stages of the experiments. The precipitate from the experiment for 134 days was chosen because it was a reaction in which calcite precipitated as determined by the XRD. Figure 15 shows the pronounced dissolution patterns in the sample after 62 days. Carbon dioxide reacts with water to produce carbonic acid (Eq.12). And carbonic acid dissociates to bicarbonate ion in the equilibrium aqueous phase (Eq.13). (Eq.12) CO2 (g) CO2 (aq) (Eq.13) H2O + CO2
H2CO3
(Eq.14) H2CO3
H+ + HCO3-
(Eq.15) HCO3- H+ + CO3 2Equation 12 is the dissolution of CO2 in water, which is highly dependent on temperature pressure and salinity (Soong et al. 2002, Drever 1988). CO2 solubility increases with
increase in pressure, decreases with increase in temperature and salinity. CO2 then reacts with water to from carbonic acid. This acid dissociates to liberate protons and bicarbonate ion. The sequential reactions in Eq.13 and 14 result in decreasing pH value (about 3.4). Dissociated hydrogen ion can dissolve calcite (mainly CaCO3) to produce calcium ion and bicarbonate (Eq. 16). (Eq.16) CaCO3 + H+ Ca2+ + HCO3Dissolution occurs mainly by the formation of few deep etch pits and some shallow ones, almost at the same position in the initial dissolution of the surface. Dissolution occurs very fast which is followed by precipitation as seen in Figure 16 (Experiment carried out for 134 days). The cations thus liberated (due to the increased acidity) react with the carbonate ion to precipitate secondary carbonates thus consuming CO2 in the process. (Eq.17) Ca2+ + CO3 2- CaCO3 Calcite formation can also be postulated by another mechanism: dissolution of the feldspars in the initial mineral matrix. For example arkosic sandstone as a carbonate reacts with hydrogen ion to precipitate calcium carbonate (Eq.18-20). (Eq.18) 2H+ + CaAl2Si2O8 + H2O Ca2+ + Al2Si2O5(OH)4 (Eq.19) Ca2+ + HCO3- CaCO3 + H+ (Eq.20) CaAl2Si2O8 + H2CO3 + H2O CaCO3 + Al2Si2O5(OH)4 The above reaction also shows kaolin deposition, which was evident in the reacted samples. This indicates that plagioclase feldspar is an active reactant in this geochemical repository. Calcite is seen growing as tightly packed polymorph of calcium carbonate. These layers and crystals of calcite are seen growing as amorphous mass intergrown with the starting minerals. These calcite crystals (Figure 17) are highly irregular in shape and show no consistency in size and shape. One calcite crystal had a width of about 10mm and most of the particles appear much smaller than this size. Hence it can be concluded not all precipitates were collected when the brine was filtered (with a filter paper of 8 mm
retention capacity). Figure 18 shows deposition of a new phase analcime or kaolinite on the reacted surface. The formation of analcime can be formulated by the transformation of albite (from the plagioclase feldspar) (Eq.21) NaAlSi3O8 + 3 H2O NaAlSi2O6.H2O + H4SiO4 Kaolinite formation from albite (from the plagioclase feldspar) can be formulated as (Eq.22) NaAlSi3O8+CO2+5.5H2O Na+ + HCO3- + 2 H4SiO4 + 0.5 Al2Si2O5(OH)4 Analcime deposits occur as large connected aggregates or as deposits on the surfaces of other minerals (Quartz). Kaolinite was not identified in the XRD analyses because it shares the same primary peak with one of the minerals in the initial assemblage, chlorite. But it was identified in the reacted assemblage. The growth was seen as aggregated units in the interstitial spaces between the primary minerals especially quartz.
Figure 15: SEM image showing calcite in the sample showing dissolution after 62days.
Figure 16: SEM image showing the growth of layers of Calcite after 134 days indicative of mineral precipitation reactions.
Figure 17: Calcite crystals growing in interstitial spaces between plagioclase feldspar.
Figure 18: SEM image showing the growth of new phase probably analcime on the surface.
Figure 19: Changes in the concentration of principal ions
Ca ion concentration shot up by about 90% after 14 days and continued to increase for 42 days after which it decreased with the final concentration approximately 47% less than the highest concentration measured. The K ion concentration exhibited similar trend to that of Ca ion with the concentration shooting up abruptly and continuing to increase till 62 days after which it decreased. The increase in the Ca ion concentration can be attributed to the dissolution of the primary carbonate minerals calcite and dolomite liberating Ca ion into the brine. The K ion concentration increase was due to the dissolution of the microcline, which is potassium feldspar, and the silicate dissolution reactions are the fastest in a low pH geochemical system. The Mg ion concentration increased by approximately 52% at 27 days and continued to increase until 62 days after which it decreased with the final concentration 42% lower than the initial concentration measured at 27 days. This increase was mainly due to the dissolution of dolomite and chlorite. Iron (Fe) concentration decreased throughout the experiment and ended up approximately 27% lower than the initial concentration. The decrease in the concentrations of the Ca, Mg and K ions by 47%, 42% and 27% in the latter stages of the experiment are indication of the fact that new minerals with the primary composition of these ions are precipitating in the solid phase. The Si concentration followed a unique trend. It increased by 14% after 27 days then decreased slightly (2%) and then increased by about 32% till the termination of the experiment. Since the increase of 2% was within the experimental and analytical uncertainty, the concentration can be considered stable from 27 days to 42 days. The Si concentration increase can be attributed to the dissolution of feldspars, microcline and andesine, which are the most sensitive minerals to decrease in pH (Dreever, 1988). Hence the silicate dissolution dominates the geochemical reactions in the system. This is evident from the fact that there is a silica coating on all the samples analyzed in the SEM with a distinctive Si peak in the EDS analyses. This is due to the deposition of Si from the brine on to the rock when the solid sample is dried prior to the analysis.. The samples collected at the end of 62 days and 134 days were selected to represent these changes and Ca ion was chosen as the principal ion undergoing the change. At the end of 62 days and also through the first 62 days Ca ion concentration increased in the brine indicating the dissolution of primary carbonate minerals calcite and dolomite in the brine. It is clearly seen in the SEM analysis of the
rock sample that is collected at the end of 62 days where calcite dissolution pattern dominates. Then, from 62 days through 134 days, the Ca ion concentration decreases, which implies that Ca bearing minerals should precipitate in the solid phase. This is clear from the SEM analysis of the sample collected at the end of 134 days where layers and crystals of calcite are seen growing as amorphous mass intergrown with the starting minerals (especially quartz). The complex trend in the brine chemistry represents the geological complexity of the system. The presence of carbonates and feldspars favors the precipitation of carbonates, thus permanently sequestering carbon dioxide. The primary reactions are dissolution of carbonates, feldspars and carbonation of plagioclase feldspar and microcline. This set of experiments was characterized by precipitation of calcite, analcime and kaolin. These changes were complemented by the changes in brine chemistry.
References Bachu,
S.,
2002.
Sequestration
of
Carbon
dioxide
in
geological
media
in
response
to
climate
change:
road
map
for
site
selection
using
the
transform
of
the
geological
space
into
the
CO2
phase
space.
Energy
Convers.
Manage
43
(1),
87‐102.
Bacon
,
D,
H.,
Sass.
M
.B.,
Bhargava.
M.,
Sminchak.
J.,,
and
.
Gupta.
N.,
2009.
Reactive
Transport
Modeling
of
CO2
and
SO2
Injection
into
Deep
Saline
Formations
and
Their
Effect
on
the
Hydraulic
Properties
of
Host
Rocks.
Energy
Procedia
(1)
3283–3290.
Bachu,
S.;
Gunter,
W.D;
Perkins,
E.H.,
1994.
Aquifer
disposal
of
CO2
:
hydrodynamic
and
mineral
trapping.
Energy
Converse.
Manage.
35
(4),
269‐279.
Bateman,
K.;
Turner,
G.,
2005.
Large
scale
column
experiment:
Study
of
carbon
dioxide,
pore
water,
rock
reactions
and
model
test
case.
Oil
and
Gas
Tech
60,
161–175.
Bergman,
P.D.;
Winter,
E.M.,
1995.
Disposal
of
CO2
in
aquifers
in
the
U.S.
Energy
Conver.
Manage
36,
523‐526
Climate
Change,
2007
An
assessment
of
the
Intergovernmental
panel
on
climate
change.
http://www.ipcc.ch/pdf/assessment‐report/ar4/syr/ar4_syr.pdf
Druckenmiller,
M.
L.;
Maroto‐Valer,
M.
M.;
Hill,
M.,
2005.
Investigation
of
carbon
sequestration
via
induced
calcite
formation
in
natural
gas
well
brine.
Energy
and
Fuels.
Dreever,
J.I.,
1988.
The
Geochemistry
of
Natural
Waters:
Surface
and
Groundwater
Environments,
Prentice
Hall,
INC.
Gunter,
W.D.;
Perkins,
E.H.;
Hutcheon,
I.
2000.
Aquifer
disposal
of
acid
gases:
modeling
of
water
rock
reactions
for
trapping
of
acid
wastes.
Appl.
GeoChem
15
(8),
1085‐1095.
Gunter,
W.D.;
Perkins,
E.H.;
McCann,
T.J.,
1993.
Aquifer
disposal
of
CO2
rich
gases:
reaction
design
for
added
capacity.
Energy
Converse.
Manage.
34,
941‐948.
Jacquemet,
N.;
Pironon,
N.;
Marc,
J.
S.,
2008.
Mineralogical
changes
of
a
well
cement
in
various
H2S‐CO2
(brine)
fluids
at
high
pressure
and
temperature.
Environ.
Sci.
Technol
42,
282–288.
Kaszuba,
J.
P.
;
Janecky,
D.
R.
;
Snow,
M.
G.,
2005.
Experimental
evaluation
of
mixed
fluid
reactions
between
supercritical
carbon
dioxide
and
brine:
Relevance
to
the
integrity
of
a
geologic
carbon
repository.
Chem
Geol
217,
277‐293.
Kaszuba,
J.
P.
;
Janecky,
D.
R.
;
Snow,
M.
G.,
2003.
Carbon
dioxide
reaction
processes
in
a
model
brine
aquifer
at
2000
C
and
200
bar:
Implications
for
geologic
sequestration
of
carbon.
Appl
Geo.
Chem
18,
1065–1080.
Lindeberg,
E.,
1997.
Escape
of
Carbon
Dioxide
from
aquifers.
Energy
converse.
Manage
38,
238–240.
Moore,
J.;
Adams,
M.;
Allis,
R.;
Lutz,
S.;
Rauzi,
S.
2005.
Mineralogical
and
geochemical
consequences
of
long
term
presence
of
CO2
in
natural
reserviors:
An
example
from
the
Springerville‐St.Johns
field,
Arizona
and
New
Mexico,
U.S.A.
Chem
Geol
217,
365–385.
Moore,
J.;
Christensen,
W.B;
Allis,
R.;
Browne,
L.R.;
Lutz,
J.;
2004.The
Minrealogical
consequences
and
the
results
of
descending
acid
sulphates
water:
An
example
from
the
Karaha
–
Telaga‐Bodas
Geothermal
system,
Indonesia.
Can.Min
vol
42.
pp
1483‐1499.
Rosenbauer,
R.
J.;
Koksalan,
T;
Palandri,
J.L,
2005.
Experimental
investigation
of
CO2
‐brine‐rock
interactions
at
elevated
temperature
and
pressure
:
Implications
for
CO2
sequestration
in
deep‐saline
aquifers,”
Fuel
Process.
Tech
86,
1581– 1597.
Span,
R.;
Wegner,
W.,
1996.
A
new
equation
of
state
for
carbon
dioxide
covering
the
fluid
region
from
the
triple
point
temperature
to
1100K
at
pressures
up
to
800
MPa.
J.
Phys.
Chem.
Ref.
Data
25
(6),
1509‐1596.
Soong,
Y.;
Goodman,
A.
L.;
Hedges,
S.
W.;
Jones,
J.
R.;
Harrison,
D.
K.,
2002.
CO2
Sequestration
and
importance
of
pH.
Am.
Chem.
Soc
47,
43–48.
Seyfried
Jr.;
Janecky,
W.E.;
Berndt,
D.R.,
1987.
Rocking
autoclaves
for
hydrothermal
experiments,
II.
The
flexible
reaction
cell
system.
In:
Ulmer,
G.C.;
Barnes,
H.L.
U.S.
Department
of
Energy,
2007,
Carbon
Sequestration
Atlas
of
the
United
States,
National
Energy
Technology
Laboratories.
U.S.
Department
of
Energy,
2008,
Carbon
Sequestration
Atlas
of
the
United
States,
Version
2,
National
Energy
Technology
Laboratories.
U.S.
Department
of
Energy,
2009,
Storage
of
Captured
Carbon
Dioxide
on
Federal
Lands,
National
Energy
Technology
Laboratories,
DOE/NETL‐2009/1358.
Geological CO2 Storage in Coal-Bearing Formation Sohei Shimada *), Zhenjie Chai and Naoto Sakimoto, Graduate School of Frontier Sciences The University of Tokyo, JAPAN *) Corresponding author: [email protected] Abstract CO2 geological storage (CGS) has been recognized as an indispensible and cost-efficient abatement measure against the Global Warming due to the CO2 emission from large-scale energy-related sources. Within the CGS, disposal of CO2 into deep saline aquifer formation (DSAF) and injection of CO2 into deep unminable coal seam (UCS) for enhanced coalbed methane recovery (ECBMR) are two promising technologies which have been widely studies. Recently, the interest of geologists and policy-makers in a specific type of subsurface formation called deep coal-bearing formation (DCBF) is booming. DCBF is characterized by the generation of relatively thin coal seam between strata of other subsurface materials, mostly sandstone. It deserves special attentions because of its intrinsic saline-aquifer-like characteristics resulting from the huge amount of storage potential all around the world as well as the presence of coal which may reduce the risk of the leakage and offset the operation cost by ECBMR. In this study, a non-isothermal multi-phase multi-component fluid dynamics simulator was developed for the comprehensive study on CO2 migration and assessment on the performance of CO2 storage in DCBF. Modeling of the thermophysical properties of fluid and the geological properties are elaborated considering the trade-off between the simplicity and the accuracy required. The newly-built simulator was substantially verified by comparing the result of this simulator with the results calculated by other sophisticated simulation code for the test problems and it shown a good agreement. The simulator was further verified by simulating CO2 injection into saline aquifer formation and tracing the migration of CO2 under the consideration of non-isothermal effect. A reservoir models which are representative in the natural world were set up. Several evaluation indexes considering every aspect of CO2 storage were introduced. Sensitivity analysis on several key parameters associated with fluid dynamic in the porous media was conducted on the models. Keywords: CO2, geological storage, coal seam, coal-bearing formation, coalbed methane 1. INTRODUCTION The countermeasures to abate the Greenhouse Gas (GHG) emission in order to prevent and mitigate the aggravation of the threatening Global Warming have been proposed, proved by laboratory experiment, demonstrated by pilot test and eventually deployed to commercialized projects. Almost all of them can be summarized into 3 categories, i.e. clean/renewable energy shifting, efficiency improvement and carbon capture and storage (CCS), regarding the reduction of anthropogenic energy-related CO2 emission from the use of fossil fuel which has been recognized as such a major contributor to Global Warming that accounts for 69% of all CO2 emissions and 60% of all GHG emissions [1]. In detail for CO2 storage, due to the physical and chemical properties of CO2, disposal of CO2 into various subsurface formations (i.e. CO2 Geological Storage) has been suggested as an appropriate and promising mean for CO2 storage and sequestration associated with the state-of-art technologies. According to the scenario studied by IEA, CO2 geological storage would contribute a cumulative storage amount of 80Gt over 1
the period from 2000 to 2050 [2]. Various subsurface formations and strategies are proposed which has been proved suitable for CO2 Geological Storage. Within it, the storage in deep saline aquifer formation (DSAF) would possess a dominant share because of its enormous global storage potential (1000 GtCO2 compared to 675 GtCO2 for oil and gas fields and 3-15 GtCO2 for Coalbed Methane) according to the lower estimate of storage capacity by IPCC [3]) and its even distribution [4] all over the world resulting in the high possibility of proximity to some large-scale CO2 sources. These obvious merits have been prompting the scientists to conduct researches on the subject of DSAF storage in the last decade and several pilot projects have been performed in progress throughout the world as a sign of successful practical application. In addition to DSAF storage, the unminable coal seams (UCS) have been proposed to be an alternative for CO2 geological storage (CGS) owing to the profitable return from the methane gas recovery via the enhanced coalbed methane recovery (ECBM) that can offset the inevitable high operation cost, as well as the comparatively well-known strata structures. It has been verified that a coal seam with relatively smaller size is capable of storing large volume of CO2 securely because of the physical adsorption enables the strong bond forces between CO2 and the enormous void space distributed in the coal matrix (micropores). Another advantage of UCS storage is that, just like DSAF, most large coal-fired power plants were built in coalfields, which can save the cost for CO2 transportation. The researches on Enhanced Coalbed Methane Recovery (ECBMR) have been carried out broadly since the 1970s and the commercialized ECBM projects in U.S.A have proved the feasibility and large potential of this technology. However, although both of these two options are very attractive for CGS, neither of them is free of problem. Regarding DSAF storage, first, as mentioned above, the cost of drilling, injection and monitoring is too large to afford and due to the non-profitable property, it is relatively difficult to provoke the interest of private ventures in the investment of the projects ruling out the considerable supports from carbon credits or governmental incentive. Second, the uncertainties with respect to the storage technology and substantial CO2 transport and chemical reaction in the reservoir including the unfavorable contamination of subsurface environment and mobile groundwater during the large post-injection time span of ten-thousands of years remain in dispute. Third, the social awareness and risk communication with regard to DSAF storage have yet been discussed, which will also probably retard the deployment of the projects. On the other hand, it is certain that UCS can make profit for investors, but according to the recent studies [5], compared to the injection of N2-CO2 mixture stream with 50%-50% composition, the injection of pure CO2 will induce the dramatic swelling of coal matrix and decline of permeability (over two orders to the small magnitude of inherent anisotropic permeability) leading to the great loss of injectivity and consequently the reduced efficiency both for CO2 storage and CH4 recovery. Moreover, Underground Coal Gasification is an innovative booming technology for the economically feasible utilization of UCS, which may turn the role of UCS from potential CO2 storage site to the simply unconventional natural resources. The last but not the least, for both options the leakage of injected CO2 back to earth surface through various possible routines (e.g. faults, roof strata) is the most important risk that all the stakeholders associated with CCS are worried about right now. These issues concerning CO2 storage either in DSAF or UCS are the vital barriers against their practical implementations and the scientists are diligently seeking for the breakthrough. Is there any effective way to take the advantages of both DSAF and UCS storage while eliminating the respective weakness as well as possible? The answer has been given by the recent interest of the geologists and policy-makers on a specific type of subsurface formation, namely deep coal-bearing formation (DCBF). 2
DCBF is such a formation that several layers of the moderately thin coal seams separately generated between the strata with other subsurface materials, such as the sandstone. Some of the superiorities of such type of strata over individual DSAF or UCS can be imagined intuitively, such as the intrinsic sand-stone like characteristics which makes it similar to DSAF formation, evidence of coexistence of coal seams and saline aquifer formation in most sedimentary basins where coal seams have been discovered [7], and the ability of CO2 adsorption in coal matrix could play an essential role as seals to lower the risk of the CO2 leakage from caprocks and the reduced coal permeability could confine the migration of gaseous and aqueous CO2 to the local storage space. However, currently while the geological surveys on DCBF are being widely conducted, no one has ever performed any numerical simulation on this type of formation to investigate the impact on the CO2 storage in such integrated formation. Undoubtedly, there must be other mutual enhancement effects or mutually adverse interaction between the coal seam and other strata on the overall performance of CO2 storage corresponding to the different injection-production profiles, time scale and geological conditions and so far, these issues remain concealed. Based on the discussion above, the objective of this study is to develop a numerical simulator and apply it on the comprehensive study on CO2 storage in DCBF to reveal the real behavior of CO2 along with other side-substances (e.g. methane, nitrogen) inside a well-defined DCBF reservoir, and to assess the overall performance of CO2 storage. The newly-built simulator needs verifying. It is verified the validity by comparing the data of variables calculated by the simulator with some experimental data during the stage of modeling. The comparison of the results from this simulator with the results calculated by other advanced simulators will be done regarding the simulation with same specification. After verification, simulation will be applied to two reservoir models to carry out the sensitivity analysis and implement the various types of injection-production profile to make assessment on the performance of CO2 storage. 2. NUMERICAL SIMULATOR DEVELOPMENT AND IMPLEMENTATION 2.1 Characteristics of Simulating CO2 Storage into DCBF Geological Characteristics As mentioned in the previous Chapter, DCBF is such a specific formation where DSAF and coal seams co-exist in a sedimentary sequence. Therefore to successfully simulate the migration of fluids across DCBF, the geological properties of both DSAF and coal seams have to be taken into account. However, almost all of the geothermal and geomechanical properties with regard to the fluid dynamics and thermodynamics for these two types of formation are significantly different from each other. This results in the requirement for the establishment of an appropriate composite reservoir model capable of customizing each of them separately and accounting for all the key factors governing the process of CO2 geological storage. In the natural world, DCBFs are found in a layered manner with the DSAF strata forming between two coal seams. Unlike the normal saline aquifer (consists of mainly sandstone) where the permeability of strata is mainly in the range of 10 - 100 mD which is ten to hundred times higher than the average permeability of coal seam found in the same depth of subsurface, the permeability of coal-like mudstone and coal-like shale in DCBF formation sometimes is lower than the permeability of overburden and underlying coal seams, depending on the sedimentary effect
3
Characteristics of Components of Fluids CO2-Water-NaCl ternary system is commonly assumed to be the primary components of system subject to the physical processes in DSAF for simulation of CO2 sequestration, neglecting the relatively minor influence possibly induced other reservoir residual gases and rock materials. On the other hand, ECBMR simulations generally focus on the system of CH4-CO2-N2-Water. CH4 is the gas adsorbed initially in the coal matrix with the concentration over 90%, while carbon dioxide and nitrogen are the main injectants stemming from the respective enhanced effect on methane recovery and the similarity to the flue gas emitted from power plants. When considering both of the CO2 storage into DCBF and the methane production from DCBF, all these substances will be involved and the correct modeling for this five-component system is necessary to precisely determine the thermophysical properties of each component and their mixtures in every thermodynamic state corresponding to the pressure, temperature and their mutual interaction. Characteristics of Injection and Production Profile On account of the extraction of natural gas from coal seam, both injection and production should be addressed when conducting the simulation of ECBMR. Moreover, many kinds of the combination of injection and production patterns (i.e. vertical wells and horizontal wells) can be selected to optimize the production of CH4 or the sequestration of CO2 according to different geological conditions, which plays an important role in ECBMR process [5]. However the injection point can only be the target coal seam. As for the simulation of CO2 storage into DSAF, no production needs to be taken into account, and the injection pattern can be chose to be vertical or horizontal injection into the prospective sandstone formation only. In the case of DCBF simulation, not only the injection and production patterns can be changed like ECBMR, but the injection point can be changed as well resulting in the different CH4 recovery and CO2 sequestration performances. Characteristics of Fluid Dynamics Like the methodology applied on the system of components, the fluid dynamics implemented in the simulation of DCBF should be the combination of all the possible fluid transports in DSAF and ECBMR. The discrepancy between DSAF and ECBMR regarding the modeling of fluid dynamics is the concentration dependent diffusion due to the mechanism of sorption which is essential for the CH4 production and CO2 sequestration in the case of ECBMR. Base on the discussion above, the difference between the framework of simulation on DSAF formation and ECBMR is summarized in Table 2.1. Table 2-1 Comparison between the domain of DSAF and ECBM modeling
DSAF
ECBM
Thermodynamic
Fluid System of
Reservoir
Mechanism of
interest
modeling
fluid transport
Single
Convective mass
Free gas and
Pure CO2,
porosity
transfer
dissolution
No production
CO2-Water-NaCl
CH4-CO2-N2-Water
Dual/triple porosity
Sorption, convective mass transfer
state of stored CO2
Adsorption, free gas and dissolution
Injection-prod uction
Flue gas, methane gas recovery
4
2.2 Scope of Numerical Simulator Considering the requirements for the description of CO2 storage into DCBF mentioned above, the general scope of this numerical simulator comprises:
a. Multi-phase fluids: gaseous phase, aqueous phase and solid phase. b. Multi-component fluids: methane, carbon dioxide, nitrogen, sodium chloride and water. c. Non-isothermal system. d. Chemical reactions are neglected. e. Injection and production with different patterns and locations. In summary, the numerical simulator developed and used in this study is a 3-phase, 5-component simulator focusing on the fluid dynamics and neglecting the relatively complicated and time-consuming mechanism of mineral trapping 2.3 Conceptual Model Based on the above that the fundamental framework for simulations including the components of interest and the mechanism of transport/storage are completely different between usual DSAF formation and ECBMR studies and the mathematical modeling designated for each of them respectively, the conceptual structure of the model is shown in Figure 2.1 with the multi-phase multi-component flow profiles on the left and the injection-projection profile on the right
Figure 2-1 Conceptual structure of the overall model The whole system in this study, as depicted in Figure 2.1, comprises as many as three phases (i.e. Gas, Liquid, Solid) and five components (CO2, CH4, N2, NaCl, Water) on the assumption of non-isothermal condition. In Figure 2.1, each ellipse represents one single phase respectively except for the upmost one which stands for the gases in adsorbed state in the coal’s secondary 5
porosity (coal matrix) and the interporosity diffusive flow (between matrix and cleats) is treated as the single-phase sink (adsorption) and/or source (desorption) term in the governing equations of CO2, CH4 and N2 . The arrow lines along with the process denotations linking the pairs of phases account for the phase transitions and phase changes that may occur dominantly in the scope of this study. As for each ellipse (3 phases and one adsorption state), all the components that may appear in that phase/state to the extent of a certain significant concentration are collected in the upper halves and the components in bold characters stand for the rich components in that phase/state. Then, the lower halves of the circles account for the main massive transport (black) and heat transfer (red) of that phase/state in the porous media of DCBF from the macroscopic point of view. On the right part of Figure 2.1, two rectangles provide an image of injection-production profiles. Similar to the way of expression in phases, the upper parts consist of the components involved and the lower parts explain the transport mechanisms near the wellbores. The fluid flows from/to the wellbores are regarded as source and/or sink terms in the governing equation like sorption process. Since the simulation is run in the infinitesimal time step, the fluid system is presupposed to keep in the state of thermodynamic equilibrium locally inside every grid (which depends on the space discretization strategy) for every time step. However the exception is only made for the processes of sorption, where the quasi-steady approximation is applied for convenience. 2.4 Primary Variables To choose the proper primary variable, the phase change should be taken into account. During the simulation, the phase changes such as the appearance and disappearance of phases may occur in any cubic grid at any moment inside the reservoir model only if the local thermodynamic equilibrium meets the critical conditions. When phase change comes up, the set of primary variables used for the previous phase state fails to function properly for the new phase state in the current loop of Newton-Raphson iteration, which may lead to the breakdown of the program. Therefore the simulator program itself should be able to recognize such phenomenon promptly and precisely by implementing monitoring. Moreover, as soon as the phase change occurs, the simulator should be able to characterize the fluid flow and change primary variables to be used in the current loop of calculation so as to make the fluid adaptive to the sudden change in its thermodynamic state and properties. The selected primary variables must be capable of correctly deriving all thermophysical properties for the phases and components which are needed to assemble the governing equation system. Besides, the newly-selected set of primary variables should also pertain to the last updated set of primary variables. To fulfill such requirement, the skill named ‘primary variable switching’ frequently applied in the multiphase flow simulations is introduced to this study. The list of the shifting scheme and the shifting conditions is shown in Table 2.2. The variable represents the NaCl vapor in gas phase. Since the evaporation of halite is out of the scope of this study, this variable is set to zero when it becomes one of the primary variables
6
Table 2-2 List of primary variables set Phase state
Primary variables 1
2
3
4
5
6
Liquid only
Pl
T
Gas only
Pg
T
Phase
Liquid+gas
Pg
T
change L app
Liquid+solid
Pl
Ss
T
check
Gas+solid
Pg
Ss
T
S app
Ss
T
S dis Ss ≤ 0
Liquid+gas+solid Pg
Sl
Sl
G app G dis
Sl ≥ 1 – Ss
L dis Sl ≤ 0
2.5 Structure and Schematic Flow of Full-scale Simulation Based on the modeling and calculation strategies presented above, the simulator was programed by C++ language with three groups of input data and two groups of output data (Figure 2.2). As a conclusion, main processes running recurrently inside the simulator are also shown in Figure 2.2.
Fig. 2.2 Scheme and flow chart of the simulator 3. VERIFICATION OF SIMULATOR In this Chapter, the simulator to be used for the assessment on the performance of CO2 storage and CH4 production in this study is verified in several aspects relevant to the fluid dynamic and thermodynamics with 7
regard to the multi-phase multi-component system including mainly CO2 and CH4. In Section 3.1, test problems concerning the mixing and mutual interaction between CO2 and CH4 are simulated. The specification of test problems follows the instruction in the simulator verification guideline reported by Pruess et al. [6]. The verification of the applicability on the primary CBM production and the corresponding ECBMR simulation are conducted in Section 3.2 using the test problems specified in the paper released by Law et al. [7]. 3.1 Simulation of Binary Gas Mixture (CO2+CH4) Pruess et al. [6] reported a guideline for the verification of simulators which are developed specially for the simulation of CO2 storage in subsurface. There are several test problems particularly designed to investigate the performance and correctness of the simulator in every aspect, such as the calculation of thermodynamic properties and the behavior of fluid flow. Each test problem represents a practical case that should be addressed in daily simulations. Four numerical simulation codes were used for the comparison. These are as follows: CHEMTOUGH, developed by Industrial Research Limited, New Zealand, a geochemical modeling extension of TOUGH2 [8], GEM, a reservoir simulator developed by Computer Modelling Group (CMG), Canada; SIMUSCOPP, a reservoir simulator developed by Institut Français du Pétrole (IFP), France; and TOUGH2/EOS7C, a special gas module for the TOUGH2 reservoir simulator developed by Lawrence Berkeley National Laboratory (LBNL), USA. As there are no space to show all comparison study results, only the result on Test Problem 2 is described here.
Figure 3.1 Schematic view of test problem 2 The schematic view of Test Problem 2 is shown in Figure 3.1. Test Problem 2 is an extension of Test Problem 1 which enlarged the scale to two-dimensional reservoir in y- and z-direction and the initial position of the reserve of CH4 and CO2 changes to be side-by-side which ensure them to mix far more dramatically than in the Test Problem 1, because the pressure-driven and gravity-driven advection play dominant role here. The strong lateral density gradient between the dense CO2 gas and the relatively light CH4 gas causes a strong density-driven flow where CO2 tends to move downward and CH4 tends to move upward to the top of the reservoir. The results of the comparison are shown in Figure 3.2 and 3.3. Figure 3.2 depicts the horizontal profile of CO2 mole fraction at the elevation of 50m at two points of time, while Figure 3.3 shows the density 8
field at t = 1 year over the entire domain
Figure 3.2 Results of Test Problem 2: CO2 mass Fraction at elevation Z=50m
Figure 3.3 Density field at t=1 year Left: GEM, Right: This Simulator 3.2 Simulation of ECBMR The fluid dynamics and storage mechanism in coal seams are quite different from other strata. Since the simulations regarding coal-bearing formations have not been studied so far, there is no reference for the verification use here. Therefore, in this study the verification of the simulations for 2 types of reservoirs, i.e. coal seams and the non-coal formations are conducted separately. First, in this section, the application on CBM is verified. As what Pruess et al. summarized in [6] with regard to the CO2 disposal into deep saline aquifer formations, Law et al. [7] set up several representative test problem to compare the results of simulations with respect to CBM recovery among numbers of advanced commercial computer codes including GEM, ECLIPES [9], COMET2 [10], SIMED II [11] and GCOMP [12]. The test problems for both primary production and ECBMR process are assumed in the 2D context with the
9
total area of 10,118 m2 (100.588m 100.588m) and are identical in all of the initial conditions. The only difference is that, in the primary production case, the injection well is completely closed, while in the ECBMR case, the injection of CO2 starts from the beginning of the simulation and the rate is equal to 28,316.82 sm3/day. Considering the symmetry due to five-spot well pattern, only one-quarter of the total area need to be taken into account which has been partitioned into 11 11 grid system. As for the fluid dynamics involved, as the mechanism of sorption and advection play essential role and the diffusion is negligible in either gas phase or liquid phase in such case, only sorption and advection are fully examined in this verification. In order to demonstrate results of the simulations for primary production and ECBMR, the rates of CBM production are shown in Figure 3.4 with the plots of results calculated by other simulators mentioned above.
Figure 3.4 CBM production rate for primary production and ECBMR In Figure 3.4, it can be clearly viewed that in the case of primary production, the monotonic and asymptotic decay with time accords with the shapes of curves from other simulators and in the case ECBMR, the saddle-like shape agrees with other curves as well. This verifies the validity of the application on CBM-related simulations, though the absolute values exhibit a slight discrepancy probably resulting from the different EOS used for the calculation of thermodynamic properties of both CH4 and CO2. 4. APPLICATION ON DCBF BY USING RESERVOIR MODEL In this Chapter, using the simulation model developed, simulations on three DCBF reservoir model are conducted to make comprehensive assessment on the performance of CO2 storage in DCBF formations and the production of CH4 by ECBMR from the coal seams by implementing various profiles of injection and production as well as the sensitivity analysis on the key parameters, e.g. anisotropic permeability of coal, shale/mudstone and sandstone. With regard to the evaluation index, for the aspect of methane production, the methane recovery (RCH4) ratio representing the produced methane amount as a proportion of its OGIP (Original Gas In Place) is introduced. However considering the higher greenhouse potential of CH4, it is undesired to elevate a significant amount of 10
free methane gas in the reservoir to increase the risk of leakage of methane from subsurface. Therefore an additional index named methane recovery safety (SCH4) expressed by the total amount of CH4 which remains in gas phase in the formation at the end of the simulation divided by the produced CH4 amount. For the aspect of secure CO2 sequestration, the free CO2 ratio (FCO2) denoting the volumetric percentage of free CO2 to its total injection amount is used. In addition, in order to evaluate the coal’s adsorption effect on CO2 storage in DCBF, the proportion of adsorbed CO2 to the total injected CO2 marked as ACO2 is also taken into account. Eventually for the overall performance of the whole project, the index of overall performance (OP) is introduced to evaluate the ratio of the total free gas amount left in the formation over the sum of produced CH4 amount and injected CO2 amount, as,
4.1 Reservoir Model and Injection Specification As shown in Figure 4.1 (side view), the reservoir model set up for simulation study consists of two types of reservoir materials, i.e. coal and sandstone. There are two layers of coal seams, since with the injection of CO2, CO2 will flow upward due to buoyancy effect and the absence of low-permeability cap-rocks. The coal seam on the top not only prevents the CO2 from further migration upward, but also can be a target layer for ECBMR. And unlike the way in normal ECBMR that CO2 flows from injection well towards production well, in this model, CO2 will flow towards the coal seam from bottom not by a stream-like flow, but may act like an up-side-down waterfall covering almost all the area at the boundary of sandstone and coal seam. Therefore what will be result of such a special flow? As well as the production characteristics from the lower coal seam is the main target of the work on this model. The overburden and underlying strata are not depicted in the Figure. However they are set to be impermeable. The injection of CO2 and production of CH4 is assumed to use vertical wellbores by the conventional 5-spot pattern (Figure 4.2). As a result, only one quarter of the total area is used for simulation and the simulation should proceed in 3-dimetional context with assumption of the zero-flux boundary condition for all boundaries in Figure 4.1 due to the spatial symmetry. The initial hydrological conditions are assumed to be 100% formation water-saturated in pores of sandstone and shale as well as the cleat system in coal, and 100% methane-saturated in coal matrix. Some other key parameters regarding the geological properties and fluid dynamics are given in Appendix, which are all around the typical values of natural DCBF.
Figure 4.1 Structure of reservoir model (unit: m)
11
Figure 4.2 5-spot pattern well profile As for the methodology, since we are not familiar with such special case, first we will try to inject the CO2 in different points. Considering the structure of model and the possible flow of CO2, three points have been chosen for the injection: 1. Injection into the lower layer of the coal seam; 2. Injection into the lower layer of the sandstone formation; 3. Injection into the upper layer of the sandstone formation. 4.2 Result and Discussion The results of implementing 3 different kinds of injection profiles are compared. Results are shown in Figure 4.3 to Figure 4.6. Unfortunately none of them could achieve a significant enhanced CH4 recovery because the concentration of CO2 in the gas phase near the production well elevated so rapidly that it reached the breakthrough before the competitive adsorption fully accomplished. Therefore, from Figure 4.3, the production of CH4 always stopped during the rising slope of the production rate curve. In addition, the enhanced CH4 flow, before the well is closed, can reach the rate as high as 8,000 sm3/day and is still rising. This implies the special flow of CO2 which mentioned above can really enhance the flow rate of CH4 more intensively than the usual CO2 injection. However, the breakthrough of CO2 makes it a waste. Another characteristic of this type of model is that after the injection stops, the strong pressure gradient and buoyancy effect on CO2 makes it enter the upper coal seam (in the manner just mentioned) from the sandstone layer and adsorbs into the coal matrix. This may explain why the adsorption amount of CO2 still increases noticeably. In order to provide the quantitative concept of how poor these projects could be for CH4 production, the evaluation index with the definition used in the last section is introduced here for comparison, including the value of primary production. According to Table 4.1, all the evaluation index with respect to the CH4 production are worse than the primary production case. Considering the large adsorption potential, it is suggested that in such case, it is better to aim at the CO2 storage rather than the recovery of CH4.
12
Figure 4.3 CH4 production rate for different injection profiles
Figure 4.4 CH4 Cumulative production for different injection profiles
Figure 4.5 Percenatge of CO2 in different phases for different injection profiles
13
Figure 4.6 Partition of CH4 in different phases at the end of simulation for different injection profiles
Figure 4.7 Percentage of CO2 in different phases at the end of simulation for different injection profiles Table 5.1 Evaluation index for different CO2 injection profile CH4
Primary
Lower
Coal
Upper
Inj
Seam
Inj
CO2
Lower
Coal
Upper
Inj
Seam
Inj
RCH4:
0.3098
0.1526
0.0986
0.0799 ACO2:
0.4128
0.3642
0.3063
SCH4:
0.2527
1.2059
1.3508
1.3736 FCO2:
0.3265
0.3781
0.4196
OP:
0.0770
0.2802
0.2870
0.3000
4.3 Sensitivity Analysis Since it is not recommended to produce the methane from this reservoir model, we will mainly focus on the CO2 storage potential in it, regardless of the evaluation index for CH4 production and the overall performance. Sensitivity analysis that we will carry out later will only apply on the injection of CO2 into coal seam.
Sensitivity Analysis on the Permeability of Coal The permeability of coal in the basic case is set to be 5 mD for horizontal permeability and 1 mD for vertical 14
one. Another 2 sets of permeability were tried for sensitivity analysis. They are 10/5 mD and 1/0.5 mD. The results are shown in Figure 4.8 to 4.11 as well as the Table 4.2. It can be clarified that the higher the permeability of coal is, the more CO2 will adsorb into coal matrix, because the higher permeability in the upper coal seam would allow CO2 to migrate into it more easily and adsorb there securely. And the higher permeability in the lower coal seam would speed up the CO2 transport in the lower coal seam, which results in the lower bottom-hole pressure and more CO2 could be injected (Figure 4.8). However the higher permeability in the lower coal seam failed to inhibit the strong buoyancy effect of CO2 leading to the free escape of CO2 from the lower coal seam into sandstone formation. That’s why the index of FCO2 for 10 mD is larger than that for 5 mD.
Figure 4.8 Cumulative CO2 injection for different permeability of coal
Figure 4.9 Percentage of CO2 in difference phases for different permeability of coal
15
Figure 4.10 Percentage of CO2 in difference phases at the end of simulation for different permeability of coal
Figure 4.11 Percentage of CO2 in difference phases at the end of simulation for different permeability of coal Table 4.2 Evaluation index in the aspect of CO2 storage only for different permeability of coal CO2
10 mD
5 mD
1 mD
ACO2
0.393954 0.364244 0.313013
FCO2
0.388846 0.378077 0.439472
Sensitivity Analysis on the Permeability of Sandstone Two sets of permeability value were used for the sensitivity analysis on the impact from the permeability of sandstones. They are 10/5 mD and and 100/50 mD. The results are shown in Figure 4.12 to Figure 4.15 and Table 4.3. The results show a preference to the low-permeability of sandstone for the CO2 storage. From the percentage graph of CO2 (Figure 4.13), CO2 in gas phase and liquid phase are nearly the same for these three cases. The only but large difference is due to the amount of adsorption which is most likely resulted from the different cumulative CO2 injection amount in these three cases. The lower permeability for the sandstone layers lying between two coal seams can help CO2 migrate into the upper coal seam, because the difference 16
between the lateral permeability in sandstone and vertical permeability at the coal-sandstone interface is not that large, otherwise, in the case of high permeability of sandstone, CO2 would be guided mostly by the relatively large horizontal permeability of sandstone, which would make it flow along the interface of coal and sandstone towards the production well where the pressure gradient is stronger and induce the breakthrough quickly. Since in the case of 10 mD, CO2 broke through much later than the case of 50 mD and 100 mD, the production well kept open resulting in the larger amount of injection of CO2.
Figure 4.12 Cumulative CO2 injection for different permeability of sandstone
Figure 4.13 Partition of CO2 in different phases for different permeability of sandstone
17
Figure 4.14 Partition of CO2 in different phases at the end of simulation for different permeability of sandstone
Figure 4.15 Percentage of CO2 in different phases at the end of simulation for different permeability of sandstone Table 4.3 Evaluation index in the aspect of CO2 storage only for different permeability of sandstone CO2
10 mD
50 mD
100 mD
ACO2
0.484552
0.364244
0.325986
FCO2
0.302644
0.378077
0.402544
Sensitivity Analysis on Salinity Pure water set in the basic case was changed to 5 wt% and 15 wt% to examine the effect of salinity on the storage of CO2. Results are shown in Figure 4.16 to 4.19 and the evaluation index for CO2 storage is shown in Table 4.4. This is another variable with contradiction. According to Table 4.4 high salinity case has high adsorption ability while the consequent low solubility allows more gases to be in the gas phase.
18
Figure 4.16 Cumulative CO2 injection for different salinity
Figure 4.17 Percentage of CO2 in different phases for different salinity
Figure 4.18 Percentage of CO2 in different phases at the end of simulation for different salinity
19
Figure 4.19 Percentage of CO2 in different phases at the end of simulation for different salinity Table 4.4 Evaluation index for different salinity Pure CO2
Water
5 wt%
15 wt%
ACO2
0.364244 0.382754 0.434869
FCO2
0.378077 0.385279 0.407716
5. CONCLUSION A non-isothermal multi-phase multi-component fluid dynamic simulator was developed to make a comprehensive study on the performance of CO2 storage into DCBF. Then the simulator was verified by comparing the simulation result to the results from other simulators. The results calculated by this newly-built simulator matched well with the results of others. A reservoir model consisted of coal seams and sandstone layers was used for simulation study and by modifying the value of some key parameters, such as permeability, salinity, we carried out a sensitivity analysis on the performance of CO2 storage by defining several evaluation indexes. It has been identified under the reservoir model this time that CO2 is too easy to breakthrough and the ECBMR is impossible, unless the permeability of sandstone is in the same order with that of coal. The results of the sensitivity analysis show the strong dependence of CO2 storage performance on the permeability of sandstones as well. The authors believe that the DCBF CO2 storage is feasible when we find the adequate geological sites with cap rock (low permeable coal seam and/or mudstone) and sandstone with rather high permeability.
20
References [1]
IPCC, 2007: Fourth Assessment Report (AR4). Geneva, Switzerland. pp 104.
[2]
IEA, 2008: CO2 Capture and Storage – A key carbon abatement option
[3]
IPCC, 2005: Special Report on Carbon Dioxide Capture and Storage. Cambridge University Press, UK. pp 431.
[4]
Bradshaw, J.B. and Dance T., 2005: Mapping geological storage prospectivity of CO2 for the world sedimentary basins and regional source to sink matching. Proceedings of the 7th International Conference on Greenhouse Gas Control Technologies (GHGT-7), September 5–9, 2004, Vancouver, Canada, v.I, 583-592.
[5]
Durucan, S., Shi, J.Q.: Improving the CO2 well injectivity and enhanced coalbed methane production performance in coal seams. Int. J. Coal Geol. 77(1–2), 214–221 (2009)
[6]
Pruess, Karsten, Garcia, Julio, Kovscek, Tony, Oldenburg, Curt, Rutqvist, Jonny, Steefel, Carl, et al.(2002). Intercomparison of numerical simulation codes for geologic disposal of CO2. Lawrence Berkeley National Laboratory: Lawrence Berkeley National Laboratory
[7]
Law, D. H., L.H.G. van der Meer, and W.D. Gunter. 2002. Comparison of numerical simulators for greenhousegas storage in coalbeds, part II: Flue gas injection. In: Gale, J. and Y. Kaya (eds.): Proceedings of the 6thInternational Conference on Greenhouse Gas Control Technologies, Pergamon, Amsterdam, Vol. 1: 563-568.
[8]
Pruess, K., C. Oldenburg and G. Moridis. TOUGH2 User’s Guide, Version 2.0, Lawrence Berkeley National Laboratory Report LBNL-43134, Berkeley, CA, November 1999.
[9]
ECLIPSE, Schlumberger GeoQuest, Abingdon, Oxon, United Kingdom.
[10]
COMET 2, Advanced Resources International (ARI), Arlington, Virginia, U.S.A.
[11]
SIMED II, Commonwealth Scientific and Industrial Research Organization (CSIRO), Kinnoull Grove, Syndal, Victoria, Australia and the Netherlands Institute of Applied Geoscience TNO, Utrecht, The Netherlands
[12]
GCOMP, BP, Houston, Texas, U.S.A.
Appendix Problem Specification of Reservoir Model Parameters and Properties of Reservoir Depth (Top)
1,000
Net Pay
m -
Initial Temperature (Top)
45
Temperature Gradient
0.35
Initial Pressure (Top)
120
bar
Initial Water Saturation
1.0
-
Initial CH4 Concentration in Coal Matrix
1.0
-
/100m
21
Initial Abs. Perm. Coal (H/V)
5.0/1.0
mD
Initial Abs. Perm. Sandstone (H/V)
50.0/25.0
mD
Initial Abs. Perm. Shale (H/V)
1.0/0.1
mD
Initial Porosity Coal
0.03
-
Initial Porosity Sandstone
0.15
-
Initial Porosity Shale
0.10
-
Pore Compressibility Non-coal
4.5×10-10
1/Pa
In-situ Density Coal
1,310
m3/kg
In-situ Density Non-coal
2,400
m3/kg
In-situ Moisture Content Coal
0.156
-
In-situ Ash Content Coal
0.0672
-
Specific Heat Capacity Coal
1,300
J/(kg K)
Specific Heat Capacity Non-coal
900
J/(kg K)
Therm. Cond. Void Pore Coal
0.335
J/(m s K)
Therm. Cond. Water Saturated Coal
1.06
J/(m s K)
Therm. Cond. Solid Saturated Coal
1.41
J/(m s K)
Therm. Cond. Void Pore Non-coal
0.20
J/(m s K)
Therm. Cond. Water Saturated Non-coal
1.60
J/(m s K)
Therm. Cond. Solid Saturated Non-coal
14.1
J/(m s K)
Liquid Residual Sat. Coal
0.03
-
Gas Residual Sat. Coal
0.005
-
Liquid Residual Sat. Non-coal
0.3
-
Gas Residual Sat. Non-coal
0.05
-
Salinity
0.0
wt%
Abs. Perm. Model Coal
P&M
-
Abs. Perm. Model Non-coal
Empirical
-
Rel. Perm. Model Coal
Gash
-
Gas Rel. Perm. Model Non-coal
B&C
λ=2
Liquid Rel. Perm. Model Non-coal
Van Genuchten
m=0.457
Capillary Pressure Model Coal
Gash
-
Capillary Pressure Model Non-coal
Van Genuchten
m=0.457; 1/α=19.61kPa
Langmuir Pressure CH4
18.160966
bar
Langmuir Volume CH4
0.01803
sm3/kg
22
Langmuir Pressure CO2
14.923657
bar
Langmuir Volume CO2
0.04449
sm3/kg
Diffusion Time Constant
38.6
day
Boundary Condition
Zero Flux
-
Inj/Pro Parameters Well Radius
0.0365
m
Well Cell Height
Net Pay of Cell
m
Skin Factor
0
-
Injection Rate
20,000
sm3/day
Max. Production Rate
100,000
sm3/day
Max. Inj. BHP
200
bar
Min. Pro. BHP
2.75
bar
Production Timing
0 - CO2 Breakthrough*
Max. 50 years
Production Point
Coal Seam(s)
-
Injection Timing
1 - 50
years
Injection Point
As written in the text
-
* when the mole fraction of CO2 in gas mixture produced exceeds 10%
Grid Setup Well Pattern
5-spot Vertical
-
Drainage Area
Figure 4.2
-
Calculation Area
Figure 4.2
-
Grid Segmentation
2×5×5
-
23
2010 International Pittsburgh Coal Conference, Istanbul, Turkey October 11 – 14, 2010
Experimental Study on Carbon Dioxide Sequestration by Mineral Carbonation Junying Zhang, Heng Yan, Yongchun Zhao, Chuguang Zheng State Key Laboratory of Coal Combustion Huazhong University of Science & Technology, 430074 China
Abstract Carbon dioxide sequestration by mineral carbonation is a potentially attractive route to mitigate possible global warming on the basis of industrial imitation of natural weathering processes. In the paper, two typical minerals serpentine and wollastonite were selected as feedstock for direct mineral carbonation experiments under lowmiddle pressure. Abundant experimental studies were performed in the paper to investigate the factors (temperature, pressure, particle size, pre-treatment and gas composition) that influence the conversion rate of carbonation reaction. The products from experiments were analyzed by x-ray diffraction (XRD), field scanning electron microscopy equipped with energy dispersive X-ray analysis (FSEM-EDX). The results show that, after antigorite mineral carbonation, the mainly mineral components include: quartz, and magnesite, few antigorite. Rhombohedra magnesite crystals and rounded particles of serpentine were also identified from SEM image. For the wollastonite carbonation, the mainly mineral in production is calcite, which was also identified by SEM analysis. Few amorphous SiO2 which can’t be identified by XRD has been found around the calcite particles in the SEM image. All of these validate that mineral carbonation is a potential technique for carbon dioxide sequestration. The method of mass equilibrium after heat decomposition was used to calculate the mineral carbonation conversion rate. The conversion rate increases with the increasing of temperature. The conversation rate of wollastonite carbonation is higher than antigorite carbonation. Pressure is also an important factor to mineral carbonation. For both of wollastonite and antigorite, the carbonation conversion rate is increase with increasing reaction pressure. Because carbon dioxide is dissolve in water solution easier in high pressure. The conversion rate of mineral carbonation after pretreatment, such as heat treatment and pulverization, is much higher than the normal mineral. Under the condition of simulated flue gas, the conversion rate of wollastonite mineral carbonation decreased sharply compared with the pure carbon dioxide. And with the addition of SO2 in the simulated flue gas, the conversion rate of wollastonite mineral carbonation decreased. A carbonation
Corresponding author. Fax: +86-27-87545526 E-mail address: [email protected]
experiment at PCO2 =4.0 MPa, T=150°C, dp< 30m results in a highest carbonation conversion rate of 89.5% for wollastonite. Compared with antigorite, wollastonite is more promising in carbon dioxide mineral sequestration.
Key Words: Carbon dioxide, sequestration, mineral carbonation, serpentine, wollastonite.
1 Introduction In the past 50 years, the increased greenhouse gases (GHGs) level in atmosphere is believed to cause global warming. Among these GHGs, CO2 is the largest contributor in regard of its amount present in the atmosphere contributing to 60 percent of global warming effect [1]. The annual report of the Word Meteorological Organization (WMO) indicated that the CO2 concentration in the atmosphere is up to 380ppm [2]. Coal-fired power plant is the largest source of CO2 emission, the power plant release over 40% of total China carbon dioxide emissions. To reduce total CO2 emission, an effective option is enhance the sequestration of CO2. Fixation of CO2 from fossil fuel combustion in the form of solid carbonates appears to be a potentially attractive and realistic option for the sequestration of CO2. Vast amounts of magnesium or calcium silicate minerals exist worldwide that may be carbonated, with magnesium and calcium carbonate as stable and environmentally harmless product [3-9]. Mineral carbonation, also referred to as mineral sequestration, is using natural silicate mineral reacting with CO2 to form geologically stable carbonate minerals. This method emulates the nature chemical transformation of CO2, for example, weathering of ultramafic rocks to form calcium or magnesium carbonates. The reactions of mineral carbonation are illustrated as below [7]: Mg2SiO4 + 2CO2→ 2MgCO3 + SiO2 Mg3Si2O5(OH)4 + 3CO2→3MgCO3 + 2SiO2 + 2H2O CaSiO3 + CO2→ CaCO3 + SiO2 However, the reaction rate of mineral carbonation under natural conditions is too slow to be used in the largescale commercial CO2 sequestration programs. So the reaction condition must improve to increase the reaction rate of carbonate reaction. Mineral sequestration which use silicate minerals as the raw materials to sequestrate CO2, was first mentioned by Seifritz (1990) [8] and discussed further by Dumsmore (1992) [9]. After that, many researches were done to provide the details and foundation for today’s research efforts. Daniel et al. got a
carbonation conversion of 80%, by using heat pre-treated serpentine feed material and a high partial pressure of CO2 (PCO2=115atm) under the condition of 185℃ and in buffer solution (0.5M Na2CO3; 1.0M NaCl or 0.5M NaHCO3) [10]. O’Connor et al. achieved a 78% conversion of preheated silicate mineral carbonation under the temperature 155℃ and higher pressure (PCO2=185atm) [4]. In summary, most of their research works were under high partial pressure of CO2, which would cost many problems for industrial applications of mineral carbonation as a carbon dioxide sequestration method. In the paper, two typical minerals serpentine and wollastonite were selected as feedstock for direct mineral carbonation experiments under middle and low-pressure. The physical-chemical characteristics of CO2 mineral feedstock and carbonation products were charactered, and the effects of experimental conditions were studied in details.
2 Experimental facility
1-CO2, 2-temperature controller, 3-pressure meter, 4, 5-gas valves, 6-high pressure reactor, 7-thermocouple Fig.1. Schematic diagram of CO2 mineral carbonation
The CO2 mineral carbonation experiments were carried out in a high pressure and high temperature reactor. The schematic diagram of experimental facility is shown in Fig.1. The experimental system is made of a high pressure reactor which has a maximum range of 20Mpa and an electrical heated system which can get a highest temperature of 300℃. The raw minerals were pulverized and sieved to get the desired particle size. After that, the mineral particles were heated under 650℃ for 2 hours before carbonation reaction. Every test, 10g mineral particles mixed with 40 ml water were used to capture pure CO2 or simulated flue gas containing 15% CO2 and 0.1% SO2. The carbonation reaction was conducted under the temperature range of 80℃ to 200℃, the pressure was set between 2MPa and 10MPa, and the reaction was last 1 hour.
3 Characteristics of mineral feedstock Two typical minerals serpentine and wollastonite were selected to react with carbon dioxide to form geologically stable carbonate minerals. The physical-chemical characteristics of CO2 mineral feedstock were charactered using x-ray diffraction (XRD) and X-ray fluorescence (XRF).
Fig.2. XRD spectra of serpentine feedstock antigorite(An), Magnesium olivine (Fo) and calcium fluoride (St)
Fig.3. XRD spectra of wollastonite feedstock
4 Characteristics of mineral carbonation products After carbonation reaction lasted for the desired time, the reactor was cooled down and the products were collected and filtered. Then the collected products were dried for further analysis. The products from experiments were analyzed by x-ray diffraction (XRD), field scanning electron microscopy equipped with energy dispersive X-ray analysis (FSEM-EDX).
Fig.4. XRD spectra of serpentine carbonation product Magnesite (Ms), antigorite(An), magnesium carbonate hydrate(Mt)
The results show that, after serpentine mineral carbonation, the mainly mineral components include: quartz, and magnesite, few antigorite and sodium chloride (Fig.4). Rhombohedra magnesite crystals (red ring) and rounded particles of serpentine (blue ring) were also identified from SEM image (Fig.5).
Fig.5.SEM image of serpentine carbonation product
For the wollastonite carbonation, the mainly mineral in production is calcite (Fig.6), which was also identified by SEM analysis (Fig.7). Few amorphous SiO2 which can’t be identified by XRD has been found around the calcite in the SEM image. All of these validate that mineral carbonation is a potential technique for carbon dioxide sequestration.
Fig.6. XRD spectra of wollastonite carbonation product
Fig.7.SEM image of wollastonite carbonation product
5 Effects of reaction conditions on CO2 mineral carbonation Abundant experimental studies were performed to investigate the factors that influence the conversion rate of carbonation reaction, such as reaction temperature, reaction pressure, particle size and pre-treatment. The method of mass equilibrium after heat decomposition was used to calculate the mineral carbonation conversion rate. 5.1 Effect of temperature
wollastonite
antigorite
Carbonation conversion rate/%
70 60 50 40 30 20 10 0 30
60
90
120
150
180
210
Temperature/℃ Fig.8 The effect of temperature on conversion of mineral carbonation
The results show that the conversion rate increases with the increasing of temperature (Fig.8). For antigorite, the carbonation conversion rate quickly rise to 30% when the temperature reaches 150℃, after that it decreases a little with temperature increasing which maybe related to the decrease of carbon dioxide solubility in high temperature. To get a high carbonation conversion, it’s necessary to choose a suitable temperature to promote the mineral carbonation. Compared the conversion of these two minerals, the conversation rate of wollastonite is higher than that of antigorite, and its temperature range is broader.
Carbontion Conversion/%
5.2 Effect of pressure
70
serpentine wollastonite
60 50 40 30 20 10 0
0
1
2
3 4 5 pressure/MPa
6
7
8
Fig.9 The effect of pressure on conversion of mineral carbonation
Pressure is also an important factor on mineral carbonation. For both of wollastonite and serpentine, the carbonation conversion rates increase with increasing of reaction pressure (Fig.9). Carbon dioxide dissolved in solution easy in high pressure. However, when the pressure increased above 6MPa, the carbonation conversion rate increased weakly with the increase of pressure, this indicates that there is a suitable pressure range to achieve high conversion rate. 5.3 Effect of mineral particle size 90 80
antigorite wollastonite
Conversion rate/%
70 60 50 40 30 20 10 0 74μm
37μm
Fig.10 The effect of particle size on conversion of mineral carbonation
Mineral carbonation conversion rate increases with the decrease of the particle size (Fig.10). this mainly due to two reasons: the first is the surface area of the mineral particles increases with particle size decreases; and at the same time, some particle crystal structure is destroyed during the crushing and screening process, this improves the activity of the mineral particles in the reaction. From these tests, it can be found that a carbonation experiment at PCO2 =4.0 MPa, T=150°C, dp<30μm results in a highest carbonation conversion rate of 89.5% for wollastonite. Compared with antigorite, wollastonite is more promising in carbon dioxide mineral sequestration. 5.4 Effect of mineral preheat treatment
45 40
Conversion rate /%
35 30 25 20 15 10 5 0 Preheated
Normal
Fig.11 The effect of preheat treatment on conversion of mineral carbonation
The impact of heat treatment on the conversion rate is shown in Fig.11. The conversion rate of mineral carbonation after heat treatment is much higher than the normal mineral. Heat treatment can reduce the water content in the mineral particles, so the quality of the relative percentage of magnesia increases. Thus speeds up the mineral particles dissolving in the solution. 5.5 Effect of the solution composition 45
Carbonation Conversion/%
40 35 30 25 20 15 10 5 0
None
0.6M NaHCO3
Fig.12 The effect of the solution composition on conversion of mineral carbonation
Figure 12 compares the results for two different types of liquid that made up the reaction slurry: one was distilled water and the other was 0.6M NaHCO3 solution. The operating conditions were the same. In O’Connor et al[4]carbonation study, they postulated that the addition of NaHCO3 (0.45–0.64 M) had a buffering effect that kept the solution pH with in a certain range, which favored carbonation and caused the NaHCO3 to act as a “catalyst.” Although the pH value of the solution was not monitored in this study during experimental runs because of the lack of a high-pressure pH sensor, we believe that the buffering effect existed in our system.
5.6 Effect of SO2 on conversion of mineral carbonation
35 simulated flue gas(without SO2) simulated flue gas
Carbonation Conversion/%
30 25 20 15 10 5 0
60
80
100
120
140
160
180
200
220
temperature/℃ Fig.13 The effect of SO2 on conversion of mineral carbonation
As shown in Fig.13, with the addition of SO2 in the simulated flue gas, the conversion rate of wollastonite mineral carbonation decreased sharply. This is because of the competition reaction of SO2 with wollastonite, which depleted Ca2+ in the solution. However, the conversion rate in 80℃ with SO2 is higher than that without SO2, which maybe because there is no competition reaction occurs in the low temperature and SO2 dissolved in the solution increases the acidity which is promoting the carbonation reaction.
6 Conclusions Mineral carbonation has been identified as a promising CO2 sequestration technology option. The conversion rate of CO2 mineral carbonation increases with the increasing of the temperature and the pressure. With the decrease of the particle size, the conversion rate of CO2 mineral carbonation increases, too. Mineral preheat treatment and change of the solution composition are helpful to increase the conversion rate of CO2 mineral carbonation. The SO2 in simulated flue gas promotes the CO2 mineral carbonation reaction at low temperature while inhibits it after temperature up to 80℃. A carbonation experiment at PCO2=4.0 MPa, T=150℃, and dp<30μm results in the highest carbonation conversion rate of 89.5% for wollastonite. Compared with antigorite, wollastonite is more promising in carbon dioxide mineral sequestration.
7 Acknowledgments The authors express their great thanks to the National Natural Science Foundation of China (Grant No. 40972102, 50906031), and National Basic Research Program of China (“973” Program) (Grant No. 2010CB227003).
References [1]. Hongqun Yang, Zhenghe Xu, Maohong Fan, Rajender Gupta, Rachid B Slimane, Alan E Bland, Ian Wright. Progress in carbon dioxide separation and capture: A review. Journal of Environmental Sciences, 20, 2008, 14-27
[2]. http://www.wmo.int/pages/index_en.html
[3]. Pan, X.; Zhang, J.; Xu, J.; Ding, F.; Wang, Z.; Zheng, C., Research status of carbon dioxide sequestration as mineral carbonation. Coal Conversion, 29(4), 2006, 78-82.
[4]. O'Connor, W. K.; Dahlin, D. C.; Rush, G. E.; Dahlin, C. L.; Collins, W. K., Carbon dioxide sequestration by direct mineral carbonation: Process mineralogy of feed and products. Minerals and Metallurgical Processing, 19(2), 2002, 95-101.
[5]. Xu, J.; Zhang, J.; Pan, X.; Zheng, C., Carbon dioxide sequestration as mineral carbonates. Huagong Xuebao/Journal of Chemical Industry and Engineering (China), 57(10), 2006, 2455-2458.
[6]. Chen, Z.-Y.; O'Connor, W. K.; Gerdemann, S. J., Chemistry of aqueous mineral carbonation for carbon sequestration and explanation of experimental results. Environmental Progress, 25(2), 2006, 161-166.
[7]. O'Connor, W. K.; Dahlin, D. C.; Turner, P .C.; Walters, R . P. Carbon dioxide Sequestration by ex-situ mineral carbonation. DOE/ARC-99-009, 7, 1999, 115.
[8]. Seifritz W. CO2 disposal by means of silicates. Nature 345, 1990, 486
[9]. Dunsmore H.E., “A Geological Perspective on Global Warming and the Possibility of Carbon Dioxide Removal as Calcium Carbonate Mineral,” Energy Convers. Mgmgt, 33(5-8), 1992, 565-572.
[10].
Daniel J F, John P B, Soong Y et al . Carbon Storage and Sequestration as Mineral Carbonates. New York : Kluwer
Academic/ plenum Publishers, 2002 :101-118
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: CARBON MANAGMENT CO2 SEQUESTRATION FOR THE SHENHUA DCL PLANT IN CHINA Qingyun Sun, Associate Director, US-China Energy Center, West Virginia University PO Box 6108, Morgantown, WV 26506-6108 USA Phone (304)293-5545, Fax (304)293-3752, Email: [email protected] Jerald J. Fletcher, Professor and Director, US-China Energy Center, West Virginia University PO Box 6108, Morgantown, WV 26506-6108 USA Phone (304)293-5499, Fax (304)293-3752, Email: [email protected]
Abstract: In 2009 the world’s 1st modern commercial direct coal liquefaction (DCL) plant was completed in China by the Shenhua Group Corporation (Shenhua), the world’s largest coal company. Based on the results obtained from initial start-up operations, the plant has undergone a series of modifications prior to full scale operation that will produce about 24,000 bbl/d of liquid fuels. When in full scale operation, the plant will generate over 3 million metric tons of highly concentrated CO2 each year, a major issue given current concerns related to the carbon footprint of the energy sector. To address these concerns, Presidents Obama (US) and Hu (China) announced a joint effort to develop a CO2 sequestration project related to the Shenhua DCL plant in November 2009 that is an initial step in addressing the carbon management issue for the coal conversion industry. The US Department of Energy (DOE) and the China National Energy Administration (NEA) are supporting a cooperative effort by a US team led by West Virginia University (WVU) and a China team supported by Shenhua to analyze the alternatives available and to develop an implementation plan for this project. These collaborative efforts include a feasibility study for the initial project and a broader study of the Ordos Basin to determine the potential for broader implementation and inclusion of other CO2 sources. Cooperative efforts are expected to address risk analysis, modeling, CO2 utilization, geoscience research on CO2 storage and carbon management planning for the Ordos Basin. As a result of initial efforts previously completed, Shenhua is implementing a series of engineering modifications to the DCL plant to increase the concentration of the CO2 stream from the initial 89% level based on the original plant design to 95% or greater to improve the sequestration efficiency. A pre-feasibility study was completed in 2009. The new effort will support the development of a full feasibility study to extend the initial efforts as well as a pilot demonstration project designed to store approximately 100,000 MT/y of CO2 in deep saline aquifers near the Shenhua DCL plant. This demonstration project is expected to be implemented during 2010. The data developed for the pilot project including data from test wells and the initial results of the pilot storage test are some of the first steps toward a comprehensive study of a CO2 sequestration effort for the Ordos Basin. If successful, this broader effort could evolve into the world’s largest current sequestration effort with an ultimate goal of sequestering 50-60 million mt/y of CO2 in the Ordos Basin. This paper provides an overview to the joint efforts between the WVU led US research team and the Shenhua led China research team to explore the potential for such a sequestration effort. The results of these efforts are expected to provide a better understanding of CO2 constraints and possible solutions relevant to future coal conversion developments around the world.
Clean Fuel Productıon Works From Çanakkale-Çan Coals Presenting Author (Name, Title, Affiliation): Oguz ALTUN, Mining Engineer, Contact Information: Mineral Research and Exploration Directorate in Turkey ([email protected]) Co-Authors: Ayşe ERDEM*, Akan GULMEZ*, Mining Engineers, Selami TOPRAK*, Assoc. Prof. in Geology, Zeki OLGUN**, Chemical Engineer Contact Information: *Mineral Research and Exploration Directorate in Turkey, **Turkish Coal Enterprises
ABSTRACT In this study, high sulfur containing Canakkale-Can coals were used to upgrade their calorific and sulfur values, with employing washing, multi gravity separation and flotation methods. For this purpose, samples were taken from the fields, and then representative samples were prepared to conduct physical, chemical, mineralogic and petrographic analysis, determine grain size distribution, proper ore processing method and suitable working conditions of the chosen suitable devices, for the experimental works. Sink-float tests were performed on the various sized prepared samples. The received data was evaluated, then concentrated coals with determined densities were obtained, then the briquettes were made out of these coals. The secondary washing cycle products and -0.5mm coals were taken to be tested with a Multy Gravity Seperator. Besides, flotation tests were carried out to decrease sulfur content of the coals. INTRODUCTION Can Lignite Enterprise Directorship (ÇLİ) is located in North Western region of Turkey. There is about 92.085.000 tons of proven reserve but, 73.983.000 ton operatable reserve of coal in Can region. The LCV of coal is 3013 kcal/kg. About 2.500.000 tons of coal is produced as open pit method, annually in the region. Almost the whole amount of the coal is consumed in a thermic power plant running with fluid bed technology to produce electricity. Since the coal is directly sent to the power plant, not any of enrichment process is conducted on the coals. In this study, washability and cleanability potential of the high sulfur and ash containing Canakkale –Can coals were investigated with technological methods. EXPERIMENTAL STUDIES Samples from Marionlar, Tarhan and Koseoglu regions are piled up in a collection area. Sample Collection The samples were being taken in whole shifts, periodically from the trucks coming from different production sites in week period. With this, the sample collection representing the whole area was realized. About 3 tons of taken sample, then, brought to Ore Processing laboratory unit of Mineral Analysis and Technology Department in MTA General Directorate.
The samples were, then, mixed to take the best representing samples to conduct chemical and petrographical analysis. Later, for sink-float tests, the samples were classified to -100+50 mm, -50+18 mm and -18+0.5 mm sizes with sieves. About 250 kg sample with -100+0.5 mm fractions were washed within 1.4 g/cm3 density, to produce the first quality clean fuel for studies. The rest material were washed within 2.0 g/cm3 density to produce 2nd quality product. In order to obtain higher quality in this process, the material was processed with a Multi-Gravity Separator (MGS). In order to carry out this, the sample sizes were decreased below 0.3 mm due to sulfur containing mineral sizes. For the ore process carried out by flotation methods, float material of 2.0 g/cm3 density were used, in order to cover the associated rocks mixed to coal during production, as well. In this way, reducing the consumption of the reactive as well as increasing flotation recovery of the process was targeted. Mineralogic Analysis According to performed mineralogic and petrographic analysis, liberation degree of coarser coals (up to +5mm) is considerably low and change between 30-45%. In finer sizes, liberation degree of coals was determined as 60-75%. The average grain size of pyrites is between 10-15 micron, and liberation degree, even in fine samples, is about 45%. Sulfur is absolutely encountered in the coarser clean coals. With the scope of conducted mineralogic analyses and previous studies, upgrading of the coals with physical methods was pointed out as to be hard for coarser sizes but there would be an experimental study with MGS for finer (-0.3 mm) Furthermore, it is thought that flotation method, which is a physico-chemical method, would be effective on the samples. Chemical Analysis Table 1. Sieve analysis and chemical analysis of the original sample Size, mm
Amount (%)
Ash %
S%
*LCV (kcal/kg)
-100+50 -50+18 -18+10 -10+5 -5+1 -1+0.5 -0.5 Total
26.8 40.0 13.9 8.5 6.4 1.9 2.4 100.0
27.3 30.7 31.6 34.3 40.9 44.2 48.2 31.5
7.0 6.5 6.5 6.3 5.8 5.8 6.6 6.6
4.653 4.379 4.308 4.095 3.509 3.163 2.915 4.304
*LCV (Lower Calorific Value) As can be seen on Table 1, the upper level of the original sample is below 100 mm and the material below 0.5 mm is about 2.4 % (as weight). Ash content of the fractions, as in dried basis, change between 27.3% and 48.2%. Sink-Float Tests When sink-float results are considered, 75.8 % of the material was floated within -100+50 mm fractions with 1.5 g/cm3 density. The ash content of the material was determined as 14.6 %, calorific value as 5594 kcal/kg, sulfur content as 7.1%. As determined in mineralogic analysis, since the liberation degree of pyrites are so small that sulfur content of the products at the end of sink-float process seems to show an increase. This is, of course, an expected result.
In -50+18 mm fraction and with 1.5 g/cm3 density, the floating amount was determined as 68%, ash content as 14.6%, calorific value as 5606 kcal/kg and sulfur content as 6.6%. In -18+0.5 mm fraction and with 1.6 g/cm3 density, the floating amount was determined as 60.1%, ash content as 15.5%, calorific value as 5451 kcal/kg and sulfur content as 5.8%. It was observed that as the size decreased the sulfur content decreases relatively as well. As can be seen from the sink-float test results, sulfur content of the sample has not encountered any significant decreases. This shows that pyrites cannot be excluded easily with physical methods. Clean Coal Production To the -100+0.5 mm sink-float test results, clean coals obtained in the first period (1.4 g/cm3 density) can be easily used in combustion and industry and the second period coals (2.0 g/cm3 density), in thermic industry. The coal properties of the floated products in -100+50, -50+18 and -18+0.5 mm fractions, within 1.4 g/cm3 density are exhibited as on Table 2. Table 2. The properties of coal having -100+0.5 mm size and 1.4 g/cm3 density Analysis
Air Dried Sample
Dried Sample
Moisture %
18.4
-
Ash %
11.6
14.3
Volatile Matter %
34.1
41.7
Fixed Carbon %
35.9
44.0
Combustible Sulfur %
4.7
5.7
Sulfur in ash %
0.2
0.2
Total Sulfur %
4.8
5.9
4414
5537
LCV, kcal/kg
Ore processing studies with MGS Experiments were planned with combining the clean coals obtained in -100+0.5 mm sizes, with 1.5-2.0 g/cm3 density and -0.5+0.02 mm size original coals to be able to lessen sulfur contents, by means of MGS. The whole samples were degraded under 0.3 mm size. Liberation degree of sulfur was considered to choose this size. With preliminary tests, feeding solid ratio, feeding amount, drum angle, wash period, drum amplitude and shaking frequency were fixed, the drum speed, and the amount of wash water were optimized to determine the proper working conditions. In MGS experiments, the work was organized to run ceaselessly and a peristaltic pump arrangement was established to feed the pulpy material, during the experiment (Figure 1). A mechanical mixer was used to stir up the material homogenously, in a plastic bucket holding about 15kg of the sample. In the conducted experiments with MGS, concentrate and waste material were taken in every 2 minutes after separation took place. After drying up the taken samples, they were weighed out and ground with a ring mill, then, their chemical analyses were carried out.
Figure 1. MGS Instrument’s arrangement Chemical analyses of the fraction -0.3 mm are exhibited on Table 3. Table 3. Chemical analysis results belonging to -0.3 mm size Analysis
Original Sample
Dried Sample
Moisture %
17.92
-
Ash %
39.58
48.22
Volatile Matter %
26.24
31.97
Fixed Carbon %
16.26
19.81
Combustible Sulfur %
5.39
6.57
Sulfur in ash %
2288
2915
Total Sulfur %
2515
3064
Table 4. Wash water analysis of -0.3 mm fraction, with 240 rpm CONCENTRATE (%)
TAIL (%)
WASH WATER amount liter/minute
Weight
Dry Ash
Sulfur
1 2
65.2 83.6
46.0 47.5
5.5 5.5
71.1 98.9
3114 2954
3
90.6
48.2
5.5
93.8
2837
Combustible LCV Weight Recovery % kcal/kg
Dry Ash
Sulfur
34.8 16.3
58.9 63.6
7.6 9.8
28.9 1.1
2023 1685
9.4
66.9
12.1
6.5
1495
Figure 2. Wash water concentrate diagram with 240 rpm
Combustible LCV Recovery % kcal/kg
Wash water amount, in the experiment, were chosen as 1, 2 and 3 l/min. As can be seen on Table 4 and Figure 2, as wash water amounts show increases, the ash contents show also increases, but LCV show decreases within 240 rpm drum speed. The best results were observed at 1l/min wash water, which its ash content was determined as 46.0 %, sulfur content as 5.5% and LCV as 3114 kcal/kg. The increase of wash water causes more discharge of lighter material, therefore the amount of material will show an increase besides an increase of ash content relatively, as well (Koca et al., 1999; 2005). Table 5. Wash water analysis of -0.3 mm fraction, with 250 rpm WASH WATER amount liter/minute
CONCENTRATE (%)
WASTE (%)
Weight
Dry Ash
Sulfur
Combustible Recovery %
LCV kcal/kg
Weight
Dry Ash
Sulfur
Combustible LCV Recovery % kcal/kg
1
68.1
45.3
5.4
73.6
3036
31.9
58.1
7.5
26.4
2074
2
72.9
46.1
5.5
78.4
2998
27.1
60.2
7.5
21.6
1899
3
74.7
47.1
5.5
80.5
2895
25.4
62.3
7.6
19.5
1756
Figure 3. Wash water concentrate diagram with 250 rpm The results of changing MGS drum speed are shown on Table 5 and Figure 3. In the conducted tests, at 250 rpm drum speed, wash water amount was increased and low ash content of about 45.3%, sulfur content of 5.4% and the highest LCV of 3036 kcal/kg were encountered. Table 6. Wash water analysis of -0.3 mm fraction, with 260 rpm WASH WATER amount liter/minute
CONCENTRATE (%)
TAIL (%)
Weight
Dry Ash
Sulfur
Combustible Recovery %
LCV Weight kcal/kg
Dry Ash
Sulfur
Combustible Recovery %
LCV kcal/kg
1
58.0
44.1
5.4
64.2
3221
42.0
57.0
7.2
35.8
2116
2
60.9
44.1
5.4
67.5
3
70.5
44.6
5.5
76.9
3135
39.1
58.1
7.4
35.5
2077
3152
29.6
90.3
7.8
23.1
1906
Figure 4. Wash water concentrate diagram with 260 rpm The results with changing drum speed as 260 rpm and wash water amount are exhibited on Table 6 and Figure 4. As can be seen on the tables and graphics exhibiting the drum speed and wash water determinations, while the drum speed increases within 1 l/min ash content in concentrate decreases but LCV increases in wash water. In 260 rpm drum speed, 44.1% of 58 % of the material participating to the experiment was determined as ash and LCV as 3221 kcal/kg. The combustible recovery is about 64. As the results are evaluated, 260 rotation/min drum speed and 1 l/min wash water amount were determined as the best proper values for the tests. When the experiments conducted with MGS are examined, ash and sulfur ratios in -0.3 mm grain size were not decreased to the desired level. Especially, when ash content is considered, 6.15% sulfur content at the entry of MGS, the sulfur content decreased to an interval between 5.37-5.52 % which indicate the sulfur reduction to be extremely difficult. In the performed mineralogic and petrographic analysis, proper liberation size of the pyrites seems to be between 10-15 microns and mostly occurred as inclusions within coals. For these reasons, as seen on the results, removing enough of pyrite minerals from coal seem to be not possible. As the results are examined in the scope of ash content, it does not seem there is enough reduction as well. The ash content at the entry of MGS was about 50% and at the end of test this amount dropped to only 44% levels. As indicated in sink-float tests of -0.5 mm size, material amount in the ±0.1 density interval seems to be considerably high which implies that application of any physical beneficiation method seems hard. Ore processing with flotation method It is meaningful to scrutinize carbon and ash content, amount of impurities and their liberation degrees when beneficiation of the coals with flotation method is considered. Carbon and ash content of coals are important parameters, directly related with their hydrophobicities. As carbon content increases (ash decreases), natural hydrophobicity increases and they are easily floated. On the other hand, alternative method removing pyrites and/or marcasites from coals which possess low carbon but high ash content, therefore low hydrophobicity, is to supress coals with starch, float the impurities with sulfhydryl (xanthate) like collectors (Fuerstenau, Miller and Kuhn, 1985). When Canakkale-Can region coals are examined physically and chemically, the second method of flotation method, as second alternative method, might be applied for sulfur removal. This alternative method was used to remove inorganically combined pyrites from Canakkale-Can coals and xanthate (KAX) was used as collector and
starch as suppress. The result of the conducted tests with different sizes of flotation method was exhibited on tables and they were examined. Flotation studies were performed with a Denver brand flotation device which is of a laboratory scale and owns a liter size flotation cell. Experimental studies were carried out with about 150 gram of enriched coal samples within heavy media and their sizes were reduced under 150 micron into three different groups (-150+100 micron, -100+50 micron, -50+10 micron). The other conditions of experiments: Solid Ratio pH Starch (To suppress coal) Xanthate (KAX) MIBC Condiotioning Period Flotation Period
: 20 % : 5.4; 5.6 and 6.3 (natural conditions) : 3330; 5000; 6670 g/ton : 500; 670; 830 g/ton : 0.075 ml : 7 min. : 5 min.
Figure 5. Flotation works The results of the applied flotation works, explained above, are exhibited on Table 7 to 11. Table 7. Effect of KAX amount on products with -150+100 micron size. Xanthate (KAX) (g/ton) 500 670 830
Product Concentrate1 Tail2 Concentrate 1 Tail2 Concentrate 1 Tail2 FEEDING
Starch: 6670 g/ton; pH: 5.4 (natural pH);
Amount %
S%
Ash %
90.9 9.1 85.6 14.4 90.0 10.0 100.0
4.3 9.1 4.2 7.9 4.9 10.4 5.5*
16.8 21.0 15.8 17.2 17.3 22.4 19.3*
LCV (kcal/kg) 4187 4723 4373 4076 4194 4592 4613*
Table 8. Effect of starch amount on products with -150+100 micron size. Starch (g/ton) 3330 5000 6670
Product
Amount %
S%
Ash %
88.8 11.2 88.6 11.4 85.6 14.4 100.0
4.5 9.7 4.9 9.7 4.2 7.9 5.5*
17.4 19.9 19.4 20.9 15.8 17.2 19.3*
Concentrate1 Tail2 Concentrate 1 Tail2 Concentrate 1 Tail2 FEEDING
LCV (kcal/kg) 4479 4684 4942 4613 4373 4076 4613*
Xanthate: 670 g/ton; pH: 5.4 (natural pH) Table 9. Effect of KAX amount on products with -100+50 micron size. Xanthate (KAX) (g/ton) 500 670 830
Product
Amount %
S%
Ash %
85.8 14.2 82.3 17.7 74.8 25.2 100.0
5.0 12.8 4.5 9.6 4.5 10.8 6.0*
26.0 28.2 22.6 22.1 24.2 24.5 24.6*
Concentrate1 Tail2 Concentrate 1 Tail2 Concentrate 1 Tail2 FEEDING
LCV (kcal/kg) 4024 4356 4173 4690 4020 4592 4332*
Starch: 6670 g/ton; pH: 5.6 (natural pH) Table 10. Effect of starch amount on products with -100+50 micron size. Starch (g/ton) 3330 5000 6670
Product
Amount %
S%
Ash %
85.5 14.5 86.3 13.7 82.3 17.7 100.0
4.7 10.2 4.3 9.6 4.5 9.6 6.0*
22.6 22.3 20.8 21.6 22.6 22.1 24.6*
Concentrate1 Tail2 Concentrate 1 Tail2 Concentrate 1 Tail2 FEEDING
LCV (kcal/kg) 4318 4702 4002 4776 4173 4690 4332*
Xanthate: 670 g/ton; pH: 5.6 (natural pH). Table 11. Effect of KAX amount on products with -50+10 micron size. Xanthate (KAX) (g/ton) 500 660 830
Product Concentrate1 Tail2 Concentrate 1 Tail2 Concentrate 1 Tail2 FEEDING
Amount %
S%
Ash %
64.1 35.9 51.6 48.4 49.3 50.7 100.0
5.0 6.9 5.1 6.9 4.8 7.1 5.9
35.2 35.1 32.4 32.5 33.1 35.5 34.0
LCV (kcal/kg) 3634 3616 3812 3630 3573 3591 3645
Starch: 6650 g/ton; pH: 6.3 (Natural pH) Explanations (Table 7 to12): 1: Sink; Clean Coal, 2: Float; pyrite, *: Correlation process were carried out and given, according to the chemical analysis of feeding. Evaluation of flotation experiment results When the results of flotation experiments are examined, conditioning studies started within natural pH environment were observed to change its pH value, after a while, to 5.4 - 6.3. The main reason for this may be organically combined pyrites in coal.
Generally in all fractions, as KAX increases, float pyrite containing tail increases and sulfur amount in the obtained concentrate (sinked material), remarkably decreased. Contrarily, ash ratio increases in the float tail as well. On the other hand, as amount of starch, to suppress coal increases, the amount of floating tail decreases and the ash content decreases too. As a result, starch plays an effective role to supress coals. CONCLUSION Canakkale-Can coals, as can be seen in mineralogic analysis, comprise of very fine pyrite grain texture. For this reason, the main purpose of this study was to reduce sulfur content of the material which was not succeeded by physical methods. The same situation is valid for ash content as well. In order to remove enough of inorganic pyrites by flotation method and to reduce sulfur content of the concentrate coal, pyrites should be ground to very fine grain sizes to be able to sufficiently liberate. Economy will come to the point then, here, therefore cost analysis of it should be carried out very well. REFERENCES Fuerstenau, M.,C., Miller, J., D. and Kuhn, M., C., Chemistry of Flotation, SME of American Institute of Mining Metallurgical and Petroleum Engineers Inc., New York, 1985. Koca, H., Koca, S., Sonmez, E. ve Ertok, C. 1999; De-Sliming and Cleaning of KütahyaSeyitömer Lignites by Multi-Gravity Separator, VIIIth Balkan Mineral Processing Conference, Belgrade, Yugoslavia, pp. 13-18. Koca, S. and Koca, H. 2005; Beneficiation of Graphite Fines from moulding Factory Wastes, Waste Management and Research, 23, pp. 338-342.
Thermoplasticity Improvement of Coal Blends by Adding Solvent-Extracted Coal Noriyuki OKUYAMA1) [email protected] Haruo KUMAGAI2) Hiroki SHISHIDO1), Koji SAKAI1), Maki HAMAGUCHI1), Nobuyuki KOMATSU1) 1) Coal & Energy Technology Dept., KOBE STEEL, LTD 2) Center for Advance Research of Energy Conversion Materials, HOKKAIDO UNIVERCITY
Abstract A coal extract, produced by thermal extraction and solvent de-ashing in 2-ring aromatic solvent, has an excellent thermoplasticity even though the parent coal appears no thermoplasticity. We named it “HPC, High Performance Caking additive”, and have been developing to utilize as a thermoplasticity accelerator to make strong coke for blast furnace. This study concerns with the effect of HPC addition to improve thermoplasticity of coal blends. Significant improvements in the thermoplasticity of coal blends were observed by HPC addition, especially with high blending ratio of slightly caking coals. The thermoplastic phenomenon of coal, softening, fluidizing and re-solidifying, was traced as the quantitative behavior of the molecule mobility of coal. The quantitative behavior of the mobile component was investigated by in-situ high temperature 1H-NMR relaxation time measurement1). Thermoplasticity appealed with generation of the mobile component, and the fluidity increased with increase in the mobile component, and re-solidified with disappearance of the mobile component. The mobile component was increased by HPC addition, and also the fluidity strongly increased. Each coal, HPC and coal blends were thermally extracted in the 2-ring aromatic solvent. The coal extraction yields were quantitatively related with the mobile component measured by the 1H-NMR relaxation time measurement. The maximum fluidity of coal was also related with the coal extraction yield. The amount of the solvent extractable component could be regarded as a scale parameter of the mobile component. The maximum fluidity of coal blends with HPC addition can be estimated from the arithmetic mean of each component coal’s extraction yield. Thus the optimum condition of HPC addition to make strong coke can be determined by this method.
Introduction We have been developing a noble coal upgrading process to make an ash-less coal by applying the solvent de-ashing technology. Coal is extracted in the coal-derived two ring aromatics at 350-400 oC. The solution is separated from the extraction residue by the gravity settling. Solvent is
recovered and recycled in the process. The extracted coal (HPC) and residue coal (RC) are obtained in this process. HPC is formed with the primary coal molecule those are released from the coal aggregated structure in the solvent extraction stage at high temperature. HPC appeals an excellent thermoplasticity even though the parent coal appeals no thermoplasticity2). HPC softens at low temperature (less than 300ºC), keeps a highly fluidity in a wide temperature range and re-solidifies at high temperature, nearly 500ºC. Therefore, HPC is available as a caking additive to make strong coke for the blast furnace. We participate in a Japanese national project called COURSE50 (CO2 Ultimate Reduction in Steelmaking process by innovative technology for cool Earth 50), in which we aim to apply HPC as a caking additive to make strong coke for the iron blast furnace, which maximize the use of hydrogen as a reductant, in order to minimize coke addition to the furnace. Thermoplasticity is one of the most important characteristics for coke making. Japanese coke and steel companies make blast furnace coke by blending various types of coals, but the strongly caking coals have to be the main components in the coal blends. On the other hand, increasing in the use of slightly caking coals is an important subject to avoid the shortage problem of strongly caking coals and to reduce accommodation cost of coal. Significant improvements in the thermoplasticity of coal blends were observed by HPC addition, especially with high blending ratio of slightly caking coals. The changing of the thermoplastic phenomenon with HPC addition is investigated in a view point of quantitative changing of the mobile component in this study. The quantity of mobile component was measured in two ways, one is the 1H-NMR relaxation time measurement and the other is thermal extraction of coal. Strong relation is observed between the mobile component and the maximum fluidity in Gieseler plastometry, and the optimum condition of HPC addition to make strong coke can be determined from those results.
Experimental (1) Materials Four kinds Australian strongly caking coals (A, B, B’, C) and a slightly caking coal (D) were used in this study. Table 1 shows the results of proximate, ultimate analysis and Gieseler plastometry (JIM M8801, similar to ASTM procedure) of the materials, and Table 2 shows the coal blending ratio of each blending base and total reflectance (Ro), total inert (TI) of each base. Base-1 consists with 75 wt% of strongly caking coals, coal-A : B: C : D = 15: 26: 34: 25, and Base-2 consists with 50 wt% of strongly caking coals, coal-A : B: C : D = 15: 30 : 5: 50. When HPC was added to coal blend, coal-C was exchanged with HPC in the case of Base-1, and coal-B was exchanged in the case of Base-2. HPC was produced using a continuous bench scale unit (BSU), 0.1 t/d of coal consumption as show in Figure 1. The feedstock for HPC was an Australian steaming coal (MO), extracted at 400 ºC for 1 hour, and insoluble matter was separated using the settling and filtration system. Table 3 shows the properties of MO-coal and HPC, and Table 4 shows the composition of the recycling solvent of BSU. This solvent was also used for the coal extraction experiments using a high-temperature and high-pressure filtration autoclave.
Table 1 Properties of blending coals Proximate analysis name
ash [wt%]db
A
10.3
VM
C
1)
Maceral analysis4) Thermoplasticity 3) ST MFT RT Log (MFD) Ro Exinite Vitrinite Inertinite
Ultimate analysis H
N
S Odiff. 2)
[wt%] ( daf )
21.2
90.2 4.6 2.0 0.6 2.6
433
[ddpm] [ºC] 481 502 1.43
[%]
[-] 1.46
0.0
86.0
14.0
TI [%] 20.7
B
9.6
23.0
90.5 4.6 1.4 0.4 3.1
435
464 486
0.70
1.31
0.4
73.4
26.2
28.4
B'
10.5
21.8
89.6 4.7 1.6 0.3 3.8
432
474 505
1.08
1.29
0.0
75.0
25.0
28.4
C
8.0
35.5
85.7 5.4 2.2 0.6 6.0
383
443 476
3.27
0.91
1.6
85.2
13.2
16.5
D
10.0
28.0
87.0 4.8 2.0 0.6 5.7
422
455 473
1.00
1.00
0.4
75.2
24.4
30.8
1) dry basis, 2) dry and ash free basis, 3) Gieseler plastometry (JIS M8801), ST: Softening temp., MFT: Maximum fluid temp., MFD: Maximum fluidity RT: Resolidification temp. 4) JIS M8816, Ro: Mean reflectance, TI: Total inert
Table 2 Coal blending rate Blending rate [wt%](d.b.) A B B' C D Ro (calc.) TI (calc.) 15 26 0 (34) 25 1.11 23.8 15 0 26 (34) 25 1.11 23.8 Base-1' 15 (30) 0 5 1.16 27.9 Base-2 50 Base-2' 20 0 (30) 0 1.19 27.6 50 ( ) :Exchanged coal with HPC in the case of HPC addition Base blend Base-1
Table 3 Properties of coal extraction solvent Ultimate analysis H N S Odiff. [%] BSU Rcycling solve 91.45 6.9 0.02 0.7 0.93 C
Atomic ratio H/C O/C 0.905 0.008
Component content, [wt%] 1-methylnaphthalene 54.23 2-methylnaphthalene 37.46 naphthalene 2.10 dimethylnaphthalene 1.92 biphenyls 0.32 benzenes (1-ring compound 0.68 paraffins 0.48 thiol (S-containing aromat 2.04 others 0.76
Coal & Energy Technology Dept, Takasago Works, KOBE STEEL
Figure 1 0.1t/d Bench Scale Unit for HPC process development
Table 4 Properties of MO-coal and HPC name
Proximate analysis ash VM C
Ultimate analysis H N S Odiff.
[wt%]db 1)
[wt%] ( daf
MO-Coal 12.2 HPC-0
0.1
2)
ST
)
Thermoplasticity MFT RT Log (MF) [ddpm] [ºC]
39.9
82.9 5.5
2.0
0.6
9.1
410
434
446
42.6
84.0 5.5
1.9
0.6
8.1
244 375-441 481
Maceral analysis Ro ExiniteVitrinite Inertinite TI [%] [-] [%]
0.78
0.71
2.2
81.8
15.8
21.4
>4.78
0.94
0.0 100.0
0.0
0.2
(2) Measurement of coal extraction yield High-temperature and high-pressure filtration autoclave was manufactured to measure the actual extraction yield as shown in Figure 3. Coal was pulverized under 1 mm of diameter and mixed with the solvent under 0.1 MPa of nitrogen pressure in the autoclave, which has 0.5 litters of volume, equipping with a magnetic induction agitator and casting electric heaters. The coal slurry was heated under 3 ºC/min of heating rate, which was same heating rate with Gieseler plastometry. After that, the heated slurry was flowed down to the filtrate receiver via the filter, which was a mesh type stainless filter (0.5 µm of pore size) equipped at the bottom the autoclave. The filtration pressure was kept about 0.05 MPa by controlling the receiver’s pressure. After that, the same solvent was injected into the autoclave to wash in the autoclave and the filtration residue on the filter. The filtration residue (RC) was completely recovered by washing with acetone and dried up in a vacuum dryer after evaporating the washing solvent. The solvent in the filtrate was removed by using vacuum distiller, and the extracted coal (HPC) was recovered after the vacuum drying. Coal extraction yield was calculated using the dry and ash-free weight value of coal and RC as follows. Coal extraction yield =
Feed coal (daf) - RC (daf) Feed coal (daf) Coal
・100 [wt%,daf]
Solvent
N2 Extraction and Filtration vessel
Filter plate
Filtrate Receiver
Fig.2 High-temperature and high pressure filtration autoclave (3) In-situ High-temperature and high pressure 1H-NMR relaxation measurement Coals were pulverized under 100 meshes and blending samples (Base-1, Base-2, Base-1 + HPC, Base-2 + HPC) were charged into quartz NMR sample tube (9.5 mm of diameter) and dried at 37º C for 24 hours. The spin-spin relaxation time, T2, was measured at 3 ºC/min of heating rate up to 520 ºC. The solid echo sequence2) was employed to generate 1H-NMR transverse relaxation signals using Techmag-Apollo NMR spectrometer equipped with a CSIRO produced high-temperature
1
H-NMR probe. 1 H-NMR relaxation measurement is a technique for evaluating the motion and mobility of molecules in terms of the relaxation behavior of protons, nuclear spins, which are attached to the molecules. The mechanism of spin-spin relaxation indicates the average distance between spins. For example, in the case of solids, each spin has nearby neighbors, and they interact each other, resulting in rapid progress of the relaxation and shorter T2 value. For liquids and gases, the relaxation process does not proceed rapidly because a much greater average distance between protons results in an enlargement of the T2 value. The obtained solid echo signals were curve fitted with one Gaussian and two exponential functions, namely the rapid decaying component, arising from the proton attached to the immobile component (G), the intermediate component (L2), and the most slowly decaying tail of the signal, arising from the proton attached to mobile components (L1). The fractional intensity of each component, f(G), f(L2), f(L1), and spin-spin relaxation time, T2(G), T2(L2), T2(L1) were calculated from deconvoluted curves1).
(4) Coal carbonization experiment The coal blends were carbonized according to the conventional metallurgical coke making procedures. Coals were pulverized less than 3 mm and HPC was pulverized less than 0.15 mm. Coal blends were charged into an iron-made container of 0.02 m3, at 800 kg/m3 of bulk density, and were heated to 1000 ºC at a rate of 3 ºC/min under inert atmosphere. Duration was 30 min. 9 containers were filled in a coke oven, which has 300 kg of coal charging capacity. Drum Test (JIS K2151) was carried out to measure the coke strength. This is common way to evaluate coke strength in Japan. A 10kg of coke sample of the +50mm square hole is placed in the tumble drum and rotated for 30 revolutions, removed, screened and replaced in the drum and given impacts by further 150 revolutions. The drum index, DI15015, means the percentage of remaining +15mm square hole after 150 revolutions. The larger number indicates the stronger coke.
Results and Discussions (1) Estimation of Gieseler fluidity of HPC Figure 2 shows the concept of Gieseler plastometry. A packed coal sample (5g, particle size 0.425mm) is heated at a heating rate of 3 ºC/min with constant torque (0.01 N・m). The fluidity curve is obtained as the changing of dial divisions per minute (ddpm) in which 100 ddpm means 1 rpm of the rod rotation. As shown in Table 4, the characteristic parameters of HPC, softening temperature (ST) and resolidification temperature (RT) were obtained, but the maximum fluid temperature (MFT) and the maximum fluidity (MF) could not be measured by using a conventional Gieseler plasto-meter, in which measurement limit was 60,000 ddpm (Log MF <4.78). Then a new plasto-meter, which expanded the measurement limit to 400,000 ddpm, was manufactured. But the maximum fluidity of HPC could not be obtained with the regulation torque. On the other hand, fluidity curves, indicating peak fluidities, were obtained with setting of several lower rotating torqueses, and the curves indicated almost same values of ST and RT as those values measured with the regulation torque.
Therefore a Gieseler curve of HPC was estimated by drawing extrapolation curves fitting to the MFT on the existent curve which obtained with the regulation torque and 200,000 ddpm of the measurement limit. Figure 3 shows the result. HPC indicates excellent fluidity, 1.24・106 ddpm (Log MF=6.09) of MF at 414 ºC of MFT. Dial divisions
Gieseler curve Constant torque Div./min
Coal
Softening
Max. fluidity (MF)
Resolidification
(ST) Molten metal bath
(MFT)
(RT)
Temp.
Figure 2 Concept of Gieseler plastometry
Log (Fluidity/ddpm)
MFT MF 414 ºC 1.24E6 ddpm
6
Extrapolation Measurement
4 2 0 200
300
400
500
Temperature [ºC] Figure 3 Expanded Gieseler curve and extrapolation curve for grasping the maximum fluidity of HPC (2) Improvement of Gieseler fluidity by HPC addition Figure 4 shows Gieseler curve of each coal. Coal-A and B are strongly caking coals. They have rather low fluidity but have higher RT than that of Coal-D, which is slightly coking coal. It suggests that they maintain the fluidity in the higher semi-coking reaction temperature range. Coal-C has high fluidity, and is used to increase the fluidity of the coal blend. HPC appeals excellent fluidity, softens at around 250 ºC and maintain highly fluidity till the semi-coking reaction range.
Figure 5 shows Gieseler curves of the coal blends and HPC additional cases. Base-1 has sufficient fluidity, but base-2 has insufficient fluidity since blending ratio of coal-C is only 5 %. In the case of HPC addition, Gieseler curves expand to wide range and higher fluidity. Especially, softening temperature shifts lower with increase in HPC addition as shown in Figure 6, and maximum fluidity strongly increases with increase in HPC addition as shown in Figure 7. It can be seen that those improvement degree for thermoplasticity are not linear, rather accelerating with HPC addition. And those effects appears more strongly on Base-2, which has weak fluidity blended 50% of slightly caking coal.
7
HPC
Log (Fluidity/ddpm)
6 5 4 3
C
2
A
D
1
B
0 200
300
400
500
Temperature [ºC]
5
5 4 3
Base-1 +HPC 10% +HPC 20%
2 1
Base-1
0 380
420 460 500 Temperature [ºC]
Log (Fluidity /ddpm)
Log (Fluidity /ddpm)
Figure 4 Comparison of Gieseler curves of each coal and HPC
4
Base-2
3 2
+HPC 20%
+HPC 10%
1 0 380
Base-2
420 460 500 Temperature [ºC]
Figure 5 Gieseler curves changing with HPC addition
RT
480 460
RT
440 420 400 380
Melting range
Melting range
ST or RT [ºC]
Base-2
Base-1
500
ST
ST
360 0
0 5 10 15 10 15 20 Amount of HPC in coal blend [wt%]
5
20
Figure 6 Expansion of the melting range by HPC addition
Log (MF/ddpm)
3 Base-1
2 1
Base-2
0 0
10
20
30
Amount of HPC in coal blend [wt%] Figure 7 effect of HPC addition on maximum fluidity (Log (MF)) (3) Correspondence with mobile component In-situ high-temperature 1H-NMR provides useful information on the modifications of the thermoplastic behavior of coal. The 1H-NMR spectra are deconvoluted into one Gaussian and two exponential functions, namely the rapid decaying component, arising from the proton attached to the immobile component (G), the intermediate component (L2), and the most slowly decaying tail of the signal, arising from the proton attached to mobile components (L1). The thermoplastic phenomena can be explained with the behavior of molecular mobility, and mobile component largely influences to the fluidity1). Figure 8 shows correspondence between fractional intensity of the mobile component, f(L1),
5 4
0.3
3
0.2
2
0.1
1
0 100
0 200 300 400 500 Temperature [ºC]
0.5 0.4
Base-2 + HPC 10%
5 4
0.3
3
0.2
2
0.1
1
0 100
Log (fluidity/ddpm)
0.4
Base-1 + HPC 10%
Intensity of mobile component [-]
0.5
Log (fluidity/ddpm)
Intensity of mobile component [-]
measured by in-situ high-temperature 1H-NMR relaxation measurement and Gieseler fluidity as a function of temperature, and compared 10 wt% of HPC addition or not. L1 component has higher molecular mobility and be regarded as the proportion of released fragment from the coal aggregated molecular structure by thermal relaxation based on hydrogen. The mobile components increase with increase in temperature, and rise up sharply in both blending base. In the case of HPC addition, mobile component increases more over. Especially in Base-2, it appeals more strongly. It can be seen that coal blend begin to soften when the mobile component increase to a certain revel, nearly 0.2 to 0.3 of the fraction intensity, and fluidity increases with increase in the mobile component. Gieseler fluidity correlates with mobile component quantitatively, and the maximum fluid temperature is almost same with the peak temperature at which mobile component achieves a maximum quantity in each case. Gieseler fluidity begins to decrease with decrease in mobile component and indicates resolidification nearly at 500ºC even though the mobile component is little bit remained with progresses of semi-coking reaction.
0 200 300 400 500 Temperature [ºC]
Figure 8 Correspondence between fractional intensity of mobile component and Gieseler fluidity as a function of temperature (4) Correspondence with coal extraction yield Thermal extraction behavior of coal in the solvent is also grasped in a viewpoint of molecular mobility. We reported that the changing in coal extraction yield with temperature was clearly explained with the information of quantitative and qualitative changing of the mobile component obtained by 1H-NMR relaxation time measurement3). Therefore, it is considered that the fluidity of coal in the thermoplastic stage corresponds with the coal extraction yield, which reflects the generation capacity of mobile component from coal. Figure 9 shows correspondence between the extraction yield of coal blend (dry and ash free basis) and the fractional intensity of mobile component as a function of temperature. The coal extraction yield increases with increase in temperature and suddenly decreased at higher than the peak temperature, which indicates the maximum extraction yield. The peak temperatures of the extraction yields almost same with the
40
0.4
30
0.3
20
0.2
10
0.1
0 100
0 200 300 400 500 Temperature [ºC]
50
Base-2 + HPC 10%
0.5
40
0.4
30
0.3
20
0.2
10
0.1
0 100
0 200 300 400 500 Temperature [ºC]
Intensity of mobile component [-]
0.5
Extraction yield [wt%] (d.b.)
Base-1 + HPC 10%
50
Intensity of mobile component [-]
Extraction yield [wt%] (daf)
peak temperatures of corresponding mobile components, and the coal extraction yields are increased by HPC addition on both blending bases. Both behaviors of coal extraction and mobile component with temperature agree at a lot of point. Especially, the maximum coal extraction yield seems quantitatively related with the maximum fluidity of coal. Figure 10 shows the maximum extraction yield of coal blends (dry coal basis) as a function of HPC amount. The extraction yield increases with increase in HPC amount linearly. As shown in Figure 7, it looks like with the increasing behavior of the maximum fluidity with HPC addition.
Figure 9 Comparison between thermal extraction yield of coal and fractional intensity of mobile component as a function of temperature
Extraction yield [wt%] (d.b.)
45 Base-1
40 Base-2
35 30 25 0
10 20 30 Amount of HPC [wt%]
Figure 10 Maximum extraction yield of coal blend as a function of HPC amount Therefore, it is considered that the coal extraction yield can be regarded as a scale parameter of the mobile component, and the maximum fluidity can be estimated from a data base of coal extraction yield. The maximum fluidity (MF) of coal blend is one of the important parameter to obtain the optimum blending conditions for making a high strength coke, but the MF value of coal
blend can not be obtained accurately by a simple arithmetic mean of each component’s MF value, especially in the case of using non or slightly caking coal, and in the case of using caking additive such as HPC. On the other hand, complex extraction yield of coal blend can be accurately obtained by the arithmetic mean of each component’s extraction yield. Figure 11 shows the relation between the arithmetic mean of each component’s maximum extraction yield (representative value) and the maximum fluidity of coal blend. There are good relationship between the maximum extraction yield and the maximum fluidity. Base-1 + HPC Base-2 + HPC
Log (MF/ddpm)
3
2
1
0 30
35
40
45
50
Extraction yield [wt%] (d.b.) Figure 11 Relation between the representative maximum extraction yield of coal blend and the maximum fluidity of coal blend (5) Coke strength Finally, carbonizing examinations were carried out to confirm the effect of HPC addition on coke strength. In these examinations, coal blends of Base-1’ and Base-2’ were prepared. In both blending bases, coal-B’ was used instead of coal-B. Figure 12 shows the changing of total dilatation of coal blend and changing of coke drum index of DI15015 as a function of MF value of coal blends, in which the MF value was estimated from each component’s extraction yield as shown in Figure 11. The drum index, DI15015, means the percentage of remaining +15mm square hole after the impact of 150 revolutions. Higher number of DI indicates higher strength of coke. In the case of Base-1’, high strength coke (DI > 86) was obtained even though HPC was not added. The DI value is increased more over to 88 by increase in MF value, and decreased in reversely. In the case of Base-2’, DI value was too weak to use for blast furnace without HPC addition, but DI was strongly increased with increase in MF value by HPC addition. Total dilation also increased with increase in MF value. The suitable dilatation brings an effect to fill the inter-particle void and to form a strong lump coke. It can be said that the optimum HPC additional rate can be determined to set up the maximum fluidity of the coal blend in the optimum zone, and coal extraction yield is a useful factor to estimate MF value in various conditions of coal blending.
Coke drum index, DI15015 Total dilatation [vol%]
( ): HPC additional rate [wt%]
90
75 60
DI (0)
(5) (10) Base-1’
(15)(20) (10)
85
(25)
(5)
80
Base-2’
(0)
TD Base-1’
40 20
Base-2’
0 0
1
2
3
4
Log (MF/ddpm) Figure 12 Relation between the maximum fluidity and total dilatation (TD) of coal blend, and coke drum index (DI15015)
Conclusions A coal extract, HPC, affects to improve coal fluidity in thermoplastic range. Thermoplasticity appealed with generation of the mobile component, and the fluidity increased with increase in the mobile component. It is confirmed that the mobile component is increased by HPC addition. The coal extraction yields are quantitatively related with the mobile component measured by the 1 H-NMR relaxation time measurement. The amount of the solvent extractable component can be regarded as a scale parameter of the mobile component. The maximum fluidity of coal blends with HPC addition can be estimated from the arithmetic mean of each component’s extraction yield. Thus the optimum condition of HPC addition to make strong coke can be determined by this method. Acknowledgement This study was carried out as a part of Japanese national project called COURSE50 (CO2 Ultimate Reduction in Steelmaking process by innovative technology for cool Earth 50). We thank NEDO for the support of this study. References 1) H. Kumagai, K. Tanabe, and K.Saito, ‘1H-NMR study of Relaxation Mechanisms of Coal Aggregate in Structure and Thermoplasticity of Coal, Nova Science, 2005, 35
2) N. Okuyama et al., Coal Preparation, 25 (4), 295 (2005) 3) N. Okuyama, K. Sakai, N. Komatsu, H. Kumagai, T. Inoue, R. Ashida, K. Miura, Proceedings in ICCS&T 07-3 (2009), South Africa
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: 3. Sustainability and Environment Carbon Capture and Integration: An Alternative Perspective to CO2 Emissions and Carbon Capture and Sequestration
Catherine A. McGanity University of Richmond 2010 International Studies: International Development [email protected]
Abstract: As carbon dioxide emissions come under greater scrutiny due to being linked with catalyzing the adverse effects of global climate change, high pollution-emitting coal power plants are called to lessen their contribution to rising greenhouse gas levels. However, the international increase in coal usage continues in the world‟s fastest growing economies, especially China, India, South Africa and Indonesia, despite such criticism (Coal Information: 2010; IEA, 2010). The potential for coal power plants to remediate their emissions through subterranean Carbon Capture and Sequestration (CCS) is economically viable, yet it does not correct the potential harm such storage procedures have on the environment or public health. Evidence analyzing harm created from water aquifers acidifying, since they are comprised of carbon-absorbing rock, is slowly bringing to light questions concerning the total human and ecological safety of CCS in the long term (Economics of geological CO2 storage and leakage; van der Zwaan & Gerlagh, 2008). Rather than seeing carbon dioxide as a pollutant, creating a procedure for Carbon Capture and Integration (CCI) can utilize these gaseous byproducts as usable secondary resources. Instead of diverting greenhouse gases from power plant smoke stacks to deep boreholes of porous stone, altering emissions flow directly to a contained environment for growing biomass (namely fast-growing and heavily carbon-dependent algae) provides a circular rather than linear method of carbon storage (Lehmann, 2009). Evidence already provides means of carbon dioxide transportation for other industries, namely industrial and chemical facilities. Grown biomass can be transformed and reintroduced into the energy production economy via two processes: 1) conversion into bio-ethanol using steam generated by excess heat from coal combustion and sold as a secondary output and 2) transformation into dehydrated algae cake and used as a solid fuel supplement should coal resource prices rise beyond sustainable profit-making levels. Outside of the energy economy, these integrative approaches enable future banking of emissions credits when larger carbon cap and trading institutions become regionally established. In both cases, Carbon Capture and Integration can be environmentally and economically beneficial for meeting the growing demand in energy currently placed on coal-based infrastructures while remaining conscious of the international repercussions from increasing carbon footprint levels.
Carbon Capture and Integration: An Alternative Perspective to CO2 Emissions and Carbon Capture and Sequestration 1. Introduction: A Valorization of Carbon Dioxide The current consumption model for coal-based electricity economies is inherently linear: extract, consume, emit and deplete. Yet energy inputs quarried from known global reserves are expended annually by the millions of barrels and billions of tons; in some cases, far outstretching the production levels of local utilities (IEA: Coal, 2010). With uncertainty future fuel availability, due to rising difficulty in finding local fossil fuel resources and increasing recognition of their inherently carbondense composition, renewable resources are gaining momentum. In response, the current coal industry is tasked with remaining economically sound while transforming its infrastructure to become environmentally aware (Hawk, 1999). Most recently, Carbon Capture and Storage (CCS) has been the standard industry response to this two-pronged challenge. To remove carbon dioxide‟s presence from airborne pollutants via CCS, a number of industrial, geological, and social prerequisites must already converge (Alie, Douglas & Davison, 2009). Targeting emissions generated by only one emitter rather than the entire coal industry, CCS continues a reactive and narrow abatement perspective rather than proactive and encompassing one. This view creates a number of issues for emitters, including proclivity towards adopting free riding mentalities and difficulty in accepting a universal carbon pricing mechanism. By applying this costly end-of-pipe technology on case analyses, the creation of carbon dioxide is not challenged but justified (Salmi, 2007). Attempts by coal engineers to plug the plethora of carbon emission leaks in the failing dam of global emissions does not address the linear chain of coal consumption, continually adding more and more pressure to the pollution abatement system. An alternative network is emerging for carbon dioxide treatment; incorporating renewable energy resources despite the assumption they are a direct threat to modern coal institutions. Building momentum for this prospective market shift towards combined coal and sustainable input markets is a real-world application of industrial ecology. Bringing this new perspective into consideration, examples of self-restoring cyclical patterns to global nutrients emphasize the opportunities for carbon dioxide consumption through the budding biomass energy market. This transformation from a linear to a cyclical market for consuming carbon inputs and byproducts, primarily through organic and industrial trading patterns, is Carbon Capture and Integration (CCI). Based on the eco-industrial parks of Europe, Australia and the United States, the handling of carbon pollutants as tradable goods provides a new outlet for coal producers‟ and consumers‟ emissions (Desrochers, 2001). Different from market dependence on fossil fuel inputs for survival, this innovative approach bypasses a potential „staples‟ trap by diversifying where and how resource demands can be met (Bridge, 2008; McGanity, 2009). With the reemergence of biomass and organic fuel resources, an opportunity for creating cleaner energy and extending infrastructure longevity has unfolded. Mainly due to algae‟s high population growth potential and its respective dependence on carbon dioxide consumption, research towards clean biofuels has spurred other discoveries in the field of organic biomass (Hammen & Osborne, 1959). With bioengineering analysis underway to raise algae efficiency in catering to liquid, gaseous and solid fuel combustion systems, microorganisms have also proven to be highly adaptable to currently utilized forms of energy generation as well as delivery methods. Furthermore, development of this new field has expanded into marketable research fields previously untouched by the energy economy, such as organic livestock feed, low-chemical wastewater
treatment, high saline aquaculture productivity, and investments in pharmaceutical and nutraceutical manipulation (Frazer, 2004). As the continual generation of energy is essential for economic activity and development, Carbon Capture and Integration enables energy providers to do what fossil fuel-trapped infrastructures cannot: provide a level of confidence in future capacity to locally provide clean electricity. Furthermore, these accomplishments can take place without the high fiscal demands for infrastructure construction typical of CCS or any other technology developed in post-industrial states. While the potential merging of carbon emissions with biomass energy markets is not likely to take place soon, the valorization of carbon dioxide as an input rather than a pollutant is gaining momentum. By creating new industrially symbiotic activities for developing clean, cyclical alternative energy resources, the modern coal and growing algae biomass industries can mutually benefit from viewing carbon emissions as usable byproducts.
2. Carbon Capture in its Present Form The greatest achievements towards lessening global carbon dioxide pollution rates will come from implementing not one but multiple proven methods for reducing aggregate carbon dioxide emissions. Devices employed in the struggle against this greenhouse gas typically assume at least one of three forms: preventives, sinks, and remediation methods. Each form is self-explanatory and looks at one aspect of CO2 emissions, analyzing how greater positive economic results can be reached while minimizing negative externalities. Within our current system of consumption, Carbon Capture and Sequestration (CCS) effectively combines preventative technologies and available sinks to tackle anthropocentric carbon dioxide emissions (Gough & Shackley, 2006). Due to the various styles of carbon capture and forms of sinks, CCS has multiple combinations readily available to handle carbon-based emissions (Gough & Shackley, 2006). Thus greater control is achieved in preventing high carbon-based emissions from polluting industries regardless of the market in which they operate. So far, it is the most detailed technical response to pollutants put into practice. As sources of man-made carbon dioxide differ by region and by industry, the array of prevention mechanisms and sinks enable CCS to customize to varying carbon-based pollutants. From power generation facilities to chemical production plants and fuel refineries, CCS has a significant ability to adapt to the needs of many carbon-based industrial activities. However, decisions on how to trap and sequester carbon depend on commercial accessibility to appropriate carbon dioxide treatment infrastructure, local transport resources, and chemical absorption potential (Bradshaw, 2009; Gough & Shackley, 2006; Peters et al, 2009). Multiple forms of capture treatment are currently under experimentation, namely pre-combustion, oxy-combustion, and post-combustion, with more innovations still yet to be tested (Grande et al, 2009). The commercial transportation of gases, while not necessarily in the form of CO 2, has undergone equal vetting through decades of conveying methane by pressurized conduits, airtight containers and lengthy hermetically sealed pipelines (Faltinson & Gunter, 2009). While capturing and moving carbon dioxide is relatively understood and achievable, its storage is contingent to external factors (Lin et al, 2007). All industrial work sites are location and scale dependent. Yet as emissions can be transported only so far before creating a larger carbon footprint, carbon emitters must confine their CCS systems within a radius of “latent abeyance.” This distance is established by a ratio of the local potential sequestration volume in natural formations to annual carbon offsets estimated by the emitter. Valued pollution levels could be extrapolated from data of previous years‟ carbon dioxide emissions.
A crucial aspect of this ratio is whether or not the provided storage volume available is finite (such as fixed magnitude of porous stone or undersea storage) or sustainable (such as botanic absorption through photosynthesis). Sinks are different from preventative measures as they cannot be altered, but are immutable environmental characteristics. While preventative methods for capturing carbon and before emissions enter the atmosphere are flexible, sinks must either react to injected emissions by absorption or by entrapment; otherwise CCS is not a practicable option. Therefore, the kind and magnitude of storage available will determine if such technologies are appropriate to apply. Assessing the economic viability of installing a CCS mechanism has greater real-world implications than described in cost-benefit or lifecycle analyses. Choosing to install a carbon storage system is directly tied to three main issues: how affordable are the necessary infrastructure components, how long will the technical investment accommodate the scale of capture required, and how accepting are local populations to the new carbon network. The preparedness of companies to invest in pre-, post- and oxy-combustion capture devices is thoroughly analyzed by Alie, Douglas & Davison; focusing on the flexibility of carbon systems, analysis only recognizes when positive assessments are reached regarding economic operability tradeoffs after incorporation, not societal (see Figure 1 below). Figure 1: Onion diagram for power plant with CO2 capture operability study
(Alie, Douglas & Davison, 2009) While there may be corporate incentives for carbon capture, transportation and storage, not all locations can provide all necessary requirement to adequately contain carbon emissions on an environmentally and socially responsible level. Looking at the needed prerequisites to efficiently and effectively install CCS with current coal-burning electricity generation, infrastructure expansion hinges most on anthropocentric categories being more qualitative rather than quantitative by nature. Locations of high-density industrial activity are typically removed from large human populations. Yet social interaction with the symptoms of manufacturing can lead to civil dissatisfaction. It is more difficult to assess the long term viability of emerging technologies when unwanted side-effects take place; funding to ensure a local community‟s drinking water quality is not appreciated until such potable resources become contaminated (Carroll et al, 2009; Frazer, 2004; Newmark et al, 2010; van der Zwaan & Gerlagh, 2008). Furthermore, moving high quantities of carbon dioxide by pipeline
to more distant sinks has its own risks. While this method of transport is environmentally viable, instances of gas leaks suffocating nearby populations could sway public opinion (Kuznetsov & Trofimov, 2006). In other cases, dislike of construction in general, adhering to the “Not In My Backyard” mentality most common in the developing wind energy or “windtricity” sector, can prevent the installation of carbon abatement or storage mechanisms (Pasqualetti, 2000). Abatement mechanisms are purportedly the silver bullet to limiting the carbon footprints of industrialization and highenergy production systems. Emissions capture from carbon-intensive operations has typically depended on one unrecognized facet: coal systems are point sources (EPA: Clean Air Act). It is easier to focus on a high emission footprint created from one location rather than thousands of locations each contributing a small percentage of carbon dioxide. Although pollutionemitting coal systems worldwide are aging, they are better targets for lessening greenhouse gas levels as point- vs. non-point sources (Hawk, 1999). With modern demand for energy still rising, Clean Development Mechanisms (CDMs) have only strengthened further coal consumption; even with abatement and capture technologies installed, global aggregate carbon emissions are increasing (Francoeur, 2009). Jevon‟s Paradox directly questioned this issue at the turn of the twentieth century: as efficiency and effectiveness rise, total coal consumption will increase beyond what past resource forecasters predicted or expected (Jevons, 1906). As electricity and energy demand is squarely placed on the shoulders of contemporary energy generators, coal consumption remains the least expensive and most plentiful input available to fuel utilities meeting energy market pressure. Yet, by designing a system around this primary input, the mechanism becomes dependent on coal resources for survival; should access to the fuel fail, the reliant infrastructure will consequently founder. This “Input-Infrastructure Trap” is evident by the current coal market: only coal will power coal-fueled power plants (Bridge, 2008; McGanity, 2009). The modern energy sector, still grounded in carbon-heavy coal resources, has fallen into the consumption trap it was predicted to encounter over a century ago (Jevons, 1906). While coal reserves are generally sound, not all states will have enough resources to empower future development and industry (Bridge, 2008; McGanity, 2009). Capturing and storing carbon does not alter the general demand for high-carbon density fuels. Nor does it confront carbon emissions already in the atmosphere; it only addresses emissions added hourly by coal-burning facilities. Utilities able to afford such an alteration in their systems are diminished in number by environmental and social constraints preventing CCS installation. Figuratively, this carbon treatment system is analogous to attaching a small band-aide to a large, bleeding wound: a small percentage of CO2 pollutants is trapped while the greater network of coal-burning power plants continue emitting copious amounts of exhaust. As this activity continues, the demand for coal will increase and exacerbate the remaining reserves available until the system no longer has any fuel left to expend. For some countries this prediction is centuries in the future, while for others it is forecasted within a decade (IEA, Coal: 2009). This emerging technology attempts to make infrastructure more sustainable; yet as it conforms to the style of consumption already practiced by the worldwide energy economy, it fails to recognize the fallibility of its most crucial input. While Carbon Capture and Sequestration is a modern adjustment to singular cases of pollution, carbon dioxide continues to manifest global consequences. This paper is not an attempt to detract from the potential benefits of Carbon Capture and Sequestration. A number of studies have provided key evidence towards the market-scale good created from offsetting carbon emissions by various storage means. While in some areas CCS technology is only in its infant stages of testing at competitive levels, evidence from ongoing research and experimentation suggest achievable future environmental benefits. Thus the
strength and constructive impact of CCS in decreasing total global CO 2 emissions provides optimism for the future ingenuity of engineers and scientists alike. Yet while positive data is available for carbon treatment, climate change requires a more pressing investigation of its weaknesses.
3. Industrial Ecology Carbon dioxide emissions are typically argued as toxic to the environment; with their general origin in fossil fuel systems, using coal to generate electricity is also becoming a topic in the carbon management debate. The one-time, sequential style of fossil fuel utilization flies directly in the face of natural ecosystems, which have integrated byproduct chemicals for millennia. However, an alternative view is gaining strength in the emerging field of Industrial Ecology (Erkman, 1997; Ehrenfeld, 2006). Emerging with formulas based on reuse and interchange, the mainstream acceptance of linear input consumption is now challenged by an alternative view infusing micro-economic models and ecological principles. Industrial Ecology encourages a systematic modeling of manufacturing and industrial activity based on those selfsustaining nutrient cycles and symbiotic relationships evident in nature (Frosch, 1992). Changing from a monochromatic system of use to a multifaceted network of resources, the benefits of carbon exchange and are only now drawing on data from global cycles in water, carbon, and nitrogen. The structure of this new organic-based blueprint agreed upon by the National Academy of Sciences (see Figure 2) demonstrates the capability for emissions to be resold for a profit within an industrial ecosystem and leads to an aggregate decrease in externalities (Jelinksi et al, 1992). Rather than continually adding inputs, resources are regenerated through multiple trade interactions and ultimately result in fewer overall emissions (Frosch, 1992). Figure 2: Type III model of the industrial ecosystem.
Jelinksi et al, 1992 Trade need not specifically takes place between materials extractors or growers, waste processors, materials processors or manufacturers and consumers; within a semi-permeable market structure, having limited resource inputs requires those businesses creating byproducts conform to the need of other local industrial demands. By catering to the needs of regional facilities, excess emissions and effluents are transformed into marketable goods.
Instances of companies using industrial ecology without creating an industrial park have taken place. One such example includes an alternative wastewater treatment system using engineered marshlands to remove harmful effluents rather than introducing foreign chemicals (McDonough & Braungart, 2002). Another model introduces the absorption of CO2 in large “new growth” forests: as the lumber industry creates a sustainable sink for offsetting carbon emissions, such processes strengthen and facilitate marketing their wood as sustainable products (Han et al, 2007; Jackson et al, 2007; Lindner & Karjalainen, 2007). Polluting companies benefit from this transaction as it allows them to organically treat emissions without investing in major CCS infrastructure or paying carbon taxes. As botanical growth is limited by access to carbon from the atmosphere, the presence and absorption capacity of this emission is integral for trees to survive; therefore this mutually beneficial practice ensures the growth of trees to meet the demands of lumber consumers and enable emitting factories to claim their operations are more environmentally friendly (Lindner & Karjalainen, 2007). This reinterpretation of the phrase “one man‟s trash is another man‟s treasure” is finding a relevant and timely place in the development of industrial symbiosis, where the wastes produced by one activity directly fuels those of another (Chertow, 2007). Emphasis is given to seeing industrial activities as companies working in concert and feeding off of each other‟s excess yield rather than autonomous consumers and emitters. Either directly or indirectly, this instance of “Industrial Symbiosis” is developing in production communities around the world (Chertow, 2007). Through incorporating the byproducts of neighboring manufacturing systems, companies have found alternative means of cutting costs and creating opportunities for higher efficiency within their own regional industrial systems (Beers et al, 2007).
4. Carbon Capture and Integration Taking a comprehensive view of the fossil fuels industry and greater global energy economy, the pernicious yet crucial use of coal is strengthened by the tenets of consumption as they are currently obeyed. Plagued by the inevitable NO X, SOX and CO2 pollutants tied to their electricity producing infrastructure, companies dependent on coal inputs limit themselves by not recognizing the demand created by other industries for these emissions. By adopting the nonstandard principles of reapplication than seclusion, a novel system of carbon reuse and market integration begins to take shape. Instead of replacing coal with another renewable resource, as is most commonly argued, I propose a greater coalescing of manufacturing and resource markets using carbon dioxide. Rather than sequestering this part life sustaining, part climate altering emission in geological formations, research can provide insight to what disconnected industries have demands for it. After all, what benefit is achieved when profit-making resources are left unconsumed by emerging market demand? With a perspective of exchange rather than sequestration, Carbon Capture and Integrations (CCI) puts into practice the concepts of localized byproduct reuse introduced by industrial ecology. Enabling an unconventional view of emissions, the cost-creating niche carbon dioxide appears to fill is challenged by a broadened market perspective in alternative utility. With this new aspect of carbon dioxide, the unrealized economic opportunities of emissions are given more prominence by looking beyond their destructive dimension. Including industries engaged in revenue-generating carbon use, the absorption of carbon dioxide by manufacturing rather than porous stone becomes more and more appealing. Facilitating this kind of application will expand both trading density and information generation beyond the current energy economy; with development rising in new linkages between apparently separate market sectors, developing a symbiotic carbon dioxide industry emerges in sectors creating the greatest demand. Research for wood- and jatropha-based biomass to meet industrial and individual energy needs
are already underway in many regions of the world (Doshi et al, 2009; Ekmann, 1998; Hooda & Rawat, 2006). However, the fastest growing industry propelling organic carbon dioxide treatment is a field frequently seen as wholly autonomous to or in direct competition with coal: algae biomass. Eco-industrial parks (EIPs) have profitably incorporated the ideas of Industrial Ecology to lower overall costs and become more environmentally aware (Desrochers, 2001). Instead of paying for waste treatment and disposal, neighboring companies mutually offset refuse materials by combining it with current operation systems to lower pollution levels and costs associated with more distant input sources. Examples in Denmark, Australia, the Ukraine, and the United States have successfully established means of incorporating excess inputs of nitrates, lime, water, and heat energy to benefit (Beers et al, 2007; Grdzelishvili & Sathre, 2007). Currently, the schematics of these parks do not include the carbon dioxide output from industrial groups‟ mechanized processes; however, plans in a majority of eco-industrial assemblies do exhibit potential for incorporating algae biomass, and thus carbon consumption. Figure 3: Schematic Adaptation of Kalundborg, Denmark Eco-Industrial Park for CCI
(Based on Ehrenfeld and Gertfer, 1997)
As CCS benefits from the inherent point-source characteristic of most coal consuming activities for capture and storage, so too can carbon integration. Since carbon dioxide output is confined to a specific outlet from coal combustion, it can then be transferred to aquaculture diffusion for growing algae in accessible ponds or photobioreactors (de Morais et al, 2007). Using the same engineering to pressurize CO2 gas in pipelines for CCS, excess heat and wastewater generated by coal consumption can also be harnessed to create an environment to grow algae (see Figure 3). Should nearby land be too valuable to spare, pipelines can carry gaseous emissions to further algae farms for consumption (Haszeldine, 2006). Stabilizing temperature, nutrient availability and water pH for commercial biomass growth can be accomplished by reusing materials found in wastewater and emissions created by coal use (Demirbas, 2007). Instead of stagnant sequestration, greenhouse gas recovery provides a constant source of food and steady essential climate variables for growing algae (Hammen & Osborne, 1959). The simplicity of Carbon Capture and Integration‟s potential application in many markets contribute to both the economic and environmental argument supporting this new system. While the emerging biomass-based fuel market currently poses a major threat to the foothold fossil fuels have in the world‟s energy economy, its blossoming research and development sector is in reality a boon for carbon-intensive industries (Gustavsson & Karlsson, 2006). As these new areas of algal research have a foundation in carbon consumption rather than generation, a maturing demand for carbon dioxide is materializing. While it is easy to argue biofuels created from algae biomass will lead to a very short-term sequestration of carbon dioxide, this is a very narrow perspective regarding its potential blooming in a variety of global markets (Herzog et al, 2003). Algae are present around the world in hundreds of ecosystems and maintain niches varying in the thousands. Therefore, the variety of chemicals created by this multitude of microorganisms is steadily addressing the needs once met by petroleumbased products. Examples of this include organic fertilizers, livestock and aquatic farm feeds, bio-plastics, biofuels, organic nutraceuticals and pharmaceuticals, and non-chemical tertiary wastewater treatment (Lal, 2004; Li et al, 2008). So long as wastewater and carbon dioxide can be efficiently diverted into neighboring algae biomass farms, both industrial and organic markets can gain benefits from this industrially symbiosis (Frazer, 2004; Hammen & Osborne, 1959). While fossil fuels are crucial raw materials for creating asphalt, green revolution fertilizers, chemical reactants, and plastics, their primary consumption for creating energy has enabled their stronghold in industrial development; yet, the commercial development of algae is projected to touch upon an equal array of markets. Through CCI, both organic and mechanical arenas in the world economy can benefit from algal biomass resale. By lengthening infrastructure longevity, lowering emissions costs, and strengthening byproduct revenue, fossil fuels can continue empowering industry while earning new profit from reselling its carbon dioxide than paying carbon taxes (Haszeldine, 2006). By constructively harnessing carbon dioxide for growing algae commercially, economic sectors previously untouched by coal can create demand for its byproducts by including this new multifunction biomass market as a resource. Taking the Schumpeterian perspective, CCI establishes a means of sustainably generating technological innovation through the creative destruction of carbon sequestration (Hart & Milstein, 1999; Hartshorn et al, 2005). While fossil fuels continue along a linear consumption model, CCI continues a millennia-long holistic and environmentally conscious, integrative and renewing network of organic resources. Whether consuming coal or moving to a less carbon-intensive fuel, the inclusion of the principles from industrial ecology could enable algae to reshape the modern world economy for the better.
5. Comparative Market Analysis In theory Carbon Capture and Integration furnishes a number of revenue-generating opportunities, yet more concrete evidence must legitimize this alternative network for carbon consumption than simple schemata and persuasion. Current funding of carbon dioxide research and development projects aim to utilize and improve upon Carbon Capture and Sequestration systems than experiment with eco-industrial parks; this is mainly because CCS has more supporting data. What information does exist regarding industrial ecology speaks more to planned structures and markets never coming to fruition than the success stories (Gibbs & Deutz, 2005). However, facts enumerating the application of multiple algal species are available to ascertain how biomass generation can fulfill the economic activities previously mentioned. Furthermore, emerging economic management studies support the long-term benefits to be gained from inter-organizational development (Fichtner et al, 2005). Extrapolation of potential algal population growth rates and habitat needs for reaching commercial scale also support the concepts introduced in those management studies. Using the genera listed below, carbon emitters can optimize CCI for nearby markets by choosing a species of algae that works best within their demands. Table 1: Algae Species and their Prospective Uses Algae Genus Spirulina Chlamydomonas Haematococcus Chlorella Dunaliella Euglena Schizochytrium Scenedesmus Xanthophyceae Chrysophyceae Emiliana
Main Byproducts Algae Genus Main Byproducts Nutraceuticals & Livestock Feed Rhodophyceae Pigments & Food Additive Biofuels & Dyes Prymnesium Biofuels Biofuels Gracilaria Wastewater Treatment & Fertilizer Biofuels & Wastewater Treatment Porphyra Nutraceuticals Nutraceuticals & Biofuel Laminaria Nutraceuticals Nutraceuticals Undaria Nutraceuticals Food additives Laminaria Fish food & Fertilizer Wastewater Treatment & Fertilizer Sargassum Fish food & Fertilizer Pigments & Food Additive Codium Nutraceuticals, Fish food & Fertilizer Pigments & Food Additive Ulva Livestock Feed & Fertilizer Carbon Abatement Based on data from papers and sources listed in References’ “Table” Subsection
While the idea of generating revenue instead of costs from carbon dioxide is environmentally and economically appealing, incorporating CCI with current market systems requires CO2 buyers recognize the benefits yielded by their market consumption. As mentioned before, large growth forests providing wood for lumber companies can absorb carbon dioxide with equal offsetting benefit as algae farms to a polluting company (Han et al, 2007; Jackson et al, 2007). Algae biomass can likewise absorb large quantities of carbon, yet benefit from faster rates of consumption (Hammen & Osborne, 1959). However, the level of reward gained by coal consumers is distinctly different between these two scenarios. While the lumber economy benefits from this transaction by marketing their stock as sustainable and responsibly grown, emissions at times must travel through the air a large distance before reaching new growth forests; by drawing carbon from the ambient air rather than pipelines or storage tanks, emissions-generated hazards such as acid rain are not prevented from taking place (Han et al, 2007; Lindner & Karjalainen, 2007).
Through the web of exchange shown in Figure 3, launching a CCI system with a coal-fired power plant also strengthens nearby companies that consume other excess resources generated by combustion. By strengthening the symbiotic ties of regional industries as well, both costs from carbon taxation are limited and revenue generated from reselling heat, fly ash and CO2 are strengthened. Yet if a coal consuming company is not a member of an eco-industrial park, producing biomass byproducts render the same profit potential (Gustavsson & Karlsson, 2006). With CCI, the carbon dioxide transportability achieved by CCS technologies are not forgotten; should carbon dioxide emitters need to send pressurized gas from 10 or 100 miles away, pipelines and storage tanks can efficiently make the journey possible (Haszeldine, 2006). The resulting goods generated by algae are still in demand by pharmaceutical companies, livestock and produce farms, wastewater treatment systems as well as coal burning facilities (see Figure 4). Within the energy economy, algae biomass can also contribute to coal companies by diversifying their energy resource holdings. “Syngas” operations, such as a Biomass Integrated Gasification Combined Cycle converts algae into bio-ethanol and bio-methane using steam generated by excess heat from coal combustion; by recycling excess heat, coal companies can provide both electricity and natural gas to their customers (Ehrenfeld and Gertfer, 1997). Should mining operations become too hazardous or coal prices rise beyond profit-making levels, utilizing dehydrated algae cake as a coal combustion additive can offset fuel costs and create greater infrastructure robustness against industrial vulnerability (Einarsson & Rausand, 1998; Keith & Rhodes, 2002). Studies expanding into this new organic fuel source have also provided data suggesting biomass can burn cleaner and provide more caloric energy per unit consumption compared to some forms of coal (Clarens et al, 2010; Ernsting et al, 2007). This not only benefits coal-firing power plants by lowering their future emissions forecasts, but also ensures combustion infrastructure remains viable as newer renewable resources gain popularity over fossil fuels (Einarsson & Rausand, 1998). In some circumstances, excess carbon dioxide can be used for preserving algae biomass for long distance shipment. To limit these groups of microorganisms to one field of interest is equal to saying petroleum can only be consumed for energy; the opposite is indeed the reality. This ease of conveyance is rising in importance as algae production plans are rising in interest to regions previously too arid to cultivate. Unable to irrigate because accessible aquifers held saline rather than fresh water or soil exhibited poor water retention, land once unfit for agricultural activity is finding new hope in algae biomass. Greatly in part to various strains‟ ability to grow in saline and hyper-saline water, poor soil can be rejuvenated using fertilizers from algae biomass that filter salts from underground water sources and introduce organic sources of nitrogen and phosphorous. Further development of algae can provide livestock feed to support livestock animals, in addition to invigorating soil quality leading to healthier grazing lands. While in industrial markets these concepts may fall on deaf ears, these facets of algae biomass are crucial for maintaining the longevity of agricultural institutions. In no way should the incentives of Carbon Capture and Sequestration be minimized when compared to Carbon Capture and Integration. Rather than storage, an economic gain has emerged through the use of waste products to catalyze the future extraction of other fossil fuels. In tandem with other fuel industries, petroleum and natural gas extraction uses carbon dioxide gases to facilitate pressurizing subterranean beds, thus making it easier for reserves located deeper within the earth to be obtained. This procedure has gained strength within the field of coal and natural gas exploration as it makes sense to increase profitability within a company‟s extraction operations than outsource cost minimization to unrealized future tax credits (Faltinson & Gunter, 2009).
Offsetting unwanted corollaries or outright abatement to avoid generating pollutants has become the ad hoc response to the coal industry‟s status quo perspective (Allenby, 2000). Retrofitting or transforming coal-based generation systems to accept cleaner burning natural gas and co-generate hydrogen as well as electricity is a preventative action that is gaining momentum under this strategy (Davison, John et al, 2009). Yet, proposing to attach any new technology to a pre-existing infrastructure must also recognize the viability of utilities commonly beyond the control the energy industry (Alie et al, 2007). Geological formations allowing carbon sequestration procedures cannot always be reached; in some instances, the capacity to harness road, railway or pipeline transport chains to convey emissions, or to appropriate mechanisms for pressurizing carbon dioxide, is not feasible (Haszeldine, 2006). The absence of international standards dictating the advancement of energy generation technology has prevented universal technological progress. Therefore the reliability of existing mechanisms cannot be guaranteed. Although at first glance investing in new infrastructure looks promising, it can prove to be a long and unfulfilling process depending on the need for systems renovation or construction. While coal usage can be economically instrumental for increasing wealth and productivity, environmental effects of using coal may also fall to the wayside from missing keystone capacities for effectively controlling carbon pollutants. While CCS uses waste products to further extract fossil fuels, this manipulation of emissions only creates an outward-expanding and self-reinforcing spiral of carbon pollution. In comparison, CCI strives to create a profitable yet closed carbon loop of byproduct resource trading and usage. While the benefits of Carbon Capture and Sequestration can be widespread, the incurred costs from construction and installation are deep. Capable of preventing large emissions streams from entering the atmosphere, CCS must be fiscally feasible before it can be technically implemented. CCS systems cost, both creation and operability, for an individual company can fall between 900 and 2100 USD/ kWh and cost up to $100 (€60)/ton of CO2 annually (Anderson & Newell, 2004; IEA: Slone et al, 2009; Pimentel et al, 2010). Additionally, CCS leads to concentrated cost for one company while millions potentially benefit from the carbon abatement; as an emitter typically spends spend millions of USD to abate carbon through CCS technology, large populations enjoy the highly diffused benefits of unaltered climate patterns and natural resource availability (Pimentel et al, 2010).
6. Potential on an International Scale Meeting international energy demands, from either gross consumption by the general public or production needs of private industry, is a difficult task. Clean development mechanisms (CDMs), which include Carbon Capture and Storage, are the response most accepted by members of the international energy community (Flues et al, 2009). Stressing the need to diminish the aggregate impacts of consumption, CDMs promote any method decreasing emissions created by the growing demand for energy (Jürgens et al, 2006; Sutter & Parreño, 2007). The repercussions from increasing total global carbon footprint levels are not tied to rates of consumption alone, but include the form of consumption as well. Yet by serving the energy needs of systems run with carbon-intensive fossil fuel inputs, the eventual rise in global carbon emissions will only be strengthened as such systems continue to support the linear consumption model of the modern coal input-infrastructure trap (Bridge, 2008; McGanity, 2009). Fossil fuels, and coal in particular, are proven through centuries of use as stable fuel inputs for electrical generation. However, issues at the forefront of energy policymakers‟ consciences engage in a continual battle: either prevent further
damage from carbon pollution to constituent property, or increase energy production to meet the rate of industrialization and development (Kverndokk et al, 2004). The conflict between carbon abatement and energy supply hinges a great deal on fuel resource availability rather than the Precautionary Principle, as the former holds larger sway in strengthening economic capacity (Lyster & Coonan, 2009). Data recording obtainable coal reserves from one year to the next and from country to country is fallible; since there is no third party checking reported data to be accurate or timely, coal deposits in reserve can fluctuate unchecked from year to year. In the case of China and Germany, both oscillation and stagnation of coal reporting has been witnessed. In the former, no new reserves have been listed since 1992 while national consumption has since risen exponentially; in the latter, drastic changes to reporting conditions to lower reported availability from a 100-year reserve to almost less than one year for Europe‟s largest exporter (Zittel & Schindler, 2007). In two of the world‟s most industrial states, inaccurate reporting of reserves makes coal, although reliable in usage, wholly unreliable in supply accuracy. International increases in coal consumption, fed by discoveries of large local or foreign coal deposits, make evident the effects population growth has towards increasing a demand for cheap energy inputs (Coal Information: 2010; IEA, 2010). The cycle of growing populations creating ever-increasing energy demand is a common assumption within the energy resource trading markets. Yet, the ramifications of fossil fuel demand on national growth are still highly variable from state to state (Bohringer & Rutherford, 2002). Evidence of greenhouse gas emissions altering global climate patterns, though obvious to citizens observing the depletion of farmland soils‟ nutrients or the rising sea water level consuming homes, is only in its infancy. Thus the development of coal industries within the recent half-century in energy sectors worldwide has made carbon emission management an issue both post-developed and developing states must resolve (Lackner & Sachs, 2005). Integrating coal from local or regional beds for energy creation demands both human labor as well as developed technical mechanisms for perpetual delivery. Once a coal-burning power plant is operational and running, it requires fuel inputs every hour of the day and every day of the year. Factors preventing a constant and comprehensive carbon dioxide management strategy most often include a lack of technology or capital to reinforce abatement infrastructure or improve upon it. In many regions of the world, offsetting such a continuous stream of carbon dioxide emissions is not possible using CCS. In some cases, restricting emissions has proven unlikely because the geological potential to store carbon is absent, because opposition has arisen due to fear of local resource contamination, because financial backing to enable CCS construction is not strong enough, or in some cases because of a combination of these factors (Bradshaw, 2009; Grdzelishvili & Sathre, 2007). In developing states especially, a constant supply of electricity for consumers it is not a viable option given lacking structural ability. Should extractable coal seams be found in a developing versus a post-developed country, how should such resources be treated? While pertinent, this question is beyond the framework underlined for this paper to answer. Governments, citizen input, market value, appropriate infrastructure availability or construction costs, industrial demand and capital availability will greatly dictate the answer (Bohringer & Rutherford, 2002; Holzman, 2010). What can be discussed is the following question: should a country decide to use its coal reserves or sell them, how can it benefit from either scenario by using CCI? The Carbon Capture and Integration network inspired by industrial ecology may provide more incentives with either choice of selling coal reserves abroad or consumption at a local level. Unlike the linkages created from carbon consumption, the assorted markets CCI touches upon all contribute to sustainable and organic development. In agrarian-based economies, expensive fossil fuel-based fertilizers prone to robbing soil of its nutrients and water retention volume over time stimulated
the green revolution (Korhonen & Seager, 2008). Using the byproducts of algae-infused CCI, organic fertilizers help rehabilitate soil content by increasing nitrogen, carbon and phosphorous compounds without harming its long-term resilience. Not only will microorganism population increase, aiding further oxygen fixation and increasing soil water retention, this augmentation of soil health will increase plant survival rates by reducing growth stressors beyond the first growing cycle. More plants reaching maturation will extend food resources for farming communities, potentially leading to surpluses taken to market for profit or to fodder livestock communities can breed for both direct consumption and dairy production. One such system, VERSATILE, has already been suggested for small developing communities, yet expands into childhood education development as well as biogas capture for household food preparation (Fernandez-Han, 2009). For this particular system, the few inexpensive inputs required for construction do not require designating a large surface area for installation; this is of major importance to small farmers dependent on their plots of arable land for earning a living. Recent tests of small aquacultures of large algae species, namely kelp and seaweed, have also proven useful for enriching soil on small island states depending upon every square acre of productive land available. Furthermore, opportunities for incorporating algae in development strategies do not need to worry about accommodating potentially invasive species as strains of algae performing various functions in a multitude of ecosystems can be found across the globe. Should a country decide to export their coal resources for a profit, CCI does not require a direct carbon dioxide input feed in order to function. As research and development of algae biomass has demonstrated, open-air tanks growing algae work equally well as direct bubbling units. The advantage to integrating carbon dioxide diffusion methods into this system is ensuring control of the algae populations‟ limiting input; however, open photobioreactors (PBRs) are equally as capable as their closed counterparts to produce large quantities of biomass, if only at a slower rate. Although it is easy to say states taking part in carbon absorption should simply have open PBRs and take carbon dioxide directly from the atmosphere for free, there are many positive externalities to be gained from purchasing coal emissions from another region or country. First and foremost, engaging in economic transactions with other states allows greater overall trade of products between national economies; imports for one state are exports for another, and in bartering deals carbon byproduct consumers gain more legitimacy in market contracts my simply sitting in the same room with large carbon dioxide emitting countries (Chadwick et al, 1985). Secondly, states can create economically symbiotic relationships with carbon emitters through sharing carbon offsets and banking carbon credits via internationally supported clean development mechanisms (CDMs), gaining further financial aid from the international community for expanding and strengthening the competitiveness of other sectors in their economy (Anderson & McKibbin, 2000; Ellerman et al, 2008). Finally, gaining exposure to international dialogue allows further advertising potential towards attracting clean technology research and development as well as foreign tourism (Chadwick et al, 1985). While many countries suffer from intellectual flight and “brain drains” as educated elites move to better paying jobs abroad, reinforcing cutting-edge technological and environmentally conscious development can assist the reverse to take place: local intellectuals can find reasons to stay within their home country as progress is established, as well as further immigration of educated researchers to stimulate further local innovations and boost national consumption. While this may at first sound much like the „pollution haven hypothesis,‟ it is entirely different (Dietzenbacher & Mukhopadhyay, 2006). When pollution havens develop, it is from manufacturing companies attempting to bypass costs incurred from taxation or being legally bound to pollution liability; in this case, CCI allows states to reap the industrial
benefits of access to free carbon byproducts without bearing the cost of making deep capital investments for developing heavy energy infrastructure. By supporting biomass development and a market demand for carbon, smaller bio-ethanol and bio-methane processing plants can provide alternative energy cleanly and at equally adequate levels than comparable coal systems. In states hoping to diversify their maturing national energy strategies, selling coal reserves to create biomass aquaculture networks may prove more economical in the long run for leapfrogging old infrastructure patterns with cleaner and more efficient technologies (Goldemberg, 1998; Hart & Milstein, 1999; Perkins, 2003).
7. Methodological Inertia Naturally, all new ideas attract both interest and criticism. Inertia stalling any momentum for this new model of carbon management will take multiple shapes and is expected to last many years before adoption of Carbon Capture and Integration is acceptable or acted upon in any form. Resistance against this idea will address to many varied yet equally viable issues: technical and economic challenges, difficulty in political ratification, lack of familiarity with industrial ecology, assumptions of future third party activity, gauged necessity of adoption, and carbon dioxide pricing schemes. Therefore, each issue must and will be methodically addressed in an effort to dissuade doubt regarding the potential future Carbon Capture and Integration systems has for future production and development. When addressing an idea‟s “bottom line,” it is difficult to designate specific figures or estimates for a system yet to be tested. However, identifying areas of concern where costs are most likely to take place is a much easier task. At this point, the economic challenge of valuating CO2 is wholly dependent on the actions of third parties in various regions around the world and emitting states‟ willingness to accept such pricing mechanisms. As the Chicago Environmental Exchange has already begun generating prices for carbon offsets markets, the European Union and the state of China have taken a different path and passed laws directly addressing carbon markets and CDMs rather than appraising carbon dioxide as a commodity (Capoor & Ambrosi, 2009; Holzman, 2010). Interest in expanding taxable carbon emissions markets from those currently taking place in the European Union and the Chicago Climate Exchange is still based on the „pollutant‟ conceptualization of carbon dioxide. However, by changing this viewpoint from harmful toxin to profit-making byproduct, government bodies interested in generating revenue by taxing carbon emissions or from the selling of carbon credits to corporations could ostensibly lose a great deal of revenue. While the past Kyoto Protocols were effective in setting benchmarks for carbon emissions and opened the door for trading between polluting countries, the lack of momentum achieved from the recent Copenhagen Summit has left a void for effective emissions oversight and enforcement. The absence of third party regulation, monetary penalization or any other form of polluter punishment creates a necessary void for CCS to operate successfully. If carbon abatement mechanisms are not reinforced as industrial goods, there will be less incentive to install such costly mechanisms and further willingness by companies to emit negative externalitygenerating carbon dioxide. After all, there will be no independent body to point the finger of blame upon polluters. CCI is different because it harnesses the money-generating capacity of market demand to enforce carbon management at a larger scale by making it a valuable trading commodity. However, the assumption has been made that CCI will also be accepted as a Clean Development Mechanism (Jürgens et al, 2006). If international bodies do not designate this new system as proactive and productive in managing carbon, many of the ideas mentioned will never find the propulsion to become successful.
The lack of across-the-board pricing or carbon trading mechanisms will allow loopholes for emitters to pollute rather than become included in commerce. Pricing within eco-industrial parks for carbon dioxide may take an insulated perspective for pricing carbon based upon the principles of local rather than national or global rates of supply and demand. While CCS is dependent on the banking carbon offsets and established carbon taxation systems to be economically viable, CCI is more dependent on the profitability of generated byproducts. Requiring third party market acceptance or enforcement of carbon dioxide combustion must also decide how and who pays for carbon taxes: coal extractors, coal burners, or energy consumers (Grainger & Kolstad, 2010). Coal-burning facilities can establish CCI and reap the rewards of biomass revenue regardless of carbon pricing, whereas plants installing CCS are directly relying on pricing potential of CO 2 within their region. If assumptions considering carbon prices are dictated by fluctuating availability of carbon and the willingness of institutions to purchase does not hold, then CCI will have an economic foundation upon which it can operate while CCS will not (Holzman, 2010). The greatest overall cost from this form of carbon management will most likely be from initial technology entrapment. While developing transport mechanisms for carbon will be first and foremost a challenge, once those systems are established and working they will mainly be utilized for this system and not subject to immediate market diversification. In this case, both CCS and CCI become subject to singular market purposes: carbon control. Consequently, pipelines transporting carbon dioxide will only transport carbon dioxide; however, railways used to move cars storing pressurized gas will have greater opportunity for alternative market use and not be subject to homogenous commodity trading. While retrofits and recycling of some components and unit groups are possible should either CCS or CCI not prove as profitable as predicted, capital turnover depend on case by case nuances unique to each adaptation. To balance this possible negative externality, the variety of markets CCI may touch over the long run can offset the inaugural costs incurred; based on the recent growth in interest and funding for algae biomass, the amount of money foreign governments and non-profit organizations are willing to designate to development infrastructure, the rising cost of fuel resources and the increasing dependence on these inputs for generating energy, it is reasonable to argue capital for development is available in the form of grants and aid spending if not accessible from local or regional banking institutions. The technical challenges may be most apparent by requiring out-of-the-box thinking when addressing unique geological features. While the apparatuses used in Carbon Capture and Integration are the same as those already employed in other gas-centered processes globally, potential real-world usage is a hurdle that must be taken into account. Rather that focus on matters already solved by hydraulic systems, such as uniform collection, containment, pressurization by volume, safe transportation, infrastructure, and distribution, appropriate application on a case-by-case basis is most important. As the natural gas industry has efficiently, safely, and effectively transported methane for consumption around the world, the difference in properties between it and carbon dioxide must be resolved before any construction can begin. Thankfully, these issues have already been analyzed and determined through the beginning research on Carbon Capture and Sequestration systems; while CCI cannot take a puzzle piece perspective of moving carbon emissions from emitter to consumer despite the pre-existence of applicable technology, the data already available from the installation and usage of CCS systems is an enormous boon to developing the multiple market opportunities of CCI. Hurdles exist beyond the realm of market integration and system adaptation; as CCI is a strategy designed for human consumption, human input at a political and social level will have a large effect on its eventual acceptance. There is a debate
growing in the international development community concerning the use of potable water to grow food and provide drinking water versus facilitating industrial activities (Frazer, 2004). One missing faculty is the control of effluents impairing those resources instrumental for citizens‟ daily lives. Either from the realization or fear of air- and water-borne toxins, human health concerns provide political pressure against industrial activities; as rates of infection and disease harm local populations regardless of wealth, race or caste, the sanitation of those crucial resources become more and more important (Jackson et al, 2007). Algae biomass incorporated in CCI bypasses this argument, as it is adaptable to potable, saline and wastewater sources (Valéry et al, 2004; Berndes, 2008). In many instances, algae need access to turbid post-consumption water in order to grow at the exponential rate it is reported to achieve (Hammen & Osborne, 1959; Kim et al, 2009). The potential prosperity created by expanding the magnitude of carbon dioxide incorporation will no doubt create greater competition within global agricultural, energy and pharmaceutical economies. Citizens and companies benefiting from subsidies within these spheres of trading will have to face new market challengers from CCI users. Despite receiving the full benefits of political buttressing via import taxes in some cases and tax rebates in others, a good deal of political pressure from special interest groups and lobbyists will put farmers with access to organic fertilizers as well as dual-input electrical generators from CCI under undue stress. Fighting against the mottos of “clean coal” and “home-grown,” commercial protectionism presents an enormous obstacle for states‟ attaining their full profit potential. Thus integrating carbon dioxide into alternative energy, wastewater treatment, and organic produce and livestock farming will most likely find increased benefits within regional markets than the world economy due to this unequal system of trade. This immobility in or inaccessibility to some of the world‟s largest national economies, which also happen to be some of the world‟s largest carbon emitters, will prevent widespread approval and excitement in the case of CCI. However, there is a silver lining: if states can develop their own markets self-sufficient from international commodities during the development of their coal reserves, such global economic access should not be necessary. Wealth can be generated despite the autonomous nature of foreign states‟ markets. Likewise, should a state chose to sell its coal reserves on the international market and purchase carbon dioxide emissions, such trading agreements can include a preferential trading status to the emissionsabsorbing party, thus bypassing the barriers constructed by economic protectionism. In this case, I assume the price of coal to be more expensive than carbon dioxide emissions; I do not see the opposite commodity prices taking place in the near or distant future, and therefore such carbon-based trading activities do not seem to be vulnerable in any way. The immediacy of investing in Carbon Capture and Integration hinges upon both recognizing and fearing the effects caused by global climate change. Decreasing consumption revolving around pollution-causing products is a double-edged sword for CCI. While abating activities requiring carbon-intensive inputs is important environmentally, the utilization of these same inputs is essential for CCI to function. As CCS was previously referenced as a metaphorical band-aid over a bleeding wound, advocates could argue CCI as the organic sutures binding the wound together, preventing further bleeding and aiding the healing process. CCI uses creative destruction as the norm towards environmentally aware and sustainable development, rather than continually attempting minor changes to “green” an industrial process (Hart & Milstein, 1999). Unlike Carbon Capture and Sequestration, the concept of Carbon Capture and Integration focuses on the long-term opportunity potential for renewable resources. The utility of algae-based development within multiple markets pervaded by fossil fuels will either face competition and thrive or not find a foothold. Despite these critiques against CCI, the greatest difference between sequestering and integrating carbon provides one main truth: CCI treats carbon dioxide already present in
the atmosphere while CCS does not. Furthermore, CCI allows for energy companies to provide electricity to their consumers should fossil fuels be removed entirely. As it will most likely take many years (if not decades) before the organization of third parties developing pricing mechanisms for carbon dioxide and market-scale algae farms reach production capacity, the inevitability of adopting CCI is unknown. And even with such institutions being formed, their ability to work in concert as described by industrial ecology is equally uncertain. Rather than an unforeseeable end to detrimental pollutant creation from linear fossil fuel dependence, the valorization of carbon dioxide as a potential input creates more than a “breath of fresh air.” Catalyzing cyclical industrial activity and energy generation, CCI establishes an alternative perspective of carbon dioxide and fosters new ideas and innovations benefiting human wealth as well as environmental balance.
8. Conclusion Slowing the rate of emissions is pivotal to checking the advance of global climate change (Bazzaz, 1990). Yet, the amount of carbon dioxide already present in the atmosphere and generated daily by modern industrial mechanisms presents a challenge for abatement efforts and environmental stability. Coal producers especially face harsh criticism for contributing to greenhouse gas emissions. Yet a conflict of interests is created for energy companies: either become environmentally friendly and limit energy generation or produce electricity regardless of pollution rates. Carbon Capture and Sequestration has been the ad hoc „greenhouse gas insurance‟ for the modern energy industry to combat emissions rates through technological achievements and research-based policy (Kosugi, 2009). Under this scenario, coal presents the most economically feasible means of meeting increasing global energy demands while seemingly addressing carbon dioxide created by the combustion process. However, CCS has a number of issues pertaining to adversely affecting global climate change. Foremost, a number of prerequisites concerning fiscal, technological, geo-environmental and social prerequisites must be addressed before CCS can ever be installed. Secondly, this system provides a case response to a global issue; as it is implemented at the case level, small overall change is achieved when comparing carbon offset by one emissions producer compared to a global industry based on coal utilization. Furthermore, CCS does not address the high levels of carbon dioxide already present in the atmosphere. While it can capture and store emissions, it cannot redress what harm has already come to fruition from high ambient carbon densities. This paper recognizes the current carbon emissions solutions researched by the coal community at large, critically analyzing the respective potential integration of technologies within the scope of our modern infrastructural capacity. Emerging with fresh takes on old problems, the field of Industrial Ecology establishes a new way of seeing technological mechanisms as networks of inputs and outputs rather than autonomous producers and consumers. With this in mind, the examples established in Denmark, Australia, the Ukraine and the United States provide visionary networks of surplus trading. This new view of emissions and effluents challenges the linear consumption model typical to production by producing an opportunity for cradle-to-cradle resource recovery (McDonough & Braungart, 2002). Rather than treating excess production of carbon dioxide as toxic waste, an opportunity for creating carbon consumption is gaining strength. Industrial Ecology critically analyzes the potential for the market resale of carbon dioxide, rather than its storage, in the form of Carbon Capture and Integration. As evidence emerges from research and development of biomass, new ecoindustrial systems and revenue-generating transactions begin to come into focus. While a number of these exchanges are many years in the future, the potential for increasing trade density among previously unrelated markets is no less possible
(Gustavsson & Karlsson, 2006). Forecasting of renewable biomass versatility could lead to commercial-scale trading of carbon dioxide emissions for the benefit of pharmaceutical, organic fertilizer, livestock and produce farming, nutraceutical, wastewater treatment, and energy generation operations. Instead of viewing CO2 as a taxable pollutant, emitters have new options available regarding not only how costs are abated, but also how revenue can be increased. The demand-creating potential for this „pollutant‟ outside of one economic sector, locally, regionally or internationally, is only the beginning of comprehending how pervasive the existence of carbon is within other systems. The possible implications of this broadened byproduct market perspective is global in nature. Creating infrastructure and fortifying industries in developing countries hinges on electricity availability; however, not all states maintain the resources to establish and maintain high levels of technological advancement (Lackner & Sachs, 2005; Bradshaw, 2009). With CCI, separate and distinct markets can become mutually strengthened through algal carbon dioxide consumption. As different states contain varying coal reserves, the benefits of establishing CCI from either developing coal infrastructure locally or selling reserves abroad are discussed, noting infrastructure necessities as well as the social obligations of governments taking part in development activities. However, the international politics of national economies is also considered; as protectionism and subsidies have become acceptable tools for safeguarding wealth within political boundaries, trade relations centered around carbon dioxide exchange analyzes the benefits for both producing and consuming states. All new ideas undergo a good deal of criticism and must overcome inertia both philosophically and physically; CCI is expected to encounter such bulwarks and outlines positive momentum bolstering the creation of a carbon market outside of direct government taxation. Since fuel-burning power plants are constrained by the inputs of their respective infrastructures, the uncertain resource availability of coal has lead to more scrutiny of the future role of fossil fuels. Additional lack of third party supervision, commercial development of algae biomass, change in status-quo fossil fuel consumption mentalities, capital investments and public acceptance have established credible hindrances to establishing CCI networks. Clean development mechanisms, while aimed at decreasing global carbon dioxide emissions, have created status quo responsive thinking to greenhouse gas emissions rather than fostering momentum towards researching, testing, and applying alternative preemptive measures (Flues et al, 2009). Pricing carbon dioxide especially has provided a key constraint towards inaugurating CCI within the coal industry. While there is a growing consensus that too much carbon dioxide has already entered the atmosphere, slowing the aggregate rate of carbon introduced into the air may prolong the time still available to study endangered natural systems and their fragile biotic webs. With Carbon Capture and Integration, the valorization of carbon dioxide engenders an alternative view of how pollutants play a larger global role both environmentally and economically. Taking holistic networks of actors rather than individual repercussions, engaging in carbon trading provides a greater potential for revenue generation than sequestering emissions underground. The commercial implementation of this concept may be many years in the making, yet the effects of coalescing activities centered around carbon dioxide are more tangible in fostering economic growth in the long term for both coal users and algal biomass industries beginning to establish a role in global markets.
9. References Figure 1: Onion diagram for power plant with CO2 capture operability study Alie, Douglas & Davison, 2009
Figure 2: Type III model of the industrial ecosystem. Jelinksi et al, 1992 Figure 3: Schematic Adaptation of Kalundborg, Denmark Eco-Industrial Park for CCI Ehrenfeld and Gertfer, 1997 Table 1 Data: De Morais et al, 2007 Kang et al, 2008 Kim et al, 2009 Li et al, 2008 Qiang, Hu et al, 2008 Tang et al, 1997 Extra information regarding tabular information can be found at the following websites: http://www.utex.org http://ccmp.bigelow.org http://www.ccap.ac.uk http://www.marine.csiro.au/microalgae http://wdcm.nig.ac.jp/hpcc.html Alie, Colin, Peter L Douglas, and John Davison. “On the operability of power plants with CO 2 capture and storage.” Energy Procedia 1.1 (2009): 1521-1526. Allenby, Brad. “Industrial ecology, information and sustainability.” Foresight 2.2 (2000): 163-171. Anderson, Kym and Warwick J. McKibbin. “Reducing coal subsidies and trade barriers: their contribution to greenhouse gas abatement.” Environment and Development Economics 5.4 (2000): 457-481. Anderson, Soren and Richard Newell. “Prospects for Carbon Capture and Storage Technologies.” Annual Review of Environment and Resources 29.1 (2004): 109-142. Andrews, Clinton and Matthew Cuddy. “Does Utility Deregulation Promote Eco-efficiency? The Case of Distributed Generation.” American Collegiate Schools of Planning Annual Conference (2001): 1-27. Aspen Publishers. “Federal Action and the Viability of CO 2 Capture and Storage.” Air Pollution Consultant 19.1 (2009): 1-4. Atimtay, Aysel T. “A Global Outlook to the Carbon Dioxide Emissions in the World and Emissions Factors of the Thermal Power Plants in Turkey.” Water, Air, and Soil Pollution 3 (2003): 325-335. Atten, Christoper, Amlan Saha, Isaac Liberman and Lea Reynolds. “Benchmarking Air Emissions of the 100 Largest Electrical Power Producers in the United States.” National Resource Defense Council (2010): 1-74. Audus, H. and P. Freund. “Climate Change Mitigation by Biomass Gasification Combined with CO2 Capture and Storage.” 7th International Conference on Greenhouse Gas Control Technologies Vancouver (2004): 1-9. Aune, Finn Roar, Rolf Golombek, and Sverre A. C. Kittelsen. “Does Increased Extraction of Natural Gas Reduce Carbon Emissions?” Environmental and Resource Economics 29 (2004): 379-400. Baris, K., A. Ozarslan, and N. Sahin. “The Assesment for CO 2 Sequestration Potential by Magnesium silicate Minerals in Turkey: Cases of Orhaneli-Bursa and Divrigi-Sivas Regions.” Energy Exploration & Exploitation 26.5 (2008): 293-309. Batten, David F. “Fostering Industrial Symbiosis With Agent-Based Simulation and Participatory Modeling.” Journal of Industrial Ecology 13.2 (2009): 197-213. Bazzaz, F. A. “The Response of Natural Ecosystems to the Rising Global CO 2 Levels.” Annual Review of Ecology and Systematics 21.1 (1990): 167-196. Béchet, Quentin, Andy Shilton, Oliver B. Fringer, Paul Muñoz, and Benoit Guieysse. “Mechanistic Modeling of Broth Temperature in Outdoor Photobioreactors.” Environmental Science & Technology 44.6 (2010): 2197-2203. Beers, Dick, Glen Corder, Albena Bossilkov, and Rene van Berkel. “Industrial Symbiosis in the Australian Minerals Industry: The Cases of Kwinana and Gladstone.” Journal of Industrial Ecology 11.1 (2007): 55-72. Benetto, Enrico, Patrick Rousseaux, and Jacques Blondin. “Life cycle assessment of coal by-products based electric power production scenarios.” Fuel 83 (2004): 957-970. Berndes, Göran. “Future Biomass Energy Supply: The Consumptive Water Use Perspective.” International Journal of
Water Resources Development 24.2 (2008): 235-245. Blomen, Eliane, Chris Hendriks, and Filip Neele. “Capture technologies: Improvements and Promising Developments.” Energy Procedia 1.1 (2009): 1505-1512. Böhringer, Christoph and Thomas F. Rutherford. "Carbon Abatement and International Spillovers." Environmental and Resource Economics 22.3 (2002): 391-417. Bradshaw, Carrie. “The New Directive on the Geological Storage of Carbon Dioxide.” Environmental Law Review 11.3 (2009): 196-203. Bridge, Gavin. “Global production networks and the extractive sector: governing resource-based development.” Journal of Economic Geography 8 (2008): 389-419. Campoy, Manuel, Alberto Gómez-Barea, Fernando B. Vidal, and Pedro Ollero. “Air-steam gasification of biomass in a fluidized bed: Process optimization by enriched air.” Fuel Processing Technology 90.5 (2009): 677-685. Cao, Yan, Yang Wang, John T. Riley, and Wei-Ping Pan. “A novel biomass air gasification process for producing tar-free higher heating value fuel gas.” Fuel Processing Technology 87.4 (2006): 343-353. Capoor, Karan and Philippe Ambrosi. “State and Trends of the Carbon Market 2009.” The World Bank (2009): 1-78. Carbo, Michiel C., Daniel Yansen, Chris Hendriks, Erika de Visser, Gerrit Jan Ruijg, and John Davison. “Opportunities for CO2 capture through oxygen conducting membranes at medium-scale oxyfuel coal boilers.” Energy Procedia 1.1 (2009): 487-494. Carroll, Susan, Yue Hao, and Roger Aines. “Transport and detection of carbon dioxide in dilute aquifers.” Energy Procedia 1.1 (2009): 2111-2118. Chadwick, Michael, Nicolas Highton, and John Palmer. “Developing Coal in Developing Countries.” Ambio 14.4/5 (1985): 249-252. Chertow, Marian R. “„Uncovering‟ Industrial Symbiosis.” Journal of Industrial Ecology 11.1 (2007): 11-30. Clarens, Andres F., Eleazer P. Resurreccion, Mark A. White, and Lisa M. Colosi. “Environmental Life Cycle Comparison of Algae to Other Bioenergy Feedstocks.” Environmental Science & Technology 44.5 (2010): 1813-9. Coal Industry Advisory Board (CIAB). “Clean Coal Technologies: Accelerating Commercial and Policy Drivers for Deployment.” International Energy Agency (2009): 1-54. Conti, John J., Paul D. Holtberg, Joseph A. Beamon, A. Michael Shaal, Glen E. Sweetnam, and Andy S. Kydes. “The National Energy Modeling System: An Overview 2009.” Energy Information Administration (2009): 1-83. Cooper, Cal (on behalf of the CO2 Capture Project). “A technical basis for carbon dioxide storage.” Energy Procedia 1.1 (2009): 1727-1733. Corella, José, José M. Toledo, and Gregorio Molina. “Calculation of the conditions to get less than 2 g tar/mn3 in a fluidized bed biomass gasifier.” Fuel Processing Technology 87.9 (2006): 841-846. Daniel, Humaira and Jose Luis Martin-Bordes (eds.). “Water Related Migration, Changing Land Use and Human Settlements.” 5th World Water Forum-“Bridging Divides for Water” Topic 1.2. (2009): 1-78. Davis, Chris, Igor Nikolic, and Gerard P. J. Dijkema. “Integration of Life Cycle Assessment Into Agent-Based Modeling.” Journal of Industrial Ecology 13.2 (2009): 306-325. Davison, John, Silvio Arienti, Paolo Cotone, and Luca Mancuso. "Co-production of hydrogen and electricity with CO2 capture.” Energy Procedia 1.1 (2009): 4063-4070. de Morais, Michele Greque and Jorge Alberto Vieira Costa. “Carbon dioxide fixation by Chlorella kessleri, C . vulgaris, Scenedesmus obliquus and Spirulina sp. cultivated in flasks and vertical tubular photobioreactors.” Biotechnology Letters 29 (2007): 1349-1352. Demirbas, Ayhan. “The influence of temperature on the yields of compounds existing in bio-oils obtained from biomass samples via pyrolysis.” Fuel Processing Technology 88.6 (2007): 591-597. Derks, P. W. J. and G. F. Versteeg. “Kinetics of absorption of carbon dioxide in aqueous ammonia solutions.” Energy Procedia 1.1 (2009): 1139-1146. Desrochers, Pierre. “Eco-Industrial Parks: The Case for Private Planning.” The Independent Review 5.3 (2001): 345-371. Dietzenbacher, Erik and Kakali Mukhopadhyay. “An Empirical Examination of the Pollution Haven Hypothesis for India: Towards a Green Leontief Paradox?” Environmental & Resource Economics 36.4 (2006): 427-449. Dong, Li, Shiqui Gao, Wenli Song, and Guangwen Xu. “Experimental study of NO reduction over biomass char.” Fuel Processing Technology 88.7 (2007): 707-715. Dong, Li, Guangwen Xu, Toshiyuki Suda, and Takahiro Murakami. “Potential approaches to improve gasification of high water content biomass rich in cellulose in dual fluidized bed.” Fuel Processing Technology 91.8 (2010): 882-8. Doshi, V., H. B. Vuthaluru, R. Korbee, and J. H. A. Kiel. “Development of a modeling approach to predict ash formation during co-firing of coal and biomass.” Fuel Processing Technology 90.9 (2009): 1148-1156. Ehrenfeld, John. “Industrial Ecology: Environmental and Economic Boon.” Communities & Banking (2006): 3-5. Einarsson, Stefán and Marvin Rausand. “An Approach to Vulnerability Analysis of Complex Industrial Systems.” Risk
Analysis 18.5 (1998): 535-546. Ekmann, J., J.C. Winslow, S.M. Smouse, and M. Ramezan. “International survey of cofiring coal with biomass and other wastes.” Fuel Processing Technology 54.1-3 (1998): 171-188. Ellerman, A. Denny, Henry D. Jacoby, and Annelène Decaux. “The Effects on Developing Countries of the Kyoto Protocol and CO2 Emissions Trading.” MIT Global Change Joint Program (2008): 1-42. Erkman, S. “Industrial ecology: an historical view.” Journal of Cleaner Production 5.1-2 (1997): 1-10. Ernsting, A., A. Boswell, N. Holland, H. Paul, C. von Weizsaecker, S. Semino, and T. Gilberston, T. “Agrofuels: Towards a reality check in nine key areas.” 12TH Meeting of the SBSTTA at the Convention on Biological Diversity Paris, France (2007): 1-34. Faltinson, John and Bill Gunter. “Integrated Economic Model: CO2 Capture, Transport, ECBM and Saline Aquifer Storage.” Energy Procedia 1.1 (2009): 4001-4005. Feron, Paul H. M. “The potential for improvement of the energy performance of pulverized coal fired power stations with post-combustion capture of carbon dioxide.” Energy Procedia 1.1 (2009): 1067-1074. Fichtner, Wolf, Ingela Tietze-Stöckinger, Michael Frank, and Otto Rentz. “Barriers of interorganisational environmental management: two case studies on industrial symbiosis.” Progress in Industrial Ecology- An International Journal 2.1 (2005): 73. Flues, Florens, Axel Michaelowa, and Katharina Michaelowa. “What determines UN approval of greenhouse gas emission reduction projects in developing countries?” Public Choice 145.1-2(2009): 1-24. Francoeur, Michel. “Coal Information: 2009.” International Energy Agency Statistics (2009): 1-515. Frazer, L. “Low Water Consumption: A New Goal for Coal.” Environmental Health Perspectives 112.5 (2004): 296-299. French, Richard and Stefan Czernik. “Catalytic pyrolysis of biomass for biofuels production.” Fuel Processing Technology 91.1 (2010): 25-32. Freund, P. “Making deep reductions in CO2 emissions from coal-fired power plant using capture and storage of CO2.” Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy 217 (2003): 1-7. Frosch, Robert A. “Industrial ecology : A philosophical introduction.” Proceedings of the National Academy of Sciences of the United States of America 89.3 (1992): 800-803. Gawande, Kishore, Robert P. Berrens, and Alok K. Bohara. "A consumption-based theory of the environmental Kuznets curve.” Ecological Economics 37.1 (2001): 101-112. Gibbs, David and Pauline Deutz. “Implementing industrial ecology? Planning for eco-industrial parks in the USA.” Geoforum 36.4 (2005): 452-464. Global CCS Institute. “Accelerating CCS.” Global CCS Institute (2009): 1-11. Goldemberg, José. “Leapfrog energy technologies.” Energy Policy 26.10 (1998): 729-741. Gough, Clair and Simon Shackley. “Towards a Multi-Criteria Methodology for Assessment of Geological Carbon Storage Options.” Climatic Change 74.1-3 (2006): 141-174. Grainger, Corbett A. and Charles D. Kolstad. “Who Pays a Price on Carbon?” Environmental & Resource Economics 46.3 (2010): 359-376. Grande, Carlos A., Rui P. L. Ribeiro, Eduardo L. G. Oliveira, Alirio E. Rodrigues. “Electric swing adsorption as emerging CO2 capture technique.” Energy Procedia 1.1 (2009): 1219-1225. Grdzelishvili, Inga and Roger Sathre. “Industrial Ecology in Transition Countries: Historical Precedent and Future Prospects.” Strategies to Enhance Environmental Security in Transition Countries. (2007): 95-115. Gustavsson, Leif and Åsa Karlsson. “CO2 Mitigation: On Methods and Parameters for Comparison of Fossil-Fuel and Biofuel Systems.” Mitigation and Adaptation Strategies for Global Change 11.5-6 (2006): 935-959. Hammen, Carl S. and Paul J. Osborne. “Carbon Dioxide Fixation in Marine Invertebrates: A Survey of Major Phyla.” Science 130.3386 (1959): 1409-1410. Han, Fengxiang X., Jeff S. Lindner, and Chuji Wang. “Making carbon sequestration a paying proposition.” Naturwissenschaften 94 (2007): 170-182. Harkin, Trent, Andrew Hoadley, and Barry Hooper. “Process integration analysis of a brown coal-fired power station with CO2 capture and storage and lignite drying.” Energy Procedia 1.1 (2009): 3817-3825. Hart, Stuart L. and Mark B. Milstein. “Global Sustainability and the Creative Destruction of Industries.” MIT Sloan Management Review 41.1 (1999): 23-33. Hartshorn, James, Michael Maher, Jack Crooks, Richard Stahl, and Zoë Bond. “Creative destruction: building toward sustainability.” Canadian Journal of Civil Engineering 32.1 (2005): 170-180. Haszeldine, R Stuart. “Deep Geological CO2 Storage: Principles Reviewed, and Prospecting for Bio-Energy Disposal Sites.” Mitigation and Adaptation Strategies for Global Change 11.2 (2006): 369-393. Hawk, E. W. “Options and Costs for CO2 Reductions at Coal-Burning Utilities.” 16th Annual International Pittsburgh Coal Conference Pittsburgh, PA: Consol Energy, 1999. 1-11.
Haykiri-Acma, H. and S. Yaman. "Interpretation of biomass gasification yields regarding temperature intervals under nitrogen-steam atmosphere." Fuel Processing Technology 88.4 (2007): 417-425. Herzog, Howard, Ken Caldeira, and John Reilly. “An issue of permanence: Assessing the effectiveness of temporary carbon storage.” Climatic Change 59.3 (2003): 293-310. Higgins, Paul A. T. “Carbon cycle amplification: how optimistic assumptions cause persistent underestimates of potential climate damages and mitigation needs.” Climatic Change 95.3-4 (2009): 363-368. Holloway, Sam. “Storage of Fossil Fuel-Derived Carbon Dioxide Beneath the Surface of the Earth.” Annual Review of Energy and the Environment 26.1 (2001): 145-166. Holzman, David C. “Climate Change Abatement Strategies: Which Way Is the Wind Blowing?” Environmental Health Perspectives 117.7 (2009): A296-A304. Hooda, N. and V. R. S. Rawat. “Role of Bio-Energy Plantations for Carbon-Dioxide Mitigation with Special Reference to India.” Mitigation and Adaptation Strategies for Global Change 11.2 (2006): 445-467. Hu, Qiang, Milton Sommerfeld, Eric Jarvis, Maria Ghirardi, Matthew Posewitz, Michael Seibert, and Al Darzins. “Microalgal triacylglycerols as feedstocks for biofuel production: perspectives and advances.” The Plant Journal: for cell and molecular biology 54.4 (2008): 621-39. Huang, Y., D. McIlveen-Wright, S. Rezvani, Y. D. Wang, N. Hewitt, and B. C. Williams. “Biomass co-firing in a pressurized fluidized bed combustion (PFBC) combined cycle power plant: A techno-environmental assessment based on computational simulations.” Fuel Processing Technology 87.10 (2006): 927-934. IEA Clean Coal Centre. “Control and Minimization of Coal-fired Power Plant Emissions: Zero Emissions Technologies for Fossil Fuels.” Information Energy Administration (2003): 1-17. Jackson, Robert B., Kathleen A. Farley, William A. Hoffman, Esteban G. Jobbágy, and Rebecca L. McCulley. “Chapter 19: Carbon and Water Tradeoffs in Conversions to Forests and Shrublands.” Terrestrial Ecosystems in a Changing World. Springer Berlin Heidelberg, 2007. 237-246. Jelinski, L. W., T. E. Graedel, R. A. Laudise, D. W. McCall, and C. K. N. Patel. “Industrial ecology: Concepts and approaches.” Proceedings of the National Academy of Sciences of the United States of America 89.February (1992): 793-797. Jenkins, B. M., J. D. Mannapperuma, and R. R. Bakker. “Biomass leachate treatment by reverse osmosis.” Fuel Processing Technology 81.3 (2003): 223-246. Jenkins, B. M., L. L. Baxter, T. R. Miles Jr., and T. R. Miles. “Combustion properties of biomass.” Fuel Processing Technology 54.1-3 (1998): 17-46. Józef, Dubinski. “Geological and geophysical aspects of the underground CO2 storage.” The 6th International Conference on Mining Science & Technology; Procedia Earth and Planetary Science 1.1 (2009): 7-12. Jürgens, I., B. Schlamadinger, and P. Gomez. “Bioenergy and the CDM in the Emerging Market for Carbon Credits.” Mitigation and Adaptation Strategies for Global Change 11.5-6 (2006): 1051-1081. Kang, Yun Hee, Jong Ahm Shin, and Myung Sook Kim. “A preliminary study of the bioremediation potential of Codium fragile applied to seaweed integrated multi-trophic aquaculture (IMTA) during the summer.” Journal of Applied Phycology 20 (2008): 183-190. Keith, David W. and James S. Rhodes. “Bury, Burn or Both: A Two-for-One Deal on Biomass Carbon and Energy.” Climatic Change 54 (2002): 375-377. Keller, Klaus, David McInerney, and David F. Bradford. “Carbon dioxide sequestration: how much and when?” Climatic Change 88.3-4 (2008): 267-291. Khan, A. A., W. de Jong, P.J. Jansens, and H. Spliethoff. “Biomass combustion in fluidized bed boilers: Potential problems and remedies.” Fuel Processing Technology 90.1 (2009): 21-50. Kim, Choong-jae, Yun-Ho Jung, Chi-Yong Ahn, Young-Ki Lee, and Kee-Mock Oh. “Adsorption of turbid materials by the cyanobacterium Phormidium parchydematicum.” Journal of Applied Phycology 22.2 (2009): 181-186. Klimmek, S., H.-J. Stan, A. Wilke, G. Bunke, and R. Buchholz. “Comparative Analysis of the Biosorption of Cadmium, Lead, Nickel, and Zinc by Algae.” Environmental Science & Technology 35.21 (2001): 4283-8. Korhonen, Jouni, and Thomas P. Seager. “Beyond Eco-Efficiency: a Resilience Perspective.” Business Strategy and the Environment 17.7 (2008): 411-419. Korre, Anna, Zhenggang Nie, and Sevket Durucan. “Life cycle modeling of fossil fuel power generation with post combustion CO2 capture.” Energy Procedia 1.1 (2009): 3771-3778. Kosugi, Takanobu. “Assessments of „Greenhouse Insurance‟: A Methodological Review.” Asia-Pacific Financial Markets (2009): 1-19. Kuznetsov, S. V., and V. A. Trofimov. “Geomechanics: Analysis of Local Fractures and Crack Growth in Coal Seams Under Compression.” Journal of Mining Science 42.2 (2006): 99-104.
Kverndokk, Snorre, Knut Einar Rosendahl, and Thomas F. Rutherford. “Climate Policies and Induced Technological Change: Which to Choose, the Carrot or the Stick?” Environmental and Resource Economics 27.1(2004): 21-41. Lackner, Klaus S. and Jeffrey D. Sachs. “A Robust Strategy for Sustainable Energy.” Brookings Papers on Economic Activity 2005.2 (2005): 215-269. Lal, R. “Agricultural activities and the global carbon cycle.” Nutrient Cycling in Agroecosystems 70.2 (2004): 103-116. Lapuerta, Magín, Juan J. Hernández, Amparo Pazo, and Julio López. “Gasification and co-gasification of biomass wastes: Effect of the biomass origin and the gasifier operating conditions.” Fuel Processing Technology 89.9 (2008): 828-837. Lara, Yolanda, Pilar Lisbona, Ana Martínez, and Luis M. Romero. “Comparative study of optimized purge flow in a CO 2 capture system using different sorbents.” Energy Procedia 1.1 (2009): 1359-1366. Law, Beverly E., Dave Turner, John Campbell, Michael Lefsky, Michael Guzy, Osbert Sun, Steven van Tuyl, and Warren Cohen. "Chapter 9: Carbon Fluxes Across Regions- Observational Constraints at Multiple Scales." Scaling and Uncertainty Analysis in Ecology. Dordrecht, The Netherlands: Springer. 2006. 167-190. Lehmann, Johannes. “Biological carbon sequestration must and can be a win-win approach.” Climatic Change 97.3-4 (2009): 459-463. Lenke, Hiltrud, Jürgen Warrelmann, Gregor Daun, Kerstin Hund, Ute Sieglen, Ulrich Walter, and Hans-Joachim Knackmuss. “Biological Treatment of TNT-Contaminated Soil. 2. Biologically Induced Immobilization of the Contaminants and Full-Scale Application.” Environmental Science & Technology 32.13 (1998): 1964-1971. Li, Siu-chung, Wen-xiong Wang, and D. P. H. Hsieh. “Feeding and absorption of the toxic dinoflagellate Alexandrium tamarense by two marine bivalves from the South China Sea.” Marine Biology 139.4 (2001): 617-624. Li, Yanqun, Mark Horsman, Nan Wu, Christopher Q. Lan, and Nathalie Dubois-Calero. “Biofuels from microalgae.” Biotechnology progress 24.4 (2008): 815-20. Licht, Stuart and Xingwen Yu. “Electrochemical Alkaline Fe(VI) Water Purification and Remediation.” Environmental Science & Technology 39.20 (2005): 8071-6. Lin, Hongfei, Takashi Fujii, Reisuke Takisawa, Toru Takahashi, and Toshiyuki Hashida. “Experimental evaluation of interactions in supercritical CO2/water/rock minerals system under geologic CO2 sequestration conditions.” Journal of Materials Science 43.7 (2007): 2307-2315. Lindner, Marcus and Timo Karjalainen. “Carbon inventory methods and carbon mitigation potentials of forests in Europe: a short review of recent progress.” European Journal of Forest Research 126.2 (2007): 149-156. Löschel, Andreas and Vincent M. Otto. “Technological uncertainty and cost effectiveness of CO 2 emission reduction.” Energy Economics 31 (2009): S4-S17. Lv, Dangzhen, Minghou Xu, Xiaowei Liu, Zhonghua Zhan, Zhiyuan Li, and Hong Yao. “Effect of cellulose, lignin, alkali and alkaline earth metallic species on biomass pyrolysis and gasification.” Fuel Processing Technology 91.8 (2010): 903-909. Lyster, Rosemary and Eric Coonan. “The Precautionary Principle: A Thrill Ride on the Roller Coaster of Energy and Climate Law.” Review of European Community & International Environmental Law 18.1 (2009): 38-49. MacKay, David J. C. Sustainable Energy- without the hot air. Cambridge, England: UIT Cambridge Ltd., 2009. Mallick, Hrushikesh. “Examining the Linkage between Energy Consumption and Economic Growth in India.” The Journal of Developing Areas 43.1 (2009): 249-280. Manancourt, Angela and John Gale. “Updating the IEA GHG Global CO 2 Emissions Database: Developments since 2002.” IEA Greenhouse Gas R&D Programme (2006): 1-11. Maria, Corrado Di and Edwin Werf. “Carbon leakage revisited: unilateral climate policy with directed technical change.” Environmental and Resource Economics 39.2 (2007): 55-74. Marland, Gregg, Bruce A. McCarl, and Uwe Schneider. “Soil Carbon: Policy and Economics.” Climatic Change 51.1 (2001): 101-117. Mascarelli, Amanda Leigh. “Algae: fuel of the future?” Environmental Science & Technology 43.19 (2009): 7160-1. McFarland, J. R., H. J. Herzog, and J. Reilly. “Economic Modeling of the Global Adoption of Carbon Capture and Sequestration Technologies.” 6th International Conference on Greenhouse Gas Control Technologies. Kyoto, Japan, 2002. 1-8. McLarnon, Christopher R., and Joanna L. Duncan. “Testing of Ammonia Based CO 2 Capture with Multi-Pollutant Control Technology.” Energy Procedia 1.1 (2009): 1027-1034. Michaelowa, Axel. “Policy Integration as a Success Factor for Emissions Trading.” Environmental Management 33.6 (2004): 765-775. Moreira, José Roberto. “Global Biomass Energy Potential.” Mitigation and Adaptation Strategies for Global Change 11.2 (2006): 313-333.
Mustacchi, C., P. Armenante, and V. Cena. “Carbon Dioxide Removal from Power Plant Exhausts.” Environment International 2.4-6 (1979): 453-456. Newmark, Robin L., Samuel J. Friedmann, and Susan A. Carroll. “Water Challenges For Geologic Carbon Capture and Sequestration.” Environmental Management 45.4 (2010): 651-61. Office of Energy Efficiency and Renewable Energy. “Algal Biofuels R&D at NREL.” National Renewable Energy Laboratory (2009). Oexmann, Jochen and Alfons Kather. “Post-combustion CO2 capture in coal-fired power plants: comparison of integrated chemical absorption processes with piperazine promoted potassium carbonate and MEA.” Energy Procedia 1.1 (2009): 799-806 Ogawa, Takashi, Uko Ohashi, Susum u Yamanaka, and Kiyoshi Miyaike. “Development of Carbon Dioxide Removal System from the Flue Gas of Coal Fired Power Plant.” Energy Procedia 1.1 (2009): 721-724. Orlic, Bogdan. “Some Geomechanical Aspects of Geological CO2 Sequestration.” KSCE Journal of Civil Engineering 13.4 (2009): 225-232. Panwar, N. L. and N. S. Rathore. “Potential of surplus biomass gasifier based power generation: A case study of an Indian state Rajasthan.” Mitigation and Adaptation Strategies for Global Change 14.8 (2009): 711-720. Parker, D., A. Hessl, and S. Davis. “Complexity, land-use modeling, and the human dimension: Fundamental challenges for mapping unknown outcome spaces.” Geoforum 39.2 (2008): 789-804. Parry, Martin. “Closing the loop between mitigation, impacts and adaptation.” Climatic Change 96.1-2 (2009): 23-27. Pasqualetti, Martin J. “Morality, Space, and the Power of Wind-Energy Landscapes.” Geographical Review 90.3 (2000): 381-394. Pearson, Peter J. G. “Energy, externalities and environmental quality: will development cure the ills it creates?” Energy Studies Review 6.3 (1994): 199-221. Pei, Qing, Jiang Dongmei, and Zhang Mengheng. “A Study of Contract Legal Issues in Chinese Carbon Market.” Ecological Economy 5.4 (2009): 314-322. Perkins, Richard. “Environmental leapfrogging in developing countries: A critical assessment and reconstruction.” Natural Resources Forum 27.3 (2003): 177-188. Peters, Glen P., Gregg Marland, Edgar G. Hertwich, Laura Saikku, Aapo Rautiainen, and Pekka E. Kauppi. “Trade, transport, and sinks extend the carbon dioxide responsibility of countries: An editorial essay.” Climatic Change 97.3-4 (2009): 379-388. Pieters, Janneke. “Using Carbon to Fight Carbon.” RenewableEnergyWorld.com. June 22, 2010. http://www.renewableenergyworld.com/rea/news/article/2010/06/using-carbon-to-fight-carbon?cmpid=BioNL-Tuesday-June29-2010
Pigosso, Daniela C.A., Evelyn T. Zanette, Américo Guelere Filho, Aldo R. Ometto, and Henrique Rozenfeld. “Ecodesign methods focused on remanufacturing.” Journal of Cleaner Production 18.1 (2010): 21-31. Pimentel, David, Rattan Lal, and James Singmaster. “Carbon capture by biomass and soil are sound: CO 2 burial wastes energy.” Environment, Development and Sustainability (2010): 1-2. Priddle, Robert (Coal Industry Advisory Board). “Coal and Sustainable Development- achieving Balance in Priorities.” World Summit on Sustainable Development (2002): 1-18. Prikhod'ko, V. P. and A. Yu. Val'dberg. “Choosing a Method for Collecting Gaseous Contaminants from Industrial Exhausts.” Chemical and Petroleum Engineering 33.3 (1997): 310-314. Rabinowitch, Eugene. “World's Energy Supplies and Their Utilization.” Proceedings of the National Academy of Sciences of the United States of America 79.4 (1951): 296-312. Rich, Bruce. “An Incoherent Climate Policy.” The Environmental Forum 27.3 (2008): 24. Richards, Kenneth R. “A Brief Overview of Carbon Sequestration Economics and Policy.” Environmental Management 33.4 (2004): 545-58. Salmi, Olli. “Eco-efficiency and industrial symbiosis- a counterfactual analysis of a mining community.” Journal of Cleaner Production 15.17 (2007): 1696-1705. Sardesai, Pushkaraj, Wayne Seames, Laura Dronen, and Evguenii Kozliak. “Exploring the gas-phase anaerobic bioremoval of H2S for coal gasification fuel cell feed streams.” Fuel Processing Technology 87.4(2006): 319-24. Schneider, Uwe A. and Bruce A. McCarl. “Economic Potential of Biomass Based Fuels for Greenhouse Gas Emission Mitigation.” Environmental and Resource Economics 24 (2003): 291-312. Schoeneberger, M. M. “Agroforestry: working trees for sequestering carbon on agricultural lands." Agroforestry Systems 75.1 (2008): 27-37. Scott, Michael J., James A. Edmonds, Matesan Mahasenan, Joseph M. Roop, Anthony L. Brunello, and Erik F. Haites. “International Emission Trading and the Cost of Greenhouse Gas Emissions Mitigation and Sequestration.” Climatic Change 64.3 (2004): 257-287.
Sengupta, Ramprasad, and Geetesh Bhardwaj. “Consumption pattern, trade, and greenhouse gas leakage in India.” Environment and Development Economics 9.3 (2004): 335-352. Senneca, Osvalda. “Kinetics of pyrolysis, combustion and gasification of three biomass fuels.” Fuel Processing Technology 88.1 (2007): 87-97. Sheehan, John, Terri Dunahay, John Benemann, Paul Roessler. “A Look Back at the U.S. Department of Energy‟s Aquatic Species Program- Biodiesel from Algae.” National Renewable Energy Laboratory (1998): 1-4. Sims, Ralph E. H. “Bioenergy to Mitigate for Climate Change and Meet the Needs of Society, the Economy and the Environment.” Mitigation and Adaptation Strategies for Global Change 8.4 (2003): 349-370. Singh, Ranjeet, M. K. Ram Reddy, Simon Wilson, Kaustubh Joshi, João C. Diniz da Costa, and Paul Webley. “High temperature materials for CO2 capture.” Energy Procedia 1.1 (2009): 623-630. Smith, Kathryn, Ujjal Ghosh, Ash Khan, Michael Simioni, Kohei Endo, Xinglei Zhao, Sandra Kentish, Abdul Qader, Barry Hooper, and Geoff Stevens. “Recent developments in solvent absorption technologies at the CO2CRC in Australia.” Energy Procedia 1.1 (2009): 1549-1555. Solomon, Susan, Gian-Kasper Plattner, Reto Knutti, and Pierre Friedlingstein. “Irreversible climate change due to carbon dioxide emissions.” Proceedings of the National Academy of Sciences of the United States of America. 106.6 (2009): 1704-9. Sondreal, Everett A., Steven A. Benson, John P. Hurley, Michael D. Mann, John H. Pavlish, Michael L. Swanson, Greg F. Weber, Christopher J. Kygarlicke. “Review of advances in combustion technology and biomass cofiring.” Fuel Processing Technology 71.1-3 (2001): 7-38. Spath, Pamela L. and Margaret K. Mann. “Biomass Power and Conventional Fossil Systems with and without CO 2 Sequestration- Comparing the Energy Balance, Greenhouse Gas Emissions and Economics.” National Renewable Energy Laboratory (2004): 1-38. Spliethoff, H. and K. R. G. Hein. “Effect of co-combustion of biomass on emissions in pulverized fuel furnaces.” Fuel Processing Technology 54.1-3 (1998): 189-205. Stables, Kay. “Educating for environmental sustainability and educating for creativity: actively compatible or missed opportunities?” International Journal of Technology and Design Education 19.2 (2008): 199-219. Stavins, Robert N. “Worried about Competitiveness?” The Environmental Forum 28.6 (2009): 18. Stephan, Gunter and Georg Müller-Fürstenberger. “Does Distribution Matter? Efficiency, Equity and Flexibility in Greenhouse Gas Abatement.” Environmental & Resource Economics 27.1 (2004): 87-107. Strømman, Anders Hammer, Edgar G. Hertwich, and Faye Duchin. “Shifting Trade Patterns as a Means of Reducing Global Carbon Dioxide Emissions.” Journal of Industrial Ecology 13.1 (2009): 38-57. Sutter, Christoph and Juan Carlos Parreño. “Does the current Clean Development Mechanism (CDM) deliver its sustainable development claim? An analysis of officially registered CDM projects.” Climatic Change 84.1 (2007): 75-90. Svoboda, Karel, Michael Pohorely, Miloslav Hartman, and Jirí Martinec. “Pretreatment and feeding of biomass for pressurized entrained flow gasification.” Fuel Processing Technology 90.5 (2009): 629-635. Swaine, Dalway J. “Why trace elements are important.” Fuel Processing Technology 65-66.1 (2000): 21-33. Tan, Raymond R. “From process integration to eco-industrial systems: Biological analogs and the road to sustainability.” Clean Technologies and Environmental Policy 10.1 (2007): 15-16. Tang, Evonne P. Y., Warwick F. Vincent, Daniel Proulx, Paul Lessard, and Joël de la Noüe. “Polar cyanobacteria versus green algae for tertiary waste-water treatment in cool climates.” Journal of Applied Phycology 9 (1997): 371-81. Tate, Cathy M., Robert E. Broshears, and Diane M. McKnight. “Phosphate Dynamics in an Acidic Mountain Stream: Interactions Involving Algal Uptake, Sorption by Iron Oxide, and Photoreduction.” Limnology and Oceanography 40.5 (1995): 938-946. Tinaut, Francisco V., Andrés Melgar, Juan F. Pérez, and Alfonso Horrillo. “Effect of biomass particle size and air superficial velocity on the gasification process in a downdraft fixed bed gasifier. An experimental and modelling study.” Fuel Processing Technology 89.11 (2008): 1076-1089. Tjernshaugen, Andreas. “Political commitment to CO2 capture and storage: evidence from government RD&D budgets.” Mitigation and Adaptation Strategies for Global Change 13.1 (2006): 1-21. Toth, Ferenc L. and Hans-Holger Rogner. “Carbon dioxide capture: an assessment of plausible ranges.” International Journal of Global Energy Issues 25.1/2 (2006): 14-59. UNFCC Chair. “Carbon dioxide capture and storage in geological formations as clean development mechanism project activities.” Subsidiary Body For Scientific And Technological Advice at the United Nations Framework Convention on Climate Change (2010): 1-4. Utaki, Takahisa. “Development of coal mine methane concentration technology for reduction of greenhouse gas
emissions.” Science in China Series E: Technological Sciences 53.1 (2010): 28-32. Ußner, Matthias and Franziska Müller-Langer. “Biofuels today and tomorrow: effects of fuel composition on exhaust gas emissions.” Accreditation and Quality Assurance 14.12 (2009): 685-691. Valéry, Loïc, Virginie Bouchard, and Jean-Claude Lefeuvre. “Impact of the Invasive Native Species Elymus athericus on Carbon Pools in a Salt Marsh.” Wetlands 24.2 (2004): 268-276. van Berkel, René. “Comparability of Industrial Symbioses.” Journal of Industrial Ecology 13.4 (2009): 483-486. van der Zwaan, Bob and Koen Smekens. “CO2 Capture and Storage with Leakage in an Energy-Climate Model.” Environmental Modeling & Assessment 14.2 (2007): 135-148. van der Zwaan, Bob and Reyer Gerlagh. “Economics of geological CO 2 storage and leakage.” Climatic Change 93.3-4 (2009): 285-309. van Egteren, Henry. “Regulating an Externality-Generating Utility: Environmental Taxes Under Limited Information.” Environmental and Resource Economics 21 (2002): 107-134. van Kooten, G. Cornelis. “Biological carbon sequestration and carbon trading re-visited.” Climatic Change 95.3-4 (2009): 449-463. Vuthaluru, Hari B. and Dong-ke Zhang. “Effect of coal blending on particle agglomeration and de-fluidization during spouted-bed combustion of low-rank coals.” Fuel Processing Technology 70.1 (2001): 41-51. Weiss, Martin, Maarten Neelis, Kornelis Blok, and Martin Patel. “Non-energy use of fossil fuels and resulting carbon dioxide emissions: bottom-up estimates for the world as a whole and for major developing countries.” Climatic Change 95.3-4 (2009): 369-394. Wilson, Elizabeth J., S. Julio Friedmann, and Melisa F. Pollak. “Research for Deployment: Incorporating Risk, Regulation, and Liability for Carbon Capture and Sequestration.” Environmental Science & Technology 41.17 (2007): 5945-52. Xiao, Penny, Simon Wilson, Gongkui Xiao, Ranjeet Singh, and Paul Webley. “Novel adsorption processes for carbon dioxide capture within a IGCC process.” Energy Procedia 1.1 (2009): 631-638. Xing, Yuquing and Charles D. Kolstad. “Do Lax Environmental Regulations Attract Foreign Investment?” Environmental and Resource Economics 21 (2002): 1-22. Yang, Fang, Bai Baojun, Tang Dazhen, Shari Dunn-Norman, and David Wronkiewicz. “Characteristics of CO2 sequestration in saline aquifers.” Petroleum Science 7.1 (2010): 83-92. Yepes, Tito, Justin Pierce, and Vivien Foster. “Making Sense of Africa‟s Infrastructure Endowment: A Benchmarking Approach.” The World Bank (2008): 1-42. Yuan, Xingzhong, Hongtao Cao, Hui Li, Guangming Zeng, Jingyi Tong, and Lihua Wang. “Quantitative and qualitative analysis of products formed during co-liquefaction of biomass and synthetic polymer mixtures in sub- and supercritical water.” Fuel Processing Technology 90.3 (2009): 428-434. Zhu, Qinghua, Ernest A. Lowe, Yuan-an Wei, and Donald Barnes. “Industrial Symbiosis in China: A Case Study of the Guitang Group.” Journal of Industrial Ecology 11.1 (2007): 31-42. Ziock, H.-J., K. S. Lackner, and D. P. Harrison. “Zero Emission Coal Power, a New Concept.” 1st National Conference on Carbon Sequestration, Washington, DC. (2001): 1-9. Zittel, Werner, and Jorg Schindler. “Coal: Resources and Future Production.” Energy Watch Group (2007): 1-47. Zulfiqar, Mohammad, Behdad Moghtaderi, and Terry F. Wall. “Flow properties of biomass and coal blends.” Fuel Processing Technology 87.4 (2006): 281-288.
Paper presented at International Pittsburgh Coal Conference 2010 http://www.engr.pitt.edu/pcc/
Future coal production outlooks in the IPCC Emission Scenarios: Are they plausible? Mikael Höök* Contact e-mail: [email protected] * Uppsala University, Global Energy Systems, Department of physics and astronomy, Box 535, SE-751 21, Uppsala, Sweden, webpage: http://www.fysast.uu.se/ges/ Abstract: Anthropogenic climate change caused by CO2 emissions is strongly and fundamentally linked to the future energy production. The Special Report on Emission Scenarios (SRES) from 2000 contains 40 scenarios for future fossil fuel production and is used by the IPCC to assess future climate change. Coal, with its 26% share of world energy, is a major source of greenhouse gas emissions and commonly seen as a key contributor to anthropogenic climate change. SRES contains a wide array of different coal production outlooks, ranging from a complete coal phase-out by 2100 to a roughly tenfold increase from present world production levels. Scenarios with high levels of global warming also have high expectations on future fossil fuel production. The assumptions on resource availability are in SRES based on Rogner’s assessment of world hydrocarbon resources from 1997, where it is stated that “the sheer size of the fossil resource base makes fossil sources an energy supply option for many centuries to come”. Regarding the future coal production it is simply assumed to be dependent on economics, accessibility, and environmental acceptance. It is also generally assumed that coal is abundant, and will thus take a dominating part in the future energy system. Depletion, geographical location and geological parameters are not given much influence in the scenario storylines. This study quantifies what the coal production projection in SRES would imply in reality. SRES is riddled with future production projections that would put unreasonable expectation on just a few countries or regions. Is it reasonable to expect that China, among the world’s largest coal reserve and resource holder and producer, would increase their production by a factor of 8 over the next 90 years, as implied by certain scenarios? Can massive increases in global coal output really be justified from historical trends or will reality rule out some production outlooks as implausible? The fundamental assumptions regarding future fossil fuel production in SRES was investigated and compared with scientific methodology regarding reasonable future production trajectories. Historical data from the past 20 years was used to test how well the production scenarios agree with actual reality. Some of the scenarios turned out to mismatch with reality, and should be ruled out. Given the importance of coal utilization as a source of anthropogenic GHG emissions it is necessary to use realistic production trajectories that incorporate geological and physical data as well as socioeconomic parameters. SRES is underpinned by a paradigm of perpetual growth and technological optimism as well as old and outdated estimates regarding the availability of fossil energy. This has resulted in overoptimistic production outlooks. Key words: Emission scenarios, CO2 emissions, future coal production, peak coal
1
1. Introduction The world’s primary energy supply is dominated by fossil energy and over 80% is derived from combustion of oil, gas and coal (IEA, 2009). Oil accounts for 34%, coal for 27% and natural gas for 21% (IEA, 2009). The use of fossil fuels is also the dominating source of anthropogenic greenhouse gasses (GHG), particularly carbon dioxide. In 2007, nearly 29 billion tons of CO2 were emitted due to fossil fuel consumption (IEA, 2009). Around 57% of all global anthropogenic GHGs derive from fossil fuel combustion, with energy supply as the largest contributing sector (Figure 1). Consequently, anthropogenic climate change caused by GHG emissions is strongly and fundamentally linked to future energy production. Studies of the future energy use and production are vital for understanding future GHG emissions. Among the fossil fuels, coal is the most carbon intensive fuel and causes the largest GHG emissions per energy unit produced. The use of coal is also the most criticised part of the fossil energy sector. Consequently, it is important to inspect what kind of expected coal usage and corresponding emissions that is reasonable in the future.
Figure 1: Shares of total GHG emissions. CO2 from fossil fuel use is dominating in terms of emissions. This strongly links energy production to GHG emissions and ultimately anthropogenic climate change to assumptions about future fossil fuel production. Derived from: IPCC (2007)
2
1.1 Aim of this study Coal is a major source of GHG emissions and an integral part of the world’s energy system for the foreseeable future. The IPCC emission scenarios (SRES, 2000) contain a number of outlooks for future coal production. This study aims to investigate those production outlooks and their feasibility, and the paper is mainly reflecting previous studies on coal and the fossil fuel use in the IPCC emission scenarios (Höök et al., 2010a; Höök et al., 2010b). This study will not venture into any form of climate change effects or spend any time reviewing or commenting the issue of man-made climate change. Such studies have already been made by others (Kharecha and Hansen, 2008; Brecha, 2008; Nel and Cooper, 2009). Instead, the focus will be strictly placed on the plausibility of future coal production trajectories. However, if assumptions regarding the most dominating source of GHG are shown to be flawed it will naturally have repercussions on the validity of the climate change projections derived from models using those emission scenarios as input. The fundamental guiding principle in modelling – garbage in, garbage out – should always be held dear.
2. Background to SRES The World Meteorological Organization (WMO) and the United Nations Environment Program (UNEP) established the Intergovernmental Panel on Climate Change (IPCC) in 1988. Its task is to assess scientific, technical and socio-economic information relevant for understanding anthropogenic climate change. The results have been published in several assessment reports and some special reports over the years (IPCC 1990; 1995; 2001; 2007). The IPCC has been using a set of scenarios, describing future development of society and emissions, to assess future climate change. The first set was published in 1990, followed by subsequent sets in 1992 and 2000. Titles, methods, classifications, assumptions have all changed over time and this has been reviewed by Girod et al (2009). All of the scenarios from 1992 where found to exaggerate one or more current climate and economic trends, leading to correspondingly exaggerated atmospheric greenhouse gas concentrations (Gray, 1998). Revisions were obviously needed and work was undertaken to develop new scenarios. The current scenario set is called the Special Report on Emission Scenarios (SRES) and were published in 2000. It is the basis for the majority of all long-term climate change projections, including those of the Fourth Assessment Report (IPCC, 2007). SRES is built on 40 different scenarios and evaluations of the associated GHG emissions. The scenarios are based on literature reviews, development of narrative storylines and the quantification of these story lines with the help of six integrated models from different countries. SRES illustrates that the future emissions, even in the absence of explicit environmental policies, very much depend on the choices that people make, how economies are structured, which energy sources that are preferred and how people use available land resources. IPCC state that “they represent pertinent, plausible, alternative futures” and derive from a descriptive and open-ended methodology that aims to explore alternative futures (SRES, 2000). The emissions scenarios are later used as an input to various climate models to depict how the climate may change under various assumptions of future anthropogenic emissions. The emission scenarios are neither claimed to be predictions nor forecasts, even though they are commonly used as such. Additionally, no probabilities or likelihoods are assigned to any of the scenarios. All scenarios are equally sound and valid, which was required by the Terms of Reference (SRES, 2000). The equal probability of each emission scenario has been described as a rather peculiar assumption (Höök et al., 2010a). In any event, this cannot be the case, since the 3
range is due to a combination of component ranges of uncertainty, and thus the extremes of this range must be less probable than the central estimate (Jones, 2001). The IPCC emission scenarios have been criticized for its inability to assign probabilities to the projections (Schneider, 2001; 2002). Additional discussion on the communicative issues and uncertainties within SRES can be found in Schenk and Lensink (2007). Fossil fuel combustion, the main source of anthropogenic GHG emissions, as well as future production, can be modelled by probabilistic methods (Kontorovich, 2009). This makes claims of equal probability for both high and low resource/production scenarios appear questionable. “Equally valid scenarios” rather materializes as an attempt to assign unjustifiably high weight to more extreme visions compared to reasonable outlooks.
2.1 Scenario overviews Each of the participating modelling teams behind the emission scenarios used computer models and experience considering long-range development of economic, technological and environmental systems to generate quantifications of the storylines, which develop the different scenarios. To simplify the procedure of depicting alternative future developments, each of the four scenario families is described by a specific storyline. The writing teams formulated the storylines in a process that identified driving forces, key uncertainties, possible scenario families and their logic (Höök et al., 2010a). Within each scenario family different variations of global and regional development and their implications for global greenhouse gas emission are explored. There is no business-as-usual scenario or disaster scenario and it was also decided that possible surprises, such as a new world war or major depression, should not be considered. This has been described as a built-in linear logic and utopian thought (Hjerpe and Linnér, 2008). The SRES storyline titles have been kept simple: A1, A2, B1 and B2. They can be shown very straightforwardly in a two-dimensional tree, which shows the global-regional focus and the economic-environmental orientation (Figure 2). Closer description of the scenario families can be found in SRES (2000). The four storylines and scenario families describe future worlds that are wealthier compared to the current world. It is important to notice that they do not include additional climate initiatives such as policies to limit GHG gases or to adapt to the expected climate change. Each scenario has a number of driving forces, such as population, economy, technology, energy, land-use, and agriculture. Different preferences and priorities lead to a huge variety of corresponding GHG emissions. The A1 storyline and scenario family describes a future world of very rapid economic growth, low population growth and the rapid implementation of new and more efficient technologies. The key plots are convergence among regions, capacity building, and increased cultural and social interactions with a significant reduction in the difference in per capita income. The A1 family is the largest and branches out in several subfamilies, each exploring an alternative future with different preferences. The A2 family contains visions of a very heterogeneous world. The key scheme is selfreliance and preservation of local identities and cultures. Fertility patterns across the globe converge very slowly, which results in high population growth. Economic development is primarily focused on regional growth and per capita economic growth and technology change is more fragmented and slower than in other scenario families. The B1 scenario family describes a convergent world with the same low population as in the A1 storyline, but with rapid changes in economic structures toward a service and information economy, with reductions in material intensity, and the implementation of clean and resource4
efficient technologies. The emphasis is on global solutions to economic, social and environmental sustainability, including improved equity, but without additional climate initiatives. Finally, the B2 scenario family describes a world in which the emphasis is on local solutions to economic, social and environmental sustainability. It results in a world with moderate population growth, intermediate economic growth and less and more diverse technological change than in the B1 and A1 storylines. B2 is oriented toward environmental protection and social equity, but also focuses on local and regional initiatives.
Figure 2. Schematic illustration of the SRES scenarios. The four scenario families are shown, very simplistically, as branches of a two-dimensional tree. Each scenario is based on a common specification of the main driving forces such as population, economy, technology, energy, landuse and agriculture. Adapted from SRES (2000) All qualitative and quantitative properties of scenarios belonging to the same family were set to match the corresponding features of the underlying storyline. Together, 26 scenarios were harmonized to have the same assumptions about global population and GDP development. The remaining 14 scenarios explore alternative interpretations of the four scenario storylines, such as different rates of economic growth and variations of population growth (Sivertsson, 2004; SRES, 2000). The main characteristics of the scenarios are presented in Table 1.
5
Table 1. Main characteristics of development in different scenario families and groups, as applied to harmonized scenarios. Adapted from SRES (2000) Family A1 A2 B1 B2 Subgroup Population growth GDP growth
A1C Low
A1G Low
A1 Low
A1T Low
A2 High
Low
B1
B2 Medium
Very high
Very high
Very high
Medium
High
Medium
Energy use
Very high
Very high
High
High
Low
Medium
Land-use changes Resource availability Technological development Change favouring
Lowmedium High
Lowmedium High
Very high Very high Low
Low
High
Medium
Medium
Medium
Mediumhigh Low
Low
Medium
Rapid
Rapid
Rapid
Rapid
Slow
Medium
Medium
Coal
Oil & Gas
Balanced
Nonfossils
Regional
Efficiency
“Dynamics as usual”
3. Future coal supply in SRES Resource availability in SRES (2000) is built around the works of Rogner (1997) and Gregory and Rogner (1998), and relies on them for detailed discussion of the estimated hydrocarbon amounts. Closer discussion on fossil fuel availability can be found in SRES (2000) and Sivertsson (2004). The message of Rogner (1997) is that the vast unconventional hydrocarbon occurrences and historically observed rates of technology change would allow hundreds of years with availability of fossil energy with low long-term costs, i.e. not significantly higher than the market price of the 1990s. Rogner (1997) also states that additional occurrences beyond the common resource base makes fossil fuels appear as an almost unlimited energy source. In summary, main point of Rogner (1997) is that “the sheer size of the fossil resource base makes fossil sources an energy supply option for many centuries to come”, provided that economy and technological progress are favourable. For coal, Rogner (1997) highlight the many fluctuations in world reserve and resource assessments. A more holistic compilation of the fluctuations can be seen in Tables 2-6 or Figure 3. There is no discussion about the historical trends in global reserve and resource assessments in the work of Rogner (1997), in contrast to Höök et al. (2010b). Table 2 and 3 show some dramatic changes in the estimated in situ resources in the published assessments by WEC and BGR. Rogner’s overview is also very brief in comparison with oil and gas, mostly calling out to Federal German Institute of Geosciences and Natural Resources (BGR) as the main source. The total coal resource is placed at 6 246 Gtoe, which would equal 8744 Gt of coal (assuming 30 GJ/ton coal). Nearly 60% of all coal is found in the most uncertain category. Rogner (1997) places the coal reserves at 1 003 Gtoe, equalling 1404 Gt of coal (assuming 30 GJ/ton coal). This is strikingly higher than the latest global reserve estimates of 826 Gt (WEC, 2009; BP, 2010). Even the latest estimate from the BGR (2009) only places total coal reserves at 503 Gtoe or 997 Gt. A full compilation of historical world coal reserve estimates can be seen in Tables 4 and 5. Rogner’s data can only be called outdated and should be replaced with more recent studies and preferably also a discussion of the historical trends in coal supply estimates. The global coal reserves reported by World Energy Council (WEC) contain a declining trend since late 1980s (Table 4). On the other hand, the BGR data does not show this 6
trend (Table 5). However, the BGR has been known to exaggerate coal reserves compared to the national agencies that provide the reports used by the WEC (Höök et al., 2010b). Table 2. Total resource estimates in Gt for selected countries and the world as published by the WEC. Data source Höök et al. (2010b) USA
China
India
FSU
Australia
South Africa
Germany
Poland
UK
Indonesia
World
1924
5 398
996
79
n.a
166
56
423
19
190
n.a
7 398
1929
3 829
1 213
79
489
166
58
423
n.a
190
n.a
n.a
1937
2 893
10 113
26
611
171
214
427
68
306
n.a
16 025
1948
2 897
n.a
n.a
n.a
n.a
90
n.a
n.a
n.a
n.a
n.a
1974
2 925
1 000
83
5 714
199
44
317
61
163
3
10 754
1976
3 600
n.a
86
6 948
372
93
n.a
n.a
190
3
11 724
1980
1 073
1 463
115
6 007
860
93
295
186
190
19
13 476
1986
1 570
2 737
115
5 502
785
133
332
197
186
23
11 990
1989
1 570
1 094
245
5 487
821
122
332
198
378
6
10 052
1992
1 570
954
245
5 487
821
122
241
207
378
64
10 236
1995
1 216
954
286
5 487
817
126
269
223
190
64
10 567
1998
1 570
n.a
n.a
4 141
454
126
308
74
n.a
n.a
n.a
2000
1 321
n.a
n.a
n.a
435
116
308
65
n.a
n.a
n.a
2004
1 564
n.a
n.a
n.a
431
115
100
170
n.a
46
n.a
2007
1 560
n.a
290
n.a
n.a
115
50
56
n.a
57
n.a
Table 3. Total resource assessments in Gt as published by the BGR. Year 1980 is seen as identical to the WEC report from the same year. Data source Höök et al. (2010b) USA
China
India
FSU
Australia
South Africa
Germany
Poland
UK
Indonesia
World
1976
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
14 185
1980
1 073
1 463
115
6 007
860
93
295
186
190
19
13 476
1988
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
11 610
1993
1725
935
201
5493
861
155
380
194
447
42
10 994
1998
1478
766
166
5499
371
55
371
168
227
18
9 636
2001
1 122
668
163
2 217
194
5
264
144
190
24
5 424
2004
1 090
1 090
259
2 295
298
164
88
70
5
11
4 668
2005
1 058
1 090
292
2 326
305
164
91
101
7
65
6 039
2006
1 359
4 367
252
2 915
194
49
85
179
190
27
9 554
2007
6 720
4 367
230
2 931
148
29
83
179
187
24
15 511
2008
8 120
5 509
278
4 336
359
31
191
401
188
93
20 767
7
Table 4. Recoverable reserve estimates in Gt for selected countries and the world as published by the WEC. Data source Höök et al. (2010b) USA
China
India
FSU
Australia
South Africa
Germany
Poland
UK
Indonesia
World
1924
5 398
996
79
n.a
166
56
423
19
190
-
7 398
1929
3 829
1 213
79
489
166
58
423
n.a
190
n.a
n.a
1937
n.a
n.a
5
309
9
8
109
15
130
n.a
691
1948
n.a
n.a
n.a
n.a
n.a
22
n.a
n.a
n.a
n.a
n.a
1974
182
80
12
136
n.a
11
65
23
4
1
591
1976
199
n.a
13
n.a
63
18
64
n.a
45
0
n.a
1980
182
140
84
233
59
33
60
39
45
1
882
1986
264
n.a
n.a
245
66
58
80
43
5
n.a
839
1989
215
731
63
241
91
55
80
40
4
3
1 598
1992
241
115
63
241
91
55
80
41
4
32
1 039
1995
241
115
70
241
91
55
67
42
3
32
1 032
1998
247
115
75
225
90
55
67
14
2
5
984
2000
250
115
84
225
82
50
66
22
2
5
984
2004
247
115
92
222
79
49
7
14
0
5
909
2007
243
115
56
222
77
48
7
8
0
4
847
2009
238
115
59
222
76
30
7
8
0
4
826
Table 5. Recoverable reserve estimates in Gt as published by the BGR. Year 1980 is seen as identical to the WEC report from the same year. Data source Höök et al. (2010b) USA
China
India
FSU
Australia
South Africa
Germany
Poland
UK
Indonesia
World
1976
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
792
1980
182
140
84
233
59
33
60
39
45
1
882
1988
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
880
1993
282
110
62
48
84
68
53
22
3
6
818
1998
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
806
2001
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
n.a
968
2004
283
115
125
206
105
49
7
10
0
5
990
2005
252
115
126
157
107
49
7
10
0
7
935
2006
247
192
96
204
79
49
41
16
0
7
1 019
2007
263
192
74
215
77
29
41
16
0
7
990
2008
263
192
81
215
77
31
41
16
0
4
997
8
Figure 3. Evolution of the worlds coal reserve and resource estimates made by WEC and BGR since 1924. Adapted from Höök et al. (2010) 9
4. Modelling future production All production numbers in SRES (2000) are presented in exajoules (EJ) or zetajoules (ZJ), which are units that are hard to grasp in layman’s terms. As a result, we have chosen to convert their figures to more commonly used units. For conversion, we have used 42 GJ as a ton of oil equivalent (toe) and expressed everything in Mtoe or Gtoe. SRES (2000) assume everything from low to very high resource availability to describe different futures scenarios. The future resource consumption is dependent on future price levels (either assumed as exogenous inputs or determined endogenously in the model) and future technology capable of mining unconventional resources. All the scenario story lines show elements of utopian thinking in particular regarding future technologies capable of decoupling economic growth from energy consumption, how globalization is assumed to even out economic differences and how access to energy is rising in the future (Hjerpe and Linnér, 2008). Regarding coal, the question is assumed to be only one of economics, accessibility and environmental acceptance. Geological availability and quality issues are simply regarded as noninteresting or solvable with new technology or higher prices. Rogner (1997) stressed the need to have dynamic upper limits and to include both anticipated technology advances and undetermined technological breakthroughs. Simplistic aggregate quantity-cost curves and learning effects are presented as the main articles for modelling energy sources, basically implying that production will “automagically” occur as rising prices makes more and more resources available. However, such an assumption fails to include the complex interaction with other potential energy sources and the fuel demand by society. Society is dependent on the flows of fossil fuel and the future production is defined through the size of those flows. The size of the tank, i.e. the resource base, is of secondary importance as it is the tap that governs the flow rate and future utilization of fossil fuels in the society. Vast but complicated resources are useless for preventing the coming of a production peak if they cannot be developed fast enough to offset the depletion of “easy” coal. Vast reserve bases have little to do with the likelihood of significant future production, as production is dependent on many more factors than just geological availability. Focusing on future coal reserve levels, while ignoring other factors that affect coal extraction, leads to an incomplete picture for future coal production. Future coal production will not be entirely determined by what is geologically available, but rather by the fraction of the practically recoverable amounts as well as the demand by consumers. One must remember that society demands energy, not just energy from coal. Hence, if this energy can be obtained less costly and more practically from other energy sources, potentially nuclear power or wind, those will be favored. Increased coal prices do not necessarily lead to increased production, increased reserves, and the transformation of resources into reserves. The price development and feasibility of other energy sources must also be considered, since it is the energy, and not the feedstock that is demanded.
4.1 A1 family This family features rich energy and mineral resources, while rapid technological progress reduces the amount of resources needed for the same level of production and increases the economically recoverable reserves. The A1 family also has some subgroups that explore variants of this rich and technological future. A1C depicts a coal-focused future where new clean coal technologies, such as sulphur removal, have made coal generally environmental-friendly with the 10
exception of GHG emissions. A1G describes an oil and gas rich future, with a rapid transition from conventional resources to rich unconventional resources such as methane hydrates. In fact, over 95% of the total gas resources in Rogner (1997) come from methane hydrates (Höök et al., 2010a). A1T describes a non-fossil future where rapid developments of solar and nuclear technologies replace fossil fuels. For coal, future production for this scenario family can be found in Figure 4. Massive increases in world production are generally expected and no peak at all is foreseen prior to 2100 in the average case. The A1C-subgroup includes a tenfold increase in world coal production. The other subgroups tend to rely on other energy sources than coal, with massive increases in global oil and gas production as a result (Höök et al. 2010a). Valero and Valero (2010) found that the proven reserves of coal, oil, and natural gas must increase by at least 56%, 139% and 288% to meet the fuel requirements on the A1-subfamily focused on fossil energy (A1F). In summary, one can only see that this scenario family is optimistic regarding future coal production and has a built-in assumption that no constraints or limitations will apply to coal extraction prior to 2100. A number of scenarios, such as A1 ASF with an oil peak at 182 Mb/d in 2020 or A1T Maria that phases out coal entirely in the long run, can be ruled out from disagreement with historical data. A number of scenarios shows a peak in coal production around 2050, in agreement with Mohr and Evans (2009) and Höök et al (2010b), but those scenarios also consume unrealistic amounts of oil and gas (Sivertsson, 2004). Table 3: Cumulative coal production 1990-2100 in the A1 family Scenario name Coal [Gtoe] A1 AIM 379.0 A1 ASF 1226.9 A1 IMAGE 533.8 A1 MESSAGE 451.4 A1 MINICAM 494.6 A1 MARIA 201.7 A1C AIM 1627.3 A1C MESSAGE 1151.9 A1C MINICAM 1427.9 A1G AIM 448.2 A1G MESSAGE 509.8 A1G MINICAM 901.4 A1V1 MINICAM 512.5 A1V2 MINICAM 540.8 A1T AIM 294.8 A1T MESSAGE 277.8 A1T MARIA 105.0 652.1 A1 AVERAGE
11
Figure 4. Projected world coal production the in A1 scenario family. Production ranges from 800 to nearly 27 000 Mtoe by 2100, with an average is 11 000 Mtoe without a peak prior to 2100. Adapted from Höök et al. (2010a)
4.2 A2 family This family describes a differentiated world which, compared to A1, is characterized by low trade flows, relatively slow capital stock turnover and slower technological progress. This family also depicts self-reliance in terms of resources and less weight on social, economic and cultural interactions between different regions. A2 is often used as the “business-as-usual” family, where economic development continues without environmental concern or major attempts to reduce the gap between rich and poor countries. Regions with rich mineral and energy supply will develop into more resource-intensive economies. Other regions with fewer resources will focus on minimizing import dependence and improve efficiency or make use of alternative inputs. The resource availability in different regions mainly determines the fuel mix used. High-income but resource-poor regions develop advanced post-fossil technologies while low-income but resource-rich regions generally rely on older fossil technology. The resource availability is generally rather conservative, and utilization is largely limited to conventional resources without venturing into the field of unconventional fuels, generally resulting in high coal consumption. Coal will become the dominant fuel since oil production peaks. From being a mere 25% of the fossil energy in the 2020s, coal grows to over 75% in 2100 (Valero and Valero, 2010). Coal production is projected with generally continue increasing trends throughout the rest of the century resulting in roughly six times higher coal production by 2100 on average (Figure 5). An exponential increase is more or less the general theme. In addition, Valero and Valero (2010) 12
also found that the proven reserves of oil, natural gas and coal would need to increase by 50%, 140%, respectively 89% in order to meet the projected demand in this scenario family. The projected outlooks would put enormous pressure on the USA, Russia, China, Australia, India and South Africa as they together hold roughly 90% of the world’s coal reserves and resources (Höök, 2010b). For example, major amounts of the Russian coal are located in the Tunguska and Lena basins in central Siberia (Thomas, 2002), and will not realistically be developed for many decades due to limited infrastructure as well as remoteness and harsh climate. Since 1860, the cumulative world production of coal has been around 165 Gtoe (Ion, 1974, 1979; Jenkins, 1989; Mitchell, 2003; BP, 2010), with 52 Gtoe alone produced in the last 20 years. With this in mind, the A2 family assumes that the cumulative production by 2100 will be many times the total historical output. Table 4: Cumulative coal production 1990-2100 in the A2 family. Scenario name Coal [Gtoe] A2 ASF 1113.4 A2 AIM 1136.6 A2G IMAGE 514.0 A2 MESSAGE 747.3 A2 MINICAM 1117.5 A2-A1 MINICAM 477.6 851.1 A2 Average
Figure 5. Projected coal production in the A2 family. Adapted from Höök et al. (2010a). 13
4.3 B1 family B1 assumed that the capital output ratio of resource exploitation is assumed to rise with increasing depletion, but this is counteracted by learning curve effects. Coal production costs are assumed to rise with increasing depth and rising labour wages but this offset by increased mechanization (for underground mining) and economic scale-up effects. Coal is significantly lower in the B1 family than in the A1 and A2 families (Figure 6). On average, coal production levels are not dramatically higher than present output. Even though, B1 ASF appears as a rather odd bird with its dramatic deviation from the other scenarios within this family. Other scenarios, such as B1 ASF with an oil production of 170 Mb/d by 2020 or B1 MARIA that features monotonically decreasing coal production after 1990, can also be ruled out due to poor agreement with actual data. Valero and Valero (2010) saw that this scenario family depletes around 70% of the worlds current coal reserves by 2100. Table 5 gives the cumulative production output to 2100. On average, most of the scenarios reach a maximum production around 2050 (Figure 6). This is in agreement with studies like Mohr and Evans (2009) or Höök et al (2010b). Table 5: Cumulative coal production 1990-2100 in the B1 family Scenario name B1 IMAGE B1 AIM B1 ASF B1 MESSAGE B1 MARIA B1 MINICAM B1T MESSAGE B1HIGH MESSAGE B1HIGH MINICAM B1 AVERAGE
14
Coal [Gtoe] 313.9 379.8 648.6 161.2 77.8 250.1 154.1 201.6 305.4 276.9
Figure 6. Projected coal production in the B1 scenario family. Adapted from Höök et al. (2010a)
4.4 B2 family Here the available fossil energy is assumed to be conservative. The availability of oil and gas expands only gradually and it not extended much beyond current conventional and unconventional reserves (as of late 1990s). However, coal is assumed to be abundant. Overall this results in relatively limited energy options for the world. Coal production is assumed to be at much higher production levels than present (Figure 7). A tripling of world coal output is the average of this scenario family, while certain individual scenarios can be even more extreme. Table 6: Cumulative hydrocarbon production 1990-2100 in the B2 family Scenario name Coal [Gtoe] B2 MESSAGE 300.7 B2 AIM 571.6 B2 ASF 785.9 B2 IMAGE 426.9 B2 MARIA 454.0 B2 MINICAM 433.7 B2HIGH MINICAM 1056.1 B2C MARIA 608.7 579.7 B2 AVERAGE
15
Figure 3: Projected coal production in the B2 scenario family. Adapted from Höök et al. (2010a)
5. Comments on the SRES production outlooks Generally, the most favourable formations have been exploited first while more challenging and/or undesirable formations have been left. In fact, various studies show that depletion can make up for technological progress in the industry (Livernois, 1988; Tilton, 2003; Rodriguez and Arias, 2008; Topp et al., 2008). In essence, the increased costs resulting from the need to mine less favourable formations, i.e. more complex geology, more remote areas, deeper or thinner coal seams, lower quality hydrocarbons and similar, have been matching the cost reductions done by increased mechanization and introduction of new technologies. However, SRES often disregard such information based on little more than optimistic expectations on new technology. The main message of Rogner (1997) directly shows misunderstanding of the actual problem as well as avoidance of the key question, namely future production. Society is dependent on fossil fuel flows and future production is about the size of those flows. The size of the tank, i.e. the resource base, is of secondary importance as it is the tap that governs the flow rate and future utilization of fossil fuels in the society. Vast unconventional hydrocarbon resources are useless for preventing the coming of a production peak if they cannot be developed fast enough. Once again it must be stressed that vast resources have little to do with the likelihood of significant future production, as production is dependent on many more factors than just geological availability. The important question for society is not the size of reserves. Rather it is the actual flow of energy that can be obtained from the reserves.
16
Other peculiar details can also be found in SRES (2000). One of those things is expectations on coal-to-liquids (CTL) technology. CTL technology costs are assumed to be very low, typically below 30 dollars/barrel and even as little as 16 dollar/barrel in some cases (SRES, 2000). Such assumptions seem rather unsound compared to more recent and updated assessments, which ends up around 48-75 US$/barrel (Vallentin, 2009). The lack of details regarding CTL conversion ratios have been highlighted by Höök and Aleklett (2009). CTL is assumed to be a vital part of future fuel supply in several scenarios. CTL normally yields 1-3 barrels per ton of coal consumed (Höök and Aleklett, 2009). For example, the world CTL production (32 Mb/d) by 2100 in the B2 MESSAGE scenario is expected to be higher than the world oil production at the same time. Such a CTLcapacity would alone require more than the present world coal consumption to fuel the synthetic fuel plants, in addition to the amounts needed for “conventional” use of coal in power plants, steel mills and other industries. Can such major expectations on CTL really be made with little more than vague arguments based on technical possibilities? Patzek and Croft (2010) highlights how 16 of the 40 coal scenarios in SRES simply grow exponentially until the year 2100. Since all scenarios are assigned equal probability, these unreasonable outliers were a weight equal to the more realistic lower scenarios. Patzek and Croft (2010) continues to point out that policy makers tend to focus on the most extreme outcomes, and these outliers have gained severe prominence as inputs to the subsequent climate models. Assigning equal probability to everything ranging from zero coal production to a tenfold increase compared to present levels cannot be a reasonable approach. In summary, we can only encourage the IPCC to involve more resource experts and natural scientists in future emission scenarios. The current set, SRES (2000), is biased toward exaggerated resource availability and unrealistic expectations on future production outputs.
6. Conclusions Perpetual economic growth is only an extrapolation from history, not a law of nature. Future depletion and the coming of a production peak are both phenomenological observations as well as results derived from physical models. However, perpetual growth is often held as a fundamental assumption for economists and used as a “rule of thumb” for extrapolation. Perpetual growth cannot be used as an underlying assumption for non-renewable energy sources, such as fossil fuels. Rotty (1979) stated that one should be able to make a more accurate analysis than simply projecting continued exponential growth in attempting to estimate the energy demands of the future. Even former technological and economic optimists are now seeing the end of an era with exponential growth (Ayres, 2006). This is hardly surprising, given the underlying arithmetic properties of growth and how quickly unreasonable values are reached for resource production and consumption even for modest growth rates (Bartlett, 1993, 1999, 2004). SRES is riddled with future production projections that would put unreasonable expectation on just a few countries or regions. Is it reasonable to expect that China, among the world’s largest coal reserve and resource holder and producer, would increase their production by a factor of 8 over the next 90 years, as implied by the A1C-scenarios? The concentration of the world’s coal reserves and resources to a few countries will result in that a few countries will play the key roles for future production. Many countries will have to rapidly increase their domestic production and consumption by absurdly large factors to fulfil the scenarios presented in SRES.
17
Resource-constrained modelling based on mathematical geology and historical time series indicate that most production trajectories in the scenarios are overoptimistic (Rutledge, 2007; Mohr and Evans, 2009; Patzek and Croft, 2010; Höök et al. 2010b). Only the B1 family appears to have any constraints on future coal production compatible with academic studies. Both Smil (2000) and Bezdek and Wendling (2002) point out that long range energy forecasters have made numerous inaccurate projections, mostly in the form of overestimations. Additionally, many inaccurate forecasts were done in good faith with state-of-the-art models, competent researchers and good funding, showing the difficulty of long-range energy forecasting. It is argued that many SRES scenarios need to be revised, generally downward, regarding production expectations from coal. Several scenarios also agree poorly with reality over the recent years and some can even be ruled out. SRES is underpinned by a paradigm of perpetual growth and technological optimism as well as old and outdated estimated regarding the availability of coal. As a result, future coal production projections in SRES (2000) are exaggerated and so are the resulting emissions. What kind of repercussions this has on the future climate is an open question which needs to be assessed from several different angles.
Acknowledgements I would like to thank Dr Michael Morbius for providing valuable inspiration. Shersti Johansson has my sincerest appreciation for proofreading and commenting.
References Ayres, R.U., 2006. Turning point: The end of exponential growth? Technological Forecasting and Social Change, 73(9):1188-1203 Bartlett, A.A., 1993. Arithmetic of growth: methods of calculation. Population and Environment, 14(4):359-387 Bartlett, A.A., 1999. Arithmetic of growth: methods of calculation II. Population and Environment, 20(3):215-246 Bartlett, A.A., 2004. The essential exponential! For the future of our planet. Center for Science, Mathematics and Computer Education, University of Nebraska-Lincoln, 302 p Bezdek, R.H., Wendling, R.M., 2002. A half century of long-range energy forecasts: errors made, lessons learned, and implications for forecasting. Journal of Fusion Energy, 212(34):155-172 BGR, 2009. Reserves, resources and availability of energy resources 2009 and previous reports. See also: http://www.bgr.bund.de/ BP, 2010. BP statistical review of world energy 2010. See also: http://www.bp.com Brecha, R.J., 2008. Emission scenarios in the face of fossil-fuel peaking. Energy Policy, 36(9):34923504 Gregory, K., Rogner, H.H., 1998. Energy resources and conversion technologies for the 21st century.
Mitigation and Adaptation Strategies for Global Change, 3 (2-4):171-229 Hjerpe, M, Linnér, B.O., 2008. Utopian and dystopian thought in climate change science and policy. Futures, 41(4), 234–245 Höök, M., Aleklett, K., 2010. A review on coal to liquid fuels and its coal consumption. International Journal of Energy Research, article in press Höök, M., Sivertsson, A., Aleklett, K., 2010a. Validity of the fossil fuel production outlooks in the IPCC Emission Scenarios. Natural Resources Research, 19(2), 63–81 Höök, M., Zittel, W., Schindler, J., Aleklett, K., 2010b. Global coal production outlooks based on a logistic model. Fuel, article in press IEA, 2009. Key world energy statistics 2009. See also: http://www.iea.org/ Ion, D.C., 1975. Availability of World Energy Resources. Graham & Trotman Ltd, London, 234 p Ion, D.C., 1979. World energy supplies. Proceedings of the Geologists' Association, 90(4), 193-202 IPCC, 1990. First Assessment Report, see also: http://www.ipcc.ch/ipccreports/assessmentsreports.htm IPCC, 1995. Second Assessment Report – Climate Change 1995, see also:
18
http://www.ipcc.ch/ipccreports/assessmentsreports.htm IPCC, 2001. Third Assessment Report - Climate Change 2001, see also: http://www.ipcc.ch/ipccreports/assessmentsreports.htm IPCC, 2007. Fourth Assessment Report - Climate Change 2007, see also: http://www.ipcc.ch/ipccreports/assessmentsreports.htm Jenkins G, 1989. Oil economists’ handbook. Taylor & Francis Group, London, 484 p Jones, R.N., 2001. An environmental risk assessment/management framework for climate change impact assessments. Natural Hazards, 23(2-3):197–230 Kharecha, P.A., Hansen, J.E., 2008. Implications of “peak oil” for atmospheric CO2 and climate. Global Biogeochemical Cycles, 22(3) Kontorovich, A.E., 2009. Estimate of global oil resource and the forecast for global oil production in the 21st century. Russian Geology and Geophysics, 50(4):237-242 Livernois, J.R., 1988. Estimates of marginal discovery costs for oil and gas. Canadian Journal of Economics, 21:379–393 Mitchell, B., 2003. International historical statistics 1750-2000. Palgrave MacMillan, London, 2580 p Mohr, S.H, Evans, G.M., 2009. Forecasting coal production until 2100. Fuel, 88(11):2059-2067 Nel, W.P., Cooper, C.J., 2009. Implications of fossil fuel constraints on economic growth and global warming. Energy Policy, 37(1), 166–180 Patzek, T.D., Croft, G.D., 2010. A global coal production forecast with multi-Hubbert cycle analysis. Energy, 35(8), 3109–3122 Rodríguez, X.A., Arias, C., 2008. The effects of resource depletion on coal mining productivity. Energy Economics, 30(2):397–408 Rogner, H.H., 1997. An assessment of world hydrocarbon resources. Annual Review of Energy and the Environment, 22:217-262 Rotty, R.M., 1979. Growth in global energy demand and contribution of alternative supply systems. Energy, 4(5), 881-890 Rutledge, D., 2007. Hubbert's Peak, the Coal Question, and Climate Change, presentation at
the ASPO-USA World Oil Conference, 17-20 October 2007, Houston, Texas, see also: http://rutledge.caltech.edu/ Schenk, N.J., Lensink, S.M., 2007. Communicating uncertainty in the IPCC’s greenhouse gas emissions scenarios. Climatic Change, 82(12):293-308 Schneider S, 2001. What is dangerous climate change? Nature, 411, 17-19 Schneider S, 2002. Can we estimate the likelihood of climatic changes at 2100? Climatic Change 52(4), 441–451 Sivertsson, A., 2004. Study of World Oil Resources with a comparison to the IPCC Emissions Scenarios, diploma thesis from Uppsala University 2004, see also: http://www.tsl.uu.se/uhdsg/Publications/Sivertss on_Thesis.pdf Smil, V., 2000. Perils of long-range energy forecasting: reflections on looking far ahead. Technological Forecasting and Social Change, 65(3):251-264. SRES, 2000. Special Report on Emissions Scenarios, report prepared by the Intergovernmental Panel on Climate Change (IPCC) for the Third Assessment Report. November 2000. See also: http://www.grida.no/publications/other/ipcc_sr/? src=/climate/ipcc/emission/ Thomas, L., 2002. Coal geology, Chichester: John Wiley and Sons, 384 p. Tilton, J.E., 2003. Assessing the threat of mineral depletion. Minerals and Energy, 18:33–42 Topp, V., Soames, L., Parham, D., Bloch, H., 2008. Productivity in the mining industry: measurement and interpretation. Productivity Commission Staff Working Paper, see also: http://www.pc.gov.au/__data/assets/pdf_file/000 5/84911/mining-productivity.pdf Valero, Al., Valero, A. 2010. A prediction of the exergy loss of the world’s mineral reserves in the 21st century. Energy, article in press Vallentin, D., 2009. Policy drivers and barriers for coal-to-liquids (CTL) technologies in the United States, Energy Policy 2008; 36(8):3198-3211 World Energy Council, 2009. Survey of energy resources interim update 2009 and previous reports. See also: http://www.worldenergy.org/
19
Manuscript Not AVAILABLE
Dr.Manfred Rumberger ER-Consult GmbH
45133 ESSEN , den 21.04.2010 Zeißbogen 63 Tel./Fax 0049 201 42885/423844 Email [email protected]
For Pittsburgh Coal Conference, Istanbul, Turkey, 11- 14 2010
The European Coal Market, a prosperous future?
In 2009 the world economy dropped significantly compared with the year before. The industry production, especially in the EU 27 broke down, which caused a bid loss of energy. Hard coal and lignite in the EU 27 have a share of 17 % the primary energy consumption, 90% of the lignite and 67 % of the hard coal are used for power generation, the remaining shares of both for heating purposes and for the iron and steel industry. 2009 all industrial and private demand of energy broke down and mainly the hard coal was hit and characterised by a reduction of domestic production and slightly rising imports. More or less all other kinds of energy were consumed on the level of the year before. After a recovery of the economy the hard coal and to a certain extend the lignite will furthermore loose importance in supplying the energy market, the domestic production of hard coal will drop significantly during the next years, the lignite will be faced with irreconcilable environmental problems, if the CO2- problem is not solved. The political attempt, initiated by the commission of the EU EU wide or by the individual member countriesm, to favour and support renewables and other clean forms of energy up to 30 % of the total primary energy supply during the next 10 – 15 years, reducing the CO2 emission of existing plants by installing very expensive new CO 2 capture and storage ( CCS) system technology, the CO2 emission trade Europe wide that starts in 2013,and developing new generations of power- and iron and steel plant equipment. Finally it is panned, to extend the lifetime of nuclear power plants, respectively to establish new reactors, as it is consideration in some member countries. All these measures will put pressure on the coal market, because coal and lignite are the biggest sources and environment and emission. The extend of this pressure will be very high, but depends on future experience with the political targets, the investment capital available and the special privileges of individual plants or projects.
1 Methane Enrichment from Anaerobic Digestion Gas(ADG) Using Polymeric Hollow Fiber Membrane Hyung-Keun Leea,*, Dae-Hoon Kima, Ki-Hong Kima, Young-Mo Ana, Hang-Dae Joa, Ki-Jun Baikb and Gang-Woo Leec a
Korea Institute of Energy Research, 71-2 Jangdong, Yuseong, Daejeon, 305-343, KOREA b
Yanbian University of Science and Technology, China c
Yoo Sung Co. R&D Center, KOREA
*
Phone: +82-42-860-3647, e-mail: [email protected]
ABSTRACT The hollow fiber membrane(HFM) was prepared by the dry-wet phase inversion method using polyethersulfone for recovery of methane from ADG application. The produced fiber was characterized by a scanning electronic microscope(SEM) and single gas permeation measurements using methane and carbon dioxide. The HFM module was manufactured in order to examine mixture gas separation behavior. Separation experiments of ternary mixtures(H2S/CO2/CH4) were carried out to investigate removal of hydrogen sulfide, and carbon dioxide and methane enrichment according to retentate flow rate and operating pressure difference. To improve separation efficiency, simulation of multi-stages separation process with recycle was conducted by numerical analysis using a MATLAB. As a result, recovery ratio of methane increased from 55 % to 90 %. 1. INTRODUCTION The production of energy from renewable sources is one of several means of reducing the emissions of greenhouse gas(GHG). Recently, the research to enrichment of methane from the biogas is progressing actively. Biogas purified and enriched in methane can be used for household applications, automobile fuel or electricity generation [1-3]. Biogas is about 55~65% methane with the remainder being mostly carbon dioxide and trace of hydrogen sulfide [1]. Gas separation using polymeric membranes is proven technology that has found wide range of applications. Flexible design, the compact space and the efficiency of membrane units compared to conventional gas separation method such as cryogenic distillation or absorption are particularly attractive [3]. In this study, hydrogen sulfide and carbon dioxide were removed and methane was enriched in order to use the biogas as a fuel of fuel cell using the membrane separation process. The hollow fiber membrane(HFM) was prepared by the dry-wet phase inversion method using polyethersulfone(PESf) and morphology was characterized by a SEM. Single gas permeation measurements were carried out using pure carbon dioxide and methane. The membrane module was manufactured in order to examine mixture gas separation behavior. Separation experiments of ternary mixtures(H2S/CO2/CH4) were carried out according to
2 retentate flow rate and operating pressure. To improve separation efficiency, simulation of multi-stages separation process with recycle was conducted based on separation behavior of a single module by numerical analysis using a MATLAB. 2. EXPERIMENTAL 2-1. Preparation of polymeric hollow fiber membranes Hollow fiber membranes were produced by a dry-wet phase inversion method, a versatile and simple technique used for the preparation of polymeric membranes [4,5]. The dope solutions contain polymer, solvent and non-solvent. N-methylpyrrolidone (NMP) as a solvent and distilled water as a non-solvent were used, respectively. The structure of produced fibers was characterized by SEM and the modules were manufactured for separation of ternary mixtures. 2-2. Gas permeation and separation The single gas permeation behavior was observed with methane of 99.999% and carbon dioxide of 99.98% according to pressures difference. For separation of ternary mixtures, 35 vol.% carbon dioxide and 1,000 ppm hydrogen sulfide balanced with methane were used. The effect of operating conditions such as pressure, and retentate flow rate was observed. Especially, pressure and retentate stream were controlled by retentate control using back pressure regulator. 2-3. Simulation of multi-stages separation by a numerical analysis Simulation program has been developed for gas separation module with counter-current flow. Pressure drop was estimates by Hagen-Poiseuille equation and internal flows of module assumed pseudo fully developed state [6]. Mass transfer equation of the retentate and permeate stream was expressed as Eq. (1)(4). dL dz dV dz
= −πDO
n i=1 Ji
(1)
= −πDO
n i=1 Ji
(2)
dx
L dzi = −πDO (Ji − x V
dy i dz
= −πDO (Ji − y
n j=1 Jj ), i
= 1,2, … , n − 1
(3)
n j=1 Jj ), i
= 1,2, … , n − 1
(4)
Where, L and V are flow rate of retentate and permeate stream. xi and yi are defined as component of i in the retentate and permeate stream. DO is membrane fiber diameters (m) of outside and Ji is flux of each component. Boundary condition of differential equation was defined as the following equations. xi = xi,f , P = Pf
at
z=0
L = L0 , V = 0, yi = yi ∗ at z=lm Where, L0 is retentate flow rate (m3/s) and lm (m) is length of membrane.
(5) (6)
3 Separation behaviors of a single module were estimated from Eq. (1)-(4) by using a MATLAB. The designed membrane process treats the gas flow of 250 LPM and concentration of hydrogen sulfide is below than 5ppm in recovered gas. 3. RESULTS and DISCUSSION 3-1. Preparation of polymeric HFM In this process a spun fiber first passes through an air gap where a controlled evaporation occurs. Due to the evaporation locally a high polymer concentration is created near the outer surface, which is necessary to form a dense membrane top layer. After the air gap the nascent fiber enters a coagulation bath where phase inversion occurs and the membrane structure is formed [7]. And water as a non-solvent additive should exhibit the best miscibility with the coagulant because water and NMP can be form a complex through multiple hydrogen bonding [8,9]. The cross-sectional structures of the hollow fibers spun by the dry-wet phase inversion method were shown in Fig. 1. The SEM photos showed that the produced fiber typically has an asymmetric structure; a dense top layer supported by sponge and finger-like substructure.
Fig. 1. SEM photos of PESf HFM. 3-2. Single gas permeation The single gas permeation behavior was observed with methane of 99.999% and CO2 of 99.98% according to pressures difference of 4-10 kgf/cm2 at 30 ℃. The respective gas permeances of carbon dioxide and methane were increased from 71-89 and 4.7-5.4 GPU, respectively, with pressure difference. As a result, CO2/CH4 perm-selectivity was increased from 15.2-16.3. The permeability of gas is determined by the solution and diffusion of gas in the membrane [7,10,11]. The diffusion coefficients of penetrants often change less than the solubility coefficients with changing pressure difference so that more soluble penetrants are more permeable. Thus, with increasing pressure difference, permeance carbon dioxide increases more significantly. Also, because the increase of carbon dioxide permeance is much more than that of methane permeance [12], due to the higher solubility of carbon dioxide in the membrane, the CO2/CH4 perm-selectivity increases with pressure difference. 3-3. Separation of ternary mixtures Separation of ternary mixtures using a single module was tested according to pressure difference of 5 and 7 kgf/cm2 and retentate flow rate of 0.4-2 LPM at 30 ℃. Fig. 2 shows
4 specifications of used a module and the separation behavior using a PESf HFM single module, respectively. Fig. 2(a) and (b) show separation behavior of ternary gas with various operating condition. Methane purity and removal efficiency of carbon dioxide and hydrogen sulfide in recovered gas were increased with pressure and stage cut while recovery ratio of methane was decreased. These phenomena can be explained by free volume theory. The concentration of gas inside the membrane increases at the higher pressure difference, and stage cut, and this causes the free volume available within the membrane to increase. This effect leads to an increase of the permeated flux of penetrants because the increasing stage cut at constant pressure difference and temperature increase the driving force. By using a PESf HFM single module, a maximum methane purity of 95 vol.% was obtained. However, recovered gas contains a significant amount of hydrogen sulfide gas with low recovery ratio of methane. The multi-stages separation process with recycle was demeaned for removal of H2S and recovery of methane.
80 60 70 40 60 20
5 kgf/cm
2
50
7 kgf/cm2 0.4
0.6
0.8
1.0
1.2
1.4
1.6
Retentate flow rate [LPM]
1.8
2.0
40 2.2
700 5 kgf/cm2 7 kgf/cm
2
600
60
500 400
40 300 200
20
100 0 0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
H2S purity in recovered gas [ppm]
90
80
0 0.2
80
CO2 purity in recovered gas [vol.%]
100
Recovery ratio of CH4 [%]
CH4 purity in recovered gas [vol.%]
100
0 2.2
Retentate flow rate [LPM]
(a) (b) Fig. 2. Result in separation of ternary mixtures using a PESf HFM module. 3-4. Simulation of multi-stages separation by a numerical analysis Computer simulation is possible to predict the performance of membrane process. In this study, the recovery ratio of methane and membrane area in the multi-stages process was estimated by simulation program using MATLAB. Fig. 3 shows the result of multi-stages simulation at 5 kgf/cm2. At this pressure, recovery ratio of methane is 90 % and 588 m2 membrane needs to get the 5ppm hydrogen sulfide in retentate stream. Fig. 4 shows membrane area and recovery ratio of methane with pressure difference. In the result of the simulation, both of the membrane area and recovery ratio were decreased according to pressure difference increase. 4. CONCLUSION PESf HFM was spun by dry-wet phase inversion method for recovery of methane from ADG application. The SEM photos showed that the produced fiber typically has an asymmetric structure; a dense top layer supported by porous, sponge- and finger-like, substructure. Gas permeation behavior of a PESf HFM module was characterized by carbon dioxide and methane according to pressure. The permeance and selectivity were increased linearly with pressure up to carbon dioxide permeance of 89 GPU and selectivity of 16.3 at 10 kgf/cm2. For separation of ternary
5 mixtures, the effect of various operating conditions such as pressure, and retentate flow rate were tested by using a single module.
Lf P2 P2+R3 1728.2 1006 38.84 61.06
Lf' 1978.2 1000 38.39 61.51
Recovery Ratio(%) Membrane Area(m2)
362.45
100
R1 671.2 210 13.05 86.93
M1
Lf’ R3
90.18 588.9
600
M2
R1
2
m
m
M3 20.89
P3
P2 524.0 267.0 16.44 83.54
R3 2
m
R3 1319.0 49.3 50.53 58.16
R2 R2 147.2 5.0 0.88 99.12
P2
P1 1294.9 1278 43.73 56.14 P3 102.8 2415 74.91 24.85
2
205.56
P1
95
500 90
80
.
400
85
4
5
6
7
8
Membrane area [m2]
Unit Lf (L/min) 250.0 ppm 1000 % 35.20 % 64.70
Recovery ratio of CH4 [%]
Item Q H2S CO2 CH4
300
Pressure difference [kgf/cm2]
Fig 3. Simulation result at 5 kgf/cm2.
Fig 4. Membrane area and recovery ratio of methane follow the pressure.
CH4 purity and removal efficiency of carbon dioxide and hydrogen sulfide in recovered gas were increased with pressure and stage cut while recovery ratio of methane was decreased because the increasing stage cut at constant pressure and pressure increase the driving force. The experimental results showed methane purity of 95 vol.% at pressure difference 7 kgf/cm2, retentate flow rate of 0.44 LPM and temperature of 30 ℃ while recovery ratio of CH4 was 46% and hydrogen sulfide remained 179 ppm in recovered gas. The multi-stages separation with recycle showed good results for removal of hydrogen sulfide and recovery of methane. Simulation results of the membrane area and recovery ratio were decreased with increasing pressure.
Reference 1. N. Abatzoglou, S. Boivin, Biofuels, Bioprod. Bioref, 3 (2009) 42. 2. Brian C. H. Steele, and A. Heinzel, Nature, 414 (2001) 345. 3. P. Bernardo, E. Drioli, and G. Golemme, Ind. Eng. Chem. Res., 48 (2009) 4638. 4. H. H. Park, B. R. Deshwal, H. D. Jo, W. K. Choi, I. W. Kim, H. K. Lee, Desalination, 243 (2009) 52. 5. H. H. Park, B. R. Deshwal, I. W. Kim, H. K. Lee, J. Membr. Sci., 319 (2008) 29. 6. H. M. S. Lababidi, Trans. IChemE., 78 (2000) 1066. 7. T. Mohammadi, M. T. Moghadam, M. Saeidi, M. Mahdyarfar, Ind. Eng. Chem. Res., 47 (2008) 7361. 8. D. Wang, K. Li, W. K. Teo, J. Membr. Sci., 176 (2000) 147. 9. D. Wang, K. Li, W. K. Teo, J. Membr. Sci., 115 (1996) 85. 10. G. Maier, Angew. Chem. int. Ed., 37 (1998) 2960. 11. M. Mulder, “Basic Principles of membrane technology”, Kluwer academic publishers (1991). 12. D. W. Wallace, Ph.D. Dissertation, Univ. of Texas at Austin (2004).
Effect of Introduction of Clean Coal Technology on Future Asian Energy Supply Sohei Shimada *) and Yuta Koyama, Graduate School of Frontier Sciences The University of Tokyo, JAPAN *) Corresponding author: [email protected] Abstract Dynamic New Earth 21 (DNE21), a long-term energy supply calculation model, was used to estimate the future energy supply up to 2050 when the CCT was introduced in Asian countries under GHG emission regulation. Three scenarios, BAU, Emission Regulation with CCT employment and NO CCT employment, were set. The statements of world main countries’ GHG reduction target were used for the CO2 emission regulation. The results showed that the introduction of CCT combined with CCS is a key technology to establish the stable energy supply for China and India under GHG emission regulation. Keywords: clean coal technology (CCT), Asia, energy supply, CO2 emission, carbon capture and sequestration (CCS)
1. Background and Objectives Coal resources are abundant and widely distributed in the world. Therefore, coal is regarded to be the indispensable primary energy resource for the long-term stable energy supply, especially in the developing countries. What is more, coal price is cheap compared to other energy resources. However, coal has some disadvantages, such as large greenhouse gas and air pollutant material emissions to the atmosphere exhausted in coal combustion process, which cause the global warming and the acid rain. In order to solve these problems, R&D on clean coal technologies (CCT) has been conducted in many countries and their developed technologies are employed in the actual coal plants. The employment of the CCT is especially important for the Asian countries from the viewpoints of an environmental measures and the future energy supply because of following three reasons. 1) Asia depends heavily on coal for energy supply. (Fig. 1.1.1) 2) Over 50 percent of all coal supply in 2008 is consumed in Asia. (Fig. 1.1.2) 3) According to some research institutes, it is predicted that energy demand in Asia will expand drastically in the near future by economic growth. However, CCT is not so commonly used in Asia, now. This study aims to evaluate the effect of the employment of CCT on the future energy supply and CO2 emission reductions in Asia. 100% 90% 80% 70% N_Gas Oil Hydro Nuclear Coal
60% 50% 40% 30% 20% 10% 0% North America
S. & Cent. America
Europe & Eurasia
Middle East
Africa
Asia Pacific
Fig. 1.1.1 Regional Primary Energy Consumption 2008 (Source: BP Statistical Review of World Energy 2009) 1
North America 18%
Asia Pacific 61%
S. & Cent. America 1% World Tota l 3303.7[MTOE]
Europe & Eurasia 16%
Middle East & Africa 4%
Fig. 1.1.2 Regional Coal Consumption 2008 (Source: BP Statistical Review of World Energy 2009)
2. Simulation Method 2.1 Simulation Model In this study, Dynamic New Earth 21 (DNE21) [1] developed by Prof. Y. Fujii of University of Tokyo and RITE (Research Institute of Innovative Technologies) is used as a simulation model. DNE21 is usable to calculate what kind of measures against global worming will be taken in the future as a most cost efficient manner from global and long-term point of view. DNE21 is cost optimization type energy system model integrating the macro economic model and climate change model. (Fig. 2.1.1)
Fig. 2.1.1 DNE21 Model Input and Output 2.2 Condition Setting of Simulation In this study following three scenarios were set. The energy supply and the amount of the CO2 emission by region at the time of from 2010 to 2050 were simulated. 1) Reference Scenario (Ref) ・NO CO2 Emission Regulation 2) Emission Reduction Strategy Scenario (ERS) 2
・CO2 Emission Regulation and Emission Rights Trade 3) NO CCT Scenario (NO CCT) ・Same Condition as ERS scenario but without Clean Coal Technologies The condition of practical use of CCT in Asia was set as follows. 1) Conventional PCC (Pulverized Coal Combustion) : Gross thermal efficiency 41–43% (2010-2050) Construction cost : 1.050 [$/kW] 2) Advanced USC (Ultra Super-Critical): Gross thermal efficiency : 43-48% (2010-2050) Construction cost : 1.100 [$/kW] 3) IGCC (Integrated Coal Gasification Combined Cycle): Gross thermal efficiency 42-52% (2020-2050) Construction cost : 1.300 [$/kW] As common conditions, maximum of installable capacities of global nuclear power generation was assumed to be 830[GW] in 2030, and 1,030 [GW] in 2050. The maximum annual world hydro and geothermal power generation was assumed to be 14,000 [TWh/year] with reference to Survey of Energy Resources 1998. The maximum annual world wind power generation was assumed to be 12, 000 [TWh/year] with reference to NCDC (National Climatic Data Center) and land use mesh data of Chiba University. The maximum annual world photovoltaic generation was assumed to be 1,270,000 [TWh] with reference to the data of the amount of sunshine from NASA SeaWIFS and land use mesh data of Chiba University. Country- specific CO2 reductions targets were set as follows in reference to CO2 reductions goal, that is stated by each countries before COP15 in Copenhagen. 1) United State of America Reducing CO2 emissions to 17% below 2005 level by 2020, and to 50% below 1990 by 2050. 2) Europe Union Reducing CO2 emissions to 20% below 1990 level by 2020, and to 50% below 1990 by 2050. 3) Japan Reducing CO2 emissions to 25% below 1990 level by 2020, and to 50% below 1990 by 2050. 4) China Reducing CO2 emissions per GDP to 40% - 45% below 2005 level by 2020, and reducing CO2 emissions to same level as 1990 by 2050. 5) India Reducing CO2 emissions per GDP to 20% - 25% below 2005 level by 2020, and reducing CO2 emissions to same level as 1990 by 2050. 6) Australia Reducing CO2 emissions to 15% below 2000 level by 2020, and to 50% below 1990 by 2050. 7) Russia Reducing CO2 emissions to 20% below 1990 level by 2020, and to 50% below 1990 by 2050. 8) South Korea Reducing CO2 emissions to 4% below 2005 level by 2020, and to same level as 1990 by 2050. 9) Developing Countries No regulation by 2020, and reducing CO2 emissions to same level as 1990 by 2050. For the basic data collection Refs [2, 3, 4, 5] were used.
3. Simulation Results 3.1 Asia 3
Fig. 3.1.1 shows the simulation result of the amount of CO2 emissions and reductions of Asia in the case of ERS. Main CO2 reductions measures from 2010 to 2020 are “Conservation”, and “Conversion”, which is conversion of fossil energy resource from coal to clean fossil fuel such as oil or natural gas. From 2020 to 2050, “Fuel switch”, which is conversion from fossil resources to non-fossil resources and CCS (CO2 capture and sequestration), becomes main CO2 reductions measures as well as “Conservation” and “Conversion”. The data in 1990-2005 in the Figure were cited from the BP Statistical Review of World Energy 2009. Fig. 3.1.2 shows the coal consumption for each scenario. In the Ref case, the coal consumption in Asia continues to increase. In the ERS case, the coal consumption in Asia decreases in 2010-2030 and increases again after 2030. This result is explained as follows. In 2010-2030 coal consumption is restricted due to its high CO2 emission intensity and promote the “Conversion” and “Fuel Switch”. In 2020-2050 coal consumption is again increases due to the employment of CCT with low CO2 emission intensity. It suggests that in order to realize the further continuous coal consumption in Asia, introduction of CCT is inevitable. Amount of CO2 Emissions and Reductions [Gton-CO2]
35.00
Emission Rights Reforestation CCS Ocean CCS Land ECBMR EOR Fuel Switch Conversion Conservation Net Emissions
30.00 25.00 20.00 15.00 10.00 5.00 0.00 1990
1995
2000
2005
2010
2020
2030
2040
2050
Year
Fig. 3.1.1 Amount of CO2 Emissions and Reductions in Asia (ERS) (*1990-2005: BP Statistical Review of World Energy 2009, *2010-2050: Simulation Results)
Coal Consumption [MTOE]
5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 1990
2000
2010
2020
2030
2040
2050
Year ERS
CCT
NO CCT
Fig. 3.1.2 Comparison of Coal Consumption in Asia Fig. 3.1.3 shows the primary energy supply balance in Asia in the ERS case. The renewable energy supply increases after 2030 and it shares about 25 % of total energy supply. As for the fossil fuels, its supply decreases a little bit in 2020-2030, but it still shares as much as 66-87%. Fig. 3.1.4 shows the electricity supply balance in Asia in the ERS case. Power generation by renewable increases and it shares about 32% in 2050. Power generation by fossil fuels decreases in 2020-2030 and it increases after that. The power 4
generation by coal increases dramatically from 4,915 [TWh] in 2030 to 13,980 [TWh] in 2050. Considering the CO2 constraint in this period, CCS plays a great role in the introduction of coal firing power generation in Asian future energy supply. 9000
Other Renewable Nuclear Oil Natural Gas Coal
Energy Supply [MTOE]
8000 7000 6000 5000 4000 3000 2000 1000 0 2010
2020
2030
2040
2050
Year Fig. 3.1.3 Energy Supply Balance of Asia (ERS)
Electricity Supply [TWh]
30000
Other Renewable Nuclear Oil Natural Gas IGCC USC Fired Coal Fired
25000 20000 15000 10000 5000 0 2010
2020
2030
2040
2050
Year
Fig. 3.1.4 Electricity Balance of Asia (ERS) Fig. 3.1.5 shows the CO2 Shadow Price for each scenario. The CO2 Shadow Price is in 2050 is 97.3 [$/t-CO2] for ERS case and 105.0 [$/t-CO2] in NO CCT case. This implies that the introduction of the CCT is economical as a CO2 emission reduction strategy. Shadow Price of CO2 [$/t-CO2]
120 105.0 100 84.0
97.3
80 79.3 55.1
60
53.9
40 20 0.0
0.0
0.0
0 2000
2010
2020
2030
2040
2050
Year ERS
NO CCT
Fig. 3.1.5 Comparison of CO2 Shadow Price 5
3.2 China Fig. 3.2.1 shows the simulation result of the amount of CO2 emissions and reductions of China in the case of ERS. Main CO2 reductions measures from 2010 to 2020 are “Conservation”, and “Conversion”. From 2020 to 2050, “Fuel switch” (CO2 capture and sequestration) becomes a main CO2 reductions measures as well as “Conservation” and “Conversion”. It suggest s that the solar cell is gradually introduced in China from 2020. Fig. 3.2.2 shows the simulation result of coal consumption of China for each scenario. In the case of Ref, coal consumption of China increases drastically compared to the other scenarios. In the case of ERS, coal consumption of China decreases from 2010 to 2030 compared to Ref, but increases from 2030 to 2050 because CCS comes into practical use as a CO2 reduction measure. In the case of NO CCT, coal consumption of China decreases after peaking in 2020. These results show it is important for China to introduce CCT, when China uses coal energy efficiently in the future. 16.00
Emission Rights Reforestation CCS Ocean CCS Land ECBMR EOR Fuel Switch Conversion Conservation Net Emissions
Amount of CO2 Emissions and Reductions [Gton-CO2]
14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 1990
1995
2000
2005
2010
2020
2030
2040
2050
Year
Fig. 3.2.1 Amount of CO2 Emissions and Reductions in China
Coal Consumption [MTOE]
2500 2000 1500 1000 500 0 1990
2000
2010
2020
2030
2040
2050
Year ERS
CCT
NO CCT
Fig. 3.2.2 Comparison of Coal Consumption in China Fig. 3.2 3 shows the energy supply pattern of China in the case of ERS. When China provides energy in most efficient manner under the CO2 regulation, the total percentages with “Nuclear”, “Renewable”, and “Other (Biomass etc.” increased from 11.8% of all energy supply in 2010 to 31.9% in 2030, and 31.3% in 2050. On the other hand, the total percentages with “Coal”, ”Oil”, and ”Natural Gas” is decreasing trend after 2010, but fossil energy use shares up to about 70%.
6
Other Renewable Nuclear Oil Natural Gas Coal
4000
Energy Supply [MTOE]
3500 3000 2500 2000 1500 1000 500 0 2010
2020
2030
2040
2050
Year
Fig. 3.2.3 Energy Supply Balance of China (ERS) Fig. 3.2.4 shows electricity supply pattern of China in the case of ERS. Electricity demand of China increases from 4,259 [TWh] in 2010 to 11,483 [TWh] in 2050 at 2.4% per annum. When China provides electricity in most efficient manner under the CO2 regulation, the total percentage with “Nuclear”, “Renewable”, and “Other (Biomass etc. ) ” increases from 18.3% of all energy supply in 2010 to 46.5% in 2030. On the other hand, the total percentage of the sum of “Coal”, ”Oil”, and ”Natural Gas” is in decreasing trend after 2010, but electricity from fossil resources made up from 53.5% to 81.5% in any result. Especially, electricity supply of China depends much on coal fired power generation. Other Renewable Nuclear Oil Natural Gas IGCC USC Fired Coal
Electricity Supply [TWh]
14000 12000 10000 8000 6000 4000 2000 0 2010
2020
2030
2040
2050
Year
Fig. 3.2.4 Electricity Balance of China (ERS) 3.3 India Fig. 3.3.1 shows the simulation result of the amount of CO2 emissions and reductions of India in the case of ERS. Main CO2 reductions measures from 2010 to 2020 are “Conservation”, and from 2020 to 2050, “Conversion” and CCS. “Fuel switch” is not introduced in this period, while “Fuel Switch” was introduced in China. Main “Fuel Switch” technology is solar cell. The reason the “Fuel Switch” is not introduced in India is the lower economy level compared to China. Fig. 3.3.2 shows simulation result of coal consumption of India in each scenario. In the case of Ref, coal consumption of India increases drastically compared to the other scenarios. In the case of ERS, coal consumption of India increases from 2010 to 2020, but decreases from 2020 to 2030 because “Conversion” is taken as a cost effective CO2 emissions reductions measures. However, after 2030, coal consumption of India increased again because CCS came into practical use as a cost effective CO2 reduction measure. In the case of NO CCT, coal consumption of China decreased after peaking in 2010. These results show it is important for India to introduce CCT, when India provides energy efficiently by using coal in the future. 7
10.00
Emission Rights Reforestation CCS Ocean CCS Land ECBMR EOR Fuel Switch Conversion Conservation Net Emissions
Amount of CO2 Emissions and Reductions [Gton-CO2]
9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 1990
1995
2000
2005
2010
2020
2030
2040
2050
Year
Fig. 3.3.1 Amount of CO2 Emissions and Reductions in India (ERS)
Coal Consumption [MTOE]
1800 1600 1400 1200 1000 800 600 400 200 0 1990
2000
2010
2020
2030
2040
2050
Year ERS
CCT
NO CCT
Fig. 3.3.2 Comparison of Coal Consumption in India Fig. 3.3.3 shows energy supply pattern of India in the case of ERS. Under the CO2 emission regulation, the total percentage with “Nuclear”, “Renewable”, and “Other (Biomass etc.” increases from 9.7% of all energy supply in 2010 to 26.2% in 2030, and 19.0% in 2050. On the other hand, the total percentage with “Coal”, ”Oil”, and ”Natural Gas” is in decreasing trend after 2010, but fossil energy use shares up to about 75%. Renewable Other Nuclear Oil Natural Gas Coal
Energy Supply [MTOE]
2500 2000 1500 1000 500 0 2010
2020
2030
2040
2050
Year
Fig. 3.3.3 Energy Supply Balance of India (ERS)
8
Fig. 3.3..4 shows electricity supply pattern of India in the case of ERS. Electricity demand of India increased from 1325.1[TWh] in 2010 to 7602.2[TWh] in 2050 at 4.4% per annum. Under the CO2 regulation, the total percentage with “Nuclear”, “Renewable”, and “Other (Biomass etc.) increases from 13.9% of all energy supply in 2010 to 36.8% in 2030. On the other hand, the total percentage with “Coal”, ”Oil”, and ”Natural Gas” is in decreasing trend after 2010, but electricity from fossil resources made up from 63.2% to 86.1% in any result. Especially, electricity supply of India depends much on coal fired power generation for electricity supply. Other Renewable Nuclear Oil Natural Gas IGCC USC Fired Coal Fired
8000
Electricity Supply [TWh]
7000 6000 5000 4000 3000 2000 1000 0 2010
2020
2030
2040
2050
Year
Fig. 3.3.4 Electricity Balance of India (ERS) 3.4 Other Developing Asia Fig. 3.4 1 shows the simulation result of the amount of CO2 emissions and reductions of Developing Asia except China and India in the case of ERS. Main CO2 reductions measures from 2010 to 2020 are “Conservation”, and “Conversion”. From 2020 to 2050, “Fuel switch” and CCS become the main CO2 reductions measures as well as “Conservation” and “Conversion”.
Amount of CO2 Emissions and Reductions [Gton-CO2]
7.00
Emission Rights Reforestation CCS Ocean CCS Land ECBMR EOR Fuel Switch Conversion Conservation Net Emissions
6.00 5.00 4.00 3.00 2.00 1.00 0.00 1990
1995
2000
2005
2010
2020
2030
2040
2050
Year
Fig. 3.4.1 Amount of CO2 Emissions and Reductions in Developing Asia except China and India (ERS) Fig. 3.4.2 shows simulation result of coal consumption of Developing Asia except China and India in each scenario. In the case of Ref, coal consumption of Developing Asia except China and India increases drastically compared to the other scenarios. In the case of ERS, coal consumption of Developing Asia except China and India decreased from 2010 to 2040, but increases a little bit from 2040 to 2050 because CCS comes into practical use as a CO2 reduction measure. In the case of NO CCT, coal consumption of Developing Asia except China and India decreased after peaking in 2010. These results show that the introduction of CCT is not so important for Developing Asia except China and India.
9
2000
Other
Enegy Supply [MTOE]
1800
Renewable
1600
Oil
1400
Natural Gas
1200
Coal
1000 800 600 400 200 0 2010
2020
2030
2040
2050
Year
Fig. 3.4.2 Comparison of Coal Consumption in Developing Asia except China and India Fig. 3.4.3 shows energy supply pattern of Developing Asia except China and India in the case of ERS. The total percentages with “Renewable”, and “Other (Biomass etc.) increased from 12.3% of all energy supply in 2010 to 38.9% in 2030, and over 50% since 2040. On the other hand, the total percentages with “Coal”, ”Oil”, and ”Natural Gas” was on decreasing trend since 2010, and fossil energy use made up fewer than 50% after 2040.
Coal Consumption [MTOE]
500.0 450.0 400.0 350.0 300.0 250.0 200.0 150.0 100.0 50.0 0.0 1990
2000
2010
2020
2030
2040
2050
Year ERS
CCT
NO CCT
Fig. 3.4.3 Energy Supply Balance of Developing Asia except China and India (ERS) Fig .3.4.4 shows electricity supply pattern of Developing Asia except China and India in the case of ERS. Electricity demand of Developing Asia except China and India increases from 1,616.6 [TWh] in 2010 to 5,687.6 [TWh] in 2050 at 3.2% per annum. Under the CO2 regulation, the total percentages with “Renewable”, and “Other (Biomass etc.)” increased from 17.9% of all energy supply in 2010 to 58.4% in 2030. On the other hand, the total percentage with “Coal”, ”Oil”, and ”Natural Gas” is in decreasing trend since 2010. This result shows Developing Asia except China and India will not depends not so much on fossil fuel for electricity supply compared to China and India.
4. Conclusions DNE21 was used for the estimation of future energy supply in Asia by the introduction of CCT under GHG emission regulation. The results of the study are summarized as follows. 1) In the case of ERS (Emission Regulation Scenario), coal consumption in Asia decreases between 2020 and 2030 because of CO2 emission regulation, but increases again after 2030, when the CCS comes into practical use. From this 10
result, it is concluded that the CCS is needed for long-term coal use. If the CCS will come into practical use, it is expected that coal will play a key role in electricity production in Asia, since over 40% of all future electricity production of Asia comes from coal fired power generation after 2030. 2) When the most cost effective energy supply will be done under CO2 emission regulation in Asia, it is expected that about 25% of the future Asia’s primary energy supply will come from renewable and biomass energy sources, but Asia will depend heavily on fossil fuel energy over 60% of all future primary energy supply in the future.
Electricity Supply [TWh]
6000
Other Renewable Nuclear Oil Natural Gas IGCC USC Fired Coal Fired
5000 4000 3000 2000 1000 0 2010
2020
2030
2040
2050
Year
Fig 3.3.4 Electricity Balance of Developing Asia except China and India (ERS) References [1] Fujii, Y. and Yamaji, K., Assessment of technological option in the global energy system for limiting the atmospheric CO2 concentration, Environmental Economics and Policy Studies, 1998 [2] BP, Statistical Review of World Energy, 2009 [3] International Energy Agency Green House Gas R&D Program, Enhanced Coal Bed Methane Recovery with CO2 Sequestration, PH3/3, 1988 [4] International Energy Agency, World Energy Outlook 2007 - China and India Insights-, 2007 [5] International Energy Agency, CO2 Capture and Storage - A Key Carbon Abatement Option-, 2008
11
INVESTIGATION OF CARBONIZATION KINETIC OF TUNÇBİLEK LIGNITE USED FOR THE PREPARATION OF ACTIVATED CARBON Burcu Özdemir1, Nilgün Karatepe1, Reha Yavuz2 1 2
Istanbul Technical University, Energy Institute, 34 469 Maslak, Istanbul-Turkey
Istanbul Technical University, Department of Chemical Engineering, 34 469 Maslak, Istanbul-Turkey Email: [email protected], [email protected], [email protected], Phone: + 90 212 285 39 40, Fax: + 90 212 285 38 84
Abstract: The main objective of this study was to investigate the carbonization of Tunçbilek lignite by TG. Thermogravimetric analyses were performed at different conditions such as heating rates of 10 and 30 °C/min, temperatures of 500, 650 and 800 °C, particle sizes of 1700-700 and 150-100 m, heating mediums of nitrogen and carbon dioxide and gas flow rates of 15 and 40 cm3/min. Based on TG curves, the kinetic parameters of the process were calculated using the methods of Coats–Redfern, HorowitzMetzger and Dharwadkar-Karkhanavala in order to illuminate the mechanism of carbonization of Tunçbilek lignite. The carbonization kinetic models of Tunçbilek lignite were set up. The main carbonization stage could be described by the second order global models for Tunçbilek lignite and activation energies varied in the range of 31.7-69.0 kJ/mol. Keywords: Carbonization, kinetics, thermogravimetry, lignite, activated carbon INTRODUCTION Activated carbons are highly porous form of solid carbons produced from a variety of raw materials. Among them, coal is the most commonly used precursor due to its low cost and large supply and also active carbon prepared from coal is superior to those derived from lignocellulosic materials in terms of mechanical properties. Basically, there are two different processes for the preparation of active carbon: physical and chemical activation [1]. Physical activation involves carbonization of a carbonaceous material followed by activation of the resulting char at a temperature between 800 and 1100 °C in the presence of suitable oxidizing gases such as carbon dioxide or steam. Chemical activation is a single step method. In this activation the precursor is mixed with a chemical agent (KOH, ZnCl2, H3PO4 vb.) and then pyrolysed between 400 and 600 °C in the absence of air [1, 2]. Therefore the preparation of activated carbon with high surface area from different materials is an interesting subject. It has been indicated that the carbonization process has great effects on the final product with both physical activation and chemical activation, and careful selection of carbonization parameters is important in order to prepare the required quality of activated carbons. The purpose of carbonization process is to enrich the carbon content and to create an initial porosity in the char. There are several critical parameters in the preparation of activated carbon that would affect its structure, one of which is carbonization temperature. It has been shown that high carbonization temperature would result in a great amount of volatiles being released from the raw material and eventually influences the product yield and porosity [3, 4]. So it is of great importance to study the carbonization behavior of the carbon materials or chemicals impregnated carbon materials in order to produce activated carbons with high quality. However, despite the significant importance of physicochemical processes associated with
the thermal degradation of the carbon materials especially coal, their mechanisms are not understood in many important features up to now [5-9]. Thermogravimetry (TG) is a convenient technique for studying the kinetics of processes involving solids, such as decomposition and gas–solid reactions, by following the weight loss and/or the rate of weight loss (DTG) of the samples with time. The kinetic parameters usually include the activation energy, the pre-exponential factor and the order of the reaction. Thermal analysis results show that activation energy is the dominating factor in the reactivity equation. Activation energy essentially affects the temperature sensitivity of the reaction rate, although the pre-exponential factor is related more with material structure. Therefore, reactivity of the material is characterized by its activation energy alone. However, these kinetic parameters are highly dependent on experimental conditions such as heating rate, temperature, sample size and the heating medium. The aim of this study was to investigate the carbonization of Tunçbilek lignite by TG. Based on TG curves, the kinetic parameters of the process were calculated using the methods of Coats–Redfern, Horowitz-Metzger and Dharwadkar-Karkhanavala in order to illuminate the mechanism of carbonization of Tunçbilek lignite and lay the theoretical foundations for the preparation of activated carbons using Tunçbilek lignite as precursors. EXPERIMENTAL Tunçbilek lignite was used as a raw material for the preparation of activated carbon in this study. Proximate and ultimate analyses of the lignite sample were reported in Table 1. The carbonization experiments were conducted in a thermal analyzer (TA Q600 model). The weight loss was recorded in the temperature range of 298-1073 K. Carbonization experiments of Tunçbilek lignite consisted of the following steps: about 60 mg of sample were heated at temperatures of 773, 923 and 1073 K under two flowing gas (FG) (N2 and CO2) for 2 h. In order to determine the effects of heating rate and lignite particle size, the experiments were carried out two different heating rates (HR) of 10 and 30 °C/min and two different lignite particle size (1700-700 m and 150-100 m) under flowing nitrogen (N2) at 1073 K. Two different nitrogen flow rates (FR) (15 and 40 cm3/min) and were also used in the experiments to determine the effect of this on carbonization stage. Table 1. Proximate and ultimate analyses of Tunçbilek lignite Sample Moisture Volatile Matter Fixed Carbon Ash C H (%) (%) (%) (%) (%) (%) Tunçbilek lignite 15.42 34.11 35.61 14.86 29.52 3.45
N O (%) (%) 2.26 64.57
Calculation of kinetic parameters Thermogravimetric analysis of the solid pyrolysis is a useful technique for developing an understanding of the pyrolysis mechanism. The rate of decomposition of a solid depends on the temperature and the amount of material. If only a single reaction is included, it is usually supposed that these functions are separable and the equation used to illuminate the progress of reaction is usually described by the following reaction [10]. d (1) k f ( ) dt Where f( ) is a function, the type of which depends on the reaction mechanism, f( ) = (1− )n, n is reaction order; k(T) the temperature dependent rate constant; T the absolute temperature, t the time and the conversion of reaction (the fractional degree of conversion) and was calculated by
w0 wt (2) w0 w f where w 0, w t, w f were the initial weight of the sample (mg), the sample weight at any temperature T or time and final sample weight (mg), respectively. The temperature dependence of the rate constant is usually described by the Arrhenius equation: E kA exp (3) T R where A presented the pre-exponential factor (min−1); E the apparent activation energy (kJ mol−1); R the universal gas constant (8.314×10−3 kJ mol−1 K−1). Under constant heating rate, dT (4) cons tan t dt Coats–Redfern method, considering the pyrolysis mechanism of independent first order parallel reactions and assuming that intra-particle heat transfer and diffusion effects are negligible, given the fine size of the sample particles used, has been used in the kinetic investigation of biomasses [11-14]. In this study, it is also assumed that, given the smaller size of the sample particles, intraparticle heat transfer and diffusion phenomena are negligible. These assumptions allow the use of Coats and Redfern method; one of the most commonly used methods and regarded as giving a good approach in thermal kinetics determination [14]. As it is required in this method, the reaction order has to be fixed in advance to allow the determination of the activation energy (E) and the pre-exponential factor (A), essential for characterizing the pyrolytic kinetics. The substitutions of Eqs. (2)–(4) and f( ) in Eq. (1), and then separate variables to obtain the following equation: d A E (5) exp( (1 ) n dT RT Integrating Eq. (5) ln(1 (1 )1n A.R 2 RT E n≠1 ln (6) 1 ln 2 E RT (1 n)T E
A.R 2 RT E ln(1 ) ln n=1 ln (7) 1 2 E RT T E Making the plot of ln[(−ln(1− )1-n)/T2(1-n)] against 1/T should give a straight line with high correlation coefficient of the linear regression analysis, from which the values of E and A can be derived from slope −E/R and intercept ln[(AR/ E)(1−(2RT/E))] of line, respectively. Horowitz-Metzger, an integral method needs a variable definition such as θ which is dependent to the reference temperature (Ts) derived by TG analysis and representing the maximum point of pyrolysis reaction curve slope which is named as deviation point and also dependent to linearly changing medium temperature (T) [15] . T Ts (8) Conversion at deviation point (αs), determines the reaction order. For n=1, reaction kinetics is defined as follows:
E ln( ln( 1 )) . C 2 R . T s
(9)
Graphics between and ln( ln(1 )) gives a line with a slope of (E/R.Ts2) and intercept“C”. For n≠1, following equation is derived:
1 n E 1 ( 1 ) ln . C 2 1 n ) R . T ( s
(10)
1(1 )1n If the reaction order is correct, the plot of θ versus ln should be a straight line, the ) (1n activation energy (E) can be obtained from the slope and “C” can be obtained from the intercept. By Horowitz-Metzger method, sample amount and heating rate is strongly effected by the change of activation enegy. In their study, Dharwadkar-Karkhanavala investigated the effect of sample amount and heating rate on solid phase pyrolysis and they rearranged the equation as activation energy is independent from sample amount and heating rate [16]. E 100 n 1 ln( ln(1 )) . . C 2 R.Ti T f Ti
(11)
1 (1 )1n E 100 n 1 ln . . C 2 (1 n) R.Ti T f Ti
(12)
where Ti and Tf are reaction initial and final temperatures respectively, C is a constant. Activation energy is calculated by the equation derived from the Horowitz-Metzger method. E slope
T f Ti 100
R Ti 2
(13)
RESULTS AND DISCUSSION The carbonization step is more important in activated carbon production since main pore structure is developed during this stage and kinetic study of carbonization is essential to understand the degradation mechanism; to know the rate of reaction and reaction parameters and to predict the products distribution. The weight loss of the studied Tunçbilek lignite occurred mainly at 423-1073 K; this temperature range was referred to as offering a clue to the key mechanistic steps in the overall lignite breakdown process and considered for the calculation of kinetic parameters. To determine the kinetic parameters, different reaction orders (n= 0, ⅓, ½, ⅔, 1, 2 and 3) were assumed and the kinetic parameters were calculated using eqs 6, 7, 9,10, 11 and 12. The reaction order is ascertained based on the best criteria. Thermogravimetric results of Tunçbilek lignite and carbonization reaction kinetic parameters were presented in Table 2 and 3 respectively. Three different temperatures of 773, 923 and 1073 K were investigated and the effect of carbonization temperature on weight loss was observed in two different atmospheres (CO2 and N2). The carbonization behavior of lignite samples in CO2 atmosphere shows that the carbonization process can be approximately described by moisture and volatile lumps. The temperature range for the volatile evolution is referred to as offering a clue to the key mechanistic steps in the overall coal breakdown process. In carbonization process, weight loss relevant to 773 and 923 K were 31.9 % and 32.9 %,
respectively. When the temperature was increased to 1073 K, weight loss became 48.8 %. At low carbonization temperatures the increment in the weight loss is negligible whereas at high temperatures the increment is quite large. The reason of this is thought to be due to the possible reactions between carbon-CO2 at different temperatures. Although the carbon-CO2 reactions start at temperatures around 923 K, reactions become dominant above 1073 K. Table 2. Thermogravimetric results of Tunçbilek lignite Sample Code A1 A2 A3 A4 A5 A6 A7 A8 A9
Carbonization Experimental Conditions T / HR /FG / FR / t) 1073/ 30 / N2 /15 /1 1073/ 30 / N2 /40 /1 1073 /10 / N2 /40 /1 1073 /30 / N2 /40 /1 1073/30 /CO2 /40 /1 773 /30 /CO2 /40 /1 923 /30 /CO2 /40 /1 773 /30 /N2 /40 /1 923 /30 /N2 /40 /1
Particle size (m) 150-100 150-100 150-100 1700-700 1700-700 1700-700 1700-700 1700-700 1700-700
Weight loss (%) 36.63 41.05 40.16 39.32 48.81 31.90 32.89 36.33 36.50
Tp (K) 619 649 615 653 679 588 605 627 615
Table 3. Carbonization reaction kinetic parameters of Tunçbilek lignite Sample Code
A1 A2 A3 A4 A5 A6 A7 A8 A9
Coats Redfern
Horowitz Metzger
C
n
r2
67.98
-0.19
2
0.965
0.974
65.08
0.47
2
0.974
2
0.976
66.51
0.07
2
0.976
0.35
2
0.975
55.90
0.35
2
0.975
36.96
0.08
2
0.976
57.87
0.08
2
0.976
0,980
46.95
0.82
2
0.987
62.19
0.82
2
0.987
2
0,961
53.59
1.36
2
0.984
68.96
1.36
2
0.984
2
0,955
31.66
2.53
2
0.982
43.32
2.53
2
0.982
2
0,959
34.09
0.58
2
0.953
51.30
0.58
2
0.953
A
(kJ/mol)
(s-1)
36.21
2.38 x1012
2
0.979
37.12
-0.19
2
0.965
36.52
3.86 x10
12
2
0.972
39.08
0.47
2
2.98 x10
12
2
0.964
38.85
0.07
3.80 x10
12
2
0,969
37.74
1.10 x10
12
2
0,993
6.24 x10
12
2
8.36 x10
12
3.02 x10
12
6.29 x10
12
38.49 34.58 38.67 40.17 34.92 37.86
E
C
n
r2
E
35.91
n
r2
Dharwadkar-Karkhanavala
(kJ/mol)
E (kJ/mol)
Intensity of the reactions between the CO2 and carbon results in then more amounts of volatiles released from the char. The temperature at which volatile evolution reaches a maximum (maximum peak temperature, Tp) during carbonization was also determined and listed in Table 2. The increase in temperature from 773 to 1073 K induced not only a weight loss, but also the peak temperatures shifted to higher values. The kinetic parameters of Tunçbilek lignite under different temperatures at CO2 atmosphere were calculated (eqs 6, 7, 9, 10, 11 and 12) and the results are listed in Table 3. As can be seen from the table, higher activation energies were found for the lower temperatures (773 and 923 K) than high temperature (1073 K) (38.67, 46.95, 62.19 kJ/mol and 40.17, 53.59, 68.96 kJ/mol versus 34.58, 36,96, 57.87 kJ/mol, respectively). Apparently, carbon-CO2 reactions which become dominant at 1073 K
influence the thermal degradation of Tunçbilek lignite. Although the calculated activation energies according to the three models showed the same behavior for the different carbonization temperatures, it can be also observed that the activation energy values according to these models were different whereas those of Dharwadkar-Karkhanavala were quite high. This might be due to the calculation differences that exit between three models. The effect of temperature was also investigated when N2 gas is used as the gaseous environment during carbonization. Weight loss was 36.3 % when the carbonization was carried out at 773 K, when temperature was increased to 923 and 1073 K weight losses were found to be 36.5 % and 39.3 %. respectively. The increase in temperature did not cause an important change in weight loss. A slight decline was observed in maximum peak temperature at temperatures of 773 and 923 K. On the other hand, the temperature was increased to 1073 K. the peak temperature shifted to slightly higher value. As a comparison, carbonization carried out at N2 atmosphere at 773 and 923 K, the weight loss is more than the carbonization carried out at CO2 atmosphere. The reason is carbon-CO2 reaction is more dominant at higher temperatures as emphasized before. Similar results are found by the study of palm oil waste [17]. Ting Yang et al. [18] also pointed out that increasing temperature causes increasing release of volatiles and increasing carbon burn off due to carbon–CO2 reaction. When the activation temperature is increased from 998 to 1048 K, the percentage of volatile content decreases continuously for both series as the release of volatiles is the dominant reaction. However, for temperature above 1048 K, the carbon burn off due to the carbon–CO2 reaction has now become dominant, resulting in decreasing percentage of carbon content but increasing percentage of volatile content with increasing temperature. Kinetics of weight losses in N2 atmosphere was slightly different to that in CO2 atmosphere. Especially at high temperature (1073 K), the calculated activation energies for all models were higher than that in CO2 atmosphere. This difference can be explained due to the carbon–CO2 reaction at 1073 K in CO2 atmosphere. The heating rate is one of the most important parameters influencing the thermal decomposition characteristics. In this study, the heating rates were varied as 10 K/min and 30 K/min (A3 and A2). Increasing the heating rate from 10 to 30 K/min increased the weight loss from 40.2 to 41.1 % because the higher rate results in a higher thermal gradient across the char sample. This large thermal gradient favors the diffusion of the N2 molecules into the carbon structure, resulting in a slightly higher weight loss. However, higher heating rate would translate into a shorter contact time between the char and the N2 gas during the temperature ramp-up stage. This would imply a shorter dwell time for the char and therefore the increment in the weight loss was not high and the effect of the heating rate was negligible. By the pyrolysis of palm oil waste, Yang et al. [17] indicated that increasing of heating rate from 0.1 K/min to 100 K/min. weight loss decreased. This decrease could be explained by the enlargement of heat transfer rate and the decrease in contact time when heating rate was increased [17-19]. The peak temperature at which volatile evolution reaches a maximum (Tp) of lignite sample under different heating rates was also determined and listed in Table 2. Table 2 showed that a higher heating rate led to increase of the temperature corresponding to the peak. For example, the maximum degradation temperatures were 615 and 649 K at the heating rates of 10 and 30 K/min. This shift of peak towards a higher temperature with an increasing heating rate was possibly caused by the effect of the kinetics of the decomposition, which resulted in a delayed degradation. This observation was consistent with that of Shie [19] and Benbouzid [20] et al, who reported that a higher heating rate resulted in a higher peak value of reaction rate and a higher temperature for its occurrence. Although the characteristic temperatures changed with the heating rate, increasing from 10 and 30 K/min, the calculated activation energies are constant (given in Table 3), indicating that the heat transfer limits are not significant for these cases.
Effect of gas flow rate on carbonization was investigated for two different gas flow rate (15 and 40 cm3/min) in N2 atmosphere (A1 and A2). Weight loss was increased by the increase of gas flow rate. For example, weight loss for the carbonization with 15 cm3/min gas flow rate was 36.6 %, whereas it was obtained to be 41.1 % in case of the gas flow rate of 40 cm3/min. Especially, the increased flow rate enhances the diffusion of the N2 molecules into the carbon structure and causes an increase of the weight loss. But since 40 cm3/min and higher flow rates are enough for a carbonization reaction to begin, weight loss stays approximately the same. Similar studies in the literature showed that the weight loss (%) is not affected by the gas flow rate [17, 18]. These differences are natural since in termogravimetric analysis where a few amount of sample such as 50 mg is used, varying the flow rate of N2 from 40 cm3/min to 120 cm3/min did not have influence on the shell pyrolysis since 40 cm3/min N2 flow rate is enough to mitigate the potential difference caused by the external heat and mass transfer in the gas phase [17]. In Yang et al. study [18], flow rate influence was also studied in order to determine on the BET surface area and micropore volume of activated carbons. It was found that at a flow rate of 100 cm3/min, the BET surface areas of activated carbons were maximum values. For CO2 flow rates of 150 cm3/min and greater, the higher flow velocities above that for 100 cm3/min reduced the contact time between the CO2 molecules and the sample surface and thereby decrease the diffusion of the molecules into the pore structures and hence reduced BET surface values were obtained. Maximum peak temperature was slightly increased by an increase in the gas flow rate. Changing the gas flow rate from 15 to 40 cm3/min, the maximum peak temperature was shifted from 619 to 649 K. The activation energies were determined by performing experiments at two gas flow rate (15 and 40 cm3/min) and listed in Table 3. Although the characteristic temperatures slightly changed with the gas flow rate, the activation energies were found as nearly the same. Lignite with two different particle size intervals (150-100 and 1700-700 µm) was used in carbonization to produce activated carbon (A2 and A4). When the particle size was decreased from 1700-700 to 150100 µm, the weight loss was increased from 39.3 % to 41.1 %. Since such a difference observed in the weight loss could also be resulted from any other effect during the experimental conditions, the difference cannot be attributed to the effect of particle size change. Because, particle-size changes over a wide range (from 150 to 1700 µm) indicates the minor influence of internal heat/mass transfer on carbonization, similar to the previous reports [17, 21]. Yang et al. [17] used palm oil waste in pyrolysis and the results show that particle-size changes over a wide range (from 250 µm to>2 mm) do not have significant influence on the shell pyrolysis. For a larger particle (>2 mm), a longer time was required to heat it and release the volatile matters due to the heat and mass transfer differences that exist between the internal and external particles. The kinetic parameters of lignite samples under two different particle sizes were also calculated and the results listed in Table 3. The activation energies for the particles ranging in size from 100 to 1700 µm are quite similar. This result is consistent with thermogravimetric result as mentioned above which particle size changes over a wide range (from 100 to 1700 µm) do not have significant influence on lignite carbonization. Conclusion The effects of carbonization conditions (temperature, heating rate, particle size, gas flow rate, etc) on the weight loss of the samples and the maximum peak temperatures were revealed in this study. For Tunçbilek lignite, experimental evidence showed that carbonization temperature, heating rate, gas flow rate and carbonization gaseous atmosphere (N2 or CO2) influenced on weight loss of the samples and the maximum peak temperature. The results of carbonization reaction kinetic parameters of Tunçbilek lignite according to Coats-Redfern, Horowitz-Metzger and Dharwadkar-Karkhanavala kinetic models
were obtained. The main carbonization stage could be described by the second order and activation energies varied in the range of 31.7-69.0 kJ/mol. REFERENCES [1] A. Ahmadpour. D.D. Do. The preparation of active carbons from coal by chemical and physical activation. Carbon. 34 (1996) 471–479. [2] C.A. Philip. B.S. Girgis. Adsorption characteristics of microporous carbons from apricot stones activated by phosphoric acid. J Chem Technol and Biotechnol. 67 (1996) 248-254. [3] W.M.A.W. Daud. W.S.W Ali. M.Z. Sulaiman. The effects of carbonization temperature on pore development in palm-shell based activated carbon. Carbon 28 (2000) 1925-1932. [4] M. Daniel. T.F. Vanessa. F. Vanessa. Activated carbons from lignin: kinetic modeling of the pyrolysis of Kraft lignin activated with phosphoric acid. Chem. Eng. J. 106 (2005) 1-12. [5] A. Ahmadpour. D.D. Do. Characterization of modified activated carbons: equilibria and dynamics. Carbon 33 (1995) 1393-1398. [6] S. Gregg. K.S.W. Sing. Adsorption Surface Area and Porosity. Academic Press. London. 1982. [7] J. Guo. A.C. Lua. Preparation of activated carbons from oil palm stone chars by microwave induced carbon dioxide activation. Carbon 38 (2000) 1985-1993. [8] V. Baliga. R. Sharma. D. Miser. T. McGrath. M.R. Hajaligol. Physical characterization of pyrolzed tobacco and tobacco components. J. Anal. Appl. Pyrolysis 66 (2003) 191-215. [9] R.K. Sharma. J.B. Wooten. V.L. Baliga. P.A. Martoglio-Smith. M.R. Hajaligol. Characterization of char from the pyrolysis of tobacco. J. Agric. Food Chem. 50 (2002) 771-783. [10] M. Garcia-Perez. A. Chaala. J. Yang. Co-pyrolysis of sugarcane bagasse with petroleum residue. Part I. Thermogravimetric analysis. Fuel 80 (2001) 1245–1258. [11] A.W. Coats. J.P. Redfern.. Kinetic parameters from thermogravimetric data. Nature 201 (1964) 68– 69. [12] J.J.M. Órfáo. F.G. Martins. Kinetic analysis of thermogravimetric data obtained under linear temperature programming-a method based on calculations of the temperature integral by interpolation. Thermochim. Acta 390 (2002) 195–211. [13] J.D. Sewry. M.E. Brown.. Model-free kinetic analysis. Thermochim. Acta 390 (2002) 217–225. [14] L.T. Vlaev. I.G. Markovska. L.A. Lyubchev. Non-isothermal kinetics of pyrolysis of rice husk. Thermochim. Acta 406 (2003) 1–7. [15] H.H. Horowitz. G. Metzger. A new analysis of thermogravimetric traces. Analyt. Chem. 35 (1973) 1464–1468. [16] S. R. Dharwadkar. M. D. Karkhavala. Calculation of activation energy of decomposition reactions from thermogravimetric analysis. Vol.II. Academic pres. New York. (1969)1049-1059. [17] H. Yang. R. Yan. T. Chin. D.T. Liang. H. Chen. C. Zheng. Thermogravimetric analysis-fourier transform infrared analysis of palm oil waste pyrolysis. Energy & Fuels 18 (2004) 1814–1821. [18] T. Yang. A.C. Lua. Characteristics of activated carbons prepared from pistachio-nut shells by physical activation. J of Colloid and Interf Sci 267 (2003) 408–417. [19] J.L. Shie. C.Y. Chang. J.P. Lin. C.H. Wu. D.J. Lee. Resources recovery of oil sludge by pyrolysis: kinetics study. J. Chem.Technol. Biotechnol. 75 (2000) 443–450. [20] M. Benbouzid. S. Hafsi. Thermal and kinetic analyses of pure and oxidized bitumens. Fuel 87 (2007) 1585–159. [21] A. Aranda. R. Murillo. T. Garc´ıa. M.S. Call´en. A.M. Mastral. Steam activation of tyre pyrolytic carbon black: Kinetic study in a thermobalance. Chem Eng J 126 (2007) 79–85.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COAL-DERIVED PRODUCTS SYNTHESIS OF NITROGEN-DOPED CARBON MATERIALS FROM COAL-TAR AND PETROLEUM PITCHES AND NITROGEN CONTAINING ORGANIC PRECURSORS Z. R. Ismagilov, Prof., Boreskov Institute of Catalysis SB RAS Pr.Ak. Lavrentieva, 5, Novosibirsk, 630090, Russia; tel.: 7-383-3306219, fax:7-383-3306219, e-mail: [email protected] Ch. N. Barnakov, Dr., A.P. Kozlov, Dr., Institute of Coal and Coal Chemistry SB RAS Pr.Sovetskiy 18, Kemerovo, 650099, Russia; tel.: 7-3842-368188, fax:7-3842-211838, e-mail: [email protected] E.I.Andreikov, Prof. , Institute of Organik Synthesis UB RAS ul.S.Kovalevskoi/Akademicheskaya, 22/20, Ekaterinburg, 620616, Russia; tel.: 7-343-3493535, e-mail: [email protected] M.A.Kerzhentsev,Dr., I.Z.Ismagilov, Dr.,Boreskov Institute of Catalysis SB RAS Pr.Ak. Lavrentieva, 5, Novosibirsk, 630090, Russia, Russia; tel.: 7-383-3306219, fax:7-383-3306219, e-mail: [email protected]
Abstract: Doping of carbon nanomaterials with nitrogen opens possibility of regulation of their functional electrophysical and adsorption properties. A series of amorphous microporous carbon materials were prepared by chemical and subsequent thermal treatment of various nitrogen containing organic precursors, or their mixtures: o-nitroaniline, 8-oxyquinoline, benzotriazole, etc. The obtained products have nitrogen content in the range of 0.5-20 wt.%. The samples have a very high specific surface area (1000-3100 m2/g) and a large fraction of micropores - over 70%. The effects of the nature of the precursor and the preparation conditions on nitrogen content and product properties were studied. It was shown that the BET area and pore volume generally decrease and the N content increases with lowering of the carbonization temperature. Another series of nitrogen doped materials were prepared by carbonization of petroleum pitch, coal-tar pitch or coal-tar pitch mixed with polyvinylchloride (PVC) as carbon precursors and polyacrylonitrile (PAN) as a nitrogen precursor. Depending on the conditions of the preparation and the PAN/pitch ratio, materials containing different amounts of nitrogen (from 5.3 to 13.6%) were obtained. The presence of PVC in coal-tar pitch was shown to increase the yield of the nitrogen containing carbonaceous product. As a result of chemical interactions between carbon precursors and PAN the aromatic structures containing nitrogen are formed. IR spectra showed the presence of pyrrole structures in the products. The carbonization of carbon precursors with PAN was shown to be an effective method of preparation of carbon materials with high nitrogen content. The state of nitrogen species in the prepared N-doped carbon materials was studied by XPS. The results showed that three types of nitrogen are generally present in these materials: with binding energies of 398.5; 400,1 and 400.8 eV. According to literature, these signals correspond to pyridine-like N, pyrrole nitrogen and bridgehead-type N incorporated into graphitic network, respectively. Acknowledgments: The authors acknowledge the support of this work by Presidium RAS (Integration Project 27.58) and by Presidium SB RAS (Integration Project 88).
1
EFFECT OF MINERAL MATTER OF BROWN COALS ON THE REACTIVITY OF CHAR STEAM GASIFICATION AND ON THE PROPERTIES OF ACTIVATED CARBONS P.N.Kuznetsov, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences 42 K.Marx str, 660049 Krasnoyarsk, Russia Fax: 7(391)2124720; E-mail: [email protected] Kolesnikova S.M., L.I.Kuznetsova Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences, 42 K.Marx str, 660049 Krasnoyarsk, Russia Fax: 7(391)2124720; E-mail: [email protected] Yu.F.Patrakov Institute of Coal and Coal Chemistry of Siberian Branch of Russian Academy of Sciences, Kemerovo, Russia Abstract. The goal of this paper is to gain a fundamental understanding about steam gasification of different brown coal chars under low temperature and high pressure. The coals from different deposits of Kansk-Achinsk Basin, Lena Basin and from Yallourn Basin (for comparison) were used to study the role of the naturally occurring metals on the physico-chemical properties of the chars and steam gasification reactivity. The data on how the preliminary decationization with solution of hydrochloric acid affects the structural characteristics of brown coal chars and their gasification reactivities are presented. The importance of controlling parameters (such as structural characteristics of carbon and naturally occurring metals) in the physico-chemical properties of the chars, in the gasification reactivity and in the textural properties of gasified chars is considered. The concentration of calcium and catalytic activity of calcium species were found to be a key factor for their reactivity for steam gasification and carbon dioxide formation. Quantitative relations were revealed. Gasification is one of the most important, clean and efficient technology to convert coal to energy carriers and to a variety of chemicals with low emissions of air pollutants, while allowing carbon dioxide to be captured [1-4]. This high temperature process is used to produce synthetic natural gas, synthesis gas, hydrogen, ammonia, methanol and other chemicals and involved in any coal processing technology which needs hydrogen and synthesis gas. Also, gasification reactions lie in the basis for the production of a variety of valuable porous carbon materials, catalyst support, sorbents, molecular sieves, in particular. Besides, power generation through combined cycle gasification offers potential advantages over conventional combustion including higher cycle efficiencies and the ready incorporation of advanced emission control technology. The gasification of coals is a two-stage process: the initial pyrolysis of carbon-containing raw materials to produce char and subsequent high-temperature partial oxidation of char with
2
steam, CO2, O2 or H2. Conversion of coal in gasification is controlled by a combination of physical and chemical phenomenon occurring in both steps, char gasification being the slower step. Different ranked coals can be gasified successfully by using the appropriate technology, however the reactivity tends to decrease with an increase in the rank of coal [6-12]. For example, a 250 fold difference in the rate of steam gasification was reported by Solano ei al. [9] depending on the precursor coal. The consensus exists that the reasons for these differences in the gasification rates of chars can be related to either intrinsic reactivity of chars, their pore structures or catalytic effect of mineral impurities, the factors controlling reaction rates being the same for steam, CO2, and O2 gasification [5, 6]. The application of the catalysts allows the gasification reactivity of coal to be increased and hence the temperature to be decreased [2, 5, 6, 13-19]. Catalytic char gasification involves an oxygen transfer step via the formation of an intermediates with catalytic species. Alkali and alkaline earth metal compounds added to brown coal in the forms of carbonate or/and other salts are well known [2, 13, 20] to exert a considerable effect on the gasification. Generally, the activity of alkali metal salts increases with the size of the cation, from lithium to caesium. However, alkali metals are inclined to react with mineral matter to form catalytically inactive compounds such as that with clays. Among the alkaline earth metals, Ca species showed usually the maximum effect on char reactivity [5, 12, 13, 17-24]. A linear relationship between the gasification rate and the calcium loading in Yallourn and Texas coals was observed in [13, 22, 23]. Ohtsuka and Tomita [13] reported the decrease in the gasification temperature by 150° C upon the addition of calcium catalytic compounds to Yallourn brown coal. This can be due to the lowering of the activation energy from 110–130 for noncatalytic gasification to 40–90 kJ/mol for calcium catalysed gasification of coal [12]. The fine CaCO3 and CaO particles generated from the calcium precursors during carbonization and steam gasification steps were found to be catalytically active species. The equilibration between these calcium forms occurred easy in the temperature range of 9231023 K depending upon the temperature, CO2 concentration and the presence of carbon, CaCO3 being the predominant species at relatively low temperaturees of 923-973K. Ohtsuka and Tomita [13] interpreted the catalysis of the gasification reaction by Ca species in terms of a carbonateoxide cycle mechanism. The following row of catalytic activity of a variety of alkali and alkaline-earth metal catalysts for steam gasification of subbituminous coal char was reported in [19]: K2CO3 Na2CO3 KCl NaCl CaCl2 CaO. The catalytic species with Mg and Ba showed lower activity as compared to Ca [12, 13, 22, 23]. Other metal species such as that containing Fe, Ni, Mn behave also catalytic activity for CO2 and steam gasification. According to Takarada [25], Ni based catalyst showed more higher activity as compared to K2CO3 for steam gasification of different coals, especially for low rank coals at a temperature of around 8000C. The following rows of catalytic activity of metal species for char gasification were obtained: Na Са Fe Mg in [12] and К Са Fe Mg in [26]. According to Tomkow et al. [27], catalytic activity of Na, Ca, Fe and
3
Mn species for gasification of chars from coal humic acids decreased in the following order Са NaFe Mn. Many researchers [5,6,9,12,13,15] studied the effects of the mineral matter naturally occuring in the coals on the behaviour of respective chars on gasification. The gasification rate of the coals with higher contents of mineral matter behaved usually higher reactivity as compared to that with low content, partially demineralized coals showing the least reactivity [5, 12, 26, 28-31]. From 5 to 25 fold differences in the gasification reactivities of the parent brown coal chars and the respective partially demineralized ones were found [9, 26, 28]. Miura and Hashimoto [5] reviewed in detail a large body of the literature data for the reactivities of coal chars for gasification with steam, CO2 and oxygen to clarify the factors controlling gasification process. Based on steam gasification data for 95 chars from 68 coals whose carbon contents ranged from 65.0 to 93.7%, the authors concluded that the gasification rates of the chars from lower rank coals (C<80 %) were affected mainly by naturally occuring metal species such as that containing Ca, K, Na and Mg acting as the gasification catalysts. Of alkali and alkaline earth metals, Ca is by far the most abundant. It occurs mostly in the form of organically bond species which were easily converted into highly dispersed catalytically active carbonate or oxide particles upon the pyrolysis and gasification. On the other hand, the gasification rates of chars from the higher rank coals (C>80 %) were controlled mainly by the intrinsic reactivity of the coal chars. From the detail analysis of the literature data, a valid measure of reactivity for all kinds of coals based on the amount of chemisorbed oxygen was discussed [5]. Kandiyoty [6] made attempts to correlate gasification conversions with specific inorganic constituents for a set of 23 coal samples. Also, the influence of demineralization and impregnation with various inorganic components on the pyrolysis and CO2 gasification behaviour of two subbituminous coals with 81-82 % Cdaf have been examined. Based on the results obtained, the authors reported that though inorganic matter clearly affects conversions during gasification, it is difficult to find systematic patterns, regarding the effect of a specific inorganic component in different coals. Actually, the effect of impregnated salts on the CO2 gasification reactivity of coals changed with reaction pressure and coal in a specific manner. For steam gasification catalysis, not only the content of catalytic metal is important but also the composition of the ash. Thus, an almost proportional increase in the gasification rate was reported in [28] as Mg content in the char increased. At the same time, no similar relations were found for usually catalytically more active Na, K and Ca species. This proved to be due to the chemical state of these elements. The X-ray diffraction patterns showed that alkali metals were associated with the aluminosilicate lattice where they losed catalytic activity, and a fraction of Ca was in a non-catalytic gypsum form. On the other hand, an anomalous increase in steam, CO2 and air gasification reactivity of the high volatile bituminous coal after it has been demineralized with HF-HCl was reported by [32]. However, Fujita et al. [33] found the reactivity of chars for CO2 gasification at 900-12000C to relate to the amounts of Fe2O3, CaO, MgO, Na2O and K2O. The sum of the amounts of Fe2O3 and К2O gave the best correlation.
4
Actually, catalysis by mineral matters has not always been found to provide adequate explanation and it has been not always rather obviously how the relative importance of the three controlling factors (catalysis, pore structure and intrinsic reactivity of carbon) changed depending on coal. According to [18, 35, 36] speculations, alkali and alkaline earth metal compounds can play the role not only of catalysts but also structure-forming agents during pyrolysis. As pyrolysis procedes, the coal structure undergoes alterations that may involve changes in plasticity, density, and porosity. Each of these structural changes, which may be influenced by the presence of mineral species, has consequences for the accessible surface area of the char formed and thus for the reactivity of the char in gasification. The authors [10, 37-39] reported that difference in reactivity of coal chars can be attributed to the difference in their physical carbon structure. Molina [40] reported the results on the effect of demineralization and metal loading on the mode of CO2 chemisorption and CO2 gasification reactivity of brown coal chars. It was found that demineralization resulted in the 5 fold decrease in the rate of char gasification. The addition of Fe and K to parent coal gave the chars whose reactivities increased by a factors of 6 and 15, respectively. However, other relations were observed when the data on char reactivities were correlated with the data on strong chemisorption of CO2 at 3000C. In this case, the ratio of the Fe and K loaded chars reactivities normalized to the surface areas which were measured from strong CO2 chemisorption, to that of char prepared from the parent coal proved to be only 4.5 (instead of 6-15). No differences between the surface normalized reactivity of Fe loaded char and that of K loaded char and also between the surface normalized reactivities of the parent and demineralized chars were observed. These findings show the difficulties in relating information on structural features of coals to gasification reactivity and in developing a predictive tool, linking catalytic effects to amounts of particular inorganic components. This needs further investigations to gain fundamentals about pyrolysis and gasification reactions of coals. It is essential to make more extensive studies on many combinations of coals and catalytic species to better understand the factors governing the extent of gasification and the yields of products in order to effectively utilize both catalyst and coal during gasification. In this respect, brown coals from the Kansk-Achinsk Basin are of particular interest. They differ from the brown coals of other large deposits in Russia and in the world, firstly, with their rather low contents of mineral matter (predominantly less that 8 %) [41]. Secondly, among the constituents of mineral matter, calcium compounds are predominant and basically connected with organic matter in the forms of carboxylate cross-links [42, 43]. These are Ca cations which affect considerably the properties, structure, and reactivity of the Kansk-Achinsk coals [43, 44]: hydrogenation reactivity with tetralin was found to be inversely proportional to the logarithm of Ca concentration [45], the properties of activated carbons were also affected dramatically by the presence of calcium [36]. Thirdly, the coals from Kansk-Achinsk Basin are characterized by low concentrations of toxic elements (for the majority of the elements, several times lower than coal clarkes, that is, the average abundances in the world’s coals [41].
5
The goal of this paper was to gain a fundamental understanding about steam gasification of different brown coal chars under low temperature and high pressure. The coals from different deposits of Kansk-Achinsk Basin, Lena Basin and from Yallourn Basin (for comparison) were used to study the role of the naturally occurring metals on the physico-chemical properties of the chars and steam gasification reactivity. The data on how the preliminary decationization with solution of hydrochloric acid affects the structural characteristics of brown coal chars and their gasification reactivities are presented. The importance of controlling parameters (such as structural characteristics of carbon and naturally occurring metals) in the physico-chemical properties of the chars, in the gasification reactivity and in the textural properties of gasified chars is considered. EXPERIMENTAL A suite of brown coal samples from the Borodino and Berezovo deposits of Kansk-Achinsk Basin, Kangalas deposit of Lena Basin was used in the paper. A sample of brown coal from Yallourn deposit in Australia was also used for comparison, since it contains few inorganic matter, the latter occurring mainly in the form of metal ions [46]. Prior to use the samples were crushed to the particle size range 0.25-0.5 mm, and dried in a vacuum chamber at 850C. Some of the coal samples were decationized by treating with 0.2 N HCl solution in a glass flask filled with an inert gas with continuous stirring using a magnetic stirrer. In 3 hours, the solution was decanted, and the treatment with fresh 0.2 N HCl solution was repeated. After completion of the treatment, the coal was filtered off from the solution, washed with distilled water to a negative reaction for chloride ions, and dried. The coal samples were subjected to proximate and ultimate analysis. The contents of metals were determined by atomic absorption spectrometry and X-ray fluorescence analysis. The carbonization of the parent and demineralized coals was carried in a quartz tube of 20 mm in diameter with no access for air at the temperature of 7000C for 1 h. The heating rate was 78 0C/min. The resulting chars were further subjected to steam gasification at 700 and 7300C for 2 hours. The reaction was carried in 0.09 ml autoclaves charged with 3 g of char and 4 g of water. The autoclaves were purged with nitrogen from the air, then hermetically sealed and inserted (5 autoclaves simultaneously) into a rotated sand-bath. The autoclave unit thus allowed five char samples to be tested simultaneously under strictly the same conditions. The heating rate of autoclaves was 7-8 0C/min. After the reaction, the autoclaves were cooled and the gaseous products were analysed by a gas chromatograph. The extents of burn-off of the chars were evaluated based on the quantity of organic matter before and after gasification and also by measuring the content of carbon in the carbonaceous gases produced. The phase composition of mineral matter and the characteristics of spatial carbon structure were analyzed by X-ray diffraction using Сu Kα- radiation and Dron-3 diffractometer. The textural characteristics (BET surface area, pore volume and average pore diameter) of chars and activated carbons were determined from the nitrogen adsorption isotherms using ASAP 2040 Micromeritics instrument, Sorptometer M analizator and gasometer GCH-1.
6
RESULTS AND DISCUSSION Coal characterization. The characterization of mineral matters of brown coals from different deposits are shown in (Table 1). The contents of ash in the coals ranged from 1.4 to 7.8 wt. %. Yallourn coal had the lowest ash content. Along with Si and Al, calcium prevailed in all the samples, 10 fold and 2 fold differences in Ca concentration showing the parent coals from different deposits and the coal samples from the same Borodino deposit, respectively. The concentrations of Mg were less than that of Ca by a factor of 4-7, except for Yallourn coal which featured almost equal weight amount of Mg and Ca. The contents of Fe varied from 0.06 to 0.55 wt. %. The alkali metals (Na and K) occurred in very low concentrations. No correlations were observed between the content of ash in the coals and contents of specific metals. Table 1. The characterization of mineral matters of brown coals from different deposits Coal deposit, sample
Аd,
Content, wt. % on daf coal
wt. %
Ca
Mg
Borodino, Bor*
7.7–4.1
1.34–0.73
0.40–0.08
Berezovo, Ber9
4.3
1.40
0.20
0.50
0.01
Kangalas, Ka20
7.9
1.0
-
0.55
0.15
Yallourn, Ya10
1.4
0.15
0.16
0.40
0.06
4.8-0.9
610-2
10-5
0.05-0.2
0
Borodino, BorD**
Fe
Na+K
0.28–0.06 0.015-0.009
* coal samples from different sites of deposit; **HCl demineralized Borodino coals. The contents of minor elements determined for the Borodino suit of coal samples occurred in trace concentrations. For most elements, except for the alkaline earth Sr and Ba, and also for Mn, the concentrations were much lower than coal Clarke numbers, i.e. the average content in the world’s coals), showing a typical feature of Kansk-Achinsk brown coals [41]. The treatment with 0.2 N HCl solution decreased the total ash content by a factor of 2– 4 (Table 1). The cations of alkali and alkaline earth metals (Ca, Mg, Sr, and Ba) were almost completely extracted in all cases. The recovery of Fe varied from 50 to 93%. Also the cations of minor elements such as Mn, V, Be readily passed into HCl solution (95 % and more). The recovery of Ni was no more than 12 %; this suggests its occurrence in a non-ion-exchangeable form in the native inorganic minerals. The composition of organic matter of selected representative coal samples are displayed in Table 2. The data show similar composition of coals, with the exception of Yallourn coal, which features a lower carbon content and increased oxygen content and also of Kangalas coal with enhanced hydrogen content. The nitrogen and sulphur contents in all the samples were close to 0.8 and 0.3 %, respectively. HCl treatment of coals hardly affect the ultimate analysis data.
7
Table 2. Ultimate analysis of representative parent coal samples. Ad, % Coal sample Borodino, Bor1 Borodino, Bor2 Borodino, Bor3 Borodino, Bor6 Berezovo, Ber9 Kangalas, Kan20 Yallourn, Ya10 Borodino, Bor6D*
4.1 6.5 7.7 5.3 4.3 7.9 1.4 1.3
Wt.% on daf coal C
H
N
S
O
69.7 70.8 71.3 70.0 70.1 71.0 66.6 71.3
4.9 4.8 4.8 4.8 4.9 5.5 4.7 4.9
0.8 0.8 0.9 0.8 0.9 0.8 0.6 0.8
0.3 0.3 0.2 0.2 0.3 0.4 0.3 0.3
24.3 23.3 22.8 24.2 23.8 21.8 27.8 22.7
*HCl demineralized Borodino coal Bor6. Char preparation. The yields of chars from the Kansk-Achinsk coals at 7000C varied from 53 to 60 % and those for Kangalas and Yallourn chars did 50% (Table 3). Kansk-Achinsk and Kangalas chars had poor porous structure with a total porous volume of no more than 0.013 cm3/g (mainly micropore) and surface areas ranged from 56 to 244 m2/g. The char from Yallourn coal had much higher surface area of 419 m2/g and pore volume of 0.15 cm3/g. The surface areas of HCl treated Kansk-Achinsk coal chars were in the same range as for the chars from the parent coals. No definite correlations were found between the textural properties of chars and composition of the coals, such as proximate and ultimate analytical data and the contents of specific metals. Table 3. The characterization of chars produced from brown coals of different deposits. Char yield,
Ad,
Surface
wt.%
wt.%
area, m2/g
Grafite-like
-component
Borodino, Bor*
53-60
6.8-12.9
56-244
74-89
11-26
Berezovo, Ber9
55
7.5
167
93
7
Kangalas, Ka20
50
14.7
42
50 56-60
2.8 1.4-7.7
419 112-217
78 91
22 9
60-72
28-40
Coal deposit
Yallourn, Ya10 Borodino, BorD **
Content, %
* coal samples from different places of Borodino deposit; **HCl demineralized Borodino coals. The X-ray diffraction patterns of the chars showed mainly quartz with different quantities to be present in all the samples (Fig1). No distinct reflections related to other inorganic species were reliably detected because of insignificant amounts, and probably due to high degree of dispersion. On the other hand, a large and broad asymmetrical reflection in the region 2 from 10 to 35° related to a specific intermolecular ordering of carbon and a weak reflection with a
8
maximum at 2 centered at 44° Карбонизат Берез. п/ион Карбонизат Берез. Карбонизат Бор.3
400
Карбонизат Австрал 300
Demineralized Berezovo Ber9D Berezovo Ber9
200
Borodino Bor3 100
Yallourn Ya10 10
20
30
40
50
60
70
2
Figure 1. The X-ray diffraction patterns of the chars from the demineralized and parent coals.
due to intramolecular ordering of the carbon were observed. Computer treatment showed the profile of the intermolecular ordering to be well simulated by a superposition of two Gaussians. They reflected rather ordered graphite-like matter of the stacked aromatic clusters and poorly
ordered
-component
located at the periphery [47]. In Table 3, the percentages of the graphite-like and
nonordered
-component
portions calculated from the Xray patterns are displayed. It is seen that the structures of the parent coal chars consist of mainly ordered graphite-like carbon matter (74-93 %) with a small portion of nonordered -component (7- 26 %), whereas the increased portion of -component (28-40 %) and diminished portion of graphite-like component (60-72 %) are present in the HCl treated chars. It thus appears that metal containing species involving ion-exchangeable metal cations, in partilular, could promote somehow the ring condensation reactions towards the more ordered graphite-like structure, although not yet reaching the condensation level in graphite. The graphite-like stacks in the chars have the diameters, with the occasional exceptions, mainly from 1.41 to 1.65 nm with the average statistical value of 1.54±0.10 nm and interlayer spacing of 0.375±0.006 nm (Table 4). However, the composition of the stacks is seen to be markedly affected. In the chars from the parent coals, the stack thickness and the number of layers ranged from 1.00 to 1.26 nm and from 3.5 to 4.3, while in chars from the HCl treated coals did 1.25-1.43 nm and 4.3-4.9, respectively. From Table 4, common trends are observed between the parameters of graphite-like stacks and the content of Ca for whole set of char samples: the number of layers and the thickness of the stacks certainly decrease as the content of Ca increases. It means that during coal pyrolysis, calcium species opposed by some way to the stacking of aromatic layers towards the more condenced graphite-like structure. Char gasification. During char gasification, the pressure in the autoclaves at 700 and 0 730 C reached up to 13 MPa due to steam and gases generated. In Table 5 are displayed the gasification data obtained at 7000C. The extents of char burn-off on a dry ash-free basis range from 18 to 32 wt. % and the yield of gaseous products from 23 to 61 wt. %, the chars from the HCl treated coals showing much lower reactivity as compared to those from the parent ones.
9
Table 4. X-ray structural data for the graphite-like carbon of chars from different brown coals. Coal char
Ca,
d002,
Number
Lc
Lа
wt. %
A
of layers
А
А
From parent coals Bor1 Bor2 Bor3 Bor4 Bor5 Bor6 Bor7 Ber9
1.26 2.15 2.24 1.74 2.19 1.83 1.30 2.44
0.383 0.393 0.379 0.376 0.378 0.369 0.380 0.377
4.3 3.5 3.9 3.9 3.9 4.2 3.9 3.7
1.26 1.00 1.12 1.08 1.10 1.19 1.09 1.02
1.53 1.60 1.44 1.55 1.47 1.46 1.42 1.65
Kan20 Ya10
1.86 0.30
0.365 0.397
4.2 3.8
1.18 1.13
1.56 1.59
From demineralised coals Bor2D Bor3D Bor5D Bor6D Ber9D
0.01 0.06 0.01 0.02 0.03
0.369 0.416 0.376 0.374 0.364
4.87 4.34 4.38 4.33 4.87
1.43 1.39 1.27 1.25 1.41
1.52 1.54 1.45 1.41 1.83
2-4 fold differences in the values of reactivity showed the o
70
Gas yield, wt.%
60
700 C o 700 C o 730 C o 730 C
50 40 30 Gas yield, %= - 16.6 +2.5 Burn-off R=0.92
20 20
25
30
35
Burn-off, %
Figure 2. The correlation between the gas yield and burnoff extent during gasification at 7000C and 7300C of the chars from the native coals (dark symbols) and HCl treated ones (open symbols).
samples from the same Borodino deposit. The increase in the temperature to 7300C results in the increase in char reactivity. From Fig.2, it can be seen that good linear correlation is observed between the extents of burn-off and the yield of gases obtained at 700 and 7300C with a correlation coefficient of 0.92. It is of importance to clarify the reason for such large variation in gasification reactivity of different char samples under the same conditions. To be particularly considered are the following factors most likely affecting the reactivity:
porous characteristics of char (surface area and the size of pores), its crystalline structure, and mineral matter effect (by catalysis of the reactions). Data on nitrogen adsorption showed all the char samples to have predominantly microporous structure and to range with the surface area from 42 to 419 m2/g (Table 5). In general, Table 5 shows a 16 fold difference in the values of gas yield calculated per unit surface area of chars (from 0.08 to 1.29 wt. % per m2). On the other hand, the values of the same order
10
Table 5. Steam gasification of chars from different coals at 7000C. Coal char
Ad,
Ca,
Char surface 2
Burn-off,
Gas yield,
Gas yield,
wt. %
wt. % db
area, m /g
wt. %
wt. % daf
wt.%/m2
From parent coals Bor1 Bor2 Bor3 Bor4 Bor5 Bor6 Bor7 Bor8 Ber9 Kan20 Yal0
6.8 11.2 12.9 8.7 7.4 8.6 6.8 8.1 7.5 14.6 2.8
1.24 2.15 2.24 1.74 2.20 1.83 1.39 1.23 2.44 1.83 0.30
130 199 56 108 130 70 244 241 167 42 419
27.4 24.1 30.0 31.8 28.0 29.1 26.5 26.3 28.1 26.9 20.2
44.3 48.6 51.1 61.0 54.2 53.4 54.5 54.0 52.0 54.4 32.4
0.34 0.24 0.91 0.56 0.42 0.76 0.22 0.22 0.31 1.29 0.08
From HCl treated coals Bor2D Bor3D Bor5D Bor6D Ber9D
5.2 7.6 1.4 2.1 1.4
0.01 0.06 0.01 0.02 0.03
114 112 154 217
20.0 18.5 18.4 20.9 19.3
31.2 23.3 28.8 34.5 33.5
0.27 0.26 0.22 0.15
(from 0.15 to 0.27) are observed for HCl treated chars. Strangely enough, for a set of char from the native coals, the gas yields and extents of burn-off tend to decrease as the surface areas of chars increase. Thus, Yallourn char with the highest surface area and pore volume proved to exhibit least reactivity as compared to other chars with much lower surface areas and pore volumes. It follows that the factor controlling the gasification reaction rate is not porous structure, and discrepancy arises from some other properties of coal chars employed. It must be born in mind that not all surface carbon atoms of chars are known [48] to have the same reactivity for gasification because of structural defects and strain energy differences within the defective carbon network. Carbon atoms out of the crystalline component have higher reactivity than those in the crystalline component. It allows the gasification process to be divided in general into two steps. Firstly, carbon atoms out of the crystalline component is preferentially oxidized, and the removal of them introduces micropores among crystallites composed of stacking aromatic layers. Secondly, these crystallites are gasified to form pore networks including mesoand macropores in the texture.
11
From comparison of the data on gasification reactivity in Table 5 (the gas yields and extents of burn-of) with the crystallographic parameters of carbon matter in Table 4 (amounts of graphite-like and nonordered γ-components) it appears that least reactive chars were those which had the higher amounts of amorphous carbon (γ-component). Less ordered chars from the demineralized coals proved to show low reactivities as compared to the parent chars with the predominantly graphite-like structure. Thus, general regularity mentioned above lacked confirmation with the chars tested. Many authors have shown [5, 6, 12] that K, Na, Ca, Mg, Fe, Ni and other alkali, alkalineearth species both added to coal and naturally occurring in the coal can catalyze the reaction of carbon gasification with steam, oxygen and CO2. The X-ray diffraction patterns of the chars showed mainly quartz 1 1- -SiO ; 2- -Fe O ; 3- CaSO ; 4- CaCO ; 5- CaO; with different quantities to be 6- Ca(OH) ; 7- Ca Al O * CaSO ; 8- FeFe O (MgFe O ) present in all the samples. No 400 Demineralized 1 Borodinsky Bor D distinct reflections related to other 1 1 2 21 11 1 1 1 1 3 2
2
3
2
3
6
12
4
3
4
2
4
2
4
5
1
2
300
6 1 4 200
7
100
0
31 4
3
2 35 6 1 5 4 5 5
10
20
5
5 3
83
30
63
3 8
75
454
Borodinsky Ber5
5
Berezovsky Ber9 3 3 5
7
8 3 33 40
3
8
8 33 3
50
8 Yallourn Ya
8
38 60
70
2
Figure 3. X-ray diffraction patterns of the chars from the demineralized and parent coals.
10
inorganic species were reliably detected because of low content and probably high degree of dispersion. On the other hand, X-ray diffraction patterns of the coal chars after ashing showed rather complicated mixture of crystalline substances. Figure 3 shows, as examples, X-ray diffraction patterns for ashes from some parent coals and from HCl treated one. The following main typical minerals were detected in the
ashes from Kansk-Achinsk and also in Kangalas coals: quartz -SiO2, gypsum CaSO4, CaO, calcite CaCO3, Ca3Al6O12-CaSO4, Ca(OH)2, αˉFe2O3 and much less amount of any other minerals. Mainly CaSO4 and FeFe2O4 (or may be MgFe2O4), and much less content of CaCO3 and CaO minerals were found in Yallourn coal ash. The prevalence of CaSO4 in Yallourn ash can be due to the presence of sulphur in excess proportion to Ca to bond almost all amount of Ca with stable CaSO4 form. On the other hand, the ash from the 0.1 N HCl treated coals consisted of mainly SiO2, while -Fe2O3 occurred in a smaller amount, and only traces of CaSO4 were detected in some ashes (Fig.3). The data in Table 5 show poor correlation between the char reactivity and total ash content. However, fairly good correlations with the coefficients of determination R2=0.91 and 0.94 are observed in Fig.4 where the values for burn-off and gaseous yields are plotted against the content of calcium only. The results obtained suggest that the reactivity of brown coal chars for steam gasification reaction seems to be controlled by the catalytic activity of calcium species (CaCO3
12
and/or CaO) that were formed from the decomposition of naturally occurring calcium compounds during pyrolysis and gasification reactions. The question may be raised as to why some deviations from the curves occur in Fig.4. Probably they could result from a number of reasons, such as nonuniform distribution of catalytically active calcium species (CaO/CaCO3) within the char, difference in the kinds of forms of calcium species due to interaction with other mineral matter with the formation of less catalytically activite calcium sulfate or even inert silicates, and also may be due to the effect of other catalytically active metal species, iron in the first place.
Gas yield (burn-off), wt.%
60
50
Gas yield, wt = 30,0+27,3 Ca -7,4Ca2 R2= 0,94; SD=3,1
40
30
Burn-off,%= 19.0+9.1Ca-2.1Ca 2 R =0.91; SD=1.47
20 0,0
0,5
1,0
1,5
2,0
2
2,5
Ca, wt.%
Figure 4. The extent of burn-off and the yield of gases dependant on the calcium content in the chars. Dark symbols - the chars from parent coals; open symbols - HCl treated coals.
The gaseous products obtained from the reaction of char gasification consisted of mainly CO2 (32-37 mol %), H2 (20-29 mol %), and CH4 (31-41 mol %) with least amount of CO (1-2 %). Fig 5 shows the yields of CO2 to correlate well with Ca content (determination coefficient R2=0.87): Yield of CO2 (mmol/g) = 3.06+1.33 Ca – 0.26 Ca2; R2=0.87; SD=0.31, where SD is the standard deviation of the regression. One can see in Fig 5 that the yield of H2 also increases with the increase in calcium content closely resembling the profile for CO2 yield. It seems reasonable to suggest that CO2 and H2 may result from the same gross reaction which can be catalyzed by calcium species: СН4 СО2 Н2 СО
Gas yield, mmol/g
5
carbon-steam reaction: CO shift reaction: gross reaction:
СО2 СН4
4
Н2
3 2
CO2 yield (mmol/g) = 3,06+1,33 Са - 0,26 Са 2
R =0,87;
1
H2 yield (mmol/g)= 2,12 + 1.22 Са -0,25 Са R =-0,69;
0,0
0,5
Nevertheless, the actual H2 molar yield is considerably less than that of CO2 and the correlation with Ca is weaker (R2=0.69):
SD=0,31
2
0
2
C+ H2O= CO+H2; CO+H2O= CO2+H2; C+2H2O= CO2+2H2.
SD=0,48
1,0
1,5
2
Yield of H2 (mmol/g) = 2.12 + 1.22 Ca – 0.25 Ca2.
CO 2,0
2,5
Ca, wt. %
Figure 5. The yields of CO2, CH4 and H2 gases dependant on the calcium content in the char.
The lack of correlation may results from a partial consumption of H2 produced in above reaction by the following methanation reaction, calcium species having no relevance to methanation because of too unsatisfactory correlation (R2=0.36). It is seen in
13
o
700
730 C o 700 C o 700 C
Surface area, m 2/g
600 500 400
Gasified chars Yallourn
300 200 Chars
100 0
5
10
15
20
25
30
Burn-off, %
Figure 6. The change in the surface area of gasified chars from different coals as a function of the extent of burn-off at 700 and 730oC.
35
Fig. 5 that large scatter for methane yields from different chars with no distinct trend is observed, partially demineralised chars producing essentially same yield of methane as the chars from the native coals. It follows from this observation that gasification to methane seems to be a non-catalytic reaction, controlled mainly by the intrinsic reactivity of the carbon. However, the contribution of the catalysis by Fe species does not appear to be excluded [49]. Partial gasification of coal chars resulted in the development of surface area of carbons from 50-270 up to 600-700 m2/g on reaching 30-35 % of burn-off (Fig. 6). The carbons from the native coals had, as a rule, higher surface areas (because of higher burn-off) as compared to that from HCl treated ones. The
development of porous structure in carbons from microporous to mainly mesoporous occured as the concentration of Ca increased (i.e. as burn-off increased) (Fig.7). From the nitrogen adsorption isotherms in Fig 8, it is seen that steam gasification under the same conditions resulted in the formation of mesopores in the chars enriched with Ca and of micropores predominantly in the chars from the corresponding HCl treated ones with no ion-exchangeable cations.
1 0,2
200 cm3/g
3
cm /g
Micropores
0,1
0,0
0,0
0,5
1,0
2
150
Mesopores
1,5
2,0
2,5
Ca, wt.%
Figure 7. The volumes of micro- and mesopores for the chars gasified at 700o C as a function of calcium content.
100 0,0
0,2
0,4
0,6
0,8
1,0
P/Po
Figure 8. The isotherms of nitrogen adsorption on the gasified at 700o C chars from the parent (1) and from HCl treated brown coals (2).
It can be concluded that the reactivity of brown coal chars from Kansk-Achinsk, Kangalas and Yallourn coals for steam gasification in the autoclave reactor under high pressure up to 13 MPa and temperature of 700-7300C is controlled by the catalytic activity of calcium species (may
14
be CaO/CaCO3) formed from naturally occurring calcium containing species. The formation of CO2, H2 and CH4 gases occurred by different ways and was affected by different mineral species. The above specific features of the composition allowed Kansk-Achinsk coals to be considered as preferable and environmentally freindly raw material naturally charged with species catalytically active for gasification and activated carbon production. Acknowledgements This work was supported in part by the Siberian Branch of the Russian Academy of Sciences (Integration Program, Project no. 104). References [1] Minchener A. Fuel 2005; 84: 2222. [2] Sharma A., Takanohashi T., Morishita K., Takarada T., Saito I. Fuel 2008; 87: 491-497. [3] Domazetis G., Barilla P., James B.D., Glaisher R. Fuel. Proc.Technology 2008; 89:68-76. [4] Stiegel G.J., Maxwell R.C. Fuel Proc. Technology 2001; 71: 79-97. [5] Miura K., Hashimoto K., Silveston P.L. Fuel 1989; 68 (11):1461-1467. [6] Lemaignen L., Zhuo Y., Reed G.P., Dugwell D.R., Kandiyoty R. Fuel 2002; 81: 315-326. [7] Collot A.-G. International Journal of Coal Geology 2006; 65: 191-212. [8] Ismail I.M.K., Walker P.L. Fuel 1989; 68: 1456. [9] Linares-Solano A., Hippo E.J., Walker P.L. Fuel 1986; 65: 776. [10] Takeda S., Kitano K., Kubota J. et al. J. Fuel Soc. Japan 1985; 64: 409. [11] Takarada T., Tamai Y., Tomita A. Fuel 1985; 64:1438. [12] Hashimoto K., Miura K., Ueda T. Fuel 1986; 65: P.1516. [13] Ohtsuka Y., Tomita A. Fuel 1986; 65(12): 1653-1657. [14] Sneth A., Sastry C. et al. Fuel 2004; 83: 557-572. [15] Li C.-Z. Fuel 2007; 86: 1664–1683. [16] Moulijin J., Kapteijn F. Carbon 1995; 33: 1155. [17] Kapteijn F., Porre H., Moulijn J.A. AIChE J. 1986; 32: 691. [18] Walker P.L., Matsumoto S., Hanzawa T., Miura T., Ismail I.M. Fuel 1983; 62:140. [19] Liu Z.-I, Zhu H.-H. Fuel 1986; 65:1334-1338. [20] Li S., Cheng Y. Fuel 1995; 74: 456-458. [21] Kepteijn F., Porre H., Moulijin J.A. AIChE J. 1986; 32: 691. [22] Hippo E.J., Jenkins R.G., Walker P.L. Fuel 1979; 58: 338. [23] Kapteijn F., Porre H., Moulijn J.A. AIChE J. 1986; 32: 691. [24] Bayarsaikhana B., Hayashia J.-I., Shimada T., Sathea C., Lib C.-Z., Tsutsumic A, Chiba T. Fuel 2005; 84: 1612–1621. [25] Takarada T., Tamai Y., Tomita A. Fuel 1986; 65(5):679-683. [26] Kyotani T., Karasawa S., Tomita A. Fuel 1986; 65:1466-1469. [27] Tomkow K., Sieminiewska T., Jankowska A., Broniek E., Jasienko M. Fuel 1986; 65: 1423-1428.
15
[28] Samaras P., Diamadopoulos E., Sakellaropoulos G.P. Fuel 1996; 75(9): 1108-1114. [29] Tromp P.J., Kaptejin F., Moulijin J.A. Nato Adv. Sci. Inst. Ser. C. 1992; 370: 75. [30] Matsuoka K., Yamashita T., Kuramoto K., Akira Tomita. Fuel 2008; 87: 885-893. [31] Hippo E., Walker P.L. Fuel 1975; 54: 245. [32] Linares-Solano. Fuel 1986; 65(10):1345. [33] Fujita H., Uba M., Nishida S. Fuel Soc. Japan 1982; 61: 301. [34] Kasaoka S., Sakata Y., Shimada M. Fuel 1987; 66(5) 697-701. [35] Li Ch-Zh. International Coal Science and Technology Conference, Okonawa, October 2005. [36] Kuznetsov P.N., Kuznetsova L.I., Kutikhina E.A. Khim. Tverd. Topl. (Moscow) 2008; 3: 36. [37] Sheng C. Fuel 2007; 86:2316-2324. [38] Beesting M., Haetwell R.R., Wilkinson H.C. Fuel 1977; 59: 319. [39] Knight A.T., Sergeant G.D. Proc of 3rd PACHEK, Seul, 1983, V.3, p.236. [40] Molina A., Montoya Mondragon F. Fuel 1999; 78: 971-977. [41] Gavrilin K.V., Ozerskii A.Yu. Kansko-Achinskii ugol’nyi bassein (Kansk-Achinsk Coal Basin), Nedra, Moscow; 1996, 271 p. [42] Birgauz R.Ya., Kukharenko T.A. in Khimiya i pererabotka topliva (Fuel Chemistry and Processing), Izd. IGI, Moscow; 1971, vol. 27, issue 1, p. 3. [43] Kuznetsov P.N. Khim. Tverd. Topl. (Moscow) 1998; 3: 53. [44] Kuznetsov P.N., Kuznetsova L.I., Bimer J. et al. Fuel 1997; 76:. 189. [45] Kuznetsov P.N., Kuznetsova L.I. Khim. Tverd. Topl. 2008; 6: 57-66. [46] Allardice D.J., Clemowb L.M., Jackson W.R. Fuel 2003; 82: 35–40. [47] Saranchuk V.I., Airuni A.T., and Kovalev K.E. Nadmolekulyarnaya organizatsiya, struktura i svoistva uglya (Supramolecular Organization, Structure, and Properties of Coal), Naukova Dumka, Kiev; 1988. [48] Byrne J.F., Marsh H. Porosity in carbons (editor J.W. Patric), Edward Amold, London; 1995, p.16. [49] Chuang-tao Guo, Li-ming Zhang. Fuel. 1986; 65(10): 1364-1367.
1
STUDY OF THE PROPERTIES OF COAL FROM MONGOLIAN SAIKHAN-OVOO DEPOSIT AND THE CHAR AND CARBONS PRODUCED B.Purevsuren, Institute of Chemistry and Chemical Technology, Mongolian Academy of Sciences, 13330 Peace ave, Bayanzurkh district, Ulaanbataar, Mongolia Fax (976-11) 453133 E-mail: [email protected] Ya.Davaajav, Kh.Serikjan, S.Batbileg, Institute of Chemistry and Chemical Technology, Mongolian Academy of Sciences, 13330 Peace ave, Bayanzurkh district, Ulaanbataar, Mongolia Fax (976-11) 453133 E-mail: [email protected] P.N Kutsnezov Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences 42 K.Marx str, 660049 Krasnoyarsk, Russia Fax: 7(391)2124720; E-mail: [email protected] Abstract—The main performance characteristics of coal from the Saikhan-Ovoo deposit in Mongolia were determined. This coal corresponds to a high metamorphic stage, which corresponds to lean coal transitional to anthracite. It is characterized by high yield of char, high calorific value, and low sulfur content. Iron compounds (60.2 %) predominate in the mineral matter of coal with a small participation of silicon, aluminum, calcium, and other metal compounds. The concentrations of the toxic metals Sr, Mn, Zn, and Cu are no higher than the maximum permissible concentrations. The supramolecular structure of the organic matter of coal mainly consists of closely packed graphite_like clusters containing on average 6.7 aromatic layers. A small fraction of carbon (about 20 %) is structured as γ_components at the periphery of clusters. Active carbon was obtained by thermal steam activation. Its structural characteristics and sorption capacity for iodine and methylene blue were determined. Oil price at the global market increases. Its dynamics is that periods of relatively stable prices are changed by an abrupt increase and then by a dramatic decrease. Because of instability on world oil market, the diversification of energy carriers is practically implemented in many countries with the involvement of various nontraditional types of organic raw materials, primarily, coal, whose reserves are much greater than oil and gas reserves [1, 2]. In Mongolia, coal is currently the main energy carrier for thermal power plants and local boiler houses [3, 4]. This country possesses large reserves of various fossil solid fuels; it is among the first 15 world countries in terms of geologically expected reserves [5, 6]. About 140 coal deposits and Carboniferous, Permian, Jurassic, and Cretaceous coal formations are well known. The Tavantolgoi deposit of high_quality coking coals; the Shariingol, Nalaikh, and Alagtogoo deposits containing anthracite; and the Baganuur and Shivee-Ovoo brown coal deposits belong to the large best studied and industrially important
2
deposits. However, in general, the coal provinces have not been adequately explored, and the reserves can be considerably increased in the future. The direct use of coal for combustion at heat and power plants and in municipal and domestic furnaces causes environmental problems, in particular, those related to greenhouse gas release, soil and air pollution due to harmful emissions. Thus, in Mongolia, attention is focused on the production of high-quality and environmentally safe energy carriers from coals. For this purpose, modern advanced technologies are developed and adapted for the efficient and environmentally appropriate use of domestic coals. Scientific research works on coal chemistry in Mongolia are mainly performed at the Institute of Chemistry and Chemical Technology, Mongolian Academy of Sciences. The main research areas are related to the acquisition of new data concerning the composition of coals from the largest deposits, the technical classification of these coals, and the development of efficient methods for coal processing [7]. The majority of studies are performed according to international collaboration programs. In particular, the coking properties of coal samples from the Tavan-tolgoi deposit were studied in cooperation with Czech researchers in order to produce high-quality coke [8, 9]. The technological properties of brown coals from the Bayanteeg, Baganuur, and Ovdogkhudag deposits were studied in pyrolysis and hydrogenation processes in cooperation with researchers from the Institute for Fossil Fuels (Moscow) [10]. Brown coals from the Ovdogkhudag and Aduunchuluun deposits were studied in liquefaction processes in cooperation with Japanese researchers [11]. Coals from the Shivee-Ovoo and Baganuur deposits were tested in Japan at a plant under liquid-phase hydrofining conditions in a solvent at 380–440°C without the use of hydrogen [12]. Coals from the Shivee-Ovoo and Baganuur deposits were also studied in gasification and pyrolysis processes in Germany [13] and the United Kingdom [14], respectively. The composition and properties of products obtained upon conversion by the above methods (coke, char, smokeless solid fuel, activated carbons, liquid substances, and gases) were characterized with the use of currently available analytical techniques and instrumentation. In this work, we report the results of a study on the composition and properties of coal from the Saikhan-Ovoo deposit in the course of pyrolysis and sorbent production. The deposit is located in the Bulgan province in northern Mongolia; it has almost not been mined up to date. EXPERIMENTAL Coal sampling at the deposit and proximate analysis were performed in accordance with procedures specified by Mongolian National Standards (UST) (moisture content, UST 656-79; ash content, UST 652-79; yield of volatile substances, UST 654-79; calorific value, UST 669-87; and sulfur content, UST 895-79).
3
The experiments on coal pyrolysis were performed in a laboratory vertical tube reactor of stainless steel with loading to 1 kg of coal with a particle size of 3–4 mm. The reactor was heated with an electric furnace; the heating rate was 20 K/min. Temperature control was performed using a Chromel–Alumel thermocouple arranged in a thermocouple well within a coal bed. The vapor–gas reaction products arrived at a water-cooled glass condenser, and the condensate (tar and pyrogenetic water) were collected in a special receiver. The yields of solid residue (char), tar, and pyrogenetic water were determined by gravimetry, and the yield of gases was found by difference. The resulting char produced was activated in a quartz tube reactor at 850°C in a flow of steam. The sample weight of carbonization products (particle size of 0.6–1.0 mm) was 10 g. The composition of the organic matter of coal was determined by elemental analysis and IR spectroscopy; the IR spectra were measured on a Vector 22 Fourier transform IR spectrometer from Bruker. The concentrations of metals in ash were determined by X-ray fluorescence analysis. The spatial supramolecular structure parameters of coal were studied by X-ray diffraction analysis. The X-ray diffraction patterns were recorded on a DRON-3 diffractometer using СuКα radiation in the 2Θ angle range from 5 to 60°. The computer processing of the X-ray diffraction patterns and the calculation of X-ray structure parameters were performed with consideration for recommendations [15–17]. Specific surface areas were measured using the thermal desorption of nitrogen, and sorption capacities were determined from the uptake of methylene blue or iodine from aqueous solutions. RESULTS AND DISCUSSION The test coal has the following performance characteristics: Ad = 10.6%, Wr = 2.24%, Vdaf = 8.2%, and a calorific value of 33.8 MJ/kg. Incineration ash is brown and mainly contains iron compounds (60.2% Fe2O3); the concentrations of silicon (17.1% SiO2), aluminum (8.2% Al2O3), and calcium (4.8% CaO) compounds are much lower (Table 1). The ratio Fe2O3 + CaO + MgO + Na2O + K2O)/(SiO2 + AI2O3+TiO2) >> 1 suggests the basic character of ash. The concentration of phosphorus compounds is 3.1%, which is unfavorable for the use of coal in the production of metallurgical coke. Table 1. Chemical composition of coal ash from the Saikhan-Ovoo deposit In ash, wt %
In coal, g/t
SiO2 Al2O3 Fe2O3 CaO
MgO TiO2 SO3
P2O5 Na2O K2O
Sr
17.1
2.5
3.1
247 316 25 130* 100* 8*
8.2
60.2
4.8
0.5
2.1
0
0.4
Note: average concentrations in world coals [18] are marked with asterisks.
Mn
Сu
Zn 225 18*
4
The concentrations of trace elements are noticeably higher than Clarke numbers (that is, the average concentrations in world coals) for toxic elements such as Sr, Mn, Сu, and Zn (by factors of 2–3). Nevertheless, their concentrations, as well as the concentrations of other elements, are no higher than the maximum permissible concentrations (MPCs). Based on the experimental data with consideration for the results of correlation analysis performed by Baryshev [19], the test coal ash is classified with medium-melting ashes; its initial softening temperature is 1250–1350°C, initial fluidity temperature is 1270–1370°C, and normal liquid slag removal temperature is 1320–1420°C. The composition of the organic matter of coal (OMC) is characterized by low hydrogen and sulfur contents (with the predominance of organic sulfur Sorg) (Table 2). Table 2. Elemental composition of the organic matter of coal (wt %) Cdaf
85.90
Hdaf
2.79
Sdaf
0.39
О + Ndaf
10.92*
Sulfur species, Sd total
Sorg
Ssulf
S SO
0.35
0.32
0.02
0.01
4
* By difference. The molecular composition of OMC was characterized by IR spectroscopy. In the IR spectrum shown in the figure, the following absorption frequency regions can be recognized: 700–900 cm–1 (out_of_plane deformation vibrations of Car–H bonds), 1000–1300 cm–1 (vibrations of bonds in various oxygen_containing groups), 1350–1470 cm–1 (deformation vibrations of methyl and
5
methylene groups), 1500–1630 cm–1 (skeletal vibrations of aromatic rings, C=O bonds in ketones, aldehydes, and quinones), 1650–1750 cm–1 (carbonyl groups in carboxylic acids, esters, and lactones), 2800–2950 cm–1 (stretching vibrations of C–H bonds in saturated aliphatic structures), 3030–3050 cm–1 (stretching vibrations of C–H bonds in aromatic and unsaturated compounds, and 3100–3600 cm–1 (stretching vibrations of hydroxyl groups in phenols and carboxylic acids). An analysis of the profile of bands in the range of 1500–1620 cm–1 suggests the presence of polycyclic aromatic structures and quinoid and carbonyl groups in coal. Aromatic rings exhibit a considerable high degree of substitution by alkyl groups, as evidenced by the following spectroscopic characteristics: weak absorption intensity at 3030 cm–1 with high intensities of absorption bands at 872 and 798 cm–1, as compared with a band at 750 cm–1, and comparatively intense absorption bands due to the stretching and deformation C–H vibrations of aliphatic groups (bands at 2853–2922 and 1456 cm–1, respectively). A band due to deformation vibrations in methyl groups (1375 cm–1) is very weakly pronounced in the spectrum. In the presence of intense absorption at 1456 cm–1, this fact suggests that methylene alkyl groups are the main constituents of aliphatic structures in coal. Oxygen-containing groups are detected by comparatively intense absorption bands at 3100–3600 and 1000–1250 cm–1, as well as at 1650–1760 cm–1, due to hydrogen bonded phenols, carboxylic acids, ethers, esters, and lactones. The low yield of volatile substances suggests a comparatively high thermal stability of the coal: upon pyrolysis at 600–650°C, the yield of tar + pyrogenetic water was only 5.3%; the yields of gas and char were 3.9 and 90.8%, respectively. In this case, sulfur (mainly as organic species) was not converted into volatile products; it remained in a bound form in char (probably, as a constituent of iron sulfide). The char produced was subjected to steam activation at 850°C. Table 3 summarizes comparative data on the X-ray diffraction characteristics of the parent coal and the carbonization products after steam activation. These data indicate that the organic matter of the parent coal contains only a small amount of low ordered γ-components (23%). The main fraction is structured Table 3. Structure parameters of the parent coal and coal after steam activation at 850°C for 1 h Sample
Parent coal Activated carbon
Specific surface area, m2/g 455
Structural component fraction, %
Graphite_like component parameters
graphite-like
γ1
γ2
d002
n
Lc
La
Сg
77 80
12 14
11 6
0.345 0.360
6.7 4.8
1.95 1.35
2.43 2.57
1.95 0.54
6
as an ordered phase as graphite-like clusters (with the thickness Lc = 1.95 nm), which consist of closely packed layers of planar polycondensed aromatic molecules (on average, 6.7 layers in a cluster) with the interlayer spacing d002 = 0.345 nm. The graphitization coefficient Сg is 1.95, as determined from the equation [15] Сg = 10–2Lc/(d002 – 0.335). Steam activation resulted in the decomposition of the least ordered γ2-component. In this case, the fraction of the graphite-like component somewhat increased and its structural characteristics changed: the interlayer distance d002 increased to 0.360 nm; the number n of layers in clusters decreased to 4.8, and the layer thickness Lc decreased to 1.35 nm; the graphitization coefficient considerably (by a factor of almost 4) decreased (to 0.54). In this case, the size of clusters Lа changed only slightly (2.43–2.57 nm). The above changes in structure parameters suggest a loosening of the OMC structure in the course of activation; as a result of this, a sufficiently high specific surface area is reached (455 m2/g). The char exhibited an insignificant sorption activity in the absorption of iodine (1.26%), and it did almost not sorb the bulky molecules of methylene blue (Table 4). After steam activation at 850°C, sorption capacity for iodine and methylene blue increased to 30.2% and 12.1 mg/g, respectively. Table 4. Sorption capacity for iodine and methylene blue of char and steam activated char (850°C, 1 h). Sample Char Activated char
Iodine, %
Methylene blue, mg/g
1.3 30.2
no sorption 12.1
CONCLUSIONS (1) It was found that coal from the Saikhan-Ovoo deposit in Mongolia corresponds to a high metamorphic rank, which corresponds to lean coal transitional to anthracite. It belongs to low-ash coals with a low sulfur content (0.35%), a low yield of volatile substances (8.2%), and a high calorific value (33.8 MJ/kg). (2) The supramolecular structure of organic matter mainly consists of ordered and closely packed graphite-like clusters containing, on average, 6.7 layers. A small fraction of carbon (23%) is structured as two γ-components at the periphery of clusters. (3) The mineral matter of coal mainly contains iron compounds (60.2%) with a small participation of silicon, aluminum, calcium and other metal compounds.
7
The concentrations of toxic elements such as Sr, Mn, Zn, and Cu are no higher than regulated MPCs. Coal ash is classified with medium_melting ashes with an initial softening temperature of 1250–1350°C. (4) Coal can be used as smokeless low-sulfur high calorific fuel. Because of its high thermal stability, coal should be combusted at an elevated temperature to prevent underburning (this is most recommended for boilers at heat stations). (5) Sorption materials with developed specific surface areas and sorption activity can also be prepared by using thermal steam activation procedure. ACKNOWLEDGMENTS This work was supported in part by the Siberian Branch of the Russian Academy of Sciences (integration program, project no. 104). We are grateful to Dr. N. Pavlenko (Institute of Chemistry and Chemical Technology, Siberian Branch, Russian Academy of Sciences) for measuring the IR spectrum. REFERENCES [1] Kandiyoti R., Herod A., Bartle K. Solid Fuels and Heavy Hydrocarbons Liquids, Elsevier, London; 2006. [2] Pashkov G.L., Kuznetsov P.N., Safronov A.F., Lyakhov N.Z. Khim. Interesakh Ustoich. Razvit. 2009; 17: 215. [3] Ochirbat P. et al. Mongolian Coal Industry in XX Century, Ulaanbaatar, 2002. [4] Brochure of Ministry of Fuel and Energy, Mongolia, 2006. [5] Tumurbaatar Z. Coal Sector of Mongolia: Book of Ministry of Fuel and Energy, Artcof’, Ulaanbaatar; 2000. [6] Purevsuren B. Abstracts of Papers, Second Korean and Mongolian Energy Conference, Yonsei Univ., Seoul; 2007, p. 13. [7] Dugarjav J., Chem. Sci. 2006; 81: 179. [8] Munkhjargal Sh. PhD Dissertation, Prague Univ., 1985. [9] Dugarjav J., Rep. Inst. Chem. Chem. Technol. Mongolian Acad. Sci., Ikh Biligt, Ulaanbaatar; 1995, p. 17. [10] Tsedevsuren Ts., Titova T.A., Dembovskaya E.A., Khim. Tverd. Topl. (Moscow) 1981; 6: 17. [11] Ueda Sh. 1994 Annual Summary of Coal Liquefaction and Gasification, Japan, 1995, vol. 19, p. 140. [12] Avid B., Sato Y., Maruyama K., Yamada Y. et al. Fuel Proc. Technol. 2004; 85: 933. [13] Avid B., Purevsuren B. et al. J. Thermal Anal. Calorimetry 2002; 68: 877. [14] Avid B., Purevsuren B., Paterson N. et al. Fuel 2004; 83: 1105. [15] Korolev Yu.M., Gagarin S.G. Koks Khim 2003; 3: 8.
8
[16] Skripchenko G.B. Khim. Tverd. Topl. (Moscow) 1994; 6: 16. [17] Saranchuk V.I., Airuni A.T., Kovalev K.E. Nadmolekulyarnaya organizatsiya, struktura i svoistva uglya (Supramolecular Organization, Structure, and Properties of Coal), Naukova Dumka, Kiev; 1988. [18] Tsennye i toksichnye elementy v tovarnykh uglyakh Rossii (Valuable and Toxic Elements in the Commercial Coals of Russia), Nedra, Moscow; 1996, p. 238. [19] Baryshev V.I. Khim. Tverd. Topl. (Moscow) 1979; 5: 81.
Research on the Preparation of Thermal Conductivity C/C Composites by One-step Hot Press Molding Jin Minglin1,*, Zhang YanWen1,2 ,Zhou XiaoLong 2, Cheng Qingzhong1 (1 Department of Material Engineering; Shanghai Institute of Technology; Shanghai 200235;2 College of Chemistry and Chemical Engineering;East China University of Science and Technology;Shanghai 200237; China)
Abstract: Along with unceasing development of aerospace technical and national defence industry,introducing C/C composites as thermal conductivity of materials can meet the demands of new technology.Compared with traditional metal materials,C/C composites have the features of light unit weight,high thermal conductivity,corrosion-proofed.This study focused on preparation of C/C composites with SCFs and thermoplastic mesophase pitch by hot press molding and discussion about the influence of SCFs’ content and distributed orientation on two-dimensional thermal conductivity and bulk density.The results showed that when TI yield level of thermoplastic pitchs was 82.1% and SCFs’ mass fraction was 7%,the hot pressed sample had high bulk density.By analysis of SEM morphology,the fiber’s distributed orientation was almost perpendicular to hot pressing direction and ordered arrangement of toothlike pitch structure was visibly found after etched with potassium dichromate,which led to thermal conductivity of longitudinal section at 127.51W/(m.K).Reversely,increasing overmany SCFs resulted in poor ordering degree and decreasing graphitization degree of matrix carbon. Key words: high thermal conductivity; mesophase pitch; short carbon fiber; C/C composites
1. INTRODUCTION Along with unceasing consummation and development of aerospace technical and national defence industry,we are in great need of fulfiling the requirements of the technology and equipment at the front.The properties of materials adapting to some special application are good strength at high temperature,high thermal conductivity,excellent corrosion resistance, good erosion resistance and so on.Therefore,the research on thermal conductivity C/C composites has arisen worldwide concerns:C/C composites of brake discs used in aeroplanes perform good frictional and wear proerties;nuclear fusion devices releasing much heat in the process have put forward higher requirement for the materials;The development of C/C composites breaks down the restriction that laptop computers select aluminum as the heat sink in order to lessen the corrosion damage.C/C composites are three dimensional multi-phases materials and have complex microcosmic structure which has multiple fractures and holes. The interface between the fiber and its surrounding matrix is found to contain a lot of debonders. The thermal conductivity of C/C composites is not only relevant to Supported by the Shanghai Committee of Science and Technology, China (No.09520500700) *Corresponding author: Fax:+86-21-64085150 E-mail addredd:[email protected](M.L.Jin)
fiber orientation distribution and variety of matrix carbon but also relating to some factors such as carrier-level of the binders,the volume fraction of SCFs,the density of materials and degree of graphitization.C/C composites are generally divided into discontinuous and continuous fiber-reinforced structure. This study focused on the matrix conductive composites made from of as the filler. This thesis aimed at preparing thermal conductivity C/C composites with SCFs and thermoplastic mesophase pitch by molding pressed. Analysis was performed on effects of the mass fraction of SCFs,fiber distributed orientation and density of C/C composites on thermal conductivity by means of rapid solidification technology.The purpose is to explore the influence of SCFs’ content and distributed orientation on two-dimensional thermal conductivity.
2 Experiment 2.1 Raw material The properties of coal tar pitch were shown in Table 1 Table 1 Properties of coal tar pitch Softening point
Vf
TI
/℃
/%
/wt%
28.
67.7
281.8
83
TS /w%
32.32
8
Short chopped carbon fibers(length 0-5mm)were shown in Fig 1:
Fig 1 SEM photos of short carbon fibers
2.2 Preparation of materials for hot pressing Mesophase pitch was synthesized from coal tar pitch in 12 tubes ( φ 20 × 200mm) which was placed one-time to a multi-tube,well-crucible stove with programmable temperature control. The temperature program comprised two steps: RT to 410℃,heating rate of 5℃/ min,and took samples under constant temperature in 410℃respectively at 3h,6h,9h.Then put synthetic mesophase pitch and SCFs into kneading pan,agitated the mixture with electric heating under certain temperature and mashed the mixture after cooling off. The procedures were repeated at least five times. 2.3 Hot press molding for the ordered growth
The mixture was compacted in a mould by a hot-pressing apparatus(RT~560,20.05MPa). The process of hot press molding: the mesophase pitch containing different mass fraction of SCFS were molded respectively at 25℃to 560℃under 20.05 MPa. zation 2.4 Carbonization and Graphiti raphitization The samples which were placed into a porcelain boat were carbonized in the tube furnace. The temperature program is as follows: first the room temperature was raised to 650℃at a rate of 2℃/min ,then programmed to rise to 800℃ at 1℃/min,finally raised to 1000℃at 2℃/min and held for 1 hours. The nitrogen atmosphere was used in the whole process. The samples were graphitized in the graphitizing furnace.The temperature program is as follows:first the room temperature was raised to 1000℃ at a rate of 15℃/min,then programmed to rise to 2700℃at 26℃/min,finally raised to 2900℃ and held for 1 hours. The atmosphere of high purity argon gas was used in the whole process. 5 Testing of properties 2. 2.5 5.1 Testing of physical properties 2. 2.5 TS content was measured according to GB/T2292-1997. The microstuture of pitch samples were observed by 500 times under Leica DMR optical microscope. The density of samples was measured according to GB/T 3810.3-2006. Measurements of the thermal diffusivity (a) and specific heat capacity (Cp) of the specimen (φ10×3mm) were carried out by using the laser-flash diffusivity instrument (Nano-Flash-Apparatus,LFA 447,NETZSCH). 5.2 Microstructure analysis 2. 2.5 X-ray diffraction (XRD) with a Cu Ka radiation (k = 0.15406 nm) was used for determining the primary structure and parameter of the graphite crystal: d002,where d002 was the graphite interlayer distance.The degree of graphitization (g) was calculated by the expression g = (0.3440-d002)/(0.3440-0.3354).Fractured micrographs of graphite blocks were observed through a field emission scanning electron microanalyser (FESEM,JEOL JSM-6360F,Japan).
3. RESULTS AND DISCUSSIONS 3.1 Analysis of hot pressed samples The two main components of C/C compositesare the matrix carbon as the binder and fibers as the reinforced filler.In this experiment,the growth of mesophase pitch was observed by polarizing (optical)microscope.As shown in Fig1:By further heating,the optical anisotropic mesophase spheres growed and coalesce with each other,finally formed in body mesophase. 0h
3h From the perspective of solubility,the matrix carbon had two phases (solid and liquid) in coexistence.Toluene solubles yield level represented thermoplastic component,which bonded with short carbon fibers in the process of hot press molding. 6h
9h
Fig
2
Polarized-light
microstructure
The Fig3 indicated that on the condition of too much thermoplastic component(TS),the reactant flow was enhanced with temperature increasing and overflowing behavior occurred. However,too little TS resulted in SCFs not being well soaked and local packing. The combining intensity between the matrix carbon was obviously decreased. Therefore the mixture was crushed into powder and hard to be sharped. The study showed that the hot pressed sample had higher apparent density and hardness as TI yield level was 82.10%. TI 67.68(0 h)~TI 71.60(3 h)
TI 82.10(6 h)
TI
86.13(9 h)
Overflow
High
bulk
density
Powder Fig 3 Hot pressed sample and TI
3.2 Effects of carbon fiber content on density of C/Cs The ordered columnar billets were prepared by hot pressing with different mass fraction of SCFs and mesophase pitch thermal-polymerizating for 6h(TI 82.10%).With increased mass fraction of SCFs from 1% to 10%,the bulk density of samples increased. This showed that in the process of hot pressing short carbon fibers as reinforcement are beneficial to the grains’ close packing and filled the pores left by the mesophase after its decomposition and release. However,overmany carbon fibers have no significant effects to increase the bulk density of samples and result in difficult mold processing. When 1.6 MSCF%=13%,mixed powder of sample remained powder and had no access to bulk materials 3
Density g/cm
Apparent Density True Density
1.4
-2
0
2
4
6
SCFm%
8
10
12
14
after process of hot pressing.This indicated that when the matrix carbon as the binder was fixed,with increasing of SCFs content and surface area,it is hard for the matrix carbon reaching enough surface coating of carbon fibers.So the interface between the fibers and its surrounding matrix became larger which had a negative influence on the bulk density of materials.By determining density curve(Fig 4),it was found that the true density increased with the increasing of SCFs content and then decreased.When MSCF%=7%,the maximum true density was 1.731g/cm3. Fig 4 Influence of SCFs on density of the samples
3.3 Analysis of SEM Fig5 was the SEM morphology of the hot pressing sample ( MSCF%=7% ) . As can be seen from interface structure of the longitudinal and cross section,The fiber’s distributed orientation was almost parallel to X-Y section and z axis direction was hot pressing direction. At the beginning of hot pressing,SCFs were arranged irregularly in the matrix carbon,then increasing load led to plane tensile behavior of thermalplastic carbon with SCFs.The plane tensile direction was almost perpendicular to Z axis.The surface of the matrix carbon had some cracks and pores.With increasing temperature and load,a large amount of small molecular gases were produced through series of chemical reactions in the matix carbon leaving some pores.Further increase in load caused the breaking at the interface between the fiber and the matrix,because the fiber had high tensile strength and then crack propagated beyond the fiber. Fig6 were the SEM morphology of the hot pressing sample (MSCF%=7%)after etched with potassium dichromate. The toothlike anisotropic pitchs were visibly found on the cross section after oxidation by potassium dichromate.The ordered arrangement of pitch structure was beneficial to increase thermal conductivity. Nevertheless,the longitudinal section showed that SCFs were coated by the matrix carbon without compact binding between the matrix and the fillers and the surface of the matrix reserved a small amount of anisotropic carbon.
(a) cross section
(b) cross section
(c) longitudinal section
(d)
longitudinal section Fig 5 SEM photos of the hot pressed samples (MSCF %=7%)
(a) cross section
(b) cross section
(c) longitudinal section
(d)
longitudinal section Fig6 SEM photos of the hot pressed samples(MSCF %=7%)after etched with potassium dichromate
3.4 Analysis of XRD With the X-ray diffraction method analysis,the principle of which is to express graphitization degree by analyzing relative difference between the ideal graphite interlayer spacing and the measured graphite interlayer spacing. The ideal graphite interlayer spacing is 0.3354nm,the observed interlayer spacing is an average value depend on the relative ratio of graphitic to turbostratic stacking,the actual investigations decrease gradually to the value of ideal gaphite interlayer spacing by heat treatment. With this theory,graphitization degree of C/C composites are calculated (P): P=
ƛ d 0.344 d 100% ƛ o 0.0086
ƛ d ---- different value between measured graphite interlayer spacing and interlayer spacing for turbostratic stacking
ƛ � ---- different value between ideal graphite interlayer spacing and interlayer spacing for turbostratic stacking Relationship between mass fraction of SCFs and graphitization degree of the materials was shown in Figure 8.The results indicated that with increasing mass fraction of SCFs,downward tendency of graphitization degree of cross section was evident.The fiber’s orientation distribution performed restrictive functions on axial ordered growth of mesophase pitch moleculars; graphitization degree of longitudinal section changed very little and fiber’s distributed orientation had little influence on radial ordered growth of mesophase pitchs,because under axial load when mesophase pitchs extended toward axial direction(Fig 7),the ordered growth was influenced
by ”intercepting function” of high tensile strength SCFs. It resulted in poor ordering degree of axial growth of mesophase pitch and decreasing graphitization degree of cross section;when mesophase pitchs growed toward radial direction,which was vertical to load direction,graphitization degree of longitudinal section was almost unchanged for SCFs’ weak “intercepting function”. axial direction radial direction
72
Degree of Graphization%
71 70 69 68 67 66 65 0
2
4
6
8
10
SCFM%
Fig7 Diagram of the matrix model in hot pressing process
Fig8 Influence of SCFs’ content on graphitization
degree
3.5 Analysis of thermal conductivity The relation curve between thermal conductivity of C/C composites and mass fraction of SCFs was shown in Fig 9:with increasing content of SCFs,there was no significant change in thermal conductivity of cross section.WhenMSCF%=7%,the value reached to 95.92W/(m.K),which indicated that there was no interfacial bonding of SCFs between matrix carbon layers(cross section);The fiber’s distributed orientation had little influence on radial ordered growth of mesophase pitchs and had obvious “bridging effect” between matrix carbon layers(longitudinal section). Therefore,thermal conductivity of longitudinal section increased with increasing content of SCFs,when MSCF%=7%,the maximum value was 127.51W/(m.K).The fiber’s distributed orientation was almost perpendicular to hot pressing direction and unidirectional fiber has very high thermal conductivity. Continuity and connectivity of crystal structure were in good condition because of interfacial bonding of SCFs.So phonon mean free path was long, moreover the conduction of phonon and electron became smooth which led to high thermal diffusivity. However,when MSCF%>7%,the binder carrying excessive amount of SCFs were in super-saturated state.This brought about SCFs’ local packing and stress concentration in materials’main pores,which produced cracks in the direction of principal stress.Consequently the thermal conductivity of materials decreased sharply with conduction disturbance of electron.
micro-scope,motion of carriers affects thermal conductivity values of graphitic samples.when orientation of fiber’s
Thermal Conductivity W/(m.K)
According to the analytical result,it was proved that C/C composites are intermediate materials between turbostratic structure and graphite crystal.High crystallinity and highly-oriented graphite structure always leads to high conductivity values. The thermal conductivity of C/C composites depends on the fibrous structure arranged orderly and axial direction orientation of fiber’s distribution.In the radial direction
distribution is the same as direction of heat conduction,thermal diffusivity path of electrons by the way of carbon length of time consumed is when carriers hitting a hindrance.
120 100 80 60 40 20
is high.The longer mobiling fiber carrier is,the higher
0 0
2
4
6
8
10
SCFm%
Fig9 Variety curve of the thermal conductivity of the samples
4. CONCLUSION (1).By hot press molding,the conclusion was shown that the hot pressed sample had higher strength and hardness as choosing mesophase pitch thermal-polymerizating for 6h(TI 82.10%). When MSCF%=7%,the maximum true density of material was 1.731g/cm3. (2).From the SEM morphology of the hot pressing sample,the fiber’s distributed orientation was almost perpendicular to hot pressing direction and ordered arrangement of toothlike pitch structure was visibly found after etched with potassium dichromate which were beneficial to increase thermal conductivity. (3).Increasing content of SCFs as the fillers had two sides to affect thermal conductivity of cross and longitudinal section.Only thermal conductivity of longitudinal section increased with increasing content of SCFs,when MSCF%=7%,the maximum value was 127.51W/(m.K). Reversely,increasing overmany SCFs resulted in poor ordering degree of axial growth of mesophase pitch and decreasing graphitization degree of matrix carbon. (4).Therefore the important issues for further study are how to change the fiber’s distributed orientation in order to increase thermal conductivity of cross section and how to change the hot pressing temperature for better ordering growth of thermoplastic carbon.
REFERENCES
[1] KATA RZYNA PB,ANDERSON KB,SZYMANSKI.T,et al.Thermal analysis of bulk carbon-carbon composite and friction products derived from it during simulated aircraft braking[J].Carbon.2007,45(3):524~530 [2]Yi Maozhong,Ge Yicheng ,Huang Boyun. Frictional surface properties and rubbing wear mechanism of C/C composites with different structure of matrix carbon[J].China nonferrous metal college journal,2006,16(6):929~936 [3]
Hino
T,Akiba
M.Japanese
development
of
fusion
reaction
plasma
components.Fusion Engineering and Design,2000;49-50:97~105 [4] Murakami M,Nishkin K,Knakam ura K et al.High-quality and highly oriented graphite block from polycomdensation.polymer films.Carbon,1992;30(2);255~262 [5]Guo Quangui,Liu Lang,Song Jinren.Research activities on carbon based materials for plasma facing components of the HT-TV superconducting Tokamak device in china.New carbon materials,2001,16(3):64~68 [6]Lu Shilin,Rand B. Preparation and Structure Characterization of Large Diameter mesophase pitch-based carbon fiber with high thermal conductivity.New carbon materials,2000;15(1):1~5 [7]Zhang Guangjing,Guo Quangui,Liu Zhanjun.Research on thermal conductivity of adulterated graphite.,New carbon materials,2001;16(1):25~28
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: 1. COMBUSTION: NOVEL TECHNOLOGIES (OXYFUEL, CHEMICAL LOOPING, ETC) EXACT TITLE OF PAPER: Comparative Study of Coal Ash and Deposits from Air and Oxy-Fuel Combustion Jost O.L. Wendt, Presidential Professor, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected] Dunxi Yu, Visiting Research Associate, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected] William J. Morris, Ph.D Graduate Student, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected] Andrew Fry, Senior Engineer, Reaction Engineering International, 77 West 200 South, Suite#210, Salt Lake City, UT 84101, USA e-mail: [email protected] Constance L Senior, Manager, Reaction Engineering International, 77 West 200 South, Suite#210, Salt Lake City, UT 84101, USA e-mail: [email protected] Abstract: Oxy-fuel combustion is a promising technology used to realize carbon capture and storage (CCS). To retrofit conventional air-fired power plants to oxy-fuel combustion, the effects of switching from air- to oxy-firing on ash deposition must be well understood. A down-fired oxy-fuel combustor was used to generate bulk ash, size-segregated ash and deposit samples under conditions of: (a) air; (b) 27%O2/73%CO2; (c) 32%O2/68%CO2. Once-through CO2 was used in oxy-fuel cases to simulate cleaned flue gas without moisture or contaminants. The excess oxygen in the flue gas for all combustion cases was fixed at 3%. For bulk ash collection, the combustion products were directed into an isokinetic water-cooled probe with a 90 mm filter housing, where ash was deposited on the filter for subsequent examination. No dilution was used. For size-segregated ash collection, the combustion products were extracted by an isokinetic, nitrogen-quenched and water-cooled sampling probe. The combination of a pre-cyclone and a Berner Low Pressure Impactor (BLPI) was used to obtain full particle size distributions and ash samples for further analyses. For ash deposit collection, a water-cooled probe fitted with a detachable tubular deposition substrate was horizontally fixed at a distance from the burner. The exposure time was about 4-5 hours. The obtained bulk ash, size-segregated ash and deposit samples were subjected to chemical analyses. Characteristic data of ash and deposits are used to identify major constituents that contribute to ash deposition. The effects of switching from air- to oxy-firing on ash partitioning between ash and deposits are investigated.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission PROGRAM TOPIC: (1. COMBUSTION- Industrial Applications, Economics, and Environmental Issues) OXY-FUEL COMBUSTION: A TECHNOLOGICAL OPTION FOR RETROFITTING EXISTING PULVERIZED LIGNITE FIRED POWER PLANTS IN TURKEY İskender Gökalp Director, CNRS-ICARE [email protected] Mücella Ersoy Chief Engineer, Turkish Coal Enterprises [email protected] Abstract: Lignite is one of the major domestic primary energy resources in Turkey. Turkish lignite reserves are estimated to 12 billion tons by 2010. However its quality is quite low and requires promotion of efficient clean coal technologies. Lignite is consumed in three sectors in Turkey: thermal power plants (85%), industry and households. Due to the oil crisis faced during 1970s, lignite gained great importance as a domestic resource and most of the existing pulverized lignite-fired power stations were installed in mid 1980s. In Turkey, the average annual energy consumption increase rate is 7%. 8079 MW of the total Turkish 44000 MW existing installed capacity concern lignite-fired power stations. While the share of lignite in electricity generation reached its highest value of 47% in 1986, this share declined to 15% in 2004 due to the use of natural gas in electricity generation. Over the last four years it has increased again to 21,1% by commissioning of new lignite-fired power plants. According to the Turkish “Electricity Market and Security of Supply Strategy Paper” issued in 2009, all existing proven lignite reserves will be used for electricity generation by the year 2023. Except one power plant based on fluidized bed combustion technology, most of the remaining pulverized lignite-fired power plants in Turkey are old and need to be retrofitted to meet environmental commitments. The thermal efficiencies of existing Turkish lignitefired power plants vary between 28% and 39% in 2008. This paper consists in two parts. The first part will summarize the past and present situation of electricity generation in Turkey using lignite power plants. This will be done by taking into account the fact that Turkey signed the Kyoto Protocol in 2009 and also is a candidate for EU membership and therefore Turkish legislation should be harmonized with EU coal, electricity and environment related legislations. New investments in the existing lignite-fired power plants to comply with EU legislations and Kyoto Protocol commitments are therefore urgently required. In the second part of the paper, retrofitting of the existing Turkish pulverized lignite-fired power plants for oxy-firing will be evaluated. Oxy-fuel combustion is a technology which, if optimized, may significantly contribute to the reduction of CO2 emissions. An important requirement is to demonstrate that this technology can be retrofitted to the many existing power plants already operating in order to prepare for implementation of CCS after 2020. This is necessary to fully evaluate the concept feasibility, address perceived commercial risks, and improve the overall cost and efficiency when the CCS readiness is taken into account. As most work to date on oxy-fuel coal combustion has been applied only at small scale and to bituminous coal, it is necessary to extend the range of fuels and the scale of the technology. Developing oxy-firing for lignite plants should ensure that clean coal technologies can be applied to these plants and should demonstrate the fuel flexibility of lignite, which increases the competitiveness of the technology.
A multi-partnership project is under development in order to retrofit a small lignite-fuelled boiler. A combined RTD/Demonstration approach is developed to validate the feasibility of retrofitting and assessing some innovative options for efficiency improvement of oxy-fuel combustion. RTD goals are to optimize the oxy-fuel combustion process, fuel handling and safety, oxy-fuel burner design, evaluation of retrofitted system, its efficiency and environmental performance, including the quality of the CO2 stream for various CCS options. Demonstration goals include a full retrofit of a pulverized lignite power plant in operation and validation of its scalability. The details of the project will be presented and discussed in the presentation.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11-14, 2010 Abstract Submission PROGRAM TOPIC: COMBUSTION OXY-FUEL COMBUSTION CHEMISTRY – IMPLICATIONS ON CORROSION RELATED ISSUES Klas Andersson, Daniel Fleig, Fredrik Normann, Filip Johnsson Division of Energy Technology, Department of Energy and Environment Chalmers University of Technology, SE-412 96, Göteborg E-mail, corresponding author: [email protected] Abstract The present work addresses recent findings on the oxy-fuel combustion chemistry and the related composition of oxy-fuel flames. The two main topics are the sulphur chemistry occurring in coal combustion including SO 2, SO3 and H2S formation for different coals and the formation and oxidation of CO in methane, propane and in various coal flames. The implications on corrosion issues in oxy-fuel boiler design are discussed, as well as future research needs. The results presented include experimental work from the literature and results by the oxy-fuel group at Chalmers. The experiments at Chalmers are performed in a 100 kWth oxy-fuel test facility, which has been used extensively since its commissioning in 2003. In addition, modeling of the gas phase chemistry is discussed and related to the experimental work. All results presented for oxy-fuel combustion are related and compared against air-fired conditions. The sulphur chemistry in oxy-fuel combustion differ from that under air-fired conditions; the main areas discussed in the paper are the Conversion of fuel-S to SO2 H2S formation in the flame at (local) fuel rich conditions SO3 formation The SO2 concentration increases drastically in oxy-fuel compared to air-fuel conditions. Both experiments and modeling show that the formation of H2S is promoted by the high SO2 concentration in the oxy-fuel environment, given substoichiometric conditions. The SO3 concentration is also expected to increase, due to the elevated SO2 concentration. Increased concentrations of SO3 in combination with a high moisture content in the flue gases (when recycling wet flue gases) will lead to a higher acid due point temperature with corrosion implications at low temperature parts of the boiler. Both experimental and modeling work indicates that the CO formation is influenced by both homogeneous and heterogeneous reactions occurring in oxy-fuel combustion. Modeling of the homogenous chemistry in methane flames has shown that the main gas-phase reaction influencing the formation is CO2 + H ↔ OH + CO
(1)
Reaction 1 gives, besides a direct effect on CO formation, also an effect on the radical pool composition by promoting the formation of OH, which influences the SO3 formation as is discussed further in the paper. High levels of CO are also found in experimental work on oxy-coal flames where gasification reactions are suggested to contribute to the CO formation, in addition to the homogenous reactions. The main reaction considered is the so called Boudouard reaction C(s) + CO2 ↔ 2CO
(2)
However, there are different opinions on the influence of gasification reactions on the CO formation in oxy-coal combustion, as will be discussed further in the paper. To summarize, high levels of CO and H2S are observed in the flame and near burner zone as well as higher SO2 concentrations in the combustion zone in general. As mentioned, the SO3 levels are expected to increase in oxy-fuel compared to air-fired conditions, but the formation of SO3 needs to be studied in more detail under oxyfuel conditions. For corrosion phenomena occurring at combustion temperatures, design principles otherwise employed
during reburning and air/fuel staging in air-fired boilers may be adopted in oxy-fuel firing. For coals with a high sulphur content and ashes with low sulphur retention capabilities, the temperature window of low temperature corrosion may be drastically influenced. Thus, oxy-fuel combustion is expected to cause a more corrosive environment than air-firing, due to the elevated concentration of combustion products.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
SESSION20 GASIFICATION: FUNDAMENTALS-1 MODELING OF COAL CHAR GASIFICATION IN COEXISTENCE OF CO2 AND H2O Satoshi Umemoto*, Shiro Kajitani, Saburo Hara Central Research Institute of Electric Power Industry (CRIEPI), 2-6-1 Nagasaka, Yokosuka, Kanagawa 240-0196, Japan *Contact Information: [email protected]
Abstract In coal gasifier, carbon dioxide gasification reaction and steam gasification reaction occur at the same time. However, the gasification models which explain the competition between carbon dioxide gasification and steam gasification are few. In this study, coal chars were gasified with carbon dioxide and steam at the atmospheric condition using a thermogravimetric analysis (TGA) at various reactant partial pressures. Langmuir-Hinshelwood type gasification model was modified to explain the char gasification by carbon dioxide and steam. In the proposed model, carbon dioxide gasification and steam gasification share active sites partially. The proposed model can describe the gasification reaction rate of coal chars in the presence of carbon dioxide and steam which was not explained by conventional models. biomass chars [4-6]. They used the simple Langmuir-
1. Introduction Integrated coal gasification combined cycle (IGCC) is be-
Hinshelwood type expression derived based on the absorption
ing developed worldwide to use coal more efficiently and
and desorption theories. Typical expressions are Eq. (1) and
cleanly. There are many types of gasifier in the world. Most of
Eq. (2).
them are oxygen blown entrained flow type. In Japan, air-
r=
blown entrained flow gasifier has been developed by CRIEPI and Mitsubishi Heavy Industries, Ltd [1]. 250 MW IGCC
r=
power plant using the air-blown gasifier was constructed and
k11 PCO2 1 + k12 PCO2 + k13 PCO k 21 PH 2O 1 + k 22 PH 2O + k 23 PH 2
(1) (2)
is in operation [2]. Furthermore, CRIEPI has proposed an O2-
These Langmuir-Hinshelwood type expressions can de-
CO2 blown gasifier for IGCC system with CO2 capture [3]. In
scribe two phenomena. The first is that the gasification reac-
any type of gasifier, three types of char gasification reaction
tion rate at high partial pressure of the reactant gas (CO2 or
occur. Those are O2 gasification (combustion), CO2 gasifica-
H2O) falls below extrapolation from the gasification reaction
tion and H2O gasification. O2 gasification rate is so rapid that
rates at lower partial pressure of the reactant gas because
latter two gasifications are regarded as the rate-determining
many active sites are occupied by the reactant molecules. The
process. Many researchers have constructed CO2 gasification
second phenomenon is that the product gas (CO or H2) can
model and H2O gasification model respectively for coal or
inhibit gasification because the product molecules can also
1
occupy the active sites. However, these equations cannot de-
a new assumption of the property of active site is proposed to
scribe the competition of the CO2 gasification and H2O gasifi-
describe the gasification rate of all coal types.
cation although in the gasifier these gasifications occur at the same time.
2. Experimental
Just a few researchers have proposed some Langmuir-
2.1. Char preparation
Hinshelwood type gasification models for coexistence of CO2
Test chars are prepared prior to gasification test. Three bi-
and H2O [7-10]. These are divided into two types according to
tuminous coals (SF coal from U.S.A, MN coal from Indonesia
property of active site. One assumes the CO2 gasification and
and DT coal from China) are used. The properties of these
H2O gasification occur at separate active sites without sharing
coals and the prepared chars are shown in Table 1. During
active sites nor interaction each other (abbreviated to model 1)
char preparation, the pyrolysis conditions have to be carefully
[7, 8], while the other assumes the all the active sites are
controlled since they have an effect on the reactivity of the
shared and common to CO2 gasification and H2O gasification
char. Pyrolysis temperature and heating rate have an influence
(abbreviated to model 2) [9, 10]. The gasification rate calcu-
on the char gasification reactivity more than pyrolysis pressure
lated based on the model 1 is definitely higher than the one
does [11]. Therefore, the test chars were prepared by rapid
calculated based on the model 2. The experimental data of
pyrolysis in inert gas (nitrogen) using a drop tube furnace at
some coal types agreed with the model 1, and the experimental
constant temperature (1400°C) at atmospheric pressure. Resi-
data of some coal types agreed with the model 2. In this study,
dence time was set at about 3 seconds.
Table 1 Properties of coal and the prepared char (Test chars were prepared by rapid pyrolysis in nitrogen using atmospheric DTF at 1400°C) SF coal SF test char MN coal MN test char DT coal DT test char Proximate analysis [wt%,dry basis] Ash 9.3 18.46 8.38 18.09 11.18 17.45 Volatile matter 41.8 40.65 27.28 Fixed carbon 48.9 50.97 61.54 Fuel Ratio (=FC/VM) 1.2 1.3 2.3 HHV [MJ/kg] 29.0 30.8 29.9 Median diameter [μm] 37.9 42.2 37.3 44.8 21.4 41.4 Ultimate analysis [wt%, dry ash-free basis] C H N S O (differential )
78.5 5.5 1.2 0.2 14.6
Ash composition [wt%] SiO 2 Al2O 3 Fe2O 3 CaO TiO 2 MgO SO3 P2O5 Na2O K 2O
50.5 10.0 5.9 18.2 0.9 3.2 6.0 0.1 3.3 0.5
97.3 0.3 1.1
81.1 5.7 2.0 0.2 11.0
<1.3
50.0 22.8 11.1 2.6 1.0 2.2 1.9 0.3 0.9 2.4
2
98.8 0.1 1.1 <0.1
83.2 4.3 0.3 0.6 11.6
58.2 20.8 8.9 1.9 0.7 0.6 0.9 0.9 0.5 1.7
97.7 0.1 0.7 <1.5
i6 → H 2O (H 2 O) ←
2.2 Char gasification test In this study, the char gasification tests were conducted in
i7
a thermogravimetric analyzer. In all cases, 5 mg of the char sample was placed in an alumina pan and heated at a heating
i8 → H 2 (H 2 ) ←
rate of 15°C/min to 900°C under a continuous argon flow of 400-450 cm3/min. The diffusion isothermal gasification of the conversion (x [-]) and gasification reactivity (r [s−1]) were cal-
i10 C + (H 2 O) → CO + H 2
culated by Eq. (3) and Eq. (4), respectively: m 0 − mt m0 − mash
(3)
r=
dx dt
(4)
(9)
i9
char was initiated by switching on to gasifying agent. Char
x=
(8)
(10)
Eq. (5) to Eq. (7) are elementary reactions concerning to CO2 gasification and Eq. (8) to Eq. (10) are elementary reactions concerning to H2O gasification. There are many other mechan-
where m0 [mg] denotes the sample mass at the start of the ga-
isms. The mechanism which is relatively simple and is able to
sification, mt [mg] the sample mass at gasification time t [s],
distinguish the difference between the active sites for CO2
and mash [mg] the mass of ash. A representative reactivity r50 (r
gasification and the active sites for H2O gasification was
at x=0.5) is used for comparison of data.
adopted in this study. If CO2 gasification and H2O gasification occur at separate active sites (model 1), the overall gasifica-
3. Results and discussion
tion reaction rate derived from the elementary reactions above
3.1. Proposal of new assumption of the property of active site
is written as Eq. (11).
An example of the mechanisms of char gasification is giv-
r=
en by the following steps [4].
1 + k12 PCO2 + k13 PCO
+
k 21 PH 2O 1 + k 22 PH 2O + k 23 PH 2
(11)
On the other hand, if CO2 gasification and H2O gasification
i1 CO 2
k11 PCO2
→ (CO 2 ) ←
occur at common active sites (model 2), the overall gasifica-
(5)
tion reaction rate is written as Eq. (12).
i2
CO
i3 → ← i4
r= (CO)
i5 C + (CO 2 ) → 2CO
(6) +
k11 PCO2 1 + k12 PCO2 + k 13 PCO + k 22 PH 2O + k 23 PH 2 k 21 PH 2O 1 + k12 PCO2 + k13 PCO + k 22 PH 2O + k 23 PH 2
(12)
The active sites are the carbon atoms which can absorb reac(7)
tant molecules [12-14]. Therefore, the active sites for CO2 gasification and the active sites for H2O gasification should be the same. However, some active sites are possibly present in the smaller pore than larger reactant (in this case, CO2) and are
3
active just for smaller reactant (in this case, H2O). In addition,
r=
CO2 is adsorbed molecularly while H2O can dissociatively be
1 + k12 PCO2
chemisorbed to form OH species [15]. Furthermore, the relation between the CO2 gasification rate and the amount of cata-
+
lytic materials is different from the one between H2O gasifica-
k11 PCO2 a a + k13 PCO + k 22 PH 2O + k 23 PH 2 c c
k 21 PH 2O
(15)
1 + bck12 PCO2 + bck13 PCO + k 22 PH 2O + k 23 PH 2
tion rate and the amount of catalytic materials [16]. It indicates
where c is the ratio of the amount of vacant active site for CO2
that the catalysts behave differently in CO2 atmosphere and
gasification to the amount of vacant active site for H2O gasifi-
H2O atmosphere. These effects depend heavily on raw coal
cation expressed as Eq.(16). ⎛ a ⎞ 1 + (1 − b)k 22 PH 2O + (1 − b)k 23 PH 2 c=⎜ ⎟ ⎝ b ⎠ 1 + (1 − a)k12 PCO2 + (1 − a)k13 PCO
type. For these reasons, a new assumption of the property of active site shown Fig. 1 is proposed. In proposed model, two
(16)
new constants are used a and b expressed as Eq. (13) and Eq. (14), respectively: a=
n share
b=
n share
3.2. Application of the proposed model to coal char gasification
(13)
ntH
Table 2 shows the fundamental kinetic parameters kij (i=1 to 2, j=1 to 2) determined from the data of three coal chars
(14)
ntC
gasification tests at 900°C and conversion, x = 0.5. The kinetic
where nshare is the amount of shared active site for CO2 gasification and H2O gasification, ntH is the total amount of active
Table 2 Kinetic parameters for three test chars (TG, 900°C, x = 0.5) SF char MN char DT char
site for H2O gasification and ntC is the total amount of active
k11 [s-1 MPa -1 ] k12 [MPa -1 ]
site for H2O gasification. Using these constants, the gasifica-
-1
36.2 -1
k21 [s MPa ]
tion reaction rate is obtained as Eq. (15):
20.7 -2
41.3
-3
17.6
model 2 CO2 gasif ication and H2O gasif ication occur at common active sites.
ntH
ntC =ntH
proposed model CO2 gasif ication and H2O gasif ication share active sites partially. nshare nshare = a ntH = b ntC ntH ntC ntH 0 ≤ a ≤ 1, 0 ≤ b ≤ 1 ntC nshare: the amount of shared active site by CO2 gasif ication and H2O gasif ication ntC : the total amount of active site f or CO2 gasif ication ntH : the total amount of active site f or H2O gasif ication
Fig. 1 Models of active sites for CO2 gasification and H2O gasification
4
31.2
3.20×10 8.93×10 1.11×10 -2
k22 [MPa -1 ]
model 1 the CO2 gasif ication and H2O gasif ication occur at separate active sites. ntC
1.26×10 -2 1.72×10 -3 2.53×10 -3
24.6
–1
Char gasification rate dx/dt|x=0.5 [s ]
parameters concerning to inhibition reaction by CO and H2 (k13 and k23) are not shown in this study because they are not needed to confirm competition of CO2 gasification and H2O gasification and to determine the parameters in Eq. (15), a and b. Fig.2 (a), (b) and (c) show the gasification reaction rate of SF char, MN char and DT char respectively. Open dots are experimental data of the gasification reaction rate in coexistence of CO2 and H2O. The sum of partial pressure of CO2 and H2O is 0.1 MPa. On the other hand, cross marks are the sum
CO2 concentration [%] 100 80 60 40 20 0 0.001 (a) SF charsum experimental data data 0.0008 model1 0.0006 0.0004
–1
Char gasification rate dx/dt|x=0.5 [s ]
The lines are gasification reaction rate predicted by the model 1 (Eq. (11)), the model 2 (Eq. (12)), the proposed model (Eq. (15)), the model for CO2 (Eq. (1)) and the model for H2O (Eq. (2)). For any chars, the experimental data of gasification reaction rate in coexistence of CO2 and H2O are lower than the prediction by model 1 and higher than the prediction by model 2. Therefore, the proposed model was used. In Fig. 2, the parameters, [a, b] are [0.42, 0.75], [0.42, 1.00] and [0.34, 1.00] for SF char, MN char and DT char respectively. Using these parameters, the proposed model describes the experimental
CO2 0 0
–1
Char gasification rate rx=0.5 [s ]
pore model In previous section, the gasification rate at x = 0.5 was used to discuss the applicability of the proposed L-H model. The random pore model has been used to describe the change of the gasification reaction rate because of the change of the char surface area by many researchers [6, 11, 17-18]. In this section, the proposed L-H type gasification model and the random pore model are combined. The gasification reaction rate equation is written as Eq.(17):
20 40 60 80 H2O concentration [%] CO2 concentration [%] 80 60 40 20
100 0.0004 (b) MN char sum data model1 0.0003
0.0002
0 0
H2O model2
0
experimental data
CO2
20 40 60 80 H2O concentration [%] CO2 concentration [%] 80 60 40 20
100 0.0004 (c) DT char sum data model1 0.0003
0.0002
100
0
experimental data
proposed model H2O model2
0.0001
0 0
100
proposed model
0.0001
data exactly.
3.3. Combination of the proposed L-H model and the random
proposed model
0.0002
of the gasification rate with CO2 without H2O (square dots) and the gasification rate with H2O without CO2 (triangle dots).
model2 H2O
CO2
20 40 60 80 H2O concentration [%]
100
Fig. 2 Comparison of experimental measurements and predictions (TG, 0.1 MPa) (a) SF char (b) MN char (c) DT char
5
shows the CO2 gasification reaction rate (CO2 = 50%) and
k11 PCO2 (1 − x ) 1 − Ψ1 ln(1 − x)
r=
1 + k12 PCO2 + k13 PCO + +
H2O gasification reaction rate (H2O = 50%). From these re-
a a k 22 PH 2O + k 23 PH 2 c c
sults, the dimensionless parameter for CO2 gasification, Ψ1 is
k 21 PH 2 O (1 − x) 1 − Ψ2 ln(1 − x)
1 and the one for H2O gasification, Ψ2 is 4 for MN char. The
(17)
1 + bck 12 PCO2 + bck 13 PCO + k 22 PH 2O + k 23 PH 2
difference between Ψ1 and Ψ2 indicates difference between the
where Ψi (i=1,2) is the dimensionless parameter indicating the
active sites for CO2 gasification and the active sites for H2O
initial pore structure. Some examples of the gasification test
gasification. Fig. 3 (b) shows the gasification reaction rate in
results using MN char are shown in Fig. 3 (a),(b). Fig.3 (a)
coexistence of CO2 and H2O (CO2 = 50%, H2O = 50%). The dot line is the experiment result. The three lines are the calculation results using model 1, model 2, and proposed model.
(a)
The proposed model is better to describe the experimental data
–1
Char gasification rate r [s ]
0.0005 0.0004
than the other models. As seen above, the proposed model is helpful to describe
0.0003
H2O Ψ2=4
though physical implications of a and b need to be considered
0.0002 0.0001
0.0005
in the future work. Some sets of a and b are possibly deter-
CO 2 Ψ1=1
0 0.0
mined to fit the calculation to the data. Physical implications help to determine these parameters. As an example of hypo-
0.2
0.4 0.6 0.8 Conversion X [–]
1.0
thesis, b is larger than a because the molecular size of CO2 is larger than the one of H2O and not able to enter smaller pore.
(b)
–1
Char gasification rate r [s ]
the competition of CO2 gasification and H2O gasification,
0.0004
4. Conclusion
model1 experimental data
0.0003
Coal chars were gasified with CO2 and H2O with a TGA.
proposed model
Conventional gasification models could not exactly describe the competition between CO2 gasification and H2O gasifica-
0.0002
tion. The Langmuir-Hinshelwood gasification model was model2
0.0001
modified to explain the interactions of char gasification by CO2 and H2O. The modified model proposed that CO2 gasifi-
0 0.0
0.2
0.4 0.6 0.8 Conversion X [–]
cation and H2O gasification share partially active sites. The
1.0
proposed model can explain the gasification rate of coal chars
Fig. 3 The change of gasification rate as reaction progress
in the presence of CO2 and H2O which was not explained by
(MN char)
conventional models. Furthermore, the combination of random
(a) CO2 gasification (CO2 = 50%, Ar = 50%) and H2O gasi-
pore model and proposed L-H model can describe of the
fication (H2O = 50%, Ar = 50%)
change of gasification reaction rate as the reaction progress.
(b) Gasification rate in coexistence of CO2 and H2O (CO2 = 50%, H2O=50%)
6
[14] Miura, K., Hashimoto, K., Silveston, Peter L., Fuel, 68,
Acknowledgement A part of the presented work was supported by New energy
1461, (1989)
and industrial technology development organization (NEDO)
[15] Montoya, A. 、 Mondragon, F., Truong, T. N., Fuel
program “Innovative zero-emission coal gasification power
Processing Technology, 77-78, 125, (2002)
generation project”, P08020.
[16] Takarada, T., Ida, N., Hioki, A., Kanbara, S., Yamamoto, M., and Kato, K. Journal of the Fuel Society of Japan, 67, 1061, (1988)
References [1] Inumaru, J., Hara, S., Hamamatsu, T., Ishikawa, H., Pro-
[17] S. K. Bhatia, and D. D. Perlmutter, AIChE J., 26, 379, (1980)
ceedings of International Conference on Coal Science. IEA, Japan, 297, (1989)
[18] Liu, G., Benyon, P., Benfell, K. E., Bryant, G. W., Tate,
[2] Watanabe, T., “Japanese Air Blown IGCC Project
A. G., Boyd, R. K., Harris, D. J., Wall, T. F., Fuel, 79,
Progress”, presented at IGCC Outlook china 2010,
617, (2000)
Shanghai, (2010) [3] Shirai, H., Proc. Clearwater Coal Conference, 44 (3), 610, (2008) [4] J.Gadsby, C. N. Hinshelwood and K. W. Sykes, Proc R Soc Lond Ser A, 187, 129, (1946) [5] Liu, Gui su, Tate, A. G., Bryant, G. W., Wall, T. F., Fuel 79, 1145, (2000) [6] Kajitani, S., Suzuki, N., Ashizawa, M., Hara, S., Fuel 85, 163, (2006) [7] Z. Huang, J. Zhang, Y. Zhao, G. Yue, T. Suda, and M. Narukawa, Fuel Proc. Tec., 91,843 (2010) [8] R.C. Everson, H.W.J.P. Neomagus, H. Kasaini and D. Njapha, Fuel, 85 ,1076 (2006) [9] D.G. Roberts and D.J. Harris, Fuel, 86, 2672 (2007) [10] H.-J. Muhlen, K.H.V. Heek and H. Juntgen, Fuel, 64 ,944 (1985) [11] Kajitani, S., Suzuki, N., “Study on coal pyrolysis property in pressurized entrained flow coal gasifier - Gasification reactivity of coal char pyrolyzed at
high temperature
and High Pressure – presented at 12th International Conference on Coal Science, Australia, (2003) [12] Radovic, Ljubisa R., Walker, Philip L., Jenkins, Robert G., Fuel 62, 849, (1983) [13] Radovic, Ljubisa R., Carbon 43, 907, (2005)
7
Page 1 of 11
INVESTIGATION OF COMPONENT RELEASE DURING PRESSURIZED, HIGH HEATING RATE DEVOLATILIZATION OF COAL AND PETROLEUM COKE David Wagner University of Utah Department of Chemical Engineering Institute for Clean and Secure Energy 50 S. Central Campus Dr., Room 3290 Salt Lake City, UT 84112 [email protected] Kevin J. Whitty University of Utah Department of Chemical Engineering Institute for Clean and Secure Energy 50 S. Central Campus Dr., Room 3290 Salt Lake City, UT 84112 [email protected] Glenn L Shoaf Sr. Development Associate Eastman Chemical Company PO Box 1972 Bldg. 231 Kingsport, TN 37662 423-229-8453 [email protected] Paul Fanning Development Associate Eastman Chemical Company PO Box 1972 Bldg. 231 Kingsport, TN 37662 423-229-8500 [email protected]
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
Page 2 of 11
Abstract A high pressure wire-mesh heater developed at the University of Utah was used to study devolatilization characteristics of two coals and one petroleum coke. The fuels were pyrolyzed under nitrogen and temperatures ranging from 1000
to 1200
at pressures to 63 bar and with
holding time at maximum temperature to five seconds. Char and volatile yields were quantified using a ‘design of experiments’ approach to assess the gasification behavior the three fuels. The findings of these data agree with similar wire-mesh studies concerning pressure, hold time, and final temperature trends.
Background In order to understand pyrolysis and gasification behavior of fuels, a method of rapid heating was devised to simulate the environment of large-scale gasification plants.
Pressure Effects – Several studies have investigated the influence of pressure on devolatilization behavior of coals1-8. Experiments at very low pressure (<1 Torr) were conducted in a wire-mesh heater in which alternating current was used to heat the system ranging from 0.1 to 5000
/s 1.
Gibbins demonstrated that heating rate effects were independent of geometry and that volatile and tar yields increased about six percent d.a.f. with heating rate in an inert atmosphere (helium) under this high vacuum. It was also shown that if the pressure or particle size was increased under rapid heating conditions, only minor pressure effects were present for lignite. With lignite, however, significant increases in hydrocarbon and char yields were evident with a decrease in tar yields, but only if the final temperature was above 800
2
. Sathe, et al. focused on super-
elevated pressures with Australian lignite and found a sharp decline in tar yield at lower temperatures and pressures up to 20 bar3. As pressures exceed 20 bar, the tar yield becomes more level and the slope of the curve decreases.
A separate study was conducted with Pittsburgh seam coal (bituminous), yielding different results. The discrepancies may be attributable to the different coal ranks or may be an inconsistency in the data and methods. To achieve an inert atmosphere for testing, helium was used along with two to ten second holding times for the samples under 1000 reaching a constant final temperature of 1000
at pressures up to 85 bar 2.
/s heating rates
Page 3 of 11
The evolved gas species themselves may influence the char conversion process and indirectly increase the volatiles yield 4. As the partial pressure of hydrogen is increased, volatiles yield is increased, but if the system pressure is increased then the volatiles yield is decreased. The occurrence of such an effect also increases volatiles yield by directly decreasing the amount of char produced, causing the reaction to be interrupted by the relatively high partial pressure of hydrogen 4. Particle Effects – It was found that higher heating rates increased overall volatiles yield 5. This effect could be attributed to the gas-gas and gas-solid reactions occurring simultaneously with higher heating rates. With lower heating rates, however, volatile yields were shown to decrease as a result of repolymerization of coal particles 6. Also, with rapid heating rates (400-100,000 /s), volatiles yields were 20 to 40% more than with slow heating rates 6, 7.
The most prominent diffusive effect from the literature was that as particle size is decreased, higher volatile yields are achieved 4. This would stem from higher fluxes across the boundary layer of the particle by lowering the surface area through which the gases travel; relatively more mass would be transported in less time thereby increasing the volatiles flux from the particle. Parallel with heating rate effects are pressure influences on the rate of diffusion of both tars and gases from fuel particles. It was found that an increase in pressure slowed rates of diffusion, especially for aromatic ring systems captured in the tars, but an increase in the partial pressure of hydrogen had the opposite effect, converting coal to gaseous and liquid products more readily, particularly benzene, toluene, and xylene.
Atmospheric Effects – In previous grid-heater studies, three environments have been used: helium, argon, and hydrogen. There were no discernable differences in the atmospheres used for the various wire-mesh studies with the exception of hydrogen, which increased volatiles yield 10 and 15 percent with respective heating rates of 5 and 1000
/s under a pressure of 70 bars.
Also, since both helium and argon are inert gases, there was no distinct difference in the level of volatiles produced, unlike any case with hydrogen 8. The instance of hydrogen being used is in agreement with other non-wire-mesh studies where an increase in partial pressure of hydrogen
Page 4 of 11 increases volatiles yield. Hydrogen is a reactive species that can (1) consume carbon via the following reaction:
2 and (2) crack aromatic hydrocarbons.
Apparatus and Procedure The pressure vessel that houses the wire-mesh heater is bolted to I-beams, which are then bolted to the floor to offer rigidity to the apparatus and keep it stationary when the flanges are being tightened. Pressures up to 1000 psig are achievable in the pressure vessel fabricated from schedule 40 six-inch stainless steel pipe and stainless steel 300-pound flanges.
Figure 1 shows a top view of the vessel where the mesh is fixed between two copper studs; no mesh is shown in this photo. The green block at the bottom is the thermocouple block that runs to the heat control hardware. The phenolic board is at a fixed height to eliminate the risk of arcing to the vessel shell. Four holes were drilled in the board to allow nitrogen to enter from the bottom and not disrupt the mesh when pressurizing.
Figure 1: Top view of the grid heater apparatus
Page 5 of 11 A schematic of the entire experimental setup is seen in Figure 2 where the SC-2345 computer module is a National Instruments signal-conditioning box with thermocouple and digital output modules SCC-TC02 and SCC-DO01, respectively.
Figure 2: Schematic of the wire-mesh heater experimental apparatus.
Fine-gauge type-R thermocouples were used to accommodate the high temperatures and have a small response time; the bare thermocouples were 0.002 inches in diameter. LabVIEW was used to accomplish a pulse-width modulation (PWM) method of controlling the power output to the mesh without overheating the sample or melting the metal.
Figure 3 shows a photograph of the wire mesh made of 304 stainless steel (SS) with the R-type thermocouple junctions. The thermocouples were spot-welded to the mesh using a small dental welder.
Page 6 of 11
Figure 3: Photograph of the 304 SS wire-mesh and R-type thermocouple leads, scaled. The arrow shows the junction point of the mesh and thermocouple leads.
Procedure – A design of experiments (DoE) plan was proposed because four factors were chosen, each at three levels. The factors chosen were fuel type, mesh temperature, vessel pressure, and hold time, each with appropriate levels to test varying effects and two or three factor interactions. These factors and levels are presented below in Table 1. Table 1: Experimental Conditions used for DoE
Factor Fuel Type Mesh Temperature, Vessel Pressure, bar Hold Time, seconds
Levels Bituminous Coal, Lignite, Petcoke 1000, 1100, 1200 0.9, 21.6, 63.0 1.0, 3.0, 5.0
The operation procedure for the University of Utah’s wire-mesh heater is very similar to those of other studies. A mesh is clamped between two conducting plates or jaws and current is run through at a proportional level to achieve a desired temperature with a possible hold time. The apparatus used R-type thermocouples (platinum/rhenium) to accommodate the higher temperatures and were welded to the mesh, ensuring a constant and local temperature reading. Under operation, the fuel was placed on the mesh-thermocouple junction so that the measurement was assured to be the fuel temperature and not just the temperature of the metal. All fuel was sieved to a particle diameter range of 38 to 75 microns. For pressurized operation,
Page 7 of 11 the vessel was sealed and nitrogen was run through the system to ensure no oxygen was present; this was verified with a gas chromatograph, manufacturer and model: Varian micro-GC CP4900. The vessel was then slowly pressurized at about six bar per minute to the desired pressure and allowed to rest until the thermocouple measurement fluctuations tapered off according to the LabVIEW interface. Within this interface, desired final temperatures, hold times, and heating rates were specified; a consistent heating rate of 1000
/s was used for all runs. After the test
was complete, the pressure was slowly released via a needle valve and the char yield was measured.
Results and Analysis Reproducibility – Following the design of experiments approach, triplicate testing was performed at the midpoint for all three fuels, a temperature of 1100
, a pressure of 21.6 bar,
and a hold time of 3.0 seconds. The average char yields of these runs are shown below in Table 2 per fuel and include error at a 90% confidence level. Table 2: DoE triplicate runs for average char yields per fuel at a 90% confidence level.
Fuel Bituminous Petcoke Lignite
Average Char Yield 69.0% 90.6% 59.4%
Error (90% CL) ± 2.10% ± 1.43% ± 4.29%
The error values from Table 2 can be applied to respective fuel types to quantify the error in each subsequent measurement without having to reproduce all conditions in triplicate (e.g. all error values for bituminous coal runs are assumed ±2.10% char yield).
The trends found in other wire-mesh studies were also observed in this study. Char and volatile yields are tabulated below in Table 3 along with corresponding run conditions. Yields that are starred (*) represent averages of more than one test under corresponding conditions; some runs were repeated under advisement of Eastman Chemical Company and some repeated based on the operator’s experience.
Page 8 of 11 Table 3: Ch har and volatilee yields of coal and a petcoke wirre-mesh heater runs with corresponding run conditions. Values that are staarred (*) are averages of more than one run at a run condition ns.
Run Conditions C Finnal Temp Prressure Holdd Time (C Celsius) (psig) (seconds)
Run
Bituminous B Volatilees Chhar Yieeld Yieldd
Petcoke Char Volatiles Yield Yield
Lignite Char Voolatiles Y Yield Yield
A
1000
0
1
*81..1%
*18.9% %
94.1%
5.9%
59.9%
400.1%
B
1000
0
5
54.11%
45.9% %
94.5%
5.5%
58.0%
422.0%
C
1000
900
1
68.22%
31.8% %
96.6%
3.4%
95.7%
4 4.3%
D
1000
900
5
61.11%
38.9% %
96.6%
3.4%
54.7%
455.4%
E
1100
300
3
*69..0%
*31.0% %
*90.6%
*9.4%
*59.4%
*440.6%
F
1200
0
1
55.44%
44.6% %
95.7%
4.3%
*54.0%
*446.0%
G
1200
0
5
52.66%
47.4% %
90.0%
10.0%
*51.0%
*449.0%
H
1200
900
1
69.99%
30.1% %
99.6%
0.4%
63.0%
377.0%
I
1200
900
5
*67..1%
*32.9% %
89.9%
10.1%
91.8%
8 8.3%
Due to a three-factorr design of exxperiments, plotting the results becoomes difficullt. Below arre two plotss of the abov ve data, one radar r plot annd one colum mn plot. Com mparing thesse two figurees with eachh other and with w Table 3, 3 trends can be found annd qualitativee conclusionns made.
Char Yield, %
Bituminous
Petcoke
Lignite
100.0% 1 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% A
B
C
D
E
F
G
H
I
Run Con nditions Figurre 4: Char Yield d (%) versus Run Conditions. The run condiitions correspon nd to specified runs r in Table 3.
Page 9 of 11
Bituminous
Petcoke
Lignite
A 100.0% I
B
80.0% 60.0% 40.0% 20.0%
H
C
0.0%
G
D
F
E
Figure 5: Radar plot of Char Yield (%) versus Run Conditions found in Table 3.
While Figure 4 offers a traditional view of data, Figure 5 shows a quick qualitative perspective of the char yield data. For each lettered run condition, the values of the three fuels are compared against each other and trends are found. For example, beginning with condition A and moving clockwise, the char yield will increase, decrease, or remain relatively constant. Depending on the direction of change for the respective char yields, conclusions are made with respect to a specific factor found in Table 1. Figure 5 shows a rigid char yield trend concerning petcoke, which would certainly be attributable to a very low volatiles percentage and high percentage of fixed carbon. For the bituminous coal tests, as found in aforementioned studies, as pressure increases the volatiles yield will decrease. This is exemplified in runs A and B versus C and D and runs F and G versus H and I, where A through D are at 1000
and F through I are at 1200
. Discrepancies in specific values or magnitudes of yields stem from a difference in fuels and even experimental technique. These reasons could be the cause of inconsistent yields when comparing the lignite runs. Here, runs C and I stand out to such a large degree that the char
Page 10 of 11 yields approach those of petcoke; since lignite has approximately four to five times more volatiles, this would be unexpected. While all testing represented in Table 3 was performed using the same standard operating procedure, it may not have been performed that same way in other studies, presenting variability and uncertainty in wire-mesh study results and methodology.
Conclusions and Recommendations The wire-mesh heater designed and fabricated at the University of Utah’s Institute for Clean and Secure Energy to study gasification behavior of coal and petroleum coke successfully reproduced trends in char and volatile yields that other studies have found. Pressure, final temperature, and holding time were varied to understand the influence of each factor on the gasification behavior of such fuels. Future testing is recommended to offer a wider range of final temperatures and offer a more complete design of experiments to fill in the gaps of the data presented herein. While qualitative results suggest a correct method, quantitative results indicate how well the method accurately presents the data and if they agree with other wire-mesh studies. Larger sample sizes are proposed for future testing to measure elemental compositions of the chars and quantify the influence of factors affecting gasification behavior to aid in industrial processes.
Acknowledgment This work was funded by Eastman Chemical Company.
References 1.
Gibbins, J. R.; King, R. A. V.; Wood, R. J.; Kandiyoti, R., Variable-heating-rate wire-
mesh pyrolysis apparatus. Review of Scientific Instruments 1989, 60 (6). 2.
Suuberg, E. M.; Peters, W., A.; Howard, J. B., A Comparison of the Rapid Pyrolysis of a
Lignite and Bituminous Coal. Thermal Hydrocarbon Chemistry 1979. 3.
Sathe, C.; Hayashi, J.-I.; Li, C.-Z., Release of volatiles from the pyrolysis of a Victorian
lignite at elevated pressures. Fuel 2002, 81 (9), 1171-1178. 4.
Anthony, D.; Howard, j.; Hottel, H.; Meissner, H., Rapid devolatilization and
hydrogasification of bituminous coal. Fuel 1976, 55. 5.
Rezaiyan, J., Gasification Technologies. 2005.
Page 11 of 11 6.
Shadle, L.; Berry, D., Coal Gasification. In Kirk-Othmer Encyclopedia of Chemical
Technology, 2009; Vol. 6. 7.
Qadar, S. A., Coal Science and Technology 8: Natural Gas Substitutes from Coal and
Oil. 1985. 8.
Stojiljkovic, D.; Radovanovic, M.; Ercegovac, M.; Jones, J. M.; Williams, A.,
Devolatilization of some pulverized Eastern European lignites. Journal of the Energy Institute 2005, 78 (4), 190-195.
High pressure TGA studies on German brown coal under carbon dioxide atmosphere
Abhishek Bhargava*, Patrick J. Masset
Freiberg University of Mining and Technology Centre for Innovation Competence Virtuhcon – Group “Multiphase Systems” Fuchsmühlenweg 9, Reiche Zeche D-09599 Freiberg, Germany
Abstract In this study, a high pressure thermo-gravimetric analyser was used to probe the partial oxidation characteristics of coal to simulate gasification like conditions. A gasification grade, German lignite coal was selected for the HP-TGA analysis at increasing values of total pressures of up to 45 bars in carbon dioxide atmosphere. Three distinct reaction zones were identified and the kinetic constants were calculated using the Coats and Redfern model. It was shown, the accessible porosity responsible for gas transport increases with the temperature and therefore the reaction rate. Computer controlled scanning electron microscope (CCSEM-EDX) was used to probe the surface morphology of coal-char particles recovered from the HP-TGA to validate the transport mechanism of the reactant gas from the surface to the un-reacted core. The evolution of solid phase porosity in coal samples with increasing temperature impacted the diffusivity of reactant gases in coal particle and thus the reaction kinetics.
Keywords: Coal gasification, High Pressure –TGA, , Kinetics, Carbon dioxide
*
Corresponding author: [email protected]
Tel/Fax: +49 3731 39 4809/4555
1. Introduction Clean coal technologies such as Integrated Gasification Combined Cycle (IGCC) power plants have attracted considerable interest for efficient and clean utilization of coal. In a gasification process, a carbonaceous fuel such as, coal is converted into a combustible gas of usable heating value. Several commercially available gasification technologies such as those of Lurgi, Shell, Prenflow etc. operate at high temperatures (1200°C-1500°C) and pressures, up-to 100 bars. Literature survey shows that several experimental studies related to thermal behaviour of coal at atmospheric pressures have been reported [1]. However, little data is available at high temperatures and pressures. The objective of this study is to explore the gasification kinetics and to study the gas-solid reactions between coal and carbon dioxide at elevated pressures. Classically, high pressure studies related to coal pyrolysis and other gas–solid reactions were done experimentally using wire mesh reactors (WMR) [2], pressurized drop tube furnaces (PDTF) [2] and high pressure thermo-gravimetric analyzers (HP-TGA) [1]. In this study, a high pressure thermo-gravimetric analyzer was utilized to explore the gasification kinetics of a German brown coal under CO2 atmosphere, constant heating rate and increasing values of total system pressure. Previously, Megatritis et al. [3] studied the relationship between the char preparation conditions and the gasification reactivity under high pressure CO2 using a two step method. They explained that the conversion rate from direct gasification was higher than, when the samples were first pyrolyzed and then gasified under CO2 in the same reactor. Molina et al. [4] studied the reactivity of coal with steam and CO2. They discussed the validity of kinetic models to describe the relation between conversion and time. Sadhukan et al. [5] have characterized the porous structure of coal char from a single devolatilized coal particle. They investigated the structural evolution during the combustion of char in a fluidized bed reactor. The development of pore structure shows the mechanism of gasification process.
In this study a gasification grade German brown coal was
procured and tested with HP-TGA under CO2 atmosphere to study the solid state reaction kinetics. Additionally, in a separate experiment, Scanning electron microscopy (SEM) was performed on the
samples recovered from HP-TGA upon reaction with CO2 to observe the evolution of surface characteristics with increasing heating rates and total system pressure.
2. Experimental 2.1. Sample preparation A gasification grade German lignite coal was selected for High Pressure – Thermo-gravimetric analysis. The sample was milled in a mortar and pestle and later sieved using Tyler screen series to recover coal particles in the size range 100-200 μm for further analysis. No thermal pre-treatments were applied. The ultimate and the proximate analyses of the coal sample were carried out in accordance with standard procedures.
2.2. Apparatus High Pressure–Thermo-gravimetric Analyser (HP-TGA) was used to perform thermal analysis on the coal sample at pressures up-to 45 bars. The device was custom made for carrying out research activities at TU Freiberg by Deutsche Montan Technologie. Figure 1 shows the schematic diagram of the HP-TGA reactor. For experiments, a known mass of the sample (approx. 300 mg) was used. For this, a cylindrical crucible made up of in-coly with closed ends was used. A thin wire mesh along the curved surface of the crucible enabled gas – solid interaction.
Microbalance
Purge gas entry
Sample lock Electricity connection Reactor
Reactor tube
Purge gas/ Product gas exit
Purge gas/ Product gas exit
Sample temperature
Figure 1: High Pressure–Thermo-gravimetric Analyzer
2.3. Techniques Coal sample prepared in the previous step was locked in the crucible and gasified with carbon dioxide non - isothermally. A constant heating rate was applied, and the sample was subject to react with CO 2 at different values of total pressure. Experiments were performed at 20°K/min at different values of
total system pressure (1-45 bars). The weight loss profile was recorded during the experiment with a microbalance as described in the previous section. The device was calibrated for temperature and weight loss using penta hydrated calcium sulphate which offered well defined mass and temperature transitions. Later, the weight loss profile was followed in non-isothermal mode under the influence of carbon dioxide at a fixed value of the total system pressure (up-to–45 bars) and temperature upto 1100 °C. The samples were later allowed to cool and recovered from the HP-TGA for further analysis. The accompanying software, corrected the data for buoyancy effects during weight measurements. Scanning electron microscopy- was used to observe the surface morphology and the porosity in coal particles. For this, scanning electron microscope Carl Zeiss - DSM 960 and EDX was used. For comparison, SEM images were recorded for original coal samples, the heat treated coals and those recovered after the reaction with carbon dioxide from the HP-TGA.
3. Results and discussion 3.1. Coal characterization The chemical analysis was carried, out on a representative sample from the bulk coal to avoid the effect of in-homogeneity. Table 1 and Table 2 show the results of the ultimate and proximate analyses of the coal sample performed in accordance with standard German DIN procedures. The values obtained are classical for a lignite type coal [6].
Table 1: Ultimate analysis of German gasification grade coal Wet Basis
Dry Basis
Element
Mass (%)
Mass (%)
C
63.36
55.76
H
4.45
4.0
O
21.16
18.62
N
0.77
0.68
S
1.16
1.02
9.00 Ash
Total = 99.9
7.92
55.73
12.00
Water (raw)
Total = 100
Table 2: Proximate analysis of German gasification grade coal Wet Basis
Dry Basis
Mass (%)
Mass (%)
Fixed Carbon
40.27
35.44
Volatile Matter
50.73
44.64
9.00 Ash
Total = 100
7.92 12.00
Water (raw)
55.73
Total = 100
3.2. Effect of operating pressure on coal gasification reactions The effect of operating pressure on coal gasification reactions was studied by fixing the value of partial pressure of CO2 (pure gas) and the varying the total system pressure over a range of 1 to 45 bars, other parameters being kept constant (particle size, heating rate and reaction gas). Figure 2 shows the weight loss profile of the coal sample in the temperature range 50-1200°C at different values of total system pressure upon reaction with carbon dioxide. The overall weight loss was found to be in the range of 93-96 percent. Figure 2 shows that higher values of total system pressure, correspond to a trend showing slightly increasing rate of weight loss. The values were corroborated using DTG curves for each experimental run. 100
Weight (%)
75
50
25
0 0
200
400
600
800
1000
1200
Temperature (°C) 1 Bar
10 Bars
20 Bars
35 bars
40 Bars
45 bars
Figure 2: High Pressure TGA of German brown coal recorded under carbon dioxide at 20°K/min
Weight Loss (Percent)
100
75
32,51
32,71
33,27
33,41
23,9
22,38 50
27,44
25,79
28,06
25 33,44
23,24
32,5
34,69
39,5
10 Bar
20 Bar
35 Bar
47,43
0 1 Bar
45 Bar
Reactant Gas Pressure Moisture Loss
Volatiles and Tars
Char Gasification
Figure 3: Percentage weight loss in German brown coal at high pressure under carbon dioxide atmosphere
Another common observation from the TGA profiles of the coal sample is the occurrence of weight loss in three distinct steps. Figure 3 summarizes the three step reaction of carbon dioxide with coal. Step one corresponds to moisture loss of 33-48 percent in the temperature range 50-250°C. The second step occurs in the temperature range of 200-550°C, with the emission of volatiles and tars comprising of polymeric aromatic compounds leaving behind the solid fuel, dry and devoid of any moisture (coal-char). The last step occurs at temperatures above 550°C. The coal-char reacts with carbon dioxide resulting in gasification of remaining carbon matrix into syn-gas (mainly CO and H2). Figure 3 shows that for each experimental run, pressure of the reactant gas showed increasing effect on the percentage moisture loss. The values of the weight loss due to inherent moisture content in the samples were found to be in the range of 32-47 mass percent. The balance 10-15 percent, content is assumed to be lost during the initial step, when the reactor is purged with carbon dioxide. The hot reactant gas (T=200°C) entering the reactor partially extracts the surface moisture from the coal sample. Overall trend shows increasing values of water loss with rising pressure. In the de-
volatilization zone, the volatiles and tar yield showed approximately constant values (22-27 percent), up-to 35 bars, but showed a decreasing trend with increasing values of total pressure. This effect appears to be minor at low temperatures (200-550°C) but, it is expected to be more pronounced at higher values. The last step corresponds to the char gasification, in which the percentage weight loss was found to be constant with increasing pressure, except at 35 bars. One possible reason for such an observation could be the inhibition effect caused due to formation of H2, also investigated by Wall. et al. [2]. Such behaviour is subject to more experimental validation. 3.3. Kinetic Parameter Estimation In the present study, the gas-solid reactions were studied in totality rather than artificially fragmenting the char gasification step. Thus, no pre heat treatments were provided to the coal sample. A modified approach was used for data analysis and evaluation of the kinetic parameters for the gasification step. The TGA curve recorded for different values of total system pressure was divided into three distinct zones i.e. – moisture loss zone, de-volatilization zone and char gasification zone respectively. The Coats and Redfern model [7] was used to fit each zone separately resulting in the computation of kinetic constants. The model assumes several values of the reaction order on hit and trial basis. This assumption points to the inference of the possible reaction mechanism involved. For, the values of n = 1, the very nature of the reaction mechanism is assumed to be elementary consisting of one or two molecular species (atoms, molecules, ions etc.). However, in view of the complex physical and chemical composition of coal particles, it is very obvious to consider the various steps involved in the reaction of active carbon sites with the reactant gas (CO2). Wall et al. [2] compiled a list of studies pointing towards the multiple step reaction mechanism considering complexities of reactant gas adsorption, and carbo-oxygen complex formation in the secondary steps. Thus, it is fairly justified to use the fractional values of orders n= 0.5, 2/3, 1, 2, 3 illustrating the multiplicity of the steps involved in the reaction mechanism.
Coats and Red-fern [7] derived the following expression:
n = 1:
1 (1 )1n AR 2 RT E log 2 (1 ) log E E 2.3RT T (1 n) 10 10
(1)
n 1:
AR (1 ) 2 RT E log log log (1 ) 2 E E 2.3RT 10 T 10 10
(2)
where k is the reaction rate constant, A the frequency factor, α the fraction of coal decomposed at any time t, n the order of the reaction, β the heating rate(K/min), R the gas constant (8.314 J mol-1K-1) and
1 (1 )1n vs. 1/T or when, n = 1, the plot of T 2 (1 n) 10
T the reaction temperature (K). The plot of log
(1 ) log log Vs 1/T should result in a straight line. The slope of such an equation should be 2 10 T 10 equal to –Ea/ (2.303R) for the correct value of n. The first term on the right hand side of equation one AR 2 RT and two- log (1 ) , is sensibly constant, for most values of activation energy in the temperature E E 10 range, in which, the reactions occur. The value of pre-exponential factor was calculated for each
experimental run by equating the first term of equation (1) and (2) on the right hand side to the intercept value obtained from Figure 4 (sample Figure for experimental run at 35 Bar).
10 9
log(-log(1-alpha)*T^-2)
8 7 6 5 4 3 2 1 0 0
0,5
1
1,5
2
2,5
3
3,5
(1000/T) Char gasification zone (n=0.6)
Moisture loss and De-volatalization zone (n=3)
Char gasification zone (n=0.5)
Figure 4: Application of Coats and Redfern Model to Coal gasification reaction under carbon dioxide atmosphere at 35 bars
40 35
Activation Energy (kJ/mol)
30 25 20 15 10 5 0 1
10
20
35
Pressure (Bars)
Mois ture Los s Zone (n=3)
De-volatalization Zone (n=3)
40
45
Char-Gas ification Zone (Averaged for n=0.6 to 1)
Figure 5: Computed values of activation energy for three step coal reaction with CO2 at increasing values of total system pressure.
1,00E+12 1,00E+11 1,00E+10 1,00E+09
Pre-exponential factor (A)
1,00E+08 1,00E+07 1,00E+06 1,00E+05 1,00E+04 1,00E+03 1,00E+02 1,00E+01 1,00E+00 1,00E-01 1,00E-02 1,00E-03 1,00E-04 1 Bar
Moisture Loss Zone (n=3)
10 Bar
20 Bar
De-volatalization Zone (n=3)
35 Bar
40 Bar
45 Bar
Char-Gasification Zone (Averaged for n=0.6 to 1)
Figure 6: Computed values of pre –exponential factor for the three step coal reaction with CO2 at increasing total system pressure. Figure 5 and Figure 6 summarize these averaged values of the kinetic constants. Regression analysis was used to find the best fit for each zone. For several experimental runs, the values of R squared show good fits at n =3 for the moisture loss and de-volatilization step. However, for the char gasification step the value of R squared showed shifting behaviour. The n values showed a decreasing effect from n = 1 to n =0.6. It can be concluded that the reaction order is a function of total system pressure and temperature. Ahn et. al.[2] have even proposed a rate law following nth- order rate equation based on shrinking core model showing direct proportionality of reaction rate with partial pressure of carbon dioxide, total system pressure, temperature and activation energy. For simplistic approach towards data treatment, averaged values of activation energy Ea, and pre-exponential factor A, were calculated from the slope and intercept values obtained from the best fits (based on R squared method). Table 3 and Table 4 show the values of activation energy and pre-exponential factor computed for the data set obtained at increasing values of total system pressure from 1 to 45 bars. These values lie in the range of 6.31 to 36.05 kJ/mol and 4.43*10-4 to 1.40*1011 respectively. For each
reaction-zone the values showed increasing trend with pressure. For comparative purposes, the values of kinetic constants were compared with the works of Yong Jean Kin [8] and Dong Zeng [9], showing close agreement.
Table 3: Activation energy computed using Coats and Red-fern model for German brown coal at increasing values of total system pressure of carbon dioxide Activation Energy (kJ/mol) Pressure (Bar) 1 10 20 35 40 45
Moisture Loss Zone (n=3) 25.2 17.4 19.5 29.9 30.1 27.4
De-volatalization Zone (n=3) 6.3 7.8 10.2 35.5 35.5 36.0
Char-Gasification Zone (Averaged for n=0.6 to 1) 11.8 14.8 10.6 29.9 25.8 30.4
Table 4: Pre-exponential factor computed using Coats and Red-fern model for German brown coal at increasing values of total system pressure of carbon dioxide Pre-exponential factor Pressure 1 Bar 10 Bar 20 Bar 35 Bar 40 Bar 45 Bar
Moisture Loss Zone (n=3) 3,26E-01 3,45E-03 2,09E-02 3,67E+10 3,85E+10 3,06E+10
De-volatalization Zone (n=3) 4,43E-04 1,88E+01 9,02E+00 8,30E+10 8,01E+10 1,40E+11
Char-Gasification Zone (Averaged for n=0.6 to 1) 9,96E-03 2,68E-02 2,13E+05 3,67E+10 1,60E+09 3,29E+09
3.4 Scanning Electron Microscopy Scanning electron microscopy showed that the porous structure of coal -char particles recovered from the crucible upon reaction with CO2. It showed the development of rough fragmented surfaces showing changing porosity upon thermal treatment. These images also indicate the possibility to reconstruct the various steps involved in coal –char gasification, which include diffusion of the reactant gas to the intricate porous char surface on the active carbon sites, and later desorption of
evolved gases upon completion of the reaction leaving behind tiny bubbles on the surface Figure 7 and Figure 8 show the SEM images of the coal - char.
A
B
C
D
Figure 7: Scanning electron micrographs of the German brown coal –char samples recovered from the HP – TGA upon gasification with CO2 at 1 bar (A) 10K/min – (500X); (B)10K/min – (9000X); (C)20K/min (500X); (D) 30K/min (500X)
The grey portion containing carbon, correspond to the carbon matrix, while the ash components are randomly distributed in it. The particles show open surface porosity which seem to be mildly affected by varying the heating rates and the total system pressure. The (A, C, D) appear to be fragmented when viewed under lower magnification (500X), and show up as a single entity having rough outside surface.
A
C
B
D
´ Figure 8:Scanning electron micrographs of the German brown coal -char samples recovered from the HP – TGA upon gasification with CO2 at 35 bar (A) 10K/min (500X); (B) 20K/min(500X); (C)20K/min (500X); (D) 30K/min (500X); Figure 8(A, B, C) show that the, particles show coherent overall structure, with rugged surface, displaying open porosity. These results are in agreement with those of Zheng et al. [9] who studied the swelling properties of intrinsic reactivities of coal chars produced at elevated pressures and high heating rates using scanning electron microscopy. They reported the evidence of bubble formation on the char surface, which possibly would have increased the diffusion of reactant gas inside the unreacted coal particle. One possible reason for such an observation might have been the escape of trapped gas molecules on the surface of coal particles leading them to swell and eventually burst. Although, with German brown coal minimal bubble formation was observed. The out side surface of each particle appears to be smooth as shown in Figure 8 (A). At higher pressure, some swelling was observed, due to formation of char resulting in voluminous mass. One extrapolation of this result could be that, at high temperatures, plasticity might increase leading to enhanced bubbling and later
coalescence, resulting in elevated surface roughness due to appearance of tiny blisters. The bubbling effect at elevated pressure and high temperature is subject to further experimental validation. 4. Conclusions The influence of pressure on coal gasification reaction under rising carbon dioxide pressure showed slight increase in carbon reactivity. The Coats and Redfern model was used to evaluate kinetic constants, such as activation energy and pre-exponential factor. The model showed that the order of reaction is a function of total system pressure. For moisture loss and de-volatilization step the values of activation energy and pre-exponential factor were found to lie between between 6.31 to 36.05 kJ/mol and 4.43 *10-4 to 1.40*1011, respectively. For char-gasification step, the values of activation energy and
pre-exponential factor were found to be in the range of 11.84 to 32.10 kJ/mol and 9.96 *10-3 to 3.69*109 respectively.
For the moisture loss and de-volatilization step the order was found to be close to three.
However, for the char gasification step, the value of n showed a decreasing trend with rising pressure. The value is likely to fall in the range of 0.6 to 1. More experimental investigations are required to validate the coal – char gasification modelling in order to propose a suitable reaction mechanism for German brown coal. Scanning electron microscopy showed internal structure of coal –char particles, showing smooth outside surface with open porosity. The effect of increasing pressure and temperature was mild on the surface morphology. The particles showed slight swelling and appeared in discreet shapes. The overall sample mass showed enhanced volume indicating conversion of coal to char.
Acknowledgment The financial support from BMBF (Federal Ministry of Education and Research) and SMWK (Saxon State Ministry for Higher Education, Research and the Fine Arts) is gratefully acknowledged. The authors thanked Dr. Schneider for the SEM analysis.
References 1. Fermoso, J., M. V. Gil, et al. (2010) "Kinetic models comparison for non-isothermal steam gasification of coal-biomass blend chars." Chemical Engineering Journal 161(1-2): 276-284. 2. Wall, T. F., G.-s. Liu, et al. (2002). "The effects of pressure on coal reactions during pulverised coal combustion and gasification." Progress in Energy and Combustion Science 28(5): 405-433. 3. Ahn, D. H., B. M. Gibbs, et al. (2001). "Gasification kinetics of an Indonesian sub-bituminous coal-char with CO2 at elevated pressure." Fuel 80(11): 1651-1658. 4. Megaritis, A., R. C. Messenböck, et al. (1999). "High-pressure pyrolysis and CO2-gasification of coal maceral concentrates: conversions and char combustion reactivities." Fuel 78(8): 871-882. 5. Molina, A. and F. Mondragón (1998). "Reactivity of coal gasification with steam and CO2." Fuel 77(15): 1831-1839. 6. Sadhukhan, A. K., P. Gupta, et al. (2009). "Characterization of porous structure of coal char from a single devolatilized coal particle: Coal combustion in a fluidized bed." Fuel Processing Technology 90(5): 692-700. 7. Akkaya, A. V. (2009). "Proximate analysis based multiple regression models for higher heating value estimation of low rank coals." Fuel Processing Technology 90(2): 165-170. 8. A. W Coats, J.P Redfern. (1964). “Kinetic Parameters from Thermogravimetric data”. Nature, . 201 (68-69). 9. Kim, Y. J., J. M. Lee, et al. (1997). "Coal gasification characteristics in an internally circulating fluidized bed with draught tube." Fuel 76(11): 1067-1073. 10. Zeng, D (2005). “Effects of pressure on coal at high heating rates and char combustion” Phd Thesis, Brigham Young University. 11. Zeng, D., M. Clark, et al. (2005). "Swelling properties and intrinsic reactivities of coal chars produced at elevated pressures and high heating rates." Proceedings of the Combustion Institute 30(2): 2213-2221.
STUDIES ON GASIFICATION OF CHAR IN FIXED BED REACTOR Ramesh Naidu Mandapati1, Sateesh D1, Narseh D1, Sanjay M Mahajani1, Anuradda Ganesh2, Sharma R.K.3, Preeti Aghalayam1* 1 Department of Chemical Engineering, IIT Bombay, Powai, Mumbai-400076, India 2
Energy Systems Engineering, IIT Bombay, Powai, Mumbai-400076, India
3
UCG Group, IRS, ONGC, Chandkheda, Ahmedabad -380005, Gujarat-, India
*Corresponding Author: [email protected] Ph.No: +91(22)2576 7295, Fax: +91(22) 2572 6895
Abstract In the present work, we study gasification of char obtained by pyrolysis of Indian Lignite coal, in a fixed bed reactor. Because of its operational flexibility, the fixed bed reactor (FBR) can be conveniently used to carry out endothermic reactions under controlled conditions. The present work is focused on how various operating parameters such as temperature, flow rate, particle size and pressure affect the extent of the steam and CO2 gasification reactions. A broad objective of this work is to develop a kinetic model, identify mass and heat transfer limitations, if any, based on this data and further support the results obtained by carrying out independent thermogravimetric analysis. Keywords: UCG, Fixed bed reactor, Gasification
Introduction The energy demand of world is increasing continuously. Coal, petroleum and natural gas are the major sources of energy generation. Depleting oil and gas reserves can be substituted with abundantly available coal1. India has huge reserves of coal. However, most of this coal is low grade, with as much as 35-50% ash content. It is therefore necessary to utilize as much coal as possible in an efficient way. Gasification is one such promising alternative to coal combustion. An efficient design of a gaisifer requires a reliable kinetic model which is normally developed based on the data generated over a wide range of operating parameters through laboratory experiments.
Thermogravimetric, grid heater and curie-point are among the different techniques which can be employed to study the kinetics of the solid-gas reactions5. In these instruments,
mass and heat transfer limitations may interfere in the analysis of the intrinsic kinetics and there is a bit of uncertainty in using this data directly for kinetic modelling. It is therefore necessary to authenticate the data in relatively large size laboratory reactors such as fixed bed reactors, fluidized bed reactors or draft tube furnaces, of which the heat and mass transfer characteristics are well known. In the present work, we perform experiments in a fixed bed reactor (FBR) to study the gasification kinetics. The gasification reaction is highly endothermic and reaction proceeds at high temperature. FBR setup can be used for endothermic reactions and maintaining isothermal conditions is possible in this setup6-10. The idea behind the present work is to study how the gasification rate is influenced by the reaction temperature and flow rate of the gasifying agent.
The quality and calorific value of product gas from the gasifier depends on the concentrations of CO and H2. These products are formed mainly in two gasification reactions viz reaction of carbon with steam and that with carbon dioxide (See Eqs. 1 and 2). The reaction kinetics is likely to also depend on the type of coal used. In this work, we study the gasification of lignite with proximate and ultimate analysis given in Table 1.
C + H2O
= CO + H2
C + CO2 = 2CO
(1) (2)
It may be noted that alongwith the gasification reaction a homogeneous uncatalysed gas phase reaction of CO and H2O (Eq. 3) called as water gas shift reaction, may take place simultaneously to alter the ratio of CO to H2 in the product gas.
CO H 2O CO2 H 2
(3)
Experimental Set-up The schematic diagram of fixed bed reactor setup is shown in figure 1. It is equipped with temperature sensors and controllers, pressure indicators, flow meters and on-line data acquisition system. The pressure and temperature ratings of the reactor are 80 atm and
12000C, respectively. The reactor in this setup can be used in both vertical and horizontal positions. To control the reactor temperature, three PID controllers are provided with sensors. In addition to these, four more thermocouples are used to measure the internal temperature of reactor at different positions. The flow of reactants can be controlled by the flow controllers. Steam temperature can be maintained at a required level with the help of a PID controller of the steam generator. The water supply to the steam generator is controlled through a pump. Two additional temperature controllers are provided to control the feed and product gas temperatures. The product gasses are passed through the double pipe heat exchanger to collect the condensables like water. The non-condensables are sent to GC for to analyse the composition of various components. The advantage of FBR setup is that an endothermic reaction can be carried out at near isothermal conditions. The steam and CO2 gasifications of coal are conducted in FBR at different temperatures and different flow rates of steam and CO2. Table 1: Proximate and ultimate analysis of coal Moisture Proximate Volatile matter Analysis Fixed Carbon
40 11 44 5
Ultimate
Ash Carbon
60.366
Analysis
Hydrogen
4.367
Experimental procedure A step-wise experimental procedure adopted to study both steam and CO2 gasification is as follows:
The reactor is detached from the setup and cleaned properly.
The coal sample is prepared and weighed.
The reactor is filled with 50 gm of coal of small size particles. Alumina balls are used as a packing material on both the sides of the packed coal in the reactor.
The reactor is positioned horizontally and thermocouples are connected at their
respective positions.
Initially the reactor is maintained at around 1500C to vaporize the moister content of the coal and inert gas is circulated through the reactor setup. This heating is carried out for two hours.
After the removal of moisture the reactor is maintained at sufficiently high temperature (~300oC) under inert atmosphere for a long time to remove the volatiles from the coal and convert it char. The gasses coming out of the reactor are analysed by using gas chromatography to ensure that the pyrolysis is over.
For steam gasification, the set point of the steam generator is adjusted to 1000C and water is sent through pump at a required flow rate. The product gas samples are withdrawn at every 10 min interval and analysed using gas chromatography. Similar procedure is applied for CO2 gasification
Gasification is carried out until the entire amount of char is consumed. Supervisory control and data acquisition system
Steam generator
Gas chromatography Gas
FIC
Gas
FIC Condenser
Heating elements with controllers Figure 1: Fixed Bed Reactor set-up
Internal thermocouples
Results and discussions The steam gasification experiments are conducted at constant temperatures from 650 to 9500C with an increment of 50oC. For all these experiments the inert gas flow is maintained at 100 ml/min. The steam partial pressure is varied by varying its feed flow rate from 2 to 6 gm/sec. It is expected that along with steam and carbon dioxide gasification, the homogeneous gas phase water gas shift reaction (Eq 3) can also take place substantially at these temperatures.
General Course of the reaction
The extent to which CO and H2 are formed is sensitive to the changes in temperature and steam flow rate. The change in composition of hydrogen, carbon monoxide and carbon dioxide in the product gas with respect to time at different temperatures is shown in figure 2. In all the cases, the concentration of the product species decreases with respect to time and eventually drops down to zero. This is because of reduction in the amount of coal available for the reaction. In general, high temperature favours formation of CO. The effects of temperature and steam flow rate on the extent of CO and H2 generated during gasification are discussed in the subsequent sections.
60
H2 CO CO2
50 40
Mole percentage
Mole percentage
60
30 20 10
H2 CO CO2
50 40 30 20 10 0
0 0
10
20
30
40
50
Time, min
(a) Steam gasification at 9500c
0
25
50
75
100
Time, min
(b) Steam gasification at 8000c
Figure 2 : Mole fractions of gasification products
125
150
Effect of Temperature on steam gasification
The rate of consumption of coal increases with increase in temperature. The rate increases by nearly 10 times when the temperature is increased from 650oC to 900oC. The time taken for consumption of 50 gm of coal is 70 min, whereas at 650oC it is as long as 500 min. As the temperature increases the hydrogen produced in gasification reaction increases. The trend is shown in figure 3.
60
950 C
50
850 C
Mole percentage
Mole percentage
60
650 C
40 30 20 10 0
950 C
50
850 C 650 C
40 30 20 10 0
0
25
50
75
100
125
150
Time, min
(a) H2 at a steam flow rate of 6 gm/min
0
25
50
75
100
125
150
Time, min
( b) H2 at a steam flow rate of 4 gm/min
Figure 3: Effect of temperature on hydrogen produced in gasification The highest mole fraction of H2 in the product gas is 43% at 650oC and steam flow rate of 6 gm /min, whereas the highest mole fraction of H2 is 58% at a temperature of 950oC for the same steam flow rate. In the product gas, the hydrogen to carbon ratio at a given reaction time varies with respect to temperature and steam flow rate. It should be noted that the enhancement in water gas shift reaction at high temperature results into higher hydrogen to carbon ratio. It is observed that at high temperature (950oC) and low steam flow rate (4 gm/min), the ratio is around 1.3. At a fixed temperature, the decrease in steam flow rate increases the hydrogen to carbon ratio (For a temperature of 650oC and steam flow rate 4 gm/min the ratio is around 2). This is because at lower partial pressure of steam, gasification is relatively slow which makes the unreacted steam available for water gas shift reaction thereby increasing the hydrogen content in the product.
High temperature enhances carbon monoxide formation. At higher temperature, most of the steam gets converted due to gasification alone and hence the proportion of CO in the product is relatively high. On the other hand, at low temperature water gas shift reaction dominates over gasification, and hydrogen to carbon ratio increases. The comparison plot of mole fraction of carbon monoxide formed at different temperatures is shown in figure 4. Based on the experimental results, it is observed that the carbon monoxide is formed in the initial period only. As the unreacted coal in the bed decreases with time, the carbon monoxide formation decreases rapidly because the availability of unreacted steam promotes the water gas shift reaction. Hence, at longer reaction time, hydrogen to carbon ratio decreases even at higher temperature. 25
950 C 850 C
15
Mole percentage
Mole percentage
20
650 C
10 5 0
950 C 850 C
20
650 C
15 10 5 0
0
25
50
75
Time, min
(a) CO at steam flow rate of 6 gm/min
100
0
25
50
75
100
Time, min
(b) CO at steam flow rate of 4 gm/min
Figure 4: Effect of temperature on Carbon monoxide produced in Gasification The highest mole fraction of carbon monoxide in the product gas is 21.61% at the temperature of 950oC and steam flow rate of 4 gm /min, which drops down to 0.55% at a temperature of 650oC for the same steam flow rate. The maximum concentrations of carbon monoxide realized at different temperatures are shown in table 2. The effect of temperature on the production of carbon dioxide at a fixed steam flow rate is shown in figure 5. Carbon dioxide is produced from the water gas shift reaction. In the lower temperature range the amount of carbon monoxide produced is very less and much of the coal is converted to carbon dioxide because of the availability of steam for water gas shift reaction.
Table 2: Maximum amount of CO produced Temperature
Maximum CO produced (mole%)
950
21.61
900
20.85
850
9.09
800
7.39
650
0.55
35
950 C 850 C
28
Mole percentage
Mole percentage
35
650 C
21 14 7 0
950 C 850 C
28
650 C
21 14 7 0
0
25
50
75
100
Time, min
(a) CO2 at steam flow rate of 6 gm/min
0
25
50
75
100
Time, min
(b) CO2 at steam flow rate of 4 gm/min
Figure 5: Effect of temperature on Carbon dioxide produced in Gasification
Effect of steam flow rate on gasification At a fixed temperature, an increase in steam flow rate (i.e. steam partial pressure) enhances conversion of coal due to gasification reaction. High steam flow rate results in an increase in the extent of water gas shift reaction. As a result, the concentration of hydrogen and hence, the hydrogen to carbon ratio in the product gas also increase with increase in steam flow rate. The effect of steam flow rate on the hydrogen produced is shown in figure 6. On the contrary, an increase in steam flow rate has an inverse effect on the carbon monoxide production (figure 7).
60
6 gm/min
50
4 gm/min
Mole percentage
Mole percentage
60
2 gm/min
40 30 20 10 0
6 gm/min
50
4 gm/min 2 gm/min
40 30 20 10 0
0
40
80
120
160
0
Time, min
50
100
150
200
250
300
Time, min
(a) H2 in steam gasification at 9000c
(b) H2 in steam gasification at 8000c
Figure 6: Effect of steam flow rate on hydrogen production in gasification 15
25
6 gm/min
6 gm/min
4 gm/min
4 gm/min
2 gm/min
20 15 10 5
Mole percentage
Mole percentage
30
0 0
40
80
Time, min
(a) CO in steam gasification at 9000c
2 gm/min
10
5
0 0
40
80
Time, min
(b) CO in steam gasification at 8000c
Figure 7: Effect of steam flow rate on carbon monoxide production in gasification
Effect of coal particle size on steam gasification It is known that the particle sizes of coal used are different for different gasifiers. Moreover, in a process like underground coal gasification (UCG) one does not have a control over the particle size. Hence, it is important to study the internal diffusion and heat transfer limitations through the effect of particle size. To study the effect of size on steam gasification, single monolith blocks of different sizes 1.5 cm, 1.25 cm and 1 cm were prepared and gasification was performed by inserting them into the FBR. A photograph of a monolith is shown in figure 8. The concentrations of H2, CO and CO2 in
the product gas are shown in figure 9. The gasification rate increases with decrease in the particle size. The rate of hydrogen generation is more for the smaller size, as shown in figure 9a. The CO produced is maximum for smaller size particle at fixed temperature and stream flow rate. These results indicate that resistance to internal transport is relevant and may affect the performance considerably.
Figure 8: Monolith particle of coal.
12 Mole percentage
Mole percentage
40 30 1.5 cm 1.25 cm
20 10 0
9 1.5 cm 1.25 cm
6 3 0
0
20
40 Time, min
60
(a) H2 in gasification of monolith
80
0
20
40 Time, min
60
(b) CO in steam gasification of monolith
80
Mole percentage
15 12 9
1.5 cm 1.25 cm
6 3 0
0
20
40 Time, min
60
80
(c) CO2 produced in steam gasification of monolith Figure 9: Profile of H2, CO, CO2 of steam gasification of monolith CO2 gasification of char in fixed bed reactor In the case when steam is used as a gasifying agent, CO2 is formed due to the presence of water gas shift reaction. It is likely that it may independently participate in CO2 gasification i.e. the Boudard reaction (Eq. 2). It is necessary to know the contribution of CO2 gasification when steam is used as a gasifying agent. Gasification reaction is carried out independently at different temperatures with variation in CO2 flow rate. The gasification experiments were conducted at constant temperatures from 800 to 10000C with an increment of 50oC. In all these experiments, the inert gas flow was maintained at 100 ml/min. The flow rate of the reactant is varied from 100 ml/min to 300 ml/min. The effects of temperature and flow rate on the amount of CO generated during gasification are discussed in the subsequent sections.
Effect of Temperature on CO2 gasification The rate of consumption of coal increases with temperature. As the temperature increases the CO produced in gasification reaction increases for a fixed flow rate of reactant gas. The variation in the formation of CO with respect to temperature for the fixed amount of CO2 flow rate is shown in figure 10. The reaction is very sensitive to temperature. Moreover, as against gasification it is relatively slow.
60
36
Mole percentage
Mole percentage
45
1000 950
27
900
18 9
1000 C 950 C 900 C
50 40 30 20 10 0
0 0
100
200 300 Time, min
400
0
500
100
200
300
400
Time, min
(a) CO2 flow rate 100 ml/min
(b) CO2 flow rate 100 ml/min
Figure 10: Mole fraction of CO in the product gas for a fixed flow rate of CO2
Effect of CO2 flow rate on gasification At a fixed temperature an increase in CO2 flow rate (i.e. its partial pressure in the feed) increases the rate of consumption of coal during gasification reaction. The concentration of CO in the product gas at a fixed temperature and different flow rates of CO2 is shown in figures 11.
70
50 300 ml/min 200 ml/min 100 ml/min
50
Mole percentage
Mole percentage
60
40 30 20 10 0 0
50
100
150
200
250
300
Time.min
(a) CO2 gasification at 10000C
350
40
300 ml/min 200 ml/min 100 ml/min
30 20 10 0 0
100
200
Time, min
300
400
(b) CO2 gasification at 10000C
Figure 11: Mole fraction of CO in the product gas at a fixed temperature
Conclusions Steam and CO2 gasification of char is studied at different temperatures and reactant flow rates. It was observed that both the gasification reactions are very sensitive to changes in temperature and the reactant gas flow rate i.e. partial pressure. The extents of the reactions are favoured by high temperature and high partial pressure of the gasifying agent in the feed. Particle size has a significant impact on the rate, and internal diffusion effects are expected to play an important role in the performance of gasifiers that deal with relatively large coal blocks (e.g. Moving bed gasifiers, underground coal gasification). The proportion of CO compared to H2 in the product gas is more at higher temperature. CO2 gasification is much slower than steam gasification, and it may not contribute significantly when steam is used as a gasifying agent. Detailed modelling is in progress to explain the results obtained.
References 1. BP Statistical Review of World Energy-2009. Technical Report. bp Global, London, UK., 2009 2. Gregg D, Edgar TF. Underground coal gasification. AIChE J 1978;24:753 3. Khadse A, Qayyumi M, Mahajani SM, Aghalayam P. Underground coal gasification: a new clean coal utilization technique for India. Energy 2007; 32:2061 4. Elizabeth
Burton,
Julio
Friedmann,
Ravi
Upadhye,“Best
Practices
in
Underground Coal Gasification”, Lawrence Livemore National Laboratory(2004). 5. Wiktorsson, L.P. and Wanzl, W., “Kinetic for coal pyrolysis at low and high heating rates – a comparision of data from different laboratory equipment”, Fuel(2000), 79(6), 701–716. 6. Feng Yan , Leguan Zhang, Zhiquan Hu, Gong Cheng, Chengcheng Jiang , Yanli Zhang, Tao Xu, Piwen He, Siyi Luo, Bo Xiao, "Hydrogen-rich gas production by steam gasification of char derived from cyanobacterial blooms (CDCB) in a fixed-bed reactor: Influence of particle size and residence time on gas yield and
syngas composition", International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2010.07.113 7. Bazardorj Bayarsaikhan, Jun-ichiro Hayashi, Taihei Shimada, Chirag Sathe, Chun-Zhu Li, Atsushi Tsutsumi, Tadatoshi Chib," Kinetics of steam gasification of nascent char from rapid pyrolysis of a Victorian brown coal", Fuel 84 (2005) 1612 –1621 8. Nicolas
Piatkowski,
Christian
Wieckert,
Aldo
Steinfeld,
"Experimental investigation of a packed -bed solar reactor for the steam-gasification of carbonaceous feed stocks", Fuel Processing Technology 90 (2009) 360–366 9. Gerardo
Gordillo,
Kalyan
Annamal ai,
"Adiabatic
fixed
bed
gasification of dairy biomass with air and steam", Fuel 89 (2010) 384 - 391. 10. Jie Wang, Mingquan Jiang, Yihong Yao, Yanmei Zhang, Jianqi n Cao, "Steam gasification of coal char catalyzed by K2CO3 for enhanced production of hydrogen without formation of methane", Fuel 88 (2009) 1572–1579
4. CARBON MANAGEMENT, PRE-COMBUSTION CAPTURE Development of dry regenerable CO2 Sorbent and WGS catalyst for SEWGS Process Joong Beom Lee, Senior Engineering Specialist, Korea Electric Power Research Institute 305-380 Munji-ro 65, Yuseong-gu, Daejeon, KOREA, [email protected], 82-42-865-5254 Tae Hyoung Eom, Jungho Ryu, Jeom-In Baek, Dong-Hyeok Choi, Keun Woo Park, Seong Jegarl, Seug-Ran Yang
Abstract: IGCC integrated with CCS has been regarded as a promising option to reduce CO 2 emission under the situation that coal is the dominant source among fossil fuels for the electricity-generating units. Recently, KEPCO Research Institute proposed the new concept for the pre-combustion CO2 capture process, which is named as one loop sorption enhanced water gas shift (SEWGS) process consisted of two fluidized bed reactors. Sorption enhanced water gas shift (SEWGS) process is combined the water-gas shift reaction with CO2 capture at the same time. In this study, Nine MgO-based dry regenerable CO2 sorbents and Seven CuO-based water gas shift catalysts were prepared by spray-drying technique to evaluate their applicability to a fluidized-bed sorption enhanced water gas shift(SEWGS) process. In these sorbents, MgO-based sorbents satisfied most of the physical requirements for commercial fluidized bed reactor process along with reasonable chemical reactivity. All sorbents had a spherical shape, an average size of 112–148 m, and a size distribution of 45–250 m, a bulk density of 0.61–0.83 g/mL. The attrition Index (AI) of all the sorbent was below 15% except P-9 sorbent, compared to about 20% for commercial fluidized catalytic cracking (FCC) catalysts. CO2 sorption capacity of PC-4 was approximately 17.6 wt% at 200℃ and 21 bar with synthesis gas conditions. Spray dried CuO-based WGS catalyst showed relatively good physical properties. Most of catalyst had a spherical shape, an average size of 121–157 um, and a size distribution of 37–250 m, a bulk density of 0.92– 1.06 g/mL. Attrition Indices (AI) of PC-4, PC-12 and PC-13was below 10%, which is suitable for fluidized SEWGS process.
Introduction Fossile fuel combustion for power generation is responsible from a quarter to up to one-third of human made CO2 emissions. The advantage that one has in power generation, with respect to other anthropogenic CO2 sources, is because emissions are generated in centralized stationary sites. There is potentially a broad range of options for dealing with them. In short term, increasing power plant efficiency will offer the most benefits, however improving efficiency is not sufficient for meeting the CO2 mitigation requirement and new option must be developed for CO2 capture and sequestration. A key requirement for such new optional technologies is that they are efficient and integrated within the power plant, so that energy needs for CO2 capture are minimized. The most direct way of capturing CO2 from power plant is by separating it from the flue gas using conventional technologies, such as absorption, adsorption, or membrane technology, etc. Recently, IGCC (Integrated Gasification Combined Cycle) power plants are considered as one of promising technology to meet future CO2 environmental targets for environmentally benign power generation. In these plants, fossile fuels such as coal or natural gas, etc are gasified by oxygen or air to convert synthesis gas, which
is then processed in water-gas shift reactor system to produce additional H2. The exit stream contained CO2 must then be further treated to capture CO2. Conventionally, CO2 separation during water-gas shift (WGS) reaction carried out minimum four or more multi stage process. However, this process (concept) is energy intensive operation and requires substantial capital costs. In order to overcome this technical problem, KEPRI proposed the new concept, which is named as one loop sorption enhanced water gas shift process that is consist of two fluidized bed reactor. In fist reactor, CO in synthesis gas is converted to CO2 and H2 under reacting with water vapor on WGS catalyst, then, CO2 is simultaneously captured by solid regenerable CO2 sorbent in the same reactor. Sorbents captured CO2 are regenerated by water vapor or CO2 and water vapor mixture to separate concentrated CO2. The fluidized bed process requires a sorbent or catalyst which maintains its structural integrity during repeated use, because the attrition of materials can result from both physical attrition such as friction and collision as well as chemical transformations like volume changes caused by reaction. Therefore, solid sorbent or catalyst must have a high attrition resistance, high sorption capacity within a reasonable contact time (2–8 s), and a good flow characteristics like high bulk density. This work aimed to develop and verify the performance of sorbent and WGS catalyst to capture CO2 from coal-derived synthesis gas at temperature range of 200–500℃ and above 20 bar. Comprehensive data, such as attrition index, particle size, bulk density, reactivity, etc., required to evaluate the developed sorbent and catalyst, P and PC series, were presented.
Experimental 1. Sorbent and Catalyst Preparation The nine MgO-based sorbents and seven CuO-based catalysts consist of an active material, a support, inorganic binders, water as solvent, organic additives as dispersants, a defoamer, and an organic binder. Commercial grade MgO and CuO were used as the active ingredient for sorbent and catalyst, respectively. Nine MgO-based sorbents and Seven CuO-based WGS catalysts formulations designated by P and PC series were produced using a spray-drying technique that is easily scalable to commercial quantities to evaluate applicable to sorption enhanced water-gas shift (SEWGS) process. The spray-drying process of the sorbent includes the selection and comminution of raw materials, colloidal slurry preparation, spray drying, and calcination steps. Among these steps, the colloidal slurry preparation is the most important step for yielding sorbent particles suitable for fluidized-bed application. Colloidal slurry must be flowable, homogeneous, dispersed, and stable. These properties are controlled by the concentration, viscosity, and pH through the addition of organic additives such as dispersants. 2. Sorbent and Catalyst Characterization The shape of calcined sorbents at 550–700℃, was confirmed by industrial microscope provided by IMT. Average particle size of the sample was measured by using a MEINZER II sieve shaker. The measuring procedure for particle size and size distribution was followed by the American Society for Testing and Materials (ASTM) E-11. Tap density of each sorbent was determined by using the Autotap instrument (Quantachrome) proposed in ASTM D 4164-88 and was obtained by dividing the sorbent mass by its tapped volume. Attrition resistance of the calcined sorbents was measured with a modified three-hole air-jet attrition tester based on the ASTM D 5757-95. The attrition resistance was determined at 10 standard L/min (slpm) over 5 h as described in the ASTM method. The attrition index (AI) is the percent fines generated over 5 h. The fines are particles less than 20 μm collected at the thimble after 5 h from the start, which was attached to the gas outlet. AI = total fines collected for 5 h/amount of the initial sample (50 g) × 100.
(1)
The corrected attrition index (CAI) is the percent of fines generated over 4 h using the following equation: CAI = (total fines collected for 5 h – fines collected for first 1 h)/(amount of initial sample – fines collected for first 1 h) × 100.
(2)
The AI (CAI) of fresh Akzo and Davison FCC catalysts, which were used as references, were 22.5(18)% and 18.4(13.1)%, respectively, under the same conditions. In a fluidized-bed CO2 capture process, materials with an AI < 30% for a transport reactor or even < 60% for a bubbling fluidized-bed reactor would be acceptable for use in a flue gas under atmospheric pressure. A lower AI or CAI indicates a better attrition resistance of the bulk particles. The CO2 sorption capacity of CO2 sorbents and CO Conversion of WGS catalysts were assessed using a HTHP TGA and a fluidized-bed micro-reactor system, respectively. The sample size of each sorbent was about 20 mg and the flow rate of the reaction gas was 15 mL/min at STP. The reaction conditions and gas compositions were the same as those shown in Table 1. The CO2 sorption capacity was determined from the breakthrough test. Table 1. Sorption Capacity Test Conditions and Gas Compositions Items
Carbonation
Regeneration
Gas composition/vol%
37.0% CO2, 5.6% CO, 47.4% H2, 10.0% H2O N2 or N2 + H2O
Temperature/C
200
400
Pressure
21 bar
21 bar
Sample load/mg
20
20
Total gas flowrate/(ml/min)
60
60
Result and Discussion A series of MgO-based sorbents and CuO-based catalysts were prepared in order to screen the suitable formulation to fluidized bed SEWGS process. 1. CO2 sorbents During the sorbent screening process using the spray-drying technique, KEPRI prepared nine kinds of MgObased sorbent formulations, which are designated as the P series. These formulations used multi ceramic binder on 8 kg solid scale and were prepared by varying the organic additives and the operation conditions of the spray drier. All the formulations yielded spherical sorbents with a mean particle size of 108–148 m and a size distribution of 45–250 m. Figure 1 shows the shape of P series sorbents.
(a) P1
(b) P3
(c) P4
(d) P5
(e) P9
Figure 1. Microscope Images of CO2 Sorbents (500X).
Table 2 summarizes the physical properties of the spray-dried CO2 sorbents. The shape, bulk density, average particle size, and size distribution of the sorbent are essential parameters for evaluation of the fluidization and solid circulation characteristics in fluidized-bed reactor process. The bulk density of sorbents is not sufficient except P-3 sorbent, which is higher than 0.8 g/mL are recommended for fluidized-bed applications. The attrition indices (AI) of P-3 and P-5 sorbents showed superior attrition resistance with relatively high bulk density. The high attrition resistance of P series sorbents, which is superior to commercial fluidized catalytic cracking (FCC) catalysts, provides a low level of additional sorbent in the process because sorbents are decreased by attrition during CO2 capture. Table 2. Physical Properties of the P Series Sorbents Calcined at 600C Properties
P-1
P-3
P-4
P-5
P-9
Average particle size/m
124
112
108
114
148
Size distribution/m
45–250
45–250
45–250
45–250
45–250
Bulk density/(g/mL)
0.63
0.84
0.76
0.68
0.61
Attrition resistance/% AI (CAI) at 10 slpm
11.9(10.5)
4.4(2.3)
2.6(1.3)
9.2(8.0)
26.4(23.2)
All the developed sorbents were tested for chemical reactivity. Figure 2 shows the TGA sorption capacities of the P series sorbents, calcinations temperature is 600C, at 200C and 21 bar. The CO2 sorption capacity showed 6.7–17.6 wt%.
CO2 Sorption Capacity/wt%
20
16
12
8
4
0 P-1 P-2 P-3 P-4 P-5 P-6 P-7 P-8 P-9
Sorbents Figure 2. CO2 Soprtion capacities of P series sorbents clacined at 600C. 2. WGS (Water Gas Shift) Catalyst WGS catalysts, which are designated as the PC series were also prepared for this study. These formulations used multi ceramic binder on 8 kg solid scale and were prepared by varying the organic additives and the operation conditions of the spray dryer. All the formulations yielded semi-spherical sorbents with an average particle size of 132–150 m and a size distribution of 53–250 m. Figure 3 shows the shape of PC-4 and PC-13 catalysts, respectively.
(a) PC-4 (b) PC-13 Figure 3. Microscope Images of WGS Catalysts (500X).
Table 3 shows a comparison of the physical properties of the WGS catalysts. The range of bulk density of catalysts is 0.92 to 1.06 g/mL. The attrition indices of the catalysts showed that the attrition index (AI) and the corrected attrition index (CAI) were range of 7.4(5.3) and 45.5(43.1). In these results, PC-4, PC-12 and PC-13 catalyst calcined at 550C is available for fluidized-bed SEWGS applications. Table 3. Physical properties of the PC series catalysts calcined at 550C Properties PC-4 PC-9 PC-10 PC-11
PC-12
PC-13
Average particle size/m
132
157
121
124
121
133
Size distribution/m
42-250
37-250
37-250
37-250
37-250
42-250
Bulk density/(g/mL)
0.93
0.94
0.93
0.92
1.05
1.06
Attrition resistance/% AI (CAI) at 10 slpm
7.4(5.3)
24.5(20.2)
36.2(32.4)
45.5(43.1)
9.2(6.9)
8.2(6.5)
Conclusion Nine CO2 sorbents and seven WGS catalysts were prepared and screened using a standard evaluation method. All the CO2 sorbents appeared to satisfy the basic physical requirements for fluidized-bed application. However, their CO2 sorption capacity is needed to improve using various methods such as incorporated additives etc. In case of WGS catalyst, the forming of WGS catalyst by spray drying technique demonstrates a successful, thereby showing the possibility for fluidized-bed process. All catalysts satisfied the basic physical requirements for the fluidized-bed reactor. Especially, PC-4 catalyst shows excellent attrition resistance of 7.0%.
Acknowledgement This work was supported by Energy Efficiency and Resources R&D program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korea government Ministry of Knowledge Economy (2008CCD11P020000), Korea Electric Power Corporation (KEPCO) and Korea Western Power co., Ltd.
References 1. Joong B. Lee, et al, “Screening of Zinc-Based Sorbents for Hot Gas Desulfurization”, Energy & Fuels 2008, 22, 1021-1026. 2. Joong B. Lee, et al, “Highly Attrition-Resistant Zinc Oxide-Based Sorbents for H2S Removal by SprayDrying Technique”, Ind. Eng. Chem. Res. 2008, 47, 4455-4464. 3. Joong B. Lee, et al, “Sodium-Based Dry Regenerable Sorbent for Carbon Dioxide Capture from Power Plant Flue Gas”, Ind. Eng. Chem. Res. 2008, 47, 4465-4472. 4. Soo C. Lee, et. al, “Development of Regenerable MgO-Based Sorbent Promoted with K2CO3 for CO2 Capture at Low Temperatures”, Environ. Sci. Technol. 2008, 42, 2736-2741.
5. Chang-Keun Yi, Sung-ho Jo, Yongwon Seo, Joong-Beom Lee, Chong-Kul Ryu, “Continuous operation of the potassium-based dry sorbent CO2 capture process with two fluidized-bed reactors”, International J. Greenhouse Gas Control, 2007, 1, 31-36
4. CARBON MANAGEMENT, PRE-COMBUSTION CAPTURE
Development of dry regenerable CO2 sorbent for Fluidized-bed CO2 Capture process from coal power plant Chong Kul Ryu, Chief Engineering Specialist, Korea Electric Power Research Institute 305-380 Munji-ro 65, Yuseong-gu, Daejeon, KOREA, [email protected], 82-42-865-5250 Joong Beom Lee, Tae Hyoung Eom, Bok Suk Oh, Jeom-In Baek, Kyeongsook Kim, Young Ho Wi, Jegarl Seong, Won Sik Jeon
Abstract:
This paper summarizes the results of performance of dry regenerable sorbent for CO 2 capture at low temperature. Each sorbents was prepared by spray drying techniques to identify potential materials with Na2CO3 and K2CO3 as active material. The physical properties, particle size distribution and average particle size, bulk density, BET, Hg porosity and shape, of the spray-dried sorbents were investigated by standard methods and the attrition resistance of the sorbents for circulating fluidized-bed application was measured with a modified three-hole air-jet attrition tester based on the ASTM D 5757-95. Sorbent chemical reactivity was measured in low temperature range (carbonation at 50oC~70oC and regeneration at 120oC~140oC) with simulated flue gas containing 10 vol% H2O and 14.4 vol% CO2 with simultaneous thermo gravimetric analyzer (TGA). The maximum CO2 sorption capacity of sorbents was approximately 12wt% and attrition index (AI) reached below 1% in case of K-based sorbents and 35% in case of Na-based sorbent. The results of physical properties and CO2 sorption capacities showed excellent characteristics for circulating fluidized-bed process application to capture CO2 from flue gas condition.
Introduction Carbon dioxide is a major greenhouse gas that is released into the atmosphere from combustion of fossil fuels used for industries such as power generation and steel production. There are three options for CO2 emission reduction effort such as efficiency improvements, conversion to fuels with less carbon emission, as well as direct carbon dioxide capture and storage. Recently, the effort for capturing CO2 is focusing on large point sources, fossil fuel fired power plants contributing approximately 40% of total global emissions [1-3]. To capture CO2, there are three technological pathways that are applicable for the power generation system: postcombustion, pre-combustion and oxy-combustion [2]. One of the advanced concepts, cost-effective and energy-efficient techniques, for the capture of CO2 is a chemical absorption process with dry regenerable solid sorbents [4]. The CO2 capture process using dry regenerable sorbents consists of two reactor, carbonation and regeneration reactor. The reaction [1] in each reactor is followed: Carbonation: M2CO3 + CO2 + H2O 2MHCO3 + ΔH (1) Regeneration: 2MHCO3 M2CO3 + CO2 + H2O (2) Where M is Na or K and the reaction enthalpy of (1) for Na and K is –31.69 and –33.74 kcal/mol, respectively. The capture process of fluidized-bed and transport reactor with dry regenerable CO2 sorbents was employed to accommodate the massive treatment of flue gas from fossil fuel-fired power plants, as well as the exothermic nature of carbonation reaction. This process has several advantages over other processes: better gas contact with
smaller solid sorbents, ease of sorbent makeup and removal, easy control of the exothermic carbonation reaction temperature, and continuous steady operation [1]. The CO2 sorbent applicable to the fluidized-bed/transport reactor process requires the characteristic which maintains the reactivity, the durability, the attrition resistance such as friction and collision as well structural transformation like volume changes caused by chemical reaction. Therefore, the characteristics must have high attrition resistance during multicycle, the high sorption capacity within reasonable contact time (2 ~ 8 s), and the good flow characteristic (high bulk density) [1, 5]. This paper will present chemical reactivity along with important physical properties of seven potassium-based solid sorbents and six sodium-based solid sorbents prepared by spray-drying technique. Experimental 1. Sorbent Preparation Each sorbents, K-based and Na-based solid sorbents, were prepared with spray dryer in batches of approximately 8kg. The raw material of sorbents was composed of an active material, Na2CO3, K2CO3 and KHCO3, a support, inorganic binders, organic additives and an organic binder. The sorbent preparation process consists of several steps: (1) mixing the raw materials in water with dispersants, (2) comminution of the raw materials and colloidal slurry preparation with a high-energy bead mill, and (3) spray drying to form a spherical-shaped green body, followed by the pre-drying and calcination at 550oC under air for 3h. Among these steps, the colloidal slurry preparation is the most important step for yielding sorbent particles suitable for fluidized-bed application. Colloidal slurry must be flowable, homogeneous, dispersed, and stable. These properties are controlled by the concentration and viscosity of the slurry through the addition of organic additives such as dispersants. The total formulation contents of the active solid materials, sodium carbonate (Na2CO3) potassium carbonate (K2CO3) and potassium bicarbonate (KHCO3), for each sorbents were approximately 30 wt% and 35 wt%, respectively.
2. Sorbent Characterization The physical properties of the thirteen sorbents prepared by a spray-drying were characterized. Scanning electron microscopy (SEM) and industrial video microscope provided by IMT was used to examine the shape. Sieve shaker (MEINZER II) method based on the ASTM E used to measure particle size and size distribution (PSD) and the bulk density was examined with a tap density meter (ASTM D 4164-88). Standard BET and Hg porosimeter were used to determine specific surface area and porosity, respectively. The attrition resistance is one of the critical parameters in the development of a sorbent for fluidized bed process, because any attrition of the sorbent causes the loss of active material resulting in lower product quality. Attrition can also result in the need for additional filtration and cause plugging as well as affect the fluidization and solid circulation properties. The attrition resistance of the sorbents for fluidized-bed applications was measured using a modified three-hole air-jet attrition tester based on the ASTM D 5757-95. The attrition index (AI) was determined at 10 slpm (standard l/min) over 5 h, as described in the ASTM method. The attrition index is the percentage fines generated over 5 h: AI = total fines collected for 5 h/amount of the initial sample (50 g) × 100.
(1)
The corrected attrition index (CAI) is the percent of fines generated over 4 h using the following equation:
CAI = (total fines collected for 5 h – fines collected for first 1 h)/(amount of initial sample – fines collected for first 1 h) × 100. (2) The AI (CAI) of fresh Akzo and Davison FCC catalysts, which were used as references, were 22.5(18)% and 18.4(13.1)%, respectively, under the same conditions. In a fluidized-bed CO2 capture process, materials with an AI < 30% for a transport reactor would be acceptable for use in a flue gas under atmospheric pressure. A lower AI or CAI indicates a better attrition resistance of the bulk particles. The chemical reactivity of each sorbents was assessed using a thermogravimetric analyzer (STA 1500, Reometircs). In a TGA test, the weight of a powder sample hanging from microbalance is monitored as a function of time. The weight gain is proportional to the extent of the reaction in the sample. The sample size of each sorbent was 10 mg and the flow rate of the reaction gas was 60 mL/min at NTP. The reaction conditions and gas compositions were the same as those shown in Table 1. The CO2 sorption capacity was determined from the increase in weight during carbonation. Table 1. TGA Test Conditions and Gas Compositions Items
Carbonation
Regeneration
Gas composition/vol%
14.4% CO2, 5.4% O2, N2 Balance (7~10 % water vapor)
N2 or N2 + water vapor
Temperature/C
50 or 70
120 or 140
Pressure
Ambient
Ambient
Sample load/mg
10
10
Total gas flow rate/ml/min
60
60
Figure 1. Schematic diagram of the TGA gas supply system: (TIC) temperature indicator and controller, (EB) electronic balance.
Result and Discussion In this study, the physical properties of the prepared sorbents have size distribution of 40~300μm, average size of approximately 100~160μm for Na-based sorbents and 90~150μm for K-based sorbents. The bulk density is estimated the range of 0.62~1.06 g/mL for Na-based and K-based sorbents. The Images of SEM and industrial video microscope showed spherical or semi-spherical shape. Figure 2 shows the shape of K-based sorbents, which were calcined at 550C for 3 hours.
(A)
(B)
(C)
(D)
Figure 2. SEM Photographs (300X) of the spray-dried sorbents: (A) Sorb K-B, (B) Sorb K-C, (C) Sorb K-D, (D) Sorb K-E. Table 2 and 3 summarizes the physical properties of the spray-dried sorbents. The shape, bulk density, size, and size distribution of the sorbent are essential parameters for evaluation of the fluidization and solid circulation characteristics in fluidized-bed reactor process. The bulk density of seven kinds of Na and K-based sorbents is sufficient, which is higher than 0.8 g/mL are recommended for fluidized-bed applications. Each sorbents could identify porous material from the results of BET and porosity. Table 2. Properties of the Na-based Sorbents Calcined at 550C Properties
Na-A
Na-B
Na-C
Na-D
Na-E
Na-F
Average particle size/m
121
104
124
158
128
129
Bulk density/g/mL
0.84
0.80
0.68
0.62
0.64
0.64
BET surface area/ m2/g
64.25
68.87
37.3
32.9
17.8
6.2
Hg Porosity/%
69.53
68.2
69.2
80.7
65.7
73.84
Table 3. Properties of the K-based Sorbents Calcined at 550C Properties
K-A
K-B
K-C
K-D
K-E
K-F
K-G
Average particle size/m
92
134
133
121
132
150
111
Bulk density/g/mL
1.06
0.88
0.95
0.84
0.66
0.69
0.81
BET surface area/ m2/g
34.92
40.7
41.6
50.0
54.9
42.0
37.0
Hg Porosity/%
63.1
69.1
55.2
71.6
66.7
64.8
60.7
The attrition indices (AI) of K-based sorbents by ASTM D 5757 (10 slpm) reached below 9% compared to about 20% of commercial fluidized catalytic cracking (FCC) catalysts. Especially, The AI of sorbents made of K2CO3 was superior to the sorbents made of KHCO3 and it is below 9%. All of the AI data for K-based sorbents was satisfied with the requirement of circulating fluidized-bed reactor. The AI of most of the Na-based sorbents, however, was not satisfied with the requirement of circulating fluidized-bed reactor. The results showed over 30% but the only two was below approximately 5% and is showed in Figure 3.
Figure 3. AI and CAI of spray-dried sorbents.
The CO2 sorption capacities of Na-based sorbents were about 7~12 wt% and the maximum utilization of the sorbents was 93%, and the results of K-based sorbents were about 2~9 wt% and 78% in the simulated flue gas. These results that Na-based sorbents are higher sorption capacity than K-based sorbents could be confirmed by theoretical capacity from thermodynamic data. Also, the CO2 sorption capacity of sorbents made of K2CO3, KA~K-D, is higher than KHCO3, K-E~K-F, could be explained with the difference in dispersion of active components by solubility into the structure of support material. Figure 4 shows the TGA CO2 sorption capacities for them. It was also observed that optimal sorption capacity of Na-based sorbents occurred at about 50oC with 10 vol% H2O while that of K-based sorbent occurred at 70oC with less than 7 vol% H2O. It is to note that the characteristics of CO2 sorption capacity of sorbent are very sensitive to water vapor content in the simulated gas.
Figure 4. CO2 sorption capacity of spray-dried sorbents. Conclusion KEPRI has been successfully spray-dried several sorbents that showed superior attrition resistance and high CO2 sorption capacity along with their high bulk density. These results are satisfied with the requirements for circulating fluidized-bed reactor as CO2 capture process using dry regenerable sorbent to reduce total capital and operating costs. Sorbent development work will continue with particular emphasis on improvement of bulk CO2 removal and chemical durability in continuous operation of real flue gas, the control of the regeneration temperature, optimization of compositions, scale-up and demonstration. Acknowledgement This research was supported by a grant from the Carbon Dioxide Reduction & Sequestration Research Center, a 21st Century Frontier Programs funded by the Ministry of Education Science and Technology of the Korean government and Korea Electric Power Corporation (KEPCO) and its five independent subsidiary fossil power companies. References 1. Joong B. Lee, Chong K. Ryu, Jeom-In Baek, Ji H. Lee, Tae H. Eom and Sung Hyun Kim, Sodium-Based Dry Regenerable Sorbent for Carbon Dioxide Capture from Power Plant Flue Gas, Ind. Eng. Chem. Res., 2008, 47, 4465-4472. 2. Jose D.Figuera, Timothy Fout, Sean Plasynski and Howard Mcilvried, Advances in CO 2 capture technology-The U.S. Department of Energt’s Carbon Sequestration Program, International Journal of Greenhouse Gas Control, 2008, 2, 9-20. 3. David A. Green, Brian S. Turk, Jeffrey W. Portzer, Raghubir P. Gupta, William J. McMichael and Thomas Nelson, Cerbon dioxide capture form flue gas using dry regenerable sorbents, Quarterly Technical Progress Report, RTI, July 2004. 4. Soo Chool Lee, Bo Yun Choi, Tae Jin Lee, Chong Kul Ryu, Young Soo Ahn and Jae Chang Kim, CO 2 absorption and regeneration of alkali metal-based solid sorbents, Catalysis Today, 2006, 111, 385-390. 5. Ryu, C. K., Lee, J. B., Eom, T. H., Oh, J. M. and Yi, C. K., Characterization of Sodium-Based Sorbents for CO2 Capture from Flue Gas, In The Proceedings of 21st Annual International Pittsburgh Coal Conference, Osaka, Japan, Sep 15-19, 2004. 6. Soo Chool Lee, Ho Jin Chae, Soo Jae Lee, Bo Yun Choi, Chang Keun Yi, Joong Beom Lee, Chong Kul Ryu and Jae Chang Kim, Development of regenerable MgO-based sorbent promoted with K2CO3 for CO2 capture at low temperature, Environ. Sci. Technol., 2008, 42, 2736-2741. 7. White, C. M.; Strazisar, B. R.; Granite, E. J.; Hoffman, J. S.; Pennline, H. W. Separation and Capture of CO2 from Large Stationary Sources and Sequestration in Geological Formations-Coalbeds and Deep Saline Aquifers. J. Air Waste Manage. Assoc. 2003, 53, 1172.
A COMPARISON OF EXPERIMENTAL AND THEORETICAL HIGH PRESSURE CO2 ISOTHERMS ON COALS FROM THE UPPER SILESIAN BASIN, CZECH REPUBLIC Zuzana Weishauptová, Oldřich Přibyl and Martina Švábová Institute of Rock Structure and Mechanics, Academy of Sciences of the Czech Republic, 182 09 Prague 8, Czech Republic Tel: +420266009305, Fax: +420284680105, e-mail: [email protected] Abstract A manometric high pressure sorption apparatus was constructed. Pilot tests of the apparatus were carried out in the subcritical region below 6 MPa pressure and the results were compared with the results of low pressure measurements extrapolated to the same pressure. Both methods provided comparable results. INTRODUCTION One way to achieve a reduction of CO2 emissions is by trapping the CO2 discharged into the atmosphere and storing it in the long term in suitable repositories, including coal-bearing strata. The choice of a suitable high-capacity locality for CO2 storage within a coal seam requires knowledge not only of the geological conditions within the seam, but also of the textural properties of the coal, which control the mechanism and extent of the storage. Sorption processes (sorption in micropores and on the surface of meso/macropores, and absorption) occurring in the porous system of coal can be considered as the main mechanism of CO2 deposition. The CO2 sorption capacity of coal is generally indirectly determined by the adsorption isotherms at pressures and temperatures corresponding to in situ conditions. Such determination can be achieved by the low pressure sorption technique, by setting the CO2 partial volumes according to its binding and storage mode. The sorption capacity is determined by extrapolation to saturation pressure as the sum of individual partial volumes (1). More common way of determination uses high pressure sorption, which allows simulating the in situ conditions by measuring the adsorption isotherm within the pressure range from ones to tens MPa (2-4). The present study has been aimed at verification of results of the high pressure isotherm measurements with a new manometric sorption apparatus by comparison with results obtained by extrapolation of low pressure measurements to higher pressures. EXPERIMENTAL The study was carried out with 3 samples from a borehole surveying in the Czech part of the Upper Silesian Basin denoted as M1, M2 (medium volatile bituminous coal), and H1 (high volatile bituminous coal). Their basic characteristics are presented in Table 1. Table 1 Basic characteristic of coal samples Sample Ad Vdaf Ro (%) M1 2.2 29.2 1.04 M2 8.9 29.1 1.10 H1 2.5 33.6 0.81 Ad – ash content (dry basis), Vdaf – volatile matter (dry and ash free basis), Ro – reflectance 1
The high pressure sorption measurements were carried out using an original manometric apparatus, whose pressure transducer works with accuracy of 0.05 % within the entire pressure range, and which maintains a constant temperature with accuracy ± 0.1 °C. A scheme of the apparatus is depicted in Fig.1 (5).
Fig.1 Schematic diagram of high pressure sorption apparatus: S – sample cell, R – reference cell, P – pressure transducer, T – temperature controlled oven, I – ISCO syringe pump, Z – vacuum pump, W – constant temperature water bath, M – pressure gauge, D – data aqcuision, V1 – V4 – 2-way valves, V5, V6 – 4-way valves, V7 – 3-way valve Sorption experiments were carried out under subcritical conditions, at the temperature 25 °C and pressure below 6 MPa. The sorbed amount was expressed by the Gibbs equation. After construction of the theoretical high pressure isotherm, micropore volume and meso/ macropore surface area were determined. Micropore volume was evaluated by the Dubinin equation of volume filling from the CO2 isotherm measured at 25 °C within the pressure range 0.01 - 0.15 MPa using a gravimetric sorption analyzer IGA 002 by Hiden. Textural parameters of meso- and macropores were determined by the high pressure mercury porosimetry using a Porosimeter 2000 Carlo Erba. The theoretical high pressure isotherm was constructed as described previously (1). RESULTS The excess adsorption isotherms of CO2 on studied coals were evaluated by fitting the adsorption isotherm data to the standard Langmuir equation. The agreement between the calculated and experimental values was good in the given pressure range where the initial excess isotherm is monotonically increasing (see Fig. 2). Comparison of the measured high pressure isotherms with the theoretical isotherms constructed on the basis of the sum of the gas adsorbed in micropores and on the surface of meso- and macropores was done using linearized Langmuir isotherms (see Fig. 3). The characteristic parameters of Langmuir equation are presented in Table 2.
2
Fig.2 Fit-curves obtained from excess adsorption data using the Langmuir equation.
Fig.3 Comparison of experimental and theoretical linearized Langmuir isotherms.
Table 2 Langmuir parameters of the experimental and theoretical isotherms (mmcb) Sample Isotherm Experimental Theoretical VL PL SL R VL PL SL R 3 2 3 2 (cm /g) (MPa) (m /g) (cm /g) (MPa) (m /g) M1 31.34 0.57 216 0.9988 29.85 0.24 203 0.9999 M2 28.01 0.72 190 0.9927 24.39 0.35 166 0.9998 H1 30.39 0.56 206 0.9994 28.82 0.23 196 0.9999 VL - volume of gas adsorbed in a monolayer, PL – Langmuir pressure, SL – Langmuir surface area, R – correlation coefficient; mmcb - mineral matter containing basis
3
The certain part of differences in the parameters of the experimental and theoretical isotherms given in Table 2 could be due to the different time required for equilibration, which is, regarding the sample masses used, in the order of minutes for the low pressure sorption and in the order of hours for the high pressure sorption. It can be assumed that in the case of long contact of carbon dioxide with the coal matrix, the amount of the gas absorbed within the coal matrix increases, and it simultaneously linearly increases with increasing pressure (2). The theoretical isotherm does not take into account the absorbed amount. That is why the adsorption capacities determined from the low pressure experiments can be considered as less exact. CONCLUSION The high pressure sorption apparatus has significantly extended possibilities of investigation of the interaction of coal matrix with carbon dioxide. The present study has verified proper function of the apparatus. Acknowledgement This work has been supported by the Czech Science Foundation under Contract No. 106/08/1146. REFERENCES 1. Weishauptová Z., Medek J.: Bond forms of methane in coal. Fuel 1998; 77: 71-76. 2. Ozdemir E., Morsi B.I., Schroeder K.: CO2 adsorption capacity of Argonne premium coals. Fuel 2004; 83: 1085-1094. 3. Goodman A.L. et al.: Inter-laboratory comparison II: CO2 isotherms measured on moisture-equilibrated Argonne premium coals at 55 oC and up to 15 MPa. Int J Coal Geol 2007; 72: 153-164. 4. Gensterblum Y. et al.: European inter-laboratory comparison of high pressure CO2 sorption isotherms. I. Activated carbon. Carbon 2009; 47: 2958-2969. 5. Přibyl O., Švábová M., Weishauptová Z.: High pressure sorption apparatus for the determination of CO2 sorption on carbonaceous materials. Submitted to Chem.listy, 2010 (in Czech).
4
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 SESSION 22 COAL SCIENCE: CBM/CARBON DIOXIDE CONVERSION OF UKRAINIAN LOW GRADE SOLID FUELS WITH CO2 CAPTURE О.М. Dudnyk, Ph.D., Senior Researcher Coal Energy Technology Institute of National Academy of Sciences of Ukraine Andriivska St., 19, Kyiv, Ukraine, 04070 Phone: +38-044-425-0419, Fax: +38-044-537-2241, Email: [email protected] I.S. Sokolovska, Researcher Coal Energy Technology Institute of National Academy of Sciences of Ukraine Andriivska St., 19, Kyiv, Ukraine, 04070 Phone: +38-044-425-0419, Fax: +38-044-537-2241, Email: [email protected] Abstract The processes of generation of synthesis gas with high hydrogen content from the Ukrainian low grade solid fuels are studied at the Fuel Cell Test Installation. In case of the use of coals (bituminous coal of Lvov-Volyn coal deposit and brown coals of Alexandria and Zhytomyr deposits), there were studied the processes: steam conversion of volatile to produce high-reactivity char; char steam gasification using lime to produce synthesis gas with high hydrogen content. In case of the use of wood waste (from Kyiv region), there were studied the processes: carbonization of wood waste pre-impregnated with water solutions of H3PO4, FeCl3, and CaCl2 to produce wood coal; wood coal steam gasification using lime to produce synthesis gas with high hydrogen content and activated char coal. A special technique for determination of CO2 absorption and desorption degrees in the steam gasifier was developed for experimental evaluation of char steam gasification. The IGCC circuits with fuel processor, electrochemical fuel cell generator, steam and gas turbines are developed. In case of utilization of bituminous and brawn coals, it is proposed to use the batch fuel processor that can operate in the modes of combustion (desorption of CO2) and steam gasification (absorption of CO2). In case of biomass utilization, it is proposed to use fuel processor consisting of the system for special preparation of wood waste, carbonizer for wood coal production, gasifier for synthesis gas and activated char coal production, system of CO2 absorption for an increase in hydrogen concentration in the product gas and CO2 desorption for absorbent revivification. The research work was carried out within the framework of two projects of the Complex Program of Scientific Researches of the National Academy of Sciences of Ukraine "Fundamental Problems of Hydrogen Energy" at the Fuel Cell Test Installation with elaboration of the before prepared approaches for the development of the Fuel Cell IGCC circuits.
1. Introduction The development and use of new high technologies with capture of carbon dioxide, sulfur compounds, and other harmful substances directly in the process of solid fuel conversion make it possible to reduce significantly the expenses for used reagents, equipment, and maintenance. In this approach, specific quantity of metal, volume of conversion devices as well as the land area needed to accommodate new equipment for capture of carbon dioxide and other harmful substances will decrease. The decision of the scientific and technical challenges of obtaining hydrogen from the coal is directly linked to the development of new energy-efficient and environmentally friendly technologies. The ultimate goal of this research is to develop new coal-fired TPPs without chimneys with full utilization of harmful emissions and carbon dioxide. Studies of steam conversion of PET coke, wood wastes, and Ukrainian power-generating high-ash coals (brown, bituminous, and anthracite) were carried out in Coal Energy Technology Institute of NASU [1-4]. This paper shows the last main results of steam conversion of Ukrainian solid fuels using lime and catalysts for generation of synthesis gas with high hydrogen content from bituminous, brown, and wood coals. The results of hydrogen generation from bituminous and brown coals are also shown in this paper. 1
Biomass can be used efficiently and environmentally friendly in the process of biomass conversion to solid fuel with following transformation of solid fuel to the hydrogen for usage in modern high-performance power installations that are based on fuel cells. Production of bio-coal from waste of biomass is the perspective process for future technology development. Obtained bio-coal does not contain tars and compounds of sulfur, and has ash content less than 10 %. Ash content of wood on the dry basis is less than 4 %. Volatile matter in recalculation on combustible weight is 70-90 %. Fresh wood contains 40-60 % moisture. To increase heat values of wood products and to decrease energy consumption it is necessary the following: - to remove moisture before long transportation of wood for reduction of transport costs; - to use modern ways of catalytic and non-catalytic carbonization for volatile carbon conversion to fixed carbon; - to prevent carbon and hydrogen oxidation by oxygen that is contained in the initial wood; - to produce high reactivity wood coal, activated charcoal and other valuable products. 2. Fuel and lime Table 2.1 shows the results of analysis of solid fuels used in the studies. Table 2.1 – The results of the fuel analysis
Fuel Bituminous coal of Lviv-Volyn coal basin Brown coal of Korostyshiv deposit of Zhytomyr region Brown coal of Alexandria deposit of Кirovograd region Freshcut sticks of birch of Kyiv region
Moisture
Proximate analysis Volatile Ash matter content wt, %
Ultimate analysis(daf) Fixed carbon
C
H
O(dif)
N
S
wt, %
6.45
28.95
26.8
37.8
81.6
5.4
10.2
1.5
1.3
11.06
36.47
25.60
26.87
63.6
5.8
26.0
0.8
3.8
9.62
47.66
6.60
36.12
62.7
5.9
26.2
0.8
4.4
40.71
48.90
0.22
10.17
48.9
6.4
44.6
0.1
0.0
The size of coal particles in the experiments was 0.63-1.0 mm. Freshcut sticks of birch were dried. After that, bark was separated on internal trunks. Trunks were cut up on plates of 10-15 mm in thickness. These plates were crushed to particles of maximum size to 17 mm. Grind building magnesia lime was used in the experiments. The size of lime particles was 0.4 to 1.6 mm. The content of active CaO in the lime was 85.1 wt. %, active MgO - 7.1 wt. %. Yield of carbon dioxide from the lime was 2.4 wt. %. 3. Description of experimental installation 3.1 Fuel Cell Test Installation Fuel Cell Test Installation (FCTI) has been created gradually and has different versions. Initially, an installation to study steam conversion of fossil fuels was created; then new devices were added to the installation: CO conversion reactor with an independent water supply system and CO2 absorption lowtemperature reactor to generate hydrogen. Combination of created devices depends on research tasks [1-3].
2
The installation for investigation of hydrogen generation from solid fuels consists of such basic equipment: water storage and supply systems; steam generator; reactor for solid fuel gasification and lime conversion; systems of purification, cooling, and drying of synthesis gas; water-shift reactor, and low-temperature reactor for carbon dioxide absorption. Fig. 3.1 shows the basic circuit of this installation. Distilled water from the tank 1 is directed to the water feeder 2. Supply of inert gas (nitrogen or argon) into the tank 1 over a level of distilled water and into the water feeder 2 is provided for installation works under pressure. Feedwater from the feeder 2 goes to the steam generator 3 for obtaining superheated steam. Steam superheated on ceramic materials passes through the three-way valve and compensator in the steam gasification reactor 10. Superheated steam is fed to the reactor 10, which located in vertical tubular electric furnace. Superheated steam that is generated in the steam generator 3 is overheated additionally on ceramic bed in the reactor 7. Solid fuel and adsorbent from the bunkers 12 and 13 is fed into the reactor 10. Bunkers 12 and 13 are used in case of need to continue research using the selected type of solid or other fuels after the experiments with preloaded material.
Fig. 3.1. The basic circuit of FCTI in the mode of hydrogen generation: 1 - tank with distilled water for gasifier, 2 - water feeder for gasifier, 3 - steam generator 4 - three-way valve, 5 - condenser, 6 - compensator, 7 - ceramic bed, 8 - reaction zone, 9 - safety valve, 10 – gasifier, 11 nitrogen cylinder, 12 - solid fuel bunker, 13 - absorbent bunker, 14 - separator, 15 - bunker with solid residue of conversion, 16 - synthesis gas cooler, 17 - tank for condensate, 18 - dehumidifier, 19 - tank with distilled water for shift reactor, 20 - water feeder for water-shift reactor, 21 - water-shift reactor, 22 - bed of ceramic, 23 - bed of middle-temperature catalyst, 24 - bed of low-temperature catalyst; 25 - gas cooler; 26 reactor for carbon dioxide absorption; 27 - burner After preliminary desulphurization and dust separating, wet synthesis gas from the separator 14 is directed to the synthesis gas cooler 16. Synthesis gas is fed to the top of cooler and passes through the inner volume of downcomers. Then the synthesis gas goes from the cooler 16 to the dehumidifier 18, filled with silica gel to remove the rest of moisture. The condensate is collected in the bottom of the cooler 16 and then goes to the tank for condensate 17. The obtained dry synthesis gas is directed to the top of the water-shift reactor 21. A tube of heat-resistant stainless steel is placed in the electric furnace of water-shift reactor. The tube is filled top-down with bed of ceramic material 22 (for steam generating and steam-gas mixture overheating), bed of middle-temperature catalyst 23 and bed of low-temperature catalyst 24 (for CO conversion by water-shift reaction). Independent 3
system of water conservation and water supply is used for water-shift reactor operation. The system consists of distilled water tank 19 and feeder 20. After CO conversion, obtained gas passes through the cooler 25 (for steam residues condensation) in the reactor for carbon dioxide absorption 26 (for hydrogen production). The obtained hydrogen is burned in the burner 27. An independent power supply of feedwater feeder, steam generator furnaces, reactor-gasifier, and watershift reactor is used to ensure uninterrupted operation of the installation. Composition of obtained dry gas after the dehumidifier 18 and before the burner 27 is determined using chromatograph. Data collection system on the base of computer is used at the installation. 3.2 Facility for wood waste carbonization Circuits of two modification of facility for wood waste carbonization are shown in Fig. 3.2. Four samples of wood (one is clean, three are impregnated with water solutions of H3PO4, CaCl2 and FeCl3) were loaded into the carbonization module, which is divided by metal walls to 4 equal sections. Each section has its own filter. Carbonization module is placed in muffle furnace on a rest. Temperature in the muffle furnace is measured by thermocouples that is connected to a personal computer by means of interface (Fig. 3.2 a). Wood coal yield is determined from difference of sample weight before and after carbonization and technical analysis.
Fig. 3.2. Facility for wood waste carbonization: a) for four different samples, b) for one sample After the determination of wood waste sample with maximum wood coal yield, all sections are filled only by this sample of wood, and counterbalance to determine the dynamics of pre-impregnated wood waste carbonization is used in the installation (Fig. 3.2 b). The obtained wood coal is used for production of hydrogen-rich gas and activated charcoal at FCTI. 4. Special technique for estimation of char steam gasification Standard methods of physical and chemical analysis of substances and own techniques developed in Coal Energy Technology Institute were used in the work. To expedite the component analysis of dry synthesis gas obtained in the process of char steam gasification, there was deduced and checked experimentally such formulae [5] to determine the volumetric concentrations of hydrogen and carbon dioxide in the dry gas: (1) [H 2 ] 66.666 0.333 [CO] 1.333 [CH 4 ] , (2) [CO 2 ] 33.333 0.666 [CO] 0.333 [CH 4 ] , [CO] and [CH4] are volumetric concentrations of CO and СH4 in dry synthesis gas,%. The equations (1) and (2) show that synthesis gas with volumetric concentrations of hydrogen to 66.7% and carbon dioxide to 33.3% (C+2H2O=CO2+2H2) can be obtained due to char steam gasification. 4
Reaction rate of fixed carbon was determined by the formula, kg / s: 12.011 V sg ([CO] [CO2 ] [CH 4 ]) , 2240 V sg is volume flow rate of dry synthesis gas, Nm3/s. Rc
(3)
The degree of carbon conversion was determined by the formula, %: n
R ci τ i Xc
i 1
100 , Cc M 100 R ci is the reaction rate of carbon during the time τ i , kg/s; M is weight of char loaded in gasifier, kg; C c is mass content of carbon in the char, %.
(4)
The degree of carbon conversion was determined by the results of technical analysis of initial char and cokeash residue after gasification, weight %: d Achr Achd
100 , (5) 100 - A dch d Achd and Achr are the ash contents in dry weight of the initial char and dry coke-ash residue after gasification, respectively, weight %. Xc
It was deduced gross formula [5] for char with steam interaction in gasifier taking excess steam during the process of reaction into account:
C n
[CO ] 2 [CO 2 ] [CO 2 ] [CH 4 ] [CO] H 2O CO CO 2 CH 4 sc sc sc sc
(6) [CO] 2 [CO 2 ] 2 [CH 4 ] [CO] 2 [CO 2 ] H 2 (n 1) H 2O sc sc n is the ratio of steam and char consumptions (n=1 for case of steam absence in obtained gas); sc = [CO] + [CO2] + [СН4], vol. %. Volumetric concentration of components in wet synthesis gas was determined by the equations, vol. %:
[CO] w
[CO] 100 , s c m wc
[H 2 ] 100 , s c m wc [CO 2 ] [CO 2 ] w 100 , s c m wc [CH 4 ] [CH 4 ] w 100 , s c m wc [H 2 ] w
[CO] 2 [CO 2 ] [H 2 O] w n - 1 100 , s m c wc [CO ] [CO 2 ] [CH 4 ] n [CO ] 2 [CO 2 ] m wc . sc
(7) (8) (9) (10) (11)
5
The degree of the use of steam was determined by the formula, %: [CO ] 2 [CO 2 ] XH O 100 . 2 n sc
(12)
Lime bed was used over a char bed for CO2 absorption in the char steam gasification reactor. Reactions of lime, carbon, and steam in the gasifier were determined by such formulae: CaO + H2O → Ca(OH)2;
H298o = - 109 kJ/mole,
(13)
С + H2O → СО + Н2;
H298o = 132 kJ/mole,
(14)
СО + Н2О → СО2 + Н2;
H298o = - 41,5 kJ/mole,
(15)
Ca(OH)2 + СО2 → СаСО3 + Н2О;
H298o = - 69 kJ/mole,
(16)
С+2H2O + CaO → СаСО3 + 2Н2; H298o is thermal effect of reactions.
H298o = - 88 kJ/mole
(17)
For estimation of hydrogen production processes, there were developed methods to determine: the degree of carbon dioxide absorption by lime in the char steam gasification reactor, the degree of CO conversion in the water-shift reactor, and the degree of carbon dioxide absorption in the low-temperature reactor [6]. Concentrations of obtained components before and after lime bed were determined at the initial stage of experimental research. The analysis of experimental data showed that, knowing the concentration of the obtained components in the dry synthesis gas after lime bed, the degree of carbon dioxide absorption by lime can be determined by calculation [6]. Concentrations of hydrogen and carbon dioxide in the obtained dry gas after char bed depend on concentrations of CO and СH4 by formulae (1) and (2). Concentrations of components in the synthesis gas before lime bed (on dry basis and without CO conversion in the lime bed) were calculated on the base of dependencies obtained by formulae (1) and (2) and synthesis gas composition after lime bed, % :
H 2 c
200
CO 3 H 2 lm
lm
lm CH 4 4 H 2 lm
200 3 H 2
,
(18)
c
COc
lm CH 4 1 4 CO lm
,
(19)
CH 4 c 200 3lm H 2 , CO 4 CH 4 lm c c 100 2 CO CH 4 c CO2 c
3
(20)
,
(21),
H 2 lm , CO lm , CH 4 lm are concentrations of hydrogen, CO, and methane in the dry synthesis gas after the lime bed, vol. %. The degree of carbon dioxide absorption by lime, %:
6
CO2 lm c CO2 CO2 lm 1 1
L X CO 2
(22),
100 ,
100
CO2 lm is concentration of carbon dioxide in the dry synthesis gas after lime bed, vol. %. Synthesis gas obtained after gasifier was fed into the water-shift reactor for CO conversion. CO conversion occurs in the presence of steam according to (15) on middle-temperature and low-temperature catalysts. The obtained mixture of gases consists mainly of hydrogen and carbon dioxide. The mixture was fed into the reactor for CO2 low-temperature absorption to generate hydrogen. As a result of a series of experimental studies, it was found that, knowing the concentration of gas components (on dry basis) before water-shift reactor and the presence of CO and CO2 impurities in the obtained hydrogen (after the reactor for CO2 low-temperature absorption), it can be determined the degree of conversion of CO in water-shift reactor and the degree of CO2 absorption in low-temperature reactor [6]. The degree of CO conversion in the water-shift reactor is determined by the formula, %:
X CO
СО2 in СО2 out 100 , 1 (23), out 100 СО 2 are concentrations of CO and CO2 in the dry synthesis gas before the water-shift
CO out 1 in СО
СО in , СО2 in
reactor, vol. %; COout , СО2 out are concentrations of CO and CO2 in the dry synthesis gas after the reactor for CO2 absorption, vol. %. The degree of CO2 absorption in the low-temperature reactor, %:
CO2 in CO2 out CO2 out 1 X CO2
Х СО
in СО
100
100
СО2 in Х СО СО
in
100 ,
(24)
100
The use of formulae (18) - (24) significantly reduced the time of analysis of processes of hydrogen generation from renewable and fossil solid fuels at Fuel Cell Test Installation. 5. Experimental results of coal and wood waste conversion 5.1. Coal conversion In the studies of volatile conversion, the fuel cell test installation consist of tank with distilled water, feeder of distilled water, electric steam generator, reactor for steam conversion of volatile, cooler of synthesis gas, low-temperature reactor for CO2 absorption with 25 % water solution of monoethanolamine, burner, and measuring system for temperatures and gas concentrations. In the mode of synthesis-gas and hydrogen generation, the tank with distilled water, distilled water feeder, electric steam generator, steam gasifier, synthesis gas cooler, water-shift reactor, reactor for CO2 absorption, burner, and measuring system for temperatures and gas concentrations were used in the Fuel Cell Test Installation.
7
5.1.1. Steam conversion of Ukrainian bituminous coal (Lviv-Volyn coal basin) with hydrogen generation Steam conversion of volatile The reactor was loaded with 496 g of coal. The obtained products of bituminous coal volatile conversion were directed from the steam conversion reactor through the cooler into the reactor for CO2 absorption. Fig. 5.1 shows the results of bituminous coal volatile steam conversion and CO2 absorption: ratio of components during the reaction (a), temperature in the bed of fuel (b), compositions of dry gases after the steam conversion reactor and absorption reactor (c), degree of CO2 absorption (d). An increase in volumetric concentration of hydrogen in the dry gas up to 81.3% was reached due to the use of the reactor for CO2 absorption. TEMPERATURE IN FUEL BED, OC
RATIO OF THE REACTED COMPONENTS STEAM/VOLATILE MATTER, KG/KG
700
1,2 1 0,8 0,6 0,4 0,2 0
650 600 550 500 450 0
10
20
30
40
50
60
0
20
40
60
Time, min
Time, min
а)
b)
CONCENTRATIONS OF COMPONENTS IN DRY GAS AFTER STEAM CONVERSION AND CO2 ABSORPTION, VOL. %
CO2 ABSORPTION DEGREE, % H2
90
CO
80
100
CO2
90
CH4
80
N2
70
30
H2
60
20
СО
70 60 50 40
10
СО2
0 0
10
20
30
40
50
60
Time, min
СН4
50 0
20
40
60
Time, min
N2
c) d) Fig. 5.1. The results of bituminous coal volatile steam conversion and CO2 absorption Steam gasification of char using lime The steam gasifier was loaded with char and lime (for capturing CO2 in gasification zone directly). 221 g of lime was loaded at the top of the reactor. The weight ratio of carbon in the coke to CaO before the experiment was 4.72 g/g. Water feed rate in steam generator was 4.2 g/min. Fig. 5.2 shows the results of the steam char gasification: the degree of carbon conversion, gas feed rate before and after the bed of lime/limestone, volumetric concentration of gaseous components before and after the bed of lime/ limestone, temperature in the bed of lime/limestone. Dry synthesis gas (H2 = 72.4-75.2, CO = 11.5-17.3, CO2 = 6.4-15.1, CH4 = 0.97 -1.00 vol.%) that is comparable by composition to synthesis gas after steam reforming of natural gas was obtained due to the use of lime in the steam gasifier of char. At the final stage of the experiments, the synthesis gas was directed into the water-shift reactor. 171.8 g of ceramics (for generating steam from feed water and heating the steam-gas mixture), 401.5 g of ironchromium catalyst (HTC-1), and 459.7 g of copper catalyst (LTC-4) were in the water-shift reactor. Fig. 5.3 shows the main operating parameters of the water-shift reactor: consumption of water in the reactor (a), the temperature in the middle of the water-shift reactor (b), the degree of CO conversion in the watershift reactor (c), the composition of synthesis gas after the water-shift reactor (d). Synthesis gas was directed from the water-shift reactor through the cooler to the reactor for low-temperature CO2 absorption. 110 ml of 8
25 % monoethanolamine solution was in the reactor for low-temperature CO2 absorption. The degree of CO2 absorption in the low-temperature reactor was 95.6-97.8 %. CARBON CONVERSION DEGREE, %
GAS FEED RATE, G/MIN
100 80 60 40 20 0
1,5
Before the bed of Ca compounds
1
After the bed of Ca compounds
0,5 0
50
100
0
150
0
Time, min
50
100
150
Time, min
а)
b)
DRY SYNTHESIS GAS COMPOSITION AFTER CHAR BED, VOL. %
DRY SYNTHESIS GAS COMPOSITION AFTER THE BED OF CA COMPOUNDS, VOL. %
60 H2
40
80 60 40 20 0
CO
20
CO2
0
CH4 0
50 100 Time, min
150
H2 CO CO2 CH4 0
50
100
150
Time, min
с)
d) O
CHAR BED TEMPERATURE, C 915 910 905 900 895 890
100
DEGREES OF CO CONVERSION AND CO2 ABSORPTION/DESOPTION, %
50
XCO XCO2
0 0 0
50
100
50
100
150
-50
150
Time, min
-100
Time, min.
e)
f) Fig. 5.2. The results of steam char gasification
WATER CONSUMPTION IN SHIFT REACTOR, KG/HOUR 600 500 400 300 200 100 0 100
TEMPERATURE IN MIDDLE OF о SHIFT REACTOR, С
600 500 400 300 200 100 0
110
120 130 Time, min
140
100
150
110
120 130 Time, min
а)
150
b) DRY SYNTHESIS GAS COMPOSITION AFTER SHIFT REACTOR, VOL. %
CO CONVERSION DEGREE IN SHIFT REACTOR, % 100 80 60 40 20 0
94 92 90 88 86 100
140
110
120
130
Time, min
c)
140
150
100
H2 CO CO2 CH4 110
120
130
140
150
Time, min
d)
Fig. 5.3. The main operating parameters of the water-shift reactor 9
Fig. 5.4 shows the summarize results of char conversion into hydrogen at the last stage of the research. HYDROGEN CONCENTRATION, VOL. %
100
80
after CO2 absorption reactor after shift reactor
70
after gasifier
90
60 100 110 120 130 140 Time, min
150
CARBON REMOVAL DEGREE, % 100 80 60 40 20 0
Total Reactor of CO2 absorption Gasifier
100
110
120
130
140
150
Time, min
Fig. 5.4 The summarize results of char conversion into hydrogen At the last stage of studies, hydrogen of 97.02 % purity was obtained with the use of steam gasifier, watershift reactor, and reactor for low-temperature CO2 absorption. 92.6-94.1% of carbon was removed in the system of hydrogen generation: 36.8-48.5% – in the gasifier, 44.1-56.3% - in the reactor for low-temperature CO2 absorption. 5.1.2. Steam conversion of Ukrainian brown coal (Korostyshiv deposit of Zhytomyr region) Steam conversion of volatile 50 g of brown coal was loaded in the reactor for steam conversion of volatile to produce char. Figure 5.5 a shows the steam consumption for the steam conversion of brown coal volatile in the process of the experiments. Figure 5.5 b shows the temperature in the coal bed and in the middle of space over the bed. Figure 5.5 c shows the composition of the obtained dry gas. During the volatile steam conversion, the hydrogen content in the obtained gas changed from 11.5 to 52.7 vol. %, methane content - from 47.9 to 19.7 vol. %, CO - from 7.4 to 15.2 vol. %, CO2 - from 2.4 to 12.4 vol. %.
a)
b)
c)
Fig. 5.5. Steam conversion of volatile The result of proximate analysis of the obtained char is: Volatile - 6.5 wt. %, Ash - 45.38 wt. %, Moisture 1.67 wt. %. Steam gasification of char using lime and catalyst of CO conversion 46 g of lime, 50 g of iron-chromium catalyst CTK-1, and 46 g of lime were loaded in consecutive order into the gasifier on the bed of the obtained char. Fig. 5.6 a shows the steam consumption in char gasification process, 5.6 b shows the temperature in the bed of char and temperature of lime/limestone on the char bed (b)
10
a)
b)
Fig. 5.6. Steam consumption in char gasification process (a) and temperature in the char bed and temperature of lime on the char bed (b) Fig. 5.7 shows the consumption of dry synthesis gas and its composition at the gasifier outlet (a, b), the reaction rate of carbon (c) and degree of carbon removal in the gasifier (d). Fig. 5.7 (b) shows that technical hydrogen (gas containing over 80 vol. % of hydrogen) was obtained at the installation during the first 22 min. of experiment, after which the hydrogen concentration decreased to 67.4 vol. %. During the first 20 minutes of the experiment, the CO concentration in the obtained gas was 1.663.25 vol. %, after which it increased to 5.69 vol. %.
a)
b)
c)
d)
Fig. 5.7. The results of steam conversion of char using lime and CO catalyst in the gasifier: consumption of dry synthesis gas and its composition (a, b), the reaction rate of carbon (c) and degree of carbon removal in the gasifier (d) During 34 min. of experiment, the absorbent removed 59.27% of carbon, which has reacted with steam 5.1.3. Steam conversion of Ukrainian brown coal (Alexandria deposit of Кirovograd region) with hydrogen generation Steam conversion of volatile 50 g of brown coal was loaded in the reactor for steam conversion of volatile to produce char. Fig. 5.8 a shows the steam consumption for the steam conversion of brown coal volatile. Fig. 5.8 b shows the temperature in the coal bed and in the middle of space over the bed. Fig. 5.8 c shows the composition of the obtained dry gas.
11
During the volatile steam conversion, the hydrogen content in the obtained gas changed from 6.8 to 66.6 vol. %, CO2 - from 14.5 to 55.1 vol. %., methane content - from 3.1 to 16.6 vol. %, CO - from 2.7 to 13.8 vol. %
a)
b)
c)
Fig. 5.8. Steam conversion of volatile The result of proximate analysis of the obtained char is: Volatile – 13.84 wt. %, Ash – 14.13 wt. %, Moisture – 9.1 wt. %. Steam gasification of char using lime and hydrogen generation 124.6 g of CO2 absorbent (lime) were loaded into the gasifier on the bed of the obtained char. Fig. 5.9 a shows the steam consumption in char gasification process, 5.9 b shows the temperature in the bed of char and absorbent.
a)
b)
Fig. 5.9. Steam consumption in char gasification process (a) and temperature in the bed of char and absorbent (b) Fig. 5.10 shows the consumption of dry synthesis gas and its composition at the gasifier outlet (a, b), the reaction rate of carbon (c) and degree of CO2 absorption in the gasifier (d). Fig. 5.10 (b) shows that hydrogen concentration in the dry gas was 82.4-84.1 vol. % during the first 15 min. of experiment, after which it decreased to 66.1 vol. %. The concentration of CO changed in a range from 7.1 to 3.7 vol. %, CO2 - from 7.7 to 29.7 vol. %, CH4 - from 1.9 to 0.5 vol. %. Fig. 5.10 (d) shows that degree of carbon dioxide removal in the gasifier was 74.5-80.8 % during the first 15 min. of gasification. Then, the degree of CO2 absorption decreased to 6.3%. During 33 min. of experiment, the absorbent removed 51.8% of carbon, which has reacted with steam.
12
a)
b)
c)
d)
Fig. 5.10. The results of steam conversion of char using lime in the gasifier: consumption of dry synthesis gas and its composition (a, b), the reaction rate of carbon (c) and degree of CO2 absorption (d) Dry synthesis gas obtained after gasifier was directed through the shift reactor to the reactor for lowtemperature absorption of CO2. The loading of the shift reactor was the same as in the case of conversion of bituminous coal char into hydrogen. Fig. 5.11 shows the degree of CO conversion in the water-shift reactor (a); volumetric concentration of hydrogen at the gasifier outlet, water-shift reactor inlet, and outlet of the reactor for CO2 absorption (b); hydrogen consumption at the outlet of the reactor for low-temperature absorption of CO2 (c); and the degree of carbon removal in the gasifier, reactor for low-temperature absorption of CO2, and the total degree of carbon removal in the installation (d).
a)
b)
c)
d)
Fig. 5.11. The results of char conversion into hydrogen: the degree of CO conversion in the water-shift reactor and the degree of CO2 removal in the reactor for low-temperature absorption of CO2 (a); volumetric 13
concentration of hydrogen in dry gas at the gasifier outlet, water-shift reactor inlet, and outlet of the reactor for CO2 absorption (b); hydrogen consumption at the outlet of the reactor for low-temperature absorption of CO2 (c); and the degree of carbon removal in the gasifier, reactor for low-temperature absorption of CO2, and the total degree of carbon removal in the installation (d) The degree of carbon removal was: in the gasifier - 5.6 to 67.1%; in the reactor for low-temperature absorption of CO2 - 23.7 to 80.2%. The degree of carbon removal in the installation was 85.8 to 91.7%. As a result, there was obtained hydrogen purity of 94.7-96.2 %. 89.7% of carbon, which reacted with steam in the gasifier, was removed during all time of experimental researches in the gasifier and reactor for low-temperature absorption of CO2. 5.2. Wood waste conversion Drying of initial wood and its impregnations with water The freshcut sticks of birch that is grown in Kiev on the bank of the river Dnepr were studied. It was shown that average density of birch with bark was 816.38 kg/m3. The moisture content of birch was 39.9-41.1 %. Firstly, all samples have been dried during one hour. Before conversion, the samples with and without bark have been extra dried at 105 оС during 24 hours. The dependence of density of birch with and without bark on moisture content is shown in Fig.5.12 a. The density of oven-dry birch depends on relative contents of bark was 518-568 kg/m3. The density of oven-dry birch without bark was 579.7 kg/m3. Fig. 5.12 b shows dependence of sample weight on residence time in water.
a) b) Fig. 5.12. Dependencies of birch density on moisture content (samples № 1-4 – with bark, sample № 5 – without bark) (a) and changing of wood weight at residence time in water (b) Division and crash of dried wood After drying, bark was separated on internal trunks. Trunks were cut up on plates of 10-15 mm in thickness. These plates were crushed to particles of maximum size to 17 mm. Percentage of bark, sawdusts, and crushed wood was 10.2/28.6/61.2 kg/kg/kg. Impregnation of crushed wood with water solutions [7] Three samples of the birch wood have been impregnated with 500 ml of 0.5% CaCl2 (converting to CaCl2), H3PO4 (converting to P2O5), FeCl3 (converting to FeCl3) solutions during 24 hours, respectively. During impregnation, samples were fully covered with solutions. After impregnation, solutions were decanted and analyzed. Samples of wood have been dried on air during 24 hours. Next, samples have been dried in drying box at 110 0С for 24 hours. Amount of CaCl2 in solutions that remained after impregnation was analyzed by gravimetrically. The solution was evaporated for water removing. Sediment has been heated at 700 0С for two hours. Obtained sediment was cooled and weighed. Amount of CaCl2 that covered wood was determined by weight difference. Amount of H3PO4 and FeCl3 was analyzed by titrimetricaly. FeCl3 solution was titrated by 0.1 N EDTA solution in the presence of 25% sulfosalicylic acid solution. H3PO4 was titrated by 0.12 N alkali solution in the presence of phenolphthalein. Titration was passed to formation of dibasic 14
salt. Amount of phosphoric acid was determined from amount of acid that can be passed from wood to impregnating solution. The wood was placed to water for 24 hours and then solutions were titrated by 0.12 N alkali solution in presence phenolphthalein. It was found that 5.9·10-5 mol/g P2O5, 1.3·10-5 mol/g CaCl2, 4.63·10-5 mol/g FeCl3 were precipitated on wood samples, respectively. Apparent density of dried and crushed wood samples without impregnation was 0.283 kg/m3. Apparent density of samples impregnated with Н3PO4, CaCl2, and FeCl3 solutions were 0.259, 0.251, 0.253 kg/m3, respectively. Carbonization of the crashed birch Table 5.1 shows results of technical analysis of initial wood samples. Table 5.1 - Results of technical analysis of initial wood samples Moisture Volatile matter Ash content # Sample wt. % 1 Initial 9.64 74.52 0.33 2 Impregnated with Н3PO4 1.24 81.88 0.37 3 Impregnated with CaCl2 1.50 82.66 1.00 4 Impregnated with FeCl3 0.98 83.18 1.11
Fixed carbon 15.50 16.52 14.83 14.73
The samples were placed into special metallic cylindrical container (module of carbonization) that has four sections for each sample. After filling, each section was closed by filter. Moreover, the cylindrical container was also closed by cover. Next, the container was moved to the muffle furnace. The thermocouple for temperature measurement was put in the middle of muffle furnace over container of carbonization. The design of carbonization container was developed on the basis of experimental carbonization of old furniture woods. Fig. 5.13 shows the change of muffle furnace temperature during carbonization of samples.
Fig. 5.13. Change of muffle furnace temperature during carbonization of samples The processes of carbonization are investigated in two modes: with and without the use of wood volatile calorific value. The heating rate was regulated by changing power of electrical heaters. When the volatile burning operating mode is used, the electrical power supply of the muffle furnace (We=889 Wt) was supported to spontaneous ignition of volatile matters at the temperature of 402оС. Next, the power supply of the muffle furnace was turned off. When the operating mode without volatile burning is used, the electrical power supply of the muffle furnace (We=889 Wt) was supported to the temperature of 325 оС. Electric power of the muffle furnace was regulated in the range of 110 to 330 W during 30 minutes with the purpose of maintenance of temperature in the muffle furnace at the level of 318-333 оС. Next, the power supply of muffle furnace was turned off. The container with samples of wood coal was cooled in the muffle furnace and pulled out. Samples were taken out from container for weighing and analysis. 15
Table 5.2 shows the results of wood conversion to the wood coal. Table 5.3 b shows results of proximate analysis of the obtained wood coal. Analysis of the results in tables 5.2 and 5.3 shows: when mode with volatile burning is used, the wood coal yield is increased by 9.9, 1.94, and 2.31 % on the dry basis due to the preliminary impregnation of birch samples with H3PO4, CaCl2, and FeCl3 solutions, respectively. Fixed carbon content in the samples of wood coal was 56.50 – 60.31 %. Mode without volatile burning gives an increase in wood coal yield by 7.03 and 2.38 % on the dry basis due to the preliminary impregnation of birch samples with H3PO4 and FeCl3 solutions, respectively. Fixed carbon content in the samples of wood coal was 46.7 – 58.3 %. The formulae of the obtained wood coals (table 5.4) were determined after the ultimate analysis. Table 5.2 - The results of wood conversion to wood coal with and without volatile burning
Sample #
Wood weight before conversion wet dry g
1 2 3 4
24.71 24.69 24.92 24.71
22.62 24.38 24.55 24.47
1 2 3 4
25.1 25.03 25.04 25.062
22.97 24.72 24.67 24.82
Wood coal weight after conversion dry g Mode with volatile burning 3.31 3.01 6.58 5.66 4.14 3.74 4.14 3.82 Mode without volatile burning 10.03 9.72 12.61 12.20 10.32 10.04 11.42 11.09
Wood coal yield on dry basis
wet
% 13.30 23.20 15.24 15.61 42.32 49.35 40.69 44.70
Table 5.3 - The results of proximate analysis of the obtained wood coal samples after conversion using modes with and without volatile burning Sample #
Moisture
1 2 3 4
9.09 14.04 9.62 7.77
1 2 3 4
3.07 3.27 2.76 2.90
Sample # 1 2 3 4
Volatile matter
Ash content
wt. % Mode with volatile burning 29.30 1.30 25.57 1.31 29.70 3.85 31.40 4.26 Mode without volatile burning 40.88 0.97 48.37 0.84 36.05 2.86 47.88 2.52
Fixed carbon
60.31 59.09 56.83 56.57 55.08 47.52 58.33 46.70
Table 5.4 - The formulae of dry ash-free matter of obtained wood coals Mode of with volatile burning Mode without volatile burning С1H0,354O0,063 С1H0,670O0,241 С1H0,228O0,039 С1H0,619O0,211 С1H0,290O0,044 С1H0,739O0,282 С1H0,451O0,113 С1H0,658O0,234
The least contents of oxygen and hydrogen and accordingly greatest contents of carbon (Table 5.4) were achieved in the samples of wood coal obtained from wood previously impregnated with H3PO4 solution. 16
The sample of birch waste impregnated additionally with 0.5% water solution of H3PO4 was chosen for generation of hydrogen-rich gas at Fuel Cell Test Installation for the reasons: the greatest stability of particles obtained during wood coal carbonization to disintegration at high heating rate, and an increase in yield of wood coal with high content of fixed carbon. For obtaining wood coal required for experiments, all sections of carbonization container were filled by crushed birch waste that was pre-impregnated with 0.5% solution of H3PO4 and dried. Weight of crushed wood waste was 108.5 g. The composition of impregnated wood in recalculation on analytical weights was: Volatile = 70.12 wt. %, Ash = 0.36 wt. %, Moisture = 10.41 wt. %. Fig. 5.14 shows changing temperature in the furnace (a) and yield of dry wood coal during carbonization (b). Table 5.5 shows the results of the conversion of crushed birch to wood coal.
a) b) Fig. 5.14. Temperature in the furnace (a) and yield of dry wood coal (b) during carbonization Table 5.5 - Results of the conversion of crushed birch to wood coal Wood weight before conversion wet dry g 108.5
97.21
Wood coal weight after conversion wet dry g 31.76
Wood coal yield on dry basis
31.34
% 32.24
As a result of conversion, there was obtained wood coal of such composition: Volatile = 14.21 wt. %, Ash = 1.23 wt. %, Moisture = 1.33 wt. %. Apparent density of the obtained wood coal was 181.6 kg/m3. Wood coal with high content of carbon (83.23 %) and low content of ash (1.23 %) was obtained due to impregnation of wood and optimal organization of carbonization process. The degree of the conversion of volatile carbon to fixed carbon was 30.4 %. LHV of obtained wood coal was 30.73 MJ/kg. Chemical efficiency of the process of producing wood coal from crushed wood was 55.6 %. Steam gasification of wood coal using lime and catalyst of CO conversion Beds of ceramic, wood coal (29.42 g), lime (153.69 g), and catalyst of CO conversion MCC-2.4 (2 wt.% CuO, 4 wt.% CeO2, 0.1 wt.% Fe3O4), developed at L.V. Pysarzhevsky Institute of Physical Chemistry of National Academy of Sciences of Ukraine, were loaded in consecutive order into the steam gasification reactor. The catalyst was placed in a specially developed section. The catalyst weight was 21.55 g. Fig. 5.15a shows a degree of carbon conversion during steam gasification of wood coal. Fig. 5.15 b shows carbon gasification rate and temperature of wood coal bed, Fig. 5.15 c - dry gas composition, Fig. 5.15 d degree of CO2 absorption.
17
a)
b)
c) d) Fig. 5.15. Degree of carbon conversion (a), carbon gasification rate (b), temperature of wood coal bed (b), dry gas composition (c) and degree of CO2 absorption (d) during steam gasification of wood coal As a result of the use of lime and the catalyst of CO conversion, concentration of CO in the obtained dry gas was low and changed in the process of experimental research in a the range of 2.1 to 4.4 vol. %. CO2 concentration increased from 3.8 to 25.7 vol. % and СH4 concentration decreased to 0.8 vol. % with increasing the temperature in the wood coal bed from 650 oC to 823 oC. Technical hydrogen (gas containing over 80 vol. % of hydrogen) was obtained at the installation outlet within 55 minutes at the temperature in the wood coal bed 660 - 757 oC. The degree of absorption of CO2 from the gas obtained after wood coal bed decreased from 93.06 to 22.56 %. 50.52 % of carbon, which has reacted with steam in the wood coal bed, was fixed with absorbent during the experimental research. Solid organic residues after steam gasification held the shape of initial wood coal, but decreased in size. Fig. 5.16 shows the BET isotherm of adsorption/desorption of nitrogen by the particles of the obtained organic residues.
Fig. 5.16. BET isotherm of adsorption/desorption of nitrogen by the particles of obtained organic residues
18
The analysis of isotherm showed that activated charcoal was obtained as a result of steam gasification of wood coal using lime/limestone at the reaction temperature of 650-823 oC. The inner surface of the activated charcoal was 1023.5 m2/g. 6. SOFC IGCC Power Plant circuits Two circuits of new coal combined-cycle plants were developed after analysis of the obtained experimental data. 6.1 SOFC IGCC Power Plant circuit for bituminous and brown coals conversion Fig. 6.1 shows the circuit of SOFC IGCC plant, which includes fuel processor, electrochemical generator (ECG), gas turbine and steam-turbine units (GTU and STU).
Fig. 6.1. Circuit of IGCC Coal Power Plant The mixture of coal and limestone is fed in circulating fluidized bed reactor through the lockage. The reactor can work in both modes of solid fuel combustion and gasification. It consists of bottom, middle, and top parts. The bottom is a zone of solid fuel combustion/gasification with recirculation of particles. The middle part is a zone of sorption/desorption of CO2 from gas obtained in the bottom part. The top is heat-exchange zone (for steam production). At first, fuel processor operates in the mode of combustion (CO2 desorption) with entering air into the bottom of the reactor. In this mode, heat needed to steam gasification is transferred to circulating solid material, and flue gas comes in the middle part to produce CO2 absorbent. Then, steam is used instead of air at the bottom of processor; the processor begins to operate in the mode of steam gasification (CO2 sorption). Heat of circulating solid material is used for steam gasification, and synthesis gas that was obtained in the bottom is cleaned from CO2 in the middle part.
19
An important factor to ensure uninterrupted operation of IGCC power plant is continuous delivery of synthesis gas with high content of hydrogen to the ECG and combustion products to the GTU inlet. At least two reactors for solid fuel conversion are used in the circuit. Synthesis gas with a high content of hydrogen is fed to the electrochemical generator, and combustion products – to the inlet of gas turbine unit with air heater. Heat of combustion products after the GTU is used to heat the air before ECG and to produce steam for operation of steam-turbine unit and fuel processor. 6.2 SOFC IGCC circuit for conversion of wood waste Fig. 6.2 shows the circuit of IGCC Power Plant with producing electricity, wood coal, and activated charcoal from wood waste.
Fig. 6.2. Circuit of bio-fuel IGCC Power Plant The circuit of IGCC Power Plant consists of unit for wood coal production, gasifier for the partial gasification of wood coal with producing activated charcoal and synthesis gas, as well as ECG, GTU and STU for electricity generation. Zone of wood waste carbonization consists of devices for impregnating wood with water solution of H3PO4, drying and carbonization. Energy of wood volatile burning is used for drying and carbonization of wood, as well as for steam generation in boiler of the steam-turbine unit. Wood coal obtained in carbonizer, steam obtained in STU, and oxygen obtained in the installation of short-cycle sorption/desorption are used in the gasifier to produce activated charcoal and synthesis gas. Synthesis gas obtained in the gasifier is fed to CO2 sorption/desorption devices. Synthesis gas with a high content of hydrogen is fed to the ECG with solid oxide fuel cells, and combustion products enriched by CO2 to the GTU inlet. Each zone of the Combined Cycle Plant is self-sufficient. Devices for wood waste carbonization can operate independently only to produce wood coal, hot water, and steam. Activated charcoal, as well as synthesis gas, which can be used for domestic use instead of natural gas, can be produced due to addition of gasifier to the circuit. Technical hydrogen for the use in transport application can be produced from synthesis gas due to addition of system of CO2 sorption/desorption.
20
Conclusions 1. New methods for calculation of CO2 absorption degree in the steam gasifier and in the reactor of lowtemperature absorption of CO2, and CO conversion degree in the water-shift reactor are developed and tested. 2. In the time of studies of steam conversion of bituminous coal volatile in the temperature range of 498683 °С, an increase in volumetric concentration of hydrogen in the obtained dry gas up to 81.3% was reached due to the use of CO2 absorption reactor (at 97 % degree of CO2 absorption). 3. Experimental studies of steam gasification of bituminous coal char with the use of lime/limestone in the gasifier made it possible to determine the degree of CO2 absorption/desorption by the bed of lime/limestone and the degree of CO conversion in the bed of lime/limestone, taking the changing composition and flow rate of dry synthesis gas into account. 4. Dry synthesis gas (H2 = 72.4-75.2, CO = 11.5-17.3, CO2 = 6.4-15.1, CH4 = 0.97 -1.00 vol.%), that is comparable by composition to synthesis gas after steam reforming of natural gas, was obtained from char due to the use of lime. At the last stage of studies, hydrogen of 97.02 % purity was obtained with the use of steam gasifier, water-shift reactor, and low-temperature CO2 absorption reactor. When lime is used in the gasifier, 92.6-94.1% of carbon was removed in the system of hydrogen generation: 36.848.5% – in the gasifier, 44.1-56.3% - in the low-temperature CO2 absorption reactor. 5. The research of steam conversion of volatile of two brown coals has shown that process of steam volatile reforming should occur in the range of 300-600 oC for the maximum yield of highly reactive char. 6. Char of brown coal was used to obtain hydrogen by two circuits: 1) steam gasifier with middletemperature catalyst of CO conversion in the bed of lime; 2) steam gasifier (using lime), water-shift reactor (using catalysts of CO conversion) and CO2 absorption reactor (using monoethanolamine solution). The research of steam conversion of brown coal char was conducted in the temperature range of 700-760 oC at initial mass ratio of CaO/carbon in char – 4.7...7.1 kg/kg. 7. Lime and iron-chromium catalyst of CO conversion were used in the gasifier to reduce CO concentration in the obtained dry gas. Technical hydrogen was obtained at the gasifier outlet at the temperature of brown coal char conversion 697-727 oC during 22 min. Throughout the experiment, the degree of CO2 capture decreased from 82.8 to 13.9%, and the degree of carbon removal - from 76.1 to 11.7%. Totally, the absorbent removed 59.3% of carbon, which has reacted with steam. 8. Hydrogen of 94.7 - 96.2% purity was obtained from char of brown coal as a result of the use of steam gasifier using lime, water-shift reactor with iron-chromium and copper catalysts of CO conversion and low-temperature CO2 absorption reactor with monoethanolamine solution. 9. In the case of the use of brown coal char at the temperature of steam gasification less than temperature of steam gasification of bituminous coal char by 230 oC, maximum volumetric concentration of hydrogen in the obtained dry gas after the gasifier was increased up to 84.3 vol. % with increasing carbon removal in the gasifier. 10. The dependence of density of initial birch with/without bark on moisture content is determined. 11. Installation for the research of simultaneous carbonization of four wood samples is created and tested in the operation. The carbonization processes of crushed pure wood and crushed wood after impregnation with 0.5% water solutions of H3PO4, CaCl2, and FeCl3 and drying with 5.9 ·10-5, 1.3·10-5, and 4.63·10-5 mol/g contents of P2O5, CaCl2, and FeCl3, respectively, were investigated in both modes with and without the use of volatile calorific value for carbonization. 12. Pre-impregnation and drying increased the share of fixed carbon in the waste of birch by 5-13 % that together with optimization of carbonization increased the efficiency in 1.3-1.8 times. The impregnation of the sample with water solution of H3PO4 gave the maximum yield of wood coal. 13. Wood coal obtained from the wood impregnated with 0.5% water solution of H3PO4 was gasified at Fuel Cell Test Installation. CO concentration in the obtained dry gas was low (2.1 - 4.4 vol.%) due to the use 21
of catalyst of CO conversion and lime. CO2 concentration increased from 3.8 to 25.7 vol. % with increasing the temperature in the wood coal bed from 650 oC to 823 oC, and СH4 concentration decreased to 0.8 vol. %. Technical hydrogen was obtained at the installation outlet within 55 minutes at the temperature in the wood coal bed of 660-757 oC. 14. The degree of absorption of CO2 from the gas obtained after wood coal bed was decreased from 93.06 to 22.56%. During the experimental studies, the absorbent fixed 50.52% of carbon, which has reacted with steam in the wood coal bed. 15. The results of organic residue analysis showed that activated charcoal with inner surface of 1023.5 m2/g was obtained as a result of steam gasification of wood coal using lime at the temperature of 650-823 oC. 16. After analysis of the obtained data, there was developed two new circuits of SOFC IGCC Power Plants using bituminous and brown coals to generate electricity and wood waste to produce wood coal, activated charcoal, and electricity. Acknowledgments Part of the presented work was carried out within the framework of projects # 24-10 and # 25-10 of the Complex Program of Scientific Researches of the National Academy of Sciences of Ukraine "Fundamental Problems of Hydrogen Energy". The authors are grateful to the National Academy of Sciences of Ukraine for support of these projects. The authors express their gratitude to Prof., Dr. Peter Stryzhak, Dr. Andriy Trypolsky, and Dr. Yevhen Kalishyn (L.V. Pysarzhevsky Institute of Physical Chemistry of NASU, Ukraine, Kyiv) for the impregnation of the waste wood samples, creation of the catalyst for CO conversion and analysis of obtained activated charcoal. References [1] O. Dudnik, I. Sokolovska, “Results of Organic Fuel Conversion at Fuel Cell Test Installation”, Fuel Cell Technologies: State and Perspectives, NATO Sci. Ser., II, Math., Phys., And Chem., Netherlands: Springer, 2005, vol. 202, pp. 163-174 (2005). [2] A.N. Dudnik, “Fuel Cell Test Installation”, Proc. of Partnership for Prosperity and Security Conference, United States Industrial Coalition, Philadelphia, PA, 17 p. (2003). [3] O. Dudnyk, “Hydrogen generation from Ukrainian coals”, Proc. of Energy Efficiency Conference, Oil and Gas Institute, Krakow, Poland, 162, pp. 63-68 (2009). [4] A. Dudnik, Y. Korchevoy, A. Maystrenko, S. Onischenko, “PCFB-FC Project”, Proc. of POWERGEN'98 International Conference, PennWell Conference & Exhibitions, Orlando, FL, Materials on CD (1998). [5] O.M. Dudnyk, I.S. Sokolovska, S.V. Zabolotny, “Steam gasification of PET char in fluidized bed”, Renewable Energy, 2008, 15(2), pp. 9-14 (2008). [6] A.N. Dudnik, “Techniques for determination of СО2 absorption and СО conversion degrees for estimation of processes of hydrogen generation from chars of renewable and fossil solid fuel”, Renewable Energy, 2009, 16(1), pp. 9-10 (2009). [7] P. Stryzhak, O. Dudnyk, A. Trypolsky, Y. Kalishyn, “Research of preparation and carbonization processes of wastes of fresh cut off birch”, Proc. of Energy Efficiency Conference, Oil and Gas Institute, Krakow, Poland, 162, pp. 203-208 (2009).
22
Manuscript Not AVAILABLE
COALBED GAS POTENTIAL IN THE MIOCENE SOMA BASIN (WESTERN TURKEY) Sedat İnan1, M.Namık Yalçın2, Ender Okandan3, Yuda Yürüm4, Ruhi Saatçılar1,5, Murat Yılmaz6, Ali Rıza Toygar6, Ayhan Kösebalaban7, Selahaddin Anaç7, Hakkı Duran7, Mehmet Onbaşı7, Mücella Ersoy7, Mehmet Atasayar7, İsmail Ergüder7, Aynur Dikbaş1, Fırat Duygun1, Semih Ergintav1, Mustafa Baysal3, Kübra Tırpan2, Korhan Yaşar2, 1) TÜBİTAK Marmara Reasearch Center, Earth and Marine Sciences Institute, Gebze-Kocaeli, Turkey 2) İstanbul University, Department of Geologoical Engineering, Avcılar, İstanbul-Turkey 3) Middle East Technical University, Department of Petroleum Engineering, Ankara-Turkey 4) Sabancı University, Faculty of Natural Sciences, Orhanlı, İstanbul- Turkey 5) Sakarya University, Department of Geophysical Engineering, Sakarya,Turkey, 6), Turkish Petroleum Corporation, Ankara-Turkey 7) Turkish Coal Enterprises, AnkaraTurkey and Soma,Manisa-Turkey
The Neogene Basins of Turkey contain as much as 9 billion tons of lignite-rank coal (Şengüler, 2001; Tuncalı et al., 2002). The Miocene Soma Basin, a rift basin trending NE-SW (approximately 20 kilometers by 5 kilometers) in the Aegean Extensional Province (EAP) of Western Turkey, is estimated to contain at the least one billion tons of lignite and about half of this reserve is present at depths greater than 600 m (Turkish Coal Enterprises, 2006). Miocene marl/limestone units and Pliocene clastics and volcanic tuffs overlie the Miocene coals of the Soma basin. There are several coal seams in the basin but the most economical and thus target seam is known as KM2 with an average thickness of about 20 meter across the basin. In the Soma Basin, Turkish Coal Enterprises (TKİ) has mined this KM2 seam by open cut coal mining and underground mining for several decades in the Northern and Central part of the basin, respectively. Recently, coal exploration activities have been extended to the Southern part of the basin by means of exploratory drillings. Recently, coal exploration activities have been extended to the Southern part of the basin by means of exploratory drillings. In this context, two boreholes encountering a coal seam (KM2) up to 20 m thick were evaluated. The KM2 coal seam was encountered between 900 and 940 m depth in two boreholes drilled approximately 1 km apart. Wellhead gas content was measured on coal cores following the USBM method (Diamond and Levine, 1981). Additionally, coal was placed in hermetically sealed canister and desorbed gas was analyzed in laboratory for chemical composition (by FTIR gas analyzer) and
13
C isotope (by GC-IRMS). Coal characterization was completed by means of Rock Eval (RE)
Pyrolysis, Proximate and Ultimate analyses, as well as microscopic analyses for typing of macerals and vitrinite (huminte) reflectance measurements. The wellhead gas content measurements (six core measurements from two boreholes) indicate that as much as 4 m3 gas / ton coal is present in the coal recovered from 900 to 940 m below the
1
surface. The rank of coal based on vitrinite (huminte) reflectance measurements is lignite to subbituminous (0.40 to 0.45 % Ro); supported by RE Tmax values of 420 0C. TOC content of the coal samples vary between 53 to 73 %. The composition of the gas is dominantly methane (more than 99.4 %) and the
13
C/12C isotope ratio of methane is 61 to 65 per mil. Considering the chemical
composition of the gas and the del 13C isotope of the methane, the source of the coal gas is biogenic probably generated by bacteria. The maceral analyses show that coal samples on average contain more than 60 % huminite(vitrinite). Adsorption on the internal coal surface is considered as the primary mechanism of gas storage in coals and the surface area, which controls the gas adsorption capacity, is in general a function of the micro-pore volume (Levy et al., 1997; Crosdale et al., 1998) which is known to be abundant in vitrinite maceral group. The vitrinite/huminite maseral content has positive correlation with gas sorption capacity (Levy et al., 1997); meaning that at a given pressure, the higher percentage of micro-pore dominated huminite/vitrinite the more gas adsorption capacity. In this context, the Miocene Soma lignites have good micro-pore properties in respect to gas adsorption. Preliminary evaluation, based on limited analyses and results summarized above, on gas potential of the Miocene Soma Basin is encouraging, yet further investigations in the Soma Basin are underway by coal desorption testing of cores received from ongoing coal exploration boreholes.
REFERENCES Crosdale, P. J., Beamish, B. B., and Valix, M. 1998. Coalbed methane sorption related to coal composition. Int. J. of Coal Geology 35:147–158. Diamond, W. P., and J. R. Levine, 1981, Direct method for determination of the gas content of coal: procedures and results: United States Bureau of Mines (USBM), RI 8515, 36 p. Levy, J., Day, H., and Killingley, J. S. 1997. Methane capacities of Bowen basin coals related to coal properties. Fuel 76(9):813–819. Şengüler, İ., 2001, Lignite and thermal power plants for sustainable development in Turkey, World Energy Council 18th Congress, Buenos Aires, October 2001. Tuncalı et al., 2002, Chemical and Technological Properties of Turkish Tertiary Coals. General Directorate of Mineral and Research Exploration of Turkey (MTA) Special Publication. ISBN: 6595-47-7, 400 p.
2
Manuscript Not AVAILABLE
UNDERGROUND COAL DETERMINATION BY INTEGRATED (REFLECTION & WVSP) SEISMIC IN THE MIOCENE SOMA BASIN (WESTERN TURKEY) Ruhi Saatçılar1,2, Murat Yılmaz3, Ali Rıza Toygar3, Sedat İnan1, Ayhan Kösebalaban4, Selahaddin Anaç4, Hakkı Duran4, Mehmet Onbaşı4, Mücella Ersoy4, Mehmet Atasayar4, İsmail Ergüder4, Fırat Duygun1, Semih Ergintav1, Aynur Dikbaş1, M.Namık Yalçın5, Ender Okandan6, Yuda Yürüm7, Emin Demirbağ8,1 1) TÜBİTAK Marmara Reasearch Center, Earth and Marine Sciences Institute, Gebze-Kocaeli, Turkey 2) Sakarya University, Department of Geophysical Engineering, Sakarya,Turkey, 3) Turkish Petroleum Corporation, Ankara-Turkey, 4) Turkish Coal Enterprises, Ankara-Turkey and Soma,Manisa-Turkey, 5) İstanbul University, Department of Geologoical Engineering, Avcılar, İstanbul-Turkey, 6) Middle East Technical University, Department of Petroleum Engineering, Ankara-Turkey, 7) Sabancı University, Faculty of Natural Sciences, Orhanlı, İstanbul- Turkey , 8) İstanbul technical University, Department of Geophysical Engineering, Maslak, İstanbul-Turkey
The Neogene Basins of Turkey contain as much as 9 billion tons of lignite-rank coal (Şengüler, 2001; Tuncalı et al., 2002). The Miocene Soma Basin, a rift basin trending NE-SW (approximately 20 kilometers by 5 kilometers) in the Aegean Extensional Province (EAP) of Western Turkey, is estimated to contain at the least one billion tons of lignite and about half of this reserve is present at depths greater than 600 m (Turkish Coal Enterprises, 2006). Miocene marl/limestone units and Pliocene clastics and volcanic tuffs overlie the Miocene coals of the Soma basin. There are several coal seams in the basin but the most economical and thus target seam is known as KM2 with an average thickness of about 20 meter across the basin. In the Soma Basin, Turkish Coal Enterprises (TKİ) has mined this KM2 seam by open cut coal mining and underground mining for several decades in the Northern and Central part of the basin, respectively. Recently, coal exploration activities have been extended to the Southern part of the basin by means of exploratory drillings. In order to aid fast and economical coal exploration activities of the Turkish Coal Enterprises (TKİ), a colloborative work has been started for development of an integrated seismic method for coal exploration. The aim is to develop a practical seismic method for TKİ to apply in its exploration activities in other Lignite coal basins of Turkey. The information gathered from tens of coal exploration boreholes already drilled in the study area will be used to calibrate the horizons to be mapped by surface seismic and WVSP methods (Tselentis and Paraskevopoulos, 2002). Surface seismic data will be collected on roughly East-West and North-South oriented lines. Furthermore, at the intersect of these lines borehole WVSP data will be collected. The seismic source for both surface and borehole WVSP seismic will be a mini-vibrator and the data will be collected contamporanously. The seismic study will be conducted between May and July 2010. Soon after, the seismic data processing work will be initiated. REFERENCES
1
Şengüler, İ., 2001, Lignite and thermal power plants for sustainable development in Turkey, World Energy Council 18th Congress, Buenos Aires, October 2001 Tselentis, G.A., and Paraskevopoulos, P., 2002, Application of a high-resolution seismic investigation in a Greek coal mine. Geophysics, 67, 50–59. Tuncalı et al., 2002, Chemical and Technological Properties of Turkish Tertiary Coals. General Directorate of Mineral and Research Exploration of Turkey (MTA) Special Publication. ISBN: 6595-47-7, 400 p.
2
SUSTAINABILITY AND ENVIRONMENT THE THERMAL TREATMENT STUDY OF PYRITE FROM SOUTH BRAZIL COAL MINING INDUSTRY
Michael Peterson*, Adilson Oliveira** Morgana Bom***, Jussara P. Pizzolo***, Gabriela B. Vieira***, Deise. Tramontin***,Keli V.S. Damin*** * Universidade do Extremo Sul Catarinense, Department of Chemical Engineering , Criciúma, Brazil - [email protected] ** Formula Chemical Company , Criciúma, Brazil - [email protected] *** Universidade do Extremo Sul Catarinense, Department of Chemical Engineering – Students.
Abstract The coal mining have been produced two effects: the region development and the environmental degradation by the pyrite (FeS2) oxidation and acid mine drainage (AMD) production. Pyrite decompositon produces gases like SO2 under oxidation conditions and S2 under inert conditions. This study is developing a system for the capture of the different species of gases for the production of products with major value. The reactions take place in a scrubber for the purification of the gases. For the SO2 scrubber reactions with a NaOH/H2O solution is used for the production of Na2SO3 that has value for chemical applications at Brazil. Other solution used was: NH4OH and H2O that produced ammonium sulphate (NH4)2(SO4) an important raw material for the fertilizer industries. The last way for the purification of the pyrite reducing it sulphur contend is to develop a system of catch of the elemental sulphur at the reactor. Characterization steps included chemical analysis with XRF spectrometer and phase analysis with XRD spectrometer. The pyrite for this work was milled under 325 mesh-tyle for the thermal treatment step. The thermal treatment took place at 750 oC for 1h and the characterization steps showed the phases: Na2SO3 ; (NH4)2(SO4). In this work are the results of XRD, particle size distribution and the pyrite heating results at different temperatures.
1. Introduction The south of Brazil has the largest coal reserves in Brazil. The states of Rio Grande do Sul and Santa Catarina have more than 91% of all Brazilian coal used for generating electricity. The city of Criciúma located in southern state of Santa Catarina is part of a region called carboniferous, due to the history of great production of coal energy for thermoelectric plants use. The production of coal increased significantly the economy of the city and region thanks to the appearance of other industrial activities such as ceramic tile industry, polymers processing, metal-mechanics and mining industries. The capital generated by mining and coal sale was used to develop those different industrial sectors. Nowadays, the city of Criciúma and region are the second largest center of tiles production and first in polymer processing, especially in the sector of packaging production. Thus one can say that coal mining in the southern state of Santa Catarina was very important to their development.
The Center for Mineral Technology (CTEM) conducted a study for the recovery of degraded areas and it indicated the characterization of coal from Santa Catarina, registering a low quality coal. For each ton of coal mined only 350 kg of coal energy is used with the production of 650 kg of coal tailings. The calorific value of coal energy varies in the range of (3000 - 4500) kcal / kg and the presence of ash is somewhere around 50%. Another major contaminant of Santa Catarina´s coal is the pyrite mineral (FeS2) responsible for the degradation of 3000 hectares (ha). It is estimated that for every ton of mined coal, 1.5 m3 of water commonly called acid is produced from the acid drainage (AD). The AD is produced by the exposure of the sulfide mineral pyrite to oxidizing conditions, such as O2 and water that transform this mineral in an aqueous liquid with high content of sulfuric acid and precipitation of a large amount of iron in the form of hydroxides. This acid effluent was responsible for the liquor/ concentration of various heavy metals, which contaminate water aquifers of the region, mainly rivers and lakes, in some cases occurring the total destruction of life. In recent years, due to a government decree that placed the carboniferous region as an area of environmental risk, the coal companies and the government carry out recovery projects for rehabilitation of degraded areas that that go by the development of new technologies for the remediation and the neutralization of acid effluents in sewage stations treatment and the specific regeneration of large areas with the restoration of the previous environment. Main chemical reactions responsible by acid drainage are described below (EVANGELOU, 1995):
FeS 2 ( s ) + 7 2 O2 ( g ) + H 2 O( l ) → Fe + + ( aq ) + 2SO4
−
( aq )
+ 2 H + ( aq )
[1]
Fe + + ( aq ) + 1 2 O2 ( g ) + 2 H + ( aq ) → Fe + + + ( aq ) + H 2 O(l ) Fe
+++
( aq )
+ 3H 2 O( l ) → Fe(OH ) 3 ( s ) + 3H
+
[2]
( aq )
4 FeS 2 ( s ) + 15O2( g ) + 14 H 2 O( l ) → 4 Fe(OH ) 3( s ) + 8SO4
[3] −−
( aq )
+ 16 H
+
( aq )
[4]
The sequence of reactions identifies a large formation of H + including the steps of oxidation of FeS2 with water and O2. The reaction (4) is the overall reaction that shows that the formation of ferric hydroxide in the form precipitation can be seen by the characteristic color of rivers in the region some years ago.
With all this background of environmental degradation, the research and development of new technologies for the transformation of pyrite into higher added value products to have a reduced environmental impact. The main point is to consider the pyrite as raw material for other industries, setting another status, no more as an environmental problem, but as a solution as well as the appearance of other industries in the region.
Everywhere, authors studied and study some possibilities of application to pyrite. JERTZ (2003) studied the rate of oxidation of pyrite in oxidizing conditions, characterizing the formation of products such as ferrous sulphate. WINTER (2004) demonstrated the characterization of inorganic compounds in coal, this study showed the ratio between the amount of present FeS2 and FeSO4. JEKIK (2006) studied the production of ferric sulphate as oxidizing agent, while PETERSON (2008) studied the formation of ferrous sulphate from pyrite under specific conditions of thermal treatment. Moreever, PETERSON and OLIVEIRA (2009) at last Pittsburgh Coal Conference evaluated the possibility of hematite utilization formed in the pyrite thermal treatment for a ceramic tile industry glaze. Now in this Project, the authors study others products from the pyrite heating and the formation of hematite in different temperatures.
2. Experimental Procedure
This work had, as a basic operating condition, the thermal treatment of a pyrite sample from southern Santa Catarina, from mine located in the municipality of Forquilhinha owned by Criciúma Ltda mining company.
The basic equipment for this experiment was a heat treatment furnace with controlled atmosphere and ceramic tube with the maximum operating temperature of 1200 °C. The heat treatments were performed at the following temperatures: 150, 300, 400, 450, 550, 650, 750, 950 and 1050 °C. The heating program was used as follows: heating rate 20 °C / min with a plateau of 4 h at each indicated maximum temperature. The gas used for heat treatments was synthetic air (21% O2 and 79% N2) at a flow rate of 60cm3/min; ensuring the required excess oxygen for total oxidation, for each temperature. The pyrite mass used in this study was around 20g for each temperature of the heating treatment; the results and discussion are about the pyrite masses used in each heat treatment.
The heat treatment was carried out with the possibility of collecting the gas resulting from the pyrite oxidation in a bubbling system with a solution of ammonium hydroxide (500 ml of water and 20 ml of ammonium hydroxide) with the main objective to precipitate the oxidized sulfur during the heat treatment under the form of ammonium sulphate. After the thermal reaction and collection, the resulting solution was dried up in a kiln at a 105 oC temperature until constant mass.
The characterizations were performed using x-ray diffraction techniques of with a Shimadzu equipment to identify the main crystalline phases formed during heat treatment and also after the ammonium sulphate solution precipitation.
The characterization of pyrite particle size used in this experiment was performed by an equipment of size particle analysis by laser diffraction - CILAS - with measuring range between 0.04 µm and 500 µm.
3. Results and Discussion 3.1 Characterization of mineralogical and particle size of the pyrite sample.
The result of x-ray diffraction of the pyrite sample used in this experiment can be seen in the x - ray graph shown below (figure 01)
600
P
P = pyrite (FeS2) R = rhomboclase (FeH(SO4)2.H2)
500
CPs
400
P P
300 P
P
P
200 P R R R R
100
R
P
P P P
0 10
20
30
40
50
60
70
80
2 theta
Figure-1: XRD of pyrite sample used in this study. P=Pyrite and R=Rhomboclase
The graph shows the main presence of the mineral pyrite (FeS2) and also a phase called rhomboclase which was possibly formed during the pyrite grinding, sense in previous studies this phase was not identified in pyrite in
natura.
Below follows the characterization of particle size by pyrite laser diffraction used in this study, as shown in Figure-2.
Figure-2: - Analysis of pyrite particle size used in this study by laser diffraction.
The result shows an average diameter of 17.25 µm and 100% of particle size inferior to 71µm of diameter. The result showed that the milling process was significant indicating particle sizes below 200 mesh-tyler, i.e., a high reactivity powder was obtained at the stage of heat treatment.
3.2. Mass Lost over Heat Treatments
The heat treatments at various temperatures show the following results as table below:
Table-1 - Results of the masses before and after pyrite thermal treatment 150°C
300°C
Sample holder
100,55g
Sample holder+Pyrite Pyrite
400°C
450°C
550°C
650°C
750°C
950°C
1050°C
100,33g
99,84g 100,15g
99,88g
100,34g
100,45g
99,36g
100,78g
121,44g
121,00g
119,27g
121,00g
120,00g
121,74g
120,57g
120,54g
121,04g
20,89g
20,67g
19,43g
20,85g
20,12g
21,40g
20,12g
21,18g
20,26g
Sample holder+Pyrite (after 120,28g heating)
119,36g
113,55g
112,62g
111,22g
111,59g
110,62g
110,93g
109,90g
Pyrite (after heating)
19,73g
19,03g
13,71g
12,47g
11,34g
11,25g
10,17g
11,57g
9,12g
Mass lost
1,16g
1,64g
5,72g
8,38g
8,78g
10,15g
9,95g
9,61g
11,14g
The results show that for temperatures up to 300 °C the oxidation reaction does not occur very intensely, from 400 °C on, we can perceive a more intense mass loss of 5.72 g indicating an oxidation a more intense reaction. Some authors like NISHIHARA AND KONDO (1958) studied the oxidation reaction and this happens according to:
2FeS2 + 11/2O2 → Fe2O3 + 4SO2 This reaction shows that the initial pyrite must lose mass in a heat treatment, which is consistent with the obtained results.
3.3. XRD Phases Identification after Heat Treatments
600 P 500 P
CPs
400
P
300
P
P
P
200 P P
100
P P
P
0 10
20
30
40
50
60
70
80
Theta 2 theta P:Pyrite
Figure-3: XRD of the annealed sample at 150°C. P=Pyrite
We can observe in Figure-3 that this XRD test showed that the pyrite phase is still present; without the occurrence of the oxidation chemical reaction with the synthetic air atmosphere. The 150oC temperature of heat treatment was not enough for the oxidation reaction to occur.
At next graph Figure-4, samples of XRD of a thermally treated sample at 300°C during the time indicated in the experimental procedure. We can check that the pyrite phase is still present, as the diffractogram shows. It is also shown that the peaks of low diffraction angle in the graph are no longer present in comparison to the 150°C thermal treatment. The clay fraction present in the pyrite sample suffered decomposition along with the possible present hydroxides. At Figure-5 x-ray diffraction of the sample annealed at 450°C can be identified the main presence of hematite phase (Fe2O3); result of complete pyrite mineral oxidation originally present in the sample before heat treatment.
600
P
500
P CPs
400
P 300
P
P
200
P P P P
100
P
P
0 10
20
30
40
50
60
Theta 2 Theta
70
80
P:Pyrite
Figure-4 - XRD of the annealed sample at 300 °C. P=Pyrite
600 H 500 H
CPs
400 300 H
200
H
H
H
HH
100
H
H
H
0 10
20
30
40
Theta 2theta
50
60
70
80
H:Hematite
Figure-5: XRD of the thermally treated sample at 450 °C. H=Hematite
At 750 oC the same pattern of oxidation may be noted, i.e. the presence of hematite in a perfect crystallization state, which is in agreement with the oxidation chemical reaction in the presence of synthetic air. Minerals of the low diffraction angle also are no longer present indicating a complete decomposition of these phases. Figure-6 shows the XRD result of the heat treatment performed at 750 °C.
800
H
700 600
H CPs
500 400 300
H
H
H
HH
H
200
H
100
H
H H
0 10
20
30
40
50
60
70
80
Theta 2theta
Figure-6 - XRD of the sample annealed at 750°C. H=Hematite
With temperature at 1050 oC the results of heat treatment was identical to the others from 450 °C on, with the crystallized hematite phase was clear, like showed at Figure-7 below:
800
H
700 600
H
CPs
500 400
H
H
300
H
200
H
HH H
100
H
H
0 10
20
30
40
Theta 2theta
50
60
70
80
H:Hematita
Figure-7: Shows the final heat treatment that occurred at 1050 °C in an identical atmosphere as the other heat treatments. H= Hematite
The results of the heat treatments indicated the preponderance of the hematite phase at elevated temperatures and in a synthetic air atmosphere. The XRD graphs shown above indicate that as the heat treatment is performed in a higher temperature only one phase appears; the hematite. This phase has the chemical formula (Fe2O3) with a reddish color. To corroborate the result of thermal treatment of pyrite coal in the State of Santa Catarina we name the authors (HU GUILIN et al, 2006) and (SK BHARGAVA et al, 2009) who show similar results about the pyrite oxidation in oxidizing conditions.
3.4. Precipitated Products
Figure-8 below shows the result of x-ray diffraction for the sample obtained in the precipitation of ammonium hydroxide solution. This solution was bubbled by the gases that exited the pyrite’s roaster furnace. According to
the chemical reaction; the gases have a significant amount of SO2 that once in contact with the formed ammonium hydroxide neutralizes forming ammonium sulphate.
S 650 14000 AS
12000
CPs
10000 8000
AS
6000
AS
4000 AS AS ASAS
2000
AS AS
0 10
20
30
40
Theta 2theta
50
60
70
80
AS: amonium sulphate
Figure-8: XRD of the precipitate collected after heat treatment. AS=ammonium sulphate
The result of Figure-8 indicates the presence of ammonium sulfate after the bubbling output of the roaster furnace gases in an ammonium hydroxide solution. The result indicates the precipitation of ammonium sulfate as a main component.
This result indicates a high possibility for heat treatment as the ammonium sulfate has an interesting commercial value since it can be used in fertilizer formulations. Anyway, the SO2 formed in the roasting reaction should be neutralized due to its high reactivity with water to the unwanted formation of acid rain.
Authors such as (JAMES LEE, and LI Rongfu, 2003) studied the structure of NH4HCO3 with the passage of carbon dioxide formed in coal combustion. The authors (HIDETSUGU NAKAMURA et al, 1994) studied the pyrite reaction with ammonium nitrate in an argon atmosphere and found the oxidation reaction of
pyrite to hematite, magnetite and (NH4)3Fe(SO4)3. The authors (FENGQI SI et al, 2009) studied the ammonium sulfate and ammonium bisulfate formation in a system for reducing NO (catalytic), this reaction occurs, according to the authors, the following reaction for a thermoelectric of mineral coal:
2NH3 + SO3 + H2O ↔ (NH4)2SO4
equation (1)
NH3 + SO3 + H2O ↔ NH4HSO4
equation (2)
These results demonstrate the possibility of ammonium sulfate formation with the bubbling gases from the pyrite roaster furnace. The formation of SO2 and subsequent SO3 formation in aqueous solution with ammonium hydroxide results in the ammonium sulfate formation, which can be seen on the Figure-9 results.
4. Conclusions The thermal processing of mineral pyrite in different atmospheres is an interesting possibility for other works. This work concluded that pyrite thermal processing in a synthetic air atmosphere formed hematite as a main component.
The temperature from which the hematite begins its formation is 450 °C, according to the thermal treatments performed in this study.
There is the possibility to neutralize the SO2 formed in the heating reaction by bubbling this gas through a solution of ammonium hydroxide to provide the occurrence of ammonium sulfate (NH4)2SO4 precipitation.
Acknowledgments We acknowledge these institutions for the possibility of being at International Pittsburgh Coal Conference: FAPESC (Santa Catarina state of research) for the project funds; Criciúma Mining Company for the pyrite sample used in this project and for Universidade do Extremo Sul Catarinense for the characterization equipments.
References BANDYOPADHYAY D. Study of Kinetics of Iron Minerals in Coal by 57Fe Mössbauer and FTIR Spectroscopy during Natural Burning. Hyperfine Interactions, v. 163, p. 167-176, 2005. BELOLI, Mário; TABLES, Joice; GUIDI, A. History of Coal in Santa Catarina. Criciúma: Oficial Press of Santa Catarina State, 2002. CARUCCIO, F. T. and GEIDEL, G. Geochemical Factors affecting Coal Mine Drainage Quality, in Reclamation of Drastically Disturbed Lands, SCHALLER, FW and SUTTON, P. (Eds), American Society of Agronomy, Madison, WI, p. 129, 1978. FENGQI SI, CARLOS E. ROMERO, ZHENG YAO, ZHIGAO XU, ROBERT L. MOREY, BARRY N. LIEBOWITZ.Inferential sensor for on-line monitoring of ammonium bisulfate formation temperature in coalfired power plants.Fuel Processing Technology, Volume 90, Issue 1, January 2009, Pages 56-66 FERROW, E. A.; KALINOWSKY, B. E.; Veblen, D. R.; SCHWEDA, P. Eur Official Mineral, v. 11, p. 999, 1999. GUILIN HU, KIM DAM-JOHANSEN, STIG WEDEL, JENS PETER HANSEN.Decomposition and oxidation of pyrite.Progress in Energy and Combustion Science, Volume 32, Issue 3, 2006, Pages 295-314 HIDETSUGU NAKAMURA, MAKOTO IWASAKI, SYUNICHI SATO, YASUTAKE HARA.The reaction of ammonium nitrate with pyrite. Journal of Hazardous Materials, Volume 36, Issue 3, March 1994, Pages 293303. JAMES WEIFU LEE, RONGFU LI. Integration of fossil energy systems with CO2 sequestration through NH4HCO3 production. Energy Conversion and Management, Volume 44, Issue 9, June 2003, Pages 15351546. KREBS, A. S. J. Conceptual Design for the Environmental Restoration Basin coal South Santa Catarina. Preparation of Project; Environmental Recovery of carboniferous region of Santa Catarina, 144 f. 2001, Tese [Doctorate]. Area of Focus: Environmental Management. Florianopolis. NASCIMENTO, F. M. F.; MENDONÇA, R. M. G., Macedo, M. I. F., Soares, P. S. M. Environmental Impacts on Water Resources of the Coal Exploration in Santa Catarina. BRAZILIAN CONGRESS OF ROOFLESS MINE & II CONGRESS OF BRAZILIAN UNDERGROUND MINE, 2, 2002 - Belo Horizonte, Anais. PETERSON, Michael. Production of ferrous sulphate from the pyrite: sustainable development. 2008. 127f. Thesis (Ph.D. in Chemical Engineering). - Universidade Federal de Santa Catarina, Florianópolis, 2008. R.R. MARCELLO, S. GALATO, M. PETERSON, H.G. RIELLA, A.M. BERNARDIN; Inorganic pigments made from the recycling of coal mine drainage treatment sludge;Journal of Environmental Management, Volume 88, Issue 4, September 2008, Pages 1280-1284.
S.K. BHARGAVA, A. GARG, N.D. SUBASINGHE.In situ high-temperature phase transformation studies on pyrite.Fuel, Volume 88, Issue 6, June 2009, Pages 988-993 SOARES, Paulo Sérgio Moreira, SANTOS, Maria Dionísia Costa dos, POSSA, Mario Valente. . Brazilian Coal: Technology and environment. Rio de Janeiro: CETEM / MCT, 2008. 300p. TRADE UNIONS OF EXTRACTION OF COAL INDUSTRY OF THE STATE OF SANTA CATARINA. A new look at coal. Criciúma, SC. 2006. Available at: . Accessed on: 27 Sep 2008. WANG, H.; BIGHAM, J. M.; TUOVINEN, O. H. Oxidation of Marcasite and Pyrite by Iron-Oxidizing Bacteria and Archaea. Hydrometallurgy, v. 88, p. 127-131, 2007.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11-14, 2010 PYROLYSIS OF THE VARIOUS TYPES OF FUELS – CREDIT CARDS CIZKOVA Adela, RACLAVSKA Helena, JUCHELKOVA Dagmar, Roubicek Vaclav VSB – Technical University Ostrava, Czech Republic, [email protected]
Introduction Only in America the 1.6 billion of credit, debit, and ATM cards was produced in 2007. But those figures don't include the billions of gift cards, loyalty cards, and store charge cards we stuff in our wallets each year. The card industry is starting to experiment with compostable PVC and biopolymers (Doha Bank Green Credit Card), but we're still a long way away from corn-flavored MasterCards. In theory, plastic credit cards could last about eight years in our wallets. Instead, we cut them up when they expire—usually after two to four years http://www.slate.com/id/2216035/. The quantity per year of shredded PVC-based plastic cards represents a significant environmental problem in their disposal or recycling. The first card was issued in 1914 by the American Telephone and the Telegraph Company ―Western Union Telegraph Company‖. It was made from sheet metal and allowed regular customers to make calls and send telegrams without immediate payment. The customer then received an inventory of telephone calls and telegrams at the end of the month (invoice), their individual and total prices, which were paid by check or statement from the bank. The first universally usable credit cards were used in 1950 by the Diners Club for its members. The card was utilized payment in restaurants and later to all contracting hotels, restaurants and shops that had closed a contract with the club. The proper "bank" credit card was issued by The Franklin National Bank of New York in 1951. Another bank, which began to issue credit cards, was the Bank of America in 1958. Bank of America use cards from plastic as the first in the world. This cards allowed payment by means of mechanical sensors imprinter. This bank made available their successful project in 1966, to other U.S. banks and one Bank in England. So the cards got into Europe. Credit card technology Currently we have three technologies, under which the cards are manufactured:
Magnetic strip card - the back of the card serves as a medium for recording the identification data in electronic transactions. Card with a chip - the card includes a built-in microprocessor which can securely store the information needed to verify the client's personal code. Hybrid card - contains a chip and magnetic strip.
Cards themselves range in thickness from 0.25mm to 0.76mm and are typically made up of two or three layers of PVC-related polymers. Other card substrate materials studied are acrylonitrile butadiene styrene (ABS), polyethylene terephthalate (PETG), polycarbonate (PC) and even corn-based polylactic acid (PLA) -- did not fare much better that PVC in terms of their environmental impact. Usually, one of the layers is screen printed, and then a clear PVC layer is laminated on top. The printed surface, sandwiched between the two plastic sheets, acts as an adhesive, so the effects of printing on the card must also be assessed. Other features, such as special finishes, magnetic strips, signature panels, holograms and smart chips will also affect the design of the card. Chemical composition Ultimate and proximate analysis provided the basic information on the composition of "waste Cards". The chip or magnetic strip form the card was removed before analysis. The results are summarized in Table 1. Table 1 Ultimate and proximate analysis of credit card C H N O S Cl Moisture
Ash content
wt %
44.55
7.68
0.50
13.61
<0.01
27.9
0.41
5.76
NCV GCV kJ/kg 20447 22029
Explanation: NCV – Net calorific value, GCV – Gross calorific value
The mineral phase composition of ash from credit cards (prepared by combustion of sample at 850oC) was determined by X-ray diffraction. In the ash predominates rutile - TiO2 (89%), in smaller amount is present tenorit – CuO (11%). Further phases are present only in traces and their evaluation is not entirely clear. This is the suboxide Ti, i.e. TiO and phase CaCuTi 4O12. The important content of Ti in the credit card has been confirmed by chemical analysis (Table 2).
Fig.1 Fly ash form credit card. Dark crystals tenorite (CuO), white compound - rutile (TiO2)
Fig.2 Coke from pyrolysis - foam
Table 2 Content of elements in the credit card - the method of X-ray fluorescence (Innow X)
Ca Ti Fe Sn Cu Cr V Ni Bi Sr Zn Rb Mo Sb
Ash from Solid residue after combustion pyrolysis (mg/kg d.m. ± standard. deviation) 6295 10325 ±213 68412 14544 ± 182 11227 2166 ± 30 93 2043 ± 44 149 1589 ± 24 2435 61 ± 10 652 126 ± 14 1817 130 ± 14 5 9±2 10 ± 1 2115 ± 56
31983 4 42 132
Enrichment Ash/pyrolysis carbon 1.64 0.21 0.19 21.96 10.66
0.02 0.19 0.07 0.55 0.06
From the results shown in the table it is clear that Sn and Cu enriched in the combustion process. Other elements are substantially enriched in the reduction processes of pyrolysis: Fe, Cr, V, Ni and Zn. Pyrolysis of credit cards Given the problematic content of PVC, cards are not suitable material for incineration, even though they have relatively high calorific value. Therefore, they are laboratory and operationally tested using pyrolysis. Pyrolysis tests were performed on laboratory equipment at the Paar company Research Center, Coal Research Center, Illinois Coal Development Park in Carterville, in the laboratories of the Southern Illinois University Carbondale. PVC is a major component of plastic cards. Pyrolysis of PVC can be described in two steps. During the first step in the temperature range 200-375 °C release hydrogen chloride and originate conjugated double bonds with polyene structures. Small amount of hydrocarbons also originate, especially unsubstituted aromatics such as benzene, toluene and anthracene. During decomposition of polyenes originate a small amount of toluene and other alkylaromatics. PVC is the most problematic for the pyrolysis of plastic compounds. In the pyrolysis of PVC is developed HCl and chlorinated hydrocarbons such as methyl chloride, vinyl chloride and chlorobenzene (Miranda et al. 1999, Duangchan, A. et al., 2008). There are many processes to prevent production of HCl. For example, it is possible to use iron oxide as a chlorine sorbent or metal oxides, which are then calcinated to obtain chlorine, and the metal oxide is regenerated (Duangchan, A. et al., 2008). Pyrolytic gas and liquid from individual compounds of plastics or their mixtures is different. During the pyrolysis of mixtures is possible to observe the interaction between plastic compounds (Williams, P. T., et al. 2007).
Table 3: Measured volumes of gas obtained during the pyrolysis of plastic Cards CO 0
CO2 2.06
CH4 3.96
obj. %
73.19
0
9.17
17.63
Products (mL/min)
ml
H2 16.45
Time (min) CO2 mL/min
H2 mL/min
CH4 mL/min
Fig.3 Plastic cards - the release of gases during the course of pyrolysis Microanalysis of pyrolytic coke was carried out in the test laboratories of MITAL Steel, Inc. using EDS - energetic dispersive analysis (SEM Tescan VEGA TS 5130 LM + OXFORD IBCA Energy 350). Pyrolytic coke from plastic cards contained different metal-phases, most of which were particulate containing Cu-Ni-Sn at different ratio and modifications with other elements (one in Figure 4). Other interesting phases were barium chloride, AgS, CuS (or chalcopyrite) and particles containing Pd. Relatively large Au particles were also found (Fig. 5). The composition of pyrolytic coke is in the table 4, and chemical character of individual particles is published in table 5. Table 4 The composition of pyrolytic coke from plastic cards C O Al Si S Cl K (%) 42.82 36.91 0.15 11.60 0.09 0.18 0.11
Ca
Ti
Fe
0.18
7.26
0.71
Table 5 Microanalysis of individual grains with content of metals (%) O
8.79
Al
Si
0.77
10.59
0.80
26.69
0.93
32.54 16.02
0.88
0.67
S
Cl
Ca
6.71
1.04 0.77
14.62
P
9.04 9.79
8.02
Ti
Fe
Ni
Cu
1.48
14.08
70.45
0.53
2.82
31.02
49.92
3.03
1.49
Pd
Sn
4.44
Sb
Au
Pb
4.31 61.15
1.26 1.23
Ag
28.69 2.99
53.53
2.33
13.44
1.24
29.66
25.02
35.61
18.54 5.29 85.38
Figure 4: Microparticles containing Cu-Ni-Sn detected in pyrolytic coke from plastic cards
Figure 5: Au particle detected in the coke from the pyrolysis of plastic cards
Conclusion Credit Card Processing using pyrolysis technology offers several advantages. Technology minimize the amount of wastes and generate gas, which can be directly used (thermal discharge path, thereby increasing overall energy efficiency in the utilization of product gas). After managing the problems with the chlorine evolved from PVC, the gas can be utilized for energetic purposes. The results of pyrolysis tests indicate the possibility of further use of the "feedstock" for the extraction of the elements enriched in reducing conditions in the pyrolytic coke (Au, Cr, Ni, V and Zn).
Acknowledgement The article was supported by the project ME09054 intensification of gas production and SV 5410051/2101 - Optimisation of Gas Production during Pyrolysis of Municipal Waste. References Duangchan, A., Samart, Ch.: Tertiary recycling of PVC-containing plastic waste by copyrolysis with cattle manure. Waste Management, vol. 28, No. 11, 2008, pp 2415-2421 Juřík P. (2006): Platebni karty, Grada, 1-296. (In Czech). Miranda, R., Yang, J., Roy, Ch., Vasile, C.: Vacuum pyrolysis of PVC I. Kinetic study. Polymer Degradation and Stability, vol. 64, No 1, 1999, pp 127-144 Williams, P. T., Slaney, E.: Analysis of product from the pyrolysis and liquefaction of single plastics and waste plastic mixtures, Resources, Conservation and Recycling, vol. 51, 2007, pp 754-769
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COAL DERIVED PRODUCTS TITLE OF PAPER: ALLIANCE DCL TECHNOLOGY FOR PRODUCTING ULTRA CLEAN TRANSPORTATION FUELS
Theo L.K. Lee Vice President & CTO, Headwaters CTL, LLC. 1501 New York Avenue, Lawrenceville, New Jersey, U.S.A 08550 ([email protected]) John Duddy Senior Technology Advisor, Axens North America, Inc. 650 College Road East, Suite 1200 Princeton, New Jersey 08540 USA
Abstract: A strategic alliance to provide a single-source solution for producing ultra-clean fuels by direct coal liquefaction (DCL) alone or in combination with refinery residues or biomass has recently been formed between Headwaters Inc. and Axens. Depending on the properties of coal of interest, the alliance can apply either slurry or support catalyst system to effectively convert coal or in combination of biomass and petroleum residues into liquid hydrocarbons using a commercial proven back-mixed reactor system. This paper provides recent results of direct liquefaction and refining studies for a bituminous coal and also for a co-feed consisting of a lignite and petroleum resid. Properties of intermediate and finished coal derived liquids as compared to petroleum derived fuels will be discussed. DCL is a hydrogen addition process and its hydrogen usage is substantially more than that for hydroprocessing of petroleum crude. This paper will compare impact of hydrogen generated by various methods, (e.g. steam methane reforming, coal and/or biomass partial oxidation, and electrolysis) on carbon footprint from DCL.
DIRECT COAL TO LIQUIDS (DCL): HIGH QUALITY JET FUELS
W. Weiss (1) *, H. Dulot, A. Quignard, N. Charon, M. Courtiade (1): IFP Energies nouvelles, 69360 Solaize, FRANCE
Summary The worldwide demand of fuels has been intensified in recent years and is expected to continue growing. To satisfy these energy requirements and diversify the source of fuels, the energy industry has to face the challenge of using alternative feedstocks in order to produce transportation fuels like jet or diesel. Direct Coal Liquefaction (DCL) process enables liquid yields higher than indirect liquefaction via Syngas and FT process (typically 3.5 bbl/T coal for the best available DCL processes compared to 2,5 bbl/T coal for indirect process, on a dry ash free basis including hydrogen production and upgrading in both cases). In order to obtain high quality transportation fuels, raw coal liquid effluents derived from Direct Coal Liquefaction need to be severely upgraded using hydrotreating and hydrocracking. This work focuses on the characterization of physical and chemical properties and composition of jet fuel cuts obtained by DCL before and after hydroprocessing, and shows that high quality jet product could be obtained using appropriate hydrocracking conditions. Keywords Direct coal liquefaction, jet fuel, quality, GC-2D
Introduction The worldwide demand of fuels has been intensified in recent years and is expected to continue growing. To satisfy these energy requirements and diversify the source of fuels, the energy industry has to face the challenge of using alternative feedstocks in order to produce transportation fuels like jet or diesel. Direct Coal Liquefaction (DCL) process, such as Axens H-CoalTS process, enables liquid yields higher than indirect liquefaction via Syngas and FT process (typically 3.5 bbl/T coal for the best available DCL processes compared to 2,5 bbl/T coal for indirect process, on a dry ash free basis including hydrogen production and upgrading in both cases). This work focuses on the characterization of physical and chemical properties and composition of jet fuel cuts obtained by DCL before and after hydroprocessing, and shows that high quality jet product could be obtained using appropriate hydrocracking conditions.
Materials and Methods A full boiling range effluent (C5-450°C) from DCL was separated into different distillation fractions which were
*[email protected]
then characterized using complementary analytical techniques. In our work, the upgrading of direct coal liquefaction effluents was carried out under severe hydrocracking (HCK) operating conditions, in a high-pressure fixed-bed reactor, under a total pressure higher than 100 bar and at relatively moderate temperatures. The liquid hourly space velocities (LHSV) are among the lower ones of HCK processes. The catalyst used in these experiments was a base metal hydrotreating / hydrocracking catalyst which was sulfided in situ. A comparative experiment in high pressure hydrotreating conditions was also performed. After reaction, effluents were cooled, condensed and separated into a gas phase and a liquid phase. Liquid phase effluents were analyzed by means of a physical distillation (ASTM D2892) and characterized using standard petroleum analysis and also using a multidimensional gas chromatography device (2D-GC or GCxGC) equipped with a flame-ionization detector and cryogenic system[1].
Results and Discussion The chemical composition of raw DCL effluent, including kerosene and diesel fractions, was found to be much more complex compared to conventional petroleum fractions. Cut
C5-380°C
2D-GC is a powerful analytical tool which give quantitative chemical families analyses. 2D-GC is now widely used in IFP Energies nouvelles for naphtha[2], kerosene, diesel[3,6,8,9] or vacuum distillate[4,5,7,10] cuts characterization. In our study, 2D-GC was used to optimize upgrading conditions in regards to petroleum analyses, by understanding structure changes that are occurring during hydrotreating and hydrocracking.
d154
g/cc
0.9255
carbone hydrogen NMR oxygen
wt %
87.26
wt %
11.24
wt %
1.28
nitrogen
wt ppm
2100
sulfur
wt ppm
100
nC14
nP + iP
wt %
4
nC13
Naphthenes
wt %
47
Aromatics
wt %
49
Tableau
1
:
Raw
Direct
Coal
Naphthenoaromatics Naphthenes Paraffins Di-aromatics +
Liquefaction
effluent
characterisation
d154 (g/cc) sulfur (wt ppm) nitrogen (wt ppm) Paraffins (wt %) Naphtenes (wt %) Aromatics (wt %) Diaromatics + (wt %) Smoke point (mm) Cetane Number Cetane Index
Monoaros DCL KERO
Jet A1 Spec.
> 0.89
0.7750.84
< 200
DCL GO
US Spec.
EU Spec.
> 0.92
-
0.8200.845
< 3000
< 500
< 15
< 10
> 1500
-
> 2000
-
-
<6
-
<3
-
-
> 40
-
-
> 55
< 35
-
> 50 > 35
-
Figure 1 : GCxGC analysis of 180-250°C cut obtained by DCL
Naphtheno aromatics Naphthenes Paraffins
Di aromatics +
<3
< 11 (< 8)
>5
(naphth alenes)
> 20
-
12-14
> 19
-
-
-
-
-
25-35
> 40
> 51
-
-
-
> 40
> 46
Tableau 2 : Characterisation of Kerosene and Diesel cuts obtained by DCL
Compared to crude oil cuts, products obtained from direct coal liquefaction are highly condensed (polynaphthenes and polyaromatics), with a very low amount of paraffinic compounds, resulting in poor combustion properties. The heteroelements analyses are also unusual with very high nitrogen and oxygen contents but low sulfur content. In order to achieve the specifications of transportation fuels, these products have to undergo a further purification and upgrading treatment that can be performed in a hydrotreatment / hydrocracking process at very high severities.
Monoaros Figure 2 : GCxGC analysis of 180-250°C cut after Hydrotreating DCL effluent
2D-GC Analysis wt %
DCL KERO
HDT DCL KERO
200-250°C
180-250°C
3.8/2.2
n/i Paraffins CnH2n+2 monoNaphthenes CnH2n di-Naphthenes CnH2n-2 tri-Naphthenes CnH2n-4 Aromatics CnH2n-6 poly-Aromatics CnH2n-8 / CnH2n-10 Aros + triNaphthenes Specific Gravity
Exemple of JET A1
2D-GC Analysis wt %
3.7/1.8
10.9/30
6.5
7.1
23
23.5
41.1
14.5
26.3
24.7
0
30
21.2
18.3
7.7
0.4
3.3
64
46.3
21.6
0.908
0.884
0.81
n/i Paraffins CnH2n+2 monoNaphthenes CnH2n diNaphthenes CnH2n-2 triNaphthenes CnH2n-4 Aromatics CnH2n-6 polyAromatics CnH2n-8 / CnH2n-10 Aros + triNaphthenes Specific Gravity
Tableau 3 : GCxGC analysis of kerosene cuts before and after hydrotreating
Hydrotreating only allows partial aromatics hydrogenation resulting in a slight density decrease. Mono-aromatics and tri-naphthenes content remains high while paraffins content is very low which does not result in a significant smoke point improvement with a typical value to be around 17 mm for a hydrotreated kerosene. Naphtheno aromatics
Naphthenes
Paraffins
Di-aromatics+
Monoaros Tableau 4 :
GCxGC analysis of 200-250°C cut after
Hydrocracking DCL effluent
DCL KERO 200250°C
HCK DCL KERO 180250°C
HCK DCL KERO 135250°C
Exemple of JET A1
3.8/2.2
2.9/5.2
3.6/11.7
10.9/30
6.5
17.8
24.8
23
23.5
56.3
49.7
14.5
26.3
14.8
8.0
0
30
3.0
1.9
18.3
7.7
0.0
0.0
3.3
64
17.8
9.9
21.6
0.908
0.857
0.837
0.81
Tableau 5 : 2D-GC analysis of kerosene cuts before and after hydrocracking
Hydrocracking allows aromatics hydrogenation which results in naphthenes content increase. The structure of hydrocracked effluent is highly poly-naphthenic but the amount of tri-naphthenes is reduced when compared to a hydrotreated product. The decreases of tri-naphthenes can be explained by ring opening reactions that occur during hydrocracking. The paraffin content increases when the initial boiling point decreases. Hydrocracking reactions allow a significant density decrease of the kerosene cut. However, with typical cut points for kerosene such as 180-250°C, specific gravity remains slightly higher than the upper limit of specific gravity specification of 0.84. This is related to the high naphthenes content as these compounds have important densities compared to the corresponding paraffins with the same carbon atoms number. However, the density can be decreased by incorporating some heavy naphtha in the kerosene pool. This operation is not detrimental to the kerosene quality and particularly the flash point which remains around 50-75°C which is well above the Jet A1 specifications of 38°C.
Aromatics (vol %) D1319 Specific gravity D4052 Flash Point (°C) D93 Freezing Point (°C) D5972 Heat of combustion (MJ/Kg) D3338 Heat of combustion (MJ/L) Sulphur (wt %) Viscosity @ 20°C (cSt) Smoke Point (mm) OR Smoke Point (mm) AND Naphthalenes (vol %) Hydrogen Content (wt %) D3343 BOCLE (mm) D5001 JFTOT ΔP @260°C (mmHg) Tube deposit rating (visual)
DCL + HCK KERO
World average Jet A, Jet-A1 , JP-8*
Jet A1
JP-8
<5
17-18
< 25
< 25
< 0.84
0.80-0.81
0.7750.84
0.7750.840
50-75
47-52
> 38
> 38
< - 65
-50
< -47
< -47
43.2
43.1
> 42.8
> 42.8
36.3
34.5
< 1 ppm wt
0.02
< 0.3
< 0.3
The mainly naphthenic structure and the related density presents the advantage of a significantly higher volumetric heat of combustion (+ 5%) when compared to the average value calculated from a World fuel sampling program[13]. A higher density jet fuel can be produced which results in a longer range for jets for a given volume of fuel.
4.0-7.6
4.0-5.0
<8 > 25
> 25
Conclusions
> 19
> 19
22-25
<1
1.5
<3
<3
13.4 – 13.7
14.1-14.3
-
> 13.4
0.6
0.6-0.65
< 0.85
<1
< 25
1
<3
point of the kerosene cut before upgrading is only in the range of 12-14 mm. In addition, Cetane Number of kerosene after hydrocracking is typically around 35-40 which allows this cut to be partially sent to the diesel pool. The diesel cut obtained by hydrocraking a DCL effluent has been previously studied[11,12] and also exhibits good combustion properties with Cetane Number higher than 50.
Detailed molecular analysis is mandatory to explain the good combustion qualities of Hydrocracked DCL products. Comprehensive 2D-GC is a powerful analytical tool which give quantitative chemical families analyses. Taking into account the good smoke point, it seems that naphthenes structure found in hydrocracked DCL kerosene can exhibit good combustion properties contrary to what is generally admitted for petroleum fractions. Kerosene cut obtained after DCL effluent hydrocracking complies with all Jet A1 or JP-8 specifications.
*World fuel sampling program, CRC Report No. 647, 2006
Tableau 6 : Petroleum analyses of kerosene cut after hydrocracking
Petroleum analyses showed that the kerosene cut complies with all Jet A1 or JP-8 specifications, i.e. density, flash point, freezing point, viscosity, smoke point, specific energy, lubricity, thermal stability. The DCL jet fuel after hydrocracking exhibits excellent cold weather properties having a freezing point of less than -65°C. Aromatics content after HCK is very low (less than 10 % wt, typically less than 5%), nitrogen and sulfur levels after HCK are extremely low (less than 1 ppm wt). With an initial boiling point around 135-150°C, the flash point is in the range 50-75°C and well above the specification which allow the refiner to have more flexibility between naphtha and kerosene production. The most severe hydrocracking conditions allow to obtain a kerosene that exhibits good combustion properties. Typical smoke point values are around 22-25 mm for a kerosene obtained after hydrocracking whereas the smoke
Direct Coal Liquefaction in conjunction with optimized hydrocracking can therefore be considered as a key process to provide alternative ultra low sulfur transportation fuels with excellent cold weather properties, high volumetric heat of combustion and good combustion characteristics, with an optimized high liquid yield on coal.
References [1]. T. Dutriez, M. Courtiade, D. Thiébaut, H. Dulot, F. Bertoncini, J. Vial, M-C. Hennion, J. Chromatogr. A., 1216 (2009) 2905. [2]. F. Adam et al, J. Chromatogr. A, Volume 1178, Issues 1-2, 2008, pp 171-177 [3] C. Vendeuvre et al,. J.Chromatogr. A, 1086 (2005) 2128 [4] T. Dutriez et al., J. Chromatogr. A, Volume 1216, Issue 14, 2009, pp 2905-2912 [5] T. Dutriez et al., Journal sep science, volume 33, issue 12, 2010, pp 1787-1796 [6] F. Adam et al, J. Chromatogr. A, Volume 1217, Issue 8, 2010, pp 1386-1394 [7] T. Dutriez et al, Anal. Chem., submitted [8] F. Adam et al, Fuel, Volume 88, Issue 5, 2009, pp 938946
[9] F. Adam et al, J. of Chromatogr. A, Volume 1148, Issue 1, 2007, pp 55-64 [10] T. Dutriez et al, 34th Int. symposium on CC and 7th GCxGC Symposium, 2010, Riva, italia [11] H. Dulot et al, ISCRE 21, 2010, Philadelphia, US [12] A. Quignard et al, World CTL, 2010, Beijing, China [13] O.J. Hadaller, J. M. Johnson, World Fuel Sampling Program, CRC Report No. 647, 2006
1
EXTRACTION OF BROWN AND SAPROPELITIC COALS WITH TOLUENE AND WATER CONTAINING FLUIDS UNDER SUPERCRITICAL CONDITIONS P.N.Kuznetsov, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences 42 K.Marx str, 660049 Krasnoyarsk, Russia Fax: 7(391)2124720; E-mail: [email protected] S.M.Kolesnikova, L.I.Kuznetsova, E.S.Kamensky, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences, 42 K.Marx str, 660049 Krasnoyarsk, Russia Fax: 7(391)2124720; E-mail: [email protected] V.A.Kashirtsev, Trofimuk Institute of Petroleum Geology and Geophysics SB RAS, 3 Koptyug ave, 630090 Novosibirsk, Russia. Fax:7(383)3332301; E-mail: [email protected] Abstract. The conversion of brown coals from the Borodino and Kangalas deposits and sopropelitic coal in an aqueous medium and in a toluene containing mixture with water and tetralin was studied under supercritical conditions over the temperature range of 375–550°C and at pressures from 13 to 40 MPa. It was found that the methanation, hydrolysis, and oxidation reactions of brown coals with the predominant formation of gaseous products (methane, carbon dioxide, and hydrogen) prevailed in an aqueous medium. Liquid substances were formed in an insignificant amount. In the toluene solvent under supercritical conditions at 440°C, the addition of a small water amount (15 %) stimulated the degradation of coals with the predominant formation of liquid products and moderate gas formation. The use of calcium oxide and sodium hydroxide as catalysts increased the yields of liquid products. It was noted that the reactivity of Kangalas coal in this process was higher than that of Borodino coal. Much higher conversions were achieved in the extraction of sapropelitic coal. In the supercritical toluene containing media with small amount of hydrogen donor tetralin cosolvent the conversion achieved 79 % at 4000C.
Under conditions of rising oil and gas prices and a gradual depletion of oil and gas deposits and propagation to hard-to-reach northern areas, high-efficiency technologies should be developed based on the use of coals as raw materials for chemical processing. The organic matter of coal (OMC) mainly has a macromolecular structure [1], and it contains molecular-size pores. Because of this, chemical transformations can be limited by the transfer of solvent molecules [2]. Along with the development of traditional technologies for coal conversion, processes based on the chemistry of extreme actions attract increasing attention. In particular, processes with the use of solvents under supercritical conditions are promising [3–14]. The solvents in a supercritical state have an increased dissolving power, a low viscosity, and high diffusion coefficients, which are close
2
to those of gases. The set of the above properties makes it possible to increase the conversion and selectivity of coal transformations into liquid products. In this case, the resulting extract can be easily separated from the solid coal residue (mineral matter and unconverted organic matter) by depressurization. Organic compounds from various classes (aliphatic and aromatic hydrocarbons, alcohols, and heterocyclic compounds) are used as solvents. Tetralin, in which the degree of coal conversion at 450°C was as high as 95% [10], exhibits high efficiency. Toluene, which has easily attainable critical temperature (318.6°C), pressure (4.11 MPa), and density (0.292 g/cm3) and is a typical component of low-boiling fractions obtained in various coal conversion processes, is of special interest to technology [4, 5, 11, 14–16]. However, as a rule, the degrees of coal conversion in toluene under supercritical conditions are low [4, 17]. According to Kershaw et al. [4], the conversion of Australian brown coals at 400°C varied from 30 to 35 %. For medium-rank coals, it was lower than 16 % at a temperature of 425°C [17]. Co-solvent additives, which actively react with organic matter of coal (OMC), are used to increase the efficiency of processes in toluene [4, 6, 8, 18]. Thus, the conversion of brown coal in toluene with the addition of only 5% tetralin under supercritical conditions increased by 10–15% [4]. Water is a unique solvent in terms of abundance and, environmental and technological safety (Tc =374°C, Pc = 21.8 MPa, and ρc = 0.322 g/cm3). It has been found that water stimulates the cleavage of ester bonds [9, 19–21] to provide the deep conversion of OMC. Processes for coal gasification with water under supercritical conditions have been proposed [22–24]. The efficiency of processes in these solvents under supercritical conditions can be increased using catalysts that can stimulate the degradation of OMC [25, 26]. In general, note that the body of published data concerning the use of solvents under supercritical conditions for coal processing is limited, and the physicochemical characteristics of occurring reactions are scantily known. In this paper, we report the results of a study of the conversion of two different brown coal samples under supercritical conditions in solvents containing water and toluene and the effect of alkaline catalysts on the efficiency of this process. EXPERIMENTAL The samples of brown coals from the Borodino deposit of the Kansk-Achinsk Basin and the Kangalas deposit of the Lena Basin and sapropelitic coal from Irkutsk deposit were used in this study. In terms of the composition of organic matter, sapropelitic coal differed from the brown coals in a higher hydrogen content (Table 1). The concentration of calcium and magnesium compounds in the Borodino brown coal ashes (Table 2) was higher (30.0 against 20.0 % in the Kangalas coal), but the concentration of alkali metals was much lower (0.6% against 2.6 %). The reaction was performed in a rotating 250 ml autoclave over a temperature range from 250 to 550°C at working pressures from 13 to 40 MPa. A 15-g portion of coal and required amounts of a
3
Table 1. Composition of coals Deposit Coal Kangalas Borodino Irkutsk
brown coal brown coal sapropelitic coal
Аd, wt. % 8.0 4.9 21.6
Ultimate analysis, % on a daf basis C H N S O 71.0 5.6 1.0 0.5 21.9 71.9 4.6 1.0 0.4 22.1 73.0 9.1 1.0 0.7 16.2
Table 2. Concentrations of the main components of brown coal ashes Deposit SiO2 Al2O3 CaO MgO Fe2O3 Kangalas 53.0 15.0 17.0 3.0 7.0 Borodino 52.0 6.0 25.0 5.0 6.0
Na2O 2.0 0.3
K2O 0.6 0.3
solvent and a catalyst were loaded in the autoclave. Calcium oxide (3 wt % on an OMC basis) and sodium hydroxide (2 wt %) served as catalysts, and water and toluene and the mixtures with water and tetralin in a ratio of 85 : 15 were used as solvents. The autoclave was made airtight, purged with argon to remove air, and heated at a rate of 10 K/min. The reaction time at a specified temperature was 1 h. After completion of the reaction and cooling of the autoclave, the volume and composition of the resulting gases were measured. The solid and liquid products were removed from the autoclave to a filter. The solid residue on the filter was exhaustively extracted with toluene in a Soxhlet apparatus. The degree of coal conversion was determined from the change in its ash content after the reaction. The yield of liquid and soluble products was calculated by difference between the overall conversion and the yield of gaseous products. RESULTS AND DISCUSSION Tables 3 and 4 summarize process performance characteristics for the conversion of brown coals in an aqueous medium at various temperatures. In the experiments with Borodino coal at a supercritical point, that is, at 375°C and a water density of 0.322 g/cm3, the coal conversion was 10 %, the gas yield being 15 wt.% (calculated on the weight of OMC), i.e. more than coal Table 3. Characteristics of brown coal conversion in an aqueous medium Deposit Water vapor Reaction Working Coal Yield, wt % on daf coal density, temperature, pressure, conversion, gas* soluble* 3 g/cm °C MPa wt % Borodino 0.322 375 22 10 15.0 (6.4) – (3.6) The same 0.380 410 35 24 21.5 (9.1) 2.5 (14.9) Kangalas 0.130 550 29 52 47.8 (29.1) 4.2 (22.9) The same 0.170 550 40 51 53.1 (30.9) – (20.1) *In parenthesis are given the values recalculated based on a carbon balance.
4
conversion. As the temperature was increased to 410°C, the density of water vapor, to 0.380 g/cm3, and the pressure, to 35 MPa, the coal conversion increased to 24 %, the yield of gaseous products to 21.5 wt.%, and the yield of solubles to only 2.5%. Table 4 shows that the gases consisted predominantly of carbon dioxide (78.5-81.5 vol.%), and it could result not only from the conversion of coal but also from conversion of water (due to oxidation of carbon). This could result in reduced yield of solubles (calculated from the difference between coal conversion and the yield of gaseous product related to only OMC). Taking this into account, the recalculation based on a carbon balance was made and percentage of carbon (of coal) converted into carbon containing gases (CH4, CO2 and CO) was estimated, and from this, the percentage of carbon converted into solubles was determined. Given in Table 3 in parenthesis are the recalculated yields based on carbon balance. The data show that 3.6 % and 14.9% of carbon of Borodino coal were converted into solubles at 3750C and at 4100C, respectively. To reach a higher degree of coal conversion, the reaction with Kangalas coal was performed at an elevated temperature of 550°C. In these experiments, supercritical conditions were not attained because at the density of water vapor even 0.13–0.17 g/cm3 the working pressure in the autoclave increased to 40 MPa. Under these conditions, the conversion of coal was as high as 51–52%. The process was accompanied by intense gas formation (yield of 47.8–53.1 on an OMC basis), methane being a predominant gas component (53.9-57.9 %) with less amounts of carbon dioxide and hydrogen (23.0–29.4 and 12.8–16.5 %, respectively) (Table 4). Based on carbon balance, the percentages of carbon converted into solubles reached 20.1-22.9 % (shown in parenthesis) at this temperature. Table 4. Composition of the gaseous products of brown coal conversion in an aqueous medium Deposit Water vapor Reaction Gaseous products, vol. % 3 density, g/cm temperature, °C СН4 СО СО2 Н2 other Borodino 0.322 375 5.6 4.0 81.5 6.8 1.8 The same 0.380 410 13.4 2.9 78.5 5.0 0.2 Kangalas 0.130 550 57.9 0.8 23.0 16.5 1.8 The same 0.170 550 53.9 0.7 29.4 12.8 3.2 From the experimental data, it appears that the decomposition of brown coals in an aqueous medium at both low (375–410°C) and high (550°C) temperatures followed by gasification reaction with conversion of carbon into solubles of less than 23%. In toluene solvent with the addition of 15 % water, Kangalas coal also underwent gasification to form mainly methane (see Tables 5 and 6) in supercritical conditions (for toluene) at a temperature of 550°C. A decrease in the temperature to 440°C resulted in a dramatic decrease in gas formation. Upon the conversion of Borodino coal, the yield of gaseous products decreased to 14.7 %. Unfortunately, it was impossible to determine the yield of soluble products based on carbon balance for the reaction of coal with the mixture of toluene and water. However, because of not high
5
yield of gases, the percentage of carbon of Borodino coal converted into solubles at 4400C could be nearly 16.3 % (Table 5) determined by conventional procedure. In the presence of calcium oxide and sodium hydroxide as catalysts, the coal conversion increased by 8–12 %, as compared with that in an uncatalyzed process (Table 5). In this case, the yield of gases (primarily, carbon dioxide and methane) was no higher than 19.1 %. The increase in conversion coal was mainly due to liquid substances (36.1–37.7 and 19.9–25.6 % for Kangalas and Borodino coals, respectively). Table 5. Characteristics of brown coal conversion in a mixture of toluene with water under supercritical conditions Deposit Catalyst Toluene and Reaction Working Coal Yield, wt % on an water vapor temperature, pressure, conversion, OMC basis density, °C MPa wt % gas soluble 3 g/cm substances Borodino – 0.36 + 0.06 440 32 31 14.7 16.3 The same CaO 0.36 + 0.06 440 32 41 15.4 25.6 NaOH 0.36 + 0.06 440 33 39 19.1 19.9 '' Kangalas – 0.22 + 0.04 550 30 42 46.9 – The same – 0.36 + 0.06 440 29 41 – – СаО 0.36 + 0.06 440 34 51 14.9 36.1 '' NaOH 0.36 + 0.06 440 32 53 15.3 37.7 ''
Table 6. Composition of the gaseous products of brown coal conversion in a mixture of toluene with water under supercritical conditions Deposit Catalyst Toluene and Reaction Composition, vol. % water vapor Н2 other temperature, СН4 СО СО2 3 o density, g/cm C Borodino – 0.36 + 0.06 440 33.1 9.0 47.4 6.5 4.0 The same CaO 0.36 + 0.06 440 30.2 11.9 49.4 7.3 1.2 '' NaOH 0.36 + 0.06 440 34.1 0.4 55.5 8.4 1.6 Kangalas – 0.22 + 0.04 550 73.2 0.8 18.3 6.0 1.7 The same СаО 0.36 + 0.06 440 39.2 7.8 39.1 9.8 4.1 NaOH 0.36 + 0.06 440 34.1 0.7 49.7 13.1 2.4 '' Without – 0.36 + 0.06 440 0.0 14.4 43.5 42.1 coal In special experiments, it was found that, upon loading a toluene–water mixture with no carbon into the reactor, the amount of formed gaseous products was smaller than that formed with coal by a factor of 20. In this case, CO2 (43.5 %) and H2 (42.1%) were the predominant constituents
6
of the resulting gas; the concentration of CO was much less (14.4 %), whereas methane was not formed at all (Table 6). Degradation products (including benzene) were almost absent from a liquid phase. The experimental data suggest that an excess of the yield of gases over the overall conversion of coal was likely due to carbon oxides, which resulted from the oxidation of coal with water. Data in Table 5 indicate that Kangalas coal differs from Borodino coal in higher reactivity in terms of overall conversions and yields of soluble and gaseous products. This can be due to both a higher hydrogen content of Kangalas coal (as compared with that of Borodino coal) and a higher concentration of alkali metal compounds in the mineral matter of coal; it was found that these compounds exert a catalytic effect on the process. Much higher conversions were achieved in the extraction of sapropelic coal (Table 7). In this case, toluene exhibited a higher efficiency (64 %) than water (60 %). As in the extraction of brown coals, the mixtures of toluene with water exhibited synergistic effect as compared to the individual solvents; the mixture of toluene with tetralin was the most efficient (79 %). Thus, the extraction of sapropelitic coals in toluene under supercritical conditions with the use of small additives of a polar solvent (water) and hydrogen donors (tetralin) allowed a deep degradation of organic matter to be achieved. Table 7. Sapropelitic coal conversion in a mixture of toluene with water at 4000C Solvent Solvent vapor Working Coal 3 density, g/cm pressure, MPa conversion, wt. % Toluene (soxhlet apparatus) 2 Toluene 0.40 13 64 Water 0.12 24 60 Toluene+water 0.36 + 0.06 28 70 Toluene+tetralin 0.36 + 0.06 13 79 ACKNOWLEDGMENTS This work was supported in part by an Integration Program of the Siberian Branch of the Russian Academy of Sciences (Project No. 106). REFERENCES [1] Krichko A.A., Gagarin S.G., Makar’ev S.S. Khim. Tverd. Topl. (Moscow) 1993; 6: 27. [2] Kuznetsov P.N., Kuznetsova L.I. Khim. Tverd. Topl. (Moscow) 2008; 6: 57. [3] Bartle K.D., Martin T.G., Williams D.F. Fuel 1975; 54(4): 226. [4] Kershaw J.R., Overbeek J., Bagnell L.J. Fuel 1985; 64(8): 1070. [5] McHugh M.A., Krukonis V.J. Supercritical Fluid Extraction: Principles and Practice, Butterworth_Heinemann, Boston; 1994. [6] Shishido M., Mashiko T., Arai K. Fuel 1991; 70(4): 545. [7] G_Hourcade M.L., Torrente C., Galan M.A. Fuel 2007; 86: 698.
7
[8] Sangon S., Ngamprasertsith S., Prasassarakich P. Proc. 12 Int. Conf. Coal Sci., Cairns, 2003, p. 3B5 on CD ROM. [9] Townsend S.H., Klein M.T. Fuel 1985; 64: 635. [10] Kershaw J.R., Jezko J. Separation Sci. Technol. 1982; 17(1): 151. [11] Brunner G. Gas Extraction: An Introduction to Fundamentals of Supercritical Fluids and the Application to Separation Processes, Springer, Darmstadt; 1994. [12] Rocha S.R.P., Oliveira J.V., Avila S.G. et al. Fuel 1997; 76: 93. [13] Yuan Q., Zhang Q., Hu H., Guo S. Fuel 1998; 77: 1237. [14] Murty Balijepalli S.N., Ravikumar Y.V.L., Dutt N.V.K. Fuel 1997; 76(2): 165. [15] Simsek E.H., Karaduman A., Caliskan S., Togrul T. Fuel 2002; 81: 503. [16] Hourcade M.L., Torrente C., Galan M.A. Fuel 2007; 86: 698. [17] Philip Cahill P., Harrison G., Lawson G.J. Fuel 1989; 68: 1152. [18] Kolesnikova S.M., Kuznetsov P.N. Khim. Tverd. Topl. (Moscow) 2008; 2: 21. [19] Tanaka T., Tsutsumi A., Yoshida K., Inaba A. Abstracts of Papers, Int. Conf. Coal Sci., Tokyo, 1989. [20] Townsend S.H., Klein M.T. Fuel 1985; 64: 635. [21] Hu H., Go S., Hedden K. Fuel Proc. Technol. 1998; 53: 269. [22] Bermejo M.D., Cocero M.J., Fernandes-Polanco F. Fuel 2004; 83: 195. [23] Vostrikov A.A., Psarov S.A., Dubov D.Yu. et al. Khim. Tverd. Topl. (Moscow) 2007; 4: 29. [24] Vostrikov A.A., Fedyaeva O.N., Psarov S.A. et al. Khim. Tverd. Topl. (Moscow) 2007; 5: 30. [25] Barton P. Ind. Eng. Chem. Process. Des. Dev. 1983; 22: 589. [26] Sato Y., Chakrabarty S.K. Liquid Fuels Technol. 1983.
Research and development of new type reactor for direct coal liquefaction HU Fating, LI Peilin, SHI Shidong, LIU Min Beijing Research Institute of Coal Chemistry, China Coal Research Institute, Beijing 100013, PR China Abstract: At present, bubble column reactor and ebullated bed reactor are applicable to the direct coal liquefaction (DCL) reactor. However, there are more or less some defects in both of them. Loop reactor is applied widely in chemistry industry, biochemical engineering, environmental engineering, petro-chemical industry and other industries for its simple structure and many merits. In order to introduce loop reactor into DCL industry, we have established a world’s first large-scale pressurized loop reactor cold-state simulation apparatus. The reactor was 0.6 m in diameter and 6.6 m in height, and its draft tube was 0.4 m in diameter and it 5.75 m in height. It simulates the fluid dynamics behaviors within the DCL reactor by using nitrogen-water system. The effects of superficial gas velocity, superficial liquid velocity, pressure and surface tension on gas holdup in the loop reactor were investigated; the anti-interference ability of the loop reactor also was investigated. The explorative experiment that the loop reactor is applied to the DCL process under actual liquefaction conditions(pressure=19.0MPa, temperature=455℃) has been conducted in the 6 ton coal/day process development unit (PDU). The test has set a precedent that loop reactor can be used as DCL reactor. The experiment results show that the loop reactor’s coal conversion rate is slightly higher than that of the ebullated bed reactor while the oil yield slightly lower than that of the ebullated bed reactor. Therefore, loop reactor can be used in the coal direct liquefaction process. Key words: loop reactor; gas holdup; cold-state simulation experiment; direct coal liquefaction
1. Introduction At present, bubble column reactor and ebullated bed reactor are applied to the direct coal liquefaction (DCL) reactor. However, there are more or less some defects in both of them. Loop reactor (LR) is a new type multiphase reactor which was developed on basis of reinforced bubble column reactor. Fluid within LR has regular circulation flow, which leads to LR have better characteristics of heat conduction, mass transfer and mixing. LR is fit for homogeneous and non-homogeneous phase reaction of gas-liquid, liquid-liquid and gas-liquid-solid [1], and was applied widely in chemistry industry, biochemical engineering, environmental engineering, petro-chemical industry and other industries for its simple structure and many merits [2]. In order to introduce loop reactor into DCL industry, we have established a world’s first large-scale pressurized LR cold-state simulation apparatus according to coal’s hydrogenated liquefaction reaction features. Making use of this apparatus, we .simulated hydrodynamic behavior in DCL reactor using nitrogen-water system. At present, the most study of LR focus on small-scale LR and in atmosphere conditions, but the researches of large-scale LR under elevated pressures conditions have not appears [1,3]. Therefore this paper has important innovation significances in the LR and DCL reactor aspect. The explorative experiment that the loop reactor is applied to the DCL process under actual liquefaction conditions has been carried out in the 6 ton coal/day process development unit (PDU). The features of bubble column reactor, ebullated bed reactor and LR were analyzed in the end.
2. Experimental
2.1 Experimental apparatus and process The cold-state simulation loop reactor(LR) which was 0.6 m in diameter and 6.6 m in height, and can bears pressure of 4.0 MPa is a kind of internal loop airlift reactor. It’s draft tube was 0.4 m in diameter and it 5.75 m in height. Below the draft tube, there is a gas-liquid distributor posses many nozzles which was 170mm in height and 5mm in inside diameter. The Fig. 1 below shows the process of cold-flow model devices. Gas Inlet and Discharge
Gas Recycle Compressor
Loop Reactor
Gas-liquid Separator
9
9
liquid Pump
Fig.1
Flow chart of cold-state simulation experimental apparatus
1. gas inlet 2.draft tube 3.gas-liquid nozzle 4.gas-liquid distributor 5.differential manometer 6.liquid inlet 7.gas flow meter 8.liquid flow meter
9.gas-liquid separator of compressor
When carry cold-flow model test, a certain amount of liquid was pouring into the LR and the separator in advance. Then a certain amount of nitrogen was introduced to this experimental apparatus that can maintain certain pressure as test required. So the gas and liquid were circulation flow inside the experimental apparatus. The density difference of fluid in draft tube and annular space drive gas-liquid performing regular circulation movement. 2.2 Measurement of gas holdup parameters Gas holdup is a important parameter that characterized hydrodynamic behavior of LR. It has great role in the LR’s scale-up design, model establishing, industrial production control. As fluid which has different gas holdup have different static pressure, we can determine the mean gas holdup of local region by the static pressure difference method. Since the liquid within differential manometer is same with that of LR, the gas holdup is defined by below formula.
g
h H
(1)
Where, Δh is numerical reading of differential manometer, and ΔH is distance between two test points of differential manometer. In this cold-flow apparatus, loop reactor was equipped with four
differential manometers which can measure the gas holdup of upper and lower inside draft tube (i.e. fluid riser region) and annulus (i.e. fluid downcomer region) respectively.
3. Results and discussion This paper first had many cold-state simulation experiments using nitrogen-water system. The gas holdups of nitrogen-water system in the loop reactor (LR) were investigated deeply. Gas holdup is the most basal characteristic parameter of gaseous phase in the LR. It is the indicators of gas-phase average residence time and gas-liquid coefficient of mass transfer, and influences other characteristic parameters directly or indirectly. Gas holdup also affects liquid-phase circulating velocity, further affects the mixing behavior of the LR [4].
3.1 Average gas holdup inside draft tube 3.1.1 The relation between gas holdup and superficial gas velocity At different gas flow rate with fixed liquid flow rate, we recorded the readings of differential manometers of upper and lower of draft tube (i.e. fluid riser region). According to average reading value of this both, we made out the relation figure (i.e. Fig. 3) between gas holdup and superficial gas velocity(hereafter, referred to as SGV) at different operating conditions, As the difference of readings between the upper and lower of draft tube is very small. 0.6 0.5
εg
0.4 0.3 0.2 0.1 0 0
0.05
0.1
0.15
Ug /m·s-1 Fig.3 The relation between gas holdup and superficial gas velocity
In Fig. 3, εg means gas holdup, and Ug means superficial gas velocity (SGV). From Fig. 3, we can see that gas holdup is increased with SGV rising. At low SGV, the gas holdup is increased linearly; at high SGV, the increased trend of gas holdup is decreased. Yang haiguang et al. attained similar research conclusions [5,6,7,8]. Therefore, the gas holdup should not be increased by improving superficial gas velocity at high SGV.
3.1.2 The effect of superficial liquid velocity We can obtain various lines of relationship of gas holdup and SGV at different superficial liquid velocities. There are four lines in Fig. 4 which the operating pressure is atmosphere pressure and Fig. 5 which the operating pressure is 0.5 MPa.
0.4
0.3
0.35
0.25
0.3
0.2
0.25
0.0074m/s 0.005m/s 0.0024m/s 0.003m/s 0.006m/s
0.15 0.1 0.05 0
εg
εg
0.35
0.003m/s
0.2
0.0072m/s
0.15
0.0042m/s
0.1
0.0051m/s
0.05
0.0061m/s
0
0
0.01
0.02
0.03
0.04
0
Ug /m·s -1
0.01 0.02 Ug /m·s -1
Fig. 4
0.03
0.04
Fig. 5
Fig.4 The change of gas holdup with superficial gas velocity (SGV) at ambient pressure Fig.5 The change of gas holdup with superficial gas velocity (SGV) at 0.5 Mpa pressure
The numbers in Fig. 4 and Fig. 5 represent the magnitude of superficial liquid velocity. Gas holdup is increased with SGV rising, and the lines distributed closely, as shown in Fig.4 and Fig. 5. The small differences between every line illustrated that gas holdup only slightly varied at different SGV. The reasons for this phenomenon may be due to higher ratio of height and diameter of the LR that lead to liquid circulation velocity is relatively small.
3.1.3 The effect of surface tension on gas holdup The surface tension of liquid in the loop reactor (LR) can be changed by adding surface active agent. When add different amount surfactant, we can obtain liquid systems with different surface tension. We carried many cold-state simulation tests in various liquid systems which has different surface tension. From these experimental results, we can analyze the effects of surface tension on gas holdup. 0.6
1.0MPa 0.033N/m 1.0MPa 0.073N/m
0.5
2.0MPa 0.073N/m
εg
0.4
2.0MPa 0.032N/m 1.0MPa 0.054N/m
0.3
1.0MPa 0.021N/m
0.2 0.1 0 0
0.01
0.02
0.03
0.04
Ug /m·s
0.05
0.06
0.07
0.08
-1
Fig.6 the effect of superficial gas velocity on gas holdup at different surface tensions
In Fig. 6, every line means different experimental liquid systems with different surface tension at different pressure. Fig. 6 shows that gas holdup decreased with the increase of the surface tension. This phenomenon can be explained by energetic theory of bubble coalescence. Because the properties of bubble coalescence varied with the change of surface tensions, when the
surface tension of experimental system is changed, the bubbles become smaller lead to velocity of bubbles slower, and consequently the gas holdup becomes bigger.
3.1.4 Influence of pressure At different gas-phase pressure conditions, we made the relation diagram of gas holdup and superficial gas velocity (SGV) when the liquid flow rate not varied. We can investigate the effect of gas-phase pressure on gas holdup from these diagrams (Fig. 7).
0.6 atm
0.5 1.0MPa
εg
0.4
1.0MPa repeat 2.0MPa
0.3 0.2 0.1 0 0
0.02
-1
Ug /m·s
0.04
0.06
0.08
Fig.7 Influence of superficial gas velocity on gas holdup at different pressures Fig.7 shows that the repeatability of experimental data at 1.0 MPa pressure is very well, and that is to say experimental data are liable. At the same SGV, the gas holdup increased with the increase of the pressure. The reason is gas bubbles diameter diminished and draft velocity decreased when gas-phase pressure increased. Zhang tongwang and Jin haibo et al. else obtained similar conclusion in the study of gas holdup in the bubble column reactor [9].
3.2 Comparison of local gas holdup in LR There were four accessory figures indicate different pressure conditions in Fig. 8 below. Four lines in every accessory figure represent gas holdup of upper of draft tube, upper of annulus, lower of draft tube and lower of annulus respectively.
Fig. 8 Effect of superficial gas velocity on local gas-holdup in different pressures Fig. 8 shows that gas holdup in draft tube and that in annulus increased with increasing of superficial gas velocity (SGV). Gas holdup of upper of draft tube was higher slightly than that of lower of draft tube, but difference not large, and the difference decreased with increasing of SGV and system pressure. Because there is diffuse procedure of gas phase in draft tube after irrupting from nozzles. Only floating a short distance, the gas phase can distribute evenly in cross section of draft tube. When making axial movement along draft tube, the bubbles diameter increased slightly with the decrease of pressure around the bubbles. The effect of diffuse procedure on gas holdup of upper and lower of draft tube became small with the increase of SGV and system pressure, as well as the bubble-diameter. Fig. 8 also shows that gas holdup of annular upper, upper and lower of draft tube higher strikingly than that of annular lower. This is largely attribute to following reasons. first, not all of bubbles which was introduced into annulus from draft tube can drop into annular lower, And the bubbles looped from annulus to draft tube is found to be negligible; second, some bubbles that was taking into annulus probably coalescence into large bubbles, and the large bubbles’ descend velocity became lower, even a portion of them can return to the upper of annulus. Therefore, there are more bubbles in annular upper than that of annular lower, and so the gas holdup of annular lower much lower than that of other local regions in LR.
3.3 Operating flexibility test of LR In industrial production process, the flow rate fluctuations of gas or liquid that was introduced into reactor always have adverse influences. When cold-state simulation unit running steadly, the liquid-level of gas-liquid separator and looping condition of LR are keep stable. In order to investigate the operating flexibility of LR, we changed the flow rate of gas or liquid deliberately. The Fig. 9 is the schematic diagram of working-condition variations in the experimental unit when gas flow rate diminished abruptly.
Fig.9 The change of working status with time when gas flow diminished abruptly
It is showed that liquid-level returned into stable position in short time (about nine minutes), and the LR still kept looping state. We observed similar results with Fig. 9 when liquid flow rate varied abruptly. Therefore, LR has excellent operating flexibility and strong anti-interference
abilities.
4. Actual direct coal liquefaction test of LR The LR (loop reactor) is a kind novel reactor for direct coal liquefaction (DCL). We established the hydrodynamics mathematical model of LR on basis of many hydrodynamics tests at cold flow model experimental apparatus. The results of this tests and mathematical model calculation showed that LR had liquid velocity of 0.2-0.5 m/s and middling gas holdup compared with bubble column reactor and ebullated reactor. Accordingly, it is wholly feasible for LR that applied to the DCL process. The DCL Process Development Unit (PDU) having a capacity of 6t of coal per day has been successfully operated during the period 2004~2007 at Shanghai, China, and possessed two ebbulated reactors which connected in series. The ebbulated reactors were modified into loop reactor, whose draft tube was 200 mm in diameter, having uniform dimension (inside diameter=320 mm, height=6 m) with original reactors. The DCL exploratory test using loop reactors of PDU were first carried out under actual DCL conditions (temperature=455℃, pressure=19.0 MPa) for more a month. Experimental results comparison of loop reactor with ebullated reactor at the same reaction conditions was given in Table 1. Table 1 Comparison of experimental results in LR with that in ebullated reactor Reactor mode Actual oil yield, daf%
Ebullated reactor
Loop reactor
49.08
48.37
water, daf%
10.83
11.65
C1-C3, daf%
6.67
6.96
COx, daf%
2.57
2.08
H 2 S, daf%
0.15
0.17
Gas yield,daf%
9.39
9.21
Hydrogen consume, daf%
4.21
4.26
85.19
87.95
Coal conversion, daf%
The experimental results show that the loop reactor’s coal conversion rate is slightly higher than that of the ebullated bed reactor while the oil yield slightly lower than that of the ebullated bed reactor. Therefore, loop reactor can be used in the coal direct liquefaction process for the industrial practice. However, the scale-up design, optimal design and theoretical study required further research.
5. Comparisons of various DCL reactors The bubble column reactor has less internal component, high gas holdup, enough gas-liquid mass transfer rate, maturated technology, and high reliability in operating. Nevertheless, bubble column reactor has two main disadvantages: firstly, liquid-phase velocity is so small slightly that is close to or less particles setting velocity that can lead to solid settlement problem after operating for long time, and required to periodically discharge residue. Secondly, as hydrodynamic restrictions, the production scale of bubbling reactor whose maximum treatment capacity of coal is 2500t/d can’t very big. The ebullated reactor has forced internal liquid circulation improves the gas-liquid-solid contact performance owing to the increase of liquid flux into the reactor. In addition, it has
advantages of uniform temperature distribution in the reactor, easy temperature control, no deposit of solid particles in reactor bottom. But ebullated reactor also has some defects. For example, it need install recycle pump which can operate at high temperature and high pressure. Moreover, ebullated reactor top must have internal components and space for gas-liquid separation. Otherwise, gas-liquid separate inadequately prone to exhausting of recycle pump. These components and space increased the structure complexity and limited the ratio of gas-liquid. Loop reactor (LR) main merits are the fluid circulated in regular direction, liquid circulating velocity fast relatively, liquid-gas attained entire backmixing mode, and no deposit of solid particles. LR has big mass transfer coefficient because Gas pass very long path in residence time, and gas bubbles have large specific surface area. LR also not need expensive recycle pump and complicated inner components, and so that reduce operating cost of reactor and operating risks for recycle pump breakdown. In summary, LR has bright prospects of application for DCL industry. Although above three kinds reactors have some differences, they have many same specialty such as chemical reaction process, slurry with flow condition, approximate reacting temperature and pressure, residence time etc.
6. Conclusions Specific conclusions of the study are the following. 1. Gas holdup increases with the increase of superficial gas velocity, however, increasing trend of gas holdup will becomes weaken when superficial gas velocity is bigger. 2. Superficial liquid velocity has little impact on relation of gas holdup and superficial gas velocity. 3. Enhancing operating pressure of gas-phase can improve gas holdup in LR. Changing surface tension of liquid systems can change gas holdup of LR. 4. The gas holdup of annular lower is far below that of annular upper and draft tube. The gas holdup has little difference between draft tube upper and lower. 5. Loop reactor has excellent operating flexibility and strong anti-interference ability. 6. Loop reactor was carried out actual feed-coal DCL experiment, and LR has better performance. 7. The advantages and disadvantages of bubbling bed reactor, ebullated reactor and loop reactor, were compared and analyzed. Loop reactor has bright prospects for DCL industry. References [1]Wang Juan, Mao Yu, Liu Yansheng, et al. Effect of gas distributor configuration on gas holdup distribution in loop reactor. Journal of Chemical Industry and Engineering, 2005, 56(1):58 [2]Zhang Yongli, Liu Yongmin, Zhang Hong.The development of investigation in loop reactor. Liaoning Chemical Industry, 2002, 31(9):410 [3]Bi Weidong, Que Guohe. Simulation Research of Gas Holdup in the Medium-Sized Continuous Residuum-Hydrocracking Slurry-Bed Reactor Under Environmental Conditions. Journal of The University of Petroleum China, 2001, 25(3):19-22 [4]Nie Dashi, Cui Yingying, Zhang Qiang et al. The characteristics and applications of airlift loop reactors. Journal of Chemical Industry & Engineering, 2004, 25(6):6-7 [5]Liu Yongmin, Zhang Qiang, Zhang Kezheng. Hydrodynamics of an air-lift loop reactor. Journal of Petrochemical Universities, 2001, 14(4):13-14
[6]Yang Haiguang, Fan Yi, Li Fei, et al. Liquid circulation velocity and local gas holdup in air-lift loop reactor. Journal of Chemical Engineering of Chinese Universities, 2003, 17(1):37-39 [7]Liu Mengxi, Lu Chunxi, Chu Ling,et al. Theoretical analysis of gas holdup in a novel air loop reactor. Journal of Chemical Engineering of Chinese Universities, 2005, 19(3):332-336 [8]Zhang Tongwang, Gao Jixian, Wang Tiefeng, et al. Hydrodynamic behavior in three-phase external loop airlift reactors. Journal of Chemical Industry and Engineering, 2005, 56(7):1213-1217 [9]Zhang Tongwang, Jin Haibo, He Guangxiang, et al. Large bubble and small bubble holdups in large-scale pressurized bubble column reactor. Chemical Engineering, 2004, 32(5):29-33
Characterization and distribution of phenolics in direct coal liquefaction oils Gao
Zhennan
Ph.D.
Mao Xuefeng
Engineer
Beijing Research Institute of Coal Chemistry, China Coal Research Institute, Beijing 100013, China (email:[email protected]) Li Wenhua
Prof.
GE China Technology Center, Shanghai 201203, China
Key words: direct coal liquefaction, phenolic, characterization and distribution, derivatization
Introduction Phenolics are the predominant oxygen-containing components in the direct coal liquefaction oil (DCLO), and account for 15% of DCLO (IBP~350℃)approximately. In a traditional DCL process, DCLO is hydrotreated unexceptionally to produce gasoline, diesel oil or kerosene; and thereby the concentration of phenolic components must be reduced sharply, in course of hydrotreating, to especially low level to come up to the requirement of quanlified products. Since phenolic compounds, however, are valuable, even rare raw material or intermediates for lots of fine chemical processes actually, the phenolics should be pre-removed, separated and purified subsequently to produce series of high profitable by-products, instead of being hydrotreated. Qualitation and quantitation of phenolics in DCLO is the fundamental step for choosing by-products and designing separation route on each compound. Because the distribution of phenolics in DCLO has been shown to affect coal degradation and reactivity, upgrading to fuels, and stability, much more attention has been paid to the analysis of these compounds during the last three decades of 20th century, and almost every analytical method available has been applied to the analysis. Limited achievement, however, has been possessed till now, because chromatographic detection technique specific for oxygen-containing compounds in DCLO is deficient at that time, and phenolics are difficult to analyse for their high polarity and strong reactivity.
Samples and DCL Bench Scale Unit Two oil samples (1# and 2#) were obtained from 0.1t/d DCL bench scale units (BSU) at China Coal Research Institute (CCRI). Sample 1# was derived from Shenli subbituminous coal from China operating in NEDOL mode, and sample 2# was derived from powder river basin (PRB) coal from U.S.A. operating in Shenhua mode. The DCL process flow sheets of NEDOL and Shenhua BSU could be found in the doctoral dissertation[1]. The samples are steady-up runs of continuous experiments, which were conducted with different types of catalysts and under different operation conditions, and thereby reflect the influence of different preparing conditions. A summary of main operation conditions is given in table 1.
Experimental Section Extraction and Derivatization. DCLO is such a complex mixture that it fails to be separated well even by state-of-the-art gas chromatography (GC) column. Therefore, the “phenolic concentrate” (PC) should be isolated from DCLO beforehand. In the manuscript, the caustic
1
extraction procedure ever introduced by Pauls[2] was optimized and used to obtain PC (1#, 2#) samples from DCLO. The oil sample was washed firstly by aqueous NaOH (weight ratio of 1/1, 10% solution, three times), then the aqueous was separated, blended and acidified with 10M H2SO4 to release a PC. After phase separation again, the aqueous layer was extracted with methylene chloride, which was removed finally by rotary evaporation, to recover additional phenolics. Analyses were performed using “derivative mixture” (DM) mainly. The DM (1#, 2#), a mixture of silica ether compounds or series of derivatives of phenolics, is prepared through a substitution reaction between a 29Si reagent and phenolics in PC. N,O-bis(trimethylsily1) trifluoroacetamide (BTSFA) was chosen as the 29Si reagent because it has been proved effective and recommended by Dadey[3]. PC was diluted with methylene chloride, then BTSFA was added with weight ratio of 1/5. Almost all phenolics in PC could react entirely and the solution become the derivativces mixture after thoroughly mixing and 24 hours standing. Analysis PC(1#~2#) and DM(1#~2#) were analyzed by combined gas chromatography/mass spectrometry (GC/MS) employing a Shimadzu GC17A and a Shimadzu QP5050 for speciating. All samples were also subjected directly to GC employing a Agilent 6890N for quantitating. Table 2 contains a summary of GC/MS conditions.
Results And Discussion The gas chromatograms of PC and DM. Figure 1 (a) and (b) shows the gas chromatogram profiles of PC 1# and DM 1#, respectively. Those of PC 2# and DM 2# are similar. At first glance, it is easy to find that the amount of dicephalous peak and shoulder perk decreases largely on gas chromatogram profile of DM, comparing to that on profile of PC. This is because the derivatives of many isomers of phenolics, which could not be separated in PC, could be separated thoroughly in DM. For example, peaks encycled by two frames on profile (a), representing C1-phenol and C2-phenol respectively, overlap widespreadly; whereas on profile (b), peaks representing isomers of derivatives of those phenols stand independently one another. Moreover, the resolution of gas chromatogram of DM is higher than that of PC. This means more accurate qualitation result could be obtained if matching mass spectra of DM. Characterization of phenolics. Results of identification of phenolics by matching mass spectra of DM 1#~2# are summarized in table 3. Because the mass spectra alone is insufficient to distinguish positional substituents, alkyl-substituted phenolics are classified and counted according to alkyl carbon number. From the data in table 3, it is obvious to find that the phenolics isolated from DCLO consist of four types of ring structures: phenols, indanols, naphthalenols and naphthols, although four DCLO samples were derived from two different types of coal under different liquefaction conditions. No other ring structures, like biphenyols or acenaphthols, were detected, which were found by Pauls [2] in several DCLO samples derived from Illinois No. 6 bituminous coal and Wyodak subbituminous coal. Besides the coal rank, we think, the different phenolics type also are the results of different liquefaction processes and different boiling ranges.
2
The statistic data in table 3 also indicate nearly 140 phenolics existing in DCLO. Among those compounds, about 90 phenols account for rough 65% of total number of compounds, and indanol and naphthalenol account for nearly 21% and 11%, respectively. Moreover, all of the phenolic compounds, essentially, are alkyl substituted with short chains. Phenols have more complex alkyl group constitution relatively. As shown in table 3, C4-phenol contains more than 20 individual phenols, and the maximum alkyl carbon no. is 6. As for indanol and naphthalenol, methyl substitutes contain the most individual compounds, as well as the maximum alkyl carbon no. is 3 and 2 respectively. The exclusive alkyl chain of naphthol is methyl. Distribution of phenolics. The areas of the gas chromatogram were normalized to provide a distribution of the phenolics. The results are reported in table 4. Results of quantitative analysis in table 4 reveal that phenol is leading component of PC, thereby the phenolic group of DCLO, accounts for approximate 78% of PC. Indanol and naphthalenol account for round 16% and 4%, and naphthol takes the rest. As for each type of phenolic, the carbon number distribution of alkyl carbons all maiximized at C1, then followed by C2. Unsubsitituted through C3-alkylated phenols accounted for over 85% of the phenols. Unsubsitituted through C2-alkylated indanols accounted for nearly 98% of the indanols. The principal naphthalenol species were methylnaphthalenols, which accounted for over 70% of the naphthalenols. Moreover, the content of phenolics without or with fewer alkyl substitutes is more than that of those with more substitutes, although the total number of compounds of the latter phenolic group is much more than that of the former phenolic group. For instance, the content of C1-phenol (including 3 isomers) is more than that of C4-phenol (including 22 isomers). Acknowledgement. We wish to acknowledge Du Shufeng for her extraction work. We are also grateful to professor Shi Shidong and Li Wenbo for discussions. We wish to thank the research team working at CCRI for samples preparation. References 1. Gao Zhennan. Study on distribution and extraction process of phenolic compounds in coal liquefaction oil [doctoral dissertation]. China Coal Research Institute. 2009 2. Richard E Pauls, Mark E Bambacht, et al. Distribution and characterization of phenolics in distillates derived from two-stage coal liquefaction. Energy & Fuels, 1990, 4(3):236-242 3. Eric J Dadey, Stanford L Smith. Determination of Hydroxyl Group Concentration in Coal Liquids by 31P NMR. Energy & Fuels, 1988, 2(3):326-332
3
Table 1
Summary of main operating conditions for DCLO preparation
Coal Catalyst Catalyst/Coal
/%
Coal slurry feed rate Flash H2 feed rate Recycle gas volume rate
1#
2#
Shenli
PRB
Fe
Fe
1
1
/kg·h
-1
8
12
3
-1
4.0
5.0
3
-1
4.0
7.0
-1
/m ·h /m ·h
G/L
/L·kg
1000
1000
Retention time
/min
120
120
Reaction Temp.
/℃
455
457
Preheater Temp.
/℃
400
380
HHPS outlet Temp.
/℃
420
400~420
Reaction pres.
/MPa
19.0
19.0
Table 2
Summary of GC/MS Conditions
GC part Column
DB-Perto
100m×0.25mm×0.5μm
Carrier gas
He(>99.99%)
Carrier gas flow rate
1.2 mL/min
Split ratio
80:1
Injection volume
1μL
Injector temp.
300 ℃
Detector temp.
300 ℃
Makeup H2
30 mL/min
Makeup Air
400 mL/min
Makeup N2
25 mL/min
Oven program
70 ℃(1min)
4 ℃/min
200 ℃(3min)
2 ℃/min
MS part Source temp.
320 ℃
Ionization mode
EI
Electron energy
70 eV
Mass analyzer
Sigle quadrupole rods
Scan rate
0.5 s/decade
Scan rage
30~300 m/z
4
250 ℃(1min)
C1-phenol
(a)
C2-phenol
(b)
Fig. 1
GC profiles of PC and DM
5
Table 3 Characterization of phenolics in DCLOs number of compounds at alkyl carbon no.
Phenolics
C0
C1
C2
DCLO 1# 9 16 22 10 2 2
Phenols Indanols Naphthalenols
1 2 3
3 15 9
naphthols
2
2
1 2
3 17
Naphthalenols naphthols
3 2
11 2
Table 4
Phenols Indanols Naphthalenols naphthols
C4
C5
C6
21
15
87 29 14 4
Phenols Indanols
Phenolics
C3
Total
DCLO 2# 9 16 25 13 2
22
16
92 34
5
19 4
Distribution of phenolics in PC Normalized % at alkyl carbon no.
C0 14.23 5.7 0.52 --
C1 21.55 6.58 2.16 --
Phenols Indanols
13.23 5.11
22.62 7.96
Naphthalenols naphthols
0.46 --
3.87 --
C2
C3
PC 1# 19.58 14.13 3.11 0.2 0.19 PC 2# 18.87 11.86 4.77 0.19 0.97
Total
C4
C5
C6
6.62
2.91
2.11
81.13 15.59 2.87 0.41
5.74
1.93
1.69
75.94 18.03 5.30 0.73
6
2010 International Pittsburgh Coal Conference DEVELOPMENT OF REAL-TIME DYNAMIC SIMULATION OF CHEMICAL LOOPING PROCESS FOR ADVANCED CONTROLS Hao Lei, Senior R&D Engineer Alstom Power Inc. 2000 Day Hill Road Windsor, CT, USA Email: [email protected] Telephone: 860-285-3284 Abhinaya Joshi, R&D Engineer Alstom Power Inc. 2000 Day Hill Road Windsor, CT, USA Email: [email protected] Telephone: 860-285-2753 Xinsheng Lou, Technology Leader Alstom Power Inc. 2000 Day Hill Road Windsor, CT, USA Email: [email protected] Telephone: 860-285-4982 Carl Neuschaefer, Director Alstom Power Inc. 2000 Day Hill Road Windsor, CT, USA Email: [email protected] Telephone: 860-285-9290 Abstract: Alstom Power Inc. (Alstom) is collaborating with the U.S. Department of Energy (DOE) in a multiphase project developing an entirely new, ultra-clean, low cost, high efficiency power plant technology for the global energy market. This new power plant concept is based on a process utilizing high temperature chemical and thermal looping technology. This chemical looping (CL) technology can be configured as a next generation power plant with a controlled stream of CO2 for use or sequestration. In coordination with the process development project efforts, Alstom is investigating and developing advanced controls for this chemical loop system under an DOE Advanced research project co-sponsorship. A key part of this controls project is development of a new computational approach to process dynamic simulations for use in the control work. The Phase I project focused on developing an early understanding of the basic transport process and the underlying process control dynamics. The project includes characterization of the chemical looping process, building solid transport process math models, and dynamic simulation software to support control investigations. This first of a kind simulator and Alstom’s experimental facilities were used in exploring advanced controls concepts such as model predictive control (MPC) for application to the chemical looping process. A paper was presented at PCC 2009, summarizing the progress of the Phase I R&D project executed by Alstom Power. In order to enable a smooth transition from the current Phase I (simulation and controls of the pilot scale process) project to Phase II (scale-up CL prototype process 3MWth, to be erected in Windsor, CT, US in 2010-11), a Phase I Extension period was created between DOE and Alstom to focus on preparations for the scale-up modeling, real-time simulation, and control specification development for the Phase 2 efforts to address the forth coming 3MWt CL prototype.
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
This paper, in addition to the off-line dynamic process simulator, presents the recent work conducted to develop a Real-time Simulator. First, the paper introduces the technology background and project status. Then, the paper presents the technical approach to first principle modeling of a hot solid multi-loop flow model, model discretization and software and numerical simulation development in the simulation environment. The simulation results and numerical problems resolved during the computational studies are presented at a proper technical level. Next, the paper shares the experiences in using a Real-Time Workshop (RTW) toolbox to convert the models into real-time target files to be used with customer specified user interfaces. The resulting real-time simulator with the basic closed-loop PID controls is then introduced in the modeling and simulation section as well. The results from the testing of Real-time advanced controllers on the simulator are presented as a validation of the real-time simulation platform. Operational optimization along with advanced controls has been established as a key milestone for this future clean power generation plant research and development project. To conclude this paper, further discussions are extended on the future phases of the project for integrated control and optimization designs along with Alstom’s clean fossil power system development. Keywords: Chemical Looping, Modeling, Real-time Simulation, Advanced Controls, Optimization 1
Introduction
Alstom Power Inc. is developing an entire new, ultra-clean power plant for the future power market. This new power plant concept is based on a chemical looping process, which is based on a process utilizing high temperature chemical and thermal looping technology. As estimated by Alstom, this new technology can provide over 90% CO2 capture with the lowest cost among all the clean coal technologies alternatives. In the future, due to environmental dispatch and integration of more and more renewable units, the power generating market will become more competitive. As such, there is a big demand to economically optimize the thermal and emission performance of the power generating units, improve load-changing flexibility and to reduce the overall plant operating cost. In order to identify the potential for improving the overall performance of the chemical looping process by integrating the application of advanced controls, Alstom’s Thermal Products R&D have launched a project of dynamic simulation and control design with the cosponsorship from the U.S. Department of Energy (DOE). It is also expected that adoption and integration of advanced controls in the early stages of process development will help to solve the inherent process control and operating challenges associated with chemical looping process. This dynamic simulation and advanced controls project runs in parallel to the chemical looping process development. At this time, the real chemical looping facility is still under design and construction. Hence for the modeling and control project, it is necessary to build a first principle based simulator, which can represent steady state and dynamic behavior of the chemical looping process with an engineering level of accuracy and then use it for both model-based operation simulations and control design evaluations. This simulator provides an affordable way of investigating and evaluating not only process control behavior, but also a large number of control schemes. Furthermore, in order to test the functionality and the performance of control system, a real time simulator should be developed. The development of controls at this early stage based on the simulators allows engineers to achieve high process reliability during process scale up and demonstration. Advanced controls technologies are to be studied to maximize thermal and emission performance and minimize operating costs. This project will deliver a dynamic simulation platform and new plant control designs that will be implemented on the prototype chemical looping facility and set the stage for future commercial-scale demonstration power plants with a goal to economically optimize the cost of electricity and CO2 capture for use or sequestration. A general summary about the technical approaches in this project is shown in Figure 1.
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
Process understanding
Mathematical modeling based on the first principles
Offline dynamic simulation with a proper software platform
Parameter estimation and model validation
Real time simulation with real time workshop
Offline control simulation studies
Control design tests on the real time simulator Control implementation on the real facilities Figure 1. A General Road Map for Technical Approaches At this stage, the Phase I dynamic simulation and advanced controls project has been completed. A Phase I Extension period was created in order to enable a smooth transition from the current Phase I (simulation and controls of the pilot scale process) project to Phase II (scale-up CL prototype process - 3MWth, to be erected in Windsor, CT, US in 2010-11). The Phase I Extension will focus on the scale-up modeling, real-time simulation and control specification development, and get prepared for dealing with the forth-coming 3MWt CL prototype in the Phase II project on simulation and control. In Phase I, several dynamic simulators have been developed for dynamic analysis and control study purposes using the first principles [3]. Based on the MATLAB/Simulink simulations, real time simulation kernels were generated using the Real Time Workshop toolbox from Mathworks. Different control strategies including benchmark PID, extended PID, linear MPC, nonlinear MPC and so on were developed and tested on the simulators to regulate the solids transport process. Some promising control algorithms were further implemented on the pilot test facilities. In this paper, the work of developing a real time simulator is presented in detail. The organization of this paper is as follows: Section 2 gives a brief overview about the dynamic offline simulation; Section 3 discusses the selection of software platform and the functionality design for real time simulation; the simulation scenario design and simulation results from the real time simulation are presented in Section 4; Section 5 concludes this paper with a brief summary on the development work and future plans. 2
A Brief Overview about the Offline Simulator for Chemical Looping Process
The chemical looping process is an ingenious fossil energy technology for clean power production with CO2 capture. The overall main chemical looping process takes place in two steps and keeps the air away from the fuel in two separate reactors with the help of oxygen carriers [1]. Currently, there are two kinds of oxygen carriers under investigation at Alstom (see Figure 2). One is Calcium based and the other is Metal Oxide based. The models presented in this paper are derived for calcium-based process.
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
Figure 2. Alstom’s Calcium Based and MeOx Based Chemical Looping Processes In order to have successive chemical reactions in the chemical looping system, it is very important to maintain a continuous circulation of the solids in the loops. Therefore, how to regulate and maintain a stable solids mass flow rate in the chemical looping process becomes one of the major tasks to operate the process. In Phase I project, the process simulation focused on studying the dynamic responses of the solids transport between loops under different operation conditions. To build a mathematical model and subsequent dynamic simulation for a process system like chemical looping, it is necessary to define boundaries, inputs and outputs. The model is built up from the cause-effect relationships between inputs and outputs at certain boundary conditions. For the chemical looping process, the inputs are the operational parameters, e.g., input gas and solids flows, design parameters such as the reactor geometry and other physical. The process model outputs are both the thermal and mechanical performance parameters, such as the change of temperature, solids mass flow and pressure in the system. The governing equations in the first approach for the gas-solids flow are described as follows: The governing equations for gas phase:
∂ ∂ (ερ g ) + (ερ g vg ) = 0 ∂t ∂z ερg vg2 f gF ∂ ∂ ∂P 2 (ερg vg ) + (ερgvg ) +ε + ερg g + + (1− ε ) ρs FD ( vg − vs ) =0 ∂t ∂z r ∂z 2 P vg ∂ ερ g h − + ρg 2g ∂t
vg2 ∂T ∂ ∂ + (ερ g v g h + )+ kg g 2g ∂z ∂z ∂z −
2h (T − T ) 6 (1 − ε )hsg (Ts − Tg ) + wall g amb + S rg = 0 dp r
The governing equations for solids phase:
∂ ∂ ((1 − ε ) ρ s ) + ((1 − ε ) ρ s vs ) = 0 ∂t ∂z ∂ ∂ ((1 − ε ) ρ s vs ) + ((1 − ε ) ρ s vs2 ) + (1 − ε ) ρ s g − (1 − ε ) ρ s FD (vg − vs ) = 0 ∂t ∂z (1 − ε )ρ s c ps ∂Ts + (1 − ε )ρ s c ps vs ∂Ts + 6 (1 − ε )hsg (Ts − Tg ) + S rs = 0 ∂t ∂z d p
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
where ε is the void fraction of gas, ρ is the density, v is the velocity, P is the pressure, h is the enthalpy, T is the temperature, Sr is the external heat source, r is the radius of the pipe, dp is the diameter of the solids particle, cp is the specific heat, and g, fgF, FD, kg, hwall, hsg are different physical coefficients in the system. The subscripts g and s stand for gas and solids phases, respectively. To simulate the dynamic process with derived mathematical models, it is desired to solve these partial differential equations (PDEs) stated above. Many numerical techniques have been developed in the literature to solve PDEs. The most widely popular numerical technique is the finite difference method [2]. The finite difference method can be applied to elliptic or parabolic PDEs and generate linear or nonlinear algebraic equations. The aim of this method is to replace the derivatives in the governing equations by finite difference approximations. However, the finite difference methods may generate high order equations due to fine mesh and is not suitable for on-line control design. The numerical method of lines [4] is another way to solve PDEs, which is similar to the finite difference method. However, unlike the finite difference method, the numerical method of lines does not discretize the time variable. As a result, it produces a set of ordinary differential equations (ODEs), which can be solved by using some standard commercial packages such as MATLAB/Simulink. An interface of an offline simulator for hot flow dual loop chemical looping process in MATLAB/Simulink is shown in Figure 3. This dynamic simulator can be used to simulate and study the solids transport between two loops under the change of temperatures. Different control scenarios related to temperatures, e.g., how to maintain the solids mass flow rate and the solids residence time in the reactors under the change of temperatures and how to regulate the reactor temperatures, etc. can be tested in this simulator. Alstom has test facilities for chemical looping cold and hot flow process development and testing. Test data were collected from these test facilities to estimate model parameters, validate the model and support the simulation work.
Figure 3. Alstom’s Dynamic Offline Simulator for Hot Flow Dual Loop Chemical Looping System
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
3
The Development of a Real Time Simulator
When a simulation is capable of executing transient solutions carried out in close to real perceived time, it is referred to as real-time (RT) simulation. The real time simulation is mainly used for testing control design and hardware-in-the-loop simulation. Having such a simulation tool is important in process analysis and control design evaluations. Real time simulation has advantages over off-line simulation. For example, more accurate analysis of system performance is possible by running the simulator and the actual physical system side-by-side. Furthermore, control design implementations for the chemical process can be tested more practically on a real time simulator, providing an opportunity to design ready-to-use control systems for the physical process. Unlike the offline simulation, where the inputs are predetermined, the inputs to the real time simulation are obtained from external devices at each sampling time. In a prescribed time interval, the real time simulator must have the abilities to solve all model equations and present the selected variables as outputs at the next sampling time. The requirement of calculation at a predetermined time brings a big computational burden for the simulation. In order to provide real time communication with external world and finish model calculation precisely within each sampling interval, the real time simulation relies on a real time kernel, which can be generated from commercially available tools. The early attempts to realize the real time simulation can be traced back to the seventies. However, due to a low speed of the hardware at that time, the development of the real time simulation was quite slow. Nowadays, the hardware and software is much fast and easy to use, and this has enabled a quick growth in the applications of real time simulations. Commercial tools, e.g., MATLAB Real-Time Workshop and dSPACE TargetLink, exist in the industrial market that can transfer offline models to real time simulations on dedicated hardware. In this work, the MATLAB Real-Time Workshop, which can convert the Simulink models to a real time kernel on dedicated hardware, is adopted. However, during the procedure, it requires that a fixed-step size ordinary differential equation (ODE) solver be used for RT simulation. While solving nonlinear differential equations, the step size needs to be very small (in the order of 10-3 s) at some instances for numerical stability and since it requires a fixed step size, RT simulations are generally computationally expensive. Therefore, for practical considerations, a RT simulation kernel is compiled as a dynamic link library (DLL) from the Simulink model and is run in a dedicated real-time target machine. A functional block diagram of the RT simulator is depicted in Figure 4. The specific DLL target for the RT simulator can be developed for different applications including MATLAB’s XPC, National Instrument’s LabVIEW, etc. External Controllers/Optimizers RT Target Development Simulink Model
Matlab Realtime Workshop
Model dll file
OPC
PID Controllers
Data Acquisition/Logging System User Interface
Real Time Target
RT Simulator
Figure 4. A Functional Block Diagram of Real Time Simulator
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
A fully functioning RT simulator for CL chemical looping process needs at least four function blocks for practical use. These function blocks are discussed below. •
RT Simulator Kernel
The RT simulator kernel is the fundamental block of the RT simulator usually in the form of a DLL file created from the simulator model. The real time kernel runs in a dedicated computer to ensure the simulation runs in real time. •
Data Acquisition and Logging System
Since the RT target runs in a dedicated personal computer, a data exchange system is necessary to write the inputs to the simulator and read the responses. The protocol used in this data communication may vary depending upon the application for which the RT target DLL is built. In addition, the simulation data needs to be saved for future use and hence an appropriate data logging function is also necessary. •
Graphical User’s Interface
An easy-to use graphical user’s interface (GUI) is important for monitoring ongoing simulation variables, applying manual inputs to the simulator and tuning basic controllers. The GUI needs to have a real-time data plotter, switches and buttons, data input functions, etc. for convenient operation of the RT simulator. A GUI built for the chemical looping real time simulator is shown in Figure 5.
Figure 5. A LabVIEW GUI Developed for a RT Simulator
•
OPC Connection
OLE (Object Linking and Embedding) for Process Control (OPC) is a standard data communication mechanism in industrial automation world. It is an important component desired in the real-time simulator to establish data connectivity with commercial
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
MPC controllers, process monitoring tools, etc. OPC connectivity is based on client-server configuration and since most of the external tools have OPC clients, the real time simulator needs an OPC server embedded to it. In view of previous experiences and ease of configuring OPC connections, National Instrument’s LabVIEW platform was chosen for the development of real-time simulation platform for CL process. With a rich library of toolboxes, it also supports development of data acquisition system and graphical user interface (GUI) for RT simulator development.
Figure 6. Hardware Setup for Real-Time Simulator The hardware setup for the real-time simulation of dual loop hot flow chemical looping system is shown in Figure 6. Several control schemes including PID and advanced control designs were tested on the real time simulator to evaluate the closed loop simulation performance. 4
Simulation Scenario Designs and Simulation Results
The real time simulation platform presented above has been utilized to design and test PIDs and a commercial model predictive control (MPC) for the chemical looping solids transport process. A set of PID controllers was designed to maintain the solids temperature in the system at a constant value. And the MPC control problem involved regulating the solids flow within and across the multiple loops while maintaining adequate solids inventories in the individual loops at high temperature conditions. In this control simulation test, the setpoints to the solids flow loading in the system represented by differential pressure (DP) were changed at different load levels while the solids inventory levels in individual loops were maintained at a safe range. Figure 7 shows results for a test case carried out in the real time simulator. The total simulation was carried out for 60 minutes. As seen in the figure, the controller was able to drive the system such the DP outputs followed their setpoints with satisfied dynamic responses. Moreover, it can also be seen that the solids
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
inventory levels in each loop were kept within a narrow range ensuring the stable operation, which is hard to be achieved with manually operation.
Figure 7. Chemical Looping Process Real Time Simulation Results for a MPC Test The control design tested on the real time simulator was further successfully implemented on a test facility. More tests are being carried out on the real time simulator as well as on the test facilities to study and evaluate different control designs at different operating scenarios so as to prepare for the advanced controls design and development for the auto-thermal prototype facility being built by Alstom.
5
Conclusions
The paper presented a result from Alstom’s dynamic simulation and advanced controls of chemical looping project. In this project, several dynamic simulators for calcium-based chemical looping pilot facilities were developed for different study purposes based on the first principles. Benchmark PID, extended PID, MPC and MPC/PID control strategies were designed and tested on both the dynamic simulators and facilities. The performance of the process under different control designs was investigated in this project. This research has laid the foundation of control design for a future large-scale prototype chemical looping plant (1-3MWth). A real time simulator is a key test platform to evaluate the control performance. Based on dynamic simulators developed in the MATLAB/Simulink at Alstom, real time simulators were developed and used for control testing and operation training. A control system with PIDs and an industrial MPC were designed for the chemical looping hot flow process. This control system was tested and tuned against the real time simulator for different solids flow loadings. Following the control simulation work, the same advanced control design has been then successfully implemented on Alstom’s pilot solids flow test facility. To develop a real time simulator for a large-scale prototype chemical looping plant (1-3MWth) will be a major objective in future work being planned for the Phase II project on chemical looping simulation and advanced controls. It can be envisioned in this scale-up case, the real time simulator will act as a very useful tool to help evaluate control designs and optimize the system operations of the whole process.
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
Acknowledgements This material is based upon work supported by the U.S. Department of Energy (DOE) under Award Number DE-FC26-07NT43095. Sincere thanks are hereby given to Ms. Susan Maley, who is the DOE/NETL Project Officer. Thanks are also owed to the management at DOE/NETL and Alstom. Many thanks should go to each team member of the chemical looping process development team (Herb Andrus, John Chiu, Paul Thibeault, et al.) and the chemical looping simulation and controls team (Dr. Wendell Mills and Jie Luo).
References [1] [2] [3] [4]
Andrus, H.E., Jr. et al., Alstom’s Hybrid Combustion-Gasification Chemical Looping Technology Development, Pittsburgh Coal Conference, September 12-15, 2005, Pittsburgh, Pennsylvania. Dahlquist, G. and Bjorak, A., Numerical Methods, Prentice-Hall, Englewood Cliffs, NJ, 1974. Lou, X., Neuschaefer, C. H., and Lei, H., Simulation and Advanced Controls for Alstom’s Chemical Looping Process, The 33rd International Technical Conference on Coal Utilization & Fuel Systems, Clearwater, Florida, USA, June 1 – 5, 2008. Schiesser, W., The Numerical Method of Lines: Integration of Partial Differential Equations, Academic Press Inc, 1991.
Disclaimer: This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
COMBUSTION TECHNOLOGIES EFFECT OF H2S ON CHEMICAL LOOPING COMBUSTION OF COAL-DERIVED SYNTHESIS GAS OVER Fe2O3 - MnO2 SUPPORTED ON ZrO2/SEPIOLITE/Al2O3 Ewelina Ksepko, PhD, Institute for Chemical Processing of Coal, 1 Zamkowa, 41-803 Zabrze, POLAND, E-mail: [email protected], Tel: 00 48 32 271-00-41 int. 149 Ranjani V. Siriwardane, PhD, U.S. Department of Energy, 3610 Collins Ferry Road, Morgantown, WV 26507-0880, USA, E-mail: [email protected], Tel: 001 304-285-4513 Hanjing Tian,, PhD, U.S. Department of Energy, 3610 Collins Ferry Road, Morgantown, WV 26507-0880, USA E-mail: [email protected], Tel: 001 304- 285-5424 Thomas Simonyi, U.S. Department of Energy, 3610 Collins Ferry Road, Morgantown, WV 26507-0880, USA E-mail: [email protected] James A. Poston Jr,. U.S. Department of Energy, 3610 Collins Ferry Road, Morgantown, WV 26507-088, USA, E-mail: [email protected], Tel: 001 304-285-4635 Marek Sciazko PhD, Institute for Chemical Processing of Coal, 1 Zamkowa, 41-803 Zabrze, POLAND, E-mail: [email protected], Tel: 00 48 32 271-00-41 int. 240
Abstract: The paper contains results of collaborative research work on novel combustion technology known as chemical looping combustion (CLC). The objective of paper was to prepare Fe 2O3 - MnO2 supported on ZrO2/Sepiolite (ICPC, Poland) oxygen carriers and to evaluate the performance (NETL, US DOE) of these for the CLC process with synthesis gas/air. Thermo gravimetric analysis (TGA) and low pressure (10 psi) bench scale flow reactor tests were conducted to evaluate the performance. Multi cycle tests were conducted in an atmospheric TGA with oxygen carriers utilizing simulated synthesis gas with & without H 2S. Effect of H2S impurities on both the stability and the oxygen transport capacity was evaluated. Multi cycle CLC tests were also conducted in the bench scale flow reactor at 800 °C with selected samples. Chemical phase composition was investigated by X-Ray diffraction (XRD) technique. Five Cycle TGA tests at 800 °C indicated that all oxygen carriers had a stable performance at 800 °C. It was interesting to note that there was complete reduction/oxidation of the oxygen carrier during the 5-cycle test. The fractional reduction, fractional oxidation and global reaction rates of the reactions were calculated from the data. It was found, that support had a significant effect on both fractional reduction/oxidation and the reaction rate. The oxidation reaction was significantly faster than the reduction reaction for all oxygen carriers. The reaction profile was changed by the presence of H 2S but there was no effect on the reaction rate due to
presence of H2S in syntheses gas. Low pressure bench scale flow reactor data indicated stable reactivity, full consumption of oxygen from oxygen carrier and complete combustion of H 2 and CO. XRD data of samples after multi-cycle test showed stable crystalline phases without any formation of sulfides or sulfites/sulfates and complete regeneration of the oxygen carrier after multi-cycle tests.
Effect of coal blending method on combustion characteristics and NOx emission in a drop tube furnace Song-Gon Kim, Chun-Sung Lee, Byoung-Hwa Lee School of Mechanical Engineering, Pusan National University, 30 Jangjeon-dong, Gumjeong-ku, Busan, 609-735, South Korea Chung-Hwan Jeon*, Ju-Hun Song, Young-June Chang Pusan Clean Center, Pusan National University, 30 Jangjeon-dong, Gumjeong-ku, Busan, 609-735, South Korea *
Corresponding Author: [email protected] Telephone: +82-51-510-3051 FAX: +82-51-582-9818 Presented at
The 27th Annual International Pittsburgh Coal Conference Istanbul, Turkey October 11-14, 2010 ABSTRACT This paper presents the dependence of combustion characteristics and emission on in-furnace blended method in Drop Tube Furnace(DTF). The experiments were performed with blending of sub-bituminous and bituminous coals (single coals and approximately 25%/75%, 50%/50% and 75%/25% blends) at different blending method under the condition of 1.34 excess air ratio. In the in-furnace blended method, the distance between injection positions was changed along the axial direction to investigate the interaction between the different rank coals. The results show that the unburned carbon ratio decreased as the blend ratio of Adaro(subbituminous coal) in Yakutugol(bituminous coal) increase when two coals fed into the furnace simultaneously(out-furnace method). But, at the 75% of sub-bituminous coal blending, emission of unburned carbon was higher than the amount which was emitted from Yakutugol(bituminous coal) burning solely. The NOx emission increased as blending ratio of Adaro increase. In the case of the 75% Adaro and 25% Yakutugol blending, in-furnace blending method was performed to compare with out-furnace blending method. In-furnace blended method led to a higher combustion efficiency and lower NOx emission than out-furnace blended method. In other words, it was seen that ignition position in blending influences remarkably combustion efficiency and NOx emission. The reason was that volatiles burning of Adaro coal which has
high volatiles content relatively launches earlier than Yakutugol coal in the out-furnace blending method. It led to increase of unburned carbon in bituminous coal by deficiency of oxygen to be reacted. Different ignition position between bituminous and sub-bituminous coals by in-furnace blended method improves burning rate of bituminous coal and reduce the unburned carbon from bituminous coal at specific blending ratio. Furthermore, NOx emission also decreased by using in-furnace blending method.
Keywords Coal, Drop Tube Furnace, In-furnace blending, unburned carbon, NOx emission, separate injection.
Introduction Since the demand for coal has been increased more and more worldwide, it is desired to use of low rank coal such as sub-bituminous coal which is low price and abundant reserve. Therefore, technical method of blending two different coals is considered to improve performance in power plants. When two types of coal are fired in a boiler, two different blending methods can be utilized. One is the line blending method, in which the types of coal are stored in a bunker on the mill. The other is the in-furnace blending method, in which each type of coal is fired by each burner without prior blending in the bunker[1]. In this study, where a blending ratio of bituminous coal(Yakutugol) and sub-bituminous coal(Adaro), is varied the emission of NOx and unburned carbon were investigated by DTF and TGA. Furthermore, the effect of the volatile matter in sub-bituminous coal on the emissions was clarified under the in-furnace blending condition.
Experimental The controlled conditions required for high temperature, blended coal oxidation experiments place great demand on a test facility. The particle reaction must proceed in a consistent manner and the temperature and composition of the gas surrounding a particle temperature, emission gas composition during reaction[2]. DTF was already used for coal blending combustion experiments by previous research[3,5~6]. Schematic diagram of experimental apparatus for coal blending is shown in Fig. 1. Unburned carbon ratio and NOx emission of blends of bituminous coal and sub-bituminous coal passing though a DTF were investigated by using a TGA and Gas analyzer. Drop Tube Furnace Drop Tube Furnace has a number of advantages for high temperature, pressure and broad
temperature regime coal oxidation experiments. The conditions encountered in pulverized fuel combustor can be closely approximated by this furnace, particle temperatures are relatively constant and it can be well characterized, gas composition are not limited and direct sampling in-situ measurements of single particles are readily accomplished. Fig.2 shows DTF which consists of injecting, reaction and collecting section. This furnace used SiC heater and can heat up to 1600℃. In the injecting section, two feeders were installed in injection probe for coal blending combustion. One coal was supplied into injection probe directly, and another coal was fed through a small feeder which is cooling by water. That feeder was installed in injection probe and length is adjustable axial direction of injection probe. Fig. 2 shows the schematic system of in-furnace blended method. The ignition position of coal supplied through center feeder can be adjusted by modifying the length of tube. This system was designed to realize actual multi-stage burner in DTF. In this experiment, in- furnace combustion is performed at 3, 5 and 7cm length
Fig. 1 Schematic diagram of coal blending experiment
Fig. 2 Schematic of in-furnace blending system Thermo Gravimetric Analyzer and Gas analyzer To calculate a mass of reacted carbon and the proximate analysis of coal, thermo-gravimetric analyzer (TGA) is used. A gas analyzer was used to measure NOx and O2 concentration on real time basis. Unburned carbon ratio was calculated by ash tracer method[2]. Experimental condition Blended coal oxidation was performed in a DTF reactor operating at 1300℃. Dry coal previously ground and sized to the 75 ~ 90µm range. The coal stream was entrained in a primary carrier gas(N2), passed through a water-cooled injection probe and injected into the center of a laminar secondary gas stream. O2 concentration was fixed at 12% for keeping the air ratio at 1.34. Pulverized samples, of single coals and blends coals were burned in DTF reactor. Table 1 shows the results of proximate and ultimate analyses for the single coals assayed. The blending coal feed rate and gas flow rate were 0.24 g/min and 5 l/min, respectively. Table 1 Properties of two coals used in this study
Coals a
M
Yakutugol Adaro a
b
Proximate analysis
Ultimate analysis
(wt. %, air-dry)
(wt. %, dry)
b
VM
c
FC
Ash
C
H
O
N
S
1.67
17.94
68.68
11.71
76.3
3.88
5.23
0.72
0.12
5.22
50.67
39.92
4.19
70.8
5.65
17.84
1.21
0.07
c
M: Moisture; VM: Volatile Matter; FC: Fixed Carbon
Results and Discussion Effect of blend ratio In the case of out-furnace method which feeds two different coals into the furnace simultaneously, as Fig. 3, the experiments performed in different blending ratio were resulted out the unburned carbon ratio decreased as the blend ratio of Adaro(sub-bituminous coal) increase, but the ratio increased at 75% of Adaro blending. This result reveals that as the blending ratio of high reactive coal which contains high proportion of volatile matter increase, the initial burning of high reactive coal can create an oxygen deficient environment leading to inefficient combustion of the low reactive coal.[3] It means that synergistic effects in the pyrolysis processes were absent at 75% Adaro blending. M. Ikeda also reported that unburned carbon concentration in fly ash becomes much higher as the moisture in sub-bituminous coal increase.[4] But, in this study, since the moisture contents were not different significantly comparing by proximate analysis, the moisture was not considered to dominant matter. In case of NOx emission, Fig. 4 shows the relationship between NOx concentration and blend ratio of Adaro. As Adaro blend ratio increase, NOx concentration became higher. A linear relationship from the gas analyses of NOx emission was found as increasing the Adaro blended ratio.
300
9 8
NOx concentration (ppm)
Unburned carbon fraction (%)
10
7 6 5 4 3 2 1 0
0
25
280 260 240 220 200 180
50
75
100
Blend ratio of Adaro (%)
0
25
50
75
100
Blend ratio of Adaro (%)
Fig. 3 Variation of unburned carbon with blen-
Fig. 4 Variation of NOx emission with blen-
ding ratio of the Adaro coal.
ding ratio of the Adaro coal.
Effect of in-furnace blended method When blending ratio was Adaro 75% and Yakutugol 25%, in-furnace method was experimented to compare the emissions of unburned carbon and NOx concentration with out-furnace method. Fig.5 shows that emission of unburned carbon along the different axial distance between coal inlet positions. The results show that initial burning of Yakutugol was not affected by lack of oxygen causing from volatile matter of Adaro. Therefore, unburned carbon ratio decreased by using in-furnace blended method.
Fig. 6 shows the relationship between NOx concentration and distance between feeders. NOx emission also decreased where 3, 5 and 7cm. The reason for the reduction of NOx is considered to be that NOx formed from Yakutugol is decomposed by the reduction matters (HCN, NH4) from the volatile of Adaro.
300
6
NOx concentration (ppm)
Unburned carbon fraction (%)
7
5 4 3 2 1 0
0
3
5
7
Distance between feeders in furnace (cm)
280
260
240
220
200 -1
0
1
2
3
4
5
6
7
8
Distance between feeders in furnace (cm)
Fig. 5 Relationship between unburned carbon
Fig. 6 Relationship between NOx emission
and distance between coal injections
and distance between coal injections
Conclusions The effects of in-furnace blending method and out-furnace blending method were investigated in DTF. The main results of this study are summarized as follows: 1. In case of out-furnace blending method, the unburned carbon ratio decreased slightly according as Adaro(sub-bituminous coal) proportion increased to 25%, 50% in Yakutugol (bituminous coal). But, when blending ratio of Adaro was 75%, emission of unburned carbon increased higher than unburned carbon ratio of burning Yakutugol solely. The reason is that oxygen consumption of Yakutugol was delayed by volatile matter from Adaro. 2. NOx emission was increased linearly as increasing the Adaro blended ratio when two coals burned simultaneously using out-furnace blended method. 3. When coals burned with blending ratio of Adaro 75% and Yakutugol 25%, in-furnace blended method resulted out less emission of unburned carbon and NOx compared with out-furnace blended method. The results revealed that Yakutugol did not affected by oxygen deficiency and NOx formed by Yakutugol was decomposed by HCN, NH4 from the volatile matter of Adaro.
Acknowledgement This work partially supported by project of the Pusan Clean Coal Center Research Grant in Pusan National University.
References [1] M. Ikeda, “Development of technology for reducing NOx emission and unburned carbon concentration in fly ash using in-furnace blending method,” CRIEPI Report, 2008. [2] K.T. Kang, 2010, “An Experimental Study on Char Oxidation Mode and Reaction Rate in Oxygen Enhanced Environments using DTF,” M.S. Thesis, Pusan National University. [3] S. Biswas, N. Chouhury, P. Sarkar, A. Mukherjee, S.G. Sahu, P. Boral, A. Choudhury, 2006, “Studies on the combustion behavior of blends of Indian coals by TGA and Drop Tube Furnace,” Fuel Processing Technology 87, 191~199. [4] M. Ikeda, H. Makino, H. Morinaga, K. Higashiyama, Y. Kozai, 2003, “Emission characteristic of NOx and unburned carbon in fly ash during combustion of blends of bituminous/sub-bituminous coals,” Fuel 82, 1851~1857. [5] C. Ulloa, A.G. Borrego, S. Helle, A.L. Gordon, X. Garcia, 2005, “Char characterization and DTF assays as tools to predict burnout of coal blends in power plants,” Fuel 84, 247~257. [6] W.H. Chen, J.S. Wu, 2009, “An evaluation on rice husks and pulverized coal blends using a drop tube furnace and a thermo-gravimetric analyzer for application to a blast furnace”, Energy & Fuels 34, 1458~1466.
TGA and DTF studies on coal blends to assess combustion performance P. Sarkar*, A. Mukherjee, S. G. Sahu, A. Choudhury, A. K. Adak, M. Kumar, N. Choudhury, S. Ghosal+ and S. Biswas* Central Institute of Mining and Fuel Research (Erstwhile Central Fuel Research Institute), Digwadih Campus, Dhanbad 828108, Jharkhand, India +
Jadavpur University, Kolkata-700032, West Bengal, India
Abstract: As the practice of utilization of high ash Indian coals for power generation by blending those with low ash coals are growing, uncertainties in resultant combustion characteristics associated with blending practices need to be assessed in depth. For this purpose, two typical blend combinations of Indian coals, representing combinations of similar and different coal rank, were chosen. Combustion studies with blends in Thermo Gravimetric Analyzer (TGA) and in Drop Tube Furnace (DTF) reveal both interactive and non-interactive behaviour of constituent coals. Type and level of interaction varies with constituent coal-type, blend proportion, combustionconditions and combustion-stages. Both positive and negative deviations from the expected weighted average values of combustion parameters were noticed. In top port (DTF) interactions are high (compared to next ports), carbon burnout values are scattered and inferior than parent coals. TGA showed lowering of activation energy in blends. Characteristic TGA parameters and burn out efficiencies in different ports of DTF’ reflected interesting features.
(Key words: coal blends, combustion, burn out behaviour)
*Corresponding Authors: Email: [email protected], Ph. 91-326-2388286, Fax: 91-326-2381113
1. Introduction
Coal blending for power generation has been adopted as routine practice in the pulverized coal power stations in India. Because it provides the flexibility of accommodating different coal types either for improvement of burning performance or to comply with the environmental stipulations. The blending strategies are commonly applied to reduce cost, control ash deposition, enhance fuel flexibility and extend the range of acceptable coals. It has been experienced that behavior of blends very often does not comply with the expected weighted average value of parameters from the pure coals comprising the blend. Therefore, interactions among the components of parent coals have to be accounted for observed deviations from such averaged behavior [1-3].
Indian power stations use mostly coals belonging to sub-bituminous category and by virtue of their depositional characteristics as well as origin they are mostly inertinite rich [4]. Detailed combustion studies with blends of Indian coals with typical petrographic mix are very few. One such study of this Institute revealed behavioral pattern blends of high ash Indian coals of similar rank [5]. To get insight into the nature and variation of interactive effects, which will affect the final combustion performance in utility boiler, detailed lab scale study (e.g., TGA) and bench scale study (e.g., DTF) are very much needed. As a matter of fact, TGA and DTF experimentations are often recommended to identify the acceptable type of coals and the appropriate blend compositions. Such prior assessment helps in decision making to avoid problems in the utility operations.
In this paper two binary blend combinations, one containing similar rank of coals and another one containing widely varied rank of coals, have been studied to observe the basic differences in the nature and extent of interaction occurring during burning of blends. Kinetic studies with the 2
coals and their blends were also taken into account to evaluate activation energy values which may enrich the level of understanding on the interactive effects in coal blends. Observations of TGA /DTF studies, activation energy values, etc. may be integrated to have an assessment of combustion performance prior to its possible utilization in boiler/ combustor.
2. Experimental Section
2.1. Sample preparation
A set of two coals (X and Y) of similar rank and another set of two coals (C and D) of widely different rank were selected for the preparation of binary blends. All the coals are of Indian origin. Low ash coals Y and D are respectively blended with their high ash counterparts X and C in five proportions (20%, 30%, 50%, 70% and 80%). The binary blend samples have been identified as X/Y and C/D respectively. The coals were crushed to 3mm size followed by grinding in a ball mill to produce a product with more than 80% passing through 200 mesh. The blend samples were prepared by mixing the 3mm size coals at the desired proportion and tumbled in a ball mill. The ground samples (<75 micron) were used for both TGA and DTF study.
2.2 Proximate and ultimate analysis and petrographic studies
Proximate and ultimate analysis of coals were done using standard procedures i.e., IS: 1350 Part-I: 1984, Part III: 1969, Part IV/1:1974, Part IV/2: 1975. Polarised light microscope (Leica made and DMRXP system) had been used for petrographic measurements. The mean random reflectance (Rr%) was measured under oil immersion (IS: 9127-Part 5: 1986/ ISO 7404-Part 5: 1994). And the macerals were measured (IS: 9127-Part 3: 2002/ ISO 7404-Part 3: 1994) using
3
both white light and fluorescent light irradiation in same system. (IS-Indian Standard, ISOInternational Standard Organization)
2.3 Thermal analysis, DSC/TGA/DTG Simultaneous thermal analyzer, model STA409C (NETZSCH, Germany) having DSC/TG sample carrier device was utilized to study combustion behaviour of four coals and the blends. About 20 mg of sample taken in the sample crucible was heated up to 750oC at a constant heating rate (10oC/min) under a constant air flow rate of 50ml/min through the sample chamber. TGA thermograms were analyzed to determine the relevant characteristic combustion parameters viz., initial temperature ‘Ti’ (Temperature at the initial phase of combustion where rate reaches 1.0%/ min), ‘T50’ (Temperature at the level of 50% loss of combustibles), and burn out temperature ‘Tb’ (temperature at terminal phase where rate reduces to just 1.0%/ min).
2.4 Evaluation of Kinetic parameters TGA data points between Ti and Tb in the interval of 1oC for the coals and their blends were analyzed using Netzsch Gerätebau kinetic software [3, 6] and curves were fitted with single step first order reaction model. The activation energy and pre-exponential factor for the coals and the blends were determined and correlation coefficients were noted. Earlier, set of exercises were carried out assuming nth order reaction, first order auto catalytic reaction, nth order auto catalytic reaction etc. Results and fit criteria of all such attempts suggested the choice of first order reaction model to be best for all the coals and their blends.
2.5 Drop tube furnace study The burning behaviour of the coals and their blends were studied in a Drop Tube Furnace. The details of the DTF has been given else where [5, 7]. The unburnt residues of all the samples were 4
collected at three different ports to observe the variation of the burning rate. The ash percentages of the samples were determined and the burnout efficiency (BE) was calculated using the following standard formula, which is based on the ash constancy approach (constancy of ash content in the parent coal and char).
BE= [1-Ao *(100-Ac)/(Ac*(100-Ao)] *100 ……………………….(1)
Where Ao is the dry ash % of the parent coal and Ac is the dry ash % of the char.
3.0 Results
3.1 Coal characteristics
Results of proximate and ultimate analysis of coals are summarized in Table 1 and those of petrographic measurement are given in Table 2. The reflectance values which indicate the rank and the maturity of the four coals under consideration varied from 0.42 to 1.24%. Coal C, having the highest rank, is basically a coking coal. X/Y combination is close in rank whereas C/D combination differs widely in maturity/ rank. Coals X, Y were inertinite rich, whereas the coal D was rich in vitrinite. The heat values of the raw coals varied from 3630 to 5010 kcal/ kg (on air dried basis).
3.2Combustion behaviour
Combustion parameters as obtained in TGA experiments with coals and their blends have been incorporated in Table 3. The kinetic parameters as obtained by analyzing the TGA results through Netzsch Gerätebau kinetic software based on single step first order reaction model have
5
been presented in Table 4. Estimates of burn out efficiency in top, middle and bottom ports of DTF are tabulated in Table 5.
4. Discussion
4.1 Combustion characteristics of coals
Combustion parameters of the coals largely depend on the maturity or rank characteristics. But there are several other parameters such as particle size/ particle size distribution, maceral distribution, char morphology, mineral matter, ash content, etc. affects the combustion characteristics in many ways. Within the above mentioned parameters inter dependency are also noticed very often and as such it is very difficult to resolve the all the issues at a time and finally to predict combustion performance of a particular blend combination. Indexes such as maceral index, petrofactor, etc.[4, 8], normally used to predict combustion performance, are difficult to work out in thermal power stations and that is why assessment based on those indexes are vary far from routine practice. Therefore it is often suggested that pre-assessment of the performance of different blend combinations can be done in TGA and drop tube furnace to select appropriate blend combination and blend composition and to avoid various operational problems in boilers.
The coals of Indian origin used in different power stations are low in rank, sub- bituminous in nature, and those are generally inertinite rich. The coals currently in use are high ash coals with typical distribution of macerals. The maceral distributions of the coals under consideration have been depicted in Table 2.
Three characteristic temperatures of TGA experiments were considered viz., Ti, T50 and Tb representing initial, major and terminal phase of the burning profile. The trends of all the above parameters in blends with increasing concentration of low ash coal (Y or D) have been presented 6
in Fig. 1(a) - (b). The trends show that both additivity and non-additivity criteria. Similarly Figs. 2(a)-(b) depict the observed trends in the burn out efficiency of the blends in different ports of DTF exhibiting both additivity and non-additivity criteria. Departure from the additivity characteristics may be due to typical intrinsic characteristics of organic components, environment prevailing in the combustor, competition among the constituents towards the available oxygen, etc. It is observed in general that the TGA and DTF combustion parameters are by and large additive when the blend proportion is 50:50. But when the blend proportions deviates from 50:50 composition additivity characteristics are rarely observed.
In DTF experiments it has been further observed that in top port deviations from the weighted average values for the blend samples are very much significant and results are scattered, whereas, in the following ports the results found to reflect more and more regularized pattern with respect to increasing concentration of low ash coal and the extent of deviation from weighted average line was also diminished as we move from top port to bottom port (Fig. 2).
TGA and DTF characteristics revealed in the Figures 1-2 also depict that when the rank of the coals in a binary blend are widely different (e.g., C/D blend), the deviation from weighted average line is more. For combination of coals close in rank said deviation is less.
Activation energy patterns with increasing proportion of low rank coal ( Fig.3) show lowering of activation energy, when entire combustion process (i.e., initial temperature to burn out temperature) is taken into consideration. The pattern for both the combinations(X/Y and C/ D) showed a minima in the blend samples containing around 50-60% low ash coals. Blends with more % of low ash coal showed rise in activation energy. This gives an indication that use of blends containing lower proportion of low ash coal may sometimes be more beneficial than that
7
containing higher proportion of low ash coals. Use of particular blend composition may be even more beneficial than using low ash coal alone.
5. Conclusion
TGA and DTF studies with different blend combinations may give meaningful inputs for appropriate choice of blend combination and blend composition to avoid problems in power station boiler and to get best possible performance out of coal blending. Both the addtivity and non-additivity criteria were shown by the blend samples. Departure from additivity characteristics of blends are difficult to predict. However, information on different blend categories (same rank, high rank-low rank, etc.) may be gathered for the benefit of power plant practices. Lowering of activation energy may be achieved through blending and sometimes blend may show lower activation energy than either of the component coals. Trends of combustion parameters and activation energy depict that more benefits may be obtained through blending a low ash coal with high ash coal in appropriate proportion compared to that which is obtainable through the use of low ash coal only.
6. Acknowledgement:
The authors are thankful to Director, Central Institute of Mining and Fuel Research, Dhanbad for giving permission to publish the paper.
7. References:
1. Su, S.; Pohl, J. H.; Holcombe, D.; Hart, J. A. Techniques to determine ignition, flame stability and burnout of blended coals in p.f. power station boilers. Progress in Energy and Combustion Science 2001, 27, 75-98.
8
2. Helle, S.; Gordon, A.; Alfaro, G.; García, X.; Ulloa,C. Coal blend combustion: link between unburn carbon in fly ashes and maceral composition. Fuel Processing Technology 2003, 80, 209-223. 3. Ulloa, C.; Borrego, A. G.; Helle, S.; Gordon, A.L.; García, X. Char characterization and DTF assays as tools to predict burnout of coal blends in power plants. Fuel 2005, 84, 247-257. 4. Choudhury, N.; Boral, P.; Mitra, T.; Adak, A. K.; Choudhury, A. ; Sarkar, P. Assessment of nature and distribution of inertinite in Indian coals for burning characteristics. International Journal of Coal Geology 2007, 72, 141-152. 5. Biswas, S.; Choudhury, N.; Sarkar, P.; Mukherjee, A.; Sahu, S. G.; Boral, P.; Choudhury, A. Studies on the combustion behaviour of blends of Indian coals by TGA and Drop Tube Furnace. Fuel Processing Technology 2006, 87, 191-199. 6. Sarkar, N. B.; Sarkar, P.; Choudhury, A. Effect of hydrothermal treatment of coal on oxidation susceptibility and electrical resistivity of HTT coke. Fuel Processing Technology 2005, 86, 487-497. 7. Choudhury, N.; Mukherjee, A.; Choudhury, A.; Biswas, S.; Sen, K. Influence of rank and macerals on the burnout behaviour of pulverized Indian coal in a Drop Tube Furnace. 20th Annual International Pittsburg Coal Conference, 2003 September 15-19. 8. Su, S.; Pohl, J.H.; Holcombe, J.A. A proposed maceral index to predict combustion behaviour of coal. Fuel 2001,80, 699-706
9
Table 1. Chemical Analysis and GCV of Coal Samples on air dried basis: Coal samples
M
Coal X Coal Y Coal C Coal D
5.3 7.5 0.9 8.3
Proximate analysis, wt% Ash VM FC
43.2 30.6 49.1 26.7
22.5 25.4 14.4 28
C
29.0 36.5 35.6 37
37.8 46.6 40.98 52.6
Ultimate Analysis, wt% H S
2.2 2.7 2.34 2.9
2 2.0 0.29 0.33
N
GCV k Cal/kg
0.8 1.0 0.82 1.35
3630 4450 3810 5010
Table 2: Petrographic Composition of Coals Coal Samples
Coal X Coal Y Coal C Coal D
Vitrinite
SemiVitrinite
Liptinite
Inertinite
Mineral Matter
Mean Ro%
24.8 26.0
0.1 0.1
10.6 10.4
32.2 40.4
32.3 23.1
0.44
22.7 53.5
3.7 0
2.1 6.4
10
43.1 22.9
28.4 17.2
0.42 1.24 0.57
Table 3. TGA Parameters of Coal and Blend Samples
Coal/ blend Sample
Ti
T50
BOT
521.2
Coal X Coal Y X : Y (20:80)
364 332.7 329.5
448.5 436 424.5
X : Y (30:70) X : Y (50:50) X : Y (70:30)
328.8 336 338
430 444 440
517 510.8 513.5 521.2 517.1
X : Y (80:20) Coal C Coal D C:D 20:80 C:D 30:70 C:D 50:50 C:D 70:30 C:D 80:20
339.9 418.8 324.3 333.5 349.7 370.6 395.2 404.5
436.3 516.7 448.2 460 453.6 472.4 496 508.6
515.4 595.8 544 561.9 570.7 583.7 596.1 588.7
Table 4. Kinetic parameters for combustion of coal and blend samples Coal/ blend Samples
Activation Energy
Pre expo-. nential
Corr. Coeff.
1.000 0.9997
Coal X Coal Y X : Y Blends
108.3752
96.41
5.37 4.60
X : Y (20:80)
93.51
4.48
0.9994
X : Y (30:70)
95.45
4.58
0.9995
X : Y (50:50)
93.22
4.28
0.9997
X : Y (70:30)
101.86
4.97
0.9998
X : Y (80:20)
98.67
4.77
0.9997
Coal C Coal D C : D Blends
122.97
5.5995
0.9997
81.4
3.3247
0.9998
C:D 20:80 C:D 30:70
72.21 65.15
2.5781 1.9969
0.9994 0.9986
C:D 50:50
66
1.9231
0.999
C:D 70:30
78.74
2.6829
0.9992
C:D 80:20
95.07
3.7824
0.9991
11
Table 5. Burnout Efficiency of coal and blend samples in DTF
Coal X
Coal Y
X/Y blends 20/80
30/70
50/50
70/30
80/20
Top
73.2
95.71
95.01
84.64
85.32
54.29
73.66
Middle
94.76
96.41
95.72
94.97
94.86
87.38
83.58
Bottom
96.47
98.48
96.93
98.31
96.86
98.19
91.21
Coal C
Coal D 20/80
30/70
50/50
70/30
80/20
C/D blends
Top
45.7
85.3
52.9
61.3
80.4
29.6
28.4
Middle
80.5
98.4
87.2
91.5
88.0
63.5
61.2
Bottom
90.2
99.8
97.7
97.2
95.0
88.8
84.5
12
Figure 1(a, b): Variation of TGA parameters with increasing proportion of low ash coal
13
X/Y blends (a) 110
Burnout efficiency (%)
100 90 Top port
80
Middle port 70
Bottom port
60 50 40 0
20
40
60
80
100
% Low ash coal (Y)
C/D Blends (b) 120.0
Burnout efficiency (%)
100.0 80.0 Top port 60.0
Middle port Bottom port
40.0 20.0 0.0 0
20
40
60
80
100
% Low ash coal (D)
Figure 2 (a, b): Variation of Burnout efficiency at different ports of DTF with increasing proportion of low ash coal 14
130
Activation Energy (kJ/mole)
120 110 100 X/Y blend
90
C/D blend Poly. (X/Y blend)
80
Poly. (C/D blend)
70 60 50 40 0
20
40
60
80
100
% Low ash coal (Y or D)
Figure3: Variation of Activation Energy of blends with increasing proportion of low ash coals
15
!
"
#
$ % $
! $ # % ! # ()* ' # +
" & & ' " # $ #$" # ' " &"$& # ) #" &" ! " ' & ) $ # " $ #$ (
$ !
!
" #
#
'
1 $ &
# " *$ !
.2
" & # !
#
$!
% &$ #
!
! !
! ! )
$
! '
(
& ' % ($
"&'' "&''
)
!!
" # ! #) # & # & " *$ & ) $! ! # & ! " " #$" # ' & , # # ./ $ ! #$"
! ! !
# # !
$% " & ' $% & " ( ( " #( + # " &( '
#
" #
" # " # " #
#
#$"
( !
" &"$&
#
&$ ' '
! # & # # #" !
" + ' ' %" $ )&
!!
( !$" & '
" "
&$
$ & !
#$" # # ! & " ! # " ) ! 345( & # ! 0 !" " $ ## " # !" "
0# 1 2 0 % & "
( " & " , ( *$ ! , ! " ( ! ! && " "
! !$" '
! && , $ (
#
#
" ! *$ !
" !) # " !) # " "& " && # $& #$" ! ! " &-) # ' ( ! " $& , #" & # & 19682 3.5(
#
" " !)$ ( " !)$ #$" && "
# ! " ! " 16782+
" #
! #
! ! #$ $+ "
# #$" #
& # &$! "
" !) # " "& 1 2( $" ! ' # " " !)$ " & " & # " " !)$
$ & " # # $ & # " ! ' ' # & 0 ! ' & ( $& + ! *$ " & "& -" & " & && " " ( "$& "& -" & " & + # " " !) # " "& + ) ! # " &) # " & " " ( + ! & !+ " & " " & " !)$ $) ( : ) & !)& " $ & " !) # " "& 1 2 & ' ! #$ ! $ " !) # " "& " & ( # !$" & ' " + !& $& , # " & 1 2 395( ' .4 " $ + " # # ! ' $ & .4 " $ #" & & ) " ! &" " # " # ' # $ & "$& # &( + # "& " & )$ $& ) " $ " " # ! ! & ! " " !) # " "& 1 2+ $ # &$ # # ) # " !)$ # # ! )& & "& $ & " "& " & " & # " # " & " & # "
#$ & ) & " !" ' ( & $ & $ & !$ & " & 365( ' " &+ 0 ! & + + &&$ . ! " & ( & " & " & & ! " 3;5( & " & + $" 1 < 2 # $& #" & = " 1 & #$ $ " !) # " "& 3;>?5( " & ' # " # 3;5(
' && " $ + ! # ') " ! # 2+ ) # " #
$
)
& #
(
$ " ! ! &
$#
" ' && "
"
! &( 3A5B
$# + '
$&
)
#$" " "
#$" "
# " ( $ ! +) $ ' C " + ) ) )& " &
$# + " " ! # & # $ # ! " " # " ' ( 4( C " ( ! & # 0 !( 0 0 C " !0 #' ) "
! # &
"
$" # $
$ + 4 . &. → $& $ + &. → $&. $& + 4 . &. → $&. , . + 4 . &. → , .& $& + , .& → $&. + , . $ + , .& → $& + , . $ + $&. → .$& $ + . , . → $, 6 $, 6 + , .& → $& + 9, .
$ C "
" "
1!
*$ & ) $! " *( 1@2+ # #
" + 47>4; ! " " &
< + 4;>.; ! $ $+ #$"
#& &&$
; ! #
!
' &
" &
"
( "$&
+'
!
!"
)
&"
$ # #$"
(
!
" !) 0 C
+ &
3@5(
$# '
'
$
&
+D
& " $ & &
"
&& '
#$" # 2
!$&
$
"
"
+
!
) #
"
0
'
#$" # !
"
& (
#$" " ! ! 0 " # (
#
' -
$ #
C
! #!0 #' ! ) & ! !(
*$ & ) $! "
)& & #(
#
" 0 ! " "
B
142 1.2 192 162 1;2 1D2 1@2 1?2 1A2
" "
" ! 1C $" # " "
#$" ' # ) & " & ! # ! !0 #' C " !0 #' ! #$" $& " !
" !)$ " + # " " ' " +@ 2 " + #A " !
"
! ' "
345B ' + *( 1?2
$!)
"
#) ' ; #? < #$ # *$ & ) $! # " ! " (
-
-
1C 2+ *( 1;2+ )& 4 !
"
$
" +?
" "
!
4( " ! "
'
"
$ C + *( 1D2+ < #$ #+ 677 # 4;77 E(
"
$
B; #
)& 4( *$ & ) $! " $ E+ 1E2 *( 1;2 677 6(7;0479 D77 .(@70474 ?77 6(760477 4777 4(9?0477 4;77 9(@047-4
!
" " & !# # " )& . $ # K # $ I) & %
)& .( 182 182 182 182 182 182 $ 182 L M 1E" &/% 2 M 1E" &/% 2
E +'
*( 1D2 @(@047-44 ;(4047-; 6(6047-. .(D.0477 D(7?047.
345 E +! *( 1?2 .(AA047; A(.60474 4(960477 A(D047-. .(;047-9
" & !# .@(46 .(7; 47(4D 7(D4 4(9D ;(6. ;9(.D .7;D .6@;
!
$ # !# .6(?. .(?. 6(4@ 7(D. ;(9; 4D(;? 6;(D6 ..?6 .DA?
#
&
$& J& #
(
&
%
& "+
K " & )& .(
'
# !
# " $% $ ! $%!" &' 9.(6? ;D(?A .(@A 6(.? 4.(A4 @(A@ 7(DD .(4D 7(?? 4(6@ D(66 49(@. 69(?6 49(;4 .DA6 ;6.6 97AD ;@.?
) '
# '
)
"
E +) *( 1@2 ;(.047-46 4(A047-D 4(4047-. 4(A70477 4(D.0479
#$" 0 ! #$ # " # ! $% # & , # & ) $% & ( " & ( & F -4 # & F -. ! # " & % ! G $HI$ ! $ I) & % H &F # " ( & F -. & "
"
(
"
& 0 # "
'
&
> *$ ! # &(
+
' # & # &$ ! # & *$ & #$ $ > # & ! #( L < ' +" # ! !, -! # $ #( & '" ' " &"$& " ! ! & ! # " #
! ! " " &+ " & + " !$& ! # & " &"$& " # ( C " + 4 # .+ # !$! &$ ) + ( $ # 0 " #+ ! # & $& & # & &$ ) " "% #( " " "
!
! ! !
" ! ) $ # $ + /$ & + " ! ! # ./ " " ) " && # ) #=$ ! + # 4 $ C " + !
" + ! # & ! $" ! " &"$& ( ' ! + C " " + !(
*$ ! + # ( 4 . && ' & # !" ! $
"
*$ N # L <( B $"" $& & 0 ! # # " !) # !, - *$ & *$ # " !! 1 O 2 " ! $ & ' $ . ! " ' $ 9( " #
.(
# !/ $ &
!
"
( "
0 +) + !$& C
" ! ) % ' # + " ! $ 4 . " &"$& " # B$ # &$ " ! ( M $&
! $ !
(
+) + ! $ ! # & & " &"$& " ! # ./ ) +
4
&$ 4
&
" &"$& # .+ # ./ + ! # # . + # # ./ " ! + #
$ "
$ ! .
/
" .
$
- !! $
0
!$& #
" &$ / +
-
" ! ! $ 1 4+ " ! 1
!
.2+
" &"$& " ! & # # 4
&&
!
1
4+
.2+
& 2
.
/ + # P
0
& !$& " !
+ ./
'
# %
! #
&$
/ + /
" !" & $& #
& & " !
" !
C
#
4
&$
&$ (
.
)
" # " " # C
! "
$ $(
#$" #( " < " !
#
)
#
1!
& ( ! $ "
&$ "
"
" !
#
#2
! $ " &
#
"
$& # " ! $ /
&$ P
1&
&$ + /" $& " !
#
"
"
"
&$
#2
(
" ! " "
+" & & (
# / (
$ $
" <
( #
&
./ *$ & # #
.( & '"
"
9( & '"
" " # C
"
! # &
! # &
* !
! " &"$& $ & " $& # $ #
" ! ! " ! " ( " # ! $ & 3445 34.5 9;7 E # .77E+ " & (C && " # ) & " ! $ ) #=$ ! C -4 " & " " ! " ) " # 4% " & # $ #
"
' " ! % )& ! ! " ! & 0( $# + ' " ) # ! & $# 347-4.5 $ # # ! $ 3475 & " # 4 !( # 4677 E+ ! 4(. # 7(6+ # ! $ C -4 # C -. & ! & $# 34.5 " ) ! ! #) + && " ! $ # &! ! C -. " ( # & # ! $ # & $# 347-4.5( $# + $& ) # $ % " & + ( ( & F -4+ & F -.+ K # $ I) & %(
&
' 61 2+ & " + !$! &$ 6) ' 1)2 # 1 2+ " ) # # ! " & ( $ $% " & ) 7(.( ' ! ./ &$ ! C " ""$ # $ $% " & &$ 7(?( $ 61 2+ 7(?( !$! &$ ! ?+ " !$! & / ' ! # . 7 )
$ 6+ 4)
!
./
7(.( 7(7(
" $ )
!
$
$
./ ./
!
' ' "
$% ! !
" & $ &&
#" $% " &
) " & # "+& %
C ( ./ !$! &$ ! C " " ( 61)2+ / ' ! # . !$! &$ ! C " ""$ &$ 4( # $ $% " & ) 7(6( &$ 7(A( $ 61#2+ ' ./ ) 7(D( !$! &$ ! C " ' ! # $ $% " ./ C " ""$ &$ 7(?( 0 ! $ ) # # ) " ) ! 7(. 7(D( $ 61 2+ $ $% " & ) 4(7( "
! # &
( $
!$! $ $
61"2+ !$!
! ""$ & ) ' ; ./
#
'
! ./
# # $ $ #
C
! !
$ 61 2( " & ( !
+
!$! &$ $ + " ) & ' + & F -4+ K +
!
1 2
1)2
1#2
$ &
$ "
& F -. & & ' " #
6(
./
& F -4+ & F -. # " ) $ 46 & # " ' & 7(.+ D(; ! 0 !$! )
' C " ( %&
# ! " # C " &
# K " & !& $& "& & ' $ ) ' F -.+ 49 & F -4 # 4. K " ) ( ! 0 !$! ) 7(6+ 6 ) &$ & F -4 " ./ 7(D+ .(; ) 7(? #
& (
12
! " !"
./
"
1"2
1 2
!$! &$ ! !$! &$ ' & # ! && & # K " & # 7(6 $ I) & ) " ! # !$! ! ) !$! &$ )& & & .8 ) ./ + !$! &$ ! 0
! C " & F -. #$" # # $ I) & % " ! & F -.+
#)
$ &
"
' " 61 2 # 61 2( $ ( 0 ! $ ) &$ ./ 7(D+ .(@; ) ) $ &$ 49 && 4(? )
$
) 61)2 !
" &
" ) # 61 2+ ) ' ""$ 7(9 #& ) ( && ' & & ' & )& (L : $ !" '( !$! ) ' $"" ! && & ( ! ) % 7(6 # # ! ) ( (
# # ) '
%$+ ' & F -4+ + ! && " ' #
#$" 0" $ I) & % & ( ./ 61)2+ ! 0 !$! &$ / " . ' 61)2 # 61 2+ " ! 0 !$! & F -. " ) $ &$ 46 ) 7(? # && . ) 4(7( ) 7(.+ D ) 7(6+ 9(@ 4(7( ! 0 !$! &$ ./
1)2
1 2
1"2 ! !
%&'(
! ! "! !"
$
! "" !" ! "# !" !"
1 2
1#2 $
;(
K
+
" 7(?
#
)
.
+
) $ &$ && 4(D; ) 7(.+ 6 )
.
+
.
4.
+
"
) 4(7( 7(6+ .(;
"
#
12
M
7(.+ ;(; ! 0 !$! )
'
) &$
!
#)
0
7(6+ 9(6
) $ I) & % " 7(? #
./
7(D+ 4(D
)
C
7(D+ .(9 ) $ && 4(4
"
) &$ )
A
4(7( ! 0 !$! &$ #$" # ) @8 ./ #& " & ( ! 0 !$! &$ ./ ! ! 7(A 7(6 #& " & ( ! " !$! ! ! 7(? 7(6 7(68 ) 7(? " !$! !$! ! 7(6 && ) + & !$! &$ ! C " $ ) )& " & & F -. ) ' $ " & $
;+
)
7(. " !$! ! ! 4 7(6 #$" # ) 98 ) 7(6 " !$! 0 !$! &$ / #$" # ) .(;8 ) 7(D . #& " & ( ! 0 !$! &$ #$" # ) ./ ! ! 7(? 7(6 #& " & ( $& " #$" && ' )& & ! ( + ./ % 7(6 #& ) #" & ( ! ! #$" ( ./
+ . + " " # .+ .+ C " ( $& ) # 4777 % & ! 0# ' ! ' # ) & F -.( 0! & , ' $ $ ) &$ #( $ ;1#2+ ) & F -.( ! 7(6+ # ! $ " " & #( $ ;1)2+ ' ' ! ' 7(6+ ' & & " 7(7. %! & 4% " & #( $ ;1"2+ # ) & F -.( ! 7(6+ " ) # 0 # & ) $ 7(7.A %! & 4% " & #( $ ;1 2+ ./ & F -.( ! 7(6+ ./ &$ " $ # #& & M ' ! ' # ) 7(6+ M " ! 0 !$! & & ) %" &/% 2( 0
M ' ! ' ) F -. & ( $ ;1 2+ " ) ! 7(6+ " ) -! 0 # & & " # ' ! ' # ! 0 !$! & & # ) $ 7(7?@ %! & 4% # ) & F -.( ! " ) -# 0 # ' ! ' & " ! 0 !$! ) , # ' ! ' # ) ) , ( $ ;1 2+ & F -. 0 C " ( ! , # ) $ 4A(?4 Q 4% " & 16@9A
" $# + #$" " C -.
! !
# $
" # #( # !
#$"
!
$ + $ & # -4 " + ) " C " 7(6 #& ) ) ' $ #$" ' ( M ) &$ , ( $& "+ +' $ ! & ) + ( () #$" # " !" & + / " ) . #=$ # "" # & ( "
+
! " ) " &
C %
' " ' && ) $ # ' ! " ' " ! $ & + ! $ C -4 # ! !+ # " & ( !$! &$ ! #" & ( ! ! )& " & & F -. ) &$ , &' " " "&$# ! ! & " " $ $ ) , ( ' + &$ ! ## # " !" & # ) " ) "
#
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
HIGH-PRESSURE AND HIGH-TEMPERATURE GASIFICATION OF UPGRADED BROWN COAL CHAR USING A MINI DIRECT HEATING REACTOR Mitsunori Makino, Xian Li, Ryuichi Ashida, Kouichi Miura* Department of Chemical Engineering, Kyoto University, Kyoto-daigaku Katsura, Nishikyo-ku, Kyoto 615-8510, Japan *Email: [email protected]
Abstract We have recently proposed an upgrading method of low rank coal which consists of treatment of coal in non-polar solvent, such as 1-methylnaphthalene, at temperatures below 350°C. The products obtained from the treatment are solvent-soluble fraction (extract) and insoluble fraction which we call “upgraded coal (UC)”. It was found that the gasification reactivity of the UC char was much larger than that of the raw coal char for the seven coals out of the nine coals tested, indicating that the proposed upgrading method can be one of the ways of enhancing the gasification reactivity of coal char without using catalyst. In Japan oxygen blown gasification with recycled CO2 has been proposed to facilitate the CO2 separation and hence to increase the gasification efficiency under the NEDO “Innovative Zero-emission Coal Gasification Power Generation Project”. To realize the gasification concept practically, it is essential to increase the CO2 gasification reactivity of coal chars under high CO2 pressure of ~2 MPa at T = ~ 1200°C. This requests the developments of methods to increase the gasification rate and to measure the gasification rate under such extreme conditions. In this work a mini direct heating reactor (mini-DHR) was successfully constructed to measure the CO2 gasification rate of coal chars at high temperature and high pressure. The validity and accuracy of the gasification rate measurement by the mini-DHR were well clarified, indicating that the mini-DHR can be a handy apparatus for measuring the gasification rate under extreme gasification conditions. Measurement of CO2 gasification rate of the UC char using the mini-DHR showed that the gasification rate of UC char was larger than that of the raw coal char by about 2 times even at high temperature and high pressure. This clarified that the proposed coal upgrading method is effective in increasing the CO2 gasification rate at high temperature and high pressure without using catalyst.
1
1. Introduction Enhancement of gasification reactivity of coal chars is very effective in increasing coal gasification efficiency. Gasification reactivity of coal chars is believed to be controlled mainly by catalytic effect of inherent minerals [1–4]. Then addition of catalyst has been performed to further increase the gasification reactivity of coal char. A more cost-effective method, however, has been desired to enhance the gasification rate. We have recently proposed an upgrading method of low rank coal which consists of treatment of coal in non-polar solvent, such as 1-methylnaphthalene, at temperatures below 350°C [5, 6]. The products obtained from the treatment are solvent-soluble fraction (extract) and insoluble fraction which we call “upgraded coal (UC)”. It was found that the gasification reactivity of the UC char was much larger than that of the raw coal char for all the three coals tested [7]. The CO2 gasification rate at 900°C of the upgraded coal char prepared from an Australian brown coal, Loy Yang coal, was surprisingly larger than the gasification rates of any other coal chars reported in the literature. Thus, it was found that the proposed upgrading method of low rank coal can be one of the ways of enhancing the gasification reactivity of coal char without using catalyst. In Japan oxygen blown gasification with recycled CO2 has been proposed to facilitate the CO2 separation and hence to increase the gasification efficiency under the NEDO “Innovative Zero-emission Coal Gasification Power Generation Project”. To realize the gasification concept practically, it is essential to increase the CO2 gasification reactivity of coal chars under high CO2 pressure of ~2 MPa at T = ~ 1200°C. This requests the developments of methods to increase the gasification rate and to measure the gasification rate under such extreme conditions. In this work a mini direct heating reactor was constructed to measure the gasification rate under the extreme conditions first. Then the validity of the proposed coal upgrading method for increasing the gasification rate under the extreme conditions was examined.
2. Experimental section 2.1 Preparation of upgraded coals Eight low rank coals including lignites from Thailand (MM) and Philippine (PH), brown coals from Indonesia (WR), Australia (LY), and Malaysia (BB and MB), and two sub-bituminous coals from Indonesia (AD and TH) were used in this work. A bituminous coal from Australia (GR) was also used for comparison purpose. The properties of the coals are given in Table 1. The detailed upgrading procedure has been described in a previous paper [5, 6]. The coal was treated in 1-methylnaphthalene (1-MN) at 350°C under 2.2 MPa 2
for 1 h and then separated into 1-MN-soluble fraction (extract) and insoluble fraction (residue) which we call “upgraded coal (UC)”. The elemental compositions and yields of the upgraded coal are given in Table 1. The raw coals and their UCs were carbonized for 30 min at 900°C in an inert atmosphere to prepare their chars. The char particles ranging from 75 to 150 μm in diameter were served to the gasification experiment. Table 1 Properties of raw coals used and upgraded coals (UCs), and yields of the UCs Ultimate analysis [wt%, d.a.f.] MM MM/UC LY LY/UC WA WA/UC BB BB/UC MB MB/UC PH PH/UC AD AD/UC TH TH/UC GR GR/UC
C 66.4 68.8 66.7 77.4 67.1 77.2 71.0 77.2 71.7 78.0 72.2 75.1 72.9 76.4 80.7 79.4 83.9 84.0
H 3.9 3.8 4.7 4.0 5.1 3.8 4.9 4.0 4.8 4.2 4.6 4.2 5.1 4.1 5.0 4.8 5.4 5.3
N 1.9 3.4 0.9 1.0 1.0 1.6 1.3 1.7 1.7 2.0 0.9 1.2 1.0 1.3 2.0 2.3 1.2 1.5
Atomic Ratio [-]
O (diff.) 27.8 24.0 27.7 17.6 26.9 17.4 22.8 17.2 21.9 15.8 22.3 19.5 21.0 18.2 12.3 13.6 9.5 9.2
H/C
O/C
0.71 0.67 0.85 0.62 0.91 0.59 0.83 0.61 0.80 0.65 0.76 0.67 0.84 0.65 0.74 0.73 0.77 0.76
0.31 0.26 0.31 0.17 0.30 0.17 0.24 0.17 0.23 0.15 0.23 0.19 0.22 0.18 0.11 0.13 0.08 0.08
Ash [wt%, d.b.] 25.8 38.5 1.5 2.3 1.5 2.9 4.1 6.0 4.6 5.9 11.8 22.9 1.8 3.6 8.2 9.4 9.6 22.2
Yield [kg/kg-d.a.f.] 0.624 0.497 0.537 0.634 0.636 0.474 0.547 0.580 0.406
2.2 Gasification experiment under atmospheric pressure (Thermobalance) The gasification rates of the chars were measured using the so-called TPR technique [8] where a small amount of char (less than 2 mg) placed in a platinum mesh basket was heated at a constant heating rate (heating rate: a = 2, 5, 10 K/min) up to 1200°C in an atmospheric CO2 stream flowing at 150 mL/min using a thermo balance (Shimadzu, TG50H). The weight decreasing curves obtained by the experiment was converted to the relationships of either the char conversion X vs. the reaction time t or X vs. the temperature T, where X was defined by X =1−
mass of remaining char (d.a.f.) mass of initial char (d.a.f.)
(1)
The X vs. t or X vs. T relationships obtained were utilized to estimate the gasification rate. 3
2.3 Gasification experiment under high pressure and high temperature (mini-DHR) Figure 1 shows the set up used for the gasification experiment. The detail of the custom made reactor, mini direct heating reactor (mini-DHR), is shown in Figure 2. The reactor was made of U-shaped SUS tubing of 3 mm I.D. The reactor itself was used as a heating element. An electric current of 75 – 150 A and a few volts were introduced to the reactor to heat up the reactor up to 900 to 1200°C. About 1 mg of char was placed in a platinum mesh basket of 1.0 mm I.D. and 10 mm high. The basket with the char sample and wrapped by quarts wool was placed just above a thermocouple in the reactor. The sample was heated up to a desired temperature in a He stream supplied through the dummy coil. After reaching the gasification temperature the gas stream was changed to the CO2 reserve coil through which a known amount of CO2 was supplied for within a several seconds. This scheme was used for measuring the initial gasification rate accurately. The conversion of char, X, was estimated by weighing the remaining char sample. When the gasification rates are relatively low, CO2 was supplied by just switching the valves V1 and V2 without using the CO2 reserve coil. By changing the reaction time, t, the accurate X vs. t relationship could be obtained over full range of X for those cases.
CO2 reserve coil Reactor
He
V1
V3
V4
P
Dummy coil V2 CO2 To vacuum pump
Figure 1 Experimental set up for gasification at high temperature and high pressure.
4
Char in mesh basket
4 - 3 mm SUS tube
Quartz wool
55 mm
Thermocouples
Gas flow
35 mm
Figure 2 Detail of the custom made mini direct heating reactor (mini-DHR). 3. Results and discussion 3.1 Gasification reactivity of upgraded coals under atmospheric pressure Figure 3 compares the 1-X vs. T relationships obtained at a = 10 K/min for the chars of raw coals and upgraded coals by the TPR gasification technique. 1-X of the UC char decreased at lower temperature than that of the raw coal char for the seven coals out of the nine coals tested, while the 1-X vs. T relationships of the UC char and the raw coal char were almost identical for the other two coals, WA and PH. This result suggests that the gasification reactivity of the chars can be enhanced by upgrading the coals using the proposed method for most coals. The 1-X vs. t or 1-X vs. T relationships measured at three different heating rates were analyzed to quantitatively estimate the gasification rates. The gasification rate (dX/dt)/(1-X) can be written as
dX / dt = k0 ( pCO2 )e − E / RT f ( X ) 1− X
(2)
where k0 ( pCO2 ) may be the function of solely the pressure of CO2, pCO2, f ( X ) is the
function of solely X, and E is the activation energy. Since pCO2 (= 101.3 kPa) is kept to be constant for all the gasification experiments, the activation energy can be obtained when we make the Arrhenius plot of (dX/dt)/(1-X) values at a specified conversion.
5
1.0
1.0
0.8
1.0 BB
0.8
0.4
1-X [-]
MM/UC
0.6 BB/UC
0.4
0.2
0.0 1.0
0.0 1.0
0.0 1.0 MB
LY
0.8 1-X [-]
0.4
LY/UC
0.6 MB/UC
0.4
0.4
0.2
0.2
0.0
0.0
0.0
1.0
1.0 0.8
0.6 WA/UC 0.4 0.2 0.0 700
0.8
PH/UC
0.6 0.4
GR
PH 1-X [-]
0.8
PH
0.2 800
900
1000
Temparature [°C]
1100
1200
0.0 700
TH/UC
1.0
WA
WA
TH
0.6
0.2
1-X [-]
1-X [-]
0.6
TH
0.8
MB
LY
1-X [-]
AD/UC 0.4
0.2
0.8
AD
0.6
0.2
1-X [-]
1-X [-]
MM 0.6
AD
0.8
BB 1-X [-]
MM
GR
0.6 GR/UC
0.4 0.2
800
900
1000
Temparature [°C]
1100
1200
0.0 700
800
900
1000
1100
1200
Temparature [°C]
Figure 3 1-X vs. T relationships obtained at a = 10 K/min under 0.1 MPa of CO2 for the chars of raw coals and upgraded coals. Figure 4 typically shows the 1-X vs. T relationships obtained at three different heating for the chars of LY and LY/UC. It is seen that the 1-X vs. T relationship shifts to higher temperatures with increasing heating rate. These data can be utilized to estimate the gasification rate quantitatively. The gasification rates (dX/dt)/(1-X) of the chars at X = 0.5
were estimated at three heating rates and their Arrhenius plots were performed as shown in Figure 5. Comparison of the gasification rates among the chars and the determination of activation energy can be made by using Figure 5. The gasification rates at 900°C were used as a measure of comparison in this work. The gasification rate of LY/UC was surprisingly 2.4 times larger than that of LY. The gasification rates of MM/UC and TH/UC were 1.7 times larger than those of MM and TH, respectively. The gasification rate of AD/UC was 1.5 times larger than that of AD. The gasification rate of GR/UC was 1.9 times larger than that of GR. Thus, we have succeeded in enhancing the CO2 gasification reactivity of the
chars prepared from both low rank coals and high rank coal by the proposed method. The activation energies estimated are listed in Table 2. The activation energies of the raw coal char and the UC char were very close for LY, MB, PH, and GR. On the other hand, the activation energies of the char prepared from the UC was larger than that of the char prepared from the raw coal for MM, BB, AD, and TH. 6
1.0
LY a = 2 K/min LY a = 5 K/min
1-X [-]
0.8 0.6
LY/UC a = 2 K/min
LY a = 10 K/min
LY/UC a = 5 K/min
0.4
LY/UC a = 10 K/min
0.2 CO2
0.0 700
800
900
1000
Temparature [°C]
Figure 4 1-X vs. T relationships obtained at three different heating under 0.1 MPa of CO2 for the chars of LY and LY/UC. 1000°C
900°C
8 6 4
-1
(dX/dt)/(1-X)|X = 0.5 [s ]
0.01
MM LY WA BB MB
MM/UC LY/UC WA/UC BB/UC MB/UC
PH AD TH GR
PH/UC AD/UC TH/UC GR/UC
2
0.001
8 6 4
2
0.0001 7.5
1000°C
8.5 -1 1/T [K ] 900°C
9.0
-4
9.5x10
8 6 4
-1
(dX/dt)/(1-X)|X = 0.5 [s ]
0.01
8.0
2
0.001
8 6 4
2
0.0001 7.5
8.0
8.5 -1 1/T [K ]
9.0
-4
9.5x10
Figure 5 Arrhenius plots of gasification rates at pCO2 = 0.1 MPa (at X = 0.5) of the chars prepared from upgraded coals and raw coals. 7
Table 2 Activation energies of CO2 gasification rates at pCO2 = 0.1 MPa (at X = 0.5) of raw coal chars and upgraded coal chars Char
E [kJ/mol]
Char
E [kJ/mol]
Char
E [kJ/mol]
MM
239.9
BB
239.5
AD
216.3
MM/UC
281.4
BB/UC
269.5
AD/UC
277.5
LY
229.5
MB
272.7
TH
255.0
LY/UC
225.0
MB/UC
285.6
TH/UC
294.1
WA
241.7
PH
273.2
GR
216.9
WA/UC
219.7
PH/UC
262.1
GR/UC
231.7
3.2 Gasification reactivity of upgraded coals under high pressure and high temperature Figure 6 shows the typical 1-X vs. t relationships obtained at 900°C under 0.1 MPa and 2.0 MPa of CO2 pressure (pCO2) for the LY and LY/UC chars. The broken lines are the
relationships obtained using a highly sensitive thermobalance (Shimadzu, TG-50H) at pCO2 = 0.1 MPa. Good agreements between the mini-DHR and TG experiments show the validity and accuracy of the gasification rate measurement using the mini-DHR. The gasification rates of the LY/UC char were higher than those of the LY char by about 2 times at both at either pCO2 = 0.1 or 2.0 MPa. This was the case within the experimental conditions tested in this work (pCO2: – 2.0 MPa; T: 900 – 1200°C). These results show that the upgrading method proposed by the authors is valid to increase the gasification rate of coal char without using catalyst even at high temperature and high CO2 pressure.
1.0 LY; 0.1 MPa LY; 2.0 MPa LY/UC; 0.1 MPa LY/UC; 2.0 MPa
900 °C
1-X [-]
0.8
0.6 TG Calculated by RPM
0.4
0.2
0.0 0
100
200
300
400
500
600
700
Time, t [ s ]
Figure 6 1-X vs. t relationships obtained at 900°C under 0.1 MPa and 2.0 MPa of CO2 using the mini-DHR. 8
Figure 7 shows the pCO2 dependency of the initial gasification rates, -rC0=dX/dt|X=0, at 900 and 1000°C. This figure shows that -rC0 is proportional to 0.32 to 0.36 powers of pCO2 for both the LY and LY/UC chars, indicating that the -rC0 value increases by 2.8 times when pCO2 is raised from 0.1 MP to 2.0 MPa for both chars. Figure 8 shows the Arrhenius plots of -rC0 values at pCO2 = 0.1 and 2.0 MPa for the LY and LY/UC chars. It is confirmed that the gasification rates measured below 1100°C were the values under the chemical reaction controlling. The pore diffusion effect could not be neglected at 1200°C. The Arrhenius plots were almost parallel each other for all the four sets of experiments below 1100°C. The activation energies estimated were 215 to 233 kJ/mol and were so close to each other. This figure shows that the -rC0 value of LY/UC char
is larger than that of LY char by 2.2 times at pCO2 = 0.1 MPa and by 1.8 times at pCO2 = 2.0 MPa. Figure 8 also shows that the gasification of the LY/UC char is completed in 4 s at T = 1200°C and pCO2 = 2.0 MPa. 4
[s ]
LY/UC; 1000°C -1
2
-rC0 = dX/dt |X = 0
0.01
LY; 1000°C
8 6 4
LY/UC; 900°C
2
0.001
8 6
LY; 900°C
4 5 67
2
3
4 5 6 7
0.1
2
3
1 PCO2 [MPa]
Figure 7 pCO2 dependency of the initial gasification rates, –rC0 =dX/dt|X=0, at 900 and 1000 °C. T [°C]
1200
1100
1000
900
LY/UC; 2.0 MPa LY/UC; 0.1 MPa
-1
dX/dt |X = 0 [ s ]
0.1
0.01
LY; 2.0 MPa
0.001
0.0001 6.5
LY; 0.1 MPa
7.0
7.5 1/T
8.0
8.5
9.0x10
-4
-1
[K ]
Figure 8 Arrhenius plots of –rC0 values at pCO2 = 0.1 and 2.0 MPa for the LY and LY/UC chars. 9
Next, to examine the effect of pore structure on the gasification rate, the random pore model (RPM) presented by Bhatia and Perlmutter [9] was applied to analyze the 1-X vs. t relationship under chemical reaction controlling regime. The RPM theoretically formulates the X vs. t relationship as
X = 1 - exp[− rC0t{1 + ψ ( − rC0 )t / 4}]
(3)
where ψ is a structural parameter reflecting pore structure of char. All the X vs. t relationships obtained under chemical reaction controlling regime were well correlated by setting ψ = 3 irrespective of T and pCO2 values for both the LY and LY/UC chars. The typical calculated X vs. t relationships are shown by the solid lines in Figure 6. This examination shows that the pore structures of the LY and LY/UC chars are very similar.
4. Conclusions
We have proposed an effective method to enhance the gasification rate of char prepared from low rank coal without using catalyst. The method we proposed is to pre-treat low rank coal in non-polar solvent, such as 1-methylnaphthalene, at temperatures below 350°C. The products obtained from the pre-treatment are solvent-soluble fraction (extract) and insoluble fraction which we call “upgraded coal”. It was shown that the gasification reactivity of the char prepared from the upgraded coal was much higher than that of the char prepared from the raw coal for the seven coals out of the nine coals tested. It was thus clarified that the proposed solvent treatment of low rank coal is an effective method to both recover value added carbon materials and to enhance the gasification reactivity of coal char without using any catalysts. A mini direct heating reactor (mini-DHR) was successfully constructed to measure the CO2 gasification rate of coal chars at high temperature and high pressure. The validity and accuracy of the gasification rate measurement by the mini-DHR were well clarified, indicating that the mini-DHR can be a handy apparatus for the gasification measurement under extreme gasification conditions. The gasification rate of the upgraded coal (LY/UC) char was found to be larger than that of LY char by about 2 times even at high temperature and high pressure. The gasification enhancement of the LY/UC char was judged to be realized by the increase of the number of “active sites” since pore structure, activation energy, and the pCO2 dependency of gasification rate were same as that of LY char. This again clarified that the proposed coal upgrading method is effective to increase the gasification rate without using catalyst. 10
Acknowledgment
This work was commissioned by the Central Research Institute of Electric Power Industry (CRIEPI) under the NEDO “Innovative Zero-emission Coal Gasification Power Generation Project”.
References
[1] Walker, P.L., Rusinko, F. and Austin, L.G. in ‘Gas Reactions of Carbon’, Advances in Catalysis, Vol.11, Academic Press, New York, 1959, p.133. [2] Essenhigh, R.H. in ‘Chemistry of Coal Utilization’ 2nd Supplementary Volume, John Wiley & Sons, 1981, p.1153. [3] Van Heek, K. H. and Muehlen, H.J. Fuel 1985, 64, 1405–1414. [4] Miura K, Hashimoto K. and Silveston P.L. Fuel 1989, 68, 1461–1475. [5] Miura K, Ashida R., Umemoto S., Sakajo A., Saito K. and Kato K. Proceedings - Annual International Pittsburgh Coal Conference (2008), 25th 464/1–464/12. [6] Ashida R., Umemoto S., Hasegawa U., Miura K, Kato K., Saito K. and Nomura S. Proceedings - Annual International Pittsburgh Coal Conference (2009), 26th 51/1–51/12. [7] Miura K, Ashida R., Makino M. and Nishida A. Proceedings - Annual International Pittsburgh Coal Conference (2009), 26th 431/1–431/10. [8] Miura K and Silveston P.L. Energy Fuels 1989, 3, 243–249. [9] Bhatia S.K and Perlmutter D.D. AIChEJ 1980, 26, 379–386.
11
CFD Simulation of Process-driven Particle Fragmentation in a Coal Bed Gasifier Franz Holzleithnera, Roland Eisla, Markus Haidera, Georg Aichingerb a
Institute for Energy Systems and Thermodynamics, Vienna University of Technology, Getreidemarkt 9 / E302, A-1060 Vienna, Austria. b
Siemens VAI Metals Technologies GmbH & Co, Turmstrasse 44, Ironmaking Technology, A-4031 Linz, Austria
Abstract: Good gas-solid contact is essential in a coal bed gasifier such as in a COREX® melter gasifier. The charged particle size distribution and the particle fragmentation behavior inside the slowly moving fixed bed strongly influence the local counter current gas flow and therefore also the rate of heat transfer between gas and coal, the rate of drying, devolatilisation and gasification. COREX® is the first commercially operating smelting reduction process, based on coal instead of coke, as alternative for industrial ironmaking route via the blast furnace. Besides coal the melter gasifier also contains reduced iron ore and additives, which are not subject of this paper. General multiphase models in commercial CFD codes are not directly applicable to the simulation of moving reactive beds considering changes in particle size distribution. A customized approach based on a combination of Eulerian and Lagrangian formulation is used to describe the flow of gas and solids as well as the physical and chemical processes across the moving bed reactor. The solids flow and the gas flow are represented by a set of Eulerian equations. So the flow of solids respectively the flow of a granular material is treated as a continuum with appropriate material properties. The balances for solids and gas flow are interconnected via source terms. The energy balance of the solids flow and the models for fragmentation and devolatilization are implemented by means of a Lagrangian formulation. The solids flow consists of a set of particle sizes including dust. This set of particle sizes changes according to local process parameters which are: solids pressure, shear stress, rate of water evaporation (coal drying), rate of devolatilization, rate of gasification and rate of temperature change. To take this into account a fragmentation model has been developed which solves a conservation equation for each particle size. The local source terms within these equations are connected to the above mentioned local process parameters. Due to the fact that the considered moving bed consists of nonuniformly sized particles the temperature of small particles will be different from the temperature of larger particles. However, the temperature of the particles is important for the rate of drying, devolatilization and gasification. Therefore an energy balance for each particle size is implemented within the presented model. In a first step the model has been used to study the impact of a changing particle size distribution on the gas flow and heat transfer between gas und solids. The effect of fragmentation on the devolatilization process has been simulated too. Next development steps are the integration of models for coal drying and gasification as well as a gas phase reaction model.
1/13
1. Introduction The main components of a typical COREX®-plant for the production of liquid iron are shown in figure 1. The two main components of the COREX® process are the reduction shaft and the melter-gasifier. In the reduction shaft iron ore is reduced up to a reduction ratio of 80 to 95% [Carpenter, 2004] by the reduction gas which consists mainly of CO and H2. The reduced iron ore coming from the bottom of the reduction shaft as well as coal and additives are charged at the top of the melter gasifier. The melter gasifier itself could be divided into four sections: the dome at the top, the slowly moving fixed bed, the hearth, and the sump at the bottom. In the dome the remaining hydrocarbons evolving from the packed bed are cracked to ensure that the reduction gas consists mainly of gases which are usable for iron ore reduction. The fixed bed consists of a mixture of pre-reduced iron ore, coal and additives. In this section the remaining iron ore is reduced completely and the drying and devolatilization processes of the coal particles take place. Due to the heating, drying and devolatilization processes, shear stress in the bed and solids pressure the coal particle size distribution is subject to change. Pure oxygen is injected at the bottom of the fixed bed in order to gasify the remaining char, melt the iron and dissolve the additives. A sufficiently coarse particle size distribution of the char matrix is important to allow counter-current flow of gas and liquids. This is also valid for the hearth underneath to ensure smooth tapping of liquid iron and slag at periodic intervals. The sump at the bottom serves as equilibration basin for liquid iron. The purpose of this work is to present a possible strategy for modeling heating-, pyrolysis- and fragmentation processes of the coal in slowly moving fixed beds similar to that of the COREX® meltergasifier. Important to mention is that only the material coal has been considered in this work. The behavior of other materials like iron, iron ore and additives as well as the interaction of these materials with the coal are not subject of this work.
Figure 1: Simplified flow sheet of a typical COREX® plant for iron making (Siemens VAI).
2/13
2. Description of the Model Multiphase models in commercially available CFD Codes are not directly capable of simulating a reacting slowly moving fixed bed with changes in particle size distribution. A lot of work has been done on the simulation of slowly moving fixed beds in blast furnaces which are to some extent similar to the slowly moving fixed bed of a COREX-Reactor. In most of the work concerning solids flow in blast furnaces three (or more) coupled Eulerian sets of equations are used for determining the flow characteristics of the solids (and fines), liquid and gas phase flow. The particle size distribution is thereby considered in the form of bed voidage and a constant mean particle diameter (e.g.: Sauter´s diameter). This is used as basis for the calculation of the relevant fluid-mechanical and chemical processes occurring in the slowly moving bed (for example [2] or [11]). The modeling approach adopted in this work is a customized combination of a Eulerian and a Lagrangian formulation. The gas phase flow is represented by using mass, species, momentum and energy balances in Eulerian notation. The solid phase (coal in this work) is depicted by means of a momentum balance in the Eulerian frame of reference. The particle specific information like mass of the particles, particle size distribution, temperature of the different particle sizes (particle energy balance), devolatilization and material properties of the particles are calculated on the streamlines of the solids flow field. The interaction between gas and solids is represented via source terms in the relevant balances. The commercial CFD-software ANSYS FLUENT 12.0.16 has been used to implement the necessary extensions to the code via user-defined functions (UDF). In the following section the considered balances and implemented models are briefly discussed. 2.1
Gas phase
2.1.1
Governing Equations:
The governing fluid dynamic equations for the gas phase are the well known equations for describing a flow of a reacting multi component mixture with diffusion in a Eulerian frame of reference: •
•
•
Overall Mass Balance:
∂ ε g ρ g ) + ∇ ⋅ ( ε g ρ g wg ) = S m ( dt
(1)
∂ ε g ρ g Yi ) + ∇ ⋅ ( ε g ρ g wg Yi ) = −∇ ⋅ Ji + Si ( dt
(2)
∂ ε g ρ g wg ) + ∇ ⋅ ( ε g ρ g wg wg ) = −∇pg + ∇τ + ε g ρ g g − Fg , s ( dt
(3)
Species transport:
Momentum balance:
The interphase momentum exchange term F is evaluated by using the classical equation of Ergun [4]:
3/13
ε s2µg ε s ρg Fg , s = 150 w − w + 1.75 wg − ws ( wg − ws ) ( ) g s 2 d εg d
(4)
Thereby d represents the so called “Sauter´s diameter“. The gas phase energy balance in Eulerian notation is given by:
∂ ε g ρ g hg ) + ∇ ⋅ ( ε g ρ g wg hg ) = ∇ ⋅ λ∇T − ∑ hi J i + Sh . ( dt i
(5)
The source term S h in Eq. (5) consists of the convective heat transfer between solids and gas phase as well as the Heats of Reactions of the devolatilzation process. The gas phase species properties like viscosity, specific heat and thermal conductivity have been included as polynomial functions of temperature and composition according to [6]. For the determination of the bulk porosity of the coal bed the model of Ouchiyama and Tanaka [12] has been implemented.
2.2
Solid phase
As stated above the momentum balance is solved in an Eulerian reference frame whereas the particle specific information is calculated on the streamlines of the solids flow. In the following paragraphs these models are briefly presented. 2.2.1
Momentum Balance:
The momentum balance for the solids flow is given by Eq. (6)
∂ ( ε s ρ s ws ) + ∇ ⋅ ( ε s ρ s ws ws ) = −∇ps + ∇τ s + ε s ρ s g + Fg ,s . dt
(6)
In the present work the emphasis lies on the influence of fragmentation on the inter phase heat transfer and the pyrolysis process. Thus the shear stresses are modeled by means of a Newtonian fluid with high viscosity. If a more proper description of the solids flow is desired, a Bingham medium or the kinetic theory of granular flow may be considered. 2.2.2 2.2.2.1
Particle Specific Information: Heat Balance of a single particle:
The considered heat balance of a single Particle is given by Eq. (7):
dH p dt
= α corr Ap (Tg − Tp ) − ∆H Pyrolysis .
(7)
The solid phase energy balance is solved for a representative particle of each particle size class. The cumulative change in enthalpy of all particles in all size classes is determined by summing up the
4/13
individual enthalpy changes of each particle size class. The resulting value represents the source term in the energy balance of the gas phase flow. Two influences have an important effect on the particle temperature. Because of the heat conduction inside the particle, especially the larger particles will have high Biot numbers. Thus the temperature distribution inside the particle needs to be taken into account. The second aspect is that heat transfer coefficients in packed beds are observed ([1], [13]) to be much lower than the heat transfer coefficient predicted by correlations like the correlation of Gnielinski [5]. Therefore the heat transfer coefficient these effects.
α calculated by the correlation of Gnielinski [5] is corrected for
The “corrected” heat transfer coefficient α corr is evaluated by means of Eq. (8) [15]
α corr = t f
1 . 1 1 dp + α κ 2λs
(8)
The factor t f takes into account that the area available for heat transfer in the packed bed is smaller than the sum of the surface area of the single particles. The term
1 dp is used to estimate the influence of κ 2λs
heat conduction inside the particle. So the particle temperatures calculated by this model are mean particle temperatures. In the current study values of t f = 0.1 [2] and 2.2.2.2
κ = 3 [15] are used.
Pyrolysis Model:
The pyrolysis model is based on the model of [8] and [9]. This model is capable of predicting the composition of Volatile matter, kinetics of evolution and heats of reaction of the coal pyrolysis process. The original model [8] deploys a parallel first order reaction model with distributed activation energies (DAE) for determining the evolution of the different volatile species. Due to the fact that the numerical integration required by the DAE approach is computationally very expensive since frequent numerical integration is necessary for all size classes in every time step over the whole range of the activation energy distribution, a simplified model has been realized. The evolution of the different volatile matter species is described by a set of parallel first order reactions given by Eq. (9):
dmi E = k0 exp − i dt RT
mi .
(9)
The considered species of volatile matter i are CH4, CO, CO2, Tar, H2, H2O, NH3 and H2S. Only the primary coal pyrolysis has been covered so the tar cracking reactions in the gas phase have not yet been included. 2.2.2.3
Fragmentation model:
The fragmentation of coal particles in slowly moving fixed beds is a governing factor for the gas flow and subsequently for the heat transfer and for the chemical processes like drying, devolatilization and gasification. The model is based on previous models for pulverized coal combustion of [14] and [10]. The
5/13
six major factors which are supposed to be the driving forces for the fragmentation of coal particles inside the slowly moving fixed bed of a COREX® melter gasifier are: • • • • • •
solids pressure (k=0), shear stress (k=1), rate of heating (k=2), rate of drying (k=3), rate of devolatilization (k=4) and rate of gasification (k=5).
The fragmentation model is formulated in terms of the number of particles in a certain size class following a distinct stream line. 3 dN 0 −C C + d1 k ∑ b0,1Sk ,1 K k 1 dt 0 d 0 k dN 1 k dt 0 −C1 − ∑ b1,1 Sk ,1 K k = k ... 0 ... ... ... ... dN n 0 dt 0
k b S K ∑k 0,n k ,n k N 0 3 d N ... ... n ∑ b1,kn Sk ,n K k 1 ... . d1 k ... ... ... ... ... ... ... Nn k 0 0 −Cn − ∑ bn ,n Sk ,n K k k
d ... ... n d0
3
(10)
The rate of change of the particle size i due to fragmentation is generally assumed to be proportional to the number of particles in each size class at the time t on a certain position on a distinct stream line. The numerical distribution of a particle size class i into smaller particle size classes j due to the process k is k
given by the elements of the fragmentation matrix bi , j . The term S k , j is the fragmentation rate constant for each process k and each size class i. The variable K k contains the driving force of the process k . For example K k for the process “heating rate” is the rate of temperature change of the particle in K/s. The variable C j accounts for the reduction of the particle size due to gasification. The fragmentation matrices
bik, j and the fragmentation rate constants Sk , j are determined by suitable experiments. 2.2.2.4 •
Material Properties of the coal particles: Specific Heat of coal:
Because of the heating, drying, devolatilization and gasification processes the temperature and the composition of the coal particles are changing during the residence time in the reactor. As a consequence the specific heat of the coal particles is not constant during these processes. To take these effects into account the specific heat model of [9] has been implemented. The reason for the selection of this model is that the specific heat could be calculated on the basis of the actual composition and temperature of the particle. The specific heat model is assumed to be valid for the whole temperature range considered.
6/13
•
Thermal conductivity:
Both, the changes in temperature and composition of the particle have a significant effect on the thermal conductivity of the coal particles. Therefore the model of [2] has been chosen to determine the thermal conductivity of coal, char and intermediate products. The thermal conductivity is estimated in terms of the “true density” of the particle and the particle temperature. The term “true denstity” in this case means the density of the daf-coal excluding the pore volumes.
3. Simulation In order to test the models presented above four steady state simulations of a slowly moving coal bed reactor have been performed. In a first step the influence of fragmentation on gas flow and heat transfer between coal and gas has been studied. In a second step the effects of fragmentation on the devolatilization process has been simulated.
3.1
Geometry
The geometry which has been used for conducting the simulations is depicted in Figure 1. The cylinder has a radius of R=4,5m and a height of H=9m. This geometry is to some extent similar to the fixed bed region of a COREX® melter gasifier.
Figure 1: Geometry of the reactor.
3.2
Boundary Conditions
The coal (daf) mass flow entering the cylinder on the top is set to 50.5 kg/s with an inlet temperature of 300 K. The analysis of the coal used for both simulations is given in Table 1.
C 89.4
H 4.9
O 3.7
N 1.2
S 0.8
Volatile Matter 25
Table 1: Analysis of the coal (wt%, daf)
7/13
The particle size distribution of the coal at the inlet is summarized in Table 2. dp [mm] X [%]
0.0001 0.1
0.25 0.5
0.5 1
5 10
8 20
16 30
20 15
25 11.5
31.5 7
50 4.9
Table 2: Particle size distribution at the top of the reactor (coal inlet).
For the simulation runs with coal devolatilization the kinetic parameters specified in Table 3 have been set.
Species
CH4
C2H6
CO
CO2
Tar
H2
H2O
NH3
H2S
Ea [MJ/kmol]
280.5
238.3
267.3
253.7
230.3
332.6
230.3
279.1
286.3
k0 [1/s]
1.x1013
1.x1013
1.x1013
1.x1013
1.x1013
1.x1013
1.x1013
1.x1013
1.x1013
Table 3: Kinetic Parameters of volatile species release.
It should be noted that for the purpose of demonstration the frequency factors are assumed to be equal and constant for all species of volatile matter. The Activation Energies of the different volatile species were set to the mean values of the activation energy distribution used by [8]. Of course for the application of the model for design or evaluation purposes the parameter setting must be further detailed. For the sake of presenting the basic characteristics of the model the current approach is taken to be sufficient. The inlet mass flow rate of gas (100% CO) at the bottom (gas inlet) of the cylinder is assumed to be 65 kg/s. The temperature of the gas at the gas inlet is set to 1500 K.
4. Results 4.1
Heat Transfer without Pyrolysis
The first two simulation runs have been conducted without considering the pyrolysis process in order to study the influence of the fragmentation process on the temperature distribution inside the reactor. The distribution of the bed voidage as fraction of volume gas per m³ reactor volume is depicted in Figure Figure 2. Due to the increasing solids pressure and the heating rate the coal particles are subject to fragmentation on the way through the reactor. The bed voidage decreases from a value of 0.407 to a value of 0.38. The reason for the decrease in voidage is that larger particles are breaking into smaller ones, which occupy more void space. As a consequence the bulk density of the coal bed increases. The evolution of the particle size distribution in Figure 3 clearly shows the increase of the volume fraction of the smaller particle size classes. The volume fraction of the larger particles with a diameter of 20 mm or bigger decreases at a height of approximately 8 m while the smaller particle size classes with a mean diameter of or below 8mm are increasing. The particle size class with a mean diameter of 16 mm is gaining volume fraction up to a height of approximately 3 m due to the fragmentation of the upper size classes. Below a height of 3 m the fragmentation to smaller size classes gets dominant over the incoming particles of the upper size classes and the volume fraction of the size class 16 mm is decreasing.
8/13
Figure 2: Distribution of bed voidage [-] inside the reactor.
Figure 3: Particle size distribution inside the reactor
The temperature distribution of the gas and the bulk phase are shown in Figure 4. The gas temperature with considered particle fragmentation is well below the gas temperature without particle fragmentation. The bulk temperature of the coal in case of particle fragmentation reaches slightly higher values at the gas inlet. This is a consequence of the increase in volume fraction of the smaller size classes. Hence, the available surface area for heat transfer is increasing too. So the heat transfer between solids and gas is enhanced.
9/13
Figure 4: Comparison of Gas and Bulk Temperature distributions with and without considered Particle Fragmentation
4.2
Heat Transfer with Pyrolysis and Fragmentation
The third and fourth simulation case illustrate the influence of particle fragmentation on the pyrolysis behavior of the coal particles. The bed voidage at the bottom of the reactor in case of fragmentation is approximately 0.395. Compared to the simulation without pyrolysis this value is slightly higher. The explanation for this effect is that as a consequence of the devolatilization process the gas mass flow is increasing from the bottom to the top of the reactor. As a result the lifting force acting on the slowly moving coal bed is increased. Thus the solids pressure and subsequently the amount of fragmenting particles are decreasing. The gas and bulk temperatures with and without fragmentation behave very similar compared to the simulation runs without pyrolysis. The effect of bulk temperature increase at the bottom of the reactor in case of particle fragmentation is somewhat smaller compared to the previous run due to the higher bed voidage at the bottom. The distributions of the mean volatile fraction of the coal bed for the cases with and without particle fragmentation are shown in Figure 5. Due to the higher bulk temperature of the slowly moving coal bed in case of particle fragmentation the pyrolysis process is slightly further progressed. Despite the fact that in the presented simulations the decrease in solids porosity from the top of the reactor to the bottom is only 2.9 % the difference in the mean volatile fraction at the bottom of the reactor is as much as 43 %. Hence, for accurate simulation results the particle fragmentation has to be considered in the simulations.
10/13
Figure 5: Comparison of mean volatile fractions with and without particle fragmentation.
5. Conclusion The particle size distribution is a governing factor for heat transfer and the chemical processes inside slowly moving bed gasifiers. A model has been presented which addresses the process of particle fragmentation by deploying a kinetic approach for the relation between the major forces acting on the coal particle and the change in particle size distribution. The influence of fragmentation on heat transfer and the devolatilization process has been determined. The results of these simulations clearly indicate that neglecting the fragmentation of the coal particles in gasifiers leads to inaccurate results especially in cases significant changes in the particle size distribution. The next development steps will include implementation of models for coal drying and gasification with respect to particle fragmentation as well as reaction mechanisms for the homogenous gas phase reactions of the different species. Notation:
Ap
Particle surface area, [m2]
bk
Progeny matrix for process k, [-]
Cj
Gasification rate of particle size j [1/s]
dp
Particle diameter [m]
Ei Fg , s
Activation energy [J/mol]
g
h Hp
∆H Pyrolysis Ji
Interphase momentum exchange [N/m3] Gravity [m/s2] Specific enthalpy [J/kg] Particle enthalpy [J] Enthalpy [J] Diffusive Flux of species i [kg/m2s]
11/13
k0 k K mi n Nj
Frequency factor [1/s] Process responsible for fragmentation Quantity describing a process responsible for fragmentation Volatile mass of species i inside the particle [kg] Number of size classes [-] Number of particles in size class j [-]
pg
Pressure of the gas phase [N/m2]
pg R Sm
Pressure of the solids phase [N/m2]
Si
Source term in the species balance [kg/m3s ]
Sh Sk , j
Source term in the gas energy balance [W/m3 ]
Universal gas constant [J/kmolK] Source term in the gas mass balance [kg/m3s ]
Rate constant of fragmentation of size class j due to process k
tf T Tg
Temperature [K] Gas temperature [K]
Tp
Particle temperature [K]
wg
Velocity of the gas mixture [m/s]
ws
Velocity of the solids [m/s]
Yi
Species mass fraction [-]
Factor for the correction of the heat transfer coefficient [-]
Greek letters:
α α corr
Corrected heat transfer coefficient [W/m2K]
εg
Volume fraction of gas [-]
εs
Volume fraction of solid [-]
ρg
Gas density [kg/m3]
ρs κ λ λp
Solid density [kg/m3]
Thermal conductivity of the particle [W/mK]
τg
Gas phase shear stress [N/m2]
τs
Solid phase shear stress [N/m2]
µg
Average viscosity of the gas phase mixture [Pas]
Heat transfer coefficient [W/m2K]
Correction factor for thermal conductivity inside the particle [-] Thermal conductivity [W/mK]
12/13
References: [1]
Atkinson, B. and Merrick, D.: Mathematical models of the thermal decomposition of coal: Heat transfer and temperature profiles in a coke oven charge, Fuel 1983, Vol. 62, 553
[2]
Austin et al.: Mathematical model of four phase motion and heat transfer in the blast furnace, ISIJ International, Vol. 37 (1997), No. 5, pp. 458-467
[3]
Carpenter, A.: Use of coal in direct ironmaking processes, IEA, London 2004
[4]
Ergun, S.: Fluid Flow through Packed Columns. Chem. Eng. Prog., 48(2):89–94, 1952.
[5]
Gnielinski V.: Wärmeübertragung Partikel-Fluid in durchströmten Haufenwerken, VDI-Wärmeatlas Kap.Gj1, Springer 2006
[6]
Kleiber et al.: Stoffwerte von sonstigen chemisch einheitlichen Flüssigkeiten und Gasen, VDIWärmeatlas Kap.Gj1, Springer 2006
[7]
Lee et al.: A development of computer model for simulating the transport phenomena in COREX melter-gasifier, ISIJ International, Vol. 39 (1999), No. 4 pp. 319-328
[8]
Merrick, D.: Mathematical models of the thermal decomposition of coal: 1. The evolution of volatile matter, Fuel 1983, Vol. 62, 534
[9]
Merrick, D.: Mathematical models of the thermal decomposition of coal: 2. Specific heats and heats of reaction, Fuel 1983, Vol. 62, 540
[10]
Mitchell, R.: An intrinsic kinetics-based,particle population balance model for char oxidation during pulverized coal combustion, Proc. of the Combustion Institute, Vol. 28, pp. 2261-2270, 2000
[11]
Nogami, H. et al.: Multi-demensional transient mathematicla simulator of blast furnace process based on multi-fluid and kinetic theories, Computers and Chemical Engineering 29 (2005) 24382448
[12]
Ouchiyama, N. and Tanaka, T.: Porosity Estimation from Particle Size distribution, Ind. Eng. Chem. Fundam. 1986, 25, 125-129
[13] Pedrag et al.: An improved model for fixed bed coal combustion and gasification, Fuel 1995, Vol. 74 No. 4 pp. 582-594 [14]
Shah KV et al. A kinetic-empirical model for particle size distribution evolution during pulverised fuel combustion. Fuel 2010
[15]
Zeisel. H: Mathematische Modellierung und numerische Simulation der Vorgänge im Hochofen, Trauner, Linz,1995
13/13
Experimental investigations into primary fragmentation of large coal particles
M. Kosowska-Golachowska, E.L. Koper*, R.L.J. Coetzer*, A. Luckos* Częstochowa University of Technology, Częstochwa, Poland * Sasol Technology R&D, Sasolburg, South Africa [email protected] Abstract Thermal fragmentation plays an important role in coal conversion processes because it accelerates devolatilization and affects the kinetics of gasification and combustion. In addition, fragmentation may generate large quantities of fine particles. Elutriation of these un-reacted coal fines lowers carbon conversion and, therefore, the thermal efficiency of the conversion process. Most of the fragmentation tests were conducted with medium-size particles in fluidized beds at temperatures 750–900°C. Little information is available in the open literature on the primary fragmentation of large coal particles typically encountered in coke ovens, smelting furnaces and fixed-bed gasifiers. In this paper the results of primary fragmentation tests conducted with single particles of three bituminous non-swelling coals in a laboratory-scale reactor at convective conditions are reported. All tests were carried out in inert atmosphere at ambient pressure and superficial gas velocity of 0.2 m/s. The effects of temperature, particle size, and mineral matter content on the fragmentation behaviour were studied. The primary fragmentation process was quantified through the probability of fragmentation and degree of fragmentation. During tests, coal particles were breaking up into pieces producing a few big pieces and a large number of fine fragments. Both particle size and gas temperature significantly influenced the fragmentation process. Only 20-mm particles tested at 400°C did not undergo the primary fragmentation. The most extensive fragmentation was observed for 80-mm particles at 600°C. The degree of fragmentation increased with increasing mineral matter content. The experimental data provide evidence that a ‘critical’ size exists below which coal particles do not experience primary fragmentation in the convective environment.
1. Introduction Thermal fragmentation plays an important role in coal conversion processes because it accelerates devolatilization and affects the kinetics of gasification and combustion. Fragmentation of coal particles during combustion is responsible for decreased burn out time, enhanced production of fly ash in fluidized-bed boilers as well as increased emission of sub-micrometer un-burnt char particles in pulverized coal combustion systems. Elutriation of un-reacted coal and char fines lowers the carbon conversion and, therefore, the thermal efficiency of the conversion process [1−3]. Because coal is not a homogeneous material, its fragmentation strongly depends on the physical properties and initial structure of the coal, and how this structure evolves during the heat-up time. Most of fragmentation tests reported in the open literature were conducted with mediumsize particles (5–15 mm) in fluidized beds at 750–900°C (e.g. [4−13]). Results from earlier studies on coal fragmentation and attrition were summarized in an extensive review by Chirone et al. [14]. Also several models have been developed to describe quantitatively the primary fragmentation process [5, 6, 15, 16]. However, little information is available on the primary fragmentation (i.e. fragmentation during devolatilization) of large coal particles (>10 mm) typically encountered in coke ovens, smelting furnaces and fixed-bed gasifiers. The primary fragmentation process is mainly associated with an increase in pressure inside the network of pores within the particle during the release of volatiles [8, 9]. This paper reports the results of primary fragmentation tests conducted with single particles of three bituminous coals in a laboratory-scale reactor at convective conditions. All tests were carried out in an inert atmosphere at ambient pressure and superficial gas velocity of 0.2 m/s. The effects of temperature (300−700˚C), particle size (20−80 mm), and mineral matter content (13−60%wt.) on the fragmentation behaviour were studied.
The primary fragmentation process was quantified fragmentation and the degree of fragmentation.
through
the
probability
of
2. Experimental 2.1. Test apparatus and procedure Primary fragmentation tests were conducted in a bench-scale unit shown in Figure 1. An electrically heated rectangular, 150×150×400 mm, chamber (1) was the main component of the unit. The chamber walls are insulated with a 30-mm layer of ceramic and glass fibre material and covered with a steel sheet. The front of the chamber was made of transparent quartz through which the fragmentation process could be directly observed. Nitrogen was supplied from cylinders (10) to a collector and then transferred through a pre-heater (7) directly into the reaction chamber. Flow rate of nitrogen was controlled by valves (11) and measured by a rotameter (12). K-type thermocouples (9) and microprocessor controllers (8) were used to measure and control temperatures inside the chamber and pre-heater with an accuracy of ±2ºC. Temperature of nitrogen was varied in the range of 300–700ºC, and the superficial gas velocity inside the chamber was 0.2 m/s for all tests conducted. Video and digital cameras were used to record the fragmentation process.
Figure 1. Test apparatus 1 – chamber, 2 – stand, 3 - ventilation duct, 4 – coal particle, 5 – computer, 6 – drawer, 7 – preheater, 8 – temperature controllers, 9 – K-type thermocouples, 10 – gas cylinders, 11 - pressure regulator, 12 – rotameter, 13 - laboratory balance Temperatures in the centre and at the surface of the coal particle and particle mass were measured using a stand (2) placed on the top of a laboratory balance (13). This stand provided support for two K-type thermocouples. The tip of the first thermocouple was located inside the particle, while the second thermocouple measured the surface temperature and served as a basket in which the coal particle was placed. The thermocouples were connected through an analogue-digital converter to a computer (5) where temperature measurements were stored. The mass of the particle was recorded every 1 s by a data logger connected to the balance (13) and computer (5). The coals used for the experiments were dense-medium separated into three narrow density ranges, to ensure approximately similar mineral matter content for each of the
particles. Spherical particles with diameters 20, 50 and 80 mm were prepared from coal lumps through mechanical grinding. A single coal particle was introduced into the reaction chamber for pre-determined time; then the sample was taken out, cooled in nitrogen and sieved through a sieve analyser (Retsch AS 200). Collected fragments of devolatilized coal particles were then subjected to several additional analyses including scanning electron microscopy (SEM), energy dispersive X-ray (EDX) spectroscopy and mercury porosimetry. Each fragmentation test was repeated three times and arithmetic mean values are presented in this paper.
2.2. Coals tested Three bituminous non-swelling coals with different mineral matter contentss were tested. Results of proximate and ultimate analyses of these coals are shown in Table 1. Table 2 lists selected physical properties of tested coals at 25°C. The laser-flash technique (Netzsch LFA-457) was used to determine thermal conductivity and heat capacity. Total porosity and average pore diameter were estimated from the results of mercury porosimetry analysis. Table 1. Proximate and ultimate analyses of coals tested C1
C2
C3
Proximate analysis (as received), % wt. Volatiles
25.6
23.7
13.9
Moisture
5.3
3.8
3.0
Ash yield
13.5
25.0
60.8
Fixed carbon
55.6
47.5
22.3
84.7
81.9
73.2
Hydrogen
3.7
3.4
2.9
Nitrogen
2.6
2.4
2.6
Sulphur
0.7
2.7
5.4
Oxygen
8.3
9.6
15.9
Ultimate analysis (dry, ash free), % wt. Carbon
Table 2. Physical properties of tested coals Property
C1
C2
C3
Thermal conductivity, W/m·K
0.22
0.26
0.55
Specific heat capacity, J/kg·K
916
1062
995
1.40–1.45
1.60–1.70
1.90–2.10
Total porosity, %
6.7
5.8
2.0
Average pore radius, nm
4.5
4.7
2.9
Relative density
3. Results and discussion 3.1. Physical properties and heating rates Figure 2 shows the relationship between volatile matter content (on dry, ash free basis) and mineral matter content (on moisture free basis) for three coals tested. The concentration of volatiles in the organic matter of coal increases approximately linearly with the mineral matter content. Since the release of volatiles during devolatilization causes a pressure build up inside the particle, this observation suggests that coals with higher mineral matter contents may be more prone to fragmentation than coals with lower mineral matter contents. Figure 3 shows results of the mercury porosimetry analysis. Both, the total porosity and the average pore radius decrease with mineral matter content. It can be expected that
7
5.0
38
6
4.5
5
4.0
4
3.5
3
3.0
36
34
32 Porosity
2
2.5
Average pore radius, nm
40
Total porosity, %
Volatile matter, % mass (daf)
flow of volatiles through smaller pores, in the case of coals with high mineral matter contentss, will be associated with higher internal pressures inside coal particles.
Pore radius
30 10
20
30
40
50
60
70
1
2.0 10
Ash content, % mass (mf)
20
30
40
50
60
70
Ash content, % mass (mf)
Figure 2. Volatile matter content as a function of mineral matter content
Figure 3. Porosity and pore radius as a function of mineral matter content
16
600
15
500
105
100
13 12
400 300
Centre temperature
95
Surface temperature 200
Mass, g
14
Temperature, °C
Compressive strength, MPa
Figure 4 shows the relationship between the compressive strength and mineral matter content. The compressive strength for coal C3 is approximately 30% lower than that for coal C1. A high concentration of minerals makes coal mechanically weaker and, therefore, more susceptible to fragmentation.
Mass
11
90
100 10 10
20
30
40
50
60
70
Ash content, % mass (mf)
0
85
0
500
1000
1500
2000
2500
3000
3500
4000
Time, s
Figure 4. Compressive strength as a function of mineral matter content
Figure 5. Typical temperature and mass profiles for an 50-mm C2 coal particle devolatilized at 500°C
Typical temperature profiles at the surface and in the centre of 50-mm particle are shown in Figure 5. This Figure clearly illustrates the existence of large temperature gradients within the particle especially at the early stage of devolatilization. Thermal stresses and shock associated with these gradients can contribute significantly to the primary fragmentation. Because the thermal conductivity for coal C3, the high mineral matter coal, is much larger than that for coal C1, the particle heats up easier and smaller temperature gradients can be expected.
3.2. The fragmentation behaviour The results of primary fragmentation tests are described in terms of the probability of fragmentation and degree of fragmentation. The probability of fragmentation, Sfp, is defined as the ratio of the number of particles that experienced fragmentation, Nfp, to the total number of particles, N, tested:
S fp =
N fp N
⋅100%
(1)
The degree of primary fragmentation [7], nf1, is the ratio of the number of fragments produced as a result of primary fragmentation, n1, to the total number of particles, N, tested:
n f1 =
n1 N
(2)
For non-fragmenting particles nf1 = 1, and for particles undergoing fragmentation nf1 > 1. The following conclusions can be drawn from the results of fragmentation tests: • The particles of tested coals were breaking up into pieces as the result of the primary fragmentation. Usually, a few big pieces and a large number of fine fragments were produced (Figures 10 and 11). • Both the particle size and gas temperature significantly influence the fragmentation process (Figures 6−8, 10 and 11). None of the 20-mm particles tested at 300 and 400˚C experienced fragmentation (Sfp=0). All 80-mm particles tested at 400, 500 and 600°C did undergo fragmentation (Sfp=1). The most extensive fragmentation was observed for the 80-mm particles at 600˚C. 140
Average degree of fragmentation
Average degree of fragmentation
12 10 8 6 4 2 0 200
300
400
500
600
700
20 mm 120 100
50 mm 80 mm
80 60 40 20
800
0 350
Temperature, °C
400
450
500
550
600
650
Temperature, °C
Figure 6. Influence of temperature fragmentation (C2 coal, 20-mm particles)
on
Figure 7. Average degree of fragmentation as a function of temperature (C2 coal) 80
400°C 50
Average degree of fragmentation
Average degree of fragmentation
60
500°C
40 30 20 10
70.33 70 60 50 40 30 20 10
0 10
20
30
40
50
60
70
80
90
Particle diameter, mm
Figure 8. Average degree of fragmentation as a function of particle diameter for C2 coal
1.67
2.00
C1
C2
0 C3
Figure 9. Influence of mineral matter content on fragmentation (20-mm particles, 500˚C)
• The extent of primary fragmentation increases with increasing mineral matter content (Figures 9 and 12). The 20-mm particles of C3 coal produced on average more than 70 fragments whereas particles of C1 and C2 coals never produced more than 4 fragments. • The primary fragmentation is strongly influenced by the total porosity and average pore diameter (pore size distribution). Coals with high total porosity and large pores do not undergo intensive fragmentation.
(a)
(b)
(c)
Figure 10. Fragments of 80-mm C2 coal particles after devolatilization at 400 (a), 500 (b), and 600˚C (c) 20 mm
50 mm
80 mm
Figure 11. Influence of particle size on fragmentation (fragments of C2 coal particles after devolatilization at 400ºC)
(a)
(b)
Figure 12. Influence of mineral matter content on fragmentation of 20-mm particles at 500˚C. Coal C1 (a), and C3 (b) • Particle ‘explosion’ was observed during devolatilization of C3 coal particles at higher temperatures. This violent behaviour can be attributed to high mineral matter content, high fraction of volatiles in the organic matter, and low porosity of that coal. Bateman et al. [17] and Eatough and Smoot [18] described similar behaviour of bituminous coal particles (5.5–7.0 mm) during devolatilization in a convective flow reactor.
4. Conclusions Primary fragmentation takes place during the devolatilization process as a consequence of internal stresses caused by the pressure of released volatiles flowing in the pore network, and by the thermal shock. The extent and pattern of primary fragmentation depends mainly on coal properties (volatile matter content, mineral matter content, porosity), particle size and heating rate. Measurements of coal properties show that there is a relationship between the amount of volatiles (daf), total porosity, compressive strength and mineral matter content. This relationship reveals that coals with high mineral matter contentss can be more susceptible to fragmentation. The existence of large temperature gradients during the initial stage of heating has been confirmed through temperature measurements at the surface and in the centre of the particle. These gradients are responsible for fragmentation via thermal shock that occurs within a few seconds after injection of the particle into the hot environment. All coals tested in this study experienced a degree of fragmentation during devolatilization. Although the fragmentation pattern varied from particle to particle, the particles usually break up into a few big pieces and a large number of fine fragments. The number of fragments produced, or the degree of fragmentation, increased significantly with increasing particle diameter and gas temperature. None of the 20-mm particles tested at 400°C experienced primary fragmentation. The most extensive fragmentation was observed for 80-mm particles. The experimental data provide evidence that a ‘critical’ size exists below which coal particles do not experience primary fragmentation in the convective environment. This critical size is a function of heating rate and, to the lesser extent, particle size and
mineral matter content. At gas temperatures below 500°C, the probability fragmentation for particles below 50 mm has been found to be very small.
of
5. References [1] Donsi, G. Massimilla, L. and Miccio, M.: Combustion and Flame, 41 (1981), 57-69 [2] Brown, R.C., Ahrens, J. and Christofides, N.: Combustion and Flame, 89 (1992), 95102 [3] Gajewski, W., Kosowska, M. and Otwinowski, H.: Powder Handling & Processing, 16 (2004), 252-257 [4] Sundback, C.A., Beer, J.M. and Sarofim, A.F.: Symp. Int. on Combustion, 20 (1984), 1495−1503 [5] Chirone, R. and Massimilla, L.: Symp. Int. on Combustion, 22 (1988), 267−277 [6] Chirone, R. and Massimilla, L.: Powder Technology, 57 (1989), 197−212 [7] Dakic, D., van der Honing, G. and Valk, M.: Fuel, 68 (1989), 911−916 [8] Chirone, R. and Massimilla, L.: Powder Technology, 64 (1991), 249−258 [9] Arena, U., Cammarota, A., and Chirone, R.: Symp. Int. on Combustion, 25 (1994), 219−226 [10] Stanmore, B.R., Brillard, A., Gilot, P. and Delfosse, L.: Symp. Int. on Combustion, 26 (1996), 3269−3275 [11] Zhang, H. Cen, K. Yan, J. and Ni, M.: Fuel, 81 (2002), 1835−1840 [12] Lee, J.M., Kim, J.S. and Kim, J.J.: Fuel, 82 (2003), 1349−1357 [13] Pinho, C.: Chem. Eng. J., 115 (2006), 147–155 [14] Chirone, R., Massimilla, L. and Salatino, P.: Prog. Energy Combust. Sci., 17 (1991), 297–326 [15] Chen, W.-Y., Nagarajan, G., Zhang, Z.-P., Shen, B.-C. and Fan, L.T.: Ind. Eng. Chem. Res., 33 (1994), 137–145 [16] Sasongko, D. and Stubington, J.F.: Chem. Eng. Sci., 51 (1996), 3909–3918 [17] Bateman, K.J., Germane, G.J., Smoot, L.D. and Eatough, C.N.: Energy & Fuels, 9 (1995), 295–301. [18] Eatough, C.N. and Smoot, L.D.: Fuel, 75 (1996), 1601–1605.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
IVESTIGATIO OF THE PYROLYSIS AD GASIFICATIO OF A TURKISH COAL USIG THERMAL AALYSIS COUPLED WITH MASS SPECTROMETRY Sibel Özdoğan*, Prof. Dr. Marmara University Faculty of Engineering Contact Information: Marmara University , Faculty of Engineering , 34722 Kuyubasi , Istanbul, Turkey Tel.: +90216 4182357/1212, Fax: +90 2163480293, [email protected] Ugur Özveren, Marmara University Faculty of Engineering Contact Information: Marmara University , Faculty of Engineering , 34722 Kuyubasi , Istanbul, Turkey Tel.: +90216 3480292/240, Fax: +90 2163480293, [email protected] Aylin Boztepe, Yildiz Technical University Faculty of Engineering Contact Information: Marmara University , Faculty of Engineering , 34722 Kuyubasi , Istanbul, Turkey Tel.: +90216 3480292/240, Fax: +90 2163480293, [email protected] Mehmet Beypınar, Marmara University Faculty of Engineering Contact Information: Marmara University , Faculty of Engineering , 34722 Kuyubasi , Istanbul, Turkey Tel.: +90216 3480292/240, Fax: +90 2163480293, [email protected] Abstract Clean conversion of coal into gaseous and liquid fuels and chemicals is of utmost importance along with its direct utilization via combustion in a sustainable manner. Pyrolysis is an important intermediate stage in coal conversion processes such as combustion and gasification. Coal gasification is applied in IGCC systems for power generation as well as for the production of a range of gaseous and liquid fuels from hydrogen to diesel. Thermal analysis results coupled with mass spectrometry (TA/MS) applied to the pyrolysis and gasification of a Turkish coal are presented here. Pyrolysis experiments have been carried out in argon atmosphere while air has been used as the gasification agent. The samples have been heated from room temperature up to 1000°C. The main evolved products have been identified through the on-line recorded mass spectra. The thermolysis behavior of the coal sample has been checked comparatively for selected gas flow rates and heating rates. The gas flow rate has been changed between 5 and 70 mL/min whereas the heating rate range has been selected as 10 to 40 °C/min. Two different sample masses (5 and 10mg) have been used to observe any effect of the mass on the thermal analysis results. Final operation conditions are based on these results. The TG curves, the on-set, shape and off-set of the DTG curves combined with the evolution temperatures of the major gaseous products such as H2, CO, CO2, CH4, H2S and SO2 for the selected appropriate operation conditions clearly show differences in the thermolysis behavior at different atmospheres, gas flow rates and heating rates. These differences in thermolysis behavior indicate when and under which conditions pyrolysis and/or partial oxidation (gasification) and/or oxidation might be occurring . The TA/MS results are also indicative of the reactivity of the coal sample to the selected atmosphere at various temperatures. TG-MS results can be utilized to determine kinetic parameters via various methods. Care should be taken to avoid mass transfer limitations. The results prove that the gasification of coal consists of two major reactions: pyrolysis and gasification of in situ formed char. The rate of the latter reaction is much slower than that of the former reaction. The volume of the gasifier is therefore primarily dependent on the gasification rate of char. For this reason, kinetics of char gasification obtained by TA/MS plays a key role by providing valuable information for the proper design and operation of gasifiers.
*Author for correspondence
ozdogan et al
1
1. Introduction Global coal reserves contain a very broad scope of coals ranging from high quality bituminous coals to low quality lignites with broad differences in organic and inorganic matrices. The qualitative and quantitative compositions of the organic and inorganic coal matrices determine the thermolysis behaviour of coals for a given atmosphere and temperature programme. Major thermolysis processes are combustion, gasification and pyrolysis [1-3]. Combustion is used to extract energy directly from coals. Gasification is a thermo-chemical process by which the organic coal matrix is converted to gaseous products with useable heating value; the gasification products can be further processed to liquid fuels or to chemicals (primarily carbon monoxide and hydrogen along with other hydrocarbons and CO2 in a controlled oxidizing atmosphere). Pyrolysis is the first chemical step in nearly all coal thermolysis processes, and has a significant influence on the subsequent process stages [4-5]. Therefore, understanding pyrolysis behavior is important in improving existing and developing new coal utilization technologies. Accurate knowledge of the whole coal thermolysis behavior is of utmost importance in the sustainable utilization of coal. Pyrolysis means the decomposition of coal while being heated to elevated temperature in an inert atmosphere and produces a hydrogen-rich volatile fraction with high hydrogen to carbon ratio which consists of gases, vapor and tar components and a carbon-rich solid residue, called char. The characteristics of the produced char control the subsequent thermolysis processes, including combustion, and in particular gasification. The reactivity and kinetics of coal char with various combustion (oxygen and/or air) and gasification (air/oxygen/steam/carbon dioxide etc.) atmospheres provide valuable information for clean coal utilization and have been widely studied [6]. Thermogravimetric analysis (TG) is one of the most common techniques used to investigate the thermal events and associated conversion kinetics during pyrolysis, gasification and combustion of coals in defined atmospheres with selected heating programmes. Non-isothermal TG measures the weight of a sample with respect to the change in temperature and represents a quantitative weight change associated with a thermally induced transition. In TG the loss in weight is recorded as a function of temperature or time while the derivative thermogravimetric analysis (DTG) represents the rate of mass loss based on TG data [7-10]. TG data can be used to obtain kinetic information. Combining TG with gas-analytical techniques significantly enhances the possibilities for interpreting the formation of thermolysis process products [11]. The knowledge about the temperatures and rates of formation of desired and undesired gaseous species in a given thermolysis atmosphere can be used for improving/redesigning /designing various processes such as thermal coal desulfurization, tar minimization during gasification, carbon combustion efficiency improvement, methane formation minimization, enhancement of hydrogen production etc. The results presented here are for Soma lignite - a quite good quality Turkish lignite with 400Mtonnes reserve. Our previous and on-going TG-MS studies indicate quite different thermolysis behaviour for Turkish coals which range from high to low quality lignites . 2. Methodology 2.1. Coal Soma lignite has a lower heating value of 4462 kcal/kg on dry basis. It has 40% volatile matter, 31% fixed carbon and 29% ash on dry basis. The total sulfur content is 1,34% on dry basis. Comprehensive information on the properties of Soma lignite are available elsewhere [12]. 2.2. Gases Argon and air are supplied by the Turkish gas company HABAS as 5 grade high purity gases (99,999% pure). 2.3. TG-MS System The TG apparatus used in this study is NETZSCH Simultaneous Thermal Analyser STA 409 PC which can be operated under selected inert, oxidative and reactive atmospheres in the temperature range of 25°C-1500°C at heating rates up to 50K/min. Inert gases (N2, Ar, He), oxidative gases (air, O2 enriched air, O2/air with inert gases), reactive gas mixtures (CO2/ CO /O2/inert gas/limited amounts of H2/H2S/SO2/COS ) can be used. It is also possible to introduce water vapor into the TG atmosphere with the other thermal analysis instrument – NETZSCH STA F3 Jupiter H2O in the Marmara University thermal analysis laboratories. Both TG systems are coupled with quadrupole mass spectrometers (QMS 403C). Hence, along with TG/DTG measurements evolved gases can be analysed simultaneously. The transfer lines between the TG and MS are heated in order to avoid condensation of the gaseous products.
ozdogan et al
2
2.3.1. Parameters Affecting the TG-MS Measurements Appropriate coal thermolysis operation parameter ranges were determined prior to analyses with the TG-MS system, primarily to achieve appropriate heat and mass transfer throughout the sample and to avoid diffusion control considering kinetic studies . 2.3.1.1. Sample Mass As the amount of the sample used increases, several problems may arise. The whole sample used may not come into contact with the surrounding atmosphere; e.g. in oxidative atmosphere the upper and bottom portions of the sample in the crucible may undergo different reactions. The usage of too small amounts of samples with coal can cause homogeneity related problems. In the literature, sample masses as low as 5 mg are used for coal based thermal analysis studies [13-16]. The weight of Soma lignite for this study was kept at about 10 mg. 2.3.1.2. Particle Size If the particle size of the sample is too small, a portion of the sample can be lost during the evacuation of the TG furnace. On the other hand diffusion limitations affect measurements, if the particle size of the sample is too large. In the literature, the particle size used for thermal analysis experiments is in the range of 100-250 µm [1722]. Soma lignite at the size of <250µm was selected for this study. 2.3.1.3. Heating Rate and Temperature Programme The main practical difference between rapid and slow heating rates is that rapid heating may reduce the time needed to attain a specified weight loss. Big differences in heating rates (1K/min and 500K/min) may also affect thermolysis mechanism. In the heating rate ranges of our studies (1-50K/min), increasing the heating rate appears to increase the on-set and off-set temperatures of thermolysis peaks rather than changing the thermolysis mechanism.Soma lignite was heated from 25°C to 1000°C at a constant heating rate of 40° C/min in this study. 2.3.1.4. Thermolysis Atmosphere and Flow Rate In pyrolysis and inert gas containing reactive atmospheres, inert gases such as nitrogen, argon and helium can be used. Helium has the highest thermal and mass diffusivity; the former property is desired, however, the latter makes helium the least suitable gas giving rise to changes of the shape and intensity of the MS signal. Nitrogen and argon are preferred to helium due to their similar properties concerning their application in TG-MS systems in pure form or as mixtures [23]. Argon usually contains less impurity than nitrogen and is preferred for quantative MS applications. Nitrogen has the same molecular mass as CO and makes CO determinations by mass spectrometry problematic [24]. In this study, Argon was used as carrier gas for pyrolysis and air as gasification atmosphere. To determine the appropriate flow rates for argon and air atmospheres the flow rate range of 5-70 ml/min has been covered for Soma lignite at a heating rate range of 10 to 40K/ min. The flow rate has only a minor effect on pyrolysis results; however, under the oxidative atmosphere the magnitude of the flow rate definitely affects the results. As the flow rate of the oxidative gas increases the thermolysis behaviour changes from gasification to combustion. A similar but definitely not so pronounced effect was observed as the heating rate decreased from 50K/min to 10K /min. Details can be found elsewhere [25]. The flowrates used in the literature with oxidative atmospheres cover a broad range from 5ml/min to 100ml/min [14,26-29]. For pyrolysis, generally, flowrates in the range of 40-50 ml/min are used [7,13,30-33]. In this study the final argon and air flow rates have been selected as 40 and 20ml/min, respectively. 2.4. MS Signals In this study the MS signals simultaneously determined along with the TG-DTG signals for the temperature range 25-1000°C are as follows: 2 (H2), 16 (CH4), 18 (H2O), 28 (CO) , 44 (CO2) and 34 (H2S), 60 (COS) and 64 (SO2) . The mass spectrometer has been checked with appropriate reference gas mixtures as well as with solid TG reference samples for the TG-MS coupled case. 3. Results Pyrolysis and air gasification TG-DTG results with Soma coal are presented in Fig. 1. Pyrolytic decomposition is observed first at 300°C and continues up to around 900°C as indicated by the TG-DTG-MS results for Soma lignite (Figures 1-3). Chemical reactions begin to occur around 250°C and end around 800°C during air gasification. Pyrolytic weight loss is much less than the weight loss of air gasification; only about 54% of the organic coal matrix is pyrolysed versus the 97% of the air gasified organic matrix. It should also be mentioned that 38% of the organic matrix of Soma lignite is converted to gaseous products up to 650°C while about 67% of the organic matrix is lost during air gasification for the same temperature interval. Moreover pyrolysis reactions
ozdogan et al
3
leading to gaseous products appear to occur primarily in two distinct peaks while air gasification appears primarily as a very broad peak followed by a sharp one around 800°C.
(a)
(b)
Figure 1. Pyrolysis (a) and air gasification (b) TG-DTG results for Soma lignite
Figures 2 and 3 present the evolution of pyrolysis and gasification products of Soma lignite. All gaseous products are observed at higher temperatures for the pyrolysis case compared to air gasification in line with the TG-DTG results. Hydrogen production covers a much broader temperature range with several peaks for gasification versus one broad peak in pyrolysis. Carbon monoxide is formed as one small broad low temperature and one sharp high temperature peak for pyrolysis while CO formation is shifted to lower temperatures for the gasification case. Carbon dioxide evolution occurs in two stages giving a small broad peak at the temperature of 427°C and a sharp peak at 735°C in pyrolysis. In air gasification CO2 evolution occurs in a much broader temperature range with a peak shoulder at 440°C and a second peak at 702°C. Methane formation is observed for both pyrolysis and gasification as one broad major peak followed by a second small one at high temperatures. Hydrogen sulfide is the sulfurous product of pyrolysis while COS and SO2 both are observed during air gasification. H2S evolves in one relatively big peak (peak temperature: 450°C) and one small peak (peak temperature: 550°C) during pyrolysis. TG-MS results for air gasification indicate that H2S production peaks around 340°C and diminishes around 450°C; some minor H2S evolution may occur above 800°C. COS begins to form around 300°C in air gasification and its evolution peaks are around 373 and 445°C while sulfur dioxide appears around 400°C, peaks at 474°C and diminishes sharply at 500°C.
4. Concluding Remarks TG-DTG-MS data for Soma coal pyrolysis in argon atmosphere and gasification in air indicate quite well the dependency of the qualitative and quantitative product formation of coals and other hydrocarbons in selected atmospheres and heating regimes. The temperature and atmosphere related formation data is of utmost importance for primary evaluation of coals towards clean coal thermolysis technologies as well as for designing custom fit processes. Data about sulfurous product formation is important in terms of their removal considering not only environmental limitations but also their poisoning effect on catalysts in further process stages following pyrolysis or gasification. Data taken for several Turkish lignites and their mixtures with biomass at various heating rates and reactive atmospheres including steam are being developed and kinetically analyzed (isoconversion , model fitting and regression ). Results are to be will be published in the near future.
ozdogan et al
4
Figure 2. H2, CO, CO2 and CH4 evolutions during Argon pyrolysis and air gasification of Soma lignite
SOMA AIR TG- MS
SOMA ARGO TG-MS
ozdogan et al
5
Figure 3. H2O, H2S, COS and SO2 evolutions during Argon pyrolysis and air gasification of Soma lignite
SOMA AIR TG- MS
SOMA ARGO TG- MS
ozdogan et al
6
Acknowledgment The authors acknowledge the support of TUBITAK within the frame of the TARAL TRIJEN Project entitled “Liquid fuel production from the mixtures of coal and biomass” (Project number:108G043) , the support of BAP of Marmara University . References [1] T. Faravelli, A.Frassoldati et al., Detailed kinetic modeling of the thermal degradation of lignins, Biomass and Bioenergy 34, 290–301, (2010). [2] C.L. Sun, Y. 0. Xiong et al., Thermogravimetric study of the pyrolysis of two Chinese coals under pressure, Fuel 76, 639-644, (1997). [3] Y. Chen, J. Duan et al., Investigation of agricultural residues pyrolysis behavior under inert and oxidative, J. Anal. Appl. Pyrolysis 83, 165–174, (2008). [4] A. Arenillas , F. Rubiera et al., Thermal behaviour during the pyrolysis of low rank perhydrous coals, J. Anal. Appl. Pyrolysis 68-69, 371-385, (2003). [5] J.J. Manya`, E. Velo et al., Kinetics of biomass pyrolysis: a reformulate three-parallel-reactions model, Ind. Eng. Chem. Res. 42, 434-441, (2003). [6] S. Sawettaporn, K. Bunyakiat et al., CO2 gasification of Thai coal chars: Kinetics and reactivity studies, Korean J. Chem. Eng. 26 (4), 1009-1015, (2009). [7] H.B.Vuthaluru, Thermal behaviour of coal/biomass blends during co-pyrolysis, Fuel Processing Technology 85, 141–155, (2003). [8] M. Otero, M.E.Sánchez et al., Simultaneous thermogravimetric-mass spectrometric study on the cocombustion of coal and sewage sludges, Journal of Thermal Analysis and Calorimetry 86, 489–495, (2006). [9] L. Zhou, Y. Wang, et al., Thermogravimetric characteristics and kinetic of plastic and biomass blends copyrolysis, Fuel Processing Technology 87, 963–969, (2006). [10] K.G. Mansaray, A.E. Ghaly, Thermal degradation of rice husks in nitrogen atmosphere, Bioresource Technology 65, 13-20, (1998). [11] S. Materazzi, Mass spectrometry coupled to thermogravimetry (TG-MS) for evolved gas characterization: a review, Applied Spectroscopy Reviews 33 (3), 189-218, (1998). [12] S. Ozdogan et al., The Determination of Selected Fuel and Additive Thermo-chemical Properties , The First Interim Report of Thermo-chemical Properties of fuels and additives, Marmara University, Project number:108G099 (January, 2010). [13] A. Arenillas, F. Rubiera et al., Thermogravimetric–mass spectrometric study on the evolution of nitrogen compounds during coal devolatilisation, Journal of Analytical and Applied Pyrolysis 65, 57–70, (2002). [14] K.E. Benfell, B.B. Beamish et al., Combustion behaviour of Bowen Basin coals, Fuel Processing Technology 60, 1–14, (1999). [15] R. Font, J.A. Conesa et al., Kinetics of pyrolysis and combustion of pine needles and cones, J. Anal. Appl. Pyrolysis, (2008). [16] C.J. Gomez, E.Mészaros, E.Jakab et al., Thermogravimetry/mass spectrometry study of woody residues and an herbaceous biomass crop using PCA techniques, J.Anal. Appl. Pyrolysis 80, 416–426, (2007). [17] K.J. Shaw, B.B. Beamish et al., Thermogravimetric analytical procedures for determining reactivities of chars from New Zealand coals, Thermochimica Acta 302, 181-187, (1997).
ozdogan et al
7
[18] M.V. Kok, Coal Pyrolysis: Thermogravimetric study and kinetic analysis, Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 25: 10, 1007-1014, (2003). [19] Y.Elbeyli and S.Pişkin, Combustion and pyrolysis characteristics of Tunçbilek lignite, Journal of Thermal Analysis and Calorimetry, (2005). [20] B.K. Saikia, R.K. Boruah et al., A thermal investigation on coals from Assam (India), Fuel Processing Technology 90, 196 – 203, (2009). [21] Y. Ming-gao, Z. Yan-min et al., Thermal analysis of kinetics of coal oxidation, Procedia Earth and Planetary Science 1, 341–346, (2009). [22] L. Chang, Z. Yan-min et al., Research on low-temperature oxidation and pyrolysis of coal by thermal analysis experiment, Procedia Earth and Planetary Science 1, 718–723, (2009). [23] M. Maciejewski, A. Baiker, Quantitative calibration of mass spectrometric signals measured in coupled TA-MS system, Thermochimica Acta 295, 95-105, (1997). [24] L.F.Calvo, M.E.Sánchez et al., TG-MS as a technique for a better monitoring of the pyrolysis, gasıfıcation and combustion of two kınds of sewage sludge, Journal of Thermal Analysis and Calorimetry, Vol.78, 587–598, (2004). [25] S. Ozdogan et al., The Determination of Selected Fuel and Additive Thermo-chemical Properties , The Second Interim Report of Thermo-chemical Properties of fuels and additives, Marmara University, Project number:108G099 (May 2010). [26] M.V. Kok, An investigation into the thermal behavior of coals, Energy Sources 24, 899–905, (2002). [27] M.V. Kok, C. Hicyilmaz et al., Effect of cleaning process on the combustion characteristics of two different rank coals, Energy&Fuels 15, 1461-1468, (2001). [28] Y. Ming-gao, Z. Yan-min et al., Thermal analysis of kinetics of coal oxidation, Procedia Earth and Planetary Science 1, 341–346, (2009). [29] A. Jess, A.K. Andresen, Influence of mass transfer on thermogravimetric analysis of combustion and gasification reactivity of coke, Fuel, (2009). [30] Y. Güldogan, T. Durusoy et al., Effects of heating rate and particle size on pyrolysis kinetics of Mengen lignite, Energy Sources 23, 337-344, (2001). [31] J. Guo, A.C. Lua, Kinetic study on pyrolytic process of oil-palm solid waste using two-step consecutive reaction model, Biomass and Bioenergy 20, 223–233, (2001). [32] A. Arenillas, C. Pevida et al., Characterisation of model compounds and a synthetic coal by TG/MS/FTIR to represent the pyrolysis behaviour of coal, J. Anal. Appl. Pyrolysis 71, 747–763, (2004). [33] M. Krzesińska, U. Szeluga et al., The thermal decomposition studies of three Polish bituminous coking coals and their blends, International Journal of Coal Geology 77, 350–355, (2009).
ozdogan et al
8
Development of Post Combustion Carbon Capture Technology,
M Hunt, F D Fitzgerald, S Black & R A Gardiner Doosan Power Systems Limited, United Kingdom
ABSTRACT Doosan Power Systems Ltd (DPS) are currently developing post-combustion and oxyfuel carbon capture technologies for early commercialisation. The purpose of this document is to highlight the work that is being carried out by DPS specifically on Post Combustion Carbon Capture. Post Combustion Capture DPS licensed an advanced solvent scrubbing technology from HTC Purenergy Inc. back in September 2008. HTC who are based in Regina, Canada, work very closely with the University of Regina who are recognised as one of the leading global research institutes in Post Combustion Carbon Capture. In addition to the licence agreement, DPS acquired a 15% stake in HTC to further cement the bond. DPS are now preparing commercial bids for fullscale solvent scrubbing plants, comprising all equipment from flue gas desulphurisation (FGD) plant outlet to CO2 compression plant inlet.
The long-term performance of HTC’s
proprietary amine-based solvent, RS2, has been assessed in verification runs at the Boundary Dam pilot plant (4 tonnes per day CO2 capture). Further work is being carried out by DPS at its purpose built solvent scrubbing pilot plant in Renfrew, Scotland. This plant is part of the Emissions Reductions Test Facility (ERTF) with an approx. 1tpd CO2 capture rate. This paper outlines Doosan Power Systems post combustion capture development and commercialisation activities.
INTRODUCTION To achieve the global target reduction in CO2 emissions of some 60% by 2050 [1], carbon dioxide capture and permanent underground storage (CCS) will be necessary.
Carbon
abatement technologies (CATs) will be required for retrofit to existing power plant and installation on new build power plant. Doosan Power Systems has taken a proactive approach to developing CATs, leading and supporting techno-economic studies on both Track 1 and Track 2 approaches of the UK DTI’s Carbon Abatement Technology Strategy [1].
The Track 1 approaches are available
now and reduce CO2 emissions per unit of electricity generated by means of improved cycle efficiency and biomass co-firing (carbon neutral).
The Track 2 approach, CCS, achieves
much larger reductions, up to 95%, by means of oxyfuel combustion, post-combustion capture (PCC) or integrated gasification combined cycle (IGCC) technologies. Initially, studies were undertaken at a high level, i.e. technical and economic feasibility. Two notable such studies were undertaken as part of the IEA Greenhouse Gas (IEA GHG) R&D programme and focused on new build power plant, with post-combustion capture of CO2 [2] and oxy-combustion processes for CO2 capture [3].
The technical and economic
aspects were investigated in detail, considering integration of the CATs, determining relative benefits of alternative approaches and reviewing the applicability of optimum solutions. Projects were also completed to evaluate and optimise how CAT retrofits can be accomplished on the UK fleet of coal-fired power plants [4], to demonstrate the potential application of advanced supercritical (ASC) pulverised fuel (PF) boiler / turbine technology and CO2 capture capabilities to the Canadian power generation market [5], and to develop new build designs of oxyfuel greenfield pulverised fuel (PF), bituminous and lignite, coalfired power plants (Enhanced Capture of CO2 (ENCAP) oxyfuel combustion sub-project) [6]. The studies concluded that post-combustion capture and oxyfuel combustion are technically feasible and represent competitive options for CO2 capture that could be applied to both new build and existing plant as retrofit.
However, research, development and demonstration of
the technologies are required in the short-term to gain operational experience and provide confidence at pilot-scale and then full-scale, and in the long-term to improve performance and reduce efficiency loss, making the CATs more attractive economically.
POST COMBUSTION CAPTURE Overview Post combustion carbon capture (PCC) uses gas-liquid scrubbing technology to remove CO2 from the flue gas produced in a conventional air-fired combustion plant. Flue gas exiting the flue gas desulphurisation (FGD) plant is cooled in a direct contact cooler (DCC) and then passes through an absorber (scrubber) column, where the CO2 is absorbed into an amine based solvent. The cleaned flue gas is exhausted to atmosphere and the CO2 rich solvent is passed to a regeneration (stripper) column via a lean/rich solvent heat exchanger. A reboiler provides the heat required to desorb the CO2 from the solvent.
The CO2 lean solvent is
cooled and returned to the absorber column and the CO2 is compressed for transport and sequestration.
Figure 1: Schematic Diagram of the Post Combustion Carbon Capture Process Post Combustion Capture Research, Development and Commercialisation Activities Amine based solvent scrubbing technology has been widely applied in the oil and gas industry for natural gas sweetening since the 1950s. Its application to CO2 removal from combustion process flue gas has been demonstrated at laboratory- and pilot-scale by many organisations.
The scale-up to full-scale power plant
(300 to 400 MWe) will be demonstrated in the UK Carbon Capture and Storage Competition, and the European Union Flagship Programme, which aims to have ten to twelve full-scale
carbon capture and storage demonstration plants (representing the leading carbon capture technologies) in operation by 2015. DPS are now preparing commercial bids for full-scale solvent scrubbing plants, comprising all equipment from flue gas desulphurisation (FGD) plant outlet to CO2 compression plant inlet. Additionally, through HTC, DPS has expertise in geological sequestration and enhanced oil recovery (EOR). DPS development activities include the pilot-scale testing and optimisation of the PCC process on the 160kWt Emissions Reduction Test Facility (ERTF). These activities include testing of solvent and packing performance and process optimisation to minimise cycle efficiency loss for large and full scale applications. The 160 kWt ERTF, Figure 2, was originally built in the early 1990s to develop in-furnace NOx control technologies (staged combustion and gas-over-coal and coal-over-coal reburn). It has subsequently been modified to include Selective Non-Catalytic Reduction (SNCR) and Selective Catalytic Reduction (SCR) reactors, an electrostatic precipitator (ESP) and oxyfuel combustion capability. A solvent scrubbing pilot plant has been installed on the ERTF and was officially opened in June 2010.
Figure 2: Schematic Diagram of Emission Control Test Facility (ERTF) The University of Regina’s Boundary Dam PCC demonstration plant has a 4 tonnes per day CO2 capture capacity. It is located at SaskPower’s 813 MWe Boundary Dam power station
in Estevan, Saskatchewan, Figure 3. The long-term performance of the proprietary aminebased solvent, RS2, has been assessed in a verification run. Preliminary results will be presented in the final version of the paper.
Figure 3: HTC’s Boundary Dam PCC Demonstration Plant
GRAPH AND/OR TABLE OF RESULTS AND BRIEF DISCUSSION.
In January 2010 DPS started work on the construction of a 100 tpd CO2 capture plant at Ferrybridge, UK. This facility which is being fast tracked will give a fully operational plant by the end of Q1 2011 and allow the consortium team of SSE, Vattenfall, DECC and Yorkshire Forward to develop further their plans for carbon capture and storage. A progress report on this project will be given at the conference along with details of other projects currently being developed such as a $5million FEED study at Antelope Valley Power station on behalf of Basin Electric. CONCLUSIONS
Excellent progress is being made in the development of post-combustion carbon capture technology, with early commercialisation a realistic target. TO BE COMPLETED WHEN TEST RESULTS AVAILABLE. ACKNOWLEDGEMENTS The authors gratefully acknowledge the grant funding provided by the UK Department of Energy and Climate Change (DECC), UK Technology Strategy Board (TSB) and the UK Engineering and Physical Sciences Research Council (EPSRC) and the technical and financial contributions made by the project collaborators: Air Products plc, E.ON UK plc, RWE npower plc, BP Alternative Energy International Limited, University of Nottingham, Imperial College London, Scottish and Southern Energy plc, ScottishPower Limited, Drax Power Limited, EDF Energy plc and DONG Energy A/S and Vattenfall AB.
The authors
also acknowledge the support of the many members of Doosan Power Systems PCC research and development, engineering and construction teams. REFERENCES 1. A Strategy for Developing Carbon Abatement Technolgies for Fossil Fuel Use, Department for Trade & Industry, DTI/Pub URN 05/844, June 2005. 2. Improvement in Power Generation with Post Combustion Capture of CO2, IEA GHG R&D Programme Report No.: PH4/33, November 2004. 3. Oxy Combustion Processes for CO2 Capture from Power Plant, IEA GHG R&D Programme Report No.: 2005/9, July 2005. 4. R. Panesar, M. Lord, S. Simpson, V. White, J. Gibbins and S. Reddy, Coal-Fired Advanced Supercritical Boiler / Turbine Retrofit with CO2 Capture, 8th International Conference on Greenhouse Gas Control Technologies, Trondheim, Norway, 19-22 June 2006. 5. Future CO2 Capture Technology Options for the Canadian Market, Carbon Abatement Technologies Programme, Department for Business Enterprise & Regulatory Reform Report No.: COAL R309, BERR/Pub URN 07/1251, March 2007. 6. G. Sekkappan, P.J. Melling, M. Anheden, G. Lindgren, F. Kluger, I.S. Molinero, C. Maggauer and A. Doukelis, Oxyfuel Technology for CO2 Capture from Advanced Supercritical Pulverised Fuel Power Plants, 8th International Conference on Greenhouse Gas Control Technologies, Trondheim, Norway, 19-22 June 2006.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
LIGNITE DERIVED CARBONS FOR CO2 CAPTURE Alan L. Chaffee, Seamus Delaney and Gregory P. Knowles CRC for Greenhouse Gas Technologies, School of Chemistry, Monash University, Victoria, Australia Author Contact: [email protected] The very low cost and very low mineral content of Victorian lignite makes this material an attractive precursor for carbon materials1. Carbon based adsorbents could potentially be used in a conventional pressure swing adsorption (PSA) or vacuum swing adsorption (VSA) process2. If fabricated into appropriate monoliths they could also be employed in electrical swing adsorption (ESA) process3. ESA is based on the realization that rapid and uniform desorption can be achieved by passing an electrical current through the adsorbent, potentially leading to improvements in cycle-time and efficiency. Carbons with substantial mesopore volumes have been targeted to facilitate rapid adsorption/desorption cycling and/or to provide the volume necessary for further functionalisation of the substrate. In this study a range of carbon based adsorbents were prepared by steam activation of titaniaimpregnated Victorian lignite as illustrated in Figure 1. The titania was removed by acid washing following activation4. Pore size analysis (using nitrogen adsorption data and the BJH method) indicated a total pore volume of 0.48 mL/g) with a significant volume present as mesopores (0.16 mL/g) and an average pore size of approximately 3.7nm (Table 1). X-ray diffraction analysis indicated no ordered symmetry.
Figure 1. Preparation of organometallic catalysed mesoporous activated carbon.
This carbon (Ti-C) was then further modified to incorporate approx 1% nitrogen. The carbon was nitrated (Ti-C-H2-NO2) and the resulting nitrate groups were reduced using aqueous ammonia (Ti-C-H2-NH2). Carbons with basic surfaces are an attractive prospect because they are expected to react with the weakly acidic CO2 molecule through an acid-base reaction. Adsorption/desorption then becomes a reversible chemisorption phenomenon and the selectivity for CO2 over other components in the feed gas might be considerably improved relative to conventional carbons. When nitration was carried out with fuming nitric acid, the pore volume dropped significantly to about half its prior value and the proportion of mesopores was reduced, suggesting that some structural degradation and pore blockage had developed during the modification. When using sodium nitrate as the nitrating agent5, this degradation did not occur such that the original pore volume distribution was substantially maintained (Table 1). The original mesoporous carbon Ti-C contained about 0.5 wt % inherent nitrogen. It can be seen that a further 0.6 wt% N was incorporated into the carbon via the functionalisation procedure Ti-CH2-NH2. Table 1. Characterisation of Ti-C and functionalised Ti-C samples Pore volume distribution
BJH pore size
N by microanalysis
m2/g
nm
wt %
ml/g
± 10
± 0.1
± 0.1
± 0.02
Ti-C
732
3.7
0.52
0.48
0.18
0.10
0.16
Ti-C-H2a
798
3.6
0.54
0.52
0.17
0.12
0.17
Ti-C-H2-NO2b
610
3.5
0.82
0.38
0.18
0.06
0.12
Ti-C-H2-NH2c
656
3.5
1.14
0.42
0.20
0.05
0.13
BET SA Substrate and functionalised substrate samples
a
Total
Small micro
Large micro
Meso
hydrogenated at 375°C; b nitrated using sodium nitrate; c reduced using aqueous ammonia
The reversible CO2 adsorption capacities of these materials were measured and compared by a thermogravimetric method. It can be seen (Table 2) that as much as 8.2 wt% (2.1 mmol/g or 41 g/L) CO2 was adsorbed on the non-functionalised mesoporous carbon from a CO2-rich stream (90% in Ar) at ambient conditions. Of this 0.8 wt% was irreversibly adsorbed at ambient conditions, but could be quantitatively recovered from the adsorbent by increasing the temperature to 120°C. Thus the working capacity for CO2 capture is reported as 7.4 wt% (1.7 mmol/g or 33 g/L) CO2. Modification of the carbon with N led to a small increase in the adsorption capacity to 9.4 wt% (2.2 mmol/g or 41 g/L) CO2. The working capacity per gram of sample was unchanged, but there was a small increase in the working capacity per gram of substrate (carbon) to 8.5 wt% (1.9 mmol/g or 37 g/L) CO2. The incorporation of N also led to a small increase in the heat of
adsorption, from 28.3 to 32.1 kJ/mol of CO2 adsorbed. These increases were smaller than expected, indicating that CO2 capture on the functionalised material was still occurring predominantly via a physisorption mechanism, not via a chemisorption mechanism as had been postulated. Table 2. Gas separation properties for Ti-C and functionalised Ti-C samples Gas separation properties (mass and volume basis) CO2 ads mass% (init) /g sample
Working capacity /g sample
Working capacity /g substrate
Hads @ 20oC (init)
%
%
%
kJ.mol-1
± 0.1
± 0.1
± 0.1
± 0.1
Ti-C
8.2
7.4
7.4
28.3
Ti-C-H2
9.1
8.2
8.0
28.1
Ti-C-H2-NO2
6.7
5.6
6.3
29.1
Ti-C-H2-NH2
9.4
7.4
8.5
31.7
Substrate and functionalised substrate samples
Key drivers for undertaking this investigation were the very low cost and very low mineral content of this lignite as a potential carbon precursor. However, the CO2 adsorption capacities of the highly mesoporous products described above were not as high as for carbons prepared from other materials or by other routes. Carbons with higher pore volumes, proportionately more microporosity and/or with more extensive modification by N gave better results, so a key challenge remains in the preparation of improved carbons with these features from Victorian lignite. Acknowledgements. We acknowledge the funding provided by the Australian Government through its CRC Program to support this research project via the CRC for Greenhouse Gas Technologies and the CRC for Clean Power from Lignite. References. 1. 2. 3. 4. 5.
Guy, P.J. and G.J. Perry, Victorian brown coal as a source of industrial carbons: a review. Fuel, 1992. 71(10): p. 1083-1086. Chaffee, A.L., et al., CO2 capture by adsorption: Materials and process development. International Journal of Greenhouse Gas Control, 2007. 1(1): p. 11-18. Burchell, T.D., et al, A novel process and material for the separation of carbon dioxide and hydrogen sulfide gas mixtures. Carbon, 1997, 35 (9), 1279-1294. Yoshizawa, N., et al, Coal-based activated carbons prepared with organometallics and their mesoporous structure. Energy & Fuels, 1997, 11(2), 327-330. Luguya, R., et al, Synthesis and reactions of meso-(p-nitrophenyl)porphyrins. Tetrahedron, 2004, 60 (12), 27572763.
PHYSICAL PROPERTIES AND REACTIVITIES OF MG-BASED DRY REGENERABLE CO2 SORBENTS PREPARED BY SPRAY-DRYING METHOD Jeom-In Baek, Tae Hyoung Eom, Joong Beom Lee, Won Sik Jeon, Chong Kul Ryu* KEPCO Research Institute 305-380 Munji-ro 65, Yuseong-gu, Daejeon, Republic of Korea Tel: 82-42-865-5250, Fax: 82-42-865-5202, e-mail: [email protected]
Abstract Integrated gasification combined cycle (IGCC) is evaluated as a cost-efficient power generation system when CO2 capture is considered together. CO2 capture at a warm temperature and a high pressure will improve the thermal efficiency of IGCC system compared with current low temperature CO2 capture process. In this work, we presented dry regenerable solid CO2 sorbents with high CO2 sorption capacity at warm temperatures and a high pressure and excellent physical properties suitable for fluidized–bed process applications. Mg-based CO2 sorbents were prepared using spray-drying technique. The physical properties of the spray-dried sorbents were characterized in terms of shape, particle size, packing density, and attrition resistance etc. The reactivity of the sorbents was measured with a thermogravimetric analyzer (TGA) and with a bubbling fluidized-bed reactor using simulated synthesis gas at 20 bar. TGA weight gain measured at the temperatures of 200 and 400 oC for absorption and regeneration, respectively, was about 11 wt%. CO2 sorption capacity measured in a bubbling fluidized-bed reactor at the same condition was around 9 wt%. Introduction The increase of natural disasters caused by the global warming has strengthened the cooperative effort to reduce the carbon dioxide emission. The largest single source of carbon dioxide emissions by human activities is coal power plants. It emits around one third of anthropogenic carbon dioxide emission. It is expected that the global electricity production will continuously increase for a few decades. Carbon capture and storage (CCS) in power plants is considered as one of the main options to mitigate carbon dioxide emission. Carbon dioxide capture technologies such as post-combustion capture, precombustion capture and oxy-fuel combustion are under development to reduce the carbon dioxide emission from power plants. Pre-combustion capture is evaluated to have less electricity cost penalty than post-combustion capture due to high concentration of carbon dioxide in the gas stream. A currently available commercial technology in pre-combustion capture is a wet scrubbing method, Selexol process, which uses a physical solvent, dimethyl ether of polyethylene glycol. The temperature of gas stream in this capture process should be lowered to around 40 oC for carbon dioxide capture. This cooling process
brings about a lot of cost and thermal efficiency loss. Another drawback of the wet scrubbing CO2 capture process is that the pressure of the separated CO2 is low, around 1.7 atm. Therefore, a great deal of energy is additionally required to compress the separated CO2 up to around 150 atm for storage. If the temperature of CO2 capture and pressure of the separated CO2 could be raised, the energy loss can be greatly reduced. To capture CO2 at warm temperatures and high pressures, several sorbent materials have been investigated in the literatures. Alkali promoted hydrotalcite-based sorbents showed breakthrough CO2 sorption capacity under 6.2 wt% at warm temperatures and high pressures (Ding and Alpay, 2000; Lee et al., 2007; Oliveira et al., 2008; Van Selow et al., 2009). Wang et al. (2008) reported 3.86.2 wt% CO2 sorption capacity for mixed oxides derived from hydrotalcite-like compounds. Siriwardane and Stevens (2009) suggested a magnesium hydroxide-based sorbents with CO2 capture capacity more than 13.2 wt% at 200 oC and high pressure over 150 psig. KEPCO Research Institute also has been developing solid sorbents which can capture CO2 at warm temperatures and high pressures in a circulating fluidized-bed process. Baek et al. (2009) presented MgO based CO2 sorbents with TGA weight gain of around 9 wt% at 200 oC and 20 bar. In this study, new spray-dried magnesium-based sorbents were designed to improve the performances of previously developed sorbents. The physical properties and CO2 sorption capacities of new sorbents were investigated to evaluate their performances. Experimental Section Sorbent Preparation: Three kinds of magnesium compounds, MgO, Mg(OH)2, and 4MgCO3Mg(OH)24H2O were well-mixed with support and multi-binder matrices in a pure water. To increase the reactivity alkali-based promoter was added. All raw materials were selected form commercially available products in a powder form. The mixed slurry was comminuted with ball mill to control particle size and make homogeneous colloidal slurry. The homogenized slurry was spray-dried to form spherical particle, i.e., green body, in an 8 kg batch scale. The green body was calcined at 550 C in a muffle oven for 5 h after pre-drying at 120 C overnight. SMO, SMH, and SMC denote the CO2 sorbents prepared using MgO, Mg(OH)2, and MgCO3, respectively. Physical Characterization: The shape of calcined sorbents was obtained using scanning electron microscope (SEM). Average particle size of the sample was measured according to the American Society for Testing and Materials (ASTM) E-11. Packing density of the sample was determined using the Autotap instrument (Quantachrome) proposed in ASTM D 4164-88 to measure the tapped density of formed catalyst. Brunauer-Emmett-Teller (BET) surface area of the sample was determined by N2 physisorption. Pore volume and porosity was obtained with Hg porosimetry. Attrition resistance of the calcined oxygen carrier was measured with a modified three-hole air-jet attrition tester based on the ASTM D 5757-95. The attrition resistance was determined at 10 standard liters per minute (slpm) over 5 h as described in the ASTM method. The attrition index (AI) is the percent fines generated over 5 h. The fines are particles collected at the thimble after 5 h from the start, which was attached to the gas outlet.
AI = [total fine collected for 5 h/amount of initial sample (50 g)] x 100%
(1)
A lower AI indicates a better attrition resistance of the bulk particles. The AI of fresh Akzo fluid catalytic cracking (FCC) catalyst, which was used as a reference, was 22.5%, under the same measurement condition. The targeted value of AI in this study was below 30%. It would be acceptable in a fluidized-bed CO2 capture process with transport reactor type. TGA Chemical Reactivity: CO2 absorption characteristics of the prepared sorbents were evaluated using a thermal gravimetric analyzer (Thermo Cahn, TherMax500) that can be used for high pressure tests. CO2 absorption was conducted at 200 C, and regeneration at 200 or 400 C. Reactor pressure was maintained at 20 bar during the tests. The simulated synthesis gas of which compositions were 36% CO2, 48% H2, 6% CO and 10% H2O was used for absorption. Regeneration gas compositions were 90% N2 and 10% H2O. The amount of samples used for tests were 50 mg scale. A total flow rate was 2 slpm. The overall schematic diagram of TGA system was shown in Figure 1. To prevent water vapor condensation, electric line heaters were installed along the gas line and gas stream was heated above saturation point. The TGA weight gain was calculated based on the stabilized weight at 200 C with flowing simulated synthesis gas in which CO2 was substituted by N2. Data acquisition
Microbalance E/B
Sytem outlet Vent
Air for pressure control
Back pressure regulator
Balance purge gas
N2 TGA furnace H2O
N2
N2
CO2 H2 CO
Vent
Furnace gas Thermocouple
H2O vapor N2
Figure 1. Schematic diagram of high pressure TGA system. To obtain more accurate CO2 sorption capacity of the prepared sorbents, cyclic CO2 absorption and regeneration tests were carried out in a batch type bubbling fluidized-bed reactor (0.038 m i.d. and 1 m height). The experimental temperatures were 200 C for absorption and 200 or 400 C for regeneration. System pressure was constantly maintained at 20 bar during the test. The compositions of simulated synthesis gas for absorption were 30% CO2, 47% H2, 8% CO and 15% H2O. Regeneration gas compositions were 85% N2
and 15% H2O. Around 350 g of sorbents was loaded in the reactor (a bed height of around 0.4 m). A total flow rate was 8 slpm. The minimum flow rate for fluidization was around 0.3–0.4 slpm. The gas composition change as a function of reaction time was measured continuously with gas analyzer. The CO2 sorption capacity was calculated by the integration of the CO2 breakthrough curve. Results and Discussion Physical Properties: The physical properties of the spray-dried CO2 sorbents were presented in Table 1. The prepared sorbents had a sphere-like shape as shown in Figure 2. The average particle size, size distribution, and bulk density were 135±10 m, 57303 m, and 0.7±1 g/ml, respectively. These physical properties were suitable for fluidized-bed applications. The BET surface area of SMC was higher than that of SMO and SMH. CO2 desorption during calcination might result in higher surface area of SMC. The attrition resistance is an important factor which decides the applicability of the sorbents to a fluidized-bed process. Low attrition resistance causes lots of sorbent loss, and consequently higher amount of the sorbents should be made up. The attrition resistance of SMH was lower than commercial FCC catalyst. Therefore, it was excluded from the reactivity tests in a TGA and a bubbling fluidized-bed reactor. Table 1. Physical properties of the CO2 sorbents calcined at 550 °C SMO SMH Avg. particle size / µm 138 127 Size distribution / µm 57 – 303 57 – 303 Bulk density / (g/ml) 0.77 0.60 BET / (m2/g) 34.5 26.2 Hg porosity / % 29.3 29.0 Hg pore volume / (ml/g) 0.41 0.61 Attrition index (AI) / % 11.9 32.9
SMC 144 57 – 303 0.72 66.6 30.7 0.45 9.5
Figure 2. SEM images of spray-dried CO2 sorbents calcined at 550 °C. Chemical Reactivity: The weight increase of SMO and SMC sorbents by CO2 capture in a TGA system were compared in Figure 3. TGA weight gain for both sorbents was higher than 10 wt% for absorption at 200 C and regeneration at 400 C, under 20 bar. SMO
showed slightly higher TGA weight gain than SMC despite less surface area. This can be ascribed to the higher MgO content of SMO than that of SMC after calcination. The narrow temperature window between absorption and regeneration can reduce energy penalty of CO2 capture process. Therefore, absorption and regeneration at the same temperature of 200 C was conducted to investigate how much sorption capacity changes. The TGA weight gain for regeneration at 200 C was around half of that for regeneration at 400 C. This indicates that the prepared sorbents can be used at a wide range of regeneration temperature between 200 and 400 C although there is the reduction of CO2 sorption capacity to some extent. The breakthrough curves of CO2 obtained from bubbling fluidized-bed tests were shown in Figure 4. The supplied CO2 was completely removed for around 5 minutes and then the concentration of CO2 sharply increased. The CO2 sorption capacities calculated from the breakthrough curves were 9.6 and 8.7 wt% for SMO and SMC, respectively. Although this sorption capacity is a little lower than that (13.2 wt%) of a magnesium hydroxide-based sorbents suggested by Siriwardane and Stevens (2009), it is much higher than that that (under 6.2 wt%) of alkali promoted hydrotalcite-based sorbents reported in many literatures. In addition to high CO2 sorption capacity, the outstanding feature of our sorbents is that they have the physical properties suitable for fluidized-bed applications .
TGA weight gain / wt%
20 (Aabsorption / Regeneation) 200 / 400 oC 200 / 200 oC
16
12
8
4
0 SMO
SMC
Figure 3. CO2 capture data in a TGA (at 20 bar) Summary The regenerable Mg-based solid sorbents for pre-combustion CO2 capture at a warm temperatures were prepared using three kinds of magnesium compounds, MgO, Mg(OH)2, and 4MgCO3Mg(OH)24H2O by spray-drying method. They had physical properties suitable for fluidized-bed applications except for the sorbent prepared with Mg(OH)2 which showed low attrition resistance. The sorbents prepared using MgO and 4MgCO3Mg(OH)24H2O showed high CO2 sorption capacity (9.6 wt%) at the reaction temperatures of 200 C for absorption and 400 C for regeneration under 20 bar. Although
CO2 concentration (C/C0)
1.0
0.8
0.6
SMO (H2O 15%)
0.4
SMC (H2O 15%)
0.2
0.0 0
5
10
15
20
25
30
Time / min
Figure 4. CO2 Capture data in a bubbling fluidized-bed reactor (absorption at 200 C and 20 bar, regeneration at 400 C and 20 bar) their CO2 sorption capacity decreased to around half when regeneration temperature was lowered to 200 C, it was still comparable to that of alkali promoted hydrotalcite-based sorbents. Detailed analysis on the CO2 sorption mechanism and reactivity will be investigated in our future work.
Acknowledgement This work was supported by Power Generation & Electricity Delivery R&D program (20101020100090) under the Ministry of Knowledge Economy, Republic of Korea. The authors also thank the KEPCO and the four fossil fuel power companies (Korea South-East, Western, Midland, and Southern Power Company, Ltds) for their support References Baek, J.-I.; Eom, T. H.; Lee, J. B.; Jeon, W. S.; Kim, J. W.; Ryu, C. K., “Characterization of dry regenerable sorbents for CO2 capture from syngas”, 26th Annual International Pittsburgh Coal Conference, September 2023, 2009. Ding, Y.; Alpay, E., “Equilibria and kinetics of CO2 adsorption on hydrotalcite adsorbent”, Chem. Eng. Sci. 2000, vol. 55, 34613474. Kee, K. B.; Verdooren, A.; Caram, H. S.; Sircar, S., “Chemisorption of carbon dioxide on potassium-carbonate-promoted hydrotalcite”, J. Colloid Interf. Sci. 2007, vol. 308, 3039. Oliveira, E. L. G.; Grande, C. A.; Rodrigues, A. E., “CO2 sorption on hydrotalcite and alkali-modified (K and Cs) hydrotalcite at high temperatures”, Sep. Purif. Technol.
2008, vol. 62, 137147. Siriwardane, R. V.; Stevens, Jr., R. W., “ Novel regenerable magnesium hydroxide sorbents for CO2 capture at warm gas temperatures”, Ind. Eng. Chem. Res. 2009, vol. 48, 21352141. Van Selow, E. R.; Cobden, P. D.; Van den Brink, R. W.; Hufton, J. R.; Wright A., “Performance of sorption-enhanced water-gas shift as a pre-combustion CO2 capture technology”, Energy Procedia, 2009, 689696. Wang, X. P.; Yu, J. J.; Cheng, J.; Hao, Z. P.; Xu, Z. P., “High temperature adsorption of carbon dioxide on mixed oxides derived from hydrotalcite-like compounds”, Environ. Sci. Technol. 2008, vol. 42, 614618.
2010 International Pittsburgh Coal Conference, Istanbul, Turkey, October 11 – 14, 2010
CARBONACEOUS PARTICLES FROM THE INCOMPLETE COMBUSTION Martina Havelcová, Ivana Sýkorová, Hana Trejtnarová, Alexandr Šulc Institute of Rock Structure and Mechanics AS CR,v.v.i., V Holešovičkách 41, 182 09 Prague, Czech Republic,
Introduction The abundance and composition of emitted products are primarily related to the fuel composition and to the characteristics of the combustion processes. Fuels used in combustion are usually fossil fuels or living or dead vegetation and they are mainly of a carbonaceous nature. The chemical composition of combustion plumes is a complex and as a result of incomplete combustion processes particles enter the atmosphere as well. As combustion whether from motor vehicles, industrial flames, or biomass burning is ubiquitous, carbonaceous particles are found virtually everywhere. Black carbon (BC) is general term used to describe carbonaceous residues produced by incomplete combustion of organic matter. The concept of BC refers to particles which are black when measured by optical techniques. In this work we characterized the chemical and petrological changes associated with the formation of char from coal, peat, wood and wheat. Since the formation of BC depends on several factors, we prepared series of carbonaceous residues in two environments. The combination of non-destructive (optical microscopy) and destructive (gas chromatography and pyrolysis-gas chromatography/mass spectrometry) analytical techniques was used for the samples characterization. The chemothermal oxidation method was employed to determine the BC content in the samples.
Samples and methods Bituminous coal (from the Czech part of the Upper Silesian Basin), lignite (from the Sokolov Basin), peat (from the Holocene Krásno peat bog in the Krušné Hory Mountains in western Bohemia), oak wood, spruce wood and wheat corn were used to prepare series of carbonaceous residues. Two environments were chosen: the first one with low access of air and the second one in nitrogen atmosphere. Four temperatures were chosen for the preparing of the carbonaceous residues: 350, 450, 600 and 800°C. CHNS analyses were performed and losses of weight were monitored during thermal action. Total organic carbon (TOC) was obtained by carbon elemental analysis after removal of inorganic carbonates by sample acidification and heating up to 80 °C. The chemothermal oxidation method (Gustafsson et al., 2001) was employed to determine the BC content in the samples: 3–5mg batch of a sample was weighed into a silver capsule and organic carbon was removed by thermal treatment at 375°C for 24 h in the presence of air. Inorganic carbonates were removed by acidification and heating up to 80°C. The elemental composition, TOC and BC was determined on a CHNS/O microanalyser Flash FA 1112 Thermo Finnigan.
2010 International Pittsburgh Coal Conference, Istanbul, Turkey, October 11 – 14, 2010
Identification and characterization of optical properties and morphology of organic matter in natural and artificial prepared samples were based on optical microspectrophotometry of polished blocks in reflected, polarized and fluorescence in accordance with the methodology published by Sýkorová et al. (2009) and Taylor et al. (1998). The reflectance of vitrinite and huminite in coal, peat, biomass and char grain polished sections was measured using a Nikon Eclipse ME 650 microscope (immersion objective, 40x magnification, oil immersion with the refractive index n = 1.515) with a LUCIA VITRINITE image analysis system for measurement of vitrinite reflectance. The maceral composition of the vitrinite, liptinite, inertinite group and composition of mineral matter in coal and biomass samples including optical anisotropy and morphology of chars were assayed using a Opton-Zeiss UMSP 30 Petro microscope (reflected monochromatic light, wavelength λ= 546 nm, immersion objective, 40x magnification, oil immersion with refractive index n = 1.518) and point-counter ELTINOR for precisely determination of studied petrographic components according to ISO 7404 (2009), Taylor et al. (1998) and Kwiecinska and Petersen (2004). The special attention was paid to petrographic characterisation of combustion residues. We determined cenosphere, network, massive and fragment texture of chars according to classifications and its modifications by Tsai and Scaroni (1987), Bailey et al. (1990), Shibaoka et al.(1996), Gentzis and Chambers (1993), Rosenberg et al. (1996), Hower et al. (2005), Suárez-Ruiz et al.(2007). The optical anisotropic texture was evaluated after system for coke (Taylor et al. 1998). Chemical analyses were used for detailed characteristics of organic compounds in samples. Gas chromatography with mass spectroscopy (GC/MS) and pyrolysis gas chromatography (Py-GC/MS) provide relatively detailed information on compounds, including degradation products (Sýkorová et al., 2009). GC/MS analysis is used for detecting and identifying substances soluble in organic solvents (Fernandes and Brooks, 2003) and thus results depends on extractability of samples tested. Powdered samples were Soxhlet extracted with dichloromethane. The extracts were analyzed by gas chromatography using a Thermo DSQ II–Trace Ultra GC equipped with a CP5 capillary column (30 m×0.25 mm×25 μm). After splitless injection, a temperature program was started from 40 to 300 °C at 5 °C min−1, with helium as a carrier gas. Greater emphasis was placed on Py-GC/MS method that allows rapid heating of the sample and its thermal decomposition to obtain compounds, and those directly identify by GC/MS. Py-GC/MS analyses were performed using CDS Pyroprobe 5150 directly connected to Thermo DSQ II–Trace Ultra GC. Approximately 5 mg of a sample was pyrolyzed at temperature 300, 400, 500, 600, 700, and 1000 °C for 20 s. The pyrolysis data indicated that pyrolysis temperature of 500°C would be appropriate for pyrolysis studies. The Py-GC interface was kept at 300 °C. The CP5 capillary column was used for separation of pyrolysis products. The temperature program was set from 35 to 300 °C at 6 °C min−1, with helium as carrier gas. Mass spectra were recorded in the electron impact mode (70 eV) in the 40–500 m/z range. Identification of compounds was based on a comparison of spectra with the NIST mass spectral library and data from literature.
Results The chemical features of the raw materials are in the Table 1. The increase in the temperature led to the forming solid residues that are different chemically and structurally from the natural materials. In the nitrogen environment the carbon content increased while the oxygen and hydrogen content decreased, indicating an increasing degree of carbonisation of chars. In the environment with low access of air were not the changes in composition so simple. The content of carbon slightly increased but at 450°C decreased and at 600°C again increased for bituminous coal, lignite and peat. The content of carbon
2010 International Pittsburgh Coal Conference, Istanbul, Turkey, October 11 – 14, 2010
slightly increased for wheat and decreased at 600°C. For the wood samples declined the carbon content from the temperature 350°C till value below 1 % at 800°C. %Wa
%Ad
%Ctd
%Hd
%Nd
%Std
%Od
%BC
BITUMINOUS COAL 0.61
5.54
81.08
4.98
1.01
0.28
7.10
0.15
LIGNITE
7.81
25.30
52.48
4.52
0.72
1.40
16.63
0.32
PEAT
11.39
3.14
54.69
6.38
1.10
0.23
34.47
0.12
OAK
7.14
0.33
49.78
5.06
0.24
0.27
44.09
0.00
SPRUCE
7.30
0.68
49.15
6.14
0.12
0.65
43.26
0.00
WHEAT
1.70
0.80
41.15
6.66
2.05
0.00
49.34
0.00
Table 1. Chemical features of the raw materials.
The highest content of BC after thermal alteration – 3.29 % - was found in bituminous coal sample at 800°C in the environment with low access of air. For all samples was found out that thermal treatment was reliable for increase of BC content but higher contents of BC were determined in the environment with low access of air. Inertinite rich bituminous coal from the Czech part of the Upper Silesian Basin contains 33 vol.% of vitrinite with random reflectance 1.09 %Rr, 11.3 vol.% liptinite and 5.6 vol.% of minerals. Coal from the Sokolov Basin is a lignite C (%Rr = 0.39) and consists of 46.6 vol.% huminite (mainly ulminite and densinite), 19.1 vol.% liptinite, 3.8 vol.% inertinite and 30.5 vol.% minerals, particularly kaolinite, siderite and pyrite. Peat sample originates from the Holocene Krásno peat bog in the Krušné Hory Mountains in western Bohemia. It represents a compressed humified and partly gelified sphagnum and sedge material with low ash with predominance of attrinite, low reflected densinite and texto-ulminite (%Rr = 0.22). Cellular texture of tissues was well preserved in samples of oak wood, spruce wood rich in resin and in wheat corn. They contain more than 90 vol. % of textinite, up to 5 vol.% of corpohuminite and resinite, and random reflectance is very low, up to 0.10% Rr. The reflectance values of our char samples rise with increasing temperature and change in relation to feed material. The rapidly rate of increase in reflectance was found in chars prepared in air atmosphere. These chars with reflectance values greater than 2.0% Rr form at temperature 450°C, with reflectance over 4.0% Rr at 600°C. Reflectance values of chars prepared in nitrogen environment climb slowly: i) up to 1% Rr at 350°C, ii) over 2% Rr at the temperature 450°C with exception lower reflectance of chars from peat and wheat, iii) above 3.0% at the temperature 600°C with exception of lower reflectance of chars from lignite and wheat. Reflectance values above 5.0% Rr were measured on char walls prepared at 800°C in both experimental conditions. Char from bituminous coal is porous anisotropic matrix which constitutes inclusions of un-fused and partly fused fusinite, semifusinite, macrinite and minerals. Char from lignite is formed by isotropic devolatilized matrix full of network and massive texture with numerous mineral inclusions. Charring wood is isotropic porous matrix of networks texture. Massive texture was found in char from wheat corn. Solvent-extractable organic matter was very low for all chars (up to 1%). All the prepared chars contained measurable quantities of PAHs in varying concentrations that were depending on carbonization temperature and raw material. Generally, the samples prepared in the environment with the low access of
2010 International Pittsburgh Coal Conference, Istanbul, Turkey, October 11 – 14, 2010
air contained lower PAHs concentrations than the prepared in the nitrogen environment and higher for coals and wheat than for woods and peat. The samples, prepared at temperature 350°C, retained the chemical properties of the original biomass. With the temperature less thermally stable compounds (alkanes, naphthalenes) were destroyed or/and evaporated. The char from bituminous coal and lignite had the highest proportion of 4-6 ring polycyclic aromatic hydrocarbons (PAHs). Pyrolysis-GC/MS provided fingerprint of the prepared samples with the point of view on organic compounds possibly connected with BC production. The data revealed that the pyrolytic products were dominated in aromatic compounds, followed by smaller amounts of aliphatic compounds. O–containing pyrolytic products were also present in all chars. The results confirmed that the materials were composed of aromatic structures that were linked by aliphatic chains and O-containing group (Song and Peng, 2009). The samples prepared in the nitrogen environment dominated by 3-5 ring compounds. The peat and wheat samples had higher content of N-containing pyrolytic products at lower temperatures that indicated the presence of proteins in these samples. Several furan derivatives were also detected particularly in oak and spruce. This is due the presence of cellulose in woods. The presence of typical lignin markers (methoxyphenols, methoxybenzenes ) pointed at the presence of lignin units in oak, spruce and peat. In the samples prepared with low access of air were found mainly 1-3 ring hydrocarbons. Ncontaining pyrolytic products were in higher content in the coals and peat than in the woods and wheat corn. Most of the compounds have been reported as BC pyrolysis products, and anthracene, phenanthrene, fluoranthene and pyrene have been reported as a strong evidence of charred organic matter (Rumpel et al., 2007; Kaal et al., 2008, 2009). Aromatic nitriles have been reported to originate from BC, and their large amounts indicate the presence of pyrogenic carbon. The nitriles may have been formed from amides, amines, quarternary or pyridinic nitrogen in polyaromatics, possibly abundant in severely charred biomass (Kaal et al., 2008, 2009). Benzonitrile was found also in samples that contain low amounts of nitrogen compounds and similar like Song and Peng, 2009, we can assume that nitrogen could be introduced into the samples from the air. In our samples benzonitrile was present in all samples and the highest relative amount was found in the bituminous coal sample prepared at 800°C in the environment with the access of air. This is in good agreement with the content of BC determined by the chemothermal oxidation method.
Conclusions The relationships between both temperatures and low air/nitrogen atmosphere during char formation and reflectance values of chars have been established. The reflectance values and BC content increase with increasing temperature. The maximum values of BC were found at temperature 800°C. The rate of these changes depends on the material and experimental conditions (air or nitrogen). Our result shows that combustion and pyrolyse at temperatures from 350°C to 800°C of bituminous coal and lignite is the dominant source of BC content (over 0.30 %). On the other hand chars produced from peat, spruce wood and wheat corn reached the BC content mostly at temperature 800°C. GC/MS and Py-GC/MS data revealed that the chars were dominated in aromatic compounds. The PAH contents varied and were depending on carbonization temperature and raw material. The data suggest that the presence of benzonitrile has relation to the BC content.
2010 International Pittsburgh Coal Conference, Istanbul, Turkey, October 11 – 14, 2010
Acknowledgements: The research was funded by the Grant Agency of the Academy of Sciences of Czech Republic (project No. IAA300460804). References Bailey, J.G., Tate, A., Diessel, C.F.K., Wall, T.F., 1990. A char morphology system with applications to coal combustion. Fuel 69, 225-239. Fernandes, M.B., Brooks, P., 2003. Characterization of carbonaceous combustion residues. II. Nonpolar organic compounds. Chemosphere 53, 447-458. Gentzis, T., Chambers, A., 1993. A microscopic study of the combustion residues of subbituminous an bituminous coals from Alberta. International Journal of Coal Geology 24, 245-257. Gustafsson, Ö., Bucheli, T.D., Kukulska, Z., Andersson, M., Largeau, C., Rouzaud, J.-N., Reddy, Ch.M., Eglinton, T.I., 2001. Evaluation of a protocol for the quantification of black carbon in sediments. Global Biogeochemical Cycles 15, 881–890. Hower, J., Suárez-Ruiz, I., Mastalerz, M., 2005. An approach toward a combined scheme for the petrographic classification of fly ash: revision and classification. Energy & Fuels 19, 653-655. Kaal, J., Cortizas, A.M., Nierop, K.G.J., 2009. Characterization of aged charcoal using a coil probe pyrolysis-GC/MS method optimised for black carbon. J. of Analytical and Applied Pyrolysis 85, 408–416. Kaal, J., Martínez-Cortizas, A., Nierop, K.G.J., Buurman, P., 2008. A detailed pyrolysis-GC/MS analysis of a black carbon-rich acidic colluvial soil (Atlantic ranker) from NW Spain. Applied Geochemistry 23, 2395–2405. Kwiecinska, B., Petersen, H.I., 2004. Graphite, semigraphite, natural coke, and natural char classification – ICCP system. International Journal of Coal Geology 67, 99-116. Rosenberg, P., Petersen, H.I., Thomsen, E., 1996. Combustion char morphology related to combustion temperature and coal petrography. Fuel 75, 1071-1082. Rumpel, C., González-Pérez, J.A., Bardoux, G., Largeau, C., Gonzalez-Vila, F.J., Valentin, C., 2007. Composition and reactivity of morphologically distinct charred materials left after slash-and-burn practices in agricultural tropical soils. Organic Geochemistry 38, 911–920. Shibaoka, M., Ohtsuka, Y., Wornat, M.J., Thomas, C.G., 1996. Investigation of brown coal gasification residues using light microscopy. Fuel 75, 775-779. Song, J., Peng, P., 2010. Characterisation of black carbon materials by pyrolysis–gas chromatography– mass spectrometry. Journal of Analytical and Applied Pyrolysis 87, 129-137. Suárez-Ruiz, I., Hower, J.C., Thomas, G.A., 2007. Hg and Se capture and fly ash carbons from combustion of complex pulverized feed blends mainly of anthracitic coal rank in Spanish power plants. Energy & Fuels 21, 59-70. Sýkorová, I., Havelcová, M., Trejtnarová, H., Matysová, P., Vašíček, M., Kříbek, B., Suchý, V., Kotlík, B., 2009. Characterization of organic matter in dusts and fluvial sediments from exposed areas of downtown Prague, Czech Republic. International Journal of Coal Geology 80, 69-86. Taylor, G.H., Teichmüller, M., Davis, A., Diessel, C.F.K., Petrology. Gebrüder Borntraeger, Berlin-Stuttgart, 704 pp.
Littke, R., Robert, P.,
1998. Organic
Tsai, C.Y., Scaroni, A.V., 1987. The structural changes of bituminous coal particles during the initial stages of pulverized-coal combustion. Fuel 66, 200-206.
Trace Element Partitioning and Leaching in Solids Derived from Gasification of Australian Coals Alexander Ilyshechkin1,1, Daniel Roberts1, David Harris1, Ken Riley2 1 CSIRO Energy Technology, PO Box 883, Kenmore, QLD, 4069 AUSTRALIA 2 CSIRO Energy Technology, PMB 7, Bangor, NSW, 2234 AUSTRALIA
Abstract Trace element concentrations vary between coals from ppb to ppm levels and can depend on the rank of the coal and its geological origins. During gasification, some of the trace elements are volatilised at high temperatures and may condense and deposit in cooler downstream parts of the system or in quench water streams. Some species may appear in condensed phases such as slag or flyash. Changes in the trace element concentrations in the slag and flyash from that of the parent coal are expected due to the reactions occurring at high temperatures and the different chemical activity of the trace element phases in the slag, flyash and syngas. Four Australian coals were used in a gasification test program conducted in the Siemens 5 MWth gasification test facility. Solid samples were collected from different points in the gasification process during each test. Compositions of these samples were analysed and the distribution of trace elements was studied. It was found that the elements can be classified as follows, according to their tendency to appear in the slag and fly ash: - Non-volatile and no partitioning between slag and flyash: Nb, U - Partitioned between slag and fly ash: Cu, W, Mo, Cd, Bi, Zn, Sn, Sb - Partially volatile and depleted from either slag or fly ash: Be, Th, Sc, Y, Li, Mn, Ni, Sr, Ba - Highly volatile (i.e. were not observed in either slag or fly ash): As, Se, B, Hg, F, Pb, V. Comparison of these experimental results with equilibrium calculations of trace element appearance in the condensed phases suggests that the modelling approach is suitable only for certain elements. For several of the trace elements of significance in this study, kinetic factors have to be considered in conjunction with thermodynamic modelling. The leaching behaviour of the trace elements in the slag was also studied. This work shows very low leachability for most of the trace elements except Zn and Sb, which, due to their relatively high volatility, reported in the slag samples in very low concentrations.
1
Corresponding author. Email: [email protected] Ph +61 7 3327 4187 Fax +61 7 3327 4455
1
1. Introduction In addition to the main mineral components in the coal (Si, Al, Fe, Ca, Mg, Ti, Mn, Na, K, P) and the minor elements such as S and Cl, there is a wide range of trace element species present at concentrations that vary strongly across coal types. The behaviour of these trace elements during gasification is of interest as their presence in solid waste streams from gasification processes can impact on disposal or utilisation strategies. It has been reported [1] that compositional changes of major elements occur in the slag and process water solids (mainly consisting of fly ash) during coal gasification. Therefore, it is likely that changes in the trace elements concentrations in the slag and flyash from the parent coal will also take place. Indeed, some of the trace elements will be volatilised at the high temperatures reached during gasification [2], or may be in liquid form. They have different chemical reactivity with the molten oxides in the slag formed on the wall of the gasifier or with molten flyash. Thus the trace elements partition between the slag, flyash and syngas and this depends on the occurrence of the elements in the coal and under gasification conditions [3]. Various aspects of the mobilisation of trace elements during gasification have been studied [2-11]. Most of the studies are focused on the prediction of the phases in which trace elements appear. Ideally, techniques for predicting trace element behaviour would model the phase transformation of the trace elements, including their volatilisation during and after gasification as a chemical kinetic process. However, this approach is not practicable due to a lack of the necessary fundamental kinetic information and, therefore, most studies have adopted a thermodynamic equilibrium approach as a best approximation. Analysis of the distribution of trace elements in the slag, flyash, and syngas is therefore useful for assessing the applicability of the thermodynamic phase equilibria approach for a particular trace element, and for determining future research needs. The scope of this study is to analyse the distribution of the trace elements in gasification solid wastes, such as slags and fly ash removed in the quench water, thereby providing some input for further development of trace element modelling under gasification conditions. Trapping of trace elements in the glassy phase has considerable consequences for the handling of solid wastes from the gasifier. In the present study, the behaviour of the trace elements in the slag was also investigated in terms of the elements leachability. 2. Experimental 2.1 Samples of slag and process water solids Solid samples of slag and fly ash were obtained from the gasification of four Australian coals: CRC701 is a low-rank (sub-bituminous) coal from Western Australia, CRC702 is a high volatile bituminous coal from the Hunter Valley in NSW, CRC703 is a semi-anthracite from the Bowen Basin in Queensland, and CRC704 is a high volatile bituminous coal from the Surat Basin in Queensland. The pilot scale test program [12, 13] involved a range of gasifier operating conditions in order that the performance of the coals could be compared under a consistent set of conditions, and
2
also that the effects of gasification conditions could be quantified for each coal in a comparable manner. Each test was performed under more than one set of steady state conditions (balance phases) during which the gasifier operated steadily—more details can be found elsewhere [12, 13]. Solids present in the discharged quench water were collected, and these are referred to as ‗process water solids‘ (PWS). Tapped slag samples were collected periodically throughout the trials, as batches using a lock-hopper system which removed slag from the bottom of the quench. PWS and slag samples from each balance phase were analysed for their chemical and physical characteristics. 2.2 Sample analysis The concentrations of the trace elements were determined by ICP-AES and ICP-MS with the exception of As, Se and Hg; these were determined using hydride generation/atomic fluorescence techniques and a cold vapour/-atomic fluorescence technique. Solid samples were either dissolved in mixed acids (including HF) or fused with lithium borate before dissolving in nitric acid. Liquid samples were acidified with nitric acid before measurement. Standards (for measurement) were prepared to match the matrix of the samples. Leaching tests were undertaken by leaching slags for three days in distilled water at a liquid:solid ratio of 20:1, consistent with previous work done in this area [14-16]. The leachates were then filtered through 0.22 um Millipore cellulose acetate membrane filters. An unacidified portion of the filtrate was analysed for pH and conductivity. Another portion of the filtrate was acidified with approximately 1% v/v AR grade concentrated nitric acid and the concentrations of trace elements determined in the leachate using ICP-AES and ICP-MS. 3. Results and discussion 3.1 Distribution of trace elements in the coal, slag and process water solids The concentrations of the trace elements in the coals and their corresponding slag and PWS (which consists mainly of flyash) is shown below in Figure 1–7. Based on these analyses, the trace elements can be categorised according to their appearance in the slag and flyash: 1. Non-volatile • Equally distributed between slag and fly ash • Same amount as in parent coal mineral matter 2. Partially-volatile, partially lost from slag and fly ash • Known to be partially volatile from the literature • Some have escaped in gas phase or water • Some amount remains in PWS and slag 3. Volatile, recaptured on fly ash • Known to be volatile or partially volatile from the literature • Present in PWS but at higher concentration than in the parent coal mineral matter 3
4. Highly volatile from slag and/or fly ash • Lost in gas or water • Low concentrations in PWS and/or slag
These categories can be further subdivided as per Table 1. Group Appearance 1.1
Experimental
Modelling
slag
slag Y
flyash Y
?
?
Y
N
N ? Y Y
Y ? N ?
Y Y ?
N Y ?
flyash
Directly passed to solid Nb, U slag/flyash Present in both phases, some loss Th, Sc, Y, Li, Mn, Ni, Sr, Ba Partially volatile: recaptured on Cu, W, Mo, Cd fly ash Partially volatile: concentrated in Cr, slag V Volatile, recaptured on fly ash Bi, Zn, Sn, Sb Highly volatile: depleted in slags As, Se only Highly volatile: significant loss B, Hg, from solid phases F, Pb Questionable: coal-specific Co, Be
2.1 2.2 2.3 3.1 4.1 4.2 5
Table 1: Distribution of trace elements in slag and PWS. “Y”- indicates that the behaviour is consistent with thermodynamic modelling, “N”- indicates that the behaviour is inconsistent with thermodynamic modelling, “?”- not clear. Arrows indicate how trace elements are re-distributed
These distributions of trace elements and discussed below, in the context of the groupings shown in Error! Reference source not found. and available thermodynamic modelling results. Group 1.1: Directly passed to solid slag/flyash Group 1.1 elements are not volatile, and have undergone no reaction during gasification. Only Nb and U can be classified in this group. Concentration of these in the coal mineral matter, slag and PWS are nearly the same in all samples (Figure 1). Some difference in the concentration is likely associated with natural variation of the element in the sample. U
Nb
Coal
80.00 70.00
14.00
40.00 30.00 20.00
Content, mg/kg
Content, mg/kg
60.00
Slag
PWS
16.00
PWS
50.00
Coal
18.00
Slag
12.00 10.00 8.00 6.00 4.00
10.00
2.00 0.00
0.00 CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
CRC703
CRC704
Figure 1 Concentration of Nb and U in coal mineral matter, and the corresponding slag and PWS
4
Group 2.1 Present in both phases, some loss The next group of trace elements (Figure 2) are partially vaporised, and present in both the slag and the PWS at reduced concentrations. Generally, their concentration in the slag and PWS is about 10-50% lower than in the parent coal mineral matter. Sr
Ba
0.90
Coal
0.70
PWS
0.60
Slag
0.50 0.40 0.30
0.45
PWS
0.40
Content, wt.%
Content, wt.%
0.80
Coal
0.50
0.20
Slag
0.35 0.30 0.25 0.20 0.15 0.10
0.10
0.05
0.00
0.00
CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
Coal
800.00
Coal
PWS
700.00
PWS
Slag
0.08 0.06 0.04
600.00
Content, mg/kg
Content, wt.%
0.10
0.02
Slag
500.00 400.00 300.00 200.00 100.00 0.00
0.00 CRC701
CRC702
CRC703
CRC701
CRC704
CRC702
CRC703
CRC704
Sc
Li 600.00
Coal
140.00
500.00
PWS
120.00
400.00
Slag
300.00 200.00
Content, mg/kg
Content, mg/kg
CRC704
Ni
Mn 0.12
100.00
100.00 80.00
Coal PWS Slag
60.00 40.00 20.00
0.00
0.00
CRC701
CRC702
CRC703
Th 70.00 60.00
CRC704
CRC701
200.00
PWS
180.00
40.00 30.00 20.00
CRC703
CRC704
Coal PWS
160.00
Slag
50.00
CRC702
Y
Coal
Content, mg/kg
Content, mg/kg
CRC703
Slag
140.00 120.00 100.00 80.00 60.00 40.00
10.00
20.00
0.00 CRC701
CRC702
CRC703
CRC704
0.00 CRC701
CRC702
CRC703
CRC704
Figure 2 Concentration of trace elements in coal mineral matter, PWS and slag: deficit in the slag and PWS
5
Manganese is known to be a low-volatile element [3], and nickel starts to vaporise at temperatures above 1100–1400°C [4-5]; however notable loss of nickel in slag has been reported at temperatures above 1300-1550°C [4-5]. At a temperature of 1800ºC 40% [5]-80% [4] of Ni is still concentrated in the condensed phases.
Group 2.2 (partially volatile) and Group3.1 (volatile), recaptured on fly ash These elements are those which have become volatile to some extent, but have been recaptured on the fly ash. Elements where a deficit in the slag is compensated by an increase in concentration in the PWS are shown in Figure 3. These elements are Cu, Mo, W, Cd, Bi, Zn, Sn and Sb. The last five are highly volatile at high temperatures and almost disappear from the slag and it is likely they are captured by the fly ash at lower temperatures downstream of the gasifier. Modelling of Cd species indicates the high volatility at gasification conditions [6]. However, Cd conservation in the slag indicates its moderate volatility in our experimental conditions. It is likely that for Cd species in the slag the residence time was not long enough to achieve equilibrium. Literature discussing the behaviour of copper suggests that it appears in the slag at temperatures above 1100 ºC, but starts to vaporise from the slag between 1200ºC [4] and1400 ºC [5], and can coexist in gaseous and condensed phases at temperatures up to 2000ºC [4, 5]. According to modelling results, Zn and Sn are mainly present in a gaseous form at temperatures above 1000 ºC [4] and 1300 ºC respectively, while Sb gradually increases in the gaseous phase in the temperature range of 1300-1900ºC [5].
6
Cu
Cd
350.00
Coal
300.00
Slag
200.00
Coal
6.00
PWS
250.00
Content, mg/kg
Content, mg/kg
7.00
150.00 100.00 50.00
PWS
5.00
Slag
4.00 3.00 2.00 1.00
0.00
0.00
CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
Coal
9.00
PWS
50.00
Slag
40.00 30.00 20.00
PWS
8.00 7.00
Content, mg/kg
Content, mg/kg
10.00
Coal
60.00
Slag
6.00 5.00 4.00 3.00 2.00
10.00
1.00 0.00
0.00 CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
CRC703
CRC704
Sn
Bi 12.00
Coal
25.00
10.00
PWS
20.00
8.00
Slag
6.00 4.00
Content, mg/kg
Content, mg/kg
CRC704
W
Mo 70.00
Coal PWS Slag
15.00 10.00 5.00
2.00 0.00
0.00 CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
Zn
CRC703
CRC704
Sb
1600.00
4.50
Coal
1400.00
PWS
1200.00
Slag
1000.00 800.00 600.00
Coal
4.00
PWS
3.50 Content, mg/kg
Content, mg/kg
CRC703
3.00
2.00 1.50
400.00
1.00
200.00
0.50
0.00
Slag
2.50
0.00
CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
CRC703
CRC704
Figure 3 Concentration of trace elements in coal mineral matter and redistributed between slag and PWS (PWS dominant)
Group 2.3 Partially-volatile, recaptured on slag
7
Only Cr has a higher concentration in the slag than in the PWS and the related coal minerals, as shown in Figure 4: this indicates a high solubility of Cr in the molten oxides at high temperatures. Thermodynamic modelling shows some losses of Cr in the condensed phase which agrees with Cr concentration in PWS, but contradicts with Cr appearance in some slags [6]. It is possible that this enhanced Cr concentration could be from materials used in the manufacture of gasifier wall materials—further work here is required. It is interesting to observe a very low concentration of V in the PWS, compared with the slag and the parent coal mineral matter. It seems that V species are captured by the molten oxides in the slag and became more chemically stable than in the fly ash. According to thermodynamic modelling [6], V2O3 is the most stable phase at temperatures higher than 500ºC. V
Cr 450.00
Coal
400.00
350.00
PWS
350.00 300.00
Slag
Content, mg/kg
Content, mg/kg
400.00
250.00 200.00 150.00
300.00
PWS
250.00
Slag
200.00 150.00
100.00
100.00
50.00
50.00
0.00
Coal
0.00
CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
CRC703
CRC704
Figure 4 Concentration of Cr and V in coal mineral matter and distributed between slag and PWS (slag enriched)
Group 4.1 Highly volatile and depleted only from the slag Two elements, As and Se, are expected from the literature to be highly volatile and are not present in the slag phase (Figure 5). As distinct from the Group 3 elements, which are also expected to be highly volatile (Bi, Zn, Sn and Sb), they are partially recaptured by fly ash. These elements are predicted to be at a very low concentration in the slag at temperatures above 1000ºC [4, 6] but would most likely condense on the heat transfer surfaces that cool the syngas [17], which may explain their concentration in the fly ash. Another reason for their presence in fly ash is that the ash residence time in the gasifier was insufficient for complete Se and As volatilisation. Se
As 120.00
Coal PWS
80.00
Slag
60.00 40.00
Content, mg/kg
Content, mg/kg
100.00
120.00
Coal
100.00
PWS
80.00
Slag
60.00 40.00 20.00
20.00
0.00
0.00 CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
CRC703
CRC704
Figure 5 Concentration of trace elements in coal mineral matter, PWS and slag: highly volatile in the slag
Group 4.2 Highly volatile: significant loss from solid phases
8
The last group consists of the trace elements that are highly-volatile and are not found in either the slag or PWS (Figure 6).The concentration of Hg, B, F and Pb are significantly lower in the PWS than originally in the parent coal mineral matter and even less in the slag: Hg and F nearly disappear from the slag. Mercury is always considered as a highly volatile element at gasification temperatures and is mostly emitted in the flue gas [6]. According to modelling, boron is highly volatile at temperatures above 1600-1700 ºC [4, 5], lead is highly volatile at temperatures above 1000 ºC [4, 6] and could be expected to be totally in the gas phase under gasifier conditions. Boron is a highly volatile element and theoretically forms gas phase species over a whole range of temperatures [6]. Appearance of mercury and boron in flyash indicates that it is likely that flyash did not have enough residence time to achieve phase equilibria in the system or they are partially captured during the cooling. Pb
120.00
PWS
100.00
Slag
80.00
1200.00
Coal
1000.00
PWS
800.00
Slag
Content, mg/kg
Content, mg/kg
B
Coal
60.00 40.00
600.00 400.00
20.00
200.00 0.00
0.00 CRC701
CRC702
CRC703
CRC701
CRC704
F
CRC703
CRC704
Hg
Coal
20000.00 18000.00
Slag
12000.00 10000.00 8000.00 6000.00 4000.00
7.00
PWS
6.00 Content, mg/kg
14000.00
Coal
8.00
PWS
16000.00 Content, mg/kg
CRC702
Slag
5.00 4.00 3.00 2.00 1.00
2000.00
0.00
0.00 CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
CRC703
CRC704
Figure 6 Concentration of trace elements in coal mineral matter, PWS and slag: highly volatile the slag or PWS
Group 5 Questionable: coal specific Trace element concentrations are different in bituminous, sub-bituminous, high volatile sub-bituminous and semi-anthracite coals. In general, this work has shown that the partitioning of the trace elements between the slag and flyash, especially those partially and highly-volatile, are very similar for all four coals. Exceptions are Co and Be, whose distributions in the coals and waste products seems to be dependant on the parent coal (Figure 7).
9
Co 180.00 160.00
120.00
Coal
40.00
PWS
35.00
Slag
30.00
Content, mg/kg
Content, mg/kg
140.00
Be
Coal
100.00 80.00 60.00
Slag
25.00 20.00 15.00
40.00
10.00
20.00
5.00
0.00
PWS
0.00
CRC701
CRC702
CRC703
CRC704
CRC701
CRC702
CRC703
CRC704
Figure 7 Concentration of Co and Be in coal mineral matter, PWS and slag.
Co is predicted to be captured most strongly by the slag [5]. Be is predicted to be in the gaseous form Be(OH)2 at temperatures above 1000 ºC and this becomes the dominant Be-species at temperatures above 1300 ºC, and it is expected to condense before gas cleaning systems [6]. On the other hand, thermodynamic modelling of Be species with other elements and phases under gasification conditions indicates that only about 2-10% of Be will be vaporised in the temperature range 1350-1900 ºC [5]. The results presented above indicate that the appearance of trace elements in the gasification solid residues does not always follow thermodynamic modelling predictions. Table 1 lists the trace elements where the thermodynamic approach is not applicable (label ―N‖) or is questionable. The relatively-long residence time of slag in a gasifier means that most of the trace elements in slag can be predicted by equilibrium modelling. Flyash, however, has a significantly shorter residence time and therefore phase equilibria may not be achieved, hence the trace elements appearance in flyash needs also to be linked with kinetic considerations. Such modelling requires further systematic experimental data to be use das the basis of ongoing development. 3.3 Slag leaching The overall leachability (or solubility in leachate) of elements from the slag is low. However, since the trace elements in the slag are different to those in the parent coal mineral matter, the leachability of each element has to be assessed in terms of the ratio to the total amount of the trace element in the slag before the leaching test. Data presented in Table 2 and Table 3 shows the absolute concentration of the trace elements in the leachates and the corresponding % of the trace element extracted from the slag. Slags s101-104 refer to slag from coals CRC701-704 respectively.
10
s101 S Mg Na K P Zn Sr Ba Ni Mn Co Li Mo B W Sc As Sb Cd Y Se Th U
s102
s103
s104
Element concentration in leachate , ppm, l:s = 20:1 0.3 0.4 0.5 1.5 0.29 0.18 0.22 0.22 0.21 0.13 0.37 0.42 0.2 0.1 0.2 0.2 0.04 0.003 0.21 0.20 0.009 0.014 0 0 0.03 0.024 0.012 0.014 0.014 0.012 0.001 0.014 0.016 0.035 0.006 0.014 0.04 0.004 0 0 0.003 0.003 0.001 0.002 Element concentration in leachate, ppb, l:s = 20:1 2.1 2.3 15.0 22.1 3.7 1.5 5.3 1.8 2.7 4.5 3.4 1.5 0.4 0.4 0.43 0.3 0.3 0.2 0.55 0.6 0.2 0.2 0.56 0.5 0.04 0.3 0.11 0.10 0 0 0.015 0.01 0.046 0.011 0.027 0.012 0 0.1 0.26 0.1 0 0 0.006 0.005 0 0 0.02 0
Table 2: Concentration of trace elements in leachates.
The leachability of major slag elements, such as Mg, Na, K, S and P, is very low: the amount of each element in the leachates is generally less than 0.001 % of the total amount in the slag. The trace elements can be grouped into categories, depending of their leachability: 1)
2)
3) 4)
High leaching elements are the elements whose amount in the leachate is ~ 1 % or more of the amount in the slag (in bold, Table 2). These elements are Mo, W, B, As, Zn, Se and Sb. However, the last five correspond to trace elements which are generally vaporised from the slag, so only molybdenum and tungsten are likely to be of concern. Elements with moderate leaching properties, where the amount of element in the leachate is between 0.01-0.5 % of the amount in the slag. These elements are Ni, Co, Li, Sc, Cd. Element with low leaching properties (< 0.01 % of amount in the slag), such as Sr, Ba, Mn, Y, Th, U. Element with leachability not detected (and not shown in the table) are: Cu, Cr, Sn, Bi, Hg, F, Pb. Within this list only Cu and Cr are still conserved in the slag; all others are highly vaporised from the slag.
11
s101 % <0.0001 0.0001 0.0001 0.0001 <0.0001 <0.0001 <0.0001 0.00001 0.115
concentration in the slag, ppm / % of leached s102 s103 s104 ppm % ppm % ppm % 16800 0.0001 11000 0.0001 10500 0.0003 5900 <0.0001 6800 <0.0001 9600 <0.0001 2000 0.0003 6000 0.0001 6700 0.0001 2300 0.0003 3600 0.0001 4200 0.0001 6200 0.0001 5800 <0.0001 3000 0.0004 1900 <0.0001 670 <0.0001 1300 <0.0001 3000 <0.0001 790 <0.0001 1200 <0.0001 690 0.00001 240 0 290 0.0001 328 0.067 120 0 341 0
S Mg Na K P Sr Ba Mn Ni
ppm 16600 10500 3300 3400 3500 2400 4600 610 244
B Li Co Y Sc Th Zn Mo U W
51 20 101 99 45 42 19 19 9.7 3.6
0.906 0.206 0.056 0.0009 0.011 0 1.0 0.38 0 0.22
219 46 52 69 52 21 17 10 5.3 2.5
0.147 0.1 0.0123 0.0003 0.01 0 1.62 0.306 0 0.34
37 182 69 47 45 15 20 5 5.0 2.1
0.155 0.165 0.029 0.0011 0.024 0.0008 0 2.12 0.008 0.41
86 127 55 62 56 21 32 18 5.7 2.5
0.073 0.348 0.086 0.0004 0.021 0.0005 0 1.0 0 0.24
Cd As Se
0.6 0.8 0.34
0.016 0.47 0.06
0.47 0.6 0.76
0.06 0.68 0.34
0.4 0.4 0.13
0.075 2.8 4.0
0.51 0.7 0.19
0.0172 1.38 2.1
0.03
0.78
0.09
6.68
0.04
0.55
0.19
1.02
Sb
Table 3: Trace element concentration in slag and the % of the element extracted. Leaching values exceeding 1% are in bold.
A summary of these leaching results is: - There is relatively high leaching of Zn from slags from CRC701 and CRC702 coals. - Sb is leached from slags from CRC702 and CRC704. - There is high leaching of As, Se and Mo from CRC703 and CRC704 slags. - Same elements with moderate leaching or low leaching were observed for almost all four slag samples. It is clear that there are coal-specific issues that are influencing the leaching results presented above. The data presented here are not suitable for an analysis of coal properties and their impacts on trace element partitioning and leaching from gasification residues; however, such information is important and will form the basis of ongoing work in this area. The appearance of trace elements in fly ash is different than in the slags. Therefore, the leaching of some of the trace elements from gasification-derived flyash could be considerably different to that from the slags. In order to be able to confidently manage the handling, disposal, or utilisation of solids from entrained flow gasification, ongoing work in this area is required.
12
4. Conclusions This work has presented some preliminary data regarding the behaviour of trace elements during entrained flow gasification of Australian coals. Based on analyses of slag and flyash samples, trace elements have been categorised into three groups based on their volatility under entrained flow gasification conditions: - non-volatile, - partially volatile and, - highly volatile and three groups based on their reactivity/solubility in the condensed phase: - present in slag only; - present in fly ash; - present in both slag and fly ash. A thermodynamic modelling approach for predicting trace element behaviour during gasification seems to be suitable for selected trace elements only, and usually only for their behaviour in the slag. Appearance of the trace elements in flyash is often very different than the appearance in the slag, and occurs in a significantly shorter timeframe; consequently, kinetic considerations have to be taken into account in order to accurately model their behaviour under entrained flow gasification conditions. Leachability of trace elements from the slag samples was found to be very low for the majority of species. Only Mo, Zn, As, Se and Sb concentrations in the leachate are in the level of a few percent of the concentrations in the slag, but the concentrations of last four elements in the slags are very low due to their high volatility. References [1] Roberts, D. R., Ilyushechkin, A.Y., Tremel, A., Beavis, P., Kochanek, M., Harris, D. J., Gasification of Australian coals, Final Report, ACARP Project C17060B [2] Helble, J. J., Mojtahedi, W., Lyyranen, J., Jokiniemi, J., Trace elements partitioning during coal gasification, Fuel 75 (1996) 931-939 [3] Font, O., Querol, X., Izquierdo, M., Moreno, N.,Diez, S., Alvarez-Rodriguez, R., Clemente_Jul, C., Coca, P., Garcia-Pena, F. partitioning of elements in a entrained flow IGCC plant: Influence of selected operational conditions, Fuel 89 (2010) 32503261. [4] Argent, B.B, Thompson, D. Thermodynamic equilibrium study of trace element mobilisation under air blown gasification conditions, Fuel 81 (2002) 75-89. [5] Thompson, D., Argent B. B. Prediction of the distribution of trace elements between the product streams of the Prenflo gasifier and comparison with reported data, Fuel 81 (2002) 555-570. [6] Diaz-Somoano, M., Martinez-Tarazona, M. R., Trace elements evaporation during gasification based on a thermodynamic equilibrium calculation approach, Fuel 82 (2003) 137-145. [7] Richaud R., Lachas, H., Healey, A. E., Reed, G. P., Haines, J., Jarvis, K. E., Herod, A. A., Dugwell, D. R., Kandiyoti, R., Trace element analysis of gasification plant samples by i.c.p.-m.s.: Validation by comparison of results from two laboratories, Fuel 79 (2000) 1077-1087.
13
[8] Alvarez-Rodriguez, R., Clemente-Jul, C., Martin-Rubi, J. A., Behaviour of the elements introduced with the fuel in their distribution and immobilization between the coal-petroleum coke IGCC solid products, Fuel 86 (2007) 2081-2089. [9] Bunt, J. R., Waanders, F.B., Trace elements behaviour in the Sasol-Lurgi Mk IV FBDB gasifier. Part1 –The volatile elements: Hg, As, Se, Cd and Pb, Fuel 87 (2008) 2374-2387. [10] Bunt, J. R., Waanders, F.B., Trace elements behaviour in the Sasol-Lurgi Mk IV FBDB gasifier. Part2 –The semi-volatile elements: Cu, Mo, Ni and Zn, Fuel 88 (2009) 961-969. [11] Bunt, J. R., Waanders, F.B. Bunt, J. R., Waanders, F.B., Trace elements behaviour in the Sasol-Lurgi Mk IV FBDB gasifier. Part1 –The non-volatile elements: Ba, Co, Cr, Mn and V, Fuel 89 (2010) 537-548. [12] Harris, D. J.; Roberts, D. G.; Frisch, F.; and Lamp, J. Linking Laboratory Scale Gasification Data with Coal Performance in a Pilot Scale Gasifier,- Proceedings International Freiberg Conference on IGCC&XtL Technologies, 3rd, 2009. [13] D.G. Roberts, D.J. Harris and A.Y. Ilyushechkin, Linking Laboratory Data with Pilot Scale Gasification Testing. Part 2: Pilot scale testing (Submitted to Fuel Proc. Techn.) [14] Ward, C.R., Dubikova, M., French, D., Jankowski, J., Li, Z.S., Groves, S. and Riley, K., Mineralogy, geochemistry and leaching characteristics of long-stored fly ashes in relation to ash dam water and groundwater quality: a case study at Wallerawang, New South Wales. Co-operative Research Centre for Coal in Sustainable Development, Technology Assessment Report 48, (2007). [15] Jankowski, J., Ward, C.R. and French, D. (2004). Preliminary assessment of trace element mobilization from Australian fly ashes for hydrogeochemical modelling of fly ash – water interactions. Co-operative Research Centre for Coal in Sustainable Development, Research Report 45, (2004). [16] Jankowski, J., Ward, C.R., French, D. and Groves, S. Mobility of trace elements from selected Australian fly ashes and its potential impact on aquatic ecosystems. Fuel 85, (2006), 243-256. [17] Major Environmental Aspects of Gasification-Based Power Generation Technologies, Final report, DOE, 2002.
14
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: Trace Elements/Emission Effect of coal rank on carbon oxides formation via the low temperature atmospheric oxidation process
Uri Green Chemistry Department, Hebrew University of Jerusalem Department of Biological Chemistry, Ariel University Center at Samaria, Ariel, Israel Contact Information: [email protected] [email protected] Haim Cohen1,2, Zeev Aizenstat3, Christoph Wiedner1,4, Sven Stark1,4 1. Department of Biological Chemistry, Ariel University Center at Samaria, Ariel, 40700 Israel 2. Chemistry Department, Ben-Gurion University of the Negev, Beer Sheva Israel 3. Chemistry Department, Hebrew University of Jerusalem, Jerusalem, Israel 4. TU Bergakademie Freiberg, Fakultät 4, Institut für Energieverfahrenstechnik und Chemieingenieurwesen, 09599 Freiberg, Germany
Imported coal travels large distances in ship-holds and is thereafter stored in strategic stockpiles. From the moment the coal is mined it undergoes degradation in a process called weathering. This process involves a series of reactions between atmospheric oxygen and the coal surface which result in several gaseous emissions. The present research aims at expanding our knowledge of the formation of low temperature (30-150C) oxidation products and their dependence on coal rank . The coals studied are those which serve for power production, namely lignites and bituminous coals. The focus of this work is primarily on the functional groups that can serve as precursors to the formation of carbon oxides. It is known that the formation of carbon oxides is correlated to the concentration of oxygen in the atmosphere. However, the precise relationship and the specific functional groups at the coal surface that are reactive are still unclear. It has been observed that coals of different rank react differently to atmospheric oxidation. That is why simulation experiments were carried out with bituminous coals on the one hand as well as lignites on the other hand. From the results of these experiments a comparison of the behavior of variously aged coals has been evaluated. Conclusions drawn from the present research can help to gain a better understanding of the reactions occurring while stored coal undergoes the oxidation process.
QUANTITATIVE DETERMINATION OF MINERALS IN COAL BY CQPAC METHOD Zdeněk Klika VŠB-Technical University Ostrava, Třída 17. listopadu 15, 708 33 Ostrava-Poruba, Czech Republic Abstract Recently a new method (CQPAC) for the quantitative determination of minerals in coal has been suggested (Klika and Ritz, 2009). This method is based on the recalculation of the elemental bulk chemical analysis on quantitative determination of minerals by optimization procedure. In this contribution the errors of quantitative calculated minerals by CQPAC method are tested using typical analysis of the bituminous coal. These errors can rise from not quite correct input data which are: a) identified minerals, b) wrong determined crystallochemical formulae of minerals and c) error in chemical analysis. The influence of those sources of errors on quantitative determination of minerals in coal by CQPAC method is critically evaluated.
1. Introduction Inorganic matter of coal consists of three basic groups (Ward, 2002). They are: a) Inorganic compounds dissolved in pore aqueous solutions of coal; b) Chemical elements bonded on organic parts of coal; c) Crystalline or amorphous forms of minerals. The first and the second groups are usually denoted as non-mineral inorganic matter, the third group as a mineral matter. The most inorganic elements are namely bonded in crystalline minerals that can be quantitatively determined usually by powder X-ray diffraction (e.g. Querol et al., 1993; Acharya, 1992; Mandile and Hutton, 1995; Ward and Taylor, 1996; Wertz and Collins 1998; Ward et al., 1999, 2001), by scanning electron microscope (SEM) methods (e.g. Finkelman, 1988; Birk, 1989; Martínez-Tarazona et al., 1990) and/or by recalculation of chemical analysis of coal on normative minerals (e.g. Nicholls, 1962; Pollack, 1979; Raask, 1985; Felgueroso et al., 1988; Brons et al., 1989; Ward and Taylor, 1996; Cohen and Ward, 1991). The CQPA method belongs to the last group of methods All above described methods have some limitations that are generally known (e.g. X-ray diffraction methods give usually low intensities of diffraction patterns, have high signal noise; SEM methods are of a point character of analysis; methods based on the recalculation of the bulk chemical analyses on quantitative mineral analyses calculate selected minerals with simple chemical formulae cascaded in selected priority;). This disadvantage is somewhat minimized in CQPAC method based on the optimization principle. Goal of this paper is to determine the errors that rise at quantitative determination of minerals by CQPAC method. 2. Chemical Quantitative Phase Analysis of Coal (CQPAC) This method is based on the recalculation of chemical analysis of coal on quantitative phase (mineral) analysis. For the recalculation the following steps are used:
1
1) 2) 3) 4)
Determination of the bulk chemical analysis of coal (CHA). Identification of minerals present in coal. Selection of chemical formulae for minerals present in coal. Calculation of mineral contents from data obtained in previous points 1, 2 and 3 using optimization procedure CQPAC. 5) Reverse control calculation of CHA from calculated mineral contents (in point 4) determined by CQPAC. These steps are illustrated graphically in Fig. 1.
Fig.1 Diagram of the CQPAC calculation.
Step 1 – Chemical analysis (CHA) X-ray fluorescence, atomic absorption spectroscopy, atomic emission spectroscopy, classical gravimetric and titration analytical methods and/or other ones can be used for the determination of all major elements or their oxides in coal. Determined concentrations of the i-th element (or its oxide) in bulk coal is denoted as (ci)exp. Step 2 – Mineral identification The identification of minerals present in coal can be performed by any available method. Especially methods like SEM, X-ray powder diffraction, FTIR spectroscopy, image analysis, optical microscopy are for this purpose very useful. For such study also heavy fraction and/or LTA obtained from original coal can be utilized. Of course at the LTA preparation we must take into account possible oxidation of some non-stable minerals (e.g. pyrite, etc.) and their transformation on new secondary minerals. Step 3 – Crystallochemical formula of identified minerals 2
For more complicated minerals as are e.g. some clay minerals (muscovite, illite, montmorillonite, chlorite, etc.) the crystallochemical coefficients must be estimated or determined. On the other hand oxides, sulphides, sulphates, carbonates phosphates have usually very simple chemical formulae (Table 1). Step 4 – Calculation of mineral contents The calculation of mineral contents is performed using the optimization procedure CQPAC (Chemical Quantitative Phase Analysis of Coal). This method is based on the assumption that concentration of chemical elements (or theirs oxides) can be express using the following equation (1): n
(ci )calc wi , j c j
(1)
j 1
where (ci)calc – is calculated percentage of the i-th element (or its oxide) in inorganic sample; wi,j is weight fraction of the i-th element (or its oxide) in the j-th mineral; cj is the calculated percentages of the j-th mineral in inorganic sample; n – is number of calculated minerals in inorganic sample. The weight fraction of the i-th element (or its oxide) in the j-th mineral (wi,j) can be calculated from crystallochemical formula of the j-th mineral. The calculation of the j-th mineral content (cj) in inorganic sample is then performed using optimization formula (2): 2
n c i exp wi , j c j min i 1 j 1 m
(2)
where (ci)exp - is percentage of the i-th element (or its oxide) in inorganic sample determined by chemical analysis; m – is number of elements or theirs oxides (in chemical analyses) used for the calculation. This procedure was originally prepared for calculation of minerals present in sedimentary or igneous rocks (Klika et. al., 1986). Step 5 – Reverse calculation of chemical analysis from calculated mineral contents. Validation of calculated mineral percentages (cj) is performed by theirs reverse calculation on chemical analyses from Eq. (1). From the difference between the calculated (ci)calc and experimental concentrations of elements in reduced analysis (ci)exp we can estimate accuracy of calculated cj percentages of minerals in coal. The program CQPAC includes about 50 various minerals identified in coal. In Table 1 are shown some of them.
3
Table 1 Examples of crystallochemical formulae of some minerals. Mineral Crystallochemical formula Quartz SiO2 Rutile/Anatase TiO2 Hematite Fe2O3 Limonite Fe(OH)3 Pyrite/Markazite FeS2 Pyrhotine FeS Calcite CaCO3 Siderite FeCO3 Dolomite/Ankerite CaMg1-xFex(CO3)2 Anhydrite CaSO4 Basanite CaSO4 0.5 H2O Gypsum CaSO4 2 H2O Apatite Ca3(PO4)2 K/Na feldspar NaxK1-xAlSiO8 Plagioclase Na1-xCaxAl1+xSi3-xO8 Kaolinite Al4Si4O10(OH)8 Muscovite/Phengite/ (Kx1Nax2)(Alx3Fex3(III)Mg x3Fex3(II)Mnx7Tix8)(Si4-xAlx)O10(OH)2 Seladonite/Illite Phlogopite/Biotite/ (Kx1Nax2)(Alx3Fe(III)x4Mgx5Fe(II)x6Mnx7Tix8)(Si4-xAlx)O10(OH)2 Annite Smectite (Kx1Nax2Cay)(Alx3Fe(III)x4Mgx5Fe(II)x6Mnx7Tix8)(Si4-xAlx)O10(OH)2 Chlorite (Alx3Fe(III)x4Mgx5Fe(II)x6Mnx7Tix8)(Si4-xAlx)O10(OH)8 Note: crystallochemical coefficients xi can be selected according to present mineral.
3. Results 3.1.
Reference coal sample
For the testing of errors of mineral analysis of coal determined by CQPA the typical bituminous coal (reference sample) was selected. Mineral analysis of this coal (MA) is given in Table 2. Using identified minerals and their crystallochemical formulae (Table 2) the related bulk chemical analyses of coal (CHA+) was calculated (Table 3). For details see step 5 (par.2) – Reverse calculation of chemical analysis from mineral contents. This CHA+, minerals and crystallochemical formulae of selected minerals were used as input data for the CQPAC calculation of mineral percentages (MA+) using steps 1-4, par.2. The calculated MA+ percentages equal to MA results (Table 2) indicating that optimization subroutine CQPAC works quite satisfactory. For the comparison in Table 3 the bulk analyses of elemental oxides (CHA) of reference bituminous coal is also enclosed. The difference between reference CHA and calculated CHA+ analyses (Table 3) can be explained by different ways. Some analyzed elements can be partly bonded on organic matter of coal (e.g. 4
Fe, Ca, Mg, Na) or minor (trace) minerals as are rutile/anatas and/or others were not identified. Of course, the reference (CHA) and/or calculated chemical analyses of coal (CHA+) can be burdened also by errors. In notes to Table 2 the crystallochemical formulaes used for the CQPAC calculation are given. Table 2 Mineral analyses of coal reference sample.
Pyrite Quartz Anhydrite/gupsum Kalcite K/Na feldspar Kaolinite illite
MA 4.8 3.5 1.7 0.3 0.85 18.0 3.3
Calculated MA+) 4.80 3.50 1.70 0.30 0.85 18.00 3.29
SUM 32.45 32.45 Notes: MA - Mineral analysis of reference coal sample. MA+) - Mineral analysis calculated from bulk CHA+) given in Table 3. Crystallochemical formulae used for CQPAC calculation: K/Na feldspar : Na0.2K0.8AlSiO8 Illite : (K0.50Na0.07)(Al1.43Fe0.05(III)Mg0.50Fe0.18(II)Ti0.04)(Si3.39Alx0.61)O10(OH)2.00 For other crystallochemical formulae see Table 1. Table 3 Bulk chemical analyses of coal Mineral SiO2 TiO2 Al2O3 Fe2O3 FeO (Fe2O3 total) MgO CaO Na2O K2O H2O+ S2SO3 (Stotal) CO2 Delta O
Bulk CHA n.d. 0.222 6.37 n.d n.d. 4.45 0.281 1.51 0.067 0.398 n.d n.d 5.7 n.d
Calculated bulk CHA+ 14.188 0.014 8.162 0.034 2.930 0.173 0.868 0.038 0.319 2.668 2.566 1.000 0.132 -0.640
SUM 32.45 Note: Delta O – originated by presence of pyrite 5
n.d – not determined 3.2
Errors due to incorrect identification of minerals
The principal problem of the quantitative mineral calculation by CQPAC consists in the determination of minerals present in sample. If they are not determined correctly then the calculated mineral contents are burdened by errors. It can be supposed that if the identification of minerals is performed very carefully using more methods (e.g. x-ray diffraction, SEM, FTIR, optical microscopy, etc) then all minerals with content above detection limits of used methods are identified; otherwise the errors in calculated minerals percentages can arise. For the testing of such errors 7 calculated variants have been performed (Table 4). Each of calculations use input data: CHA+ (Table 3) and selected minerals with crystallochemical formulae given in notes below Table 2. Calculations differ by selection of minerals. In the first variant pyrite is not included, in the second quartz, in the third anhydrite, etc. For better comparison the reference mineral analysis (MA) of reference coal sample is included (Table 4). The results show that biggest errors originate if pyrite is not included into the calculation. Then very bad results of kaolinite and illite are obtained and K/Na feldspar is even of negative content. Also if quartz (2), kaolinite (6) and/or illite (7) are not included into calculation big errors arise.
Table 4 Influence of identified minerals on calculated MA+ Mineral Pyrite Quartz Anhydrite Kalcite K/Na feldspar Kaolinite illite SUM
Reference MA 4.80 3.50 1.70 0.30 0.85 18.00 3.29
1 3.47 1.70 0.30 -6.04 11.83 17.26
32.45
28.53
2 4.80 1.70 0.30 6.87 16.18 3.49
Calculated MA+ 3 4 5 4.80 4.80 4.78 3.50 3.50 3.64 1.83 1.70 1.07 0.30 0.85 0.85 18.00 18.00 17.66 3.29 3.29 4.38
6 4.28 2.01 1.70 0.30 -11.03 36.56
7 4.85 3.51 1.70 0.30 2.48 19.46 -
33.33
31.52
33.81
32.29
32.28
32.46
Notes: Not identified: 1 – Pyrite; 2 – Quartz; 3 – anhydrite; 4 – kalcite; 5 ortoklas/mikrokline; 6 – kaolinite; 7 – illite; Crystallochemical formulae (see in notes below Table 2) 3.3
Errors due to incorrect crystallochemical coefficients of minerals
The next sources of errors are not correctly determined crystallochemical coefficients of minerals. In our coal sample two minerals have variable crystallochemical coefficients (illite/muscovite and K/Na feldspar). In contrast to original crystallochemical coefficients of 6
minerals (see notes below Table 2) calculated minerals MA+ in Table 5 have been obtained for different crystallochemical formulae of muscovite/illite (1 – 3) and for K/Na feldspar (4). Comparison of the calculated mineral contents MA+ (1 – 4) with reference MA shows that changes of crystallochemical coefficients of calculated minerals results in not striking differences among the results. Table 5 Influence of crystallochemical coefficients of minerals on calculated MA+ Mineral Pyrite Quartz Anhydrite Kalcite K/Na feldspar Kaolinite illite SUM
Reference MA 4.80 3.50 1.70 0.30 0.85 18.00 3.29 32.45
1 4.83 4.43 1.74 0.30 0.15 17.99 2.81
Calculated MA+ 2 3 4.85 4,80 4.11 3.90 1.70 1.70 0.30 0.30 1.06 0.05 18.68 18.41 1.53 3.32
4 4.79 3.54 1.70 0.30 0.30 17.65 4.19
32.22
32.22
32.47
32.48
Notes: Changes in crystallochemical formulae: 1 – muscovite 3 instead of illite; 2 – muscovite 2 instead of illite; 3 – muscovite 1 instead of illite; 4 – crystallochemical coefficient of K/Na feldspar changed from 0.2 on 0.8. Used crystallochemical formulae: 1Muscovite 3: (K0.86Na0.10)(Al1.90Fe0.02(III)Mg0.06Fe0.10(II)Ti0.01)(Si3.02Al0.98)O10(OH)2.00 2Muscovite 2: (K1.00)(Al3.00)(Si3.00Al1.00)O10(OH)2.00 3Muscovite 1: (K0.87Na0.07)(Al1.43Fe0.05(III)Mg0.50Fe0.18(II)Ti0.04)(Si3.39Al0.61)O10(OH)2.00 4K/Na feldspar: Na0.8K0.2AlSiO8 Illite : (K0.50Na0.07)(Al1.43Fe0.05(III)Mg0.50Fe0.18(II)Ti0.04)(Si3.39Alx0.61)O10(OH)2.00 3.4.
Errors due to incorrect determined chemical analyses.
Not correctly determined bulk chemical analysis of coal is the last source of tested errors. For the testing of these errors the identified minerals MA and their crystallochemical formulae (Table 2) have been used while the bulk chemical analysis (CHA+) has been changed as shown in the legend below Table 6. For the first MA+ calculation (variant 1) all oxides of CHA+ have been increased about 5 % so that sum of them is 34.04 %. In Table 6 calculated contents of minerals are also increased about 5% as well as sum of minerals (variant 1). Analogously for the second MA calculation (variant 2) all oxides of CHA+ have been decreased about 5 % and calculated contents of minerals are decreased about 5% as well as sum of minerals. In next calculations (variants 3 – 6) only the changes in one oxide of CHA+ are performed (see legend below the Table 6). Comparison of the calculated mineral contents (variants 3 – 6) with original MA show that changes of oxide percentages in bulk analysis result in not striking differences of calculated minerals.
7
Table 6 Influence of incorrect determined chemical analyses on calculated MA+ Reference MA Pyrite 4.80 Quartz 3.50 Anhydrite 1.70 Kalcite 0.30 Ortoklas/Microcline 0.85 Kaolinite 18.00 illite 3.29
1 5.04 3.67 1.79 0.31 0.89 18.90 3.46
SUM
34.07
2 4.56 3.33 1.62 0.28 0.81 17.11 3.12
Calculated MA+ 3 4 5 4.80 4.78 5.06 4.92 2.30 3.50 1.70 1.70 1.70 0.30 0.30 0.30 0.85 0.95 0.53 18.0 18.97 17.71 3.29 4.55 3.95
6 5.02 3.50 1.70 0.30 1.17 18.29 2.65
30.83
33.87
32.63
Mineral
32.45
33.58
32.75
Note: Relative errors of oxides: 1: +5% positive error of all element oxides; 2:–5 negative error of all element oxides; 3: 10 % positive error of SiO2 (SiO2=15.61%); 4: 10 % positive error of Al2O3 (Al2O3= 8.98 %); 5: 10 % positive error of FeO (FeO= 3.22 %); 6: 10 % positive error of S2- (S2- = 2.82 %). Crystallochemical formulae – see in notes to Table 2. 4. Conclusion The methods for the quantitative minerals determination in coal are usually burdened with relatively big errors. It is also true for the methods based on the recalculation of bulk chemical analysis of inorganic element oxides into quantitative phase (mineral) analysis. In this paper the CQPAC method (Klika and Ritz, 2009) is used and validated. The validation of this method has been performed using typical bituminous coal sample for which the mineral and chemical analyses have been independently determined. Three possible sources of errors in mineral calculation by CQPAC were tested. They are: 1) Incorrect identification of minerals present in coal 2) Incorrect crystallochemical coefficients of minerals 3) Errors in bulk chemical analysis of coal The results show that biggest errors origin in case that some of major mineral has not been included into the CQPAC calculation. With high probability this example can not come because major minerals are usually reliably determined with x-ray diffraction or others instrumental or optical methods. Omitting of some of minor mineral brings about much fewer errors in rest mineral calculation. The incorrect selection (approximation) of crystallochemical coefficients and/or fewer errors in bulk chemical analysis of elements in coal bring about relatively much less errors in quantitative calculation of minerals by CQPAC method. Acknowledgment We acknowledge the financial support of the Grant Agency of the Czech Republic (grant No.105/08/0913).
8
References Acharya B. S., 1992. Quantitative determination of minerals in Indian coal by X-ray diffraction. Fuel. 71, 346-348. Birk D., 1989. Coal Minerals: Quantitative and descriptive SEM-EDX analysis. J. Coa. Qual. 8, 55-62. Brons G., Siskin M., Botto R.I., Güven N., 1989. Quantitative mineral distribution in Green River and Rundle oil shales. Energy & Fuels. 3, 85-88. Cohen D., Ward C. R., 1991. SEDNORM – a program to calculate a normative mineralogy for sedimentary rocks based on chemical analyses. Compt. Geosci. 17, 1235-1253. Felgueroso J., Martínez-Alonso A., Martínez-Tarazona M., Tascón J.M.D., 1988. The determination of mineral matter content of low-rank coals. J. Coal Qual. 7, 127-131. Finkelman R. B.,1988. The inorganic geochemistry of coal: a scanning electron microscopy view. Scanning Electron Microsc. 2, 97-105. Klika, Z., Weiss, Z., Chmielová, M., 1986. A method of quantitative mineralogical analysis of rocks from their elementary chemical analysis. Proceedings from The ninth conference on clay mineralogy and petrology, Ostrava. 11-18. Klika Z., Ritz, M. 2009. A new method for the determination of minerals in coal. Proceedings of International Conference of Coal Science and Technology, Cape Town, in: CD-ROM, section 8/08-2. Mandile A. J., Hutton A. C., 1995. Quantitative X-ray diffraction analysis of mineral and organic phases in organic-rich rocks. Int. J. Coal. Geol. 28, 51-69. Martínez-Tarazona M. R., Palacios J. M., Tascón J. M. D., 1990. SEM-EDX characterization of inorganic constituents of brown coals. Inst. Phys. Conf. Ser. 98, 327-330. Nicholls G. D., 1962. A scheme for recalculating the chemical analyses of argillaceous rock for comparative purposes. Am. Miner. 47, 34-46. Pollack S. S., 1979. Estimating mineral matter in coal from its major inorganic elements. Fuel. 58, 76-78. Querol X. A., Alastuey A., Chinchón J. S., Fernández-Turiel J. L., López-Soler A., 1993. Determination of pyritic sulphur and organic matter contents in Spanish subbituminous coals by X-ray power diffraction. Int. J. Coal. Geol. 22, 279-293. Raask E., 1985. The mode of occurrence and concentration of trace elements in coal. Prog. Energy Combust. Sci. 11, 97-118.
9
Ward C. R., Taylor J.C., 1996. Quantitative mineralogical analysis of coals from the Callide basin, Queensland, Australia using X-ray diffractometry and normative interpretation. Int. J. Coal. Geol. 30, 211-229. Ward C. R., Soeasr D. A., Booth C. A., Staton I., Gurba L. W., 1999. Mineral matter and trace elements in coals of the Gunnedah Basin, New South Wales, Australia. Int. J. Coal. Geol. 40, 281-308. Ward C. R., Matulis C. E., Taylor J. C., Dale L. S., 2001. Quantification of mineral matter in Argonne Premium Coals using interactive Rietveld-based X-ray diffraction. Int. J. Coal. Geol. 46, 67-82. Wertz D. L., Collins L. W., 1998. Using X-ray methods to evaluate the combustion sulphur minerals and graphitic carbon in coals and ashes. Am. Chem. Soc. Div. Fuel. Chem. 33, 247252.
10
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
PROGRAM TOPIC: COAL SCIENCE (CHEMISTRY) THE CATALYTIC EFFECT OF ADDED SODIUM- AND POTASSIUM CARBONATE TO AN ACID TREATED INERTINITE RICH SOUTH AFRICAN BITUMINOUS COAL CHAR Christien A Strydom, Prof Chemical Resource Beneficiation, North-West University, Potchefstroom, South Africa. [email protected] Lucinda Klopper,Me; John R Bunt, Prof* Chemical Resource Beneficiation, North-West University, Potchefstroom, South Africa. * Seconded by Sasol Technologies South Africa to North-West University and Harold H Schobert, Prof The EMS Energy Institute, Penn State University, University Park, Pennsylvania 16802 USA and Chemical Resource Beneficiation, North-West University, Potchefstroom, South Africa.
Alkali salts in the mineral part of coal decompose, react and vaporize as the temperature increases. The degree of vaporization of alkali metals or compounds, as well as other inorganic species, depends on the mode of occurrence of the mineral matter, the temperature and the grain size in the coal. The condensed inorganic species deposit onto the boiler tubes of the furnace at lower temperatures and give rise to a fouling problem. The condensed alkali species also lead to agglomeration of remaining coal and ash particles within the reactors. Potassium and sodium compounds also enhance the gasification rate due to catalytic activity of these compounds. Alkali species react with the coal phenolate- and carboxylic groups and catalyze the formation of ether cross-links within the carbon matrix of coal. The amounts of potassium and sodium species within coal thus change the reactivity of the coal.
A South African inertinite rich bituminous coal, containing 23% ash was treated with HF and HCl to reduce the mineral content to less than 2%. The coal char was prepared by heating the coal samples under nitrogen up to a temperature of 900 ºC and constant mass. Between 0.25% and 4% (mass percentages) of Na2CO3, K2CO3 and a blend of these compounds were added to the coal samples before charring. The same amounts were added to separate charred coal samples. Comparative CO2 reactivities of the acid treated coal char samples were determined using a thermogravimetric method. The catalytic effects of potassium carbonate and sodium carbonate on the reactivity of the treated coal char samples were investigated. These large percentages of added compounds enhance the influence thereof and lead to clearly identified trends. A substantial increase in the reactivity was observed after addition of these catalysts. SEM analyses of the char that formed indicated changes in char porosity and surface structure. BET surface area analyses were performed and reported. The contribution of the increased surface areas to the increased reactivity was established and the influence of
the added catalysts thus validated. The samples were also heated under CO2 to 900 ºC in a tube furnace, and XRD and FTIR analyses performed on the products to analyze the composition of the samples.
Pre-Combustion Cleaning of Pulverized Fine Coal at Power Plant Using Novel RTS Dry Separation Technology Daniel Tao 234 MMRB, University of Kentucky, Lexington, KY 40506, U.S.A. (859) 257-2953 (o); (859) 323-1962 (f); [email protected] Ahmed Sobhy 234 MMRB, University of Kentucky, Lexington, KY 40506, U.S.A. (859) 257-4124; (859) 323-1962; [email protected] Qin Li 234 MMRB, University of Kentucky, Lexington, KY 40506, U.S.A. (859) 257-4124; (859) 323-1962; [email protected] Rick Honaker 234 MMRB, University of Kentucky, Lexington, KY 40506, U.S.A. (859) 257-1108; (859) 323-1962; [email protected]
ABSTRACT Coal cleaning technologies are vitally important for producing clean coal to supply abundant and affordable energy for the nation’s economy under the increasingly stringent environmental regulations. They are generally categorized into pre-combustion and post-combustion processes. Coal cleaning prior to combustion is often accomplished using wet physical processes at coal preparation plants. However, the resultant coal product still contains a significant amount of impurities such as ash, sulfur, mercury that are not well liberated from coal particles mostly in the size range of several millimeters to inches. The pulverization of coal prior to combustion at the power plant provides an ideal feed of well liberated fine coal particles for physical cleaning. If a cost-effective dry separation process is developed to further clean pulverized fine coal at the power plant prior to its combustion, the cost of expensive post-combustion chemical cleaning processes such as flue gas scrubbing can be reduced considerably. In this study, a novel rotary triboelectrostatic separator (RTS) was investigated for dry cleaning of pulverized Illinois fine coal samples. The proprietary RTS technology is characterized by an innovative high efficiency rotary charger, charger electrification, laminar air flow, and specially designed electrodes. The separation performance of this technology was evaluated in terms of ash content reduction as well as mercury and sulfur content reduction. Important process parameters such as charger material, potential and rotation speed, feed flow rate and air flow rate were investigated for their effects on the separation performance. Key Words: Fine Coal; Liberation; Mercury; Particle Charging; Rotary Charger; Sulfur; Triboelectrostatic Separation.
1
INTRODUCTION Run-of-mine coal often contains high levels of impurities such as ash, mercury, and sulfur that have to be removed before coal can be utilized by electric utility companies or the steel making industry. This process is referred to as coal cleaning, which is accomplished at the coal washing plant usually located at or near the coal mine. Most of run-of-mine coal is cleaned by low-cost gravity and heavy medium separation processes in water with particles of up to several inches in size. However, it is well known that coal particles larger than even one millimeter still contain a significant amount of ash, mercury, and sulfur due to poor liberation of minerals. Thus, some of the “clean” coal product from the coal preparation plant is still considered to be non-compliance coal and has to be blended with higher quality coal before combustion or the utility company has to resort to expensive post-combustion cleaning processes to meet air emission standards. At power generation utility plants, coal is often pulverized to –200 mesh (75 µm) size prior to combustion. Pulverized coal has an average particle size of about 5 µm. At such a fine size, ash-forming minerals including mercury- and sulfur-containing pyrite are essentially liberated from the coal matrix since these minerals in Illinois coal are mostly liberated at or slightly below 1 mm or 1000 µm. Thus, pulverized coal provides an ideal feed to the dry triboelectrostatic separation process to further upgrade the coal by removing additional ash, mercury, and sulfur. The dry triboelectrostatic cleaning of fine coal has recently received intensive attention worldwide. Lindquist et al. (1995) successfully applied electrostatic separation to pyrite removal from the Illinois No. 6 bituminous coal. Stencel et al. (1998) carried out a detailed study of triboelectrostatic cleaning of fine coal with the U.S. coal samples. Dwari and Rao (2009) studied the separation of an Indian thermal coal using a cylindrical fluidized bed tribo-charger dry and reported that the process reduced the -300 m coal feed ash from 43% to 18% and 33% with a yield of about 30% and 67%, respectively. Bada et al. (2010a, b) investigated triboelectrostatic cleaning of two South African coals and also reported promising results. For example, the technique reduced the coal ash from 32% to 12% and at the same time lowered the sulfur content from 2.8% to 0.4% at a recovery of 9%, which was much better than the flotation separation performance. Similar studies were conducted by Trigwell et al. (2003) with the Illinois No. 6 and Pittsburgh No. 8 coals to remove ash and sulfur. Electrostatic beneficiation of fine coal was also explored by Chinese researchers (Fan et al., 2006) who reported encouraging separation results. Nevertheless, electrostatic cleaning of fine coal has not been commercialized successfully due to its perceived drawbacks of relatively low throughput, poor efficiency, etc. The present study was performed to develop a novel rotary triboelectrostatic separator (RTS) for dry cleaning of Illinois coal after pulverization at the power plant to further reduce ash, mercury, and sulfur content in coal. The cleaned product can be directly fed to the boiler without further treatment. The RTS is characterized by an innovative high efficiency rotary charger, charger electrification, laminar air flow, and specially designed electrodes. Experimental tests have shown that compared with a conventional separator, the innovative separator has increased particle charge density, improved solids throughput, enhanced process efficiency, and reduced overall energy consumption. The RTS technology provides an efficient dry separation process for cleaning or purification of coal fines and other fine particles. 2
EXPERIMENTAL Laboratory Prototype RTS Separator A continuous laboratory prototype separator illustrated in Figure 1 was fabricated for pulverized fine coal cleaning. Rotary chargers made of stainless steel and copper were used for fine coal cleaning. An electrical potential was applied to the rotating charger via a copper brush which was connected to the charger via a unique mechanism. The separator had an overall dimension of 6” (L) 3” (W) 50” (H) and a capacity of approximately 10 lb/h. The rotary charger had dimensions of about 3” in diameter and 4” in length. The rotary triboelectrostatic separation system included a vibratory sample feeder, a rotary charger or charge exchanger, a separation chamber, an injection gas unit, and two high voltage DC supplies. Samples were fed by the feeder into the rotary charger. A small amount of transport gas was injected with the feed. Gas-particle flow interacted with the rotary charger. Due to particle-to-charger or particle-to-particle collisions, particles became charged negatively or positively, depending on their work functions. Charged particles then passed through the separation chamber and reported to one of three cyclones attached to the system. A honeycomb collimator was used to minimize turbulence and vibration electrodes were deployed to reduce particle deposition in the separation chamber. One of the two high voltage sources in the system was for particle charging and was attached to the rotary charger. The other was for the separation of charged particles and is attached to the separation chamber.
(2) Heating Lamp (1) sample Feeder
(5) Motor
(11) Splitter
(7) Speed Controller
(16) Injection Flow
(8) Separation Chamber
(12) Cyclone
(6) Roller Charger
(15) Tailings
(14) Middlings
(13) Concentrate
(12) Cyclone
(12) Cyclone
(4) Sample Distributor
(3) Eductor
(9) DC Power for Roller
(10) DC Power for Electrodes
Figure 1. The laboratory triboelectrostatic separation system.
3
Fine Coal Sample The pulverized coal sample used in this study was acquired from a power plant that used a pulverizer to reduce coal size feed to the boiler. The proximate analysis showed the coal sample had 16.4% moisture, 21.0% ash, 32.4% volatile matter and 27.2% fixed carbon. Table 1 shows the size-by-size weight and ash distribution data of the coal sample. Approximately 20% of the sample was smaller than 44 m and this sample had a d50 of 32 m.
Size (mm) 2.38 2.38-1.19 1.19-0.595 0.595-0.250 0.250-0.125 0.125-0.074 0.074-0.044 -0.044
Table 1. Particle size-by-size weight and ash distribution Cumulative Underscreen Wt (%) Ash (%) Wt (%) Ash (%) 15.55 15.98 100.00 20.81 12.09 13.78 84.45 21.70 12.67 11.69 72.36 23.02 18.43 9.68 59.69 25.43 11.46 12.33 41.26 32.46 5.74 15.15 29.80 40.20 4.48 19.96 24.06 46.18 19.58 52.17 19.58 52.17
Parametric Investigation of RTS Technology A systematic parametric investigation of the proposed RTS separation technology was conducted for fine coal separation performance evaluation. The coal sample was kept in an oven to minimize moisture absorption. Approximately 100 grams was processed in each experiment with feed metering provided by the action of a vibratory feeder. Three products were collected from the first stage; namely R (concentrate), C (middling), and L (tailing). Typically, a two-stage processing protocol was used. Each product from the first stage was further processed to generate nine final products. Unless otherwise specified, all separation tests were conducted using the copper charger, and under the following conditions: charger rotation speed: 3000 rpm; injection air velocity: 2.5 m/s; co-flow air velocity: 3.1 m/s; feed rate: 800 g/h; separation voltage: 22.5 kV; charger voltage: –2.5 kV; temperature: 24 °C. Process parameters including feed rate (0.4, 0.8, 1.2, 2.0, and 3.6 kg/h), charger material (copper and stainless steel), charger rotation speed (2000, 3000, 4000, and 5000 rpm), applied charger voltage (–5, –2.5, 0, +2.5, and +5 KV), injection air flow velocity (0.9, 1.5, 2.5, 3.7, and 4.9 m/s), and co-flow rate (0.9, 1.5, 2.5, 3.7, and 4.9 m/s), were examined. All samples were analyzed for ash content. Coal samples generated under optimum conditions were also analyzed for sulfur and mercury content, which were then used to determine ash, sulfur, and mercury rejection as well as combustible recovery for the process. RESULTS AND DISCUSSION The separation performance data in this paper is presented in terms of combustible recovery vs. product ash or normalized product ash which is defined as the percentage of product ash to 4
feed ash. The normalized product ash is preferred in this study since the wet and sticky feed samples were very difficult to homogenize. Also, significant fluctuations in feed ash were observed in the tests. Effect of Feed Rate Figure 2 shows the effect of feed rate on the coal separation performance. The data was normalized with respect to feed ash so that the shift in grade vs. recovery curves can be solely attributed to changes in feed rate. The curve that is closer to the upper left corner represents a better separation. Figure 2 shows that as the feed rate increased from 800 to 2000 g/h, the separation performance improved. This suggests that with this coal sample, particle-particle contacts played a significant role in acquiring surface charge since higher feed rate always result in more particle-particle contacts. It should be noted that during the tests, concentrate moved toward the negative electrode and tailings were deflected to the positive electrode, suggesting that carbon particles were positively charged and minerals were negatively charged, which is in agreement with previous studies (Lindquist et al., 1995; Stencel et al., 1998; Trigwell et al., 2003, Bada et al., 2010a).
Figure 2. Feed rate effect on separation efficiency. Effect of Charger Rotation Speed Tribocharging is dependent on the relative motion speed between particles and the rotary charger, with higher speed resulting in stronger electrification. The easiest way to control this is to adjust the rotary charger rotation speed. Figure 3 shows the effect of charger rotation speed on the separation performance curve. At 5000 rpm, ash content can be reduced by 20% at a combustible recovery of about 58%. This cleaning effect is considered to be significant since the feed was very fine coal that had been cleaned with traditional coal washing techniques with the remaining ash components mainly in middling particles known to be difficult to clean. This data 5
clearly show that better separation was accomplished at higher rotation speeds of 4000 or 5000 rpm, which is consistent with established theory that higher surface charge results when relative motion speed between charger and particles increases (Stencel et al., 1998).
Figure 3. Charger rotation speed effect on separation efficiency. Effect of Applied Charger Voltage One of the unique features of the rotary triboelectrostatic separator is the applied potential to the charger to enhance the particle charging process. Figure 4 shows separation efficiency curves at different charging voltages ranging from –5 kV to +5 kV. It is quite clear from these figures that separation performance significantly increased as charging voltage varied from +5 kV to –5 kV, and compared to 0 V or no charging voltage, which is the case with the conventional triboelectrostatic separation, the separation at -2.5 or –5 kV was substantially more efficient. For example, a product ash reduction of 30% was obtained at 0 V charging voltage with only less than 10% recovery. However, it was obtained at –5 kV charging potential with a combustible recovery of more than 45%. Comparing the separation curves at +5 kV and –5 kV suggests that the application of -5 kV increased the combustible recovery by up to 80%, which clearly illustrates the utmost importance of controlling the applied charger potential.
6
Figure 4. Applied charger voltage effect on separation efficiency. Effect of Injection Air Flow Rate It is known that injection air flow rate or velocity affects particle speed and residency time in both charging and separation chambers and thus, particle charge density and separation efficiency. A higher injection air flow rate results in a faster velocity at which particles strike the charger but it causes shorter charging and separation times. Figure 5 shows separation efficiency curves at different injection flow velocities. As the injection air velocity decreased from 4.9 to 1.5 m/s, the separation performance curve moved upward consistently, indicating the separation performance was improved. This was possibly a result of longer residence time in the separation chamber at lower air flow velocity. However, too low a velocity of 0.9 m/s gave rise to a poorer performance than at 1.5 m/s, mainly due to lower charging efficiency.
7
Figure 5. Injection flow rate effect on separation efficiency. Effect of Air Co-Flow Rate The gas flow that enters the separation zone on both sides of the connector between charging and separation chambers is referred to as co-flow. It is used to comb misplaced particles and force them to deflect to the desired product stream. Compared to conventional triboelectrostatic separators on the market, this is another distinct feature of the new rotary triboelectrostatic separator. Figure 6 shows separation efficiency curves at different co-flow rates. The injection flow rate was kept equal to the co-flow rate to minimize air turbulence in the separation chamber in these tests (Tao et al., 2008). The separation curves moved consistently toward the upper left corner as the air co-flow rate increased from 1.5 m/s to 3.1 m/s, indicating that better separations were achieved at higher co-flow rates. As the co-flow rate further increased to 3.7 m/s, the separation curve shifted substantially toward the lower right corner, suggesting a poorer separation performance. Thus, the optimum co-flow rate for this coal sample was 3.1 m/s at which a 30% reduction in ash content from 21% to less than 15% was achieved with a 70% recovery.
8
Figure 6. Injection flow rate effect on separation efficiency Effect of Charger Material Two chargers, one made of copper and the other made of stainless steel, were tested for fine coal cleaning to investigate the effect of charger material on triboelectrostatic separation performance. Copper is known to be a good material for the charger while stainless steel is a common wear-resistant material. Figure 7 shows separation efficiency curves obtained with both copper and stainless steel chargers. For this particular coal sample, the copper charger produced substantially better separation than the stainless steel charger. It should be noted, however, that the importance of the charger material for the separation efficiency heavily depends on the coal sample to be processed. With some coal samples we tested, the stainless steel almost produced the identical separation.
9
Figure 7. Charger material effect on separation efficiency. Comparison of RTS Separator with Release Analysis Release analysis was often used to provide a measure of ultimate separability of a coal sample by froth flotation technique. Its result represents the best possible separation that can be achieved by froth flotation under ideal conditions. Figure 8 shows the release analysis and RTS separation results. The RTS performance was better than the release analysis data at lower product ash but was slightly inferior to the release analysis result at higher product ash.
Figure 8. Comparison of release analysis and RTS separation performance.
10
Figure 9. Comparison of ash, sulfur, and mercury rejection efficiency of the RTS process. Sulfur and Mercury Rejection Sulfur and mercury rejection efficiency is shown in Figure 9 together with ash rejection with the –7+325 mesh fraction coal sample. The sulfur and mercury rejection correlates well with the ash rejection. The RTS technology was effective in reducing the amount of sulfur and mercury although ash was the easiest to remove. The sulfur rejection was almost the same as the Hg rejection. CONCLUSIONS The following conclusions can be drawn based on the data shown above: 1. For the coal sample evaluated in this study, coal particles were positively charged while ash particles were negatively charged. Experimental results have demonstrated that the RTS technology achieved an effective further cleaning with a pulverized clean coal with a d50 of 32 m and an ash content of 20%. For example, a cleaner coal product of 15% ash was produced at a recovery of more than 70%. 2. Process variables including feed rate, injection air flow rate, air co-flow rate, and applied charger potential all affected separation efficiency. The appropriate applied charger potential was critical for separation efficiency. For the tested coal sample, a charger potential of –5kV produced much better separation than 0 V or positive voltages. Injection flow rate and co-flow rate should be approximately equal to result in the most favorable flow condition in the separation chamber.
11
3. The copper rotary charger performed better than the stainless steel rotary charger in upgrading fine coal particles. The degree of improvement by the copper charger depends on the coal sample. 4. The RTS process can not only reduce the ash content in the coal but also reject sulfur and mercury effectively, although sulfur and mercury were more difficult to remove from coal. ACKNOWLEDGEMENT The authors would like to sincerely acknowledge the funding of the Department of Commerce and Economic Opportunity of the state of Illinois through the Office of Coal Development and the Illinois Clean Coal Institute under the project Number of 08-1/4.1A-1, which made this work possible. Technical assistance and professional advice from the project manager Joseph Hirschi is deeply appreciated. REFERENCES Bada, S. O., Tao, D., Honaker, R. Q., Falcon, L. M. and Falcon, R. M. S., 2010b. “Parametric study of electrostatic separation of South African fine coal”, Journal of Mining Science and Technology, July, 20(4), pp.535-541. Bada, Samson; Falcon, L.M.; Falcon, R.M.S.; Honaker, Rick; Tao, Daniel, 2010a. “A study of rotary Tribo-electrostatic separation of south african fine coal”, International Coal Preparation Congress 2010, Conference Proceedings, p 596-607. Dwari, R.K., Hanumantha Rao, K., 2009. “Fine coal preparation using novel tribo-electrostatic separator”, Minerals Engineering, v 22, n 2, p 119-127. Fan, M., Chen, Q., Zhao, Y., Tao, D., Luo, Z., Zhang, X., Tao, X., Yang, G, 2006. “Fine coal dry classification and separation”, Minerals and Metallurgical Processing, v 23, n 1, p 17-21. Lindquist, D.A., Mazumder, M.K., Tennal, K.B., McKendree, M.H., Kleve, M.G., and Scruggs, S. 1995. “Electrostatic beneficiation of coal”, Materials Research Society Symposium Proceedings, v 371, p 459-463, Advances in Porous Materials Stencel, J. M., Schaefer, J. L., Ban, H., Li, T. –X., and Neathery, J. K. 1998. “Triboelectrostatic Cleaning of Coal In-line Between Pulverizer and Burners at Utilities”, Int. J. Coal. Prep. Util., 19(1): 115-129. Tao, D., Fan, M., and Jiang, X., (2008), “An Innovative Rotary Triboelectrostatic Separator for Fly Ash Purification,” Annual SME Meeting Preprint 08-008, Salt Lake City, UT, February 25 – 28. Trigwell, Steve, Tennal, Kevin B., Mazumder, Malay K., Lindquist, David A., 2003. “Precombustion Cleaning of Coal by Triboelectric Separation of Minerals”, Particulate Science and Technology, v 21, n 4, p 353-364.
12
The prediction of caking propensity of Gondwanaland coals using petrography Daniel Van Niekerk, Johan Joubert and Trudie Brittz Sasol Technology Research and Development, Coal and Syngas Technologies, 1 Klasie Havenga Road, Sasolburg, 1947, South Africa
Abstract The caking propensity of a coal (the extent of agglomeration) is an important factor to consider in coal conversion processes. In general, the caking propensity of Northern Hemisphere Carboniferous-aged coals can be predicted using vitrinite content: The higher the vitrinite content, the greater the caking propensity. Predicting caking propensity for Permianaged Gondwanaland coal is problematic due to the high inertinite content. Inertinite consists of both a reactive component (exhibit behaviour similar to that of the corresponding vitrinite) and an inert component. Therefore, it is postulated that the amount of reactive macerals will correlate to the caking propensity. Current characterization of reactive inertinite is conducted by visual recognition during maceral point-count analysis. Visual recognition is highly subjective and therefore prone to errors. To reduce the amount of bias from the maceral analysis, a new multi-maceral reflectogram model is proposed. A multi-maceral reflectogram consisting of reflectance values from vitrinite, semifusinite and inertodetrinite were constructed for various coals. From the multi-maceral reflectogram the amount of reactive macerals could be determined. The amount of reactive macerals obtained from the reflectogram exhibited a linear relationship with caking propensity: The greater the caking propensity, the higher the amount of reactive macerals. Therefore, the multi-maceral reflectogram is proposed as a petrographic method to predict caking propensity of Gondwanaland coals.
Introduction The caking propensity of a coal (extent of agglomeration) is an important factor to consider in coal conversion processes.1 In general, the caking propensity of coals can be predicted using vitrinite content: The higher the vitrinite content, the greater the caking propensity.2 Predicting caking propensity of inertinite-rich coals is problematic due to the presence of both inert- and reactive inertinites. Reactive inertinites (semifusinite and inertodetrinite) have reflectance values similar to the accompanying vitrinite and also exhibit behaviour similar to that vitrinite.2,3 Therefore, it is postulated that the amount of reactive macerals will correlate to the caking propensity. A new multi-maceral reflectogram model is proposed consisting of reflectance values from vitrinite, semifusinite and inertodetrinite.
Experimental Samples Various bituminous coal samples were collected and prepared for petrographic and caking analysis. Petrography Vitrinite random reflectance analyses (ISO 7404-5) and maceral composition analyses (ISO 7404-3) were conducted on all the coal samples. In addition, random reflectance analyses were conducted on 100 inertinite particles (semifusinite and inertodetrinite). Classification of coals was conducted according to ISO 11760-2005. Classification of inertinite was based on the International Committee for Coal and Organic Petrology (ICCP) classification system.4 Caking propensity of coal The various coal samples were heated, under pressure, to approximately 600 ºC. The caking propensity (%) was determined using Equation 1.
% Caking =
Agglomerated Mass × 100 Total Mass
Equation 1
Results and discussion Standard petrographic maceral point-count and random reflectance (RoV%) analyses were conducted and are summarized in Table 1. Coal Coal A Coal B Coal C Coal D Coal E Coal F Coal G Coal H Coal I
Table 1: Petrographic results for the various coals
Vitrinite 13.9 22.1 35.0 13.4 8.2 15.4 81.8 67.0 11.0
Liptinite 1.4 3.4 1.9 0.6 0.8 1.0 1.8 4.0 7.0
Inertinite 69.5 63.3 55.3 73.1 70.8 72.4 10.8 8.0 47.0
Minerals 15.3 11.2 7.8 13.0 20.3 11.2 5.6 21.0 35.0
RoV% 0.67 0.63 0.60 0.62 0.75 0.65 0.61 0.61 0.52
s 0.09 0.07 0.06 0.09 0.14 0.00 0.07 0.04 0.07
Caking % 4.7 29.2 53.9 21.7 9.7 4.5 100 100 66
According to mean random vitrinite reflectance (RoV%), all the coals are all ranked highvolatile bituminous. Coal I, however, lies on the border of high-volatile bituminous and subbituminous. The distribution of the inertinite macerals are summarized in Table 2. Table 2: Inertinite composition of the various coals Coal
RSF
ISF
Fusinite
Micrinite
Coal A Coal B Coal C Coal D Coal E Coal F Coal G Coal H Coal I
16.0 5.8 8.8 12.6 5.4 12.2 1.4 2.0 7.0
3.5 1.0 2.3 3.6 0.8 1.6 0 3.0 18.0
4.7 18.5 13.6 8.8 14.3 10.8 3.2 1.0 4.0
3.5 8.8 4.3 4.2 6.0 4.6 3 0 0
R Inertodetrinite 31.5 20.1 18.1 34.3 35.4 35.4 2.4 0 6.0
I Inertodetrinite 10.2 9.2 8.2 9.6 8.9 7.8 0.8 2.0 12.0
Total Inertinite 69.5 63.3 55.3 73.1 70.8 72.4 10.8 8.0 47.0
RSF - Reactive semifusinite, ISF - Inert semifusinite Distinction between reactive and inert inertinite (semifusinite and inertodetrinite) was conducted visually. Inertinite macerals with a visual reflectance similar to the accompanying vitrinite were labeled as reactive inertinite. It is noted that visual recognition of reactive inertinite is highly subjective and reproducibility could only be accomplished if there is regular interaction between petrographers.2 Several methods have been proposed to determine the percentage of reactive inertinite in coals.2,3,5 Using vitrinite content to predict caking propensity It is generally accepted that there is a relationship between caking propensity and the vitrinite content of a coal (the higher the vitrinite content the greater the caking propensity). The vitrinite content versus caking propensity of all the coals was evaluated and a linear relationship was observed between coals A to H (R2 of 0.93). Coal I did not exhibit the same vitrinite vs. caking trend than the other coals (Figure 1).
100.0 90.0
% Vitrinite (mmf) _
80.0 70.0 60.0 50.0 40.0 30.0 20.0 Coal I
10.0 0.0 0
10
20
30
40
50
60
70
80
90
100
110
% Caking
Figure 1: Vitrinite content versus caking propensity Therefore, the vitrinite content versus caking propensity method can be applied to all the high-volatile bituminous coals with a reflectance greater than 0.60 % (RoV% >60 %). Multi-maceral reflectogram method to predict caking propensity It is generally accepted that a linear relationship exists between caking propensity and reactive maceral content for a given coal (the higher the reactive maceral content the greater the caking propensity). Reactive macerals include: vitrinite, liptinite, reactive semifusinite and reactive inertodetrinite. Various methods have been proposed to determine the percentage of reactive inertinite in a coal sample.2,3,5 The reactive content versus caking propensity for the coals is shown in Figure 2. 100.0
% Petrograpic reactives
90.0 80.0 70.0 60.0 50.0 40.0
R2 = 0.2587
30.0 0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0 110.0
% Caking
Figure 2: Reactive content versus caking propensity for all coals No significant correlation between caking propensity and reactive macerals were observed. This may be due to over and/or under estimation of the reactive inertinites during maceral point-count analysis. Visual recognition of reactive inertinites is highly subjective and therefore prone to errors.2,3 To reduce the amount of bias from the maceral analysis, a multimaceral reflectogram model was proposed. The multi-maceral reflectogram was generated from the reflectance values from vitrinite, semifusinite and inertodetrinite. Random reflectance analyses were conducted on 100 vitrinite particles and 100 inertinite particles (semifusinite and inertodetrinite) and compiled into V-types (for vitrinite) and Itypes (for inertinite). For example: vitrinite reflectance values between 0.3 and 0.39 are classified as V3, vitrinite reflectance values between 0.4 and 0.49 are classified as V4, etc. The resulting multi-maceral reflectograms are displayed in Figure 3 and 4.
18
18
16
16
14
14
Frequency (%)
Frequency (%)
20
12 10 8 6
12 10 8 6
4
4
2
2
0
0 3
4
5
6
7
8
9
3
10 11 12 13 14 15 16 17 18 19 20 21 22
Reflectance types
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22
Reflectance types
Figure 3: Multi-maceral reflectogram of two non-caking coals In non-caking coals the reflectograms (Figure 3) indicate two distinct populations: vitrinite population (V3 to V8) and an inertinite population (R9 to R22). Therefore, in non-caking coals the vitrinite and inertinite populations are distinctly separated. 40
20
A
35
16
Frequency (%)
Frequency (%)
25 20 15 10
B
18
30
14 12 10 8 6 4
5
2
0
0 3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22
Reflectance types
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Reflectance types
Figure 4: Multi-maceral reflectogram of two caking coals The reflectograms of caking coals (Figure 4) are either vitrinite-rich (Figure 4A) or the vitrinite and inertinite populations overlap (Figure 4B). To quantify caking propensity from the reflectograms, the following is assumed: vitrinite and a portion of the inertinite (semifusinite and inertodetrinite) is reactive and will result in caking. The calculation of the cut-off value between reactive and inert portions in a reflectogram is well discussed in literature. 3,4,6 The reactivity can be determined from these reflectograms as follows (Figures 5 to 7): 35
Figure 5
Frequency (%)
30 25 20 15 10 5 0 3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22
Reflectance types
Reflectograms are generated by compiling all reflectance values (vitrinite and inertinite) into R-types (reflectance-types). The R-type distribution was normalized to correspond to the maceral point-count data (vitrinite, semifusinite and inertodetrinite). The resulting multi-maceral reflectogram is shown in Figure 5.
35
Frequency (%)
All vitrinites are reactive (reactive portion of reflectogram shown in grey). The inertinite that overlaps with the vitrinite are also assumed to be reactive (overlapping inertinite shown in stripes). Therefore, the reactive portion of the reflectogram is all the vitrinite and the overlapping inertinite (inertinite with the same reflectance to the corresponding vitrinite).
Figure 6
30 25 20 15 10 5 0 3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22
Reflectance types 35
Figure 7
Frequency (%)
30 25 20 15 10 5 0 3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22
It is assumed that all inertinite with a reflectance of 4 R-types above the highest vitrinite reflectance will be reactive (reactive inertinite). This assumption is based on trail and error in linking caking to reactivity. Therefore, the reactive portion of the reflectogram (shown in grey in Figure 7) consists of vitrinite, overlapping inertinite and reactive inertinite. The rest of the reflectogram is assumed to be inert. The percentage of reactives (reflectogram reactivity) can now be determined.
Reflectance types
The reflectogram reactivity (percentage of reactives) was calculated from the multi-maceral reflectogram for each coal. A linear correlation between reflectogram reactivity and caking propensity was observed (Figure 8).
Reflectogram reactivity (%)
100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0
R2 = 0.9581
20.0 0
10
20
30
40
50
60
70
80
90
100
110
Caking (%)
Figure 8: Reflectogram reactivity versus caking propensity Therefore, if the reflectogram reactivity (R) can be determined, the caking propensity can be predicted using linear Equation 2. % Caking = 1.4486(R) - 42.203
Equation 2
This method is complex due to the determination of the reflectogram reactivity value (R), but can be applied as a tool to predict caking propensity.
Conclusions A multi-maceral reflectogram consisting of reflectance values from vitrinite, semifusinite and inertodetrinite were constructed for various coals. From the multi-maceral reflectogram the amount of reactive macerals could be determined. The amount of reactive macerals obtained from the reflectogram exhibited a linear relationship with caking propensity: The greater the caking propensity, the higher the amount of reactive macerals.
References 1. 2. 3. 4. 5.
Taylor, G.H.; Teichmuller, M., Davis, A., Diessel, C.F.K., Littke, R., Robert, P., Organic Petrology, 1998, Gebruder Borntraeger, Berlin. Snyman, C.P., 1989. The role of coal petrography in understanding the properties of South African coal, International Journal of Coal Geology 14, 83–101. Kershaw, J.R., Taylor, G.H., 1992. Properties of Gondwana coals with emphasis on the Permian coals of Australia and South Africa, Fuel Processing Technology 31, 127–168. International Committee for Coal and Organic Petrology (ICCP), 2001. The new inertinite classification (ICCP System 1994), Fuel, 80, 459-471. Kruszewska, K., 1989. The use of reflectance to determine maceral composition and the reactive-inert ratio of coal components, Fuel, 68, 753-757.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Manuscript Submission TOPIC 6: COAL SCIENCES SESSION 29: COAL SCIENCE: BENEFICIATION - 1 DRY COAL CLEANING USING THE FGX SEPARATOR Mehmet Saracoglu, Ph.D. Candidate, University of Kentucky, Department of Mining Engineering University of Kentucky, 131 Mining and Mineral Resources Building, Lexington, KY, 40506-0107, [email protected] Rick Q. Honaker, Ph.D., Chair and Professor, University of Kentucky, Department of Mining Engineering University of Kentucky, 230 Mining and Mineral Resources Building, Lexington, KY, 40506-0107, [email protected] Abstract: The extraction of coal typically results in the recovery of pure rock that ranges from very small to very large quantities depending on the seam thickness and other characteristics. In the eastern U.S. coalfields, a significant amount of out-of-seam rock is being extracted in order to recover coal from relatively thin seams. As a result, the amount of run-of-mine feed that is rejected as rock material at the coal cleaning plant can range from 50%-70%. The haulage, processing, and storage of the rock represent significant energy inefficiency and high operating costs. Removing the pure rock material near the extraction point would provide significant economic and environmental benefits. The removal of pure rock using a relatively high-density separation of around 2.0 is referred to as “deshaling”. Wet-based technologies are the most commonly employed cleaning units for removing rock from coal; however, these processes are generally massive and immobile, while also requiring water addition and a slurry treatment system. This study focused on evaluating a novel dry separation technology, the FGX Separator, at several coal mining operations throughout the U.S. The FGX Separator applies table concentration principles using air as the medium. Parametric evaluations were conducted at nearly all tests sites using a 3-level experimental design in an effort to realize optimum performances. Regardless of coal type, table frequency and longitudinal slope were found to be the most critical factors in controlling product ash content over a range of energy recovery values. Additionally, the amount of fluidization of air applied through the deck was critical for ensuring optimum energy recovery. 1. Introduction: 1.1 Background The three basic steps in coal mining are: (1) face extraction of the coal, (2) haulage of the broken coal to the preparation facilities, and (3) cleaning of the coal to produce a marketable product. The extraction of coal typically results in the recovery of pure rock that ranges from very small to very large quantities, depending on the seam thickness and other characteristics. To illustrate the importance, one can consider the high-quality, federally compliant coal extracted in the eastern U.S. coalfields. It has been estimated that the eastern U.S. coalfields retain nearly 85% of the original coal reserves after two centuries of escalating production. The eastern Kentucky coalfield is considered one of the largest resources of low-sulfur, high-BTU coal among the eastern U.S. coalfields, with a 64 billion ton reserve base. The high-quality reserve also extends into southern West Virginia and southwestern Virginia. However, Weisenfluh et al. (1998) reports that nearly 52% of the remaining eastern U.S. coalfield resources are located in
coal seams, which are 14 to 28 inches in thickness, while 31% are in 28 to 42 inch-thick seams. Likewise, a study on the Virginia coalfields found that 30% of the total reserve base or 203 million tons exist in seams with a thickness less than 28 inches (Saracoglu, 2007). Extraction of coal from these seams is a significant challenge both technically and economically. To economically extract coal from thin seams, it is necessary to use equipment with high advancement rates, which generally requires equipment that is oversized compared to the seam height. As a result, a significant amount of out-of-seam rock is recovered, loaded and hauled to the preparation plant. At the plant, high rejection rates of 50%-70% are common, which equates to double the amount of electricity and other fuels being used to recover a ton of clean coal as compared to past practices in thick seams. At several operations in eastern Kentucky, the recovery of 1 ton of clean coal results in 2 tons of reject that requires transportation and storage at a significant financial and environmental cost. Although the three basic coal mining steps, those mentioned above, have traditionally been viewed as independent operations, integration of these steps can provide significant improvements in overall energy efficiency and economical benefits (Luttrell et al., 1996). This goal can be achieved by employing a mobile cleaning process near the extraction point or working face of a mining operation, which will provide an efficient, high-density (≈2.0 S.G.) separation for the removal of pure rock, with ash values higher than 80% and very low or no heating (Btu) values. As previously discussed, the content of high-density rock in the raw feed is increasing in U.S. coalfields. In eastern Kentucky, current values average around 60%-70%. Removing the pure rock material near the extraction point would reduce the energy required for material haulage from the mine site to the preparation plant or load-out facility and from preparation plant to the refuse disposal area. The removal of well-liberated extraneous pure rock from run-of-mine coal is referred to as “deshaling”. The greatest impact of deshaling on economics may be realized when the pure rock is rejected from a coal that bypasses the processing plant and is subsequently blended with processed cleaned coal to achieve the final market requirement for quality. This benefit can be realized by one mining operation or from the global operations of a company, whereby the coal from a mine at one operation is blended with another to meet contract grade requirements for one end user. As shown in Figure 1, if one volume unit of pure rock is rejected by deshaling, four times the volume in middling particles containing 25% ash by volume can be added to the product. After considering the relative density differences, this implies that about 3 tons of middling material can be recovered for every ton of pure rock rejected (Saracoglu, 2007). Honaker et al. (2006) also showed that the treatment of low-ash run-of-mine coal to remove the small amount of rock using a dry separator prior to blending with washed coal has significant economical benefits. Middling Units
Pure Rock
= Coal Rock
4 volume units of middlings = 1 volume unit of rock 3 tons of middlings = 1 ton of pure rock Figure 1. Schematic showing the equivalence of recovering one ton of rock versus three tons of middling particles.
1.2 Overview Dry particle separators have a long history of application in the U.S. coal industry. According to Arnold et al. (1991), the amount of coal processed in the U.S. through dry cleaning plants reached a peak in 1965 at 25.4 million tons. The largest dry-based cleaning plant located in Pennsylvania processed 1400 t/hr of -1.9 mm coal using a total of 14 units. The last complete dry cleaning plant operating in Kentucky was closed in the late 1980’s. Reasons for the decline in dry coal production to less than 4 million tons annually by 1990 include increased run-of-mine moisture levels that resulted from dust suppression requirements and the demand for higher quality, compliance coal which required efficient, low density separations. The effective top size particle for most of the separators was around 50 mm and the effective size ratio for which good separation was achieved was between 2:1 and 4:1. It is noted that this effective particle size range is much smaller than most wet, density-based separators. The reported probable error (Ep) values vary due to the particle size ranges treated. For example, within a small particle size range of 4:1, the Ep values ranged from 0.15-0.25 whereas a 50:1 ratio provided Ep values around 0.30. These values indicate that the air-based systems are much inferior in separation efficiency as compared to wet coarse coal cleaning units. However, dry coal cleaning devices typically have lower capitol and operating costs, no waste water treatment and impoundment requirements, lower product moisture values and less permitting requirements. If a high-density separation provides the desired effect on coal quality, dry cleaning separators are an attractive option (Honaker et al., 2007). Interest in dry coal cleaning has increased significantly in recent years mainly due to a need to clean lignite and sub-bituminous coals in the western and Gulf Coast regions of the U.S. Expectations are that requirements for cleaning Powder River Basin coal will increase in the near future due to changing geological conditions that include the presence of intrusions in some of the major coal seams. An additional application is the removal of high-density rock from run-of-mine coal in the Central Appalachia coalfields prior to shipment to a processing facility. Current interest is to identify an efficient and cost-effective dry cleaning technology for the purpose of achieving the deshaling objective. As a result, several processing technologies used during the peak years of dry coal preparation have been recently modified and successfully commercialized. The Allair Jig, for example, is a modification of the Stump Jig technology and is commercially represented by Allminerals Llc. (Kelly and Snoby, 2002). The unit has been successfully applied in several applications within and outside the U.S. for coal cleaning (Weinstein and Snoby, 2007). Chinese researchers and manufacturers have applied basic fundamentals including computational fluid dynamics to the redesign of dry particle separators including those employing dense medium and tabling principles. The FGX Separator, developed by Tangshan Shenghou Machinery Co. Ltd. is an example of a Chinese dry, density-based separation technology that has several hundred commercial installations (Lu et al., 2003, Li and Yang, 2006). A mobile pilot scale unit of the FGX Separator has been evaluated recently for the cleaning of run-ofmine coal sources of different ranks in the U.S. This publication will present and discuss the separation performances achieved on the U.S. applications. 2. Experimental: 2.1 Process Description The FGX Separator, a dry coal cleaning device, employs the separation principles of an autogenous medium and a table concentrator as shown in Figure 2. The feed is introduced into a surge bend from which the underflow is controlled using an electro-magnetic feeder. The separation process generates three product steams, i.e., clean coal product, middlings, and tailing streams. Two dust collection systems are employed to clean the recycled air and to remove the dust from air being emitted into the atmosphere. The separating compartment consists of a deck, vibrator, air chamber and hanging mechanism. A centrifugal fan provides air that passes through holes on the deck surface at a rate sufficient to transport and fluidize the particles. Riffles located on the deck direct material toward the back plate. The deck width is reduced from the feed end to the final refuse discharge end. Upon introduction of feed coal into the separation chamber, the upward flow of air forms a fluidized-bed of fine reject, which causes the light coal particles to migrate to the top of the particle bed. The presence of about 10-20% minus 6mm material in the feed is needed to develop a fluidized autogenous medium particle bed. Low-density particles (such as coal) move toward the right front part of the table due to the downward slope (from the back of the table to the front) of the table and form the upper layer of particles that are collected along the front length of the table. The upward fluidization velocity is insufficient to fluidize the high-density particles (such as rock) and thus they maintain contact with the table surface where both vibration and the continuous influx of new feed material force the material by mass action to travel toward the narrow end of the table where the final refuse is collected. The particles that overflow the front of the table are directed into product, middling or tailing bins by splitters that are adjusted to achieve a clean coal product with high-ash content tailings. The material from each of the three bins is transferred from the FGX unit by conveyors.
Figure 2. Schematics showing the basic operation of the FGX Separator unit. 2.2 Test Program Description A 5 t/hr pilot-scale FGX Separator unit utilizing a 1 m2 air table was used at the various test sites. The feed to the unit was supplied directly from the mine or pre-screened, using an inclined vibrating screen, to remove the minus 6mm material from the feed prior to treatment. Underground coal sources required pre-screening due to the relatively large amount of fines and the amount of surface water present due to dust suppression activities. The purposes were to ensure that the separation process was not affected by the presence of an excessive amount of fine material and to minimize the negative effects of feed moisture. Since the minus 6mm material is required to form a fluidized medium, the screening process was operated in a manner that allowed a portion of the fine fraction to report to the unit. After initiating the feed, a representative sample was collected from each process stream after three minutes of steady operation until a bed is formed on the separating deck. A second sample cut was taken after an additional minute and a third cut was obtained after another minute. When three streams (clean coal, middlings and tailings) were collected through the splitters, each sample was collected for 10 seconds period, with a total of 30 seconds sampling for each stream. The three sample cuts from a process stream were combined for a total representative sample under the given operating conditions. The tests were conducted in continuous mode with no recycling of middling stream. 2.3 Parametric Studies Parametric studies were conducted on bituminous, sub-bituminous and lignite coals to optimize the operating conditions of the FGX Separator using the following variables: •
•
•
•
Air Blower Frequency (hz.): Air is introduced to the FGX Separator via a centrifugal fan underneath the deck surface, which passes through the holes at a sufficient rate to transport and fluidize the particles. Air blower frequency is crucial to maintain fluidization of the light (clean coal) particles at an elevated level and to allow high-density (rock) particles to settle onto the separation deck. Table Frequency (hz.): Vibration is introduced to the table through asynchronous motors that are attached to the back of the air table. Vibrating motion of the table is perpendicular to the feed entry and shakes the air table from front to back. Table frequency is the variable that controls the vibration of the table, which is important for the transportation and discharge of the material on the deck surface. Cross-Table Slope (degree): Cross-table slope is the inclination of the table deck from the back plate to the front discharge plate. The normal range for the cross table slope is between 8º-10º. Cross-table slope is important for the heavy (reject) material traveling along the back plate to the reject end of the table with the vibration. Cross-table slope should be maintained at a sufficient inclination for the discharge of the heavy and light material reporting to the front of the air table. Longitudinal Slope (degree): Longitudinal slope is the inclination of the table deck from the reject end to the feed end of the FGX Separator. The normal range of the longitudinal slope is between 0º and 2º. High longitudinal slope values (1.5º) keep the fluidized light (clean coal) material near the feed end of the table thereby resulting in high energy recovery values and high-density separations. On the other hand, low longitudinal slope values (0.5º) results in lower ash content in the product but greater replacement of coal in the reject stream. As a result, longitudinal slope is the key variable for the quality of the clean coal material.
Using the Design-Expert software, developed by Stat-Ease. Inc., parametric studies were conducted to evaluate the parameter and parameter interactions. The objectives were: • •
Quantify the effect of the four variables and their associated interactions on the separation performance achievable by the FGX Separator. The experimental data was used to form models, which described their effects on the response variables, i.e., yield, recovery, product ash content. Identify the set of parameter values providing the optimum separation performance for a given coal using the models.
At the end, a three-level statistically-designed Box-Behnken test program was chosen to be performed on each coal type to determine if the magnitude of the parametric effects vary as a function of coal rank and the amount of reject material in the feed. Based on the preliminary tests conducted on the FGX Separator in China and in the U.S., the lowest (-1 level), the mid-point (0 level) and the highest (+1 level) parameter values for all variables were determined, as shown in Table 1. At the first testing site, the test program on the FGX pilot-scale unit varied all four variables, which involved 29 tests. An evaluation of the parametric models, which will be discussed later, showed that the cross table slope has the least impact on the separation of coal and rock. Therefore, this variable was kept at a constant value of 8.5° and was removed for the parametric tests at all subsequent sites, which reduced the number of tests for optimization to 17. By varying the parameter values systematically, the optimum test performances and the corresponding conditions were identified. Table 1. Range of parameter values evaluated during testing programs of the FGX Separator Range
Operating Parameter Air blower frequency (%) Table frequency (%) Cross table slope (degree) Longitudinal slope (degree)
-1 level
0 level
+1 Level
50 40 8.0 0.5
55 45 8.5 1.0
60 50 9.0 1.5
2.4 Coal Characteristics 2.4.1 Bituminous Coal The first test program was conducted at a western bituminous coalfield in Utah. The objective at this testing site was to evaluate the potential of the FGX Separator to reduce the ash and total sulfur contents of the coarse fraction (+6.4 mm) comprised in the run-ofmine bituminous coal feed to a level that will produce a marketable clean coal product when blended with the untreated fine coal. The required product ash content from the treatment of the +6.4 mm coal was understood to be around 10%. Maximization of mass yield to the product was targeted to ensure an economically feasible application. The particle size and density distribution for the run-of-mine feed material is shown in Table 2. The particle size distribution clearly shows that the majority of the ash-forming components existed in the +6.4 mm material.
Table 2. Characterization data obtained from analysis of the run-of-mine coal at the western bituminous coalfield. Size Fraction (mm) 51 x 25.4 25.4 x 13 13 x 6.4 - 6.4
Total
Particle Size Analysis Weight Ash Total (%) (%) Sulfur (%) 12.85 31.62 1.63 34.71 19.90 1.69 41.97 15.62 1.67 10.47 11.81 4.19
100.00
18.76
1.93
Density Fraction (R.D.) 1.4 Float 1.4 x 1.6 1.6 x 1.8 1.8 x 2.0 2.0 x 2.2 2.2 Sink Total
Weight (%) 75.06 8.61 2.42 1.44 1.84 10.63 100.00
Washability Analysis Ash Total (%) Sulfur (%) 6.25 1.85 21.32 1.63 37.12 2.85 53.24 3.79 71.40 3.11 87.10 2.12 18.76 1.93
2.4.2 Sub-Bituminous Coal For Powder River Basin (PRB) sub-bituminous coals in Wyoming, coal preparation typically involves simple crushing. However, during the extraction of coal from the pit, some coal gets contaminated with out-of-seam material. As a result, the contaminated material is kept in the pit as fill material. Estimates of the amount of lost coal due to contamination are between 1 and 10 million tons annually, which represents a significant loss in revenue. Tests were performed to evaluate the feasibility of the FGX Separator to reject the out-of-seam rock and produce a marketable clean coal product. During the parametric tests, 152 x 0 mm (6 x 0-in.) raw coal feed was collected before screening at 51mm (2-in.) and 6.4 mm (¼-in.) sieves to obtain the appropriate particle size range of 51 x 6.4 mm (2 x ¼-in.) for the FGX Separator. A particle size analysis of a representative feed sample revealed that 78% of the total coal had a size greater than 6.4 mm, as shown in Table 3. Considering that the material is currently disposed at the mine site, the ash content was relatively low. The size analysis shows that the highest amount of ash-bearing material exists in the -6.4 mm range and for the +6.4 mm fraction, the amount of ashbearing material was around 23.89%. Interestingly, the majority of the large rock was found to easily breakdown and dissolved into ultrafine particles after wetting. This behavior of the shale particles eliminated the possibility of conducting the washability analysis. Table 3. Raw feed (152 mm x 0) size distribution from the PRB sub-bituminous coalfield. Size Fraction (mm)
Weight (%)
Ash (%)
+51
65.64
25.37
51 x 6.4
12.76
16.27
-6.4
21.60
31.23
Total
100.00
25.47
2.4.3 Lignite Coal Similar to the PRB coal, coal processing at a Gulf Coast lignite operation in Texas, typically involves crushing and screening. However, significant amount of the deposit contains large pyrite inclusions that are a carrier for mercury. Therefore, the FGX Separator was evaluated for its ability to reject pyrite and other ash-bearing materials, as well as for the reduction of mercury. The objective for this site was to evaluate the potential of the FGX Separator to reduce the ash, total sulfur and mercury contents of the coarse fraction (51 x 6.4 mm) comprised in the run-of-mine feed. Mercury reduction was a requirement for use at a nearby power generating facility and the sulfur reduction and energy value enhancement provide economical benefits. Maximization of mass yield to the product was targeted to ensure an economically feasible application with possible minimum product ash content.
3. Results and Discussion: 3.1 Bituminous Coal 3.1.1 Separation Performance The initial test program was performed on western bituminous run-of-mine coal and the objective of the test program was to produce a clean coal product that was sufficient in quality to meet specifications for a steam coal contract. The required operating conditions to meet the objective vary from coal-to-coal due to the unique physical characteristics of the coal including the particle size and density distributions, particle shape of the coal and reject material and the quantity of high density rock in the feed. As such, the values of four critical operating parameter values, as shown in Table, were varied over 29 tests according to a three-level statistical design. Clean coal-middlings and middlings-reject splitter positions were maintained at pre-determined settings throughout the program. The amplitude of the table was maintained at 8 mm while the air valve supplying air to the table was set at the full open position. Feed rate was maintained at a level of around 5.6 tons/hr (5.0 metric tons/hr). The tests were conducted in continuous mode with no recycling of the middling stream. Mass yield values were determined based on the middlings stream being combined with the tailings stream to represent the total reject. Timed samples of the feed, product, middling and tailing streams were collected from each test.
A mass balancing method was utilized for the raw data through the minimization of weighted sum of squares adjustment approach. Mass yield to the product was determined based on sample weights collected over the same time period from each process stream, while the energy recovery (%) was calculated using the yield value and the heating values (Btu) of the clean coal and feed. Although the dust sample data is omitted from the table, the data were included in the overall yield assessment. The average ash reduction achieved during the test program by the FGX unit was from 18.21% to 10.76% while recovering 76.8% of the total feed weight. The total sulfur content was decreased from 1.61% to 1.49%. This equates to ash and total sulfur rejection values of 53.9% and 29.0% on average, respectively. As a result, the average heating value of the coal was upgraded from 11513 Btu/lb to 12691 Btu/lb. Product ash and total sulfur contents realized from the FGX unit were as low as 9.55% and 1.39%, respectively, with minimal effect on overall product yield. The average ash content in the tailings stream was 72.70% and values greater than 80% were realized from several tests. Also, sulfur was found to be effectively concentrated into the tailings stream as indicated by the average sulfur content of 2.67%. 3.1.2 Parametric Evaluation The data from the statistically designed test program were used to develop empirical models describing effect of the operating parameter values on the response variables, i.e., product ash and energy recovery. In the 29-test program, five repetitive tests (Tests 2, 11, 15, 16, and 18) were conducted corresponding to the central parameter values. Using these values, normal error is calculated associated with the experimental procedure and sample analysis to show the degree of repeatability of the test results. The variability for Energy Recovery (%) and Product Ash (%) responses resulting from experimental error are shown in Table 4. Table 4. Statistical evaluation of the experimental error realized from the western bituminous coal parametric test. Test Number 2 11 15 16 18
Energy Recovery (%) 84.66 87.24 88.47 86.76 80.88
Product Ash (%) 10.44 11.58 10.47 10.79 10.69
Mean
85.60
10.79
95 % Confidence Interval
2.61
0.41
The data shows the existence of a small degree of randomness in the response variables measured upon the parameter values. This is evident in the measured Energy Recovery (%) value of Test 18 and Product Ash (%) value of Test 11, which marginally fall out of the determined confidence interval corresponding to 85.60% 2.61 and 10.79% 0.41, respectively. The other measured response variables show a lower degree of variability due to experimental error. The software suggested the quadratic model for the prediction of the response variables. The quadratic model can be generalized as:
Y = A1 + A2 (ABF) + A3 (TF) + A4 (CTS) + A5 (LS) + A2 (ABF 2 ) + ...+ Ai (CTSxLS)
(1)
where Y is either the energy recovery or the product ash and Ai the corresponding coefficients, ABF the air blower frequency, TF the table frequency, CTS the cross-table slope and LS the longitudinal slope.
€
3.1.2.1 Model Significance A statistical model suggests that the observed results are a direct product of the changes in the parametric values rather than experimental error. To prove this, the variances due to the experimental procedures should be less than the variance in the response variables for all tests; otherwise a statistical model cannot exist. The Model F-statistic (F-value) is used to evaluate the effects of parameter changes on a response variable by testing the null hypothesis using the following equation:
SSyy F =
SSE
df r
(2)
df e
where SSyy is the sum of squares residuals for the response variables, dfr the degrees of freedom associated with the calculation of SSyy, SSE the sum of squares of the experimental error and dfe the degrees of freedom associated with the determination of the SSE value which is obtained€by subtracting the measured responses (n=29) from the model parameters in the quadratic model (k=14). The null hypothesis suggests that the variance observed for all tests can be explained by pure error. In order for the model to be accepted, the null hypothesis should be rejected. Comparing the F-value with the calculated value of 95% confidence interval, F0.05, can reject the null hypothesis. If the F-value exceeds the calculated F0.05, the model is accepted. F0.05 is determined from the F-Table using the model parameters (k=14) and difference between the measured responses and the total parameters (n-k+1) in the quadratic model, where n=29. Both of the Model F-statistic values shown in Table 4.4 are greater than the calculated F0.05 value of 2.48 and therefore the models were considered adequate. One of the most common practices in null hypothesis testing is to report the significance level of a test using the probability value, Prob > F. This value represents the probability of falsely rejecting the null hypothesis. The Prob > F value less than 0.05 indicate a strong relationship between the parameter changes and the response variables. As shown in Table 5, both models had Prob > F values less than 0.05, which are considered to have a significant effect on the response variable. Table 5. ANOVA table generated for the response variables associated with the FGX Separator tests in UT. Degree of
Source
F Value
Freedom
Model
14
Lack-of-Fit
10
Model
14
Lack-of-Fit
10
Prob > F
Energy Recovery Model 7.77 0.0002 1.40
0.3990
Product Ash Model 4.13 0.0061 1.12
0.4986
R2
Radj2
0.89
0.77
0.81
0.61
Another method to test the null hypothesis is using the Lack-of-Fit F-statistic (FLOF), which determines the significance of the individual parameter effects and associated interactions in the model. The Lack-of-Fit analyses the null hypothesis, which means the corresponding coefficients in the empirical model are zero, using the following equation:
SSE FLOF =
SSR
k
(3)
([ n − (k + 1)]
where SSE is the model sum of squares, SSR the sum of squares of the residuals, n the number of data points and k the number of model parameters. The calculated Lack-of-Fit F-statistic value should be less than F0.05 value for a model to pass the lack-of-fit test. F0.05 is determined from the F-Table using the lack-of-fit (n-k-m=10) and the pure error (m=4) degrees-of-freedom values in the € quadratic model within the 95% confidence interval. Both of the calculated Lack-of-Fit F-statistic values shown in Table 5 were less than the F0.05 value of 5.96 with low Prob > F values, suggesting that the Lack-of-Fit F-statistic is non-significant and the quadratic model is an adequate fit. The coefficient of determination, R2, is used as a measure of the ability of a model to accurately predict the response parameter value. The R2 value was determined using the following expression:
R
2
∑ (y = 1− ∑ (y
i
− yˆ i ) 2
i
− y i )2
(4)
where y i is the observed value, yˆ i the predicted response variable value for each test and y i the mean experimental value obtained from all available data. An R2 value of 1.0 indicates that 100% of the response parameter variability within the parameter value ranges tested can be explained by one or more of the parameters and parameter interactions considered in the model. As shown in Table 5, € the combustible recovery and product ash models have high R2 values of 0.89 and 0.81, respectively.
€
€
The R2 value could be increased by adding parameters or interaction terms and could be forced to a value approaching to 1.0. However, a high R2 value is not the only parameter to quantify the ability of the model to predict the response variable. The adjusted coefficient of determination, Radj2, is used to account for the presence of excessive terms within the model and is calculated and compared to the R2 value using the following expression: 2
Radj = 1−
n −1 (1− R 2 ) n − (k + 1)
(5)
where n is the number of data points and k the number of parameters in the model. By comparing Equations (4) and (5), one can see that Radj2 must always be less than R2, which indicates the excessive number of parameters from their relative difference. If the R2 and Radj2 values differ by a small number, the models are adequate for the prediction of the response variable values. As shown in Table 4.4, the energy€recovery and product ash Radj2 values of 0.77 and 0.61, respectively, were considered reasonable with respect to R2 values. The comparisons between the actual and predicted values for the energy recovery and product ash models show a good prediction of the experimental data (Figure 4). a)
b)
13
Predicted Product Ash (%)
98 Predicted Energy Recovery (%)
€
91
84
77
69
12
11
10
9 69
77
84
91
Actual Energy Recovery (%)
98
9
10
11
12
13
Actual Product Ash (%)
Figure 4. Comparison of the experimental and empirical models for: a) energy recovery (%) and b) product ash (%) of the western bituminuous coal.
3.1.2.2 Model Simulation and Discussion In the parametric studies, the significance of a model parameter is indicated by Prob > F, with a value less than 0.05. For the western bituminous coal testing, the quadratic model was found have a significant effect in the determination of the energy recovery and the product ash content with respective Prob > F values of 0.0002 and 0.0061. From the operating parameters, both the table frequency (TF) and the longitudinal slope (LS) of the table were found significant with the corresponding Prob > F values of <0.0001 and 0.0002, respectively, for energy recovery and 0.0028 and 0.0073, respectively, for product ash response variables. On the other hand, cross-table (CTS) slope was found insignificant for both energy recovery (Prob > F=0.3969) and product ash (Prob > F=0.0721) variables. As a result, a decision was made to take out CTS as one of the operating parameters for the parametric tests at all subsequent sites and kept constant at 8.5º, which was found to be optimum value. During the quadratic model simulation, air blower frequency (ABF) was found to have a significant effect on energy recovery (Prob > F=0.0179). However, this was not the case for product ash content (Prob > F=0.34), which may be explained based on the effect of air on the table, enabling only the low-density material (product) to fluidize and discharge near the feed end of the table, when sufficient amount of air is distributed to the table deck. On the other hand, the energy recovery drops when insufficient amount of air enables the low-density material (product) to hit the table surface, which eventually reports to the middlings or even reject part of the table. Additionally, the square interaction of longitudinal slope (LS2) was found to be statistically significant effect on determining the product ash content (Prob > F=0.0095). Interestingly, the same interaction was found non-significant for the energy recovery model (Prob > F=0.5203). The reason for this difference may be explained by the importance of the longitudinal slope for determining the product ash. As discussed before, the longitudinal slope is the inclination from the reject end to the feed end of the table. As such, increasing the longitudinal slope (positive values, i.e. 1 - 1.5º) increases the resistance for material to transport toward the reject end of the table, which typically results in rock reporting in the product. On the other hand, a horizontal slope (i.e., 0º longitudinal slope) leads to a low product ash content. Simulated interactive effects on the response variables are shown in Figures 5 and 6. From the ANOVA table, table frequency and longitudinal slope were both found to have significant effects on FGX Separator performance for Energy Recovery (%) and Product Ash (%) variables. The contour lines represent the tendency of the energy recovery values and product ash contents with respect to the operating parameters and correspond to the identified values on the y-axis. Model predictions indicate that increasing the longitudinal slope from the reject end to the feed end and decreasing table frequency improved energy recovery, as shown in Figure 5. Energy recovery was increased from about 74% to 86% with decreasing table frequency from 50 hz. to 40 hz. Decreasing the table frequency in the aerated table eliminates the probability of the heavy material (reject) to moving toward the front of the table with the excessive movement. This helps the light material (product) to stratify along the bed without misplacing reject particles, which could result in poor separation performances. As such, energy recovery can be further increased from about 86% to 98% by using the lowest table frequency value of 40 hz, while simultaneously increasing the longitudinal slope from 0.5 degrees to 1.5 degrees. The vibration of the table applied through the control of table frequency had an interesting effect on the product ash content as shown in Figure 6. As expected, the product ash content initially improved with a decrease in table frequency due to better stratification along the feed end of the table. However, as the table frequency is decreased further, the high-density (tailings) material was not able to report to the back plate and travel along the table length, thereby creating product ash gradients on the aerated table with significant amount of ash-bearing material close to the feed end of the table. A more significant effect on product ash is realized from longitudinal slope adjustments. Over a longitudinal slope range from 0.5 to 1.5º, product ash is increased from about 10% to 12%, as a result of high-ash middlings material reporting to the product stream. 3.2 Sub-Bituminous Coal 3.2.1 Separation Performance For the Powder River Basin (PRB) sub-bituminous coal, tests were performed to evaluate the feasibility of the FGX Separator to reject the out-of-seam rock during the extraction process and produce a marketable clean coal product from the contaminated “rib” coal. As previously discussed, the FGX Separator best performs in narrow size ranges, with a lower size limit of around 6.4 mm (¼-in.). This is evident from Table 3 that, the closest size range (12.7 x 6.4 mm) had the lowest feed ash content for the PRB coal and the highest amount of ash-bearing material exists in the minus 6.4 mm range. Due to the high-ash content of the fine fraction (minus 6.4 mm), the average product ash was increased from 6.9% with the -6.4 mm not included to 8.4%, when processing 51 x 0 mm coal fraction. However, the average yield to the product was almost the same for both feed size distributions, which was around 80%. Based on the experimental data, the best ash separation performance was achieved by Test 6 for both feeds, with product ash contents at 9.1% and 7.2% and corresponding yield values of 90% and 89.1%, respectively.
Energy Recovery (%)
98.4 92.4 86.4 80.4 74.4
50 1.50
48
1.25 1.00 Longitudinal Slope (hz.) 0.75
45 43
Table Frequency (hz.)
40 0.50 Figure 5. Predicted effects of table frequency and longitudinal slope on energy recovery achievable from the treatment of western bituminous coal.
12.0
Product Ash (%)
11.4 10.8 10.2 9.7
50
1.50
48
1.25
45 Table Frequency (hz.)
1.00 43
0.75
Longitudinal Slope (hz.)
40 0.50 Figure 6. Predicted effects of table frequency and longitudinal slope on the product ash for the western bituminous coal.
During the PRB parametric studies, the table products were sampled in 5 splits to provide mass and quality distributions of the material exiting the table. Product and reject streams were calculated as the combination of Splits 1 through 3 and Splits 4 through 6, respectively. Split 4 averaged about 47% ash from all 15 tests, which indicates the presence of misplaced coal that could be recovered. Removing Split 4 from the tailings material resulted in a reject stream containing an average of 83.74% ash-bearing material. However, the goal of the study was to simply determine if a high quality coal could be produced from the current waste material and a corresponding estimate of yield. The FGX Separator proved the ability to achieve a product quality sufficient for meeting market requirements. The ash content was reduced on average from 20.79% to 8.40% while recovering 59.1% of the current waste material. In some tests, the mass yield approached 90% while achieving product ash contents below 10%. 3.2.2 Parametric Studies The parametric studies for the PRB sub-bituminous coal were conducted using the Box-Behnken statistical program and the results were used to derive empirical models describing the response variables, i.e., product ash, product yield and energy recovery, as a function of the operating parameter values. The evaluation involved using the same parametric test program used for the western bituminous coal, with the exception of cross-table slope (CTS) as one of the operating parameters. As a result, the parametric tests were reduced to 17 tests including three were repetitive tests (Tests 10, 14 and 15), which were conducted at the central operating parameter values to quantify the experimental error. Table 7 shows the error associated with the experimental procedure and sample analysis for the Product Ash (%), Product Yield (%) and Energy Recovery (%) response variables. The data revealed the existence of a small degree of randomness for the Product Yield (%) and Energy Recovery (%) measured values in Test 15 and 10, respectively. These values were found to be on the edge of the confidence interval margins with corresponding values of 72.44% 5.82 and 86.19 6.82, respectively. The small difference between the margins and the experimental values were found to be acceptable to be included in the program, +0.12 and +0.07 for Product Yield (%) and Energy Recovery (%), respectively. Table 7. Statistical evaluation of the experimental error realized from the PRB sub-bituminous coal parametric tests. Test Number
Product Ash (%)
Product Yield (%)
Energy Recovery (%)
10 14 15
9.26 9.29 8.57
78.38 69.24 69.71
93.08 81.92 83.56
Mean
9.04
72.44
86.19
95 % Confidence Interval
0.46
5.82
6.82
During the parametric evaluation, the quadratic models were found to be suitable for the prediction of the response variables. However, the non-significant model terms with the highest Prob > F values were excluded from the Product Ash (%) empirical models to meet the Prob > F significance criteria, which was determined to be less than 0.05 for 95% confidence interval. After excluding these model terms, the modified quadratic model was found to be significant for the Product Ash (%) with a Prob > F value of 0.0447. On the other hand, the Product Yield (%) and Energy Recovery (%) quadratic models were found significant including all model terms. The quadratic model with three operating parameters can be derived from Eq. 1 and generalized as:
Y = A1 + A2 (ABF) + A3 (TF) + A4 (LS) + A5 (ABF 2 ) + A6 (LS 2 ) + A7 (ABFxTF) + A8 (ABFxLS) + A9 (TFxLS) (6) where Y is either the energy recovery or the product ash and Ai the corresponding coefficients, ABF the air blower frequency, TF the table frequency and LS the longitudinal slope.
€
3.2.2.1 Model Significance The significance levels of the parameters with the corresponding models for the PRB sub-bituminous coal tests are given Table 8 including the F-statistic values from ANOVA. The F-value was calculated using Eq. (2). The null hypothesis should be rejected in order for the model to be accepted. If the F-value determined from the F-Table exceeds the calculated F0.05 value, which corresponds to the 95% confidence interval, the model is accepted. As shown in Table 8, the F-values of 7.22 and 5.83 exceeded the calculated F0.05 value of 3.29, for Product Yield (%) and Energy Recovery (%), respectively and thus they are accepted. However, for the Product Ash (%) model, the null hypothesis is accepted due to a lower F-value than the calculated confidence interval. The modified quadratic Product Ash (%) model is not significant enough to meet the first criteria of the statistical test and thus it is rejected. The
insignificance of the Product Ash (%) model can be explained by the small range for the product ash values, ranging only from 6.77% to 10.69%, which is apparent from Figure 10 (a) and (b). To determine the significance of the operating parameters and corresponding interactions, the Lack-of-Fit F-statistic (FLOF) was calculated using Eq. (3). Accepting the null hypothesis indicates that the coefficients present in the modified quadratic models area all equal zero. The null hypothesis is rejected and the models are accepted if the calculated Lack-of-Fit F-statistic is less than F0.05 value. Using the degrees-of-freedom values of lack-of-fit and the pure error, the Lack-of-Fit test F0.05 was determined from the F-Table. As shown in Table 8, the F-statistic values of 0.57 and 0.091 are both less than the calculated F0.05 value of 8.73 for the Product Yield (%) and Energy Recovery (%), respectively. This eliminates the probability of falsely rejecting the null hypothesis and confirms the model. As expected, the FLOF value is higher than the calculated F0.05 for the Product Ash (%) model, which confirms its insignificance. Table 8. ANOVA table generated for the response variables associated with the FGX Separator tests from the PRB sub-bituminous coal parametric tests. Source
Degree of Freedom
F Value
Prob > F
F0.05
R2 Radj2
Product Ash Model Model
7
3.43
0.0474
3.68
0.73
Lack of Fit
5
10.77
0.0195
4.72
0.52
Product Yield Model Model
9
7.22
0.0081
3.29
0.90
Lack of Fit
3
0.57
0.6618
8.73
0.78
Energy Recovery Model Model
9
5.83
0.0149
3.29
0.88
Lack of Fit
3
0.091
0.9613
8.73
0.73
To measure the ability of the models for accurately predicting the response values, the R2, coefficient of determination, values were calculated for the PRB sub-bituminous tests. Using Eq. (4), the R2 values for the Product Yield (%) and the Energy Recovery (%) models were found as 0.90 and 0.88, respectively. High values of coefficient of determination values for these models showed a good prediction of the experimental data. Using the same expression in Eq. (5), the Radj2 values for the Product Yield (%) and the Energy Recovery (%) models were calculated to be 0.78 and 0.73, respectively. The comparison between the R2 and the Radj2 values were not close as expected, however, the correlation of the actual and the predicted values were found reasonably close except for a couple outlier values as shown in Figure 11 (a) and (b). 3.2.2.2 Model Simulation and Discussion As discussed earlier, the Prob > F value in the ANOVA table shows a significant model parameter in the statistical analysis for values less than 0.05. For the PRB sub-bituminous coal testing, the quadratic model was found have a significant effect in determining the product yield and the energy recovery values with respective Prob > F values of 0.0081 and 0.0149. From the operating parameters, both the table frequency (TF) and the longitudinal slope (LS) were found significant with the corresponding Prob > F values of 0.0011 and 0.0085, respectively, for the product yield and 0.0011 and 0.0184, respectively, for the energy recovery response variables. Likewise, the air blower frequency (ABF) term was found insignificant for product yield and energy recovery models with corresponding values of 0.468 and 0.8771, respectively. The square interaction of table frequency (TF2) for the product yield and the air blower frequency (ABF2) energy recovery models were also found significant with the corresponding Prob > F values of 0.0129 and 0.0381, respectively.
Three-dimensional plots were constructed for the PRB sub-bituminous coal by simulating the operating parameters and the response variables. As shown in Figures 12 and 13, the effect of significant parameters (LS and TF) on product yield and energy recovery, respectively, showed similar a trend as expected because of their close relationship in the calculation. However, by looking at both figures one can see a more curved plot for the product yield than the energy recovery. This may be explained by the difference of the significance levels for the longitudinal slope parameter between the two models. According to model fitness values explained above, the LS parameter for the product yield model was found more significant than that of energy recovery model with a Prob > F value of 0.0085. As such, the three-dimensional plot shows a more curved interactive effect on the longitudinal slope side of the axis. a)
b)
9 1 Predicted Energy Recovery (%)
Predicted Product Yield (%)
100
95
90
85
80 80
85
90
95
Actual Product Yield (%)
100
8 5 8 0 7 4 6 8 6 8
7 8 8 4 Energy0 Recovery 6 (%) Actual
9 2
Figure 11. Comparison of the experimental and empirical models for: a) product yield (%) and b) energy recovery (%) values of the PRB coal. Similar trends for both models show that decreasing the vibration and increasing the inclination of the table have a positive effect on the yield and the recovery values. The effect of the table inclination, from the reject side to the feed side, on energy recovery is in agreement with that of western bituminous coal tests. As discussed previously, increasing the inclination of the table on the reject side allows misplaced high-quality coal in the middlings reporting to the product side of the table and thus increasing the yield to the product and the energy recovery values. The trends developed from the western bituminous coal models support the improving effect of the lower table frequency on the energy recovery values. To explain the effect of the table frequency, the effect of the longitudinal slope should also be considered. With a table inclined from the reject end to the feed end, the vibrating action of the table needs to be restricted to hinder the movement of the ash-bearing particles from the middlings side to the product side of the table. Decreasing the table frequency will limit the ability of the movement of the low-density (coal) particles to the other parts of the table and thereby will allow them to only report to the product side with the help of the table inclination. This results in more misplaced high-quality coal reporting to the product side of the table and thus higher energy recovery and yield values will be achieved. Model predictions indicate a better significance for the longitudinal slope parameter in the statistical analysis. However, the threedimensional simulations show that the effect of this parameter on both response variables were minimal, improving yield and recovery values by only 2% and 5%, respectively. On the other hand, the table frequency of the table has the greatest impact on separation performance of the FGX Separator, providing a greater improvement on yield and recovery, 17% and 10%, respectively. This explains the importance of the movement of coal on the aerated table.
Mass Yield (%)
91 85 80 75 70
50
1.50
48
1.25 1.00
45
43 0.75 Table Frequency (hz.) Longitudinal Slope (º) 0.50 40
Energy Recovery (%)
Figure 12. Predicted effects of table frequency and longitudinal slope on the product yield for the PRB coal.
100 95 90 85 80
50
1.50
48
1.25 1.00
45
43 0.75 Table Frequency (hz.) Longitudinal Slope (º) 0.50 40 Figure 13. Predicted effects of table frequency and longitudinal slope on the energy recovery for the PRB coal.
3.3 Gulf Coast Lignite Coal 3.3.1 Separation Performance For the Gulf Coast lignite coal, tests were performed to assess the feasibility of the FGX Separator first to reduce the total sulfur content (lb/M-Btu) of coal that is not marketable as a direct ship product. The following priorities were decreasing the mercury content (lb/T-Btu) and improving the heating value (Btu). The coal was relatively low in ash content but high in total sulfur. Same testing program and operating parameters were used with the other sites to quantify the parametric efforts and to obtain optimum separation performance levels. The average separation performance achieved from 17 tests provided a significant sulfur reduction from 1.9% to 1.2%. The sulfur reduction was due to the presence of large coal pyrite particles. Ash reduction was limited and it was only decreased from 6.6% to 4.9%, due to the low amounts of high-density rock in the feed. As such, the improvement in the heating value was minimal with the average increase from 7710 Btu/lb to 7817 Btu/lb. Despite the low feed ash, the average mass yield to the product stream was 79%. The yield values were reflective of misplaced coal in the middlings stream, which was combined with the tailings stream in the analysis. Interestingly, the tailings ash content was lower than expected at 42.25%. Observations during the tests indicated the presence of mostly high-density rock in the tailings. After conferring with the mine personnel, it was learned that the reject was comprised of carbonate minerals, which have significant amount of volatile matter. This fact explains the lower than expected tailings ash values. The concentration of sulfur was very apparently visible and confirmed with an average total sulfur value of 5.17% in the tailings stream. Since pyrite is the main source, the total pyrite content was 10.34% on average. At this testing site, the most significant impact provided by the FGX Separator was the reduction in sulfur and mercury contents. Unlike the ash and sulfur reduction calculations, the total sulfur (SO2) and the mercury (Hg) contents with respect to the energy value (Million Btu = M-Btu and Trillion Btu = T-Btu, respectively) were used to determine the separation performance, rather than using the percent sulfur and mercury content. The average sulfur content in the tailings stream was 5.2% and values greater than 5.3% were realized from several tests, representing a great SO2 (lb/M-Btu) reduction of 35.8% on average. Although the feed sulfur content varied, the FGX Separator provided a consistent product SO2 content of 3.2 lbs/M-Btu. It is generally known that the mercury content in coal is generally associated with the pyritic minerals. This well established observation is apparent in the Gulf Coast lignite coal, as indicated by a large Hg reduction of 54.4%. Although mercury content varied significantly throughout the testing program, mercury content less than 10 lbs/T-Btu was generally achieved from the material in the product stream. The total mercury content in the product of the FGX Separator was as low as 6.65 lb/T-Btu, with the highest yield to the product value of 85.8%, while the tailings mercury content was 570 lbs/T-Btu. The average mercury content in the tailings stream was 350 lb/T-Btu and values greater than 400 lbs/T-Btu were realized from several tests. 3.3.2 Parametric Studies The FGX Separator experimental results for the Gulf Coast lignite coal obtained from the statistically designed 17-test program together with the operating variable values (Table 3.9) were entered into the Box-Behnken statistical program to develop empirical models. These models were then used to describe the response variables (i.e., product ash, energy recovery and mercury reduction) as a function of the operating parameter values (i.e., air blower frequency, table frequency and longitudinal slope). In the 17-test program, five repetitive tests (Tests 1, 2, 3, 5, and 6) were conducted at the central parameter values to show the degree of repeatability of the test results by calculating the normal error associated with the experimental procedure and sample analysis. The variability in the Mercury Reduction (%), Energy Recovery (%) and Product Ash (%) responses resulting from experimental error are shown in Table 9. According to the data in Table 9, randomness in the response variables indicates the presence of experimental error. The measured response variables show some degree of variability due to the experimental error for the corresponding margins of 64.07% 3.48, 4.66% 0.28 and 82.10% 4.04 for Mercury Reduction (%), Product Ash (%) and Energy Recovery (%), respectively. Given the magnitude of the response variables, the experimental error is respectable. A quadratic expression was found to be significant for the prediction of the Mercury Reduction (%) values, in the same form as outlined in Eq. (6). On the other hand, for the product ash and energy recovery response variables, quadratic model was modified to be significant (Prob > F value less than 0.05) by excluding some terms from the equation. For the Product Ash (%) quadratic model, the interaction between the air blower frequency and the longitudinal slope (ABF*LS) was found to have the highest Prob > F value of 0.528, representing a non-significant model term. Excluding this model term resulted in a modified significant quadratic model with a Prob > F value of 0.0467. For the Energy Recovery (%) quadratic model, the squared interaction of the table frequency term (TF2) and the interaction between the table frequency and the longitudinal slope (TF*LS) were found to have non-significant Prob > F values of 0.737 and 0.910, respectively. As a result of excluding these model terms, a modified significant quadratic model with a Prob > F value of 0.0259 was used to predict the Energy Recovery (%) values. Table 10 shows the significance levels of the parameters with corresponding models for Gulf Coast lignite parametric tests.
Table 9. Statistical evaluation of the experimental error realized from the Gulf Coast lignite parametric tests. Mercury Reduction
Energy Recovery
Product Ash
(%)
(%)
(%)
1
62.29
75.03
4.50
2
67.66
82.67
5.03
3
67.12
84.28
4.89
5
65.24
87.46
4.23
6
58.05
81.08
4.64
Mean
64.07
82.10
4.66
3.48
4.04
0.28
Test Number
95 % Confidence Interval
Table 10. ANOVA table generated for the response variables associated with the FGX Separator tests from the Gulf Coast lignite parametric tests. Source
Degree of Freedom
F Value
Prob > F
F0.05
R2 Radj2
Mercury Reduction Model Model
9
13.04
0.0013
3.68
0.94
Lack of Fit
3
1.48
0.3462
6.59
0.87
Product Ash Model Model
8
3.53
0.0467
3.44
0.78
Lack of Fit
4
0.34
0.8420
6.39
0.56
Energy Recovery Model Model
7
4.15
0.0259
3.29
0.76
Lack of Fit
5
0.44
0.8021
6.26
0.58
3.3.2.1 Model Significance The Gulf Coast lignite test results showed that a significant sulfur reduction was achieved by rejecting the pyritic particles in the feed. However, due to very narrow range of product sulfur values, 1.12% to 1.31%, a significant model fit for product sulfur and sulfur reduction was not found. On the other hand, the Mercury Reduction (%) model was found to be the most significant among others, including all the model parameters. Due to the modifications on the quadratic model to achieve a significant model for product ash and energy recovery responses, degrees-of-freedom values are different between the response variables as shown in Table 10. Again, the F-Value is evaluated by testing the null hypothesis using Eq. (2). The model is accepted and thus the null hypothesis is rejected if the F-Value exceeds the calculated F0.05. F0.05 is determined from the F-Table using the model degrees-of-freedom (k) and the difference between the measured responses and total degrees-of-freedom values (n-k+1) in the quadratic model. As shown in Table 10, all the F-Values are greater than the calculated F0.05 values and therefore the models were considered adequate.
The null hypothesis was also tested using the Lack-of-Fit F-statistic (FLOF), using Eq. (3). For a model to pass the lack-of-fit test, the calculated FLOF, should be less than F0.05 value. F0.05 for the lack-of-fit test is determined from the F-Table using the lack-of-fit and the pure error degrees-of-freedom values. The degrees-of-freedom value for the pure error is same for all models (m=4), however the lack-of-fit values are different due to the modifications on the quadratic model. The calculated FLOF values for all models, shown in Table 10, were less than the F0.05 value with low Prob > F values. As a result, the FLOF is found to be non-significant and the models showed an adequate fit. The empirical models were tested to estimate the response variables using the coefficient of determination, R2. The empirical models were found to accurately estimate the mercury reduction values with a value of 0.94. R2 values for the product ash and energy recovery models were found to show poor performance to explain the parameter interactions considered in the models due to low significance levels, as stated above. These models were forced to be significant by excluding some model terms, which also affected the coefficient determination, R2. The R2 values were dropped from 0.79 to 0.78 and from 0.77 to 0.76 for the Product Ash (%) and Energy Recovery (%) models, after the non-significant model terms were removed from the quadratic models. The adjusted coefficient determination, Radj2, was calculated for the Gulf Coast lignite parametric tests by taking into consideration the presence of excessive terms within the quadratic model. Using the Eq. (5), Radj2 is calculated and compared to R2 value, where a small difference between these values is realized as an adequate prediction of the models for the response variable values. Table 4.36 shows the R2 and Radj2 values for the corresponding response models. Clearly, the Mercury Reduction (%) model can be described as the best-fit model for the parameter values. Despite the fact that the other two models were adjusted to meet the significance level, they showed good predictions to fit the models. The correlation between the actual and the predicted response variable values for Mercury Reduction (%) and Energy Recovery (%) are shown in Figures 14 a) and b). a)
b)
89.18 Predicted Energy Recovery (%)
Predicted Hg Reduction (%)
68 58
48
38
83.87
78.56
73.24
28 67.93 28
38
48
58
68
Actual Hg Reduction (%)
67.93
73.24 78.56 83.87 Actual Energy Recovery (%)
89.18
Figure 14. Comparison of the experimental and empirical models for: a) product yield (%) and b) energy recovery (%) values of the Gulf Coast lignite coal.
3.3.2.2 Model Simulation and Discussion As stated previously, the significance of a model parameter in the parametric studies is decided based on the Prob > F value of less than 0.05. For the Gulf Coast lignite coal testing, the quadratic model was found have a significant effect in determining the mercury reduction, product ash and the energy recovery values with respective Prob > F values of 0.0013, 0.0467 and 0.0262. Mercury reduction model has shown the best correlation between the experimental results and model predictions with air blower frequency (ABF) and its square term (ABF2) having the greatest impact with corresponding Prob > F values of 0.0009 and 0.0006, respectively. As predicted from the three-dimensional simulation, increasing both the table frequency of the table and the air blower frequency had a positive effect on the mercury reduction, as shown in Figure 15. While keeping the longitudinal slope parameter at a constant value of 1.0º, mercury reduction was improved from about 27% to 40% by increasing the table frequency from 40 to 50 hz. Increasing the table frequency allows the pyrite particles to report to the back plate easily and follow the helical pattern where only the high-density particles were discharged to the reject end of the aerated table.
66 Hg (lb/T-Btu) Reduction (%)
56 46 36 26
60 58 55
Air Blower (hz.)
53 50 50
43
40
45 48 Table Frequency (hz.)
Figure 15. Simulated effects of air blower and longitudinal slope on the Hg reduction for Gulf Coast lignite coal. During the Gulf Coast lignite parametric tests, it was noted that there was a significant amount of misplaced middlings material, containing both the high-density (rock) and light-density (clean coal) particles. As such, in order to have a better stratification, the simulated plot suggests decreasing the longitudinal slope on the aerated table to achieve a clean coal product, as shown in Figure 16. For the product ash model, instead of the general rule-of-thumb 95% confidence interval; 90% confidence interval (Prob > F value of less than 0.1) was found to be significant to describe the effects of longitudinal slope and table frequency parameters with corresponding Prob > F values of 0.0954 and 0.0986, respectively. While keeping the air blower frequency at a value of 55 hz., very high and very low table frequencies showed increasing values in the product ash content. This shows the same trend with the western U.S. bituminous coal simulations where decreasing the table frequency initially improves the product ash content due to a better stratification on the feed side of the table. However, further decreasing the vibration of the table creates product ash gradients and more misplaced material in the middlings, which increases the product ash content values.
Product Ash (%)
5.4 5.1 4.8 4.5 4.2
50 48 45 Table Frequency 43 (hz.) 40
1.25
0.50
1.50
1.00 0.75 Longitudinal Slope (º)
Figure 16. Simulated effects of air blower and longitudinal slope on the product ash for Gulf Coast lignite coal.
86 Energy Recovery (%)
82 78 74 70
50 47 45 Table Frequency 43 (hz.) 40
1.25
0.50
1.50
1.00 0.75 Longitudinal Slope (º)
Figure 17. Simulated effects of air blower and longitudinal slope on the energy recovery for Gulf Coast lignite coal.
Longitudinal slope and table frequency parameters were found significant for the energy recovery model corresponding Prob > F values of 0.0204 and 0.0131, respectively. As shown in Figure 17, the energy recovery was increased from around 70% to 85%, while increasing the longitudinal slope from 0.5 to 1.5 degrees. The energy recovery improvement may be explained due to an increasing resistance for particle movement down the table, which results in coal in the middlings stream shifting to the product stream. In addition to this, the effect of the table frequency on the energy recovery was also realized where almost 10% improvement occurred with a decrease of the vibration of the table from 50 hz. to 40 hz. This was expected since the low table frequency limits the action of the particles moving immediately to the other parts of the table, allowing more time for stratification process which results in high energy recovery values. 4. Evaluation of the Parametric Studies 4.1 Effects of Parameters During the parametric studies conducted on eastern bituminous, sub-bituminous and lignite coals, three parameters (i.e., table frequency (TF), longitudinal slope (LS) and air blower frequency (ABF)) were used to test their effect on the response variables. In the first test site for western bituminous coal, the effect of cross table slope (CTS) parameter was also investigated. After evaluating the parameters with the associated model fitness values, cross-table slope was removed from all subsequent testing sites, on the basis of its overall impact on performance and a need to reduce the number of parametric tests. When evaluating four parameters, a 3-level statistically-design test program required a total of 29 experiments. After removing the cross table slope, the optimization test program was reduced to seventeen tests. Although their effects have been discussed in detail in earlier sections of the thesis, a summary of their impact on separation performance is provided in the following paragraphs. Table frequency (hz.) controls the vibration of the table, which is important for the transportation, and discharge of the material on the deck surface. Decreasing the table frequency (slowing the vibration) of the aerated table reduces the transportation rate of material toward the discharge end that tends to increase product ash but improves energy recovery. The motion also assists in the stratification process. A high table frequency value results in an elevated transfer rate of material toward the reject end of the table, which could lead to a loss in energy recovery. As such, a feed coal containing less than 50% ash-bearing material should use a moderate table frequency (i.e., TF≈45 hz.) whereas a high value should be used for high ash coals where a clean product is desired. Longitudinal slope (degree) controls the horizontal inclination of the table along the length of the table, i.e., from the reject end to the feed end. Increasing the longitudinal slope (LS > 1º) increases the resistance to the material moving toward the reject end of the table. This condition is desired when the objective is to remove as much rock as possible while recovering near 100% of the 1.6 float material. On the other hand, a flatter table deck (LS ≈ 0º) leads to a low product ash content by enhancing the movement of high density particles toward the reject side and spreading the middlings material toward the reject end. A near horizontal deck position is desired when producing a clean coal product from high ash feed coal. For example, in the western bituminous and Gulf Coast lignite coal operations, the objective was to produce a marketable product by removing the out-of-seam dilution material and reducing the sulfur and mercury content, respectively. In this scenario, to move all the ash-bearing material down the table without mixing with product, the parametric study results suggested a near flat table deck (LS ≤ 0.5º) as shown in Figure 6. On the other hand, from the eastern bituminous and PRB sub-bituminous coal testing, the objective was to maximize the energy recovery with no loss of coal to the other streams. In this case, the product ash content was less important. Therefore, high table slope values (LS ≥ 1º) were found to be optimum to capture all the possible misplaced coal particles. Air blower frequency (hz.) is important to fluidize and transport the low-density particles to the product stream. The fluidization of the coal particles is crucial for an effective separation. An insufficient amount of air to the table will reduce the fluidization of the light particles thereby, resulting in poor separation efficiencies. Interestingly, during the PRB sub-bituminous coal testing, increasing the table frequency while decreasing the air blower frequency was found to significantly improve the energy recovery. This demonstrates the importance of the parameter interactive effects on the separation performance achievable by the FGX Separator. 4.2 Summary of Significant Parameters The significant parameters for all the parametric studies conducted on various coal types are summarized in Table 11. The parameters shown with an asterisk (*) are found to be significant for 95% confidence interval (Prob > F value of less than 0.05) while the remaining parameters were included based on a significance level of 90% (Prob > F value of less than 0.1). For the Gulf Coast lignite coal, including the 90% confidence interval was essential to describe the product ash model; otherwise the model would be limited to other two parameters. Please note that squared interaction of the same parameter and interactions between different parameters are not listed in the table below to show the effect of the significant model parameters in the program only and for simplicity of the analysis.
Table 11. Significant parameters (parameter interactions are not included) for the corresponding models for each testing site. Testing Site
Model – Significant Parameter
Western Bituminous
Product Ash – TF*, LS*
Coal (UT)
Energy Recovery – TF*, LS*, ABF*
PRB Sub-Bituminous
Product Yield – TF*, LS*
Coal (WY)
Energy Recovery – TF*, LS*
Gulf Coast
Product Ash – TF, LS
Lignite Coal
Energy Recovery – TF*, LS*
(TX)
Mercury Reduction – ABF*, TF*
It is clear from Table 11 that table frequency and longitudinal slope are critical parameters in the control of product ash content for bituminous coal. However, air blower frequency or the amount of fluidization air is also important for the determination of product ash for lower grade coals, which tend to have a lower particle density tan bituminous coals. As expected, all three parameters are important for maximizing recovery for all coal ranks that were tested. In addition, the most significant parameter tested in all statistical programs is found to be the table frequency (TF), which fit all the models in each testing site. Longitudinal slope (LS) was also found to be very significant except the mercury reduction model in the Gulf Coast lignite coal. In this model, the longitudinal slope itself was not significant; however, its interaction with the air blower frequency met the significance criteria. This can be explained by the behavior of the sulfur- and mercury-bearing pyritic particles on the table. These particles, with specific gravities around 5.0, remain in contact with the table upon entering the table and report to the end of the table with other reject material. Increasing or decreasing the longitudinal slope will not change the direction of their movement, which follows the riffles on the table. On the other hand, when the table slope is decreased and the frequency is increased, only the clean coal particles will be able to report to the product discharge. The significance of the ABFxLS interaction, which is not listed above, with a Prob > F value of 0.0036 explains the importance of these parameters for the product sulfur and mercury contents. 4.3 Optimization Studies The optimization test results for all coal types were performed using the Design-Expert statistical program. In the optimization program, the optimum performances and corresponding parameter values were determined according to the priority of the test program, i.e., maximum rock rejection while recovering nearly 100% of the coal for eastern bituminous coal tests. Lower and upper limits of the parameter and response values were kept in the operating model and experimental result range, respectively. Another set of response variable results were found by using the parameter mid-points in the testing program. Additionally, the “Point Prediction” statistic of the design program allows the determination of the model parameter values for a chosen response variable. Based on the desired response variable results, the parameter values needed to achieve the objective were determined. The initial test program was performed on bituminous run-of-mine coal at an operation in Utah. The objective of the test program was to produce a clean coal product that was sufficient in quality to meet specifications for a steam coal contract. To meet this objective, the optimum operating model was found by minimizing the product ash content and maximizing the energy recovery in the experimental result range. The maximum priority was given to the product ash model, since the objective was to achieve a clean coal product. As shown in Table 12, low table slope and high air frequency values were found to significantly improve product ash and energy recovery values. As expected, cross-table slope values around 8.5º provided low product ash values. During the 29-test program, the product ash values ranged from 9.5% to 12.5%. Thus, this range was used predict the operating conditions for the given product ash values. The resulting energy recovery product ash values follow the general “recovery versus ash” curve. Third optimum model prediction resulted in the best among others, appearing on the elbow of the curve. In summary, high table frequency and longitudinal slope values are needed to achieve a clean coal product with low ash content. High fluidization rates are required for recovery purposes.
Table 12. Simulated optimum performances over a range of product ash values when treating western U.S. bituminous coal tests; average feed ash = 18.2%. ABF (hz.) 60
TF (hz.) 50
CTS (º) 8.6
LS (º) 0.6
Product Ash (%) 9.5
Energy Recovery (%) 77
60
48
8.5
0.9
10.5
83
60
41
8.5
1.5
11.5
93
60
40
8.9
1.5
12.5
95
The processing of PRB sub-bituminous coal with the FGX Separator evaluated the feasibility of rejecting the out-of-seam rock due to contamination during extraction. The production of a marketable clean coal product was the main objective. However, the product ash model was found insignificant in the statistical model due to its very narrow range of values, i.e. from 6.7% to 10.7%. The narrow range signifies the ease of which the FGX Separator rejected the out-of-seam material, thereby consistently providing low product ash content values. The optimization process targeted maximization of energy recovery. The simulated results in Table 13 agree with previous results in that, low table frequencies and high longitudinal slopes provide maximum recovery values. Table 13. Simulated optimum performances over a range of product ash values when treating PRB sub-bituminous coal tests; average feed ash = 20.8%. ABF (hz.) 56
TF (hz.) 49
LS (º) 0.6
Energy Recovery (%) 80
57
44
1.1
90
58
40
1.5
100
The objectives associated with the treatment of the Gulf Coast lignite were to maximize sulfur and mercury rejection and upgrade the energy value, while achieving a respectable energy recovery. In the statistical program, the sulfur reduction model was found insignificant due to the relatively small changes in the absolute sulfur content values. However, sulfur rejection was significant, with over 30% removed. Likewise, the ash content of the feed was low at 6.6%, which limited the amount of ash reduction and heating value improvement that could be achieved. The mercury reduction model was found to be the most significant in the statistical analysis. Based on this outcome, the objective of the optimization exercise was to maximize mercury reduction while achieving high recovery values, as shown in Table 14. Table 14. Simulated optimum performances over a range of product ash values when treating Gulf Coast lignite coal tests; average feed ash = 6.6%. ABF (hz.) 51
TF (hz.) 44
LS (º) 1.0
Mercury Reduction (%)
Product Ash (%)
Energy Recovery (%)
45
4.9
84
52
45
1.2
55
4.8
83
54
45
1.3
65
4.7
82
5. Summary and Conclusions: The FGX Separator provides a dry, density-based separation that utilizes the combined separating principles of an autogenous fluidized bed and a table concentrator. The project evaluated the dry particle separator at several mining operations across the U.S. for the treatment of several ranks of coal. The objectives of the testing programs at each site varied and included: 1) the production of clean coal having qualities that meet contract requirements and 2) maximization of the amount of high-density rock rejected prior to transportation and processing. The study involved the installation of a 5-t/hr pilot-scale unit of the FGX Separator for the in-field tests and a detailed parametric study at each site to ensure that optimum performances were realized for each coal. The objective of the parametric studies was to evaluate the effect of air blower frequency, table frequency, longitudinal slope and cross-table slope on the separation performance. A Box-Behnken statistical program, which resulted in seventeen tests for most test programs, used to evaluate the various operating parameters considered. Utilizing the results of the parametric studies, quadratic models were developed to describe the parameter and parameter interaction effects on the product ash content and energy recovery. For the special case of Gulf Coast lignite treatment, mercury reduction was also investigated and correlated with the parameter effects. For coals containing little or no material having a density between 1.6 RD and 2.0 RD, the FGX Separator was evaluated to produce a clean coal product that meets utility contract specifications. For PRB sub-bituminous coal, the ash content was reduced from about 20.8% to 8.4%, on average. Similar results were obtained for bituminous run-of-mine coal at a western bituminous coal operation. The dry air table separator also reduced the total sulfur and mercury contents of Gulf Coast lignite coal by 35% and 54%, respectively. The results from the best performances achieved over a range in product ash contents for the western bituminous coal tests showed that the FGX Separator has the ability to reduce ash content to values below 12% while maintaining a high level of recovery. In addition, rock is effectively rejected into the tailings stream without losing coal as indicated by the high tailings ash. And the results indicated that the middlings stream required retreatment to recover the misplaced, low-density coal particles. The parametric tests for western bituminous, PRB sub-bituminous and Gulf Coast lignite coals showed that the most significant factors to determine the product ash and energy recovery values were the table frequency (TF) and the longitudinal slope (LS). Increasing the table frequency (TF≥45 hz.) was found to eliminate the probability of the reject material staying near the front end of the table due to insufficient movement. This improves the stratification of the clean coal material along the feed end of the table and decreases the product ash content. Additionally, decreasing the longitudinal slope (LS≤0.5º) to form a flatter table deck enhances the removal rate of ash-bearing material toward the reject port located at the end of the table. The product ash model was found insignificant for the PRB sub-bituminous coal due to very close range of the experimental data. In nearly all tests, the ash content was reduced from 20.8% to around 8.4%. The variability in the product ash content was very small, which indicates that composition of the feed includes little or no middling-type particles. To achieve high energy recovery values, increasing the table slope from the reject end to the feed end (LS≈1.5º) is required. Decreasing the table frequency (TF<45 hz.) while increasing the air blower frequency (ABF>55 hz.) was found to have a positive effect on the energy recovery, which allows the coal particles to sufficiently fluidize on the table without excessive movement. For the Gulf Coast lignite coal, one of the main objectives was to reduce the sulfur and mercury content of the coal by rejecting the coarse pyrite particles from the feed. However, the sulfur reduction model was found insignificant due to large variation in the feed sulfur from test to test. While evaluating the mercury reduction model, in addition to table frequency and longitudinal slope variables, air blower frequency was also found significant. Increasing the air blower frequency (ABF>55 hz.) allows only the clean coal particles to fluidize and helps the stratification on the table. This also minimizes the misplacement of coal particles in the middlings stream thereby providing a better separation from the sulfur/mercury-bearing particles. The commercial FGX Separator has the ability to treat particles from 80 mm to 3 mm, however, the optimum upper and lower size limits for the 5-t/hr pilot-scale unit was found to be 51 mm and 6.4 mm, respectively. It was noted that, in the eastern bituminous coalfields, greater amounts of high-density material exist in the larger particle size fractions of run-of-mine coal. On the other hand, when treating the ‘rib’ coal, a significant amount of fine material (<6.4 mm) is being screened out because of the current optimum size range for the pilot-scale unit. Therefore, modifications need to be made to extend the lower and upper size limits. 6. Acknowledgements: The research discussed in this publication was funded in part by the U.S. Department of Energy through the Mining Industry of the Future program (DE-FC26- 05NT42501). The authors greatly appreciate Ed Thompson, Robert Bratton, Dr. Datta Patil and Carlos
Munoz-Diaz for their work on the field and the assistance provided by the personnel of the mining companies at each of the test sites. 7. References: Arnold, B. J., Hervol, J. D. and Leonard, J. W., 2003, “Dry Particle Concentration,” Coal Preparation 5th Edition, Society of Mining, Metallurgy and Exploration, Littleton, CO, Chapter 7, Part 3, pp. 486-496. Honaker, R. Q., Luttrell, G. H. and Lineberry, G. T., 2006, “Improved Coal Mining Economics Using Near-Face Deshaling”, Minerals and Metallurgical Processing Journal, Vol. 23. No. 2, pp. 73-79. Honaker, R. Q., Saracoglu, M., Thompson, E., Bratton, R., Luttrell, G.H. and Richardson, V., 2007, “Dry Coal Cleaning Using The FGX Separator”, Proceedings, 24th Annual International Pittsburgh Coal Conference, Johannesburg, South Africa. Kelley, M. and Snoby, R., 2002, “Performance and Cost of Air Jigging in the 21st Century,” Proceedings, 19th Annual International Coal Preparation Exhibition and Conference, Lexington, Kentucky, pp. 175-186. Li, G. and Yang, Y., 2006, “Development and Application of FGX Series Compound Dry Coal Cleaning System”, China Coal, Technology Monograph of the Tangshan Shenzhou Machinery Co., Ltd., pp. 17-28. Lu, M., Yang, Y. and Li, G., 2003, “The Application of Compound Dry Separation Technology in China”, Proceedings, 20th Annual International Coal Preparation Exhibition and Conference, Lexington, Kentucky, pp. 79-95. Osborne, D. G., 1988, “Pneumatic Separation,” Coal Preparation Technology, Graham and Trotman, Norwell, Massachusetts, pp. 373386. Saracoglu, M., 2007, “Dry Coal Cleaning Using The FGX Separator”, Master’s Thesis, University of Kentucky, Lexington, Kentucky. Weinstein, R. and Snoby, R., 2007, “Advances in Dry Jigging Improves Coal Quality,” Mining Engineering, Vol. 59, No. 1, pp. 2934.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
Improving the Efficiency of Lignite Drying Wayne S. Seames, Ph.D. Professor Department of Chemical Engineering 241 Centennial Drive Stop 7101 Grand Forks, ND 58202-7101 Phone: (701)777-2958 Fax: (701)777-3773 Email: [email protected] Carlos J. Bucaram Graduate Student Department of Chemical Engineering 241 Centennial Drive Stop 7101 Grand Forks, ND 58202-7101 Phone: (701)777-4244 Fax: (701)777-3773 Email: [email protected] Steven A. Benson, Ph.D. Professor Department of Chemical Engineering 241 Centennial Drive Stop 7101 Grand Forks, ND 58202-7101 Phone: (701)777-5177 Fax: (701)777-3773 Email: [email protected] Srivats Srinivasachar, Sc.D. President Envergex, LLC 10 Podunk Road Sturbridge, MA Phone: (508) 347-2933 Email: [email protected]
Abstract: The feasibility of using mechanical vapor recompression (MVR) to increase the efficiency of drying North Dakota lignite coal has been examined. MVR is used to recover the latent heat of vaporization by compressing the moisture-laden process streams from the drying process. The work conducted includes: 1) compiling and analyzing lignite coal data from lignite-fired power plants; 2) identifying achievable drying goals for the plant, and 3) performing computer simulations to determine potential efficiency gains through the use of MVR. Scenarios examined included drying the lignite from an initial moisture content of 38% to less than 10% using direct MVR, indirect MVR, and combined indirect and direct MVR. In addition, options for utilization of waste heat in the system were examined.
Introduction North Dakota lignite on average has a moisture content of 38 percent, with a range between 28.1 and 45.1 percent. The goal of the work is to determine the technical feasibility of using mechanical vapor recompression technology to increase the drying efficiency on small-scale (20 to 50 MW) lignite-fired combined heat and power systems. A method to improve the energy efficiency of moisture reduction in brown coal is a technology called mechanical vapor recompression, or MVR (Stamatelopoulos, 2007). This method recovers the latent heat of evaporation, and uses its energy to further dry more coal. The process efficiency is boosted by recompressing the steam driven out of the coal, and reusing its latent heat to further dry the coal. This paper reports the results of computer simulations of using the MVR technology to dry coal for a small North Dakota lignite-fired combustion system. Materials and Procedures ChemCAD 6.1.0 simulation tool was used in order to determine the thermodynamic feasibility of a pre-drying plant sized for a 20 MW combined heat and power plant. Heat and material balances were performed on the drying system in order to determine the unit operations necessary for the drying operation. In addition, the operational parameters needed to run the pre-drying technology for future testing was also determined. The simulations were conducted for two different scenarios for a 20 MWe equivalent plant. The first scenario involved the drying of lignite from an initial moisture content of 38 percent to a final moisture content of 28 percent. The second scenario involved drying of lignite from to a final moisture content of 8 percent. In all of the scenarios, lignite with physical and chemical characteristics shown below was used as input into the model. The composition of the lignite used in the simulations is summarized in Table 1. Table 1. Properties of the lignite used in simulations
Results and Discussion Several simulations were performed and can be found elsewhere (Bucaram, 2010). As an example, the simulation for the drying from 38 to 8 percent moisture level is shown on Figure 1. In this case, the wet lignite is introduced into a crusher where its mean particle size is reduced to less than about 3 mm. After a series of heat exchangers, the temperature of the wet lignite is raised from 20oC to about 77oC, before going into the dryer. After the dryer, the dried coal is cooled to about 61oC. The water that was evaporated from the coal leaves as steam from the dryer and is then recompressed to about 3.2 bar. This increases the temperature of the steam and the saturation temperature at which it condenses. In the dryer, the high pressure steam is condensed and the heat of condensation is indirectly transferred to provide for the latent heat of evaporation of the moisture in the wet coal to be dried.
Figure 1. ChemCAD simulation for a 20 MW lignite coal pre-drying plant. Scenario: Initial moisture content-38%; Final moisture content-8%
Table 1. Simulation results for the reduction of moisture in coal from 38 to 28% and 38 to 8%
Figure 3. Power requirements for drying ND lignite to selected moisture level and comparison to work conducted by Kakaras and others (2002) Conclusions The simulations using mechanical vapor recompression showed a reduction on the ratio of energy consumption per ton of evaporated water compared to the case where the evaporated moisture is exhausted into the atmosphere. For the scenario where the moisture content of the lignite was reduced from 38 to 28 percent, the calculated energy consumption was 102 kWhe/ton H2O evaporated. In the second case, where the moisture content was reduced from 38 to 8 percent, the energy consumption was calculated to be about 89 kWhe/ton H2O evaporated. The electrical energy demand when converted to a thermal energy equivalent would translate to about 474 Btu/lb of water evaporated (102 kWhe/ton H2O) using MVR technology. As a complement it is worth noting that the above equivalent thermal energy demand is much below the typical demand of about 1300 Btu/lb of water evaporated, which evaporation systems usually need to apply, if the energy in the evaporated water is not recycled.
Acknowledgement The authors thank the Lignite Energy Council and the North Dakota Industrial Commission for providing us with funds to carry out this project.
References Bucaram, C.J., Computer Simulation Study for the External Pre-drying of Lignite Coal, Master Thesis, Chemical Engineering, University of North Dakota, 2010. Kakaras, E., P. Ahladas, and S. Syrmopoulos, Computer simulation studies for the integration of an external dryer into a Greek lignite-fired power plant. Fuel, 2002. 81(3): p. 583-593. Stamatelopoulos, G. WTA offers bid efficiency gains. Modern Power Systems, 2007: p. 17-21.
DEACTIVATION OF IRON BASED FISCHER‐TROPSCH CATALYST: A CRITICAL PROBLEM Deena Ferdous and Belma Demirel Rentech Inc., Rentech Energy Technology Center, 4150 East 60th Avenue, Commerce City, CO 80022, USA [email protected] Abstract Fischer‐Tropsch synthesis (FTS) catalytically converts syngas, a mixture of CO and H2 to hydrocarbons through a surface polymerization reaction. Iron‐based FTS catalyst precursors consist of Fe2O3 nanocrystals. Promoters are often added to these nanocrystals in order to improve catalyst performance. However, the exact structural composition of the active sites and the deactivation mechanism of this catalyst are still not clearly understood. Iron based Fischer‐Tropsch catalysts undergo two types of deactivation during use. One is due to the physical degradation of the catalyst because it destroys the integrity of the catalyst. This is considered as a serious problem because it forms fine particles that cannot be easily separated from the heavy wax products. It also increases the viscosity of the slurry which may cause extreme pressure drops. The other deactivation pathway can be classified as being due to chemical or phase change. Moreover, carbon deposition, pore blocking, sintering and deposition of other chemicals such as sulfur on the catalyst are also considered as the possible reasons of Fe based Fischer‐Tropsch catalyst deactivation. 1. Introduction FTS is a strongly exothermic reaction. Often the catalysts and reaction conditions are tuned in such a way to obtain a wax consisting of long hydrocarbon chain products (C20+). A typical catalyst contains promoters like copper to enhance catalyst reducibility, potassium to improve CO dissociation and silica or zinc oxide to improve the attrition resistance of the catalyst which in turn also improves the amount of iron atoms interacting with the gas phase. Catalysts are normally treated in H2, CO or syngas to convert Fe2O3 to its active form. It is generally recognized that a complex mixture of iron phases such as metallic iron, carbide iron and iron oxide is formed after activation and during FTS reaction. The iron based catalysts lose activity with time on stream and the rate of deactivation depends on the presence/absence of promoters. It is believed that carburized iron is an active catalyst. However, it is still disputed in the literature whether or not bulk carbide phases themselves play an active role during FT synthesis. During pretreatment and FT synthesis of iron based FT catalyst bulk iron carbide is formed because the activation energy for carbon diffusion (43.9‐69.0 kJ/mol)1 is lower than for the FT reaction (89.1 ± 3.8 kJ/mol). Many researchers claim that iron carbides are absolutely necessary for a FTS catalyst to be active. On the other hand other claims that metallic iron2, Fe2O33 and Fe3O44 are the active phases in FTS. The iron based FT catalyst system is one of the oldest and perhaps most studied systems known in the field of heterogeneous catalysis. However, even after more than of 80 years of research many important questions remain unanswered. The purpose of this paper is to review the deactivation mechanisms of iron based FT catalyst. 2. Iron Carbide Formation and FT Reactions For iron based FT catalyst, iron carbides are normally formed during activation treatment or under FT reaction conditions. Hexagonal ε‐Fe2C can be formed by carburizing finely divided iron or iron oxide powder at low temperature (typically 170°C). Hexagonal ε‐Fe2C is transformed into χ‐Fe5C2 Hägg carbide above 250°C, which is eventually decomposed into θ‐Fe3C cementite above 450°C5.
Dependent on the gas composition and temperature, the following reactions takes place during catalyst pretreatment5. FexOy + yH2 xFe + yH2O (1) FexOy + yCO xFe + yCO2 (2) CO can be adsorbed on the metallic iron active surface sites (represented by *) and subsequently gets dissociated. The carbon species can then react with iron species and form bulk iron carbide. *+ CO* C* + O* (3) yC* + xFe FexCy (4) Equation (4) takes place readily on iron at typical FT reaction temperatures because of the low apparent activation energy of this reaction. Dissociated carbon can also take part in two other reactions. First, in the presence of dissociated hydrogen, surface carbide species can be hydrogenated according to the following reaction. C* + xH* CH*x + x* (5) The CHx species can then further react with each other on the surface and form longer chain products. Secondly, when more and more carbon species are available on the surface and adsorbed hydrogen atoms are sparse, a side reaction can take place. C* + C* Ci* (6) Where, Ci is the inactive surface carbon species. The build‐up of these species on the catalyst surface can lead to the blocking of the active surface and lowering the overall catalyst activity. 3. Iron Catalyst Deactivation During FTS, Fischer‐Tropsch catalysts undergo deactivation due to the physical degradation and also due to the chemical change. In the FTS, inherent catalyst deactivation and operational catalyst deactivation will occur. Inherent catalyst deactivation occurs due the exposure of the catalyst to the realistic (clean) Fischer‐ Tropsch conditions such as temperature, partial pressures of H2, CO, CO2 and H2O and the possible product compounds whereas additional deactivation routes may play a role in the operation of the catalyst. Four main mechanisms of deactivation have been described in the literature5. 1. Transformation of active iron phases (χ‐carbide, ε‐carbide, ε’‐carbide, more generally FexC of metallic ‐Fe) to less active or inactive phases. Most groups report that the active phase is gradually oxidized to magnetite (Fe3O4). On the other hand, the other groups believe that catalyst deactivation occurs due to the interconversion of one kind of iron carbide to other kind of iron carbide. 2. In the second mechanism, it is proposed that the deactivation occurs due to the pore plugging and deposition of inactive carbonaceous compounds (graphite carbon, amorphous carbon coke) on the catalyst surface which in turn limiting the contact between reactant gases and catalytically active phase. 3. Some groups claim that sintering, the loss of catalytic surface area due to reopening or migration and coalescence of the iron phase is a main cause for the loss of catalyst activity. 4. Finally, iron based catalyst could be poisoned and deactivated by different chemical compounds present in syngas such as sulfur, carbonyl, ammonia, chlorine etc. The first three mechanisms are known as inherent catalyst deactivation mechanism whereas the 4th mechanism is considered to be as operational deactivation6. Some researchers claim that a single mechanism is responsible for catalyst deactivation; on the other hand, most groups propose that catalyst deactivation occurs due to the contributions of more than one mechanism. 3.1. Deactivation through phase change For iron based catalyst whether the catalyst exists as carbides or as oxides depends on the reducing behavior of the atmosphere present in the reactor7. Catalyst is exposed to an atmosphere of hydrogen, carbon
monoxide, and the reaction products such as water, carbon dioxide etc. It is believed that the oxidation of iron carbide and metallic iron occurs in the presence of water and carbon dioxide. O’Brien et al.7 from the thermodynamic data analysis on the iron/iron oxide system, reported that oxidation conditions exist at a H2O/H2 ratio>0.03 or a CO2/CO ratio>2.8, and the Fe3O4 is the stable phase compared to metallic iron. However, they have not reported whether these conditions are applicable for high temperature or low temperature FTS. During FTS, iron carbide can be oxidized to Fe3O4 according to the following equations if H2O/H2 or CO2/CO ratios are high enough8. FexCy + yH2O yC* +yH2 + FexOy (7) FexCy + yCO2 yC* +yCO + FexOy (8) The above reactions sum up is considered as the one of the main proposed mechanisms of deactivation of iron based catalyst during FTS. It has been reported that iron carbides are considerably more resistant to oxidation by water than metallic iron5. It has also been reported that smaller catalyst particles carburize more rapidly and are not oxidized during reaction whether larger particles showed higher oxide : carbide ratio8. In contrast, Rustad et al.9 reported that oxidation of nano‐sized iron carbide thermodynamically more feasible than the oxidation of bulk iron carbide, due to higher surface energy of iron carbide compared to magnetite. Therefore, the oxidation of small crystallites is possible to occur at much lower water partial pressure. It is observed that the relationship between the catalyst activity and catalyst time on stream strongly depends on the composition of the catalyst, the activation conditions and the reactor configuration. For example, Bukur et al.10 reported a continuous decrease in catalyst activity when the catalyst was pretreated by carbon monoxide or syngas whereas better catalyst stability was observed when it was pretreated by hydrogen. Their hypothesis was that the higher surface concentration of hydrogen on these catalysts somehow stabilized catalyst activity. The hydrogen reduced catalyst showed higher selectivity toward methane, lower hydrocarbons, isomers and olefins as compared to carbon monoxide or syngas pretreated samples. Deactivation of the carbon monoxide or syngas pretreated catalyst was attributed to the conversion of the χ‐Fe5C2 to less active oxide phases and the build‐up of carbonaceous material. From literature it is clearly observed that the oxidation of iron carbide and the catalyst activity solely depends on the catalyst composition11. Davis11 studied time on stream phase composition of 100 Fe/3.6 Si/0.71 K catalyst under FT reaction conditions (270°C, 175 psig, H2/CO= 0.7 at a flow rate of 3.4 nl/h/g Fe) using Mössbauer. This catalyst was pretreated at the temperature, pressure and the flow rate of 270°C, 175 psig and 2.0 nl/h/g Fe, respectively. The holding time was 24 hours. This catalyst was quite stable for more than 450 hours when the conversion was maintained ~82% and the oxides on the catalyst surface was more than 80% after 350 hours of time on stream. The total carbide and super paramagnetic at this time on stream was 0 and 18%, respectively which clearly shows that for this catalyst there is no correlation between catalyst activity and the iron carbide. These authors claimed that the activity of the catalyst is coming from the super paramagnetic materials and minimum amount of iron carbide is required to maintain stable catalytic activity. However, Raje et al.12, Zhang et al.13 and Hayakawa et al.14 supported that deactivation of iron based catalyst occurs by the oxidation of iron carbide. 3.2. Deactivation due to pore plugging and the deposition of carbonaceous material on the catalyst Deactivation of iron based catalyst may also occur due to the deposition of carbon on the catalyst surface. This is also known as the inherent cause of iron based catalyst deactivation.6,15 It has been reported that under Fischer‐Tropsch reaction conditions, carbon deposition is thermodynamically favorable and the rate of carbon deposition increases with the increase in crystalline size.6 Two types of carbon species are mainly formed on the catalyst surface (1) amorphous; (2) graphite. Amorphous carbons are mainly formed during low temperature (<280°C) Fischer‐Tropsch reaction. On the other hand, the graphite (insoluble) carbon formation is favorable at the temperature above 280°C which leads to coke deposition.
During FT synthesis, dissociation of CO causes the formation of carbon atom which in turn forms surface carbon species due to the direct coupling according to Eq. (6) which is known as amorphous carbon. These carbon species reduce the catalyst activity by blocking the active catalyst sites. However, these carbon species can be removed by hydrogen treatment at elevated temperature (>350 °C). Eliason and Bartholomew16 reported that under mild steady state reaction conditions, Eq. (6) is not a serious competitor for carbon atoms. It starts to play an important role as the reaction temperatures increase. In the beginning of FT synthesis, the effect of significantly less active carbon species becomes evident immediately with the increase in the formation of hydrocarbon products and the catalyst activity decreases. They have reported to decrease in 40% activity within 2 hours at 280°C for unpromoted catalyst. It was proposed that at higher temperature the deactivation of catalysts occur due to the following reactions. C* + yC* C* y+1 (polymeric β‐carbon) [‐C*y+1‐] (graphite δ‐carbon) (9) Fe2.2C (ε’) Fe5C2 (χ) Fe3C (θ) (10) Carbon deposition may also occur due to the Boudouard reaction which causes the formation of insoluble carbon on the catalyst surface C* +2CO C‐C* + CO2 (11) C* + CO + H2 C‐C* + H2O (12) Jung and Thomson17 speculated that carbon atoms precipitated by the ε’ to χ transformation may serve as nucleation sites for the Boudouard reactions (Equations 11 and 12). However, Eliason and Bartholomew16 indicated that these carbon atoms formed in carbide transformations may serve as nucleation sites for polymerization to β‐carbon and/or crystallization to graphitic δ‐carbon. Hao et al.18 found that the carburization is enhanced with increasing reduction temperature and further carburization takes place during FT reaction. They have reported that the deposition of free carbon might partition the catalyst surface into some “island”, restricting the diffusion of monomers for chain growth. Therefore, the increase in deposited carbon might result in the decline of the selectivity of the heavier products. This group found that the catalyst activity decreased with the increase in the reduction temperature from 240 to 280°C. However, the stability of the catalyst markedly improved with the increase in reduction temperature. They interpreted that the decrease in activity could be due to the content of working catalyst loss in the heavier wax. Davis11 reported to increase in carbon deposition tenfold with the increase in temperature from 285 to 338°C indicates the strong temperature sensitivity for deactivation by carbon deposition. However, from their work, no correlation was observed between catalyst activity and carbon deposition. 3.3. Deactivation by poisoning Poisoning is referred as the strong chemisorption of reactants, products or impurities on the catalytic sites. Poisoning can be slow or fast depending on the poison’s concentration and it can be either reversible or irreversible depending on the strength of the poison. Poisoning is a complex process and occurs due to the some or all of the following reasons 15: (1) physical blockage of one or more catalytic sites by strongly adsorbed poison; (2) electronic modification of nearest neighbor atoms and sometimes even next‐nearest neighbor atoms; (3) restructuring of the adsorbent surface, and (4) hindering surface diffusion of adsorbed reactants, thereby preventing reaction. Sulfur poisoning of iron catalysts is considered as the one of the most serious problem in FT circles. A small concentration of sulfur can deactivate the catalyst very quickly. Typical sulfur levels currently allowed for Fe‐based FT are less than 5 ppb in syngas feed to the reactor.19 It has been indicated that the low level of sulfur poisoning through the pretreatment of reduced iron catalysts have been shown to enhance the activity of the catalyst, whereas high levels led to deactivation.6 Poisoning by H2S is slower on iron carbide than on metallic iron20 and the transformation of iron to iron sulfide which is difficult to reduce is thermodynamically feasible. During poisoning by sulfur, H2S adsorbs on the catalyst surface and deactivates the catalyst by accumulating it on the catalyst. Literature shows that sulfur not only deactivates the catalyst but also the selectivity shifts toward less olefinic material.19 Shultz et al.21 reported that alkali promoters, particularly potassium provide
resistance to sulfur deactivation. Manganese and boron were reported as sulfur sinks by several authors22. However, the effect of sulfur on the catalyst surface depends on the nature of sulfur species23. Several researchers indicated that the sulfur enhanced catalyst life and olefin selectivity21, 26, on both iron24 and iron carbide.25 For example, Bromfield and Coville26 in their work prepared a series of catalyst by adding small amounts Na2S to precipitated iron. Na2S concentrations were varied from 500‐20000 ppm. Catalyst with low sulfide concentrations were up to four time more active in FT reaction than a sulfur free catalyst, while catalyst with high sulfide loadings were apparently poisoned. Different compounds such as ammonia, HCN and carbonyl mainly act as a kinetic inhibitor during FTS. 27 Rausch indicted that co‐feeding of ammonia to FTS did not result in decrease in conversion for iron catalyst. It has been reported to decrease in CO conversion over cobalt catalyst in the presence of HCN28. No information is available on the impact of HCN on iron catalyst. Deactivation of cobalt catalyst by carbonyl material and chlorine was reported but no data is available on the impact of these compounds on iron catalysts. 3.4. Deactivation by sintering Sintering of catalyst occurs because of the loss of catalytic surface area due to crystallite growth of the catalytic phase, and loss of support area due to support collapse and of catalytic surface area due to pore collapse on crystallites of the active phase. Smit and Weckhuysen5 stated that all iron based FTS catalysts to some extent deactivate by sintering of the active phase. In contrast, other group explicitly concluded that sintering is not playing a significant role in the deactivation of FT catalyst.16 Many groups reported that sintering is one of the reasons for FT catalyst deactivation. Mansker et al.29 reported that the change in carbide phase composition from Fe7C3 to χ‐Fe5C2 was obtained by a progressive growth in crystallite size after ~350 hours on stream. Sintering is driven by the thermodynamic force to minimize the exposure of highly energetic surfaces, which precede either through atomic migration and/or crystallite migration and this process is generally accelerated by the presence of water vapor15. Sintering rates depend on the temperature, atmosphere, metal type, metal dispersion, promoters/impurities and support area, texture and porosity.15 Sintering mainly occurred when the temperature is above the Tamman temperature. When the temperature is above the Tamman temperature, atoms on the surface of crystallites are expected to become mobile and as a result crystallites can sinter29. The Tamman temperature for iron is 633°C which is much higher than the typical FT reaction temperature (200‐400°C). As the FT reaction is very exothermic, there is a possibility that the local temperature of crystallites could be very high. Van Steen and Claeys6 reported that sintering can be reduced by improving the distribution of catalytically active component. On the other hand, Rao et al.30 indicated that sintering can be prevented by stabilizing the active phase using the binders.
4. Conclusion The main challenge in the design of novel iron‐based FTS catalysts is to overcome their notoriously high deactivation rates. Even after more than 80 years of research on iron‐based FTS catalysts, no straight forward relationship between catalyst structure and performance is available. In‐situ time on stream characterization could be useful to find the possible reason of FT catalyst deactivation.
5. References 1. Baker, R. T. K. (1979), “In‐Situ Electron Microscopy Studies of Catalyst Particle Behavior”, Catal. Review, 19, pp. 161‐209. 2. Shashkin, D. P., Shiryaev, P. A., Chichagov, A. V., and Morozova, O. S. (2004), "Structural Transformation of Iron Oxides during the Formation of Fischer‐Tropsch Catalysts", Kinet. Catal., 33(2), pp. 744‐749. 3. Grzybek, T., Papp, H., and Baerns, M. (1987), "Fe/Mn Oxide Catalysts for Fischer‐Tropsch Synthesis ‐ Part V. XPS Surface Characterization of Calcined and Reduced Samples", Appl. Catal., 29(2), pp. 335‐350.
4. Mahajan, D., Gütlich, P., Ensling, J., Pandya, K., Stumm, U., and Vijayaraghavan, P. (2003), "Evaluation of Nanosized Iron in Slurry‐Phase Fischer‐Tropsch Synthesis", Energy & Fuels, 17(5), pp. 1210‐1221. 5. de Smit, E. and Weckhuysen, M. B. (2008), “The Resistance of Iron‐Based Fischer‐Tropsch Synthesis: on the Multifaceted Catalyst Deactivation Behavior”, Chem. Soc. Rev. 37, pp. 2758‐2781. 6. Van Steen, E. and Claeys, M., (2008), “Fischer‐Tropsch Catalysts for the Biomass to Liquid (BTL)‐Process, Chem. Engg, Tech. 31(5), pp. 655‐666. 7. O'Brien, R. J., Xu, L., Spicer, R. L., and Davis, B. H. (1996), "Activation Study of Precipitated Iron Fischer‐ Tropsch Catalysts," Energy & Fuels, 10(4), pp. 921‐926. 8. Dry, M. E. (1981), "The Fischer‐Tropsch Synthesis," in "Catalysis, Science and Technology", (J.R. Anderson, M. Boudart, eds.), Springer‐Verlag, 1, pp. 159‐256. 9. Rustad, R. J., Wasserman, E., and Flemy, R. A.(1999), “A Molecular Dynamics Investigation of Surface Reconstruction on Magnetite (001)”, Surface Science, 432 , pp. 583‐588. 10. Bukur, D. B., Koranne, M., Lang, X., Rao, K. R. P. M., and Huffman, G. P. (1995), "Pretreatment Effect Studies with a Precipitated Iron Fischer‐Tropsch Catalyst", Appl. Catal., 126, pp. 85‐113. 11. Davis, H. B., (1999), “Technology Development for Iron‐Fischer Tropsch Catalysis,”, DOE Contract DE‐AC22‐ 94PC94055, Final Report, March. 12. Raje, A., O'Brien, R. J., Xu, L., and Davis, B. H. (1997), "Deactivation of Iron‐Based Catalysts for Slurry Phase Fischer‐Tropsch Synthesis," in "Catalyst Deactivation", Stud. Surf. Sci. Catal., 111, pp. 527‐533. 13. Zhang, Y., Sirimanothan, N., O’Brien, J. R., Hamdeh, H. H., and Davis, H. B. (2001), “Study of Deactivation of Iron‐Based Fischer‐Tropsch Synthesis Catalysts”, Studies in Surface Science and Catalysis, 139, pp. 125‐132. 14. Hayakawa, H., Tanaka, H., and Fujimoto, K. (2006), “Studies on Precipitated Iron Catalysts for Fischer– Tropsch Synthesis”, App. Catal., 310, pp. 24‐30. 15. Bartholomew, C. H. (2001), "Mechanisms of Catalyst Deactivation", Appl. Catal. A, 212(1‐2), pp. 17‐60. 16. Eliason, S. A., and Bartholomew, C. H. (1999), "Reaction and Deactivation Kinetics for Fischer‐Tropsch Synthesis on Unpromoted and Potassium‐Promoted Iron Catalysts," Appl. Catal. A, 186, pp. 229‐243. 17. Jung, H., and Thomson, W. J., (1992), “Dynamic X‐ray Diffraction Studies of an Unsupported Iron Catalyst ij Fischer‐Tropsch Synthesis”, J. Catal., 134, pp. 654‐667. 18. Qing‐Lan Hao, Q.‐L., Liu, F.‐X., Wang, H., Chang, J., Zhang, C.‐H., Bai, L., Xiang, H. W., Li, Y.‐W., Yic, F., and Xu, B.‐F., (2007), Effect of Reduction Temperature on a Spray‐Dried Iron‐Based Catalyst for Slurry Fischer– Tropsch Synthesis”, J. Mol. Catal. 261, pp. 104‐111. 19. Kritzinger, J. A. (2002), "The Role of Sulfur in Commercial Iron‐Based Fischer‐Tropsch Catalysis with Focus on C2‐Product Selectivity and Yield", Catal. Today, 71(3‐4), pp. 307‐318. 20. Karn, F. S., Shultz, J. F., Kelly, R. E., Anderson, R. B. (1964), "Hydrogen Sulfide Poisoning of Nitrided and Carbided Iron Catalysts in the Fischer‐Tropsch Synthesis", Ind. Eng. Chem. Prod. Res. Dev., 3(1), pp. 33‐38. 21. Shultz, J. R., Hofer, L. J. E., Karn, F. S., and Anderson, R. B. (1962), "Studies of the Fischer‐Tropsch synthesis", J. Phys. Chem., 6(6), pp. 501‐506. 22. Bartholomew, C. H., and Bowman, R. M. (1985), "Sulfur Poisoning of Cobalt and Iron Fischer‐Tropsch Catalysts," Appl. Catal., 15, pp. 59‐67. 23. van Dijk, W. L., Niemantsverdriet, J. W., van dar Kraan,A. M., and van der Baan, H. S. (1982), “Effects of Manganese Oxide and Sulphate on the Olefin Selectivity of Iron Catalysts in the Fischer Tropsch Reaction, Appl. Catal. 2(4‐5), pp. 273‐288. 24. Chaffee, A. L., Campbell, I., and Valentine, N. (1989), "Sulfur Poisoning of Fischer‐Tropsch Synthesis Catalysts in a Fixed‐Bed Reactor", Appl. Catal., 47(2), pp. 253‐276. 25. Madon, R. J., and Taylor, W. F., (1979), “Effect of Sulfur on the Fischer‐Tropsch Synthesis: Alkali‐Promoted Precipitated Cobalt‐Based Catalysts”, chapter 8 in “Hydrocarbon Synthesis from Carbon Monoxide and Hydrogen” (E. L. Kugler, F. W. Steffgen, eds.), ACS Adv. Chem. Ser. 178, pp. 93‐112.
26. Bromfield, T. C., and Coville, N. J. (1999), "The Effect of Sulfide Ions on a Precipitated Iron Fischer‐Tropsch Catalyst", Appl. Catal., 186(1‐2), pp. 297‐307. 27. Rausch, A. K. (2008), “C‐N‐Kopplung and Heterogenen Kobalthaltigen Katalysatoren”, PhD thesis Carl von Ossietzky Universität Oldenburg. 28. Oliva, C., van den Berg, C., Niemantsverdriet. J. W., and Curulla‐Ferre, D. (2007) “A Density Functional Theory of HCN Hydrogenation to Methylamine on Co(111)”, J. Catal., 248, pp. 38‐45. 29. Mansker, L. D., Jin, Y., Bukur, D. B., and Datye, A. K. (1999), "Characterization of Slurry Phase Iron Catalysts for Fischer‐Tropsch Synthesis", Appl. Catal., 186(1‐2), pp. 277‐296. 30. Rao, V. U. S, Stiegel, G. J., Cinquegrane, G. J., and Srivastave, R. D., (1992), "Iron Based Catalysts for Slurry Phase Fischer‐Tropsch Process: Technology Review", Fuel Proc. Tech., 30, pp. 83‐107.
Product Distribution and Reaction Pathways during Fischer-Tropsch Synthesis on an Iron Catalyst Dragomir B. Bukur*a,b and Lech Nowickic a Chemical Engineering Program, Texas A&M University at Qatar, P.O. Box 23874, Doha, Qatar b Department of Chemical Engineering, Texas A&M University, College Station, TX 77843, U.S.A. c Faculty of Process and Environmental Engineering, Technical University of Lodz, 90-924 Lodz, Poland *Corresponding author: (974)4423-0134, [email protected]
Abstract: A precipitated iron Fischer-Tropsch (F-T) catalyst was tested in a stirred tank slurry reactor under different process conditions. The extent of water-gas-shift (WGS) reaction increased with increase in conversion of the limiting reactant, indicating that the WGS is a consecutive reaction with respect to water formed in the F-T synthesis reaction. The experimental results indicate that 1-olefins participate in secondary reactions (e.g. 1-olefin hydrogenation, isomerization and readsorption). Secondary hydrogenation and isomerizationof 1-olefins increased with increase in partial pressure of hydrogen. Gas residence time had significant effect on selectivity of ethylene and other gaseous 1-olefins. Chain growth probability factor increased with increase in molecular weight. Keywords: Fischer-Tropsch synthesis, iron catalyst, product distribution. 1. Introduction The present study is a part of broader effort to develop a comprehensive kinetic model for slurry phase F-T synthesis on iron based catalysts, which would be capable of predicting composition of reaction mixture for a given set of process conditions. To this end we conducted experiments in a stirred tank slurry reactor (STSR) over a wide range of process conditions and obtained comprehensive data base for future kinetic studies. Here we provide qualitative discussion of experimental results and draw conclusions about the nature of some of the reactions (primary vs. secondary reaction pathways).
2. Experimental Three tests (runs SB-21903, SB-26203 and SB-28603) were conducted in a 1 dm3 stirred tank slurry reactor (Autoclave Engineers). The feed gas flow rate was adjusted with a mass flow 1
controller and passed through a series of oxygen removal, alumina and activated charcoal traps to remove trace impurities. After leaving the reactor, the exit gas passed through a series of high and low (ambient) pressure traps to condense liquid products. High molecular weight hydrocarbons (wax), withdrawn from a slurry reactor through a porous cylindrical sintered metal filter, and liquid products, collected in the high and low pressure traps, were analyzed by capillary gas chromatography. Liquid products collected in the high and atmospheric pressure traps were first separated into an organic phase and an aqueous phase and then analyzed using different columns and temperature programmed methods. The reactants and noncondensible products leaving the ice traps were analyzed on an on-line GC (Carle AGC 400) with multiple columns using both flame ionization and thermal conductivity detectors. Further details on the experimental set up, operating procedures and product quantification can be found elsewhere [13]. A precipitated iron catalyst (Ruhrchemie LP 33/81 with nominal composition 100 Fe/5 Cu/4.2 K/25 SiO2 in parts by weight) calcined in air at 300°C and sieved between 140-325 mesh was loaded into the reactor filled with 300-320 g of Durasyn 164 oil (a hydrogenated 1-decene homopolymer, ~ C30 obtained from Albemarle Co.). The catalyst was pretreated in CO at 280°C, 0.8 MPa, and 3 NL/g-cat/h for 12 hours. After the pretreatment the catalyst was tested initially at 260°C, 1.5 MPa, 4 NL/g-Fe/h (where, NL/h, denotes volumetric gas flow rate at 0°C and 1 bar) using CO rich synthesis gas (H2/CO molar feed ratio of 0.67). After reaching a stable steady state value (~60 h on stream) the catalyst was tested at different process conditions (total of 25 sets of process conditions in three tests/runs).
3. Results and Discussion 3.1 Effects of Process Conditions and Conversion on Water-Gas-Shift Reaction and Hydrocarbon Product Distribution The catalyst was tested under 25 sets of different process conditions. The following values (or ranges) of process conditions were utilized in these three STSR tests: Reaction temperature (T):
220, 240 and 260°C
Reaction pressure (P):
8, 15 and 25 (22.5) bar
Feed composition (H2/CO ratio):
2/3 or 2/1
Gas space velocity (NL/g-Fe/h):
0.5 – 23.5 2
Effect of temperature and conversion of the limiting reactant (H2 for H2/CO = 2/3 feed gas, CO for H2/CO = 2/1 feed gas) is shown in Figures 1-3. As shown in Figure 1 the usage ratio (UR) decreases whereas the CO2 selectivity increases with increase in conversion (at constant temperature) or with increase in temperature (at constant conversion of the limiting reactant). This trend is the same regardless of feed composition (H2/CO = 2/3 in Fig. 1a, or H2/CO = 2/1 in Fig. 1b). The effect of conversion is consistent with a concept that the WGS reaction is a consecutive reaction. The increase in WGS activity (higher CO2 selectivity and lower UR) with increase in temperature (at constant conversion) is a kinetic effect (Fig. 1). Methane selectivity increases whereas the selectivity of high molecular weight hydrocarbons +
(C5 ) decreases with increase in temperature (Figure 2). Experimental data at 220°C and 240°C do not follow this trend, due to experimental errors. At a given temperature methane conversion increases with increase in conversion (H2/CO = 2/3 feed in Fig. 2a) whereas C5+ selectivity decreases. However, this trend was not observed with H2/CO = 2 feed gas at 220°C and 240°C (not shown). As shown in Figures 3, 1-olefin content decreases, whereas 2-olefin content and n-paraffin content increase with increase in conversion of the limiting reactant. This indicates that 1-olefins are consumed in secondary reactions, whereas n-paraffins and 2-olefins are formed in part in secondary reactions. Carbon number dependences of selectivities (at constant conversion) show the following trend: 1-olefin selectivity passes through a maximum and n-paraffin selectivity passes through a minimum at C3, whereas 2-olefin selectivity increases with carbon number. It is well known that ethylene is more reactive than other 1-olefins and it can initiate chain growth or be incorporated into the growing chains, and thus its selectivity is lower than that of 1-propene and other low molecular weight (MW) 1-olefins [4-7]. Described carbon number effects have been ascribed [814] to secondary reactions of 1-olefins (1-olefin readsorption, hydrogenation and/or isomerization) and increase in residence time with increase in molecular weight. 3.2 Carbon Number Product Distribution Typical carbon number product distributions at different process conditions are shown in form of Anderson-Schulz-Flory (ASF) plots (Figure 4). Carbon number distributions could not be described by uniform chain growth probability factor, , which would result in a straight line
3
(ln xn vs Cn). Experimental data were fitted using a three-parameter model of Huff and Satterfield [15]: xn = (1-1) 1n-1 + (1 - ) (1 - 2) 2n-1
where: xn = mole fraction of products containing n carbon atoms (hydrocarbons and oxygenates), = fraction of type 1 sites, 1 = chain growth probability on type 1 sites, and 2 = chain growth probability on type 2 sites. Experimental data are reasonably well represented by this type of model. The model parameters were estimated using a nonlinear regression.
4. Conclusions Increase in the extent of WGS reaction, manifested in increase of CO2 selectivity and decrease in UR, with increase in conversion of the limiting reactant, is consistent with the concept that the WGS is a consecutive reaction with respect to water that is formed in FTS reaction. Increase in the extent of WGS reaction with increase in temperature, decrease in total pressure, or decrease in H2/CO feed ratio (i.e. decrease in CO partial pressure) is due to kinetic effects. Decrease in 1-olefin content and increase in 2-olefin and n-paraffin contents with increase in conversion are consistent with the concept that 1-olefins participate in secondary reactions, whereas 2-olefins and n-paraffins are formed in these reactions (e.g. 1-olefin hydrogenation, isomerization and readsorption). Secondary hydrogenation and isomerization reactions increase with increase in partial pressure of hydrogen. Gas residence time has pronounced effect on selectivity of ethylene and gaseous 1-olefins, but it is less pronounced for higher MW olefins (C10+). The residence time of high MW hydrocarbons is much longer than that of gaseous hydrocarbons and is determined by the rate of liquid (wax) removal from the reactor. Methane selectivity decreases (and that of C5+ hydrocarbons increases) with decrease in temperature, increase in reaction pressure, and/or decrease in H2/CO feed ratio. The effect of conversion (i.e. gas space velocity) on hydrocarbon selectivity was relatively small in most cases, and there were no clearly discernible trends. Carbon number product distribution could be
4
described by so-called “double alpha” model (two distinct chain growth probabilities, one for lower MW products and the second one for higher MW products).
Acknowledgements This work was supported by US DOE (University Coal Research Program) grant No. DE-FG2602NT41540.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]
Bukur, D. B., Patel, S. A. and X. Lang, Appl. Catal., 61, 329 (1990). Zimmerman, W. H. and Bukur, D. B., Can. J. Chem. Eng., 68, 292 (1990). Bukur, D. B., Nowicki, L. and Lang, X., Chem. Eng. Sci., 49, 4615 (1994). Novak, S., Madon, R. J. and Suhl, H., J. Chem. Phys., 74, 6083 (1981). Novak, S., Madon, R. J. and Suhl, H., J. Catal., 77, 141 (1982) Iglesia, E.; Reyes, S. C.; Madon, R. J.and Soled, S. L., J. Catal., 129, 238 (1991). Komaya, T. and Bell, A. T., J. Catal., 146, 237 (1994). Kuipers, E. W., Vinkenburg, I. H. and Osterbeck, H., J. Catal., 152, 137 (1995). Schulz, H.; Rosch, S.and Gokcebay, H. Selectivity of the Fischer-Tropsch COhydrogenation, in Coal: Phoenix of '80s, Proc. 64th CIC. Coal Symp. Vol. 2, Ottawa, 1982; pp. 486-493. Schulz, H., Beck, K. and Erich, E., Fuel Proc. Technol., 18, 293 (1988). Dictor, R. A. and Bell, A. T., J. Catal., 97, 121 (1986). Donnelly, T. J. and Satterfield, C. N., Appl. Catal., 52, 93 (1989). Iglesia, E.; Reyes, S. C.and Madon, R. J., Adv. Catal., 39, 221 (1993). Madon, R. J. and Iglesia, E. J. Catal., 139, 576.(1993). Huff, G. A. and Satterfield, C. N., J. Catal., 85, 370 (1984).
5
1.8
50 a 45
1.6
35
Usage ratio
1.2
30 1.0 25 0.8 20 0.6
15
UR @ 260 °C UR @ 240 °C
0.4
10
UR @ 220 °C CO2 @ 260 °C CO2 @ 240 °C
0.2
P = 15 bar H2/CO = 2/3
CO2 @ 220 °C 0.0 0
20
CO2 Selectivity (%)
40
1.4
40
60
80
5 0 100
Conversion of H 2 (% )
2.0
40 b
1.8
35
1.6
Usage ratio
25
1.2 1.0
20
0.8
15 UR @ 260 °C
0.6
UR @ 240 °C 0.4
CO2 Selectivity (%)
30 1.4
10
UR @ 220 °C CO2 @ 260 °C
0.2
CO2 @ 240 °C
P = 15 bar H2/CO = 2/1
CO2 @ 220 °C
0.0 0
20
5
40
60
80
0 100
Conversion of CO (%)
Fig. 1 Effect of temperature and conversion on WGS reaction (15 bar) a) feed H2/CO ratio 2/3, b)feed H2/CO ratio 2/1. 6
5.0 a
CH4 Selectivity (C atom %)
4.5
4.0
3.5
3.0
2.5
2.0 260 °C
1.5
240 °C
P =15 bar H2/CO = 2/3
220 °C 1.0 0
20
40
60
80
100
Conversion of H 2 (% ) 90 b
C5+ Selectivity (C atom %)
85
80
75
70 260 °C P = 15 bar H2/CO = 2/3
240 °C 220 °C
65 0
20
40
60
80
100
Conversion of H 2 (% )
Figure 2. Effect of temperature on hydrocarbons selectivity (15 bar, H2/CO: 2/3) a) Methane selectivity, b) Higher hydrocarbon selectivity.
7
1
a
Mole fraction
0.1
T = 240 °C P = 15 bar H2/CO = 2/3 SV = 2.0 NL/g-Fe/h
0.01
0.001
1= 0.71 2= 0.92 = 0.82
0.0001
0.00001 0
10
20
30
40
50
Carbon number 1
b
Mole fraction
0.1
T = 240 °C P = 15 bar H2/CO = 2 SV = 5.8 NL/g-Fe/h
0.01
0.001
1= 0.62 2= 0.87 = 0.67
0.0001
0.00001 0
10
20
30
40
50
Carbon number
Figure 4. Carbon number product distribution according to extended Anderson-Schulz-Flory model (T = 240C).
8
2010 International Pittsburgh Coal Conference Program Topic: Combustion Smartsheet Tool applied to Boiler Performance Analysis and Economic Optimization of a Circulating Fluidized Bed Boiler Abhinaya Joshi, R&D Engineer Alstom Power Inc. 200 Great Pond Drive, P.O. Box 500 Windsor, CT, USA Email: [email protected] Telephone: 860-285-2753
Xinsheng Lou, Technology Leader Alstom Power Inc. 200 Great Pond Drive, P.O. Box 500 Windsor, CT, USA Email: [email protected] Telephone: 860-285-4982
Carl Neuschaefer, Director Alstom Power Inc. 200 Great Pond Drive, P.O. Box 500 Windsor, CT, USA Email: [email protected] Telephone: 860-285-9290
Paul Panos, Process Engineer Alstom Power Inc. 200 Great Pond Drive, P.O. Box 500 Windsor, CT, USA Email: [email protected] Telephone: 860-285-3844 Weikko Wirta Boiler and Engineering Superintendent AES Thames (Currently, Plant Manager - AES Huntington Beach) Email: [email protected]
Abstract In the current deregulated competitive electricity market and with tighter environmental regulations, it is an important goal that power plant boilers are operated and controlled in the most efficient, economic and cleanest manner while fulfilling the grid load demand requirements. This paper presents the development and testing results of a boiler performance analysis and optimization tool referred to as “Smartsheet”, capable of assisting plant owners and operators in meeting the stated optimization goals. The Excel based tool includes a neural network process model, economic relationships, an optimization solver and a number of user functions and interface for use in the analysis and the operational optimization of a circulating fluidized bed (CFB) boiler. In view of the fact that a CFB boiler process is nonlinear with strong interactions among process variables, an artificial neural network (ANN) based model was developed to capture the nonlinear relationships between the variables representing operating
1
conditions and the variables that relate to operating costs. The tool has been tested at the AES Thames, station in Uncasville Connecticut. The validation testing involved developing two orthogonal test matrices based on Design of Experiments (DoE) methods that were then used to carry out two independent tests in the Alstom supplied CFB boiler (100MW). The test data collected from the first set of tests was used to train the ANN model and the data from a second set of tests was used for the model validation. The model was combined with the other software developed functions into a tool to support power plant engineers and operators in boiler and total plant performance analysis. The tool is specifically designed to assist in generating economically optimal operating plant settings based on a utility’s specified current cost and emission credit factors. The optimization results to date show that the optimized operating settings can save, on average, more than 2% of operating costs over the current operating conditions, which have been fine tuned by almost 20 years of operating experience. Additional validation tests of the optimal operating conditions suggested by the optimization tool have been planned with the customer to further validate the tool. Keywords: CFB boiler, Economic Optimization, Neural network, Genetic algorithm.
1. Introduction In the current deregulated competitive electricity market and with tighter environmental regulations, it is an important goal that power plant boilers are operated and controlled in the most efficient, economic and cleanest manner while fulfilling the grid load demand requirements. Boiler operators and engineers can use the wealth of knowledge they have about boiler design and operating parameters and their effect on key economic variables such as fuel consumption, emissions, etc. to improve the operating profits. Nevertheless, determining the “best” combination of operational settings for multiple variables out of all the possible condition combinations that will achieve maximum operating profit is a very challenging human problem. To determine all the possible operating condition, honor plant constraints and then find the set of many variables that establishes the operating conditions that minimize operating costs requires the aid of a computational tool. In addition, ever changing with time electricity prices, fuel costs and emission credits make the boiler optimization problem much more difficult where a software tool will aid the operator in selecting optimal parameters. This paper presents the development and testing results of a boiler performance analysis and optimization tool, “Smartsheet”, capable of assisting plant owners and operators in meeting the stated optimization goals. The Excel based tool includes a nonlinear process model based on neural network, economic relationships, a global optimization solver and a number of user functions and interface for the analysis and the operational optimization of a circulating fluidized bed (CFB) boiler. The tool can be either used in an offline engineering mode or in online supervisory control mode. 2. Circulating Fluidized Bed Boiler Plant Circulating fluidized bed (CFB) (Semedard, et al., 2005) combustion technology is relatively new, yet mature solidfuel firing technology developed in last forty years. CFB boilers can be designed to fire essentially any combustible material, and also can be designed to fire a wide range of fuels in a given boiler (Abdulally and Darling, 2007). While there are many differences between CFB and other available boiler technologies, one of the most important is furnace temperature, with CFBs operating at much lower furnace temperatures than pulverized coal boilers. The relatively low furnace temperature is the source of most key benefits of fluidized bed firing; NOx levels are low because no atmospheric nitrogen is converted to NOx, SO2 levels are low because SO2 can be efficiently absorbed via limestone injection to the furnace, and fuel flexibility results because the ash stays in solid form, so fuel ash characteristics do not impact the design. CFB boilers also have a high solids residence time in the furnace, which compensates for the relatively low furnace temperature, allowing high combustion efficiency and limestone utilization. The circulating fluidized bed (CFB) boiler under consideration is a coal-fired 100 MW Alstom designed drum-type subcritical boiler (see Bozzuto, 2009; Marchetti et al., 2003) similar to the diagram shown in Figure 1. It also supplies steam (in the range of 100 K lb/hour) to a processing plant.
2
Figure 1. Elevation View of an Alstom CFB Boiler 3. Performance Analyses and Economic Optimization Tool The performance analyses and economic optimization tool, Smartsheet, developed and applied here for a CFB boiler hinges on two components: an artificial neural network based steady state nonlinear boiler model and a genetic algorithm based global optimization solver. Using these two components and economic relationships, different boiler performance analyses and economic optimization functions can be performed. These Smartsheet functions are discussed in the sub-sections below. A glimpse of the Excel based Smartsheet tool and its ranges of functionality is shown in Figure 2. The tool can be either used in an offline engineering mode or in online supervisory control mode.
3
Figure 2. A Glimpse of the Excel Tool 3.1. Artificial Neural Network Model The artificial neural network (ANN) based nonlinear process model of the CFB boiler is one of the key components of the Excel Smartsheet tool. The Boiler process has been identified as a time-varying nonlinear process over the different loads and operating conditions. Therefore, the artificial neural network, which has been widely and successfully utilized to model nonlinear relationships between variables in boiler process (Joshi, et al., 2008), was adopted to develop a CFB process model. Neural network model coefficients are estimated (referred to as neural network training) by using test data and numerous studies available in literature have found that the ANN has abilities to capture input-output relationship in the data accurately. The development of a neural network boiler model for boiler optimization typically involves three steps: 1) Defining input-output model structure, 2) Test matrix design & boiler test execution, and 3) Neural network model training & validation using test data. Defining the input-output structure for the ANN model is an important step and requires deep process and boiler equipment understanding. Exclusion of a key boiler process variable in the model may affect the operating cost calculations and inclusion of an unnecessary variable in the model prolongs the test and model building time. In addition, the input-output list was also verified and trimmed based on thorough analysis of historical operation data. A neural network input-output model structure for the CFB boiler is depicted in Figure 3. As seen in the figure, there are two categories of inputs to the model: 1) Manipulated and 2) Disturbance. Manipulated variables (MV’s) are those variables that can be changed by the operator when necessary. Disturbance variables (DV’s) are those that affect the defined model outputs but cannot be changed at will, such as ambient air temperature, external steam flow, etc.
4
PA Flow O2_Level
NN Model
SA Head Pressure MVs
Furnace DP
Inputs
Coal Feed
Low Bed Temp NOx
Coal Distribution
Outputs
LS distribution LS Feed
FARI On/Off Ambient Temp DVs
MW Demand Ext. Steam Flow Figure 3. A neural network input-output model structure for a CFB boiler
In the next step, two sets of test matrices with statistically independent test cases in each test matrix were designed based on a design of experiments (DoE), by including all the manipulated variables that are identified as inputs to the neural network model. Each test matrix included a series of tests with a limited number of values for each manipulated variable that span its full range. Two test datasets were collected from the tests conducted on the CFB utility boiler and then processed. The first dataset was utilized for the model development and the second for checking the developed model for its accuracy. The neural network structure used for the model development was so called multi-layer perceptron (MLP) with one hidden layer, i.e. there are three layers – an input layer, a hidden layer and an output layer as seen in Figure 3. The development of the neural network model includes iterative estimation of model coefficients (referred to as training) with a checking mechanism (called cross-validation) and model testing by using the dataset. The crossvalidation mechanism helps avoid the neural network model from over-fitting the dataset during its training. Model over-fitting may give very good predictions for the particular dataset but will perform very poorly for other datasets not used in training. The dataset is normally divided into three data subsets: training, cross-validation and testing data for this purpose. The prediction data from the resulting neural network model and the actual measured plant data are plotted in Figures 4-6. As seen in the figures, the prediction from the neural network model gives a good match with the actual plant data. The actual values of the outputs have not been shown, as they are confidential information.
Figure 4. Predicted and actual plant data for NOx
5
Figure 5. Predicted and actual plant data for Total Fuel
Figure 6. Predicted and actual plant data for Total Limestone 3.2. Genetic Algorithm for Global Optimization The genetic algorithm (GA) based global optimization solver is the next important component of the Smartsheet TM tool. As mentioned above the neural network boiler model is nonlinear, and therefore, the resulting optimization problem to be defined in Section 3.4 becomes nonlinear and often tends to be non-convex, i.e. it can have multiple locally optimal solutions across the range of input values. Since conventional gradient-based optimization solvers are designed to converge to a local optimal point, they often fail to find the global optimum solution among the various local optimal points. Therefore, a genetic algorithm based optimization solver was developed and incorporated in the tool. A genetic algorithm is a class of stochastic search techniques influenced by evolutionary biology, such as inheritance, mutation, selection, and crossover. The steps by which an optimization problem is solved using a genetic algorithm (refer to Gen and Cheng, 2000 for more details) are as listed below. The steps are also presented as a flow chart in Figure 7. Step 0: Initialize. Choose GA parameters, e.g., size of the population, mutation rate (a parameter that represents the probability of occurrence of mutation, which is explained in Step 2), maximum number of iterations, minimum difference in the value of the objective function (as defined in equation 1 in Section 3.4) of the best solution in successive iterations, etc. Minimum and maximum bounds of the inputs also needs to be defined. Go to Step 1. Step 1: Create initial population of chromosomes. A ‘chromosome’ is an array of input values (as seen in Figure 3) included for optimization. Individual elements in a chromosome representing individual inputs are called ‘genes’. In this step, a population of chromosomes is generated by randomly picking a value for each gene (input) within its
6
lower and upper bound. For each chromosome, its associated objective function value is calculated and the chromosomes are arranged in the order with the best solution first. Go to Step 2. Step 2: Crossover, Mutation and Elitism: This step produces a next generation population (group) of chromosomes with an expectation of improvement in the best solution from the last generation by the mechanisms of crossover, mutation and elitism. During a ‘crossover’, two chromosomes (parents) from the last generation population are randomly selected and a new chromosome is created by randomly selecting a gene from either of the parents. During a mutation, a chromosome is randomly selected from the last generation population and undergoes a mutation condition test where a random number is generated to see if it is below the mutation rate (chosen in Step 0). If the mutation test is passed, a new chromosome is created from the selected one by randomly selecting a gene and altering it with a random number within the bounds of the particular input. The selection of chromosomes from the last generation for crossover and mutation is done in such a way that there is a higher probability for a chromosome with better objective function value to be selected. ‘Elitism’ means to transfer the chromosome with the best objective function value from the last generation population to the new generation population. Since genetic algorithm is a stochastic search method, the solution it gives may or may not be a stationary point with slightly better solution in its immediate neighborhood. To address this issue, the elitism mechanism has been extended such that not only the chromosome with the best objective function value but also few more chromosomes that are slight variants of the current best are introduced to the next generation population. Go to Step 3. Step 3: Evaluation. The objective function value for each chromosome in the current generation population is calculated and the chromosomes are arranged in the order with the best solution first. If the improvement in the objective function value of the best chromosome in the current generation population and compared to the best chromosome in the last generation population is larger than a minimum difference value (chosen in Step 0), and if the maximum number of generations (iterations) has not been surpassed, then go to Step 2. Otherwise go to Step 4. Step 4: Termination. The chromosome with the best objective function value is the optimal solution to the optimization problem. Initialize
Create Initial population of solutions
Crossover, Mutation, and Elitism
Evaluation
No Solution satisfies the termination condition?
Yes End
Figure 7. A Genetic Algorithm flow chart
7
3.3. Boiler Sensitivity Analysis The boiler sensitivity analysis function allows engineers and operators to understand the relationships between all the input and output variables. Although power plant engineers with their wealth of experiences have a good understanding of some of these relationships, there are relationships among some variables that are hard to discern. Two types of sensitivity analyses methods were incorporated in the tool: 1) Basic - one input one output sensitivity curves and 2) Advanced - sensitivity curves with an additional input variable in the sensitivity plots. The basic sensitivity curve gives one-to-one functional relationship between an input variable and an output variable while other inputs are fixed to a nominal value. The basic sensitivity curves for the three outputs with respect to temperature are shown in Figure 8. Basic sensitivity curves give an idea of where outputs will be at different input settings.
Figure 8. Basic sensitivity curves for three outputs with respect to an input The advanced sensitivity curve gives functional relationships between an input and an output while a third input is at different discrete values, and other inputs are fixed to a nominal value. An advanced sensitivity curve for an output with respect to an input for five discrete values of another input is shown in Figure 9.
Figure 9. Advanced sensitivity curve for NOx with respect to PA Flow at five discrete values of O2 Level 3.4. Boiler Economic Optimization The foremost objective of developing the Smartsheet tool was to carry out economic optimization of the boiler/plant and compute the most profitable operating settings for the given power demand and input/output prices. Power plant engineers and managers with a their know-how and experience can change the boiler/plant input settings towards more profitable conditions in terms of, say, two or three variables. However, it is very difficult to find an overall true mathematically based economic optimal operating settings that involves many input and output variables. Furthermore, with constant changes in prices of outputs (electricity) and inputs (fuel, sorbent, etc.),
8
finding the most profitable condition at frequent intervals is almost impossible without an appropriately designed optimization tool. The economic optimization problem (with an objective function and constraints) for this work is formulated as
min J , with J = C NOx y NOx + C Fuel y Fuel + C LS y LS + C Ash y Ash ,
(1)
u
subject to
y = f NN (u , θ ) , umin ≤ u ≤ umax , where y = [yNOx, yFuel, yLS, yAsh] is a vector of model outputs; u is a vector of model inputs, e.g., O2 level, bed temperature, etc.; CNOx, CFuel, CLS, and CAsh are unit costs of the model outputs y; fNN is the neural network model; umin and umax are minimum and maximum bounds of inputs, respectively; and θ is a set of model coefficients. Although yAsh (total ash output) is not predicted by using the neural network model, it can be calculated based on total fuel and limestone values, their composition, and estimated calcination and sulfur capture ratio. The unit cost CNOx of NOx output can be positive or negative (with different values) depending on NOx production is above or below the allowed limits, respectively. The cost function used in the optimization problem above may not be the actual total cost of producing a unit of power, nevertheless, it covers major costs of operation in the form of fuel, limestone, and NOx costs/credits. The Smartsheet tool developed can provide basically three types of optimization solutions: 1) Enumeration solution, 2) Globally optimal solution and 3) Guided optimal solutions. For each solution type, a subset of inputs can be selected for optimization. The first solution type is based on enumeration, which gives a solution based on an exhaustive search of the input space with a number of discrete input points. This is a crude and simple method and may work with few variables and when a reasonably good solution is sufficient, but the problem size and the computation time increases very quickly with increase in variables and required resolution. The second solution type is globally optimal and computed based on genetic algorithm as discussed in Section 3.2. To provide more freedom for the engineers and operators, the global optimal solution is displayed with a number of near-optimal solutions from the last population for their reference in operational decision making. The operators may have preferences for one set of input settings that provides almost as much cost savings as the globally optimal set of input settings. The third solution type is similar to the second type but guided to a small number of partitioned regions. Since many experienced operators feel comfortable running a boiler with certain variables in certain preferred regions (e.g. operation of boiler at lower O2 in order to stay away from ID fan motor limits during the summer), this solution type provides solutions from searches at different regions in the input space. 4. Optimization Results The Smartsheet tool presented in Section 3 has been extensively tested to generate economically optimal input setting for different optimization cases by using current prices of fuel and limestone and NOx credits. The tool was also used to run optimization cases to evaluate potential improvement that could have been made in past operations. An improvement test case over duration of a month is shown in Figure 10, which compares the incurred total cost based on current settings and the optimized total cost. The cost numbers are confidential and hence are not revealed. As seen in the figure, the operating cost could have been reduced for all but one sample. Extrapolating the improvement results, the Smartsheet tool suggests savings of, on average, more than 2% of the current operating costs.
9
Figure 10. Total cost: Actual and Improved The results within the tool have been very encouraging to both Alstom and the AES power plant team members. To validate the optimization results in the field, about a month long plant tests have been planned for the near future. 5. Summary and Conclusions Smartsheet is an Excel based tool for the performance analyses and economic optimization of operating PC or CFB boilers. The initial application testing completed for a CFB plant was presented. Enabling the powerful optimization functionality of the tool are its computational engine components including: neural network based process model and a global optimization solver. The neural network model was trained and validated using actual boiler plant test data, which was derived and collected by implementing a systematic DoE orthogonal based set of test matrices. The resulting model is used for Boiler process behavior prediction and sensitivity analyses. The Smartsheet tool also includes economic relationships tailored to the Utility operations, which are used together with the model and the solver to provide different analyses of the boiler performance and to carry out economic optimization cases to derive the best operating points. The results from the tests so far show that this new tool is capable of supporting engineers in carrying out boiler performance and optimization analyses. The results also suggest that there are tangible savings that can be made by adjusting some of the boiler parameters even on units that have been working over many years to optimize operations. Additional Plant tests have been planned to further validate the optimization results.
References [1]
I. Abdulally and S. Darling, “CFB Technology: Economic, Clean and Efficient Steam and Power Generation for the Oil Sands Industry,” Oil sands and Heavy oil technologies conference, Calgary, Canada, July 18-20, 2007.
[2]
C. Bozzuto (Editor), “Clean Combustion Technologies: A Reference Book on Steam Generation and Emission Control,” Alstom Power, Inc., Windsor, Connecticut, Fifth Edition, 2009.
[3]
M. Gen and R. Cheng, “Genetic Algorithms and Engineering Optimizations,” John Wiley & Sons, New York, NY, 2000.
[4]
A. Joshi, X. Lou, C. Neuschaefer, and P. B. Luh, “Neural Network Modeling and Nonlinear Model Predictive Control for a Gas Fired Boiler Simulator,” Presented at the 18th Annual Joint ISA POWID/EPRI Controls and Instrumentation Conference, Scottsdale, AZ, USA, 8-13 June 2008.
[5]
M. M. Marchetti, T. S. Czarnecki, J. C. Semedard, S. Devroe and J. M. Lemasle, “Alstom’s Large CFBs and Results,” Proceedings of 17th International Conference on Fluidized Bed Combustion, ASME, Jacksonville, FL, USA, May 2003.
10
[6]
J. C. Semedard, R. Skowyra, and J. Seeber, “Circulating fluidized bed technology: clean fuel-flexible generation for China's energy future,” China Power, August 10-11, 2005, Beijing, China.
11
A MODEL OF PRIMARY FRAGMENTATION OF COAL PARTICLES IN FLUIDIZED-BED COMBUSTION Monika Kosowska-Golachowska1, Adam Luckos2 1 Czestochowa University of Technology Armii Krajowej 19c, 42-201 Czestochowa, Poland T: +48 34 3250579; F: +48 34 3250604; E: [email protected] 2 SASOL Technology R&D 1 Klasie Havenga Road, PO Box 1,Sasolburg 1947, South Africa T: +27 16 9605573; F: +27 11 2194122; E: [email protected] ABSTRACT Fragmentation of coal particles plays a significant role in fluidized-bed boilers because it accelerates devolatilization and influences the size distribution of char particles in the bed. A model is proposed to describe heating, devolatilization and fragmentation processes for large coal particles burnt in a fluidized-bed combustor. Results of tests carried out in a bench-scale bubbling-bed unit at conditions similar to those prevailing in fluidized-bed boilers are compared with predictions of the model. INTRODUCTION It is important to understand changes in particle size distribution of fed coal during fluidized-bed combustion because they affect the devolatilization and combustion rates, distribution of volatiles inside the furnace and thermal efficiency of the boiler. Fragmentation of coal is the result of many physical and chemical processes taking place inside burning particles. It reduces the size of particles, accelerates the release of volatiles and affects the chemistry of pollutant formation and reduction. At least three mechanisms of fragmentation can be considered. The first one, called thermal fragmentation, occurs immediately after a particle is introduced into the hot furnace and is associated with thermal shock caused by large temperature gradients developed inside the particle. If thermal stresses generated by these temperature gradients exceed the mechanical strength of coal the particle cracks into several fragments. The second mechanism, called primary fragmentation, is associated with an increase in the pressure inside the network of pores during the release of volatiles. The third mechanism, called secondary fragmentation, occurs during char combustion when bridges connecting different irregularly-shaped components of the char structure burn out. In the case of bituminous coals, the primary fragmentation is the main cause of breaking down of coal particles during fluidized-bed combustion. The secondary fragmentation of char particles is not very extensive and, therefore, its contribution to the total fragmentation is relatively very small (1). Several models have been developed to predict the fragmentation behaviour of large coal particles, e.g. Chirone and Massimilla (2); Sasongko and Stubington (3); Dakic et al. (4). In this study a model developed by Kosowska (5) is proposed to
predict the behaviour of large coal particles during devolatilization and primary fragmentation in a bubbling fluidized-bed combustor. DESCRIPTION OF THE MATHEMATICAL MODEL Temperature profiles, volatile matter release profiles and pressure profiles inside the coal particle are calculated as a function of time and radial position according to the block diagram shown in Figure 1. The main assumptions of the model are (5): • Coal particles are spherical, • Coal behaves as an elastic-brittle material, fracture sections being parallel to geological bedding planes, • During devolatilization coal particles experience both thermal stresses and stresses caused by pressure increase inside pores when volatiles are released. The effect of thermal stresses is neglected.
Figure 1. Block diagram of the model Heating of coal particle Temperature profiles, T(r, t), for a spherical coal particle is calculated from the transient energy balance equation given by:
ρC p
∂T (r , t ) 1 ∂ 2 ∂T (r , t ) = 2 r k ∂t ∂r r ∂r
(1)
The initial and boundary conditions for Equation (1) are: at
t = 0;
at
r = R, t > 0;
at
r = 0, t ≥ 0;
T ( r ,t ) = T0 ∂T (r , t ) k = h[Tb − T (R, t )] ∂r ∂T ( r , t ) =0 ∂r
(2) (3) (4)
The correlations proposed by Stubington and Sumaryono (6) have been used to estimate the values of the heat conduction coefficient, k, and specific heat capacity, ρCp, for coal and char particles. The heat transfer coefficient, h, between the fluidized bed and a coal particle was calculated from the empirical correlation suggested by Dakic et al. (7). Release of volatiles Following Anthony et al. (8), it is assumed that devolatilization is the result of a large number of simultaneous, independent and irreversible first-order reactions with a distribution of activation energy. The total amount of volatiles released in any spherical shell is estimated by adding the contributions of all individual species with assumption that the distribution of activation energy can be described by the continuous Gaussian distribution function:
∞ t E dt f (E )dE V (r , t ) = V 1 − ∫ exp − ∫ k 0 exp − ( MR)T (r , t ) 0 0 *
f (E ) =
where
1
σE
(5)
(E − E 0 ) exp − 2σ E2 2π
Using equation (5), the rate of devolatilization and the molar flux of volatiles are calculated as:
Γ(r , t ) =
∂V (r , t ) ∂t
N V (r , t ) =
ρ 1 MV r 2
(6) r
∫ r Γ(r , t )dr 2
(7)
0
Intraparticle transport of volatiles It is assumed that volatiles (tar + gas) are transported from the interior of a spherical particle towards its external surface by viscous flow through macropores. It is also assumed that the viscosity of volatiles is constant and their accumulation in any spherical shell is negligible. The molar flux of volatiles, Nv, is ((2), (3), (9), (10)):
r pc2 φ pc N V (r , t ) = − 8(MR )T (r , t )τ pc µV
∂P(r , t ) P(r , t ) ∂r
(8)
Equation (8) gives the pressure distribution inside the particle, P(r, t), as a function of radial position and time after integration with the initial condition for t = 0 at the particle surface
P(R, t ) = P0
(9)
Primary fragmentation A mechanical model similar to that proposed by Chirone and Massimilla (2) is proposed to describe the primary fragmentation of spherical coal particles. The starting point for predicting the occurrence of fragmentation is a simple onedimensional stress field, consisting of internal forces due to volatile pressure and resistance forces of the material. At a segment base, Sθ, parallel to bedding planes (Figure 2), solid elements connecting the spherical segment to the remaining part of the sphere, at given time during devolatilization, balance the resultant of applied load, FA(θ,t), with a resistant force, FR(θ,t).
Figure 2. Coal spheres divided into spherical segments with base parallel to bedding planes (2, 5) Both forces are orthogonal to Sθ. This latter hypothesis implies that fragmentation is determined by a simple one-dimensional stress field. Accordingly, the sphere breaks when the resultant of applied force, FA(θ,t), overcomes the maximum value of the resistance force, FR(θ,t), i.e. when
FA (θ , t) > FR (θ , t)
(10)
FA(θ,t) is the result of volatiles pressure. Due to the direction of volatile flow it is a tensile force expressed as
F A (θ , t ) =
∫ P ( r , t )φ θ
θ
θ
( rθ , t ) dS
(11)
Sθ
where
φ (r , t ) = φch* +
1 (V (r , t ) − V * )(φch* − φc ) * V
(12)
The resultant FR(θ, t) of resistance forces acting on the surface Sθ is
FR (θ , t ) = ∫ σ [1 − φθ (rθ , t )]dS Sθ
(13)
This force depends on mechanical strength of coal and porosity as a function of radial position and time. RESULTS OF NUMERICAL SIMULATIONS Equations (1), (5)–(8), (11) and (12) are integrated numerically using parameters listed in Tables 1 and 2. Detailed solutions of these equations and results of numerical simulations can be found in elsewhere (5). Table 1. Values of parameters used in the model Ambient temperature Bed temperature Atmosphere Diameter of coal particles
T0 = 20°C Tb = 700–900°C air 0.002–0.01 m
Mean activation energy Standard deviation of activation energy Pre-exponential factor Average molar mass of volatiles
E0 = 228.32×106 J/kmol σE = 12.55×106 J/kmol 14 k0 = 8.57×10 1/s Mv = 40 kg/kmol
Tortuosity factor of convective pores Dynamic viscosity of volatiles Pressure in fluidized bed
τpc = 3 µV = 3.62×10-5 Pa·s P0 = 105 Pa
Table 2. Properties of tested coals Brown coal Density, kg/m3 Volatile matter content, % wt. Fraction of convective pores, % Average pore radius, nm Total porosity of coal, % Maximum char porosity, %
950 42.0 20.0 1.9×10-7 37.77 60
Hard coal Julian Wujek 1260 1320 37.5 34.0 1.0 1.0 5×10-8 4×10-8 13.75 11.4 40 35
Anthracite 1700 3.0 0.1 2×10-10 5.27 10
Figures 3–5 show temperature profiles, volatile matter release profiles, and devolatilization rate profiles calculated for a 5-mm Julian coal particle burnt in a bubbling fluidized-bed at 850°C. Figure 3 shows large temperature gradients existing inside the particle and caused by the resistance to heat transfer. These results suggest that the internal heat conduction is the controlling mechanism for devolatilization of large particles in fluidized-bed combustors. The devolatilization process is relatively slow and requires approximately 20 s to release volatiles contained in the particle (Figure 4). At the initial stage of devolatilization, the devolatization rate passes through the maximum near the surface of the particle (Figure 5). With time, this maximum moves towards the centre and its value decreases. Figure 6 shows the maximum calculated pressures in the centre of 5-mm particles and times when they occur for different coals. The highest value of ~51 MPa was obtained after 13 s for Wujek coal. This result corresponds to the extensive primary fragmentation observed during combustion tests with particles of this coal. The
lowest values were obtained for brown coal and anthracite; ~0.5 and ~6.0 MPa respectively. These two coals did not undergo the primary fragmentation at all (5). 0.40
1000
0.30
V, kg/kgcoal
Temperature, ˚C
800
600
400
5s
r=0 r = 1.25 r = 1.75 r=2 r = 2.25 r = 2.5
200
0.20
10 s
0.10
15 s 20 s 0.00
0
0.0 0
10
20
30
0.5
40
Figure 3. Calculated temperature profiles for a 5-mm particle burnt at 850°C
1.0
1.5
2.0
2.5
Radius, mm
Time, s
Figure 4. Calculated volatile matter release profiles for a 5-mm particle
0.12
60
0.10
50
0.08
40
12 s - Julian hard coal 13 s - Wujek hard coal 15 s - brown coal
Pressure, MPa
Γ, kg/kgcoal•s
17 s - anthracite
0.06 5s 0.04
10 s
30
20
15 s 0.02
20 s
10
0.00 0.0
0.5
1.0
1.5
2.0
Radius, mm
2.5
0 0.0
0.5
1.0
1.5
2.0
2.5
Radius, mm
Figure 5. Calculated devolatilization rate Figure 6. Calculated maximum pressures profiles for a 5-mm particle at 850°C for different coals at 850°C VERIFICATION OF THE MODEL Figure 7 shows a comparison of model predictions and experimental data for temperature profiles as a function of time for a 10-mm hard coal particle burnt in a bubbling fluidized-bed at 850°C. At the early sta ge of heating, some discrepancies between measured and calculated temperatures are seen. These discrepancies are caused by the uncertainty in estimation of the heat transfer coefficient at the surface of the particle (influence of radiation), and measurement errors due to heat conduction through the sleeve of the thermocouple that measured temperature in the centre of the particle. Figure 8 shows measured and predicted values of devolatilization time as a function of particle size at 850°C. The agreement between exper imental data and calculated values is good especially for smaller particle sizes. The comparison of model predictions and experimental data of fragmentation time for 5 and 10-mm coal particles from Julian and Wujek mines, burnt at temperatures 800–900˚C is shown in Table 3. Again, the agreement between measurements and predictions is good.
The accuracy of the model predictions indicates that the model is extremely capable of predicting the devolatilization and fragmentation processes for large coal particles during fluidized-bed combustion. 1000
60 Julian - experiment
Devolatilization time, s
800
Temperature,˚C
Sobieski - experiment
50
600
400 Centre - model Surface - model
Julian - model Sobieski - model
40
30
20
10
200 Centre - experiment Surface - experiment
0
0
2 0
20
40
60
Time, s
80
100
4
6
8
10
Diameter of coal particle, mm
Figure 7. Experimental and predicted Figure 8. Experimental and predicted temperature profiles for a 10-mm particle values of devolatilization time Table 3. Experimental data and model predictions for fragmentation time of 5 and 10 mm coal particles from Julian and Wujek mines burnt in a bubbling fluidized-bed at temperatures 800–900˚C
Experiment, tf1, s Model, tf1, s (calculated pressure, MPa) Experiment, tf1, s Model, tf1, s (calculated pressure, MPa) Experiment, tf1, s Model, tf1, s (calculated pressure, MPa) Experiment, tf1, s Model, tf1, s (calculated pressure, MPa) Experiment, tf1, s Model, tf1, s (calculated pressure, MPa) Experiment, tf1, s Model, tf1, s (calculated pressure, MPa)
Coal type Diameter of coal hard coal hard coal particles, mm Julian Wujek Temperature 800˚C 5 13-18 12-16 15 14 5 (27) (47) 10 20-25 32-38 22 34 10 (36) (58) Temperature 850˚C 5 10-15 9-15 12 13 5 (32) (51) 10 22-30 25-34 27 28 10 (41) (65) Temperature 900˚C 5 8-13 9-14 11 12 5 (34) (53) 10 22-27 22-28 24 26 10 (47) (68)
NOTATION Cp E, E0 FA, FR h
- specific heat capacity, J/kgK - activation energy, J/kmol; mean activation energy, J/kmol - resultant of tensile force, N; resultant of resistance force, N - heat transfer coefficient, W/m2·K
k k0 MV (MR) NV P, P0 r R t, tf1 T V * V
- heat conduction coefficient, W/m·s - pre-exponential factor in Arrhenius rate expression, 1/s - molar mass of volatiles, kg/kmol - universal gas constant, J/kmol·K 2 - molar flux of volatiles, kmol/m ·s - pressure inside the particle, Pa; ambient pressure, Pa - radial position, m - particle radius, m - time, s; time of primary fragmentation, s - temperature, K - volatiles released, kg/kgcoal - maximum amount of volatile matter that can be released, kg/kgcoal
Greek symbols
φ,φc φ pc, φ *ch Γ
ρ
σ σcr σE
τ pc
- porosity %, porosity of coal, % - porosity of convective pores,%; maximum porosity of char, % - devolatilization rate, kg/kgcoal·s - density, kg/m3 - stress, N/m2 - compressive strength, MPa - standard deviation in activation energy, J/kmol - tortuosity of convective pores
REFERENCES (1) Kosowska-Golachowska M. and Luckos A. (2009) An experimental investigation into the fragmentation of coal particles in a fluidized-bed combustor, Proc. of the 20th Inter. Conf. on Fluidized Bed Combustion, G. Yue, H. Zhang, C. Zhao and Z. Luo (Eds.), Tsinghua Univ. Press, Beijing, 30-334. (2) Chirone R., Massimilla L. (1989) The application of Weibull theory to primary fragmentation of coal during devolatilization, Powder Technology 57, 197-212. (3) Sasongko D., Stubington J.F. (1996) Significant factors affecting devolatilization of fragmenting, non-swelling coals in fluidized bed combustion, Chemical Engineering Science 51 (16), 3909-3918. (4) Dakic D., Grubor B., Oka S. (2000) Statistical model for prediction of primary fragmentation of coal particles in fluidized bed combustion – comparison with experimental results, 41th IAE Fluidized Bed Conversion Meeting, Italy. (5) Kosowska M. (2005) The thermal fragmentation of coal during combustion process. PhD Thesis. Czestochowa University of Technology. (6) Stubington J.F., Sumaryono (1984) Release of volatiles from large coal particles in a hot fluidized bed, Fuel 63, 1013-1019. (7) Dakic D., Ilic M., Oka S. (2000) Pressure increase inside a coal particle during devolatilization in a fluidized bed combustor – new results of mathematical modeling, 41th IAE Fluidized Bed Conversion Meeting, Italy. (8) Anthony D.B., Howard J.B., Hottel H.C., Meissner H.P. (1976), Rapid devolatilization and hydrogasification of bituminous coal, Fuel 55, 121-128. (9) Gavalas G.R., Wilks K.A. (1980) Intraparticle mass transfer in coal pyrolysis. AIChE 26, pp.201-211. (10) Bliek A., van Poelje W.M., van SwaajW.P.M., van Beckum F.P.H. (1985) Effects of intraparticle heat and mass transfer during devolatilization of single coal particle, AIChE 31, pp.1666-1681.
International Pittsburgh Coal Conference 2010 Istanbul, Turkey October 11-14, 2010 Abstract Submission PROGRAM TOPIC:
FLUIDIZED-BED COMBUSTION AND CO-FIRING
MAIN PROBLEMS CONCERNING CO-FIRING BIOMASS MIXTURE WITH HARD COAL IN PULVERIZED-FUEL BOILERS Krzysztof JESIONEK, Prof. DSc., PhD. Eng Marcin KANTOREK, PhD Student Wrocław University of Technology, Faculty of Mechanical and Power Engineering Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, Poland [email protected] Henryk KARCZ, PhD. Eng Wrocław University of Technology, Faculty of Mechanical and Power Engineering, ZBUS-TKW Combustion, Sp. z o.o. ul. Sikorskiego 120, 95-015 Głowno, Poland [email protected]
Abstract: The methods of burning used in Poland are limited mainly to stocker-fired and fluidic fuel boilers application as well as to co-burning coal powder mixtures in pulverized fuel boilers. The co-burning of timber with coal results however in lowering thermal efficiency of steam producing installation or impairs its operational reliability i.e. increases its running costs and lowers its safety or service. Generally speaking it results from the difference in kinetic characteristics at timber and coal. Into consideration were taken typical pulverized fuel boilers working according to the method established for Polish power engineering for feeding with hard coal. Operating indexes of boilers were analyzed (fired with mixture of biomass and hard coal) bound with combustible portion ratio in slag and volatile ash. The analysis was made for the boilers OP-1050, OP-210, OP230, OP-430, OP-650, applying the normal combustion chambers. Results obtained concern the maximum boiler heat-load and maximum number of working mills defined as optimum quantity for a given type of boiler. The unburned part in slag and volatile ash has been marked as that ment for a standard boiler fuel and for various mixtures of biomass and hard coal.
1. Introduction Large electric and thermal energy providers such as power plants and thermal-electric power stations strive to fulfil their commitments as regards meeting the limits defining the share of energy produced from renewable resources in relation to the total production of energy from conventional sources. These restrictions limit the amount of combusted biomass. In the case of small energy units, biomass combustion does not interfere significantly with the boiler‟s basic operating parameters, i.e. efficiency (r|). However, in the case of large energy units, this parameter may be significantly reduced. Other operating parameters such as operating reliability, dependability and installation safety may also deteriorate seriously. Differences regarding physical-chemical properties are present both in organic and mineral substances. This results in diametrically different behaviour of the substances in both mill and boiler systems, which, as a consequence, results in a significant deterioration in the operating values of the power unit. So far the solid-state biomass combustion methods used in Poland have been limited mainly to co-combustion of a mixture of forest biomass and coal in stoker-fired boilers and fluidised-bed boilers, and co-combustion of dust mixtures in pulverised-fuel boilers. In the case of stoker-fired boilers and fluidised-bed boilers, the combusted wood is in the form of sawdust, pellets, briquettes or chips, while pulverised-fuel boilers use dust for combustion. In all these cases, co-combustion of wood, regardless of its form, with coal results in reducing the boiler‟s thermal efficiency, increasing the operating costs and impairing the safety of operation. This is partly a result of the fact that the kinetic characteristics of wood differ significantly from those of coal. When biomass-coal mixtures are combusted in stoker-fired boilers, the combustion of the wood on the grate is much more rapid and results in increased chimney loss (Sk) through the increase in the amount of the air remaining unused on the grate and the increase in the waste gas exhaust temperature caused by the higher contamination of the boiler‟s heating surfaces by flue dust, mainly from the mineral substances present in wood. Reduced thermal efficiency of fluidised-bed boilers is also caused by contamination of boiler‟s heating surfaces by flue dust, slagging of the chamber walls with fluid cinder and, as a result, an increase in the waste gas exhaust temperature. The increase in the degree of boiler heating surface contamination results mainly from the biomass ash physical structure. The structure and physical properties of biomass ash is completely different from coal ash due to mineral substances. Biomass ash has a much lower apparent density and more developed external surface. It has a structure of flocculent flat plates that create larger agglomerates, a softening temperature of 1000-1200 °C, but does not have a granular structure with smooth surfaces and higher softening temperature. Intensive contamination of stoker-fired and fluidised-bed boiler heating surfaces is a feature characteristic for boilers fired by wood biomass. Another significant loss resulting from biomass combustion is the incomplete combustion loss in fly-ash. This loss is caused by removing fly-ash particles from the fluidised bed by the waste gas stream. This loss sometimes exceeds the permissible 5 % threshold of combustible
element content in ash. The incomplete combustion loss in fly-ash results mostly from very low apparent density of carbonisate caused by rapid wood pyrolysis in the fluidized bed. As the buoyant force of carbonisate granules is much higher than that from gravity, fly-ash granules float freely in the waste gas stream. As far as it may be viable to combust other forms of plant biomass than wood (in special cases even animal biomass may be used), in stoker fired boilers and particularly fluidised-bed boilers, and in pulverized-fuel boilers only wood biomass may be used. In the case of the latter, biomass from different sources may not be due to transport and milling obstacles resulting from its physical structure and physical-chemical properties. If wood biomass is co-combusted with coal dust, difficulties are encountered due to the lack of wood milling properties. When wood is present in a coal pulveriser, its substance is not milled but pulped and crushed by the milling elements. Since division and separation of granules in the mill is achieved on the basis of differences between particle weights, even large wooden splinters may not be separated as the “mesh fraction” but transported to the boiler‟s chamber with the dust stream as the so-called “minus mesh”. This phenomenon results from the fact that the apparent density of wood in the dry state may be 2-3 times lower than that of coal. The analysis of the combustion of biomass in the form of sawdust, chips and pellets was carried out using standard pulverized-fuel boilers operating in Polish power engineering units. The analysis was carried out on the basis of collected data regarding incomplete combustion losses in cinder and flue dust of standard boilers fuelled with hard coal. The analysis was carried out for OP-150, OP-210, OP-230, OP-410, OP-430 and OP-650 boilers with a standard combustion chamber and OP-650 with an enlarged combustion chamber and for the BP-1150 shaft boiler.
2. Carbonisate content in cinder and flue dust in pulverised-fuel boilers of various thermal power values The results of the designation of fly-ash content in cinders depending on the biomass mass fraction in coal dust based on boiler nominal thermal load are presented in figure 1, and the fly-ash content in flue dust is shown in figure 2. On the basis of the results in figure 1, it is clear that the content of uncombusted material in the cinder mostly depends on the boiler type and the biomass share in the fuel supplied to the combustion chamber. As the boiler chamber diminishes together with the decrease in its efficiency, the uncombusted element loss in the cinders grows as the boiler efficiency decreases. Based on the dependencies shown in figure 1, one may conclude that for OP210 and OP-230 boilers the permissible share of the coal biomass, remembering that the permissible value of fly-ash in cinder must not be exceeded, should be lower than 6 %; for OP-380 and OP-430 boilers this should be lower than 8 %; for OP-650 boilers this should be lower than 9 % and for OP-650 boilers with an enlarged combustion chamber this should not exceed 10 %. For BP-1150 shaft boilers the mass fraction of co-combusted biomass should be lower than 11 %.
Fig. 1. Carbonisate (bkz) content in the cinders depending on the combusted biomass mass fraction (bb) and boiler type
Fig. 2. Fly-ash content (bkp) in flue dust depending on the combusted biomass mass fraction (bb) and boiler type
The fly-ash content in flue dust is shown in figure 2. The uncombusted material loss in the flue dust, depending on the boiler type and the amount of co-combusted biomass, is similar to the case of carbonisate content in the cinders. The permissible fly-ash content in flue dust is, however, exceeded with a much higher biomass share in the combusted coal. For example, the permissible fly-ash content in flue dust of an OP-210 boiler is exceeded for biomass co-combustion above 9 % of the mass fraction, which is 3 % more than in the case of cinders. The influence of biomass mass fraction on the fly-ash content in the flue dust is noticeable only when the biomass share in the combusted fuel is considerable. For high-power boilers (over 400 t/h of vapour) the permissible limits of fly-ash content in the flue dust is observed only when the biomass content is above 10 %, and for shaft boilers this value must be above 13 %, Fig.. 2.
3. Mechanism of uncombusted material creation in cinder and flue dust The coal-biomass dust provided by the burner undergoes the processes of drying, degassing and combustion of thermal decomposition products (pyrolitic gas and carbonisate) in the dust flame. The pyrolitic gas combusts in the initial flame stage, creating a „flame front‟. The carbonisate (solid remains after degassing) is ignited in the flame front and combusts in the final part of the flame. The characteristics of particle size distribution of the dust supplied to the burner show a clear tendency towards a dimensional particle size distribution in the dust duct section and in the powdered-fuel burner nozzle. In the middle nozzle section occur mostly dust granules of the smallest dimensions; however, granules of the largest dimensions occur on the rim. As the combustion rate of the smallest granules is much higher than the large granule combustion rate, the carbonisate burnout degree in the flame core is full and in the peripheral section the large carbonisate granules remain uncombusted. As the peripheral part of the dust flame is adjacent to the boiler chamber wall zones, where the temperature and oxygen content are relatively low, there are sometimes no conditions for the large carbonisate granules to combust fully. In adverse conditions, they are very often extinguished and fall into the cinder hopper in the form of carbonisate. In the wall zone, where the oxygen concentration is very low and temperatures are lower than the flash-point, large carbonisate granules are not ignited and instead fall to the cinder hopper thus constituting a combustion loss to the cinders. The process of dropping the fly-ash into the cinder hopper is particularly facilitated by the negative pressure present in this area (from 0 to -100 Pa) and the lack of flow of gases from the cinder hopper to the upper part of the boiler chamber. Thus, the presence of fly-ash in the cinder hopper is caused mainly by relatively large and heavy carbonisate granules from the crushed xylem. The flow process of carbonisate into the cinder hopper is characterised by the carbonisate granules stream falling into the hopper in those areas directly adjacent to the screening wall outlines or in its direct vicinity. The density of the stream of falling carbonisate granules in the cinder hopper axis drops almost to zero.
The test results show that the amount of falling carbonisate in the falling cinder flow shown in figure 3 was obtained from an OP-410 boiler fired with hard coal and biomass dust.
Fig. 3. The carbonisate drop distribution in the longitudinal and lateral axis of the cinder hopper horizontal section The process of carrying away fly-ash particles as a result of biomass granule degassing is a typical problem in pulverised-fuel boilers. This is due to the low apparent density of carbonisate granules and the high resistance factor, which is shown by a considerable prevalence of the buoyancy force over the gravitation force and means that particles are easily lifted by flowing exhaust gases. This problem is particularly important in boilers with corner (tangential) burners where, in the middle of the chamber, a central whirl is created in which the gas flow velocity is much higher than in other chamber sections. As a result of differences in the flow velocity, temperature field and gas concentration, a large content of biomass fly-ash flows through the central whirl. All the larger uncombusted biomass particles located in the central whirl operation zone are captured by the rapidly flowing exhaust gases and carried away from the boiler chamber.
4. Possibility of reducing the fly-ash content in flue dust The fly-ash carried by exhaust gases is highly reactive and may be combusted in the upper boiler part before the inlet to the platen superheater. The process of complete combustion of fly-ash must be, however, stabilised and catalysed by the liquid fuel flame, preferably by a liquid fuel belonging to renewable energy sources. The waste liquid fuel in this case, which in the future may be regarded as a renewable energy source, is technical glycerine obtained as a waste product from biodiesel fuel manufacturing using vegetable oils, e.g. rape oil, tall oil or animal fat. The content of fly-ash in flue dust sampled from the charging hopper of the electrofilter is shown in figure 4. The diagram shows the dependency of the uncombusted fly-ash elements on the boiler heat load.
Fig. 4. Uncomplete fly-ash combustion loss in flue dust depending on boiler power (N) Curve 1 shows the content of carbonisate granules in the flue dust without activating the glycerine burners. The uncombusted material comes mainly from the xylem. Separated carbonisate granules have an oblong form with clearly visible fibres featuring very well developed external surfaces. Microscopic examinations have showed a structure of loosely located fuzzy layers with lengthwise-directed orientation, being very porous with a very low specific gravity value. The uncombusted material amount (Slk) grows rapidly from 1.5 % with a 130 MW boiler load to 4.5 % with a 220 MW load. With this configuration and
when only the powdered-fuel burners are active, the minimum limit of safe operation is approximately 130 MW, which is equal to approximately 60% of the heat load. The content of uncombusted material in the flue dust when the glycerine burner located in a boiler‟s corner is showed by curve 2. The amount of uncombusted material (Slk) in flue dust has been almost halved for higher boiler heat loads. The decrease in the uncombusted material losses in the flue dust for low boiler load is inconsiderable at approximately 0.5 %, while the technical minimum value of the boiler has also been decreased. The boiler operated stably and reliably with a 120 MW heat load, which is approximately 54 % of the nominal load. Installing a glycerine burner in the boiler wall axis considerably decreased the content of solid combustible particles in the flue dust (curve 3). The amount of carbonisate in the flue dust was decreased to about a quarter across the entire heat load range. Curve 3 depicts the percentage dependency of the combustible particles share in flue dust depending on heat load, which is almost flat. The uncombusted material value changes from 0.25 % with the lowest load to 0.62 % with the maximum loads. Locating the glycerine burner in the boiler wall axis enabled the burning glycerine flame to reach the boiler chamber area where the highest fly-ash concentration in flowing waste gases was present. The tangential location of the powdered-fuel burners creates a flame in the central part of the combustion chamber, whirling around the boiler axis and gradually transforming into a block of exhaust gases containing a considerable amount of flue dust and fly-ash, and creating a “chimney” moving towards the combustion chamber outlet, which is confirmed by mathematical modelling based on a 3D model, [1]÷[3]. The flow of the main exhaust gases mass, which contains almost 90% of the flue dust granules and carbonisate, thorough the central zone of the higher combustion chamber section ensures full catalytic combustion of almost all the flyash when the glycerine burner is located on the chamber wall axis. This may be achieved even without providing any additional amount of air to fully combust the volatile carbonisate. The oxygen necessary for full combustion of the carbonisate comes from the waste gases (on average 6÷8 %). The results confirmed the assumption and the importance of catalytic action of the liquid fuel flame zone located near the flame, which is rich in oxygen and hydroxide radicals as a carbonisate combustion initiation factor catalyzing the C oxidation. A vital problem in the carbonisate catalytic combustion process is the range of coverage of the reactive space by the flame from the combusted liquid fuel, [4],[5]. The intensive process of carbonisate combustion takes place mostly in the zone located near the flame, which is rich in oxygen and hydroxide radicals catalyzing the C oxidation. This regularity forces us to create conditions in the boiler chamber where the largest possible part of the chamber cross-section is covered with the liquid fuel flame (a glycerine flame in this case), especially the zone where the highest fly-ash concentration is present, [6], [7]. Taking into account the fact that the concentration of fly-ash is the highest in the central part of the chamber (known as the “chimney” zone, featuring the highest carrying velocity, the largest draft in the chamber outlet direction), the glycerine burner flame should permeate this zone of the furnace chamber.
5. Conclusions •
•
•
The amount of biomass combusted depends on the boiler power. In OP210 and OP-230 boilers the maximum amount of combusted biomass is 6 % of the mass fraction, in OP-410 and OP-430 this maximum value is 9 %, in OP-650 boiler with a standard chamber this value must not exceed 10 %, in OP-650 boilers with an enlarged chamber the maximum value is 12 % and in the case of OB-1150 pulverized-fuel monotube boilers this vale must not exceed 13 % of the mass fraction. Decreases in the uncomplete combustion loss in fly-ash is very effective and considerably lowers the value of this loss through the introduction of a liquid fuel burner into the boiler chamber zone between the powderedfuel burners and the platen superheater. As far as corner powdered-fuel burners are concerned, the degree of full fly-ash combustion is the highest when an oil burner is located on the boiler chamber wall axis.
Bibliography [1] Karcz Głąbik R., Kantorek M., Modliński Zb., Rączka P.; „Efekty ekonomiczne spalania biomasy w przedpalenisku kotła energetycznego”. Energetyka 2010, nr 2 (luty), s. 115-127. [2] Głąbik R., Rzepa K., Modliński Zb., Sikorski Wł., Kosiorek-Herbuś A., Karcz H.; „Obliczenia cieplne kotła OP-150 dla różnych udziałów energetycznych biomasy w paliwie” Prace Naukowe ImiUE Pol. Śląskiej 10th International Conference on Boiler Technology, Seria: Monografie Konferecje, 2006, z. 16, t. 1. [3] Karcz H., Modliński Zb., Głąbik R., Kosiorek-Herbuś., Kurzelewski J.; „Wpływ spalania biomasy w przedpalenisku na aerodynamikę kotła energetycznego”, Konferencja Naukowo-Techniczna pt” Nowoczesne technologie spalania węgla i paliw odpadowych, Szczyrk 15-17 listopada 2006, s. 107-120. [4] Karcz H., Butmankiewicz T., Machnikowski J., Andryjowicz Cz.: „Sposób i instalacja wykorzystania produktów niecałkowitego spalania”, Patent do zgłoszenia P361112 z dnia 07.07.2003 r. [5] Karcz H., Butmankiewicz T.: „Sposób zmniejszania straty niecałkowitego spalania koksiku w energetycznym kotle pyłowym oraz energetyczny kocioł pyłowy do realizacji tego sposobu”, Zgłoszenie Patentowe P388591 z dnia 21.07.2009. [6] Karcz H., Chzmaljan D.: Kinetika sžiganija ugolnych zioren sodieržauščich letuše veščestva. Archiwum Termodynamiki i Spalania 1979, Vol 10, Nr 1, s. 85-103. [7] Standa J., Karcz H.: Neue Aspekte der diffuskinetoischen. Verbennungs teorie von festen Bernnstoffen im Lichte der Untersuchungen der inneer Struktur von verbrannten Kohlenkőrnern, VDI-Gesellschaff Energietechnik. Verbernnung Und Feuerungen, 16 Deustscher Flammentag Tagung. Clausthal, 14 um 15 September 1993, VDI verlag, s. 227-342.
TECHNICAL AND ECONOMIC EVALUATION OF THE DESULPHURIZATION PROCESSES AT POWER STATIONS USING LIGNITE Hasancan Okutan, Prof.Dr. İstanbul Technical University, Chemical-Metallurgical Faculty, Chemical Engineering Department 34469, Maslak-İstanbul – TURKEY Bülent D. Çift, Dr. İstanbul Technical University, Chemical-Metallurgical Faculty, Chemical Engineering Department 34469, Maslak-İstanbul – TURKEY Abstract Due to industrial development and population growth Turkiye’s energy demand is increasing every year. Although renewable energy sources are becoming widespread these sources are not sufficient enough to meet all of the energy demand of Turkey in the near future. Therefore fossil fuels are going to be used for a long time for domestic heating and electricity production. The disadvantage of using fossil fuels is sulfur dioxide (SO 2) emissions during the combustion according to the sulfur content of the fuel. Turkey’s energy consumption depends heavily on fossil fuels. Therefore SO2 emission level is going to be higher than the legal limits according to Industrial Based air Pollution Control Regulation since our lignite has a low calorific heating value and high sulfur content which are 1807 kcal/kg and 1.85 % respectively. Although there are disadvantages of using lignite for electricity production it’s our native energy source and by applying appropriate desulphurization methods it is possible to reduce emission of SO2 to lower levels. Removal of SO2 is not only legal obligation but also an act for protection of the public health and environment. Because an important effect of SO 2 is it’s contribution to acid rain. Acid aerosols cause chlorosis, the loss of chlorophyll and plasmolysis, tissue collapse of leaf cells in plants. In this study, most prevalent way to prevent SO2 emission, desulphurization flue gas after combustion of fossil fuel especially lignite, is elaborated wet scrubbing, spray dry and dry sorbent injection methods are investigated. A computer program is written to determine the lowest capital investment among those three methods of desulphurization is found as dry sorbent injection capital investment cost model for certain methods are developed as a function of plant capacity and sulfur content for trona sorbent.
1. Introduction: Due to industrialization, Turkey’s energy demand is increasing every year as in all over the world. In order to supply this enegy demand Turkey is searching new and renewable energy sources in electricity production but neglect using lignite, a native source, that have a reserve of 8.3 billion tons. Turkish lignites have low calorific value (65 % is less than 2000 kcal/kg) and have high sulphur content (1.4 - 4.65 %). Because of these two properties lignites have a disadvantage for air-environment pollution and for human health (Ekinci,1994). To overcome this disadvantage those methods are used; sulphur removal from lignite, sulfur dioxide (SO2) removal during the combustion and flue gas desulfurization (FGD) after combustion. In sulfur removal from coal, conventional and advanced physical techniques are used such as crushing, grinding, jigler, heavy medium tank, cyclones and floating 10-40% of total sulfur can be removed and in order to remove organic sulfur from coal chemical and biological methods are necessary. By using physical methods it costs 1 to 5 $ per ton coal and it takes 1-2 years to install a plant (Özsoy, 1998). As a result desulfurization after combustion seems to be the most economical way to remove SO2 from flue gas. Coal’s share in world’s electricity production is 39%. And in Turkey coal’s share is 25% (Vatan, 2006). And totally 53 % of EÜAŞ’ (Turkey Electricity Generation Company) have FGD units (Çift, 2008). In 47% of EÜAŞ’ thermal power plants don’t have FGD units. In Seyitömer, Soma, Kangal1-2, Tunçbilek and Afşin-Elbistan A coal fired thermal power plants FGD units are absent. In Table 1 required desulfurization rates according to the thermal thermal heating value, coal sulfur content and capacity of the power plant using lignite and hard coal at EÜAŞ’ thermal power plants. Calculations are based on these
1
assumptions ; 37% thermal efficiency and 100 % of sulfur in the coal is converted to SO 2 (In fact 75% of sulfur is combustible (EPA,1985)).
Quantity of Coal (ton/h)/ Unit
2.2
312
1657000
93
1765 2473 2700 2550 3720 2080 3040 2870 2470
2.5 3.17 2.1 3.5 1.06 3.3 1.54 2.2 3.17
277 141 181 137 97 235 50 170 141
1026700 732200 1026200 733600 757700 1026000 320000 1025000 732000
90.3 89.3 83 90 53 91.4 65 82.1 89.3
Required Minimum Desülfurization (%)
Sulphur %
2530
Amount of Flue Gas 3 (Nm /h)/Unit
Thermal Heating Value of Coal(kcal/kg)
1. Afşin-
Installed Capacity (MW)
Name of Thermal Power Plant
Table 1. Required desulfurization rates EÜAŞ’ thermal power plants
2. Yeniköy 3. Kangal 4. Orhaneli 5. Çayırhan 6. Soma B 1-2 7. Kemerköy 8. Tunçbilek 3 9. Yatağan 10. Kangal
3x340+ 1x335 2x210 2x150 1x210 2x150 2x165 3x210 2x65 3x210 3x150
11.
Seyitömer 1-2-3
3x150
2600
1.31
134
732000
73
12.
Çatalağzı (Hard Coal)
2x150
3870
0.7
90
732 000
25
Elbistan B
In this study tecnical and economical investigation is aimed for certain desulfurization methods which remove/reduce the adverse effets of sulfur dioxide (SO2) on environment and human health arise from emissions of lignite using thermal power plants. Thus, lignite, native energy source of Turkey can be evaluated in a more effective way. Orienting to this aim, existing situation of the reserves of lignite and it’s average physical and chemical properties are given, the post-combustion desulfurization units used in the world and in Turkey are investigated. Among the desulfurization units; the most prevalent wet method, becoming widespread spray dry method and developing dry sorbent injection methods are selected and examined. As a sorbent that is going to be used in these methods , lime, limestone and trona are selected. In order to lower the sorbent cost which is the most important parameter effecting the economics of desulfurization processes and to increase the efficiency of SO 2 reduction, evaluation of BEYPAZARI trona reserves is researched. For economic comparison among the desulfurization techniques capital investments of them are calculated by using Gutrie Method. A computer programme in Turbo Pascal programming language is written in order to calculate the capital investment and operation cost for the selected methods using selected sorbents by taking the capacity of thermal power plant calorific value and total sulfur content of the coal as indepentend parameters. By using this programme capital investment and operating costs can be calculated for three different methods. The capital investment costs of desulfurization processes are demonstrated by 3-dimensional graphics as a function of the thermal power plant capacity and coal sulfur content and capital invesment models for these methods are developed. 2.
Computation of Capital Investment and Operating Costs of Desulfurisation Methods For estimating the capital investment and operating costs of three major FGD systems (dry injection, wet and dry methods) following steps are persecuted (Ulrich, 1984);
2
a. Detailed process flow sheet is prepared, b. Material balance is prepared, c. Equipment list with capacities is obtained, d. Utility consumption is listed. Due to the flow diagram, material balance and equipment list, equipment capacities and purchase equipment costs are determined. In determining purchase cost Module Cost Method which is developed by Gutrie is used. This method is accepted as the most accurate method to determine the preliminary cost estimation. This cost model depends on purhase equipment costs under some base conditions. In the case of deviation from base case the base cost is multiplied by coefficients which are determined by equipment type, system pressure and construction material. Formula which is given below is used to calculate the bare module cost (Turton, 2003); CBM = Cpo*FBM
(1)
CBM : Bare module equipment cost FBM : Bare module cost factor Cpo: Purchase cost generally for equipments constructed by carbon steel under normal atmosphere pressure Capital investment of FGD unit is calculated by direct and indirect costs as a function of equipment purchase cost. Direct costs include instrumentation and control, freight,site preparation, sales tax, electrical and piping and indirect costs include engineering, installation, start-up, contractor fees and contingencies. Purchase equipment cost for another capacity is calculated by using an exponential factor of an equipment and cost index (Heinsohn, 1999). Cvr =Cur*(v/u)a
(2)
Cvr: Purchase equipment cost which have the capacity of v in the year r Cur : Purchase equipment cost which have the capacity or size of u in the year r Cvs = Cvr(Is/Ir)
(3)
Cvs : Purchase price of equipment in the year s Cvr: : Purchase price of equipment in the year r Is, Ir : Index of the years s and r respectively (from CE, M&S, ENR, etc.) And in the Formula given below, total module cost and grass root can be calculated;
CTM
N i 1
N
CTM, i
1.18
CBM, i
(4)
i 1 N
C GR
C TM
0.50
C BM, i
(5)
i 1
CTM : Total module cost refers to the cost of making small to moderate expansions or alterations to an existing facility. CGR : Grass roots cost refers to a completely new facility in which we start the construction on essentially undeveloped land, a grass field (Turton, 2003). A computer programme is prepared in Delphi 7.0 programming language. As a base case, for spray dry system 200 MW , for dry injction 100 MW, for wet system 210 MW installed capacity is chosen for a power station using 2000 kcal/kg heating capacity and
3
2% sulfur content of coal. Assumptions made for calculations are as follows; 40% excess air used in combustion chamber, in calculation of equipment capacities 2% total sulfur is accepted as combustible (in real 75 % of total sulfur is converted to SO2 (EPA,1985) ). The other assumptions are as follows; in dry injection method, approach to saturation temperature is 400C, Ca/S and Na/S molar ratio in sorbent feed is 2, flue gas pressure drop is 43 mbar, 12 days of sorbent stock , limestone density is constant and 2500 kg/m3, thermal efficiency is 37% (in burner), efficiency of electrical motors is 0.7, ash density is constant and 2800 kg/m3, flue gas exit temperature is 700C, purchase equipment cost datum are for base case (carbon steel and atmospheric pressure), trona as an ore have 84% purity, Ca/S=1.05 in wet method and 1.1 in spray dry method, thermal power station works with load of 100%, desulfurization rate in dry method is 75% for limestone, 85% for limeand 90 % for trona. In spray dry method desulfurization rate is 90% for lime, 93% for trona (Soud, 1994). In wet method desulfurization rate is 95% for limestone and trona (Barbour, 2002). Electricity price is 8 cent/kwh and water cost is 1$/m3 , waste disposal cost is 5 $/ton (for spray and dry method) (Jons, 1986). 8 seconds of residence time in spray dryer chamber (Sedman, 1986). One important point is that SO2 emission rate is limiting the sorbent selection because of the legislative obligation , SO2 emission rate should not exceed 2000 mg/m3 for the capacities smaller than 100 MW, 1300 mg/m3 for capacities between 100 and 300 MW, 1000 mg/m3 for capacities higher than 300 MW. Since in July 22, 2006, Regulation of Industrial Based Air Pollution has become effective and emission limits mentioned above can not be exceeded. And absorbents used especially in dry injection method are of great importance for fulfilling the emission limits. Because in dry injection SO2 reduction capability can be increased up to 90%. Also the sulfur content, the thermal heating value of coal and installed capacity of the power plant play an important role in emission rate. Therefore these variables are chosen as key parameters for the computer programme. 3. Results and Discussion By using computer programme with the above assumptions capital investment costs of the 3 different types of FGD units for different sorbents are calculated (Çift, 2008). Also annual operating costs related to these methods and sorbents are calculated. The parameters are installed capacity (MW), coal sulfur content (S%), thermal heating value of coal (kcal/kg) and sorbent type. In dry injection method for a 100 MW installed capacity, 2% of total sulfur in coal and 2000 kcal/kg heating value of coal 6.3 million dollars capital investment is required when lime is used as sorbent. For the same parameters in spray dry metod 13 million dollars of capital investment is found for lime as a sorbent. For the same capacities and coal characteristics wet-limestone’s capital investment is 18 million dollars. If limestone is used as a sorbent in dry injection method the capital investment decreases to 5.5 million dollars for the same parameters. In Figure 1 the total capital investment (TCI) of three methods (dry, spray and wet) for trona sorbent as a function of installed capacity of power plant is given. As can be seen from Figure 1 the below capital investment model is obtained and dry injection method has the lowest value (for 2000 kcal/kg and 2% sulfur coal). TCI = 375 900* (MW)0.6146
(6)
If limestone is used as a sorbent in dry injection method this model is obtained; TCI = 230 600* (MW)0.6969
(7)
In spray dry method exponantial value of installed capacity is 0.64 and for wet system this is 0.63.
4
Figure 1. Total Capital Investment of Three different method using trona as a sorbent 70
0,6334
Capital Investment (Million $)
y = 1,1011x 2
R = 0,9997
60 50
y = 0,7179x0,6438 2
R = 0,9999
40
Dry Injection Method Spray Dry Method
30 y = 0,3759x0,6146
20
2
Wet Method
R = 0,9857 10
Üs (Wet Method)
0 0
100
200
300
400
Capacity (MW)
500
600
700
Üs (Spray Dry Method ) Üs (Dry Injection Method)
If the limitations of desulfurization rates of sorbents is taken into consideration these results are obtained ; for 100 MW installed capacity and trona is used as a sorbent in dry injection case total sulfur content of coal should not exceed 5% for 2000 kcal/kg heating value of a coal of thermal heating value of 1500 kcal/kg used in this case total sulfur should be below 4.6%. For capacities between 100 and 300 MW total sulfur of the coal used should be below 4% for 2000 kcal/kg heating value. For 1500 kcal/kg sulfur of coal should not exceed 3.3%. And for capacities higher than 300 MW sulfur contaent of coal should be below 2.8 % (for 2000 kcal/kg) and 2.5 % (for 1500 kcal/kg). At lower capacities upto 50 MW the difference between dry, spray and wet methods vary in the range of 4-5 million dollars respectively. For 50 MW, dry system have 4.1 million dollars investment cost, spray dry have 8.9 million dollars and wet system have 13.1 million dollars investment cost. And with the increasing capacity difference increases to 13-14 million dollars. For 300 MW wet system’s investment is 41.5 million dollars, spray’s is 27 miilion dollars and dry’s is 14.1 million dollars (Çift, 2008). In Figure 2 annual operating costs of those three methods are shown comparatively in a graph for trona used as a sorbent. Operating costs are nearly the same upto 50 MW. But for capacities higher than 300 MW different become seizable and reaches to 2 million dollars between methods ,wet, spray and dry in the increasing row respectively. While wet method has the lowest value for annual operating costs, for invesment cost this method has the highest value. And for capacity of 600 MW difference between methods reaches to 7-8 million dollars. This means that higher than 300 MW capacities although the investment cost is higher using wet method seems to be profitable in the long term (15 years is the breaking point ). Among the parameters of annual cost (electricity, water and sorbent consumption, operator,maintenance & repairement and waste disposal) sorbent cost have the biggest percentage. For example in dry injection method if trona is used as a sorbent for 100 capacity , sorbent cost forms the 66.5% of the annual costs. And for 200 MW capacity in spray dry method sorbent cost forms the 32% of the annual cost. Since sodium based sorbents are more expensive compared to calcium based sorbents (limestone is the cheapest, 20 $/ton) obtaining trona (as a sodium based sorbent) from national reserve, as cheap as limestone, can bring a great cost advantage. Because sodium based sorbents are more active than calciun based sorbents and therefore achieve higher desulfurization rates than calcium based sorbents (Doğu, 1984).
5
Annual Operating Cost (million $)
Figure 2. Comparison of Annual Costs of three different FGD Methods.
80 70
Dry Method
60 50
Spray Dry Method
40 30
Wet Method
20 10 0 0
100
200
300
400
500
600
700
Capacity (MW)
In Figure 3, trona based dry injection method’s capital investment model is shown as a function of both thermal power station capacity and coal’s sulfur content. Auto 2 Fit 3.0 programme is used and Quasi-Newton method is chosen in forming the capital investment model, TCI, which is given below (78 data set is used that are obtained from the prepared computer programme) ; TCI = 321 917* (MW)0.61* (S%)0.3
(R2 =.977)
(8)
Figure 3. Three dimensional Capital Investment graph of dry injection method when trona is used as a sorbent.
6
In Figure 4 trona based spray dry method’s three dimensional capital investment model is shown as a function of both capacity (MW) and coal’s sulfur content (S%). In spray dry method capital investment reaches to 2 times the dry injection method. The capital investment model is given below (87 data set is used); TCI = 631 513* (MW)0.64 * (S%)0.197
(R2=0.9997)
($)
(9)
As can be seen from the Formula capital investment varies with capacitiy’s exponential factor of 0.64 and coal’s sulfur content’s exponential factor of 0.197. Figure 4. Three dimensional Capital Investment graph of spray dry method when trona is used as a sorbent.
And in Figure 5 trona based wet system’s three dimensional model is shown (79 data set is used). Mathematical model is given below (for 2000 kcal/kg coal); TCI = 909 759* (MW)0.63 * (S%)0.26
($)
(R2=0.9992)
(10)
Figure 5. Three dimensional Capital Investment graph of wet method when trona is used as a sorbent.
7
4. Conclusions : Lignite as an ore, Turkey’s national energy source, it is going to be one of the major energy source of Turkey in the future. Since Turkish lignite has low calorific value and high sulfur content, lignite based thermal power plants should install FGD units which brings extra cost to the electricity production. Therefore it is important to install FGD units in a more cheaper way. For this aim Beypazarı trona reserves can be evaluated in dry sorbent injection method (because of it’s low capital investment and ease of installing). By this way cheaper way of installing FGD units can be achieved. For the comparison of capital investment costs of different FGD methods and different sorbents a computer programme is prepared in Delphi 7.0 programming language. The rate of capital investment of dry injection, spray dry and wet method is found as 1: 2.1: 3.2 respectively. Also annual operating costs of these methods when trona is used as a sorbent are determined as 28, 32 and 36 million dollars respectively for 300 MW and 2000 kcal/kg coal of 2% sulfur content. The model function of capital investment for dry injection is found to be as follows; 321 917* (MW)0.61* (S%)0.3. Also dry injection is the most suitable method for retrofit application to the existing plants which don’t have FGD units. But in dry injection required desulfurisation rate can not be achieved if calcium sorbents are used (75 %). In order to overcome this disadvantage Beypazarı trona reserve can be evaluated. As known limestone is the most common and the cheapest sorbent used in desulfurisation processes. If trona can be obtained from Beypazarı trona reserve as the same cost with limestone it is going to be advantageous to use trona. Since it is more active than calcium based limestone in desulfurisation. Beypazarı trona reserve can be evaluated for this aim.
5. References Barbour,W., Oommen, R., Shareef, G.S., Vatavuk, W.M., (2002). Wet Scrubbers for Acid Gas, EPA Air Pollution Control Cost Manual, Sixth Edition. Çift, B.D., (2008). Economic and Technical Evaluation Analysis of the Desulfurization Processes at Lignite Using Power Stations, Doctorate Thesis, İstanbul Technical University, İstanbul, (Turkish). Doğu, T., (1984). Effect of Pore Structure on the Mechanism of SO2 Sorption on Activated Soda in Frontiers in Chemical Reaction Engineering, Eds. Mashelkar, R.A.,John Wiley Publ.Co. Ekinci, E., Tırıs, M., Okutan, H., (1994). Thermal Power Plants and Emission Control, Gökova Gulf Environmental Problems and Environment Management Symposium, 2830 June (Turkish). EPA, (1985). Flue Gas Desulfurization Inspection and Performance Evalution Manual, PEI Associates. Inc., Ohio,USA. Heinsohn, R.J., Kabel, R.L.,(1999). Sources and Control of Air Pollution, Prentice Hall Inc., Upper Saddle River,New Jersey. Özsoy, H., Köksal,S., Öztürk,A., Tütünlü,F., (1998). Advanced Technologies for Increasing Efficiency in Energy Production and Environment Protection, Energy Policy Working Group Report, TÜBİTAK-TTGV Science Tecnology-Industry Conversations Platform, Ankara (Turkish).
8
Sedman, C.B., Sparks, L.E., (1986). Mini Spray Dryers for Low Cost SO2 Control, Transfer and Utilization of Particulate Control Technology Symposium, New Orleans, Indiana, USA. Soud, H.N., (1994). FGD Installations on Coal-Fired Plants, IEA Coal Research IEACR/71, June, London. Ulrich, G.D., (1984). A Guide To Chemical Engineering Process Design and Economics, John Wiley&Sons, New York. Turton, R., Bailie, R.C., Whiting, W.B., Shaewitz, J.A., (2003). Analysis, Synthesis and Design of Chemical Processes, Prentice Hall International Series, New Jersey. Vatan, B., (2006). Evaluation of Applications of Reorganisation in Energy Sector, Turkey 10th Energy Congress, p. 311-315 (Turkish).
9
Desulfurization Characteristics of Powdered Hydrated Lime for Flue Gas Sorbent Injection Process
Hyok Bo Kwon Professor, Department of Environmental Engineering, Kyungnam University, 449 Masan Korea, +82 55 249 2249, +82 55 249 2244, [email protected]
Sang Whan Park Principal Research Scientist, Material Science and Technology Division, KIST, Seoul, Korea, +82 2 958 5472, +82 2 958 5409, [email protected], Hyung Taek Kim Professor, Division of Energy Systems Studies, Ajou University, Suwon, Korea, +82 31 219 2321, +82 31 219 2692, [email protected]
ABSTRACT In this study, the effect of property change of hydrated lime by methanol treatment and adsorption temperature change on desulfurization efficiency in flue gas sorbent injection process was determined. Desulfurization efficiency of methanol treated hydrated lime increased and it was affected by BET surface area change. In addition, reaction temperature did not affect desulfurization efficiency at normal operating temperature range of 70-150℃. Desulfurization efficiency increased significantly when water was added to sorbent. Desulfurization efficiency increased as moisture content of untreated sorbent increased; however, it decreased as moisture content of methanol treated sorbent increased. In conclusion, desulfurization efficiency increases by adding water and moisture content should be modulated by sorbent property for optimal desulfurization efficiency.
1. INTRODUCTION Coal cleaning and flue gas desulfurization are technologies that remove SO2 from flue gases of power plants when burning coals. Flue gas desulfurization is by far the most effective method use today; however, more cost effective desulfurization technology can be achieved by combining those two methods. In flue gas desulfurization system, SOx is removed from flue gas by absorption, oxidation, reduction, and adsorption. The wet desulfurization, using lime or limestone, is the most common process today among approximately 130 methods developed. However, the chief drawback of wet desulfurization process is that it requires a large installation area when applied to existing coal-fired power plants. Flue gas sorbent injection technology or in-duct desulfurization process has been widely studied (1-3) because of its low capital and operating costs, and ease of applying to existing facility . In this technology, slurry or dry powder of hydrated lime(Ca(OH)2) or limestone are used as the sorbent, and then they are sprayed into the duct downstream of the air heater to remove calcium
sulfate in dry form. This process is of great interest in regards to power plant retrofitting due to the short construction time and the small amount of capital required. Therefore, it offers advantages for the cases of limited installation area for wet desulfurization process or outworn plant facility. However, it occurs high sorbent cost because of its lower desulfurization efficiency of sorbent injection than existing wet desulfurization technology. Accordingly, it can be economically attractive if the problem of low efficiency of sorbent is solved. In this study, the effect of change in sorbent property and operating condition, such as temperature and humidity, on desulfurization efficiency was examined in flue gas desulfurization process where hydrated lime was sprayed into the duct downstream.
2. EXPERIMENT 2-1. METIREAL The hydrated lime was used as a reagent and its chemical composition is shown in Table 1. A methanol which was used to change sorbent property was absolute alcohol of 99.9% purity.
2-2. PROPERTY CHANGE OF SORBENT A methanol was used to change sorbent property and treatment procedure is as follows. We poured 10 g of (Ca(OH)2) and 50mL of methanol into 100mL of round-bottom flask and then mixed them with magnetic bar for 3 hours at 25℃, 40℃ and 55℃ under nitrogen atmosphere. A mixture was cooled under nitrogen atmosphere, dried in vacuum oven, and sealed. Methanol treated samples went through 150 mesh. Adequate amount of water was added by using micro pipet to demonstrate the effect of moisture content on desulfurization.
2-3. EXPERIMENTAL APPARATUS AND PROCEDURE The experimental apparatus for desulfurization efficiency used in this study consists of gas regulator, humidity controller, reactor, and peripheral device, etc. and its schematic is shown in Fig. 1. 1000ppm of SO2 (N2-balanced) at flow rate of 2.0mL/min was injected into the reactor. In order to maintain 60% of relative humidity inside reactor, nitrogen gas for dilution purpose was saturated by passing through 500mL of Pyrex gas washing bottle. The temperature of gas washing bottle for maintaining humidity was controlled by heating tape. The reactor was made of stainless steel and consists of inner tube and outer tube. Two stainless steel Quick flanges, each of them are 80mm long and 10mm in internal diameter, were connected together for inner tube and outer tube is 90mm long and 350mm in internal diameter. The temperature of inner tube was controlled by heating tape. PP tube and stainless steel Swage lock were used to connect apparatus and temperature was measured by K type thermocouple. The quartz wool of 100mm x 150mm which 120mg of sorbent was evenly sprayed on was rolled and then inserted into the inner tube. After reaction, the sorbent was washed off with acetone and dried. We measured sulfur content of sample by sulfur analyzer, calculate S/Ca mole ratio, and then compared desulfurization efficiency.
3. RESULT AND DISCUSSION 3-1. EFFECT OF CONTACT TIME Table 2 shows that the effect of contact time of hydrated lime with SO2 on desulfurization efficiency. We determined that overall desulfurization efficiency is considerably low and the effect of increasing contact time is also insignificant. However, we set up 30 minutes to provide enough contact time for (4-6) sorbent with gas for the rest of this experiment studies .
3-2. EFFECT OF TEMPERATURE Table 3 shows that the effect of reaction temperature on the desulfurization efficiency. This result indicates that the desulfurization efficiency slightly increases as temperature increases. Therefore, it can be concluded that temperature has no significant impact on desulfurization efficiency within this experiment temperature range and set the reaction temperature at 95℃ for the rest of this experiment.
3-3. EFFECT OF PROPERTY CHANGE Table 4 shows that the effect of BET surface area change of methanol treated sorbent on adsorption 2 efficiency. BET surface area of methanol treated sorbent (18-41 m /g) tends to increase considerably 2 compare to untreated sorbent (16 m /g). BET surface area increases as methanol treatment temperature increases but it tends to decrease at high temperature. SEM photo of sorbent shown in Fig.2 also describes changes by methanol treatment. As methanol treatment temperature increases, 5-10㎛ of untreated sorbent were chopped into small pieces that have 0.2㎛ thickness and 1㎛ long. However, as shown in Fig.2-b, BET surface area of methanol treated sorbent at 25℃ when sorbent begins to be chopped is bigger than untreated sorbent. BET surface area of methanol treated sorbent at 55℃ when surface area is well developed (Fig.2-d) is smaller than methanol treated sorbent at 40℃. The sample which methanol was applied on for 3 hours at each temperature was reacted for 30 minutes at 95℃ and its desulfurization efficiency change tends to be the same as BET surface area change (Table 4). However, reaction temperature did not affect the sorbent that is treated for 3 hours at 55℃ as shown in Table 5. Therefore, from this result and the effect of reaction temperature on untreated sample (Table 3), it can be concluded that reaction temperature did not affect desulfurization efficiency at normal operating temperature range of 50-150℃ in flue gas sorbent injection technology (Fig.3).
3-4. EFFECT OF MOISTURE CONTENT In flue gas sorbent injection technology, moisture control of inside the pipe is known as an extremely 1-4) important factor. To determine the effect of moisture on desulfurization efficiency, adequate amount of distilled water was evenly added to the sorbent by using micro pipet, and then desulfurization experiment was conducted for 30 minutes at reaction temperature of 70℃. Desulfurization efficiency increased considerably when water was added. Desulfurization efficiency increased as more amount of water was added, and it was more than six times greater than the average efficiency of 0.92% when water was not added (Table 6). Table 7 also shows that desulfurization efficiency increased considerably when water was added to methanol treated sorbent and reacted for 30 minutes at 70℃. It reached 7.66% which is average eight times greater than average efficiency of untreated sorbent (0.92%). This result indicates that desulfurization efficiency is average 1.35 times greater than when 1.5wt % of water was added to the untreated sorbent (5.64%). However, as shown in Fig.4, desulfurization efficiency tended to decrease as methanol treatment temperature increased. A methanol treated sorbent that was reacted for 3 hours at 40℃ showed the highest desulfurization efficiency. As shown in Table 8, this sorbent showed higher desulfurization efficiency when water was added gradually and reacted for 30 minutes at 70℃. Desulfurization efficiency of 13.54% was gained by adding 0.5wt % of water, which is 15 times greater than the average desulfurization efficiency of untreated sorbent (0.92%). Desulfurization efficiency decreased significantly as moisture content increased; however, it still remained high and
corresponded with the desulfurization efficiency when same amount of water (1.5wt %) was added shown in Table 7. However, Fig.5 indicates that the effect of moisture content on methanol treated sorbent is opposite to the effect of moisture content on untreated sorbent (Table 6). In other words, desulfurization efficiency increases as moisture content of untreated sorbent increases; however, desulfurization efficiency increase as moisture content of methanol treated sorbent decreases. In conclusion, desulfurization efficiency increases by adding water and it is greater when methanol treated sorbent was used. Furthermore, moisture content should be modulated by sorbent property.
4. CONCLUSION The effect of change in sorbent property and operating conditions, such as temperature, moisture content, on desulfurization efficiency in flue gas sorbent injection process is as follows. First, BET surface area change affected desulfurization efficiency of methanol treated sorbent. Second, reaction temperature change did not affect desulfurization efficiency at normal operating temperature range of 70-150℃. Third, desulfurization efficiency increased significantly as moisture content of sorbent increased. Forth, desulfurization efficiency increased as moisture content of untreated sorbent increased; however, it decreased as moisture content of methanol treated sorbent increased. Fifth, desulfurization efficiency increased by adding water and it was greater when methanol treated sorbent was used. Thus moisture content should be modulated by sorbent property.
5. REFERENCES 1. Srivastava RK, Jozewicz W. Flue gas desulfurization: the state of the art. J Air Waste Manag Assoc 2001;51:1676-88. 2. Srivastava RK, Jozewicz W, Singer C. SO2 scrubbing technologies: a review. Environ Prog 2001;20:219-27. 3. Hu GL, Dam-Johasen K, Wedel S. Hansen JP. Review of the direct sulfation reaction of limestone. Prog Energy Comb Sci 2006;32:386-407. 4. Withum JA, Yoon H. Treatment of hydrated lime with methanol for in-duct desulfurization sorbent improvement. Environ Sci Technol1989;23:821-7. 5. Kim H, Kim D, Kim H, Kwon H. Microstructural investigation of calcined and sulfated limestone for the utilization in the AFBC environment. J Korean Surf Eng1993;26:255-62. 6. Kim H, Kwon H. Kinetic study of limestone calcination and sulfation reaction under AFBC environment, Environ Eng Res1998;3:105-113.
Table 1. Hydrate analysis Ash (wt %)
75.10
Moisture (wt %)
0.41 Ash analysis (wt %)
SiO2
1.27
Al2O3
0.35
CaO
94.52
Fe2O3
0.02
K2O
0.19
MgO
2.86
Na2O
0.11
TiO
0.02
SO3
0.04
wet screen
wt %
+150
0.6
200 × 325
1.9
-325
97.5
Table 2. Effect of contact time on sorbent activity. Reaction temperature(℃)
93
96
96
94
96
Contact time (min)
5
10
20
30
60
0.65
0.88
0.87
0.93
0.78
Ca utilization (%)
Table 3. Effect of reaction temperature on sorbent activity. Reaction temperature(℃) Ca utilization (%)
70
95
115
135
150
0.85
0.93
0.78
0.97
1.05
Table 4. Effect of methanol treatment temperature on sorbent activity. untreatment
25
40
55
surface area (m /g)
16.74
18.18
41.47
28.43
Ca utilization (%)
0.92
1.23
1.83
1.70
Treatment temperature(℃) 2
Table 5. Effect of reaction temperature on methanol treated sorbent activity. Reaction temperature(℃) Ca utilization (%)
70
95
115
135
150
1.85
1.70
1.63
1.86
1.72
Table 6. Effect of moisture contents on sorbent activity. Added H2O (wt %)
0
0.45
0.58
1.00
1.52
Ca utilization (%)
0.92
1.85
3.13
5.64
5.64
Table 7. Effect of methanol treatment temperature on water added sorbent activity. Added H2O (wt %)
0
Treatment temperature (℃) untreated Ca utilization (%)
0.92
1.05
1.05
1.05
1.05
1.05
25
40
40
55
55
8.74
7.18
8.58
6.65
7.16
Table 8. Effect of moisture contents on methanol treated sorbent activity. Added H2O (wt %)
0
Treatment temperature (℃) untreated Ca utilization (%)
0.92
0.50
1.00
1.50
1.50
40
40
40
40
13.54
8.74
8.58
7.18
Fig 1. Schematic diagram of experimental apparatus.
a)
c)
b)
d)
Fig 2. SEM image of hydrated lime. a) untreated hydrate b) methanol treated hydrate(3hrs mixing at 25℃) c) methanol treated hydrate(3hrs mixing at 40℃) d) methanol treated hydrate(3hrs mixing at 55℃).
3
Raw sample
Ca utilization (%)
2.5
Methanol treated sample
2
1.5
1
0.5
0 0
50
100 temperature (℃)
150
Fig 3. Effect of reaction temperature on sorbent activity.
200
10
Ca utilization (%)
7.5
5
2.5
0 0
20
40
60
temperature (℃) Fig 4. Effect of methanol treatment temperature on water added sorbent activity. 1.5wt % water added to the sorbent.
Ca utilization (%)
15
10
5
0 0
0.5
1
1.5
Added H2O (wt %)
Raw sample
Methanol treated sample
Fig 5. Effect of water addition on sorbent activity.
STATUS OF LARGE CIRCULATING FLUIDIZED BED BOILER OPERATION IN CHINA Jie Yu, Senior Engineer, Beijing Research Institute of Coal Chemistry, China Coal Research Institute Contact Information: NO.5 Qingniangou Road, Hepingjie, Beijing 100013, P.R.China E-mail: [email protected] Abstract: circulating fluidized bed (CFB) boiler technology has been rapidly developed in China due to its fuel flexibility including low rank coal and low cost emission control. In the past twenty years, the CFB technology was developed by the local institute and companies. In the sane time, it was also introduced from foreign famous companies.
Now, China has the largest
installed number and capacity of CFB boiler in the world. The total number of CFB boilers in China is over 3000, of which capacity ranges from 6MWe to 330 MWe. The fuel in the CFB boilers includes anthracite, lean coal, bituminous, lignite, gangue, petroleum coke, biomass, municipal solid waste, and many kinds of industrial waste. The total installed capacity of CFB boiler in China is 66GWe, sharing about 15% of the total thermal power generation, among which, the number of 135MWe CFB boilers is over 200, while that of running 300MWe CFB boilers is 26, and there are other more than 60 units of 300MWe CFB ordered. Now, the 600MWe supercritical CFB boiler demonstration is being constructed. The operational performance of large CFB boilers including 135MWe and 300MWe was investigated in the present paper. And the reliability, economy, emission are also discussed as well as the problems in these units. Key word: China, large circulating fluidized bed (CFB) boiler, operation performance, status
0
Introduction
Circulating fluidized bed boiler has been paid more attention on because of its excellent fuel flexibility and low cost emission control (Lu Q, et al, 2006). Forced by energy and environment, the CFB was used widely in China. The technology of CFB boiler made in China was not only from the local institute or companies, but also introduced from foreign famous companies. The total number of CFB boilers in China is more than 3000, which is over one third of the total thermal power plant in unit number. The unit capacity ranges from 6MWe to 330 MWe. The fuel fired in these CFB boilers covers anthracite, lean coal, bituminous, lignite, gangue, petroleum coke, biomass, municipal solid waste, and many kinds of industrial waste. The total installed capacity of CFB boiler in China is 66GWe in 2009 sharing about 15% of the total thermal power generation ( Li J, et al., 2009b). In these CFB boilers, the number of 135MWe CFB boilers is more than 200; most of them were put into operation. And the number of running 300MWe CFB boilers is 26 and 17 of them are manufactured with introduced technology from Alstom, 1 of them is made with Xi’an Thermal Power Research Institute, 8 of them are developed by Dongfang Boiler Co. Ltd supported technically by Tsinghua University. There are other more than 60 units of 300MWe CFB ordered by the end of 2009 ( Li J, et al., 2009a). Now, the 600MWe supercritical CFB boiler demonstration is under construction. The CFB boiler plays an important role for clean coal power generation in China. However, the CFB boiler operation performance is not as good as expected due to the shortage of experience because the age of CFB combustion technology is younger than pulverized coal (PC) fired boiler. This paper investigates the operational performance of large CFB boilers including 135MWe and 300MWe in China. And the reliability, economy, emission are also discussed as well as the problems in these units
1
Operation of 135MWe CFB boiler in China
1.1 Reliability The economy depends on the unit reliability strongly. In this paper, two indexes are used to descript the boilers reliability. One is unplanned outage times caused by various accidents and the other is longest continuous running period. Figure 1(a) shows the unplanned outage times of 135MWe CFB boilers in recent years. The unplanned outage time decreases significantly year by year because the boiler design was improved. Unfortunately, the times of unplanned outage for some boilers is still over 9 in 2008, see Fig.1 (b).
16
6 5
Unit number
Unplanned outage times
7
4 3 2
2008
12 8 4
1 0
0 2004
2005
2006 2007 Year (a) Fig.1
2008
0
1
2 3 4 5 6 7 Unplanned outage times (b)
8
9
Unplanned outage times of 135MWe CFB boiler
When CFB boilers appeared in China twenty years ago, it was called “week boiler” because it frequently forced to shutdown by various accidents, such as the leakage of heating surface erosion, defluidization (Mao J, et al., 2006). With the improvement in design, manufacture, construction, commissioning, operation and management, the continuous running period became longer and longer, shown in Figure 2. To some CFB power stations, the longest running time is near 400 days.
1.2
Economical
Economic index includes carbon content in fly and bottom ash, ignition oil, electric power consumption. Figure 3 shows that the carbon content (LOI) in fly ash for 135MWe CFB boilers decreases in the recent years. 15 Carbon content in fly ash %
Longest continuous running period, days
400 300 200 100 0 2004 Fig. 2
2005
2006 Year
2007
2008
Longest continuous running period of 135MWe CFB Boiler
12 9 6 3 0
2004
Fig. 3
2005
2006 Year
2007
2008
Carbon content in fly ash of 135MWe CFB boiler
Most of the 135MWe CFB boilers mentioned in this paper were put into operation in 2004. During the commissioning and initial operation, the carbon content in fly ash of these CFB boilers is pretty higher. And after two years operating experiences, the average LOI decreased to less than 7%. This is not satisfied compared with the oversea CFB boilers. It was believed to be caused by the fuel property (Sahu R, et al., 1988; Xiao X, et al., 2005). The fuel in the CFB boilers in China is generally hard coal or even coal stone with lower volatile matter content and high ahs content. The net heating value of the coal is lower than 3000 kcal/kg. For such fuel, it is difficult to be fired in PC boiler. For some CFB boiler fried lignite, the carbon content in fly ash is near 0 (Wang Y, et al. 2003), which confirmed the CFB boiler design. This indicates that the combustion efficiency of CFB boiler is not always excellent but higher reactivity coal. The carbon content in bottom ash is often less than 3%, which is very stable. In the CFB boiler, there is a large amount of material in the bed with high temperature, which supplies the heat for coal particle ignition. Thus the low loading without oil of the 135MWe CFB boilers can reach 35% of B-MCR. However, during the startup, the oil consumption is larger than that in PC boiler due to the material. With the increasing of the experiences (Zhang Q, et al., 2005; Meng L, et al., 2007), the oil consumption for startup decreased year by year, shown in Figure 4. The material is fluidized by the primary air in a CFB boiler, which means the primary air pressure is higher. And the back
pressure of the furnace at the secondary air nozzle is high, which leads to the secondary air pressure should also be high (Lu J, et al., 2003; Yue G, et al., 2009). Thus the fan motor power is high. And if the bottom ash cooler is fluidized, or there is external fluidized bed heat exchanger, the high pressure fan is necessary (Wang W, et al., 2009). So the electric power consumption rate of CFB boiler is high, generally 9% for condensed steam unit and 10% for cogeneration of heat and power unit, see Fig.5. 10.5 Electric power consumption %
Oil consumption t/year
500 400 300 200 100 0
10.0 9.5 9.0 8.5
2004
Fig. 4
1.3
Cogeneration of heat and power Condensed steam unit
2005
2006 Year
2007
2004
2008 Fig. 5
Oil consumption for startup of 135MWe CFB boilers
2005
2006 Year
2007
2008
Electric power consumption rate of 135MWe CFB Boiler
Emission
3
SO2 and NOx emission, mg/Nm
Desulphurization efficiency %
One of the outstanding advantages of a CFB boiler is desulfurization with limestone during combustion. The desulfurization efficiency of 135MWe CFB boilers in China is higher than 85%, and many of them are even higher than 95%, see Fig.6, it can be seen that SO2 emission of CFB boilers meets the national emission standard of 400mg/Nm3 from Fig. 7. 100 400 80 60 40 20 0 2004 Fig. 6
2005 2006 Year
2007
300 200 100
2008
Desulphurization efficiency of 135MWe CFB boilers
SO2 emission NOx emission
0 2004
Fig. 7
2005
2006 Year
2007
2008
SO2 and NOx emission of 135MWe CFB Boiler
From Fig. 7, it can be seen that NOx emission of these CFB boilers is far below the Chinese national standard of 500mg/Nm3. This is contributed by the low temperature combustion, air staging and high content of unburntout carbon in the material to reduce the NOX.
2. Operation of 135MWe CFB boiler in China The number of running 300MWe CFB boilers in China is 26. 17 of 26 are manufactured with Alstom technology which is very well known, 1 is developed by Harbin Boiler workers Co. Ltd and Xi’an Thermal Power Research Institute, 8 are developed by Dongfang Boiler Co. Ltd supported technically by Tsinghua University. There are other more than 60 units of 300MWe CFB ordered by the end of 2009; most of them are domestic technology.
2.1 Reliability The reliability of 300MWe CFB boilers is relatively low. For example, the average unplanned outage times is 5.625 times in 2008, far exceeds that of 300MWe PC boilers (which is 0.89 time in2008). The continuous operational period of 300MWe CFB
boilers in 2007 were only 154 days, not only far lower than that of 300MWe PC boilers, but also lower than that of 135 MWe CFB boilers. The reason for the poor reliability of 300MWe CFB boilers was due to the lack of the operational experience of large capacity CFB boiler. With the accumulation of operational experience and better understanding of large CFB boilers, the continuous operational period of 300MWe CFB boilers increase. In 2009, it becomes, for some 300MWe CFB boilers over 200 days. Especially for some CFB boilers with domestic technology, because the process and equipment is simplified, the reliability is much better. In Baolihua Power Plant, the continuous operational period is more than 300 days in 2009.
2.2
Economical
The net coal consumption of 300 MWe CFB boilers is 353.86gec/kWh in 2009, which is lower than that of 135 MWe CFB boilers but higher than that of 300 MWe PC boilers. This is because the difference of coal combustion technology. In CFB boiler, the coal property is relative lower, and the electric power consumption in CFB is higher than that in PC. In 2007, the electric power consumption in subcritical 300MWe PC with flue gas desulphurization is about 5.67%. However, it is about 9~11 for CFB boiler. Thus to decrease the electric power consumption in CFB is an important issue. Fortunately, Tsinghua University suggests a simplified process for 300MWe CFB. And the demonstration developed by Dongfang Boiler Co. Ltd and located in Longyan, Fujian Provence has been put into operation in end of 2009. In this 300MWe demonstration plant, the electric power consumption is about 4.5%. This achievement encourages the deep investigation and optimization of 300 MWe CFB. Same as the 135MWe CFB boilers, the coal in the 300MWe CFB is also low rank coal. Due to the coal consumption in 300MWe CFB is double, the coal is better in net heating value. Because the volatile matter content in the coal is also higher, the carbon content in fly ash is lower. Because there are massive inert bed materials in the bed, the oil consumption for startup in 300MWe CFB boilers is larger. Thus, it is of great significance to study the heating by neighborhood boiler, under bed ignition, etc to save oil (Li J, et al., 2009b).
2.3
Emission
Emission of 300MWe CFB boilers is very low. The SO2 and NOX emission is far lower than Chinese national standards. The desulfurization efficiency with limestone of all the 300MWe CFB boilers is over 90%, higher than that of 135MWe CFB boilers when the Ca/S is about 1.7 to 2.1, leading to the SO2 emission in the exhaust flue gas is less than 350mg/Nm3 but Baima Power Plant of which the sulfur content in the coal is more than 3.5%. The NOX emission ranges from 50 to 150 mg/Nm3.
3
large CFB development in China
3.1 300MWe CFB with local technology The demonstration of the first 300MWe CFB boiler developed with local technology supported by Tsinghua University is in Baolihua Power Plant. It is natural circulation with single drum. There is no fluidized bed heat exchanger in the boiler. In the front part of upper furnace, there are twelve panels of wing wall superheater and six pieces of wing wall reheater. There are eight coal feeders in front wall of the furnace. The primary air box located under air distributor is made of membrane water wall expanded by rear water wall of the furnace. Secondary air is introduced into the furnace on the taper area in 2 layers from front wall and rear wall. Four bottom ash discharge openings are arranged in the rear wall above air distributor, corresponding to four roller ash cooler. Three cyclone separators are located between furnace and secondary pass. The auto-balance loopseal, which has one input and two outputs to improve the return of circulating material, locates under the cyclones, as shown in Fig. 8. The reheat steam temperature is controlled by the flow rate of flue gas. The low-temperature reheater is located in the front pass of the parallel passes, while the final superheater and primary superheater are located in the rear pass. The economizer is arranged in the combination pass above the horizontal tube air preheater (Guo, Q, et al., 2009). Deeside the 300MWe CFB boiler developed by Dongfang Boiler mentioned above, Harbin boiler workers Co. ltd also developed their own 300 MWe CFB. One is technologically supported by Xi’an Thermal Power Institute. The other is own developed referenced Alstom experiences.
And Shanghai developed their own 300 MWe CFB similar with that of Dongfang.
3.2
600MWe supercritical CFB development
A once-through 600MWe supercritical CFB boiler is being developed for Baima Power Plant. It is designed according to the characteristic of the local fuel. The rated steam output of this boiler is 1900 t/h, with the main steam of 25.5 MPa in pressure and 571 oC in temperature, and the reheat steam 569oC in temperature. Once-through vertical membrane water walls are arranged in the furnace. Twin furnaces are combined to ensure the secondary air jet penetration depth. Shown in Fig. 9, six cyclones as well as six loopseals, six external heat exchangers (EHEs) under them are installed beside the furnace, three sets on each side. The secondary superheater and high temperature reheater are immersed in the EHEs, while the final superheater is the wing wall type hanging in the upper furnace. The fuel gas out of cyclones is collected, and introduced from two sides into the secondary pass. The primary superheater and the low temperature reheater are installed in line in the heat recover pass, which is packed by steam cooled membrane wall. Below it, the economizer and the air preheater are installed (Li Y, et al., 2009). Now, this boiler is under construction. It will be put into operation in the nest year. And more large CFB development is under consideration (Yue G, et al., 2009).
4
Summary
The circulating fluidized bed combustion develops rapidly in China due to its advantage of fuel flexibility and low cost for emission control. But its net power generation efficiency is not as good as that of PC for the 135MWe CFB and 300MWe CFB, especially the reliability is not satisfied. Thus to increasing the reliability and efficiency is very important. The achievement of simplified CFB process is very encouraged. In the same time, the supercritical CFB is another issuer. The demonstration of 600MWe supercritical CFB boiler will be put into operation in the next year in China.
Fig. 9
Schematic diagram of the 600MWe supercritical CFB boiler
1-Steam-water separator 4-Cyclone separator 7-Primary reheater 10. Loopseal. 13-EHE (SH II_1) 16–Hot gas generator
2-Water vessel 5–Gas pass 8-Economizer 11-Bottom ash cooler 14-EHE (SH II_2)
3-Coal bunker 6-Primary superheater 9-Coal feeder 12-EHE (RH II) 15-Secondary air
Reference: [1] Guo Qiang, Zheng X, Zhou Qi, et al.. Operation experience and performance of the first 300MWe CFB boiler developed by DBC in China. In: Yue G, Zhang H, Zhao C, et al. eds. Proceedings of the 20th International Conference on Fluidized Bed Combustion. Xi’an, Beijing: Springer Press. 2009: 237~242 [2] Li Jianfeng, Mi Jianhua, Hao Jihong, et al. Operational status of 300MWe CFB boiler in China. In: Yue G, Zhang H, Zhao C, et al. eds. Proceedings of the 20th International Conference on Fluidized Bed Combustion. Xi’an, Beijing: Springer Press. 2009a: 243~246 [3] Li Jianfeng, Yang Shi, Hao Jihong, et al. Operational status of 135MWe CFB boilers in China. In: Yue G, Zhang H, Zhao C, et al. eds. Proceedings of the 20th International Conference on Fluidized Bed Combustion. Xi’an, Beijing: Springer Press. 2009b: 219~223. [4] Li Yan, Nie Li, Hu Xiukui, et al. Structure and performance of a 600MWe supercritical cfb boiler with water cooled panels. In: Yue G, Zhang H, Zhao C, et al. eds. Proceedings of the 20th International Conference on Fluidized Bed Combustion. Xi’an, Beijing: Springer Press. 2009: 132~136 [5] Lu Junfu, Feng Junkai. On Design of CFBC Boilers. In: Xuchang Xu ed. Proceeding of the 5th International Symposium on
Coal Combustion, Nanjing China, 2003: 309~313 [6] Lu Qinggang, Song Guoliang, Sun Yunkai, et al. Independent research and development of 600MWe supercritical circulating fluidized bed boiler technology. In: Yue G, Hao J eds. Proceedings of the 1st CCCCFB & the 6th annual meeting of CFBCEC, 2006: 56~67. [7] Mao Jianxiong. Availability of circulating fluidized bed boilers. Boiler Technology, 2006, 37(3): 24~28 [8] Meng Luowei, Xu Zhengquan. Measures to reduce auxiliary power consumption rate for large-scale CFB boilers. Electric Power, 2007, 40(9): 65~68 [9] Sahu R, Levndis Y, Flagan R C, et al. Physical Properties and Oxidation Rates of Chars from Three Bituminous Coals. Fuel, 1988, 67: 275~283. [10] Wang Wei, Li Jinjing, Yang Shi, et al. Experimental study on heat transfer in a rolling ash cooler used in tee CFB boiler. In: Yue G, Zhang H, Zhao C, et al. eds. Proceedings of the 20th International Conference on Fluidized Bed Combustion. Xi’an, Beijing: Springer Press. 2009: 1147~1151 [11] Wang Yu, Li Weiqi, Lu Junfu, et al. Modeling Research on Char Particle Combustion Behavior in CFB Condition. In: Xuchang Xu ed. Proceeding of the 5th International Symposium on Coal Combustion, Nanjing China, 2003: 120~124. [12] Xiao Xianbin, Yang Hairui, Zhang Hai, et al. Research on carbon content in fly ash from circulating fluidized bed boilers. Energy and Fuels.2005,19 (4): 1520~1525 [13] Yue Gangxi, Yang Hairui, Lu Junfu, et al. Latest development of cfb boilers in China. In: Yue G, Zhang H, Zhao C, et al. eds. Proceedings of the 20th International Conference on Fluidized Bed Combustion. Xi’an, Beijing: Springer Press. 2009: 3~12 [14] Zhang Quansheng, Tian Weidong, Wang Baojun, et al. Discussion on the energy saving technique of CFB boiler units. Electrical Equipment, 2005, 6: 77~79.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11–14, 2010
Performance of a 500 kWth Pressurized Entrained-Flow Coal Gasifier Kevin J. Whitty, Randy Pummill, David R. Wagner, Travis Waind and David A. Wagner Institute for Clean and Secure Energy The University of Utah Salt Lake City, Utah, USA
ABSTRACT In order to acquire industrially-representative data on performance of coal gasification systems, the University of Utah has commissioned a small pilot-scale, pressurized, oxygen-blown entrained flow coal gasifier. This system has a maximum throughput of approximately 1.5 tons coal/day (approx. 500 kWth) at the maximum design pressure of 31 bar. Operationally, the system performs well, achieving good conversion and producing hydrogen- and carbon monoxide-rich syngas. Operation at very low pressures (< 5 bar) is challenging, primarily because the system was designed for operation at 10-30 bar. Initial results from experimental campaigns suggest that a residence time on the order of 5-6 seconds is necessary to achieve good conversion.
INTRODUCTION Gasification of coal is an attractive technology for both power production (through an IGCC configuration) and chemicals production (through catalytic processing of the synthesis gas product). Interest and investment in coal gasification technology has increased dramatically in recent years, particularly in developing regions such as China. A majority of the new and planned coal gasification systems are high temperature, pressurized, oxygen-blown entrained-flow systems offered by the likes of GE, ConocoPhillips, Shell, Siemens and Uhde. But despite rapid growth and billions of dollars being spent on new gasification-based facilities, there is still very little public information available regarding real-world performance of this type of system, and detailed characterization of the environment in industrial gasifiers is essentially non-existent. Development of a database for gasification performance versus operating conditions has been identified as a key need for development of the technology [1]. But such information is unlikely to come from industrial gasifiers and few small-scale entrained-flow coal gasification facilities exist in the world today [2-5]. As part of a larger overall effort to improve the understanding of and develop simulation tools for high temperature coal gasification, the University of Utah recently completed construction of a small pilot scale pressurized, oxygen blown, downwards-fired entrained flow coal gasifier. The gasifier is being used for study of system performance with different fuels, configurations and operating conditions. It will also provide an opportunity to acquire validation data from a well-characterized system for the gasifier simulation tools being developed at the University of Utah. The University of Utah’s coal gasifier has undergone several weeks of commissioning, with each campaign successively increasing throughput and operating pressure. This paper describes experience to date with the system, including operating experience and performance data.
Page 1
EXPERIMENTAL The University of Utah’s pressurized entrained-flow gasifier was originally designed and built in 2005 under a project funded by the U.S. Department of Energy’s Office of Biomass Programs (OBP) for processing black liquor, a byproduct of the papermaking process. However, OBP changed priorities as construction was nearing completion and the project did not continue to the operation phase. In 2008, the gasifier received new life under a research program funded by the U.S. Department of Energy’s National Energy Technology Laboratory (NETL). Several modifications were necessary to make the system operable on coal. These modifications and remaining construction were completed in 2009 and the system was first started up in December that year. Entrained Flow Gasifier System An overall schematic of the entrained-flow gasification system is presented in Figure 1. Typical and maximum operating conditions are given in Table 1. The gasification system comprises several subsystems for feed, reaction, quench and syngas and slag handling. These are detailed below.
Figure 1. Schematic of the entrained-flow gasification system.
Coal Feed System. Currently the gasifier is configured to feed coal as a slurry. Future efforts may target development of a dry feed system. Slurry is prepared by mixing coal and water in a large (~ 1 m3, 270 U.S. gallon) tank equipped with an agitator and a recirculation pump. Fluxing agent (e.g., limestone), surfactants and emulsifiers may also be added. Because of the small geometries in the injector, the coal is ground quite fine (60-120 micron) to minimize the risk of plugging the injector. It should be possible to feed larger particles, but currently operability is the primary objective. It is possible to achieve slurry concentrations of 60% by weight without chemical additivies. The slurry tank is water cooled to minimize heating and associated water evaporation resulting from input of mechanical energy to keep the slurry well mixed.
Page 2
Table 1. Typical and maximum operating conditions for the entrained-flow gasification system. Parameter Operating temperature (°C) System pressure (bara) Coal slurry feed rate (liter/hr) Coal feed rate (kg/hr dry basis) Slurry concentration (wt%) Oxygen feed rate (kg/hr) Syngas production rate (m3/hr dry)
Typical
Maximum
1425 18 70 42 59 50 80
1700 31 150 80 65 150 150
From the discharge side of the recirculation pump, a tee connects to a 3.5 kW progressive cavity pump which is the main feed pump for the gasification system. The pump is capable of feeding over 125 liter/hr of slurry, which corresponds to roughly 75 kg/hr of dry coal, at 31 bar pressure. The slurry flow rate is measured with an ultrasonic flow meter, check valve and then into the injector. The coal injector itself is a twin-fluid design, with coal slurry flowing through the center channel and high pressure oxygen flowing through an annulus concentrically around the slurry. The oxygen channel is angled inwards at the exit to impart a shear force on the slurry, thereby atomizing the fuel. The injector sits inside a water-cooled shroud. Details of the injector design and associated performance are provided elsewhere [6]. Oxygen Feed System. Oxygen comes from a 20 ton liquid oxygen tank at the facility. The oxygen is pressurized to 23 bar, which limits the normal operating pressure to about 18 bar. If higher pressures are desired, an external high pressure oxygen supply (e.g., a high pressure oxygen trailer) can be coupled to the system. The oxygen flows through a regulator, flow control valve, Coriolis flow meter and into the inejector. Gasification Reactor. The reactor itself is constructed of refractory within a 0.75-m pressure vessel. The internal dimensions of the reactor are 1.6 m length by 0.21 m inside diameter. This is much different than commercial downflow entrained-flow gasifiers, which typically have a height-to-diameter ratio of 2 to 3. The high height-to-diameter ratio of the Utah gasifier “stretches out” the reaction zone, allowing sampling and characterization of the environment as the coal is being converted. The reactor has six pairs of opposing sample ports spaced every 0.20 meters down the upper 70% of the reactor. (Flanges in the transition section make it impossible to have sampling ports in the lower 30%.) Five type B thermocouples are spaced every 0.25 down the length of the main section of the reactor. Two additional ports near the burner can be fitted with high pressure windows for visualization of the flame at low pressures. Quench System. The synthesis gas and slag products from the gasification reactor are cooled and separated in a 2-stage quench system (Figure 3). The lower section of the refractory that forms the reactor rests on a 0.07-m tall cooled ring with an inner diameter approximately 0.1 m larger than the inner diameter of the refractory. A 1.6-m long, 0.36-m diameter “dip tube” hangs from this ring. Immediately below the ring, water is injected into the product stream through four
Figure 2. Gasifier.
Page 3
flat spray nozzles. This direct cooling comprises the primary quench. The degree of cooling can be adjusted by changing the flow rate of water spray. A water bath in the bottom of the quench system comprises the secondary quench. The gas flows downwards through the dip tube, then bubbles through the secondary quench and flows back up between the dip tube and the outer vessel of the gasifier and out of the vessel. Slag and large suspended solids remain in the quench bath. The inventory of the bath is large enough that it can hold slag from an entire day’s operation. The water in the quench bath circulates through a high pressure centrifugal pump, through one of a pair of parallel bag filters, through a cooler and back to the quench bath. This keeps the water bath cool and allows collection of entrained char. A radar-type level sensor on a vertical pipe offset from the quench bath monitors the water level in the quench bath. Periodically, a drain valve on the recirculation line opens to allow quench water to flow to a flash tank in order to maintain the target quench water level. Syngas Handling System. After exiting the gasifier, the syngas flows through a pressure control valve and into the facility’s thermal Figure 3. Quench system. oxidizer, where it is combusted. The combustion gases are cooled and pass through a scurbber, ID fan and out through the stack. The syngas gas can be routed through a hot gas filter between the gasifier and pressure control valve, but this is generally bypassed due to operational problems related to condensation in the filter, since the syngas becomes saturated with water in the quench system. System Operation The gasification system is designed for operation at elevated pressure, 10 to 30 bar, and operating at low pressures creates challenges. Target feed rates of fuel and oxygen are based on an estimated 5 to 6 second residence time in the reactor, so the capacity of the system increases proportionally with pressure. Because heat loss through the reactor refractory is relatively constant, regardless of pressure, the fraction of fuel energy lost as heat through the walls is comparatively high at low pressure when fuel feed rates are low. Estimated heat loss based on measured shell temperatures is 10 kW, which accounts for roughly one-fourth of energy input at 2 bara pressure, but is less than 2% of energy input at 25 bara pressure. The consequence of this is that, for similar residence times and oxygen/fuel ratios, the system operates several hundred degrees cooler at very low pressure. In practice, this is overcome by operating at a higher oxygen/fuel ratio at low pressure. Guidelines for operation on a typical bituminous coal (Illinois #6) are indicated in Table 2. These assume an overall stoichiometry of 0.6 (corresponding to an O2/fuel ratio of 1.3 for this particular coal), a 6 second residence time and a slurry concentration of 60 wt%, with a resulting average reactor temperature of 1450°C.
Page 4
Table 2. Operation guidelines for Illinois #6 coal. Assumes average reactor temperature of 1450°C, 1.3 kg O2/kg coal, 60% slurry and 6 sec. residence time. Pressure (bara)
Coal feed rate (kg/hr)
Slurry flow rate (liter/hr)
Oxygen feed rate (kg/hr)
Syngas flow (dry, Nm3/hr)
Thermal input (kWth)
1
2.2
3.6
2.7
3.7
17
2
4.4
6.9
5.5
7.3
34
5
11
17
14
18
84
10
22
35
27
37
168
15
33
52
41
55
253
20
44
69
55
73
337
25
55
87
68
92
421
30
66
104
82
110
505
Startup of the gasifier involves preheating the system with a 70 kW natural gas burner for 48 to 72 hours to ensure that the refractory has achieved a steady state temperature and the reactor thermocouples indicate a temperature of at least 1250°C. The natural gas burner is then exchanged for the coal injector. Rather than igniting using the coal slurry, alcohol is initially fed into the reactor. After introduction of alcohol is confirmed, oxygen flow is carefully initiated and increased to the target rate (still substoichiometric). The highly volatile alcohol ignites from the heat of the refractory and the two flame detectors indicate a good, stable flame. The system is heated to approximately 1350°C while feeding alcohol and the system is pressurized to 4 bara pressure. During pressurization, the feed rates of alcohol and oxygen are slowly increased. Once stable, pressurized operation on alcohol has been achieved (20-30 min), the feed is swapped from alcohol to coal slurry using a 3-way valve. The system smoothly transitions to coal gasification, which is observed initially by a decrease in reactor temperature and later by a change in syngas composition. Once the system is operating on coal, the pressure is increased to the target pressure while slurry and oxygen flows are increased. Shutdown of the system is approximately the reverse of the startup procedure. The system is depressurized to approximately 4-5 bar while flow rates are decreased. Once at that pressure, coal slurry is swapped back to alcohol. Nitrogen is then fed through the oxygen line. Once the flame is observed to go out, the flow of alcohol is stopped. The system continues to be purged with nitrogen as it is depressurized. If additional tests are planned, the burner is exchanged for the natural gas standby burner and the system is kept hot. Before igniting the natural gas burner, the quench bath is drained to remove slag that has accumulated. For final shutdown, the coal injector is removed and the system is allowed to cool naturally while a small flow of nitrogen is allowed to flow through the reactor.
RESULTS Results from the first series of commissioning tests are described below. Overall System Performance In general, the gasifier operates very stably once it has reached target conditions. There has never been any problem with plugging of the coal injector, which was a significant concern as the system was being designed. Plugging has also never been encountered with the slurry preparation and recirculation system, the quench recirculation system or syngas lines. The slurry and oxygen flows are stable, which results in stable operation of the reactor. Figure 4 shows a representative data log of system temperature and syngas composition during the first hour of a typical test.
Page 5
Figure 4. Reactor temperatures (left) and concentrations of major syngas species (right) during the first hour after the gasifier reached its target pressure for a typical run. Thermocouples T1 through T5 are positioned in order from top to bottom in the reactor. The variation in reactor temperature on thermocouple T1 at time 20:20 is associated with adjustment of a purge flow on the flame detectors. Syngas compositions are given on a normalized nitrogen-free basis.
The one issue that has been presented challenges during operation is pressure control. Generally the pressure is able to be controlled to within 0.5 bar of the target pressure, but feedback is slow and the valve actuation has not been as smooth as desired. This is an issue that is under review and is expected to be improved over time. Influence of Oxygen/Coal Feed Ratio One of the primary “knobs” used to adjust conditions in the gasifier is the oxygen/coal feed ratio. Increasing this ratio results in higher temperatures and lower syngas heating value, since additional oxygen consumes hydrogen and carbon monoxide that would otherwise be useful contributors to the syngas. Optimally, one would like to operate a gasifier at as low of an O2/coal ratio as possible while still maintaining good conversion. Doing so not only maximizes energy value in the syngas, but also helps reduce degradation rates of the gasifier wall and injector. A campaign was conducted in which the gasifier was operated at relatively low pressure (4.5 bara) and the feed rate of coal slurry was held constant while the oxygen flow rate was varied over quite a wide range. Each set of conditions was held for approximately 15 minutes, which arguably is not sufficient time to reach full steady state but is sufficient to the overall influence of the O2/fuel ratio on system performance. Temperatures and syngas compositions were recorded at each set of conditions. These are displayed in Figure 5. Only two temperatures are shown, those of thermocouple T2, which is the hottest in the system and near the flame, and T4, which is approximately 60% down the length of the reactor. (The other 3 thermocouples showed similar trends.) Syngas composition was measured by GC and is shown on a dry, nitrogen-free basis. The calculated heating value is also shown in the figure. As one would expect, increasing the amount of oxygen fed to the gasifier results in higher reactor temperatures as exothermic combustion reactions occur to a higher degree. Correspondingly, the CO2 concentration increases and the concentrations of syngas components H2 and CO decrease. This results in an overall decrease in gas heating value. Additional testing in subsequent campaigns will involve maintaining each set of conditions for longer so that the system can come to full steady state. Carbon conversion will also be assessed by measuring the amount of char captured in the quench system, as it is expected that the coal was not fully converted at the low O2/fuel ratios.
Page 6
Figure 5. Temperatures (left) and concentrations of major gas species (right) as a function of oxygen/coal ratio.
Influence of Pressure/Residence Time A campaign was performed during which the flow rate of coal slurry and oxygen were kept constant while the overall pressure of the system was increased from 4.3 bar to 7.8 bar (absolute). At 4.3 bar, the estimated residence time was 3.4 seconds. As the pressure increased, so did residence time so that it was roughly 6.2 seconds at 7.8 bar. It took approximately 10 minutes for the pressure to increase. The resulting temperatures (at positions T2 near the flame and T4 approximately 60% of the distance through the reactor) and concentrations of H2, CO and CO2 (dry, nitrogen-free basis) are shown in Figure 6 for the 40 minute period after the pressure increase was initiated.
Figure 6. Variation in temperatures at thermocouples T2 and T4, as well as resulting H2, CO and CO2 compositions (dry basis) as the pressure was increased from 4.3 to 7.8 bar while keeping coal and oxygen feed rates constant. Instability in the temperature was partly due to challenges holding pressure constant as described above.
Page 7
Clearly, gasification efficiency increased significantly as pressure increased. The carbon dioxide concentration decreased while hydrogen and carbon monoxide concentrations increased notably. This is much more an effect of the increased residence time than of pressure. Although the system was not configured to allow direct measurement of carbon conversion in these tests, it is reasonable to assume that with the longer residence time, the fuel was more efficiently converted. The increase in temperature also is indicative of better conversion.
CONCLUSIONS The University of Utah's recently commissioned pressurized entrained-flow gasifier operates well and is able to produce good quality synthesis gas during extended operation. Overall, few technical challenges have been encountered with the system, and those that do remain (e.g., pressure control) are not insurmountable. Reactor temperatures and syngas properties vary with the O2/fuel feed ratio and pressure/residence time as one would expect. Future testing will involve continued assessment of how operating conditions (fuel feed rate, O2/fuel ratio, pressure) affect system performance. Each set of conditions will be run for enough time for the system to achieve steady state and mass and carbon balances over the system will be calculated by including measurements of char and slag capture in the quench system.
ACKNOWLEDGEMENT This material is based upon work supported by the Department of Energy under Award Number FC2608NT0005015.
REFERENCES 1. Clayton, S.J., Stiegel, G.J., Wimer, J.G., "Gasification Technologies: Gasification Markets and Technologies — Present and Future. An Industry Perspective," U.S. DOE Report DOE/FE-0447 (2002). 2. Choi, Y-C., Park, T-J., Kim, J-H., Lee, J-G., Hong, J-C., Kim, Y-G., "Experimental Studies of 1 Ton/Day Slurry Feed Type Oxygen Blown, Entrained Flow Gasifier," Korean J. Chem. Eng. 18(4):493-498 (2001). 3. Hughes, R., "Entrained Flow Gasifier R&D at CanmetENERGY," Proc. 2010 Workshop on Multiphase Flow Science, 4-6 May 2010, Morgantown, VA (2010). 4. Cousins, A., McCalden, D.J., Hughes, R.W., Lu, D.Y., Anthony, E.J., "Entrained Flow Gasifier Fuel Blending Studies at Pilot Scale," Can. J. Chem. Eng. 86:335-346 (2008). 5. Wang, Z., Yang, J., Li, Z., Xiang, Y., "Syngas Composition Study," Frontiers of Energy and Power Eng. in China 3(3):369-372 ( 6. Waind, T., Whitty, K.J., "Characterization of a Small Scale Slurry-Fed, Oxygen-Blown Entrained Flow Gasifier: How Injector Geometry Affects Flame Stability and Performance," Proc. 2010 Int'l Pittsburgh Coal Conf., 11-14 October 2010, Istanbul, Turkey (2010).
Page 8
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 CHARACTERIZATION OF A SMALL SCALE SLURRY-FED, OXYGEN-BLOWN ENTRAINED FLOW GASIFIER: HOW INJECTOR GEOMETRY AFFECTS FLAME STABILITY AND PERFORMANCE Travis Waind, Graduate Research Assistant The University of Utah 50 S. Central Campus Dr., MEB 3290, Salt Lake City, UT 84112 [email protected] Kevin Whitty, Associate Professor The University of Utah 50 S. Central Campus Dr., MEB 3290, Salt Lake City, UT 84112
Abstract: The University of Utah operates a 1 ton/day pressurized, slurry-fed, oxygen-blown entrained-flow coal gasifier. To determine the ideal design settings for the coal slurry injector in this system, a number of factors were evaluated. The injector is a twin-fluid design with coal slurry flowing through the central channel and oxygen entering at high velocity through a concentric, angled annulus around the slurry, providing atomization. The injector is positioned in a water-cooled sleeve to keep the injector cool. The tip of the injector can be removed from the injector and replaced. The angle that the oxygen tube is tapered can be adjusted to change the pattern of the injector discharge. The tip of the injector can also be adjusted so that the oxygen flow path is small or large, which affects the pressure drop across the nozzle. A balance needs to be stricken between the pressure drop across the nozzle (atomizing velocity) and the angle of the injector discharge. The geometries of the injector design should be optimized to provide a thorough devolatilization of the coal-water slurry. Through cold-flow testing, a higher pressure drop was shown to provide better atomization. By increasing the impingement angle, the water droplets in the plume appeared to be smaller than those of a smaller impingement angle. The smaller impingement angle provided the smallest water spray angle, giving it a larger velocity component in the downward direction. Changing the gas and liquid flow rates for a given impingement angle and pressure drop did not appear to affect atomization. During tests on the entrained flow gasifier, a higher impingement angle caused a slight increase in gasification products. Also, higher pressure drops led to an increase in gasification products as well as a slight increase in temperature.
INTRODUCTION Injector performance is important for gasification. An efficient injector design will provide the proper atomization of a coal-water slurry mixture so that the fuel can completely devolatilize into the gasification products. Improper injector design can lead to unreactant fuel and a loss in efficiency. An efficient injector will also ensure stable flame operation for the system (Pohl, et al., 1985). The University of Utah operates a 1 ton/day pressurized, slurry-fed, oxygen-blown entrained-flow coal gasifier. During poorlycontrolled operation, the temperature profile of an EFG can often be the opposite of what would be expected. If the flame from the injector is pushed too far into the reactor, a myriad of problems can occur. Downtime and repairs can be expected from problems such as excessive refractory erosion, erratic syngas composition, overheating of the EFG body, or others. If the flame is pushed too far into the reactor, the cause can most likely be identified as the injector, assuming that feed flow rates are correct. The injector geometries might be such that the oxygen stream does not interact enough with the coal-water slurry stream to provide proper atomization at a distance that is not down inside the reactor. In order to determine which factors affect the position of the flame inside the EFG, several tests were performed involving injector settings, and hot and cold operation of the injector. The goal in operating the gasifier is to acquire a thorough characterization of this particular EFG. By achieving some degree of repeatability, error will be able to be quantified in such a way to develop a working numerical model of the gasification process. The injector testing provides an important facet to this model. Because oxygen and coal-water slurry enter the EFG through the injector, the injector becomes a boundary condition defining the behavior of these two substances as they enter the EFG. THEORY Atomization is the process of a liquid stream being broken up into small droplets. For entrained-flow gasification coal injectors, a gas stream (typically oxygen or steam) is used to break up the liquid stream. Different mechanisms exist that cause atomization, particularly in injector nozzles. With one type of nozzle, the gas stream has no normal direction relative to the liquid stream. This type of nozzle lends itself to a wavy-sheet formation mechanism of atomization (Lefebvre, 1992). The liquid is sent in the same direction as the gas. As the liquid leaves the injector nozzle, it has time to develop waves. As these waves propagate, they eventually come in contact with the parallel airstream. This contact assists the already developed waves in breaking the liquid stream into smaller parts, drops (Lefebvre, 1992). The relative velocities of the liquid and gas streams play an important role in performance of this type of atomizer. The nozzle used for testing directs the gas stream into the liquid stream as they exit the nozzle’s tip. The gas stream velocity is orders of magnitude faster than the liquid stream. This velocity difference provides the energy needed for a thorough atomization of the liquid stream. However, as the droplets move away from the injector, they will begin to coalesce (Hedley, and Yiu, 1985). This type of atomization is known as gas-assist atomization because the high-velocity gas stream provides the energy to break apart the coal-water slurry (Lefebvre, 1992). Gas-assist atomization is enhanced in this injector due to design. Because the space that the gas stream flows through is decreased as it impinges on the liquid stream and exits the nozzle, the velocity of the gas stream is greatly increased by the Venturi effect. The increased velocity gives the stream a higher kinetic energy. This increase in kinetic energy provides better atomization. However, the increased velocity is at the expense of a pressure drop due to Bernoulli’s principle. That fact is of little concern, as the fluids have reached their destination. In fact, this pressure drop is exploited to achieve the velocity increase and atomization. EXPERIMENTAL Two systems were used to evaluate performance of the injector in this study: a cold flow system and the actual gasifier. Injector Design The injector is a twin-fluid design. It allows coal-water slurry to flow down the inner tube, and the oxygen to flow down the outer tube. The injector neck is surrounded by a water jacket that keeps the injector from damage at the higher operating temperatures of the EFG. An air or nitrogen purge is provided for the small gap between the water jacket and the slurry/oxygen tubes to prevent soot build-up and assist heat removal. A schematic of the injector is shown in Figure 1.
Figure 1: Above is a rough schematic of the coal-water slurry injector used for EFG testing at the University of Utah. The diagram on the left shows an overall view of the injector, while the diagram on the right shows some detail of inside the injector. The injector has a removable nozzle that directs the oxygen flow into the coal slurry stream. Three nozzles have been constructed, with impingement angles of 25°, 45°, and 65°. The angle of impingement is measured from the vertical. Each nozzle is denoted as an “angle”-nozzle representative of its impingement angle. The 25° nozzle has not been used in testing on the EFG at this point. A profile of the 45° nozzle is shown in Figure 2. The inner tube shown in Figure 2 is the coal slurry tube. The distance between the coal slurry tube and the nozzle is referred to as the gap width. By changing that distance, the oxygen or air pressure drop across the nozzle is changed by forcing the gas through a smaller or larger annulus.
Figure 2: Schematic of the 45° nozzle that was used for testing. The angle is measured with the vertical as a reference.
Cold-Flow Testing A system was constructed for atmospheric pressure cold-flow testing of the injector. Scaffolding was constructed to support the injector in a position far enough off the ground to allow visibility of the water spray and allow adjustments of the injector gap width. House air was connected to the injector through pressure regulator, a rotameter, and a pressure gauge. Water was also connected to the injector through a pressure regulator, a rotameter, and a pressure gauge. A schematic of the cold-flow setup used for testing the injector on water and air is shown in Figure 3.
Figure 3: Above is the setup of the cold-flow testing that was done on the injector. The injector was suspended on scaffolding so the water spray could be seen.
To analyze how the injector was pushing the flame into the bottom of the EFG, a pressurized test apparatus was created to evaluate injector settings. The body of the injector test apparatus was constructed of clear polyvinyl chloride (pvc) with the idea that the plume from the injector could easily be seen through the clear pvc. Unfortunately, upon commencing trials on this apparatus, the visibility was found to be inadequate. A visualization of the spray profile could not be achieved as the atomized water very quickly clouds the sides of the vessel. This setback led to atmosphere, cold-flow evaluation of the injector. By suspending the injector and water jacket in the air, the plume from the injector was readily visible and photographed. Water and air lines were run to the injector as well as the appropriate regulating devices in order to quantify the feed flow rates and pressures. The water and air flow rates and pressures were set to the desired values while simultaneously adjusting the gap width within the nozzle to achieve the desired pressure drop. Once the pressure drop and feed flow rates were set, photographs were taken to document the shape and profile of the water plume. This process was repeated with various flow rates, pressure drops, and impingement angles. Entrained-Flow Gasifier Reactor The pressurized entrained-flow gasifier EFG is approximately 1.5 m (5 ft) tall and encompasses two sections: the reaction zone and the quench bath. The actual reaction zone inside the EFG is approximately 20 cm (8 in) in diameter. There are five thermocouples positioned in the reaction zone that provide a temperature profile and give an idea of the flame position. A cooling ring serves as a buffer between the reaction zone and the quench section. The reaction zone can approach temperatures around 1500°C (2700°F) at the refractory surface. Details of the EFG are given elsewhere (Whitty, et al., 2010). The injector was evaluated during EFG operation over a multiple day period. For the first day of testing, the 45° nozzle was used in the injector. The EFG system pressure for this run was kept at 4.5 bar absolute pressure or bara (65 psia). Once the EFG had reached a steady-state on the coal slurry, the oxygen pressure drop across the injector was set to 2.1 bar (30 psi). The system was allowed time to move towards equilibrium before the pressure drop was changed again. This was done for pressure drops of 2.1 bar (30 psi), 1.4 bar (20 psi), and approximately 1.0 bar (15 psi). The 65° nozzle was used in the injector on the second day of testing. The EFG system pressure was kept at 4.5 bara (65 psia), similar to the first day’s run. Once the EFG had reached a steady-state on the coal slurry, the pressure drop was set to 2.1 bar (30 psi). The system was allowed time to move towards equilibrium before the pressure drop was changed again. The pressure drop set points that were evaluated on the first day were repeated on the second day.
RESULTS Cold-flow Testing Cold-flow analysis was performed on the injector in order to evaluate the spray profile of the various nozzles at various flow rates and pressure drops. The three factors that were evaluated during the cold-flow analysis were (1) the impingement angle, (2) the flow rates of oxygen and water, and (3) the pressure drop across the nozzle. Influence of impingement angle. Three nozzles were available for analysis. The angles of the three nozzles were 25°, 45°, and 65°. A schematic of the 45° nozzle is shown above in Figure 2. For evaluating the effect of the impingement angle, each nozzle was run at a set flow rate (corresponding to the appropriate flow rates into the EFG at 4.5 bara, 65 psia) and pressure drop. For the case shown in Figure 4, the air pressure drop chosen was 1.4 bar (20 psi). The nozzle resulting spray profile was photographed, and the nozzle was changed out. Figure 4 shows a comparison of the three nozzles operating at the same pressure drop and flow rates (air and water). The 45° and 65° nozzles had similar spray patterns, while the 25° nozzle produced a notably narrower spray. The bulk spray from each nozzle is highlighted by the dark blue lines that encapsulate the majority of the spray. Both the 65° nozzle and the 45° nozzle have a somewhat uniform fine mist surrounding the bulk spray as shown by the mist outside of the blue lines. The 25° nozzle produced larger droplets in addition to the fine mist, presumably due to less aggressive atomization due to the shallower impingement angle. The streams of mist coming from the 65° nozzle appear to be fairly controlled. They appear to be symmetric around the bulk spray, and careful inspection of the photographs indicates that the streams of mist have the same relative size water drop on each side of the bulk spray. The streams of mist coming from the 45° nozzle also appear to be symmetric with slightly more diversity in the droplet size. The 25° nozzle seems to randomly scatter water droplets, and the droplets don’t appear to be any constant size.
Figure 4: The injector on the left has the 65° nozzle. The injector in the middle has the 45° nozzle. The injector on the right has the 25° nozzle. Each green square represents an area that is approximately 1.27 cm (0.5 in) by 1.27 cm (0.5 in).
The green grid shown in Figure 4 is in place to gauge the dimensions of the water plume. Each green square represents an area that is approximately 1.27 cm (0.5 in) by 1.27 cm (0.5 in). The width of each nozzle is 2.54 cm (1.0 in). The plume from the 65° nozzle travels approximately 7 cm (2.75 in) to achieve a width of approximately 1.91 cm (0.75 in). This gives the plume from the 65° nozzle an angle of approximately 15.5° from one blue line to the other. The plume from the 45° nozzle achieves a width of approximately 2.41 cm (0.95 in) over the same distance. This gives the plume from the 45° nozzle an angle of approximately 19.6°. The plume from the 25° nozzle achieves a width of approximately 0.76 cm (0.30 in) at the same distance from the nozzle. This gives the plume from the 25° nozzle an angle of approximately 6.2°. It should be noted that the spray angle for coal-water slurry atomization will be larger than that of water atomization (Meyer, and Chigier, 1985). The water spray angles comparisons are useful to determine the change due to impingement angle, but the spray angle will not be the same for coal-water slurry atomization in the EFG (Meyer, and Chigier, 1985). Each nozzle’s spray profile appears to have a “pinch point” or neck just below the exit point of the injector. The smallest neck of the three nozzles is in the spray profile from the 25° nozzle. The next smallest neck is from the 65° nozzle. The largest neck is from the 45° nozzle. This is surprising. One would expect that as the angle of the nozzle increased, the resulting spray profile’s neck would decrease from the higher radial momentum of the gas stream. The 45° nozzle appears to be the median at which the impacting air streams and water stream maintain the proper balance that gives a large neck in the spray profile. The sharp angle of impingement on the 65° nozzle inherently creates the small neck in its spray profile. The neck in the 25° nozzles spray profile is far less pronounced than the other two nozzles because the impingement angle is less normal to the water stream. Influence of flow rates. The gas and liquid flow rates were also evaluated. The injector’s spray profile was evaluated at flow rates corresponding to the appropriate flow rates into the EFG at pressures ranging from 1 bara (15 psia) to 7.9 bara (115 psia). These flow rates are calculated based on some assumptions. The assumptions are that the product gas flow has a residence time of five seconds, the coal-water slurry is 60 weight percent solids, the stoichiometric ratio is 0.6, and that the EFG is operating at 1400°C (2600°F). For evaluating the effect of flow rate, a nozzle and a pressure drop were chosen. In the case shown in Figure 5, the 45° nozzle and a 2.8 bar (40 psi) air pressure drop were chosen. The flow rates were then set. The resulting spray profile was photographed, and the flow rates were changed. Figure 5 shows a comparison of two nozzles operating with different flow rates. Both nozzles are the 45° nozzles. Both nozzles have an air pressure drop of 2.8 bar (40 psi). The flow rates of water and air in the photograph on the left correspond to an EFG system pressure of 1.7 bara (25 psia). The flow rates of the nozzle on the right correspond to an EFG system pressure of 7.9 bara (115 psia). There does not appear to be many differences between the two nozzles. The higher flow rates of the nozzle on the right are apparent in the increased density of water droplets shown in Figure 5. However, the size of the water droplets coming from the left injector appears to be approximately the same as the water droplets coming from the right injector. The gap width between the nozzle and the inner water tube is larger for the injector on the right than that of the injector on the left. Without changing that gap width, a flow rate increase alone would increase the air pressure drop across the nozzle. Therefore, in order to maintain the 2.8 bar (40 psi) air pressure drop, the gap width must be increased to allow more through flow.
Figure 5: Both injectors have the 45° nozzle. The injector on the left has a flow rate corresponding to an EFG system pressure of 1.7 bara (25 psia). The injector on the right has a flow rate corresponding to an EFG system pressure of 7.9 bara (115 psia). Both injectors are operating at a 2.8 bar (40 psi) air pressure drop. The angles of the water plumes from each nozzle in Figure 5 appear to have the same angle. There seems to be no noticeable difference between the spray profiles when the flow rates are increased. Of course, EFG operation would surely change with the increase flow rate as residence time of the resulting syngas would change. This could lead a different syngas composition. Considering the cold-flow testing, the only difference between the two flow rates is the operational change of increasing the gap width of the injector to maintain the same air pressure drop. Influence of pressure drop. The influence of pressure drop across the nozzle was also evaluated. The injector’s spray profile was evaluated at pressure drops of 0.7 bar (10 psi), 1.4 bar (20 psi), 2.1 bar (30 psi), 2.8 bar (40 psi), and 3.5 bar (50 psi). For evaluating each pressure drop, a nozzle was chosen and the water and air flow rates were set (corresponding to the appropriate flow rates into the EFG at 4.5 bara, 65 psia). In the case shown in Figure 6, the 45° nozzle was chosen. In Figure 6 below, a comparison is shown between a 25° nozzle operating at a 0.7 bar (10 psi) drop and the same nozzle operating at a 3.5 bar (50 psi) drop. The resulting spray profile was photographed, and the pressure drop was changed. By increasing the pressure drop across a nozzle, a noticeable change occurs. The spray profile on the left, 0.7 bar (10 psi) drop, of Figure 6 appears to contain much larger water droplets than that of the spray profile on the right, 3.5 bar (50 psi) drop. The spray profile on the right, 3.5 bar (50 psi) drop, appears to be more forceful, or the overall velocity of the spray appears larger. The angles of the plumes appear to be the same for both pressure drops. However, the higher pressure drop, 3.5 bar (50 psi) drop, appears to produce fewer large droplets surrounding the bulk spray.
Figure 6: Both injectors shown have the 65° nozzle. The injector on the left is operating at a 0.7 bar (10 psi) pressure drop, whereas the injector on the right is operating at a 3.5 bar (50 psi) pressure drop. Water and air flows are the same for both cases. As described above, the pressure drop is increased across the nozzle by decreasing the size of the annulus that the air flows through. This decrease in size and corresponding increase in pressure create an increase in velocity at the nozzle’s exit. This velocity increase is the cause of the smaller water droplets coming from the left injector shown in Figure 6. The increased velocity provides a higher kinetic energy that shears the water droplets from the water stream coming from the injector. The large droplets shown on the left of Figure 6 are due to smaller velocities of air, and thus, a smaller shearing kinetic energy. Entrained-Flow Gasifier Reactor Operation In addition to the cold-flow testing, the operation of the EFG was analyzed with various injector settings. The 65° nozzle and 45° nozzle were both used. Also, the pressure drop across the injector was varied from approximately 0.7 bar (10 psi) to 2.1 bar (30 psi) at 0.7 bar (10 psi) intervals. To evaluate the effect the injector settings had, the resulting temperature profile and syngas composition were analyzed. The coal slurry contained 56% pulverized coal by weight and was fed at 20 kg/hr. The oxygen flow rate during testing was approximately 15 kg/hr. The pressure drop with respect to time is shown in Figure 7 for the first day of testing on the EFG. The temperature profiles from the five thermocouples in the reaction zone are shown with respect to time in Figure 8. The syngas composition for each pressure drop is shown in Table 1. Figure 7, Figure 8, and Table 1 all are the result of the 45° nozzle.
Figure 7: Oxygen pressure drop over the injector versus time. The oxygen pressure drop is shown in Figure 7 for the EFG testing with the 45° nozzle. Three distinct pressure drops were targeted: 2.1 bar (30 psi), 1.4 bar (20 psi), and 0.7 bar (10 psi). As seen on the right in Figure 7, an oxygen pressure drop of 0.7 bar (10 psi) was not achievable, so 0.8 bar (12 psi) was targeted instead. The respective oxygen pressure drops were maintained for an average of 35 minutes. This amount of time is not enough to ensure the EFG has reached an operational equilibrium. However, a general trend can be made by looking at the decreasing oxygen pressure drop over time. Figure 8 shows the average temperature profile throughout the reaction zone of the EFG for testing with the 45° nozzle for injector oxygen pressure drops of 2.1 bar (30 psi), 1.4 bar (20 psi), and 0.8 bar (12 psi). The temperature of every thermocouple increased as the pressure drop decreased, with the exception of TC4.
TC1 TC2 TC3
2.1 bar 1.4 bar
TC4
0.8 bar
TC5 1100
1200
1300
1400
1500
1600
Temperature (°C)
Figure 8: Reactor temperature profiles during operation with the 45° nozzle at specific injector oxygen pressure drops. The thermocouples are number in order of their position from the top of the reaction zone, so TC1 is closest to the injector and TC5 is furthest away. The thermocouples are 29 cm (11 in) from the adjacent thermocouple. Figure 8 shows the temperatures generally increasing with the decreasing oxygen pressure drop. The time allowed for each respective pressure drop was not adequate to allow the EFG to completely reach equilibrium. However, the data shows the temperature increasing with the decreasing pressure drop. The flame is thought to be closest to TC2 for all pressure drops shown in Figure 8. It was thought that the flame position might change as the pressure drop was changed, but this was not observed. Table 1 shows the average syngas composition for each targeted pressure drop during EFG testing with the 45° nozzle. The carbon monoxide and hydrogen appear to change proportionally to the pressure drop, whereas the carbon dioxide appears to be inversely proportional. The decrease in pressure drop worsens the degree to which the coal slurry is atomized (see Figure 6), allowing some of the slurry to fall through the EFG without reacting and resulting in an effecting increase in the oxygen/fuel ratio. This would explain the slight increase in CO2 at the expense of CO in Table 1 and the slight increase in temperature shown in Figure 8. Table 1: Syngas composition (mol %, dry, normalized nitrogen-free) from EFG testing on 8/18/10. Pressure Drop, bar (psi) H2 CO 0.8 (12) 25.9% 35.7% 1.4 (20) 27.1% 37.3% 2.1 (30) 27.9% 38.3%
CO2 38.2% 35.4% 33.6%
During the second day of testing, the 65° was not used. Figure 9 shows the pressure drop with respect to time for the second day of testing on the EFG. Figure 10 shows the temperature profiles from the five thermocouples in the reaction zone. Table 2 shows the syngas composition. Figure 9, Figure 10, and Table 2 are all the result of the 65° nozzle.
Figure 9: Oxygen pressure drop over the injector during the second day of testing. The oxygen pressure drop is shown in Figure 9 for the EFG testing with the 65° nozzle. Three distinct pressure drops were targeted: 2.1 bar (30 psi), 1.4 bar (20 psi), and 0.7 bar (10 psi). As seen on the right in Figure 9, an oxygen pressure drop of 0.7 bar (10 psi) was not achievable, so 1.0 bar (15 psi) was targeted instead. The respective oxygen pressure drops were maintained for an average of 36 minutes. This amount of time is not enough to ensure the EFG has reached an operational equilibrium. However, a general trend can be made by looking at the decreasing oxygen pressure drop over time. Figure 10 shows the average temperature profiles throughout the reaction zone of the EFG for testing with the 65° nozzle for injector oxygen pressure drops of 2.1 bar (30 psi), 1.4 bar (20 psi), and 1.1 bar (15 psi). The temperature of every thermocouple increased as the pressure drop decreased.
TC1 TC2 TC3
2.1 bar 1.4 bar
TC4
1.0 bar
TC5 1000
1100
1200
1300
1400
1500
1600
Temperature (°C)
Figure 10: Reactor temperature profiles during operation with the 65° nozzle at specific injector oxygen pressure drops. The thermocouples are number in order of their position from the top of the reaction zone, so TC1 is closest to the injector and TC5 is furthest away. The thermocouples are 29 cm (11 in) from the adjacent thermocouple. As seen in Figure 10, performance versus pressure drop was similar to that observed on day one with the 45° nozzle. TC2 indicated the highest temperature, and temperatures overall increased with decreasing pressure drop. Table 2 shows the average syngas composition for each targeted pressure drop during EFG testing with the 65° nozzle. The carbon monoxide and hydrogen increase with pressure drop while CO2 is highest at the lowest pressure drop. As with the 45° nozzle, the suspicion is that low pressure drops result in poor atomization, allowing more of the fuel to pass through the reactor without reacting, resulting in an effective increase of the oxygen/fuel ratio.
Table 2: Syngas composition (mol %, dry, normalized nitrogen-free) from EFG testing on 8/19/10. Pressure Drop, bar (psi) H2 CO 1.0 (15) 27.8% 39.7% 1.4 (20) 28.6% 40.34% 2.1 (30) 30.0% 41.5%
CO2 32.4% 30.9% 28.3%
Comparing the cold-flow and gasification tests, there are notable similarities, such as the worse atomization that results at low oxygen pressure drops. One must be cautious drawing too many conclusions, as there is a definite physical difference between water and coalwater slurry. The coal-water slurry has significantly different rheological properties, most notably viscosity and surface tension (Meyer, and Chigier, 1986). The coal-water slurry requires much more energy to thoroughly atomize. The spray from coal slurry will have a larger spray angle than that of just water. The drop size from coal slurry atomization will be larger than that of water with the same injector conditions. Considering the fact that coal-water slurry requires much more energy to thoroughly atomize than water and considering that lower pressure drops seem to favor combustion more than higher pressure drops, it is not difficult to deduce that a lower limit exists for efficient pressure drop over an injector nozzle during gasification. Figure 4, shows a comparison of the 65° nozzle, the 45° nozzle, and the 25° nozzle. Theoretically, the 25° nozzle provides a much higher axial velocity. This high velocity could be enough to cause some of the coal-water slurry to rapidly flow through the reaction zone without fully undergoing conversion. Future tests will aim to measure unconverted material in the quench of the reactor. Issues that will be pursued in further testing are quantifying droplet size in the spray as well as modifying the injector design to more easily achieve concentricity between the nozzle and the inner coal slurry tube. By quantifying droplet size, the degree of atomization for each injector setting can be quantified numerically instead of visually. By making the injector tubes more concentric, an easily
repeatable spray pattern is possible. Also, an idea of quench water contents would help in re-affirming the results shown from these preliminary tests. If any unreactant coal is found in the quench water, a lower limit on oxygen pressure drop and impingement angle can be set. A method of sampling the quench bath, or dumping it completely, would have to be established to allow a check for coal while the EFG is idling at atmospheric pressure. CONCLUSIONS By performing cold-flow tests with the coal slurry injector, a good visualization of the resulting atomized water plume was established. This allowed for a better understanding of what is happening inside the EFG during operation, despite a coal slurry spray being wider than a water spray. The change is spray characteristics with changes in injector factors was able to be seen. By making changes to the injector, operation of the EFG can hopefully be improved and optimized. The gasification products, CO and H2, were increased by both a higher impingement angle and a higher oxygen pressure drop across the nozzle. From the cold-flow testing, it was shown that these factors (high impingement angle and high gas pressure drop) both provide a more complete atomization per visual analysis. This suggests that by increasing the impingement angle and oxygen pressure drop, the coal-slurry was able to more thoroughly devolatilize in the EFG, although the water spray and coal slurry spray will behave differently. REFERENCES Hedley, A.B., and S.M. Yiu. “Droplet Size Distribution Changes During the Atomization of a Coal Water Slurry.” Seventh International Symposium on Coal Slurry Fuels Preparation and Utilization. 1985. 377-392. Lefebvre, A.H. "Energy Considerations in Twin-Fluid Atomization." Journal of Engineering for Gas Turbines and Power. 114.1 (1992): 89-96. Meyer, P.L., and N. Chigier. “Photographic and Malvern Analysis of Coal-Water Slurry Atomization.” Seventh International Symposium on Coal Slurry Fuels Preparation and Utilization. 1985. 402-429. Meyer, P.L., and N. Chigier. "The Atomization Process in Coal-Water Slurry Sprays." Eight International Symposium on Coal Slurry Fuels Preparation and Utilization. 1986. 144-161. Pohl, J. H., J. Sepulveda, and L. B. Rothfield. “Correlation of the Spray Characteristics of Coal-Water-Fuels.” Seventh International Symposium on Coal Slurry Fuels Preparation and Utilization. 1985. 357-376. Whitty, Kevin J., Randy Pummill, David R. Wagner, Travis Waind, and David A. Wagner. "Performance of a 500 KWTH Pressurized Entrained-Flow Coal Gasifier."27th Annual International Pittsburgh Coal Conference. Istanbul, Turkey. 2010. ACKNOWLEDGEMENTS This material is based upon work supported by the Department of Energy under Award Number FC26-08NT0005015.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: GASIFICATION ANALYSIS OF FINES PRODUCED FROM NON-SLAGGING COAL GASIFIER AND EVALUATION OF ECONOMIC USAGE Yongseung Yun, Director, Institute for Advanced Engineering [email protected] Seok Woo Chung, Team Leader, Institute for Advanced Engineering [email protected] Na Rang Kim, Senior Researcher, Institute for Advanced Engineering [email protected]
Abstract: Most coal gasifiers apply the slagging method for coal ash under high temperature, which gives an advantage on ash disposal with smaller volume and non-leaching of heavy metal components. Slagging was related to many operational problems that had been occurred in pilot and demonstration-scale coal gasifiers. Slagging causes distinctive problems in gasifier operation such as slag-tap plugging and the accumulation of a sticky fly-slag on the radiant cooler and in syngas passage. If the fly-ash from the gasifier can be used as an economically viable material such as cement filler or construction raw material, the coal gasification on non-slagging mode might be an attractive option which replaces coal slag to fly-ash as a final product. In particular, coals that contain ash components exhibiting a high melting temperature such as many Australian coals can be utilized more widely in gasification. Coals of high ash melting temperature should add a certain fluxing agent to reduce the ash flowing temperature. If the fly ash from the non-slagging gasification can be used as a useful material as much as slag, development of a versatile gasifier that can be applicable to wide range of coals would be possible. For usage of fly-ash for other applications, it should meet a certain criteria. Typically, the combustion fly-ash should contain less than 5 wt% remaining carbon to be useful as cement filler, etc. However, fines from the one-stage entrained-bed coal gasifier of pilot-scale 1 ton/day contain 20-70% remaining carbon. In this case, final fines might be applicable for another low-grade fuel if combustibility and other criteria are satisfactory, or a specialty fuel that does not contain volatiles with high surface area. The pilot-scale coal gasifier which was designed to complete the conversion in one pass through the gasifier normally produced 1-5 wt% of feed coal powder as entrained fines. In terms of carbon conversion, typically more than 97% has reached. Although the internal recycle amount of reacted fines to the gasifier can reach more than 50% in two-stage gasifiers, actual amount of fines and slag remains at the ash amount in feed coal plus a small portion of un-reacted carbon which is below than 1-3 wt% of total carbon. In the study, remaining carbon in fines and their physical characteristics were evaluated for the samples obtained from a non-slagging gasifier with slags from the same coal. In short, in order to compare and verify the applicability of fines from the non-slagging gasifier as a useful by-product, an Indonesian subbitumious coal was gasified in the 1 ton/day scale non-slagging gasifier. Samples of entrained fines from the gasifier were analyzed by SEM, EDX, ICP-OES, particle size analyzer, and TGA. Leaching result of heavy metals on the entrained fines was compared with that from gasification slag on the same coal. And a preliminary evaluation on the possible usage of entrained fines as a useful material is discussed.
DEVELOPMENT OF GAS, POWER AND TAR CO-GENERATION SYSTEM WITH CIRCULATING FLUIDIZED BED TECHNOLOGY Presenting author: Qinhui Wang, Professor, State Key Laboratory of Clean Energy Utilization Institute for Thermal Power Engineering, Zhejiang University, 310027, Hangzhou, China; E-mail: [email protected]; Tel: 0086-571-87952802; FAX: 008657187951616; Co-authors: Mengxiang Fang, Zhongyang Luo, Mingjiang Ni, Kefa Cen State Key Laboratory of Clean Energy Utilization; Institute for Thermal Power Engineering, Zhejiang University, 310027, Hangzhou, China
ABSTRACT A gas, power and tar cogeneration technology, which combines a circulating fluidized bed (CFB) boiler and a fluidized bed gasifier to realize gas, power and tar co-generation in one system, has been developed. A 12 MWe gas, power and tar co-generation demonstration plant fired Huainan bituminous coal has been operated successfully and the demonstration operation results showed that the constructed co-generation system may be operated continuously under the design requirements. A series of test operation results showed that the operation temperature in the gasifier has great influence on the pyrolysis gas composition and tar yield. The tar yield reaches a maximum value at a temperature range from 550 to 600℃, and hydrogen content in the produced gas increases with increasing the operation temperature in the gasifier. Key words: Co-generation, gas, tar, circulating fluidized bed, coal pyrolysis
1. Introduction A gas, power and tar cogeneration technology has been developed by Zhejiang University since 1990’s [1-2]. In this system, a fluidized bed gasifier is combined with a circulating fluidized bed boiler to realize gas, power and tar cogeneration. The bituminous or lignite coal is first pyrolysed and gasified in the fluidized bed gasifier by mixing with the high temperature ash from the CFB combustor. The produced gas from the gasifier is purified and used for industrial purpose or as gas turbine fuel. The tar is collected during purification process and may be processed to extract monocyclic aromatic hydrocarbons (MAHs) and polycyclic aromatic hydrocarbons (PAHs) etc., or to make liquid fuels by hydrorefining. The char residue is fed into the circulating fluidized bed combustor together with the recycle material by the recycle device. The steam generated in the CFB combustor is used for power generation and heat supply. It will increase the utilization value and system efficiency of coal conversion process. Based on the research works on the 1MW pilot plant in the laboratory, a 12 MWe demonstration plant fired the bituminous coal has been constructed and operated since August, 2007. A series of tests have been completed to get the related performance data. In this paper, the design features and main operation results of the 12MW gas, power and tar co-cogeneration demonstration plant will be introduced.
2. Description of the 12 MWe cogeneration demonstration plant
The 12MWe gas power and tar co-generation plant is retrofitted from a 12MWe circulating fluidized bed power plant and the main design parameters are shown in Table 1. Huainan bituminous coal is used as the design coal and the required particle size of the feed coal is 0-8 mm. Table 1. The design parameters Item
Parameter
Rated steam capacity Main steam pressure Main steam temperature Gas yields
75t/h 3.82Mpa 450℃ 3000Nm3/h
Item
Parameter
Feed water temperature Air temperature Coal feed rate Gas calorific-value
150℃ 25℃ 16.5 1.2t/h
The processes of the 12 MWe gas, power and tar co-generation plant are shown in Fig. 1. The system combines a 75 t/h CFB boiler with a fluidized bed gasfier which can produce about 3000Nm3/h gas with medium calorific value and about 1.2 t/h tar. Bituminous coal is fed into the fluidized bed gasifier and mixed with high temperature ash (about 900℃) from the standpipe of CFB boiler cyclone. The fed coal is pyrolysed around 600℃ in the gasifier. The pyrolysis gas is separated firstly in a cyclone and cooled in water quench tower. Part of tar will be captured by the water quench tower and most tar will captured by the tar electrical precipitator. Few tar mist and water in the gas will be captured by the secondary electrical precipitator after further cooling in a water cooler. Part of the pyrolysis gas is recycled to the fluidized bed gasifier as the fluidization medium and the produced gas is supplied to the users. The ash particles mixed with the un-reacted char in the gasifer are introduced into the boiler furnace together with the separated fine particles from the gas cyclone by a loopseal recycle device. The un-reacted char particles are burned in the CFB furnace and part of the high temperature ash separated in the hot flue gas cyclone is introduced into the gasifier as the heat carrier. Steam
User Stack
Drum Cyclone Separator
Heat Exchanger Electrostatic PrecipitatorFan Gas Tank
Boiler Indirect Cooler
Coal Hopper Cyclone Separator
Electrical Tar Precipitator
Quench Tower
Electrical Tar Precipitator
Fan
Gasifier Loop Seal Tar Loop Seal
Coal Hopper
Water Pool Loop Seal
Tar Pool
Circulating Light Water Oil Pool Pool
Fan
Primary Air Burner
Fig. 1
The 12 MW gas, power and tar co-generation demonstration plant
The operation requirements of the co-generation system are: (i)may operates in two modes: with operating the gasifier or without operating the gasifier (i. e., the CFB boiler may continue to operate when the gasifier is stopped for maintenance.) (ii) Good load adjustment whether the gasifier is operated or stopped.
3. Operation of the 12 MWe co-generation demonstration plant 3.1 Typical operation performance The 12 MWe gas, power and tar co-generation demonstration plant was operated since August, 2007. One typical Huainan bituminous coal with high ash melt temperature was used, and its proximate analysis and ultimate analysis are shown in Table 2. The continuous operation showed that the demonstration plant may be operated stably and realize gas, power and tar co-generation from coal in an integrated production system. It mainly showed that: (1) the CFB boiler and the fluidized bed gasifier may be operated well in coordinated, and the high temperature ash including the char particles may be recycled between the boiler and gasifier stably and reliably; (2) the developed coal gas purification and tar recovery system may be operated continuously in stably, and meet the design requirements of the gas and tar capture; (3) the system may be operated steady in different operation conditions: (i) no any coal was fed into the CFB boiler furnace and all fuel of the boiler was the char particles from the gasifier; (ii) the load of the gasfier may be adjusted easily by decreasing the coal feed rate to the gasfier and feeding part of coal into the boiler furnace directly when the boiler operate in full load; (iii) the CFB boiler may be operated well when the gasifier is stopped. Table 2. The coal ultimate and proximate analysis Proximate analysis % Mar Mad Aad
Vad
FCad
Qnet,ar - MJ·kg 1
Ultimate analysis% Cad
Gray-King assay tar yield,%
Had Nad St,ad Oad
Tar,ar
3.87 2.07 30.41 28.78 38.74 21.40 54.72 4.65 0.98 0.25 6.92
9.70
A series of tests have been completed to get the effect of operating parameters such as temperature in April, 2008. Coal, gas, tar, char and ash are sampled during different operation conditions. Tar sampling was carried out according to literature [3]. Gas was led from the gasifier via sampling line to gas bottles put in the cold bath. The organic compounds in the gas were absorbed by solvent in the bottles. The temperature and flow rate of the sampled gas flow through the equipment were measured. The tar yield was obtained after the solvent and water was removed by distillation. Tar and gas compositions and concentrations were measured using GC-MS and gas chromatograph, respectively. The typical operating parameters were briefly shown in Table . Part of coal is fed into boiler due to the limit of the performance of coal feed system of the gasifier in typical full load operation condition. As shown in Table 3, the tar yields is about 10% of coal and is slightly higher than the analysis data of Gray-king assay. The main reason should be that the pyrolysis process in the fluidized bed gasifier is flash pyrolysis. The main composition of the produced gas is hydrogen and methane. The gas may be used as city gas or raw material of chemical synthesis processes. The oxygen and nitrogen content in the gas is
low and it shows that almost no air is leaked to the gasifier and gas treatment system, and almost no flue gas leakage between the gasifier and the boiler furnace. In additional, there exists some difference between the design data and the measurement data because different coal is used during the test. Table 3 Typical operating parameters of the 12MW co-generation demonstration system Item
Unit
With operating the gasifier
Steam capacity Steam Pressure Power generation Coal feed rate of the gasfier Temperature at boiler furnace exit Temperature in the gasifier Gas yields Gas Composition H2 CH4 C2H4 C2H6 CO CO2 N2 O2 H2S Tar
t/h Mpa MW t/h ℃ ℃ Nm3/h
70 3.82 12 10.4 900 ~600 ~1950
% % % % % % % % mg/Nm3 kg/h
27.3 36.9 3.54 6.9 7.1 10.1 6.56 0.11 7215 ~1000
3.2 Effects of the operation temperature in the gasfier 错误!未找到引用源。 shows the effect of the operation temperature in the gasfier on the gas composition. As shown in Fig. 2, the content of H2, CO and C2H4 increase with increasing in the operation temperature in the gasifier, but the increment decreases. H2 should be mainly from the saturated hydrocarbons cracking reaction at lower temperatures process, as well as condensation of aromatic hydrocarbons occurred at high temperature process during the pyrolysis process. With increasing in the operation temperature, condensation of aromatic hydrocarbons is carried out quickly and promotes the growth of the H2. While the temperature rose to a certain extent, the degree of condensation would not be deepen, the H2 growth rate is gradually slow [4]. CO may be mainly from the following two ether components: ether bridges connecting each unit of coal, which is the main sources of CO under low temperature process, and the other is diarylether compounds breaking down under high temperature process [5]. With increasing the operation temperature, diarylether compounds of coal, carbonyl functional groups of tar, as well as oxygen heterocyclic ring and OH-component begin to be cracked and release CO. The higher operation temperature will promote the crack reaction. Therefore, higher temperature, higher CO [6].
Unlike to the H2 and CO, CH4 content increases firstly and then reduced with increasing in the pyrolysis temperature. It reaches a maximum content around 630 ℃. CH4 is usually from the cracking of methyl and methoxyl group at a relatively low temperature and decomposition of methylene connecting various units of coal at high temperature [5, 6]. With an increase in the operation temperature, more cracking reactions will occur gradually, and the degree of reaction to in-depth, which would help to raise the CH4 content. On the other hand, dehydrogenation of tar, in particular, aromatic hydrogenations, gradually increases and more tar begin to be formed into a coke as the pyrolysis temperature is increased. When the pyrolysis temperature is higher than 630 °C, the dehydrogenation of aromatic hydrogenations were dominate and the CH4 concentration is decreased. Tar is the important liquid products of co-generation system. The tar yield is affected greatly by the operation temperature in the gasifier. As shown in , the tar yield increases markedly with increasing the operation temperature at lower temperature and reaches a maximum yield in a temperature range from 550 ℃ to 600℃. As increasing the temperature continuously, the tar yield will be reduced. This is in good agreement with those obtained by the reference 7 and 8. So it is necessary to control the operation temperature in order to produce tar with optimum quality and quantity. During the coal pyrolysis process, only part of coal is decomposed at lower temperature. With increasing the pyrolysis temperature, more fractions of coal will be devolatilized and the coal pyrolysis rate is speeded. So the tar yield gets a marked increment in this stage. When the pyrolysis temperature reached to the certain temperature range, the tar yield will reach a maximum rate. Then the tar yield will reduces with increasing the temperature further. It is mainly due to the following two reasons: (1) With increasing the pyrolysis temperature, more tar will be cracked to generate gas although much tar is released. (2) As shown in Fig. 2, H2 concentration in the pyrolysis atmosphere in the gasifier increases with increasing the operation temperature. The existence of H2 will reduce the polymerization of free radicals and the combination opportunity for free radical to generate tar [9]. 12 45 40
10
H2
Concentration (%)
30
O2 25
N2
20
CH4
15
CO CO2 C2H4
10
Tar yields(%)
35
8
6
4
2
5 0 550
600
650
700
Pyrolysis Temperature(℃)
0 400
450
500
550
600
650
700
Pyrolysis temperature(℃)
Fig. 2 Effect of pyrolysis temperature on gas compositions
Fig. 3
Effect of pyrolysis temperature on tar yields
4. Conclusions A new coal-based gas, power and tar co-generation system had been developed, which combined combustion and pyrolysis process with circulating fluidized bed technology. The operation and test results of the 12 MWe demonstration plant fired Huainan bituminous coal showed the co-generation system may be operated stably and realize gas, power and tar co-generation from coal in an integrated production system. The results of a series of tests showed that the gas composition and tar yield are affected greatly by the operation temperature in the fluidized gasifier. The tar yield reaches a maximum rate at a temperature range from 550 to 600℃, and the hydrogen content in the produced gas increases with increasing in the pyrolysis temperature while CH4 content decrease.
Acknowledgement
The authors gratefully acknowledge the financial support from the National High Technology Research & Development Program of China (No. 2007AA05Z334)
References 1.
Mengxiang Fang, Zhongyang Luo, Xuantian Li, Qinghui Wang, Mingjiang Ni, and Kefa Cen, A multi-product cogeneration system using combined coal gasification and
combustion. Energy, 23(3), (1998) 203-212. 2. Wang Qinhui, Luo Zhongyang, Fang Mengxiang, Ni Mingjiang, Cen Kefa. Design of a 12MW Tri-cogeneration System of Gas, Heat and Power. In: Proceedings of the 16th international conference on fluidized bed combustion. Edited by: Donald W., Geiling, P. E., ASME, May 13-16, Reno, Nevada, Paper No. FBC01-0082 3. Simell Pekka, Stahlberg Pekka, Kurkela Esa, Albrecht Johannes, Deutsch Steven, and Sjostrom Krister, Provisional protocol for the sampling and anlaysis of tar and particulates in the gas from large-scale biomass gasifiers. Version 1998. Biomass and Bioenergy, 18(1) (2000)19-38. 4. Haibin Li, Zhiyuan Yang, Hong Lu, Yang Wang, and Bijiang Zhang, Pyrolysis of coal in a fluidized bed reactor Ⅱ: effects of temperature and residence time in freeboard on gaseous product composition. Journal Of Fuel Chemistry And Technology, 26(4) (1998)339-344. 5. Sada Eizo, Kumazawa Hidehiro, and Kudsy Mahally, Pyrolysis of lignins in molten salt media. Ind. Eng. Chem. Res., 31(1992) 612-616. 6. Xu Wei-Chun and Tomita Akira, Effect of temperature on the flash pyrolysis of various coals. Fuel, 66(5) (1987) 632-636. 7. Tyler Ralph J., Flash pyrolysis of coals. 1. Devolatilization of a Victorian brown coal in a small fluidized-bed reactor. Fuel, 58 (1979) 680-686. 8. Jieguang Wang, Xuesong Lu, Jianzhong Yao, Weigang Lin, and Lijie Cui, Experimental Study of Coal Topping Process in a Downer Reactor. Ind. Eng. Chem. Res., 44 (2005) 463-470. 9. Xuedong Zhu, Zibin Zhu, lihua Tang, and Chengfang Zhang, Fundamental Study on the Pyrolysis of Coals I. Effect of Atmosphere and Temperature on Pyrolysis. Journal Of East China University Of Science And Technology, 24 (1998) 37-41
Thermal Chemical Process Study on Chemical Reaction Network of Jet-fluidized Bed Gasifer Reaction System Xue-Cheng Hou, Jie Feng, Wen-Ying Li*, Xiao-Hui Chen Key Laboratory of Coal Science and Technology for Ministry of Education and Shanxi Province; Taiyuan University of Technology, Taiyuan 030024,China
Abstract To study and optimize gasification characteristics of jetting fluidized bed coal gasifier, we studied the influence law of two gasification parameters, oxygen feed rate into center nozzle and coal feed rate, on gasifier gasification process and gas composition variations using established reactor network model, and analyzed the calculation results. Results show that the gasifier temperature distribution is one of the most critical factors affecting gas compositions for jetting bed coal gasifier, and oxygen feed rate into center nozzle and coal feed rate result in obvious changes of gasifier temperature distribution. Within the calculation range, with the increase of oxygen feed rate into center nozzle, jet region increased nearly twice, temperature increased by 306 K and high-temperature region moved up, carbon conversion efficiency rose from 68% to 97% (jet region temperature had exceeded the ash melting temperature), and in generated gas, CO and H2 content changed obviously. With the increase of coal processing capacity, jet region temperature decreased by 252 K, the gasifier overall temperature decreased, and carbon conversion efficiency reduced from 98% to 74%. Dilute phase region has a great effect on effective gas composition in final products, especially H2 content. Keywords reaction network model; gasification parameters; jetting fluidized bed
Introduction To study the influence process of gasification parameters on gasification characteristics of jetting fluidized bed coal gasification system and to illustrate the scaling relationships[1-2] of jetting fluidized bed coal gasifier in scale-up process, we established the equivalent reactor network model of jetting fluidized bed coal gasifier, and verified the model’s reliability and practicality. Based on this, we used the model to study and analyze the variations of gasification reaction process and gas compositions when gasification parameter changes. The objective is to investigate how the sensitivity parameters (oxygen feed rate into centre nozzle, coal feed rate and temperature etc.) affect the process of coal gasification.
1 Calculation statement *
Corresponding author. Email: [email protected] for W.Y. Li; [email protected] for J. Feng.
As the gasifier temperature distribution is the key factor on gasification characteristics[3-21], in this paper, we investigate the two most effective operational parameters for the gasifier temperature distribution, they are oxygen feed rate into nozzle and coal feed rate, and discuss the effect of these two parameters on jet, dense-phase and dilute-phase regions in the gasifier as well. In the ash-agglomerating pilot plant experiments, the range of oxygen feed rate into nozzle is 19.54~35.07 Nm3·hr-1, however, in the calculative process, it ranges from 18.55 to 37.09 Nm3·hr-1 with other operational parameters unchanged. In the ash-agglomerating pilot plant, the designed feed rate of coal is 2.4 t·d-1, that is, 100 kg·hr-1, but 75.87~123.30 kg·hr-1 in the calculative process, other conditions being unchanged.
2 Model calculation and analysis 2.1 Influence of oxygen feed rate into nozzle 2.1.1 Influence of oxygen feed rate into nozzle on jetting region When the JinCheng (JC) anthracite feed rate is 94.5 kg·hr-1, operational pressure is 0.53 MPa, the air feed rate and steam feed rate of distributor is 107.94 Nm3·hr-1, 97.92 kg·hr-1 respectively in ash-agglomerating pilot plant calculative process, we discussed the variations of jet region under three conditions that oxygen feed rate into nozzle is 18.55 Nm3·hr-1, 30.91 Nm3·hr-1, 37.09 Nm3·hr-1 respectively. The influences of oxygen feed rate into nozzle on jet region are listed in Table 1. Table 1 Oxygen feed rate into nozzle remains unchanged Oxygen feed rate Velocity of into nozzle nozzle 3 -1 /(Nm ·hr ) /(m·s-1)
Lj
Dj/
/(cm)
(cm)
Volume of jetting zone/ (cm3)
Temperature of jetting zone/K
18.55
6
53.64
4.10
708
1453
30.91
10
62.43
5.57
1522
1645
37.09
12
65.90
6.22
1999
1759
With the increase of oxygen feed rate into nozzle, high-temperature jet region volume increases from 708 cm3 to 1999 cm3, and combustion reaction intensifies, and high-temperature jet region temperature rises from 1453 K to 1759 K in the gasifier. The operational temperature of jetting bed coal gasifier should be controlled near the coal ash
softening temperature
[22]
. If the temperature is too high, far exceeding the ash melting
temperature, there will be large pieces of slag block which can break gas flow and result in operational difficulties
[23]
. JC anthracite softening temperature (ash agglomeration
temperature) is 1724 K [24]. As shown in Table 1, the minimum temperature in the jet region is 1453 K and the overall temperature tends to be low, as well as low degree of coal gasification in the gasification process calculated in this paper. Jetting region temperature is near ash melting temperature 1645 K under the normal operation condition. When inlet oxygen feed rate keeps increasing, the jetting region’s temperature is 1759 K, beyond the JC coal’s ash melting temperature, so, this is a non-normal operation. It is well known that in jetting bed gasifier, the ash mainly melts in the high-temperature jetting region; therefore, the jetting region temperature control is the key factor ensuring the normal operation [25]. We will discuss temperature variations, carbon conversion and gas composition changes along axial distance in the gasifier under the three operational conditions. 2.1.2 Influence of oxygen feed rate into nozzle on gasifier temperature distribution As shown in Fig. 1, with increasing oxygen feed rate into nozzle, share of pulverized coal combustion inside gasifier increases, and an overall temperature rises, of which jetting region temperature rises from 1453 K up to 1759 K, and the gas outlet temperature rises from 958 K up to 1180 K. Besides, the maximum temperature point of axial distance rises from 26.82 cm up to 32.95 cm because of the increasing penetration depth [26-28]. But when oxygen feed rate into nozzle increased to 37.09 Nm3·hr-1, the temperature of jetting region had exceeded the JC anthracite’s fusion temperature, in this way, it is not conducive to the stability of operation. Therefore, oxygen feed rate into nozzle should be controlled in this range in actual operation.
1800
3 18.55Nm /hr 3 30.91Nm /hr 3 37.09Nm /hr
outlet of dense-phase zone
1600
Temperture/K
1400 1200 1000
outlet of dense-phase zone
800 600
outlet of gasifier
400 0
50
100
inlet
150
200
250
300
Axial distance/cm
Fig. 1 Influence of oxygen feed rate into nozzle on temperature distribution 2.1.3 Influence of oxygen feed rate into nozzle on carbon conversion Efficiency of carbon conversion/100%
100 90 80
outlet of dense-phase zone
70 60
3 18.55Nm /hr 3 30.91Nm /hr 3 37.092Nm /hr
50 40
outlet of dense-phase zone
30 20
outlet of gasifier
10 0 0
inlet
50
100
150
200
250
Axial distance/cm
Fig. 2 Influence of oxygen feed rate into nozzle on carbon conversion As shown in Fig. 2, with the increase of oxygen feed rate into nozzle, the carbon conversion increases intensively from the initial 68.16% up to 97.38%, and coal has been basically completely converted. But from the results discussed above, we know that ash easily slags under this operational condition, which can not ensure stable continuous operation, so the operational condition is inappropriate. When oxygen feed rate into nozzle is 18.55
Nm3·hr-1, though oxygen consumption is less, the carbon conversion of coal is only 68.16%, and gasification degree is very low, therefore, it is not suitable for large-scale operations. In view of the above, the effect of oxygen feed rate into nozzle on gasifier’s carbon conversion is significant. Oxygen feed rate into nozzle should be controlled in a reasonable range to make sure that there is not only a higher carbon conversion, but the jetting temperature is not too high and operation is stable. 2.1.4 Influence of oxygen feed rate into nozzle on gas composition When oxygen feed rate into nozzle is 18.55 Nm3·hr-1、30.91 Nm3·hr-1、37.09 Nm3·hr-1, the changes of gas composition along axial distance are respectively listed in Fig.3, Fig.4, and Fig.5. 0.50
CH4
3
18.55 Nm /hr
CO2
0.45
CO H2O
0.40
H2
O2
Mole fraction
0.35 0.30
outlet of dense-phase zone
outlet of gasifier
0.25 0.20 0.15 0.10 0.05 0.00
50
53.64
Fig. 3
100
150
200
250
300
Axial distance/cm
Gas composition changes when oxygen feed rate into nozzle is 18.55 Nm3·hr-1
0.5 CH4
3
30.91 Nm /hr
CO2
CO H2O
Mole fraction
0.4
H2
O2
0.3
outlet of gasifier
0.2
outlet of 0.1 dense-phase zone
0.0
50 62.43
100
150
200
250
300
Axial distance/cm Fig. 4 Gas composition changes when oxygen feed rate into nozzle is 30.91 Nm3·hr-1 0.50 CH4
0.45
CO2
3
37.09 Nm /hr
CO H2O
0.40
H2
Mole fraction
0.35
O2
outlet of 0.30 dense-phase zone 0.25 0.20
outlet of gasifier
0.15 0.10 0.05 0.00
50 65.9
100
150
200
250
300
Axial distance/cm Fig. 5 Gas composition changes when oxygen feed rate into nozzle is 37.09 Nm3·hr-1 As shown in Fig.3, Fig.4, and Fig.5, the increase of oxygen feed rate into nozzle has a significant effect on dense-phase region outlet gas composition, especially on CO and H2O content. CO content increases along with the increase of oxygen feed rate into nozzle, and mole fraction is 9.34%, 17.62% and 20.64% respectively; while H2O content decreases, mole fraction is 50.40%, 40.70% and 38.486% respectively. Meanwhile, with the increase of oxygen feed rate into nozzle, the changes of outlet gas composition are also great as follows:
H2 content decreases and mole fraction is 30.54%, 29.67% and 27.104% respectively; CO content increases and mole fraction is 12.46%, 21.96% and 24.95% respectively; CO2 content decreases and mole fraction is 21.71%, 19.57% and 18.31% accordingly; CH4 content decreases and mole fraction is 3.08%, 2.75% and 2.66%. The reason is that, due to the increase of oxygen feed rate into nozzle, combustion intensifies, resulting in the increase of CO2 content. However, the gasifier temperature rising is conducive to Boudouard reaction of coke and CO2 as well as coke and steam water-gas reaction and CH4 cracking reaction, meanwhile, with the increasing temperature, the equilibrium of CO shift reaction is driven toward the left. All above resulted in that steam content, CO2 content, H2 content and CH4 content decreases in different degree, CO content increases [12]. The increase of oxygen feed rate into nozzle has little influence on outlet gas composition in dilute phase region. In dilute phase region, changes of gas composition are as follows: H2O content decreases intensively; H2 content increases; CO content increases first then has a slight decrease; CO2 content is almost unchanged; CH4 and O2 content are very low (CH4 less than 3.5%, O2 less than 2%), and has a slight decrease. But the increase of oxygen into gasifier raises the overall temperature which is conducive to endothermic reaction. The change of each produced gas component is different in dilute phase region: H2O content reduced by 19.78%, 15.92% and 12.84 correspondingly; H2 content increased by 13.36%, 10.49%, 8.33% respectively; CO content increased by 3.12%, 4.34% and 4.31% separately; CO2 content increased by 3.95%, 1.61% and 0.62%; while, CH4 content reduced by 0.32%, 0.27% and 0.23% respectively. Analysis shows that the sharply reduction of H2O content is the significant reason for the increase of H2 content. At higher temperature, water gas reaction and water gas shift together result in a great increase of H2 content, while CO content increase extent is limited. As described above, with the increase of oxygen feed rate into nozzle, the changes of gasifier temperature, gas composition and carbon conversion are great. And the increase of oxygen feed rate into nozzle also has a great effect on gas composition changes. This shows that the influence of oxygen feed rate into nozzle on gasification characteristics of jetting bed coal gasifier is significant. Jetting depth temperature region is the critical one affecting gasification process of jetting fluidized bed gasifier, supplying heat to gasification reaction.
Its temperature and region changing is the key factor on gas composition changes. Thus, in jetting fluidized bed design and optimization process, the nozzle design should be considered significantly. In gasifier scale-up process, it must be paid more attention to the significant influence of central jetting changes on stable operation as well as produced gas composition. 2.2 Influence of coal feed rate on gasification characteristics 2.2.1 Influence of coal feed rate on jetting region and total temperature in gasifier The ash-agglomerating pilot plant operational conditions are as follows: operational pressure is 0.53 MPa; air feed rate from distributor is 107.94 Nm3·hr-1; steam feed rate from distributor is 97.92 kg·hr-1, and oxygen feed rate into nozzle is 30.91 Nm3·hr-1. This section studies jetting region temperature changes when coal feed rate is 75.87 kg·hr-1, 94.50 kg·hr-1, 123.30 kg·hr-1 respectively. When coal feed rate is 75.87 kg·hr-1, far below design feed rate, jetting region temperature is 1718 K, very close to ash melting temperature; when coal feed rate is 94.50 kg·hr-1, according with design feed rate, jetting region temperature is 1645 K; when coal feed rate is much greater than design feed rate, the jetting region temperature is 1566 K. The influence of coal feed rate on total temperature in the gasifier is listed in Fig.6. 1800
75.87kg/hr 94.50kg/hr 123.3kg/hr
1600
Temperture/K
1400 1200 1000
outlet of dense-phase zone
800
outlet of gasifier
600 400
inlet
0
50
62.43
100
150
200
250
300
Axial distance /cm
Fig. 6 Influence of coal feed rate on temperature inside gasifier In the jetting region it mainly takes pulverized coal combustion reaction supplying energy to entire gasification process and maintaining gasification temperature inside the gasifier [25,
29-31]
. When oxygen feed rate into nozzle is certain, heat from pulverized-coal combustion is
certain. With the increase of coal feed rate, gasifier’s total temperature reduces. The reason is that, with the increase of cold material in dense-phase annulus zone inside gasifier, the overall jet and annulus region convective heat transfer and the radiant heat transfer from jet region to annulus region intensify, resulting in the reduction of jetting region temperature. When coal feed rate increases and coal combustion amount is certain, gasification amount increases accordingly, and the heat absorbed by gasification reaction increases, which is one of the significant reasons resulting in the reduction of gasifier total temperature. As shown in Fig.6, when coal feed rate is 75.87 kg·hr-1, gas production is less, heat absorbed by gasification reaction is also less, and the gasifier temperature decreases gently; while coal feed rate is 123.30 kg·hr-1, gas production is larger, heat absorbed by gasification reaction accordingly increases, and the gasifier temperature decreases greatly. 2.2.2 Influence of coal feed rate on carbon conversion
Efficiency of carbon conversion
1.0
0.8
0.6
75.87kg/hr 94.50kg/hr 123.3kg/hr
0.4
outlet of dense-phase zone
outlet of gasifier
0.2
0.0
inlet
0
50
62.43
100
150
200
250
300
Axial distance/cm
Fig. 7 Influence of coal feed rate on carbon conversion As shown in Fig.7, with the increase of coal feed rate, carbon conversion reduces sharply. When coal feed rate is 75.87 kg·hr-1, share of pulverized coal combustion is greater and gasifier temperature is 1714 K, which is conducive to gasification reaction, and carbon conversion is 98%, almost complete conversion. When coal feed rate is 123.30 kg·hr-1, carbon conversion is only 74.4%. Thus, coal feed rate should be strictly controlled in actual operation.
When coal feed rate is less, jetting region temperature easily exceeds the ash melting temperature, resulting in gasifier slag and affecting operation stability. When coal feed rate is larger, gasification level is low and is not helpful to material conversion. 2.2.3 Influence of coal feed rate on gas composition in gasifier When coal feed rate is 75.87 kg·hr-1, 94.5 kg·hr-1, 123.30 kg·hr-1, the changes of gasifier gas composition are respectively listed in Fig.8, Fig.9, and Fig.10. 0.50
Mole fraction
CH4
75.87kg/hr
0.45
CO2
0.40
CO H2O
0.35
O2
H2
0.30 0.25 0.20 0.15
outlet of dense-phase zone
0.10
outlet of gasifier
0.05 0.00
50
100
150
200
250
300
Axial distance/cm
62.43
Fig. 8 Gas composition changes when coal feed rate is 75.87 kg·hr-1 0.50 0.45
CH4 CO2 CO H2O H2 O2
94.5kg/hr
Mole fraction
0.40 0.35 0.30 0.25 0.20 0.15
outlet of gasifier
outlet of dense-phase zone
0.10 0.05 0.00
50
62.43
100
150
200
250
300
axial distance/cm
Fig. 9 Gas composition changes when coal feed rate is 94.50 kg·hr-1
0.50
CH4
0.45
CO2
123.3kg/hr
CO H2O
Mole fraction
0.40
H2
O2
0.35 0.30 0.25 0.20 0.15
outlet of dense-phase zone
0.10
outlet of gasifier
0.05 0.00
50
62.43
100
150
200
250
300
Axial distance/cm
Fig. 10 Gas composition changes when coal feed capacity is 123.3 kg·hr-1 We can know from the three figures above that the increase of coal feed rate results in significant changes of H2O and H2 content in dense-phase region gas composition. Under the three conditions, H2O content reduces, and mole fraction is 44.113%, 40.70% and 39.07% respectively; H2 mole fraction is 16.72%, 19.18% and 20.88% respectively. The main reason is that sufficient water-gas reaction leads to the large H2O consumption and the increase of H2 content because of the increase of coal feed rate. CO mole fraction is 16.57%, 17.62% and 17.01% respectively. CO content did not change significantly due to coke and CO2 Boudouard reaction which raises CO content, while water-gas reaction consumes CO, the two reactions co-act to make CO content almost unchanged. In outlet effective gas composition, H2 content changes are greater, and mole fraction is 24.35%, 29.77% and 32.69% respectively, which shows that, with the increase of coal feed rate, H2 content in effective gas composition rises. CH4 content rises slightly and CH4 mole fraction is 2.37%, 2.75% and 3.40% respectively, due to the increase of CH4 content generated from coal pyrolysis, with the increase of coal feed rate. CO content almost remains unchanged and CO mole fraction is 20.05%, 21.90% and 21.23% respectively, which shows that the influence of coal feed rate on CO content in outlet effective gas composition is not obvious.
In comparison of the three kinds of coal feed rate, variation of effective gas components in dilute phase region is similar, but different in extent. H2 content rose and mole fraction respectively rose by 7.63%, 10.59% and 11.81%; CO2 content rose slowly and mole fraction respectively rose by 0.95%, 1.69% and 2.32%; CO content increased first then had a slight decrease, and mole fraction respectively rose by 3.48%, 4.28% and 4.23%; CH4 content was very low, less than 4%, and also decreasing degree was a few. It shows that dilute phase region has a great effect on gas composition, and the increment of coal feed rate has an obvious effect on H2 content variation and a limited effect on CO, CO2 and CH4 content variations. From the influence of coal feed rate on gasification characteristics of jetting bed, we can learn that coal feed rate has an obvious effect on gasifier temperature, carbon conversion and H2 content in outlet syngas composition. Coal feed rate also has a significant effect on the stable operation of gasifier, especially when beyond gasifier’s capacity, gasification level will descend sharply, leaving gasifier out of operation.
3 Conclusions The influences of oxygen feed rate into center nozzle and coal feed rate on coal gasification process were simulated for the pilot jetting fluidized bed coal gasifier. The simulations provide insight into the gasification characteristics of the jetting fluidized bed coal gasifier. The following conclusions can be drawn from the above analysis. (1) For jetting fluidized bed coal gasifier, high-temperature jetting region is the key which provides heat for gasification reaction and maintains gasifier temperature, and moreover, it has a critical effect on gasification products and stable operation of gasifier. (2) The changes of operational conditions will lead to obvious variations in gasification characteristics of jetting bed coal gasifier: with the increase of oxygen feed rate into nozzle, the gasifier overall temperature rises and the carbon conversion also increases greatly, but jetting region temperature should be controlled within the coal ash softening temperature; with the increase of coal feed rate, the gasifier overall temperature reduces, however, when coal feed rate is beyond design feed rate greatly, the carbon conversion also decreases greatly and gasification level will descend sharply.
(3) The jetting and dense-phase regions have a significant effect on gas compositions. However, under certain operating conditions, dilute phase region also has a great impact on gas compositions of final gasification products.
Acknowledgment This work is financially supported by National Basic Research Program of China (973 Project 2005CB22120102).
References [1] Glicksman LR. Fluidized bed scale-up [M]. William Andrew Publishing, 1999. [2] Matsen JM. Scale-up of fluidized bed processes: Principle and practice [J]. Powder Technology, 1996, 88(3): 237-244. [3] Pan YG, Velo E, Roca X, Manya JJ, Puigjaner L. Fluidized-bed co-gasification of residual biomass/poor coal blends for fuel gas production [J]. Fuel, 2000, 79(11): 1317-1326. [4] Fang YT, Chen FY, Wang HY, Zhang JM, Wang Y, Zhang BJ. Study on gasification reaction of coal and char in circulating fluidized bed (CFB) II. Effect of temperature and oxygen concentration to gasification reaction in CFB [J]. Journal of Fuel Chemistry and Technology, 1999, 27(1):23-28. [5] Xiao R, Jin BS, Zhou HC, Huang YJ, Zhong ZP, Zhang MY. Coal gasification using high temperature preheated gasifying agent in a pressurized spout-fluid bed [J]. Proceedings of the CSEE, 2005, 25(23):100-105. [6] Li QJ, Zhang MY, Shi AY, Song ZL. Experiment study on coal gasification in pressurized spout-fluid bed [J]. Journal of Southeast University (Natural Science Edition), 2006, 36(5): 765-768. [7] Wang ZH, Zhou JH, Wang QH, Fan JR, Cen KF. Thermodynamic equilibrium analysis of hydrogen production by coal based on coal/Cao/H2o gasification system [J]. International Journal of Hydrogen Energy, 2006, 31(7): 945-952. [8] Zhou SM, Jin BS. A numerical model for pressurized fluidized bed coal gasifier. Coal Conversation, 2006, 36(5): 765-768. [9] Ocampo A, Arenas E, Chejne F, Espinel J, Londono C, Aguirre J, Perez JD. An experimental study on gasification of Colombian coal in fluidized bed [J]. Fuel, 2003, 82(2): 161-164. [10] Huang YJ, Jin BS, Zhong ZP, Zhou HC, Kong HL. Experiment and research on partial coal gasification in circulated fluidized bed [J]. Coal Science and Technology, 2003, 31(12): 20-22. [11] Huang YJ, Jin BS, Zhong ZP, Xiao R, Zhou HC. Factors affecting fluidized bed partial-gasification [J]. Boiler Technology, 2005, 36(6):8-14. [12] Wu XC, Wang QH, Luo ZY, Fang MX, Cen KF. Modelling on effects of operation parameters on entrained flow coal gasification(II): Model prediction and analysis. Journal of Zhejiang University (Engineering Science), 2004, 38(11): 1483-1489. [13] Xiao R, Zhang MY, Jin BS, Huang YJ. High-temperature air/steam-blown gasification of
coal in a pressurized spout-fluid bed [J]. Energy Fuels, 2006, 20(2): 715-720. [14] Pinto F, Franco C, Andre RN, Tavares C, Dias M, Gulyurtlu I, Cabrita I. Effect of experimental conditions on co-gasification of coal, biomass and plastics wastes with air/steam mixtures in a fluidized bed system [J]. Fuel, 2003, 82(15-17): 1967-1976. [15] Ye DP, Agnew JB, Zhang DK. Gasification of a south Australian low-rank coal with carbon dioxide and steam: Kinetics and reactivity studies [J]. Fuel, 1998, 77(11): 1209-1219. [16] Kim YJ, Lee JM, Kim SD. Coal gasification characteristics in an internally circulating fluidized bed with draught tube [J]. Fuel, 1997, 76(11): 1067-1073. [17] Veraa MJ, Bell AT. Effect of alkali metal catalysts on gasification of coal char [J]. Fuel, 1978, 57(4): 194-200. [18] Herguido J, Corella J, Gonzalez-Saiz J. Steam gasification of lignocellulosic residues in a fluidized bed at a small pilot scale. Effect of the type of feedstock [J]. Industrial & Engineering Chemistry Research, 1992, 31(5): 1274-1282. [19] Lv PM, Xiong ZH, Chang J, Wu CZ, Chen Y, Zhu JX. An experimental study on biomass air-steam gasification in a fluidized bed [J]. Bioresource technology, 2004, 95(1): 95-101. [20] Narvaez Ian, Orio Alberto, Aznar Maria P., Corella Jose. Biomass gasification with air in an atmospheric bubbling fluidized bed. Effect of six operational variables on the quality of the produced raw gas [J]. Industrial & Engineering Chemistry Research, 1996, 35(7): 2110-2120. [21] Liu GS, Tate AG, Bryant GW, Wall TF. Mathematical modeling of coal char reactivity with CO2 at high pressures and temperatures [J]. Fuel, 2000, 79(10): 1145-1154. [22] Qu LJ. Progress of research on the fluidized bed coal gasification technology. Coal Conversation, 2007, 30(2): 81-85. [23] Tang Z, Zhang JM, Wang Y, Zhang BJ, Bi JC, Toshinori K. Simulation of a large scale jet-fluidized bed gasifier (I) model set up. Journal of Fuel Chemistry and Technology, 1997, 25(5): 438-444. [24] Ji XG, Bu XP, Wang P, Xin SH. Experiment summary on pressured fixed bed gasification of Guizhou anthracite [A]. Proceedings of the 13th annual meeting of the National Gas Making Technology [C], 2005. [25] Bi JC, Kojima T. Prediction of temperature and composition in a jetting fluidized bed coal gasifier [J]. Chemical Engineering Science, 1996, 51(11): 2745-2750. [26] Guo QJ, Yue GX, Zhang JY, Liu ZY. Flow regime transition and penetration depth of large jetting fluidized bed [J]. Journal of Chemical Industry and Engineering, 2001, 52(9): 803-809. [27] Guo QJ, Yue GX, Zhang JY, Liu ZY. Hydrodynamic characteristics of a two-dimensional jetting fluidized bed with binary mixtures [J]. Chemical Engineering Science, 2001, 56(15): 4685-4694. [28] Zhong WQ, Zhang MY. Jet penetration depth in a two-dimensional spout-fluid bed [J]. Chemical Engineering Science, 2005, 60(2): 315-327. [29] Bi JC, Luo CH, Aoki K, Uemiya S, Kojima T. A numerical simulation of a jetting fluidized bed coal gasifier [J]. Fuel, 1997, 76(4): 285-301. [30] Luo CH, Aoki K, Uemiya S, Kojima T. Numerical modeling of a jetting fluidized bed
gasifier and the comparison with the experimental data [J]. Fuel Processing Technology, 1998, 55(3): 193-218. [31] Kimura T, Bi JC, Uemiya S, Kojima T. Analysis of local reactions in a laboratory scale jetting fluidized bed coal gasifier [J]. Fuel Processing Technology, 1993, 36(1-3): 219-225.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: CARBON MANAGEMENT EXACT TITLE OF PAPER: An Efficient Membrane Process to Capture Carbon Dioxide From Power Plant Flue Gas Bilgen Firat Sercinoglu , Research Engineer, Membrane Technology and Research, Inc. 1360 Willow Road, Suite 103, Menlo Park, CA, US, [email protected], 650-543-3349 Tim Merkel, Director of Process Research and Development, Membrane Technology and Research, Inc. 1360 Willow Road, Suite 103, Menlo Park, CA, US, [email protected] 650-543-3362 Xiaotong Wei, Senior Research Engineer, Membrane Technology and Research, Inc. 1360 Willow Road, Suite 103, Menlo Park, CA, US, [email protected], 650-543-4691 Haiqing Lin, Senior Research Engineer, Membrane Technology and Research, Inc. 1360 Willow Road, Suite 103, Menlo Park, CA, US, [email protected], 650-543-3359 Jenny He, Senior Research Scientist, Membrane Technology and Research, Inc. 1360 Willow Road, Suite 103, Menlo Park, CA, US, [email protected], 650-543-3352 Richard Baker, Founder, Membrane Technology and Research, Inc. 1360 Willow Road, Suite 103, Menlo Park, CA, US, [email protected], 650-328-2238 Abstract To mitigate the harmful effects of global climate change, CO2 in power plant flue gas must be captured and sequestered. Current absorption technologies proposed to capture CO2 from flue gas are costly and energy intensive. Membrane technology is an attractive CO2 capture option because of advantages such as energy-efficient passive operation, no use of hazardous chemicals, tolerance to acid gases and oxygen, a small footprint, no additional use of water, and no steam use requiring modifications to the existing boiler and steam turbine. Working with DOE, MTR has developed new membranes and process designs to recover CO 2 from power plant flue gas. MTR Polaris membranes have CO2 permeances ten times higher than standard commercial membranes, which greatly reduces the cost of a membrane capture system. These membranes are combined with a novel process design that uses incoming combustion air as a countercurrent sweep to maximize membrane module performance and recycle CO2 to the boiler. Design calculations estimate that this membrane process can capture 90% of the CO2 in flue gas as a supercritical fluid using <20% of the plant power, at a cost of $20$30/ton of CO2 captured. This translates to an increase in the levelized cost of electricity (LCOE) of about 40%. Currently, MTR is testing this membrane CO2 capture process with Arizona Public Service (APS) at natural gas and coal-fired power plants. The test at APS’ Cholla power plant uses commercial-scale membrane modules and captures about 1 ton CO 2/day. Technical results from the field tests and future plans will be discussed in this presentation.
CO2 capture by Condensed Rotational Separation R.J. van Benthum, H.P. van kemenade, J.J.H. Brouwers, M. Golombok Eindhoven University of Technology, Netherlands ∗ Corresponding
author e-mail: [email protected] www.tue.nl/ptc Abstract
Condensed rotational separation is a technique in which flue gas is cleaned by condensation of the CO2 and mechanical centrifugal separation. It requires a purification of CO2 in the flue gas, prior to separation. This purification can be realized with existing techniques like oxygen enriched coal combustion or CO2 separating membranes. Combined with an enrichment technique, condensed rotational separation provides an answer that can compete with promising conventional techniques for CO2 capture, like oxy–fuel combustion or amine absorption. These conventional techniques produce a waste stream with a high CO2 purity that can be compressed to supercritical pressure for transport and storage. It is shown that energy consumption of CRS is only slightly more than gas compression of a sequestration stream resulting from conventional separation techniques.
1
Introduction
Application of CO2 capture and storage (CCS) in conventional air blown coal combustion power plants involves the separation of CO2 from the flue gas at low concentration and low partial vapour pressure. Currently the method of choice in industry is the absorption of CO2 with amine. This process requires energy however, mostly in the form of low pressure steam to regenerate the solvent. This steam cannot be used for electricity generation and the costs, including compression of the CO2 to supercritical pressure, are estimated at 1150 MJ per ton of separated CO2 for 90% CO2 removal [1]. Another approach, currently investigated at pilot scale by Vattenfall [2], is the combustion of coal with pure oxygen and is referred to as oxy–fuel combustion. Since no nitrogen is present in the combustion air, flue gas consists of almost pure CO2 . Oxy–fuel enables the capture of CO2 by direct compression of the flue gas but requires an air separation unit (ASU) to supply the oxygen. The energy penalty of oxy–fuel consists of the air separation and compression of the flue gas and is estimated at 990 MJ per ton of separated CO2 , for 90% CO2 removal [1]. In this study we explore a breakthrough compact CO2 capture method that can be either retrofitted to existing power plants or incorporated in traditionally designed new ones: condensed rotational separation, abbreviated as CRS. The technology relies on separation by cooling, a technology pioneered by von Linde in the 1900’s [3]. Gases with a positive Joule–Thomson coefficient decrease in temperature and condense when throttled. By controlling how much the temperature decreases by adiabatic cooling and expansion components can be separated as they have different boiling temperatures. Several processes have been developed on this principle, for example the Total Sprex process to remove H2 S from natural gas [4], Cryocell by CoolEnergy to remove CO2 from natural gas [5], Controlled Freeze Zone by Exxon for sour gas fields [6] and the Alstom anti-sublimation process for flue gases [7].
2
Rapid cooling of binary– or multi–component mixtures of gases to temperatures where one, or some of the components preferentially condense, leads to a mist of very small droplets with diameters of 1 to 10 micron. The phenomenon is known to occur by aerosol formation in flue gases of biomass combustion installations [8], condensate droplets resulting from cooling of wet natural gas [9] and has also been measured in experiments with CH4 /CO2 mixtures and CO2 /N2 mixtures [10]. For a process which relies on fast phase change as a means of separation to be economical and practical, it is necessary to have a device capable of capturing micron-sized droplets with high collection efficiency, low pressure drop and small building space. Cyclones are the standard for liquid/gas separation in hydrocarbon processing plants [11, 12], but can only handle condensing droplet sizes above 15 µm [13–15] . A more feasible solution is provided by the rotational particle separator, abbreviated by RPS. The RPS consists of a cylindrical pipe in which a rotating element is placed. The rotating element is a simple rotating body consisting of a very large number of axial channels of a few millimeters in diameter. In the channels the micronsized droplets are centrifuged to form a liquid film at the walls which is ripped of at the exit of the channels in the form of droplets typically 20 µm or larger. The RPS thus agglomerates droplets to a size, such that they can be separated according to the working principle of ordinary axial cyclones [16–18]. The rotating element can receive its momentum for rotation by pre-rotation of the gas entering the rotating element, and/or by external drive through an electrical motor which is indirectly connected through a magnetic field. In this paper we do not go into the RPS separation technology, which is standard, but assess the possibility of CRS for CO2 capture in coal combustion. To that end, the thermodynamics of CRS are discussed in §2 and translated into separation conditions, separation performance and the involved energy penalties. In section 3 a first attempt is made to get an estimation for the energy costs of CRS. As reference point the 500 MWe supercritical pulverized–coal power plant described in the MIT study "The Future of Coal" [1] is used. After a discussion of the results and future work, the main conclusions are summarized in section 5.
2
Condensed Rotational Separation
The essence of CRS is the mechanical separation of a partially condensed gas mixture into a vapor and a liquid phase by means of cooling and centrifugal separation. The concept of CRS was originally developed for cleaning contaminated natural gas (CH4 , CO2 , H2 S) [16]. In natural gas (see figure 1A), contaminated gas from a high pressure gas well (typically 100 bar) is cooled by expansion (EXP1) such that the gas mixture enters the vapor–liquid region. The condensed mixture contains at this point a CH4 rich vapor phase and a CO2 rich liquid phase. During the expansion the liquid phase is formed as a mist of micron–size droplets as a result of heterogeneous condensation in the gas. After separation by the RPS, the cleaned natural gas (the product) is compressed to pipeline pressure (typically 100 bar). The liquid waste is pumped to supercritical pressure (typically 100 bar) [10, 18–20].
2.1
single stage CRS
In contrast to natural gas (figure 1A), flue gas (N2 , CO2 ) from power plants is usually at ambient pressure (1 bar) and moderate temperature. This means that cooling cannot be provided by expansion. A heat exchanger has to be used to cool down the gas mixture (figure 1B), and a chiller has to be provided (HEX1). This is the first energy penalty. In figure 2A the phase envelope is shown in the p–T diagram for a 21/79 %mole CO2 /N2 composition, corresponding to a typical dry flue gas. This flue gas composition is derived from the stoichiometric combustion of coal (represented by carbon) with normal air (O2 ,N2 ).
3
A Feed
B Feed
EXP1
(CH4, H2S, CO2)
(N2, CO2)
Product (vapor)
COM1
Product (vapor)
(CH4 rich)
(N2 rich)
HEX1 RPS
RPS
Waste (liquid)
Waste (liquid)
(H2S, CO2 rich)
(CO2 rich)
figure 1: Comparison of single stage CRS in natural gas and in flue gas. The dotted line below the heat exchanger presents a part of the waste stream that is separated by wall condensation and drainage in the heat exchanger.
An extended equation of state program is used, based on a cubic equation of state of the Soave– Redlich–Kwong type with pure component parameters fitted to vapor pressures and liquid densities along with a composition dependent mixing rule. A freeze out model to predict the boundary of solid CO2 formation is incorporated. The motivation for developing an in–house code instead of using commercially available programs is that flexibility in the optimization studies was needed. Although the vapor–liquid–solid regime is depicted by the grey areas in figures 2 A and B, we do not acces this region with CRS because of the forming of solid particles. It is known that for temperatures below the freeze–line, solid formation can occur on the walls in the heat exchanger, thereby clogging the heat exchanger. It can also occur in the bulk flow, which is comparable to heterogeneous condensation. Separation of a mixture in the vapor–liquid–solid regime may result in a vapor phase containing small solid particles and a liquid phase in which solid particles are dissolved. This liquid–solid slurry can clog the channels of the RPS and other down–stream equipment. Although it might be possible, no research is done sofar on operating CRS in the vapor–liquid–solid area.
A
250
CO 2 Recovery 100
B
250
90 200
200
80
60 50
100
40
Pressure [bar]
Pressure [bar]
70 150
150
100
30 50
20
50
10 0 −80
−70
−60
−50 −40 temperature [
o
−30 C]
−20
−10
0
0
0 −80
−60
−40 −20 temperature [ o C]
0
figure 2: Phase diagrams of a A:21/79 CO2 /N2 mixture and B:70/30 CO2 /N2 mixture. The white solid line represents the boundary of the vapor–liquid region. The grey area and the white dashed line denote the region in which solid formation occurs and the solids line of CO2 . The color scale denotes the CO2 recovery on a scale of 0 to 100%.
20
4
With this restraint, the point with the lowest pressure at the interface between the vapor–liquid regime and the solids–line (white dotted lines in figures 2 A and B), is at a temperature of -58 o C and 31 bar. This point is mentioned as the lower triple point and is shown in figures 2 A and B by ’TR’. The consequence is that the flue gas has to be compressed (COM1 in figure 1B) in order to apply CRS, thus adding an additional energy penalty to the previous mentioned energy for cooling. The separation performance of CRS can be expressed in two parameters: The waste CO2 purity and the CO2 recovery. The waste CO2 purity is the concentration of CO2 in the waste stream. The recovery is defined as the fraction of the total amount of CO2 in the feed that ends up in the waste. An expression for the CO2 recovery is given as: ˙ liq xCO M xCO2 zCO2 − xCO2 amount o f CO2 in waste 2 . RCO2 = = = (1) ˙ f eed zCO amount o f CO2 in f eed zCO2 yCO2 − xCO2 M 2 ˙ f eed and M ˙ liq denote the feed and waste flow in [mole s−1 ], zCO , xCO and yCO the In equation 1, M 2 2 2 mole fractions of CO2 in the feed, waste and product stream. Strictly defined targets concerning the capture and sequestration of CO2 from flue gases are not well documented. However, in agreement with the EU framework proposal for CCS a target for waste CO2 purity (the CO2 concentration in the waste stream) can be set at ≥ 95% and we aim for a CO2 recovery range of 70–90% [21, 22]. This pipeline specification for transport and storage of CO2 for sequestration is of great importance since it determines the separation conditions of CRS. In figures 2A and 2B the color scale denotes the CO2 recovery. The colors in the figures show the CO2 recovery that can be achieved over the range of pressure and temperature in which the vapor– liquid regime is found. For typical flue gas (figure 2A) the maximum CO2 recovery is a meagre 37 % and not worth the effort of applying CRS. From eq. (1) it can be derived that an increase in feed CO2 concentration is required to increase the CO2 recovery. The flue gas entering CRS must contain a higher CO2 concentration; higher than 21%. To support this conclusion the phase diagram of a 70/30 CO2 /N2 mixture is depicted in figure 2B. For this example a maximum recovery 70–90 % can easily be achieved. Next to temperature and pressure a third parameter is now introduced that affects the CO2 recovery: the CO2 concentration in the feed stream. From figures 2A and 2B and the evaluation of phase diagrams in between, four important conclusions can be drawn: • For a binary mixture within the vapor–liquid regime, the mole fractions of CO2 and N2 in both the vapor and liquid are only dependent on pressure and temperature. • Highest CO2 recovery is found in the vapor–liquid region at the interface between the solids line and the vapor–liquid region. • The solids line (dashed white line) is positioned around -60 o C for CO2 /N2 mixtures and is not much influenced by the CO2 concentration in the feed. • The increase of CO2 concentration in the feed results in extension of the phase diagram towards higher temperatures. As a result, the CO2 recovery increases overall.
The first conclusion can be proven with the Gibbs Phase rule. The conclusion implies that CO2 and N2 fractions in the vapor and liquid phase do not change if mole fractions in the feed are changed while pressure and temperature are kept constant. The total amount of CO2 and N2 in the vapor and liquid phase however do change. The second and third conclusion lead to the choice of the separation temperature as close as possible to the solids line (white dashed line). Since the boundary line of the solids area is hardly affected by the CO2 concentration, the separation temperature can be set at -55o C for any binary CO2 /N2 mixture. As an implication of the first three conclusions, 95% waste CO2 purity is found at -55o C and 36 bar, independent from the the feed concentrations. Two parameters that determine separation performance remain: separation pressure and the CO2 concentration in the CRS feed
5
stream. The separation pressure influences both CO2 recovery and the CO2 purity of the waste stream. The CO2 concentration in the feed only influences the CO2 recovery. Evaluation of the separation performance for different feed concentrations and different separation pressures is shown in figure 3. Figure 3 presents the separation performance as CO2 recovery versus the waste CO2 purity for different feed concentrations. The pressure increases by clockwise direction of the lines. It shows that a 40–70% CO2 concentration in the feed is required to achieve ≥ 95% waste CO2 purity and a CO2 recovery in the range of 70 to 90% (green area in figure 3). The results of figure 3 lead to the requirement of CO2 enrichment: The CO2 concentration in the flue gas, prior to the CRS process needs to be increased. CO2 enrichment can be achieved for example by combustion of the fuel with oxygen enriched air, or by CO2 /N2 separating membranes after combustion and prior to the CRS process. This enrichment step is the third additional energy penalty.
100 95 90
x CO2 (% mole )
85 80 75 70
20 % mole CO 2
65
30 % mole CO 2 50 % mole CO 2
60
70 % mole CO 2 80 %
55 50
0
20
mole
CO
2
40 60 CO 2 Recovery (%)
80
100
figure 3: CO2 recovery against waste CO2 purity (xCO2 ) for a single stage CRS process. The legenda shows the CO2 concentrations in the feed from 20 to 80%. The green rectangle in the upper right corner depicts the targets for CCS: 70–90% CO2 recovery and a waste CO2 purity ≥ 95%. Note: Following the lines in clockwise direction corresponds to an increasing separation pressure.
In summary, there are three energy penalty’s involved with CO2 capture by CRS: cooling by a chiller, compression, and the CO2 enrichment of the flue gas, prior to CRS. We identified that the separation performance of CRS can be expressed as a CO2 recovery and a waste CO2 purity. The CO2 recovery of CRS is determined by the separation pressure and the composition of the CRS feed stream. The waste CO2 purity is solely dependent on pressure, as long as the feed stream consists of only CO2 and N2 .
2.2
two stage CRS
Single stage CRS operating at a separation pressure of 36 bar results in a waste CO2 purity of 95% as stated in section 2.1. At this pressure, the maximum CO2 recovery is however not reached, as can be derived from figure 3. At separation pressures above 36 bar, the CO2 recovery is increased but a lot of nitrogen is dissolved into the waste stream and the requirement of ≥ 95% is not met. At low
6
separation pressures (< 36 bar), the amount of condensed matter is small. However the waste CO2 purity is above 95%. It thus meets the requirement for waste CO2 purity, but only a small part of the CO2 is recovered from the feed in the waste stream. Single stage CRS operating at a pressure above 36 bar can be improved with a second stage in the waste stream that operates at 36 bar (figure 4A). This second stage produces a waste CO2 purity of 95%. Single stage CRS operating at a pressure below 36 bar can be improved by a second stage added to the vapor stream (figure 4B), which operates at a pressure above 36 bar. Both options can be realized as a sequential process (figures 4A and 4B) or a recycle process (figures 4C and 4D). In the latter, the product (vapor) or the waste stream (liquid) from the second stage is fed back to the feed stream.
A) Liquid-Sequential
B) Vapor-Sequential Product (N2 rich)
Product (N2 rich)
1
Feed (CO2 , N2)
2
(T , P1)
2
(T , P2 < P1)
1
Feed (CO2 , N2)
(T , P2 > P1)
(T , P1) Waste (CO2 rich)
Waste (CO2 rich)
C) Liquid-Recycle
D) Vapor-Recycle Product (N2 rich)
Product (N2 rich)
1
Feed (CO2 , N2)
(T , P2 < P1)
(T , P1)
(T , P2 > P1) Feed (CO2 , N2)
2 Waste (CO2 rich)
2
1
(T , P1) Waste (CO2 rich)
figure 4: The different combinations of staging that can be distinguished in two stage CRS.
The analysis of separation performance for all depicted configurations in figure 4 is similar. As an example we discuss the vapor–sequential CRS layout: A single stage CRS process operating at a pressure below 36 bar produces a waste CO2 purity ≥ 95% in the waste stream. However, a great part of the CO2 leaves the single stage CRS process trough the product stream, resulting in low CO2 recovery. Adding a second stage to the product stream that operates at a higher pressure could improve the amount of captured CO2 . Such a two stage vapor–sequential CRS layout, is shown in more detail in figure 5. The second stage consists of compression (COM2) and isobaric cooling (HE2), to reach the separation condition of the second stage. The first stage (psep < 36 bar) results in a liquid CO2 concentration > 95%. The second stage (psep > 36 bar) results in a liquid CO2 concentration < 95%. The second stage separation pressure is determined by the CO2 concentration of the combined waste stream (waste CO2 purity). The procedure for finding the second stage pressure is as follows: a first stage pressure is fixed and the second stage pressure is varied over an interval of pressures, such that it operates in the vapor–liquid region of the second stage gas mixture. The interval is further limited by the ≥ 95% requirement of the CO2 concentration (waste CO2 purity) in the combined waste stream. The second stage pressure is determined by finding the maximum CO2 recovery within this limited interval. The operating limits of the vapor–sequential layout are determined by the boundaries of the
7
vapor–liquid region (white solid line in figures 2A and 2B) and by the requirement for 95% waste CO2 purity of the combined waste stream. In comparison with single stage CRS, vapor–sequential CRS results in improved CO2 recovery, while the CO2 concentration in the waste stream is kept at 95% or more. Feed (CO2 , N2)
Product (vapor) (N2 rich)
COM1
COM2 HE1
HE2 RPS1
RPS2
PUMP
Waste (liquid) (CO2 rich)
figure 5: Vapor–sequential CRS: The second CRS stage is added to the product stream (vapor) of the first stage. The pump illustrates that both liquid streams are at equal pressure before the two streams merge.
In figure 6 the investigated single and two stage processes are compared for the CO2 recovery as a function of the CO2 concentration in the feed. All layouts, depicted in figure 4 produce a waste CO2 purity of ≥ 95%. The CO2 recovery is thus determined by a variation of the separation pressure(s), such that the requirement of the waste CO2 purity is satisfied. For the two stage layouts, only the maximum CO2 recovery at a given CO2 concentration in the feed is depicted. 95 90
2
CO recovery
(%)
85 80
Single Stage Liquid Sequential Vapor Sequential Liquid Recycle Vapor Recycle
75 70 65 60 55 50 45 20
30
40 50 60 CO concentration in CRS feed (% 2
mole
)
70
80
figure 6: Maximum recovery that can be achieved with the different CRS layouts, while producing a waste stream with a CO2 purity of ≥ 95%.
All two stage processes have a clear advantage over the single stage process in terms of CO2 recovery that can be reached at similar CO2 concentration in the feed. In the range of 70–90% CO2 recovery, the vapor–sequential CRS process reaches highest CO2 recovery and is therefore the layout of choice.
8
3
Energy costs
In section 2.2 it is found that a vapor–sequential CRS layout has the advantage over the other layouts in terms of recovery. To achieve a CO2 recovery of 90% and a waste CO2 purity of 95%, the CO2 concentration in the CRS feed must be at least 64% (see figure 6). For this case an example of oxygen enriched coal combustion combined with vapor–sequential CRS is discussed below. The assumptions for this example can be found in table 1. tabel 1: Assumptions for the energy calculations of CRS Power plant electrical power output 500 MWe thermal–to–electric efficiency 40% (CO2 capture excluded) coal type Carbon (HV: 32800 kJ/kg) composition normal air 21/79%mole O2 /N2 CRS ηT/C(isentropice f f iciency) / η pump 80% (per stage) max. compression ratio 4 (per stage) heat exchanger effectivity 80-95% inlet conditions feed: 1 bar , 40o C exit conditions product:1 bar, 40o C waste: 100 bar, 40o C
The oxygen enriched air for combustion is assumed to be provided by mixing of normal air and a high purity oxygen stream from an air separation unit, often denoted as ASU. Although it is unlikely to be the best method to get partial enrichment, it is a proven technique and therefore used as reference for pre–combustion enrichment. The required high purity oxygen for combustion purposes can be achieved by an air separation unit (ASU) using cryogenic distillation. Typical costs of cryogenic distillation to produce low pressure high purity oxygen (∼90% O2 ) from ambient air can be estimated at a basic separation energy of 576 kJ/kg O2 [23]. This basic separation energy does not include the efficiency of the compressors, heat of regeneration of driers and the power consumption of the cooling system of the ASU. However it gives a basic magnitude of today’s feasible energy costs for oxygen enrichment. For this example a process overview for the CRS installation is constructed with use of Aspen Plus (figure 7). The CRS feed stream, which is the mass flow of dry flue gas from the power plant, is compressed in three stages (COM1, COM2 and COM3) with equal pressure ratio to first separation pressure, such that the compression ratio of each compressor is <4. The compression ratio between the vapor stream from RPS1 and the second separation pressure is also <4. Single stage compressors as commonly used in LNG and air separation industry can thus be applied. Separation of the vapor and the formed CO2 rich liquid during cooling takes place in rotating particle separators RPS1 and RPS2 (see figure 7). In the first CRS stage, most liquid is formed and separated, leaving only a small vapor stream for the second CRS stage. After compressors COM1 to COM3 the flue gas is intercooled in heat exchangers HE1, HE2 and HE3. Cooling water is used to bring the flue gas down to 40o C. HE4 in figure 7 is cooled against the high pressure waste stream (HE12) and the low temperature expanded product streams from the second and third expansion stage (HE9, HE10, HE11). The outlet temperature of HE4 and the outlet pressures of expanders EXP1, EXP2 and EXP3 are chosen such, that the inlet and outlet temperatures of HE9, HE10 and HE11 match with the inlet and outlet temperature of HE4 and that thermal power of HE9 to HE12 match with the thermal power of HE4. In HE6 the vapor stream from RPS1 is internally cooled against HE8 in the product stream (vapor) from RPS2. The outlet temperature of HE8 is set by the inlet temperature of HE6. The outlet temperature of HE6 is set such that the thermal power of HE6 and HE8 match.
9
The additional required cooling is provided by chillers. These are represented by HE5 and HE7 in figure 7. The chillers work as reversed Carnot cycles and their efficiency is assumed equal to the inverse Carnot efficiency. The liquid waste stream is pumped to a pressure of 100 bar by PUMP1 and PUMP2. For 90% CO2 capture with a waste CO2 purity of 95% by combination of vapor–sequential CRS and oxygen enriched coal combustion, an energy penalty is paid of 729 MJ/t separated CO2 . This is equal to 18.4% of the electric power output or 7.4% of the produced thermal power. The energy consumption per component is given in table 2: tabel 2: Energy operating costs of two stage vapor–sequential CRS combined with oxygen enriched coal combustion, for a 500MWe power plant with CO2 capture. (nett electrical output: 500MWe )
Oxygen enrichment Compressor Expander Chiller Pump total
[ MJ/t CO2 ] 246 547 -111 41 7 729
%Pelec 6.2 13.8 -2.8 1 0.2 18.4
The results from the example in table 2 show that the total energy requirement is dominated by the CO2 enrichment by oxygen enriched coal combustion (34% of energy consumption) and the compression of the CO2 within the CRS process (75% of energy consumption). Enrichment is therefore an important contributor to the overall energy penalty. The second conclusion that can be drawn from this example is that the total energy requirement CRS combined with CO2 enrichment is of the same order but somewhat better as other reported technologies like oxy–fuel combustion (990 MJ/t, CO2 ) and amine absorption (1150 MJ/t, CO2 ) with CO2 compression [1], that produce 90% CO2 recovery and a CO2 purity of 95%. To achieve a CO2 recovery of <90%, requires less CO2 enrichment, compared to 90% CO2 recovery, as can be seen in figure 6. The energy penalty for CO2 enrichment is thus expected to be less. If additionally, the requirement for waste CO2 purity is increased, the separation pressure(s) decrease, along with a further decrease in CO2 recovery. This implies reduced compression costs. Using CRS in combination with an enrichment technique as a bulk separator (CO2 recovery <90% and/or a waste CO2 purity ≥ 95%), can thus substantially lower the costs.
V
64 %mole, CO2 36 %mole, N2 mass flow: 232.5 kg/s pressure: 1 bar temperature: 40oC
mixture:
FEED STREAM:
40oC
COM2
3.2 bar 160oC
HE11
COM3
S
( to sequestration) 95 %mole, CO2 5 %mole, N2 159.3 kg/s 100 bar 40oC
33.5 bar 162oC
HE3
1 bar -52oC
-2.96.104 kW
6.67.103 kW
2.59.104 kW
WASTE STREAM:
10.4 bar 161oC
2.64.104 kW
40oC
2.65.104 kW
COM1
HE2
-2.75.104 kW
-2.69.104 kW
HE1
( to chimney ) 16.2 %mole, CO2 83.8 %mole, N2 73.2 kg/s 1 bar 40oC
PRODUCT STREAM:
4.19.104 kW
HE12
Heat Exchanger 1
HE4
-5.71.104 kW
3
100 bar -52oC
PUMP1
V+L
1.09.103 kW
-55oC
V+L
HE5
22.4 %mole, CO2 77.6 %mole, N2 82.7 kg/s 33.5 bar -55 oC
-1.07.104 kW
40oC
.
HE9
7.25 10 kW
EXP2
-6.24.103 kW
-43oC
5.1 bar -52oC
Heat Exchanger 1
40oC
40oC
HE10
6.73.103 kW
EXP3
-6.56.103 kW
100 bar
L
RPS1
V
25.1 bar -52oC
.
.
3
HE8
7.47 10 kW
HE6
-6.86.103 kW
Heat Exchanger 2
95.5 %mole, CO2 4.5 %mole, N2 149.8 kg/s 33.5 bar -55oC
83.5 bar 22oC
3
COM4
5.51 10 kW
21oC
EXP1
-4.32.103 kW
-40oC
PUMP2
V+L
-55oC
1.86.101 kW
HE7
-3.79.103 kW
100 bar
L
RPS2
V
V = Vapor L = Liquid S = Supercritical
87.3 %mole, CO2 12.7 %mole, N2 9.5 kg/s 83.5 bar -55oC
16.2 %mole, CO2 83.8 %mole, N2 73.2 kg/s 83.5 bar -55oC
10
figure 7: Flow and energy diagram of the two stage vapor–sequential CRS process for oxygen enriched coal combustion.
11
3.1
Energy costs versus recovery
The example calculation in the beginning of § 3 is only a point calculation in terms of CO2 feed concentration (64%) and CO2 recovery (90%). It is however of interest to take a look at the behavior of the energy penalty caused by CRS only, over a range of CO2 recoveries and for different CRS layouts. For derivation of the energy penalty, the feed mass flow of CRS is required. To derive the CO2 enriched feed mass flow for CRS, CO2 enrichment is assumed to be a black box in which N2 is separated from the flue gas stream to obtain the required CO2 concentration in the CRS feed. The feed mass flow is found by solving the componentwise mass balance of this black box. In figure 8 the energy penalty of single stage CRS is evaluated for 60, 70 and 80% CO2 feed concentration and a waste CO2 purity between 99 and 95%. For single stage CRS, operating at a fixed CO2 feed concentration, a decrease in CO2 recovery comes together with an increase in CO2 concentration in the waste stream (waste CO2 purity), as can be seen in figure 3.
12
11
%P electric
[%]
10
80%
9
70%
60%
8
7
6
5
0
10
20
30
40
50
60
CO 2 recovery [%]
70
80
90
100
figure 8: Energy penalty of CRS versus CO2 recovery for 60, 70 and 80% CO2 concentration in the feed.
From figure 8 it can be concluded that energy costs of CRS decrease for increasing CO2 feed concentration and for decreasing CO2 recovery. Further investigation towards possibilities of enrichment (pre– and post combustion) is needed to complete the total energy requirement picture. For every CO2 recovery an optimum is expected in total energy requirement versus CO2 concentration in the CRS feed. In figure 6 it is seen that vapor–sequential CRS results in higher CO2 recovery than single stage CRS for the same CO2 concentration in the feed. In the end, it is however not the CO2 recovery that determines the layout of choice, but the energy requirement of CRS for the same CO2 recovery and same CO2 concentration in the feed. For 64% CO2 concentration in the CRS feed (equal to the example in the beginning of § 3) an example is illustrated in figure 9, where the energy penalty of single stage CRS is compared to vapor–sequential CRS. The line in figure 9 corresponding to vapor– sequential CRS is given over the range of minimum–maximum recovery that can be reached, while producing a 95% CO2 concentration in the waste. Figure 9 shows that the CO2 recovery of single stage CRS, at which the waste CO2 purity is 95% (86% recovery), can also be approximated with vapor–sequential CRS but with a smaller energy requirement (∼1%). Vapor–sequential CRS has therefore not only the advantage of higher CO2 recovery, compared to single stage CRS, but also the advantage of a lower energy penalty at equal CO2 recovery.
12
12
Vapor-Sequential
11.5 11
%P
electric
[%]
10.5 10
Single stage
9.5 9 8.5 8 7.5 7 50
55
60
65
70
75
CO 2 recovery [%]
80
85
90
95
figure 9: Comparison of the energy penalty between single stage and vapor–sequential CRS as a function of recovery for a CO2 concentration of 64% in the feed stream. The line for vapor–sequential CRS is calculated for the minimum and maximum recovery that can be achieved while producing a waste of 95% CO2 purity.
3.2
Compression of the waste CO2 stream
As pointed out previously in this paper, CRS produces a CO2 waste stream as a liquid which is then pumped to supercritical pressure. Oxy–fuel combustion and amine absorption for example also produce a high purity gaseous CO2 waste stream, but at atmospheric pressure. For transport and underground injection, the CO2 waste stream must still be compressed to supercritical pressure. In general all carbon capture processes that produce CO2 for transport and injection have to deal with this compression of the CO2 waste stream to supercritical pressure. The penalty of this compression is basically equal for every carbon capture technique. It is therefore interesting to investigate how much extra energy is required by CRS, compared to gas compression of the CO2 waste stream from conventional techniques. For transport and underground injection, the CO2 waste stream must satisfy the pipeline specification of ≥ 95% CO2 concentration. Since this is almost pure CO2 , the CO2 waste stream from conventional CO2 capture processes can be assumed as a pure CO2 stream. Gas compression of the waste CO2 stream can be assumed as isothermal compression of an ideal gas. More realistic is however a four stage isentropic compression of a real gas with intercooling and an isentropic efficiency of 80%. Ideal gas isothermal compression represents the absolute minimum amount of energy required to bring the sequestration stream from atmospheric to supercritical pressure. Both the energy costs of isothermal compression and 4–stage compression are depicted in figure 10, together with the energy costs of single stage CRS from figure 8. Figure 10 shows that if 90% of the CO2 in flue gas is captured, about 7–9% of the electric power output of the plant is required to compress it from 1 bar to a sequestration pressure of 100 bar. To do the same with single stage CRS requires 11% for the CO2 concentration in the feed for which 90% CO2 recovery can just be achieved. The extra energy required by CRS in this case, is only 2% of the electric power of the power plant. When the CO2 concentration in the feed increases or the CO2 recovery decreases, in figure 10, the energy consumption of CRS approaches the energy consumption of the 4 stage gas compression of the CO2 waste stream. This has two causes. To explain these two causes, it is convenient to describe a single stage CRS process as two physically separated streams (illustrated in figure 11): a N2 rich stream and a CO2 rich stream. Both streams are compressed and cooled to separation pressure and temperature. After separation the N2 rich stream is heated and expanded to the same pressure
13
12
60%
%P
electric
[%]
10
70% 80%
8
6
4
2
0 0
Isothermal compression ideal gas 1−>100 bar 4−stage compression 1−>100bar, ηisenth =0.8 10
20
30
40
50
60
CO recovery [%]
70
80
90
100
2
figure 10: Energy penalty of CRS for 60, 70 and 80% CO2 concentration in the feed stream, compared to gas compression of the sequestration stream as a function of CO2 recovery. The gas compression from 1 to 100 bar is calculated for a pure CO2 stream and for both isothermal compression of ideal gas and 4 stage isentropic compression with an isentropic efficiency of 0.8.
COM1
HEX1
COM2
HEX3
COM4
HEX5
N2 rich stream
A CO2 rich stream
B
CHIL
separation
and temperature as the feed. For the CO2 rich stream there are two options: Case A: heating of the CO2 rich stream and gas compression to supercritical pressure. Case B: pumping the liquid to supercritical pressure and then heat it up to the same output temperature as case A. Suppose there are no irreversible losses. Then the thermal cooling energy of the N2 rich stream before separation in HE1 can be fully exchanged with the thermal energy after separation in HE2. Similarly the compression work before separation in COM1 can be produced completely by expansion after separation in EXP. Therefore the N2 rich stream is energy neutral.
HEX2
EXP
HEX4
COM3
PUMP
HEX6
figure 11: A schematic presentation of CRS in which the product and waste streams are depicted as physically separated streams. For the CO2 rich stream, case A corresponds to gas compression of the CO2 rich waste stream. Case B represents the liquid pumping as occurs in CRS.
14
For case A in the CO2 rich stream, heat before (HE3) and after separation HE4 can be fully exchanged. Heat exchangers HE3 and HE4 can therefore be neglected from the picture and only a gas compression of the CO2 rich stream (COM2 and COM3) is left. In case B, the liquid is pumped to supercritical pressure, thereby heating up only a few degrees. The thermal energy in HE6 is therefore less, compared to HE4, when heated up to the same outlet temperature as case A. If heat before (HE5) and after (HE6) is exchanged, there is a lack of cooling power before separation. This cooling power has to be provided at the separation temperature by a chiller (CHIL). In this paper we suppose chillers that work on mechanical power and water cooling. For the conditions used in this paper a COPR of 2 is used, so the mechanical power required by the chiller is 0.5 times the thermal energy that must be cooled from the CO2 rich stream. The energy that must be extracted from the chiller by cooling water is 1.5 times the thermal energy that must be cooled from the CO2 rich stream. A lower CO2 recovery in CRS corresponds to a lower separation pressure, as can be derived from the phase diagrams (figure 2). The temperature after compression will therefore be lower, which means less cooling power is needed to cool down the CO2 rich stream to separation temperature. As a consequence, the required cooling power by the chiller (CHIL) decreases in case B. The difference between the mechanical energy required in case B (chiller, compressor, pump) and case A (compressor) becomes therefore smaller with decreasing separation pressure and thus with decreasing CO2 recovery. One can interpret this as a result of the heat leakage by water cooling in the chiller in case B, which is not present in case A. Water cooling has a zero energy penalty, since water pumping requires a negligible amount of energy compared to the total costs of CRS and it is usually available at power plant locations. The difference in heat leakage between CRS (case B) and gas compression (case A) is the primary effect causing the energy penalty of CRS to approximate 4 stage gas compression for decreasing CO2 recovery. The secondary cause is related to the efficiencies in the system, to describe irreversible losses. CRS operates with a compression efficiency of 80% and heat exchanger effectiveness of 0.8–0.95. The 4–stage gas compression operates at 80% isentropic efficiency. The total efficiency can therefore differ between CRS (case B) and gas compression (case A). Irreversible losses in the N2 rich stream also explain the shift of the CRS energy penalty towards 4 stage gas compression in figure 10. A higher CO2 concentration in the CRS feed implies that less nitrogen is present. The N2 rich mass flow is therefore smaller. Nett irreversible losses in the N2 rich stream (figure 11) thus become less and the energy consumption of CRS tends more towards 4 stage gas compression of the pure CO2 stream (figure 10). It can be concluded that CRS adds only a small extra energy penalty to the gas compression of the CO2 waste stream. This extra energy penalty decreases if the CO2 concentration in the CRS feed increases or if the CO2 recovery is allowed to be lower then 90%. If one wants to apply a conventional technique like oxy–fuel for the capture of CO2 , then one can think of replacing the ordinary gas compression by CRS. A small energy penalty is then added by CRS, but the required CO2 concentration resulting from oxy–fuel combustion is allowed to be lower then the pipeline specification. Energy can then be saved on the air separation unit to provide oxygen enriched air for oxy-fuel combustion and/or costs can be saved on the quality of the fuel without loss of overall power plant efficiency.
15
4
Discussion
Condensed rotational separation can only capture a considerable amount of CO2 (CO2 recovery between 70 and 90%) and bring it to a pipeline specification of ≥ 95%, if combined with a method to enrich the flue gas in CO2 prior to CRS to a CO2 concentration of 40 to 70%. Furthermore the application of CRS is most effective when bulk separation of CO2 is required (CO2 recovery ∼ 70 to 90%). In that case, the energy penalty for CRS is about 9 to 12% of the electric power output, as can be seen in figure 10. Estimated from figure 10 and the example calculation in § 3.1, the energy costs of CRS combined with oxygen enriched coal combustion for 70 to 90% CO2 removal are between 16 and 19% of the electric power output. This energy penalty range may vary somewhat for a different concentration of CO2 in the CRS feed and/or for a different CRS configuration. All carbon capture processes require compression of the CO2 stream to approximately 100 bar for transportation and/or subsurface injection. CRS itself consumes only marginally more energy than the gas compression of a sequestration stream from other separation techniques, as was shown in § 3.2. However, by using CRS instead of conventional compression, the sequestration stream can be brought to pipeline specifications (≥ 95%). As a consequence, conventional techniques like oxy–fuel combustion, become a purification step prior to CRS, which have to produce a flue gas stream with only a moderate CO2 purity. The implication is that the CO2 concentration in the flue gas after oxy– fuel combustion (the CRS feed stream) does not have to satisfy the pipeline specification anymore. A CO2 concentration in the CRS feed steam of 40 to 70% is satisfactory. This means that we can relax the requirements of oxy–fuel combustion on the leakage of ambient air into the boiler, the fuel quality and the air separation. The operating and capital costs can therefore be potentially lower, then applying oxy–fuel with CO2 as a CO2 capture method. From sections 2 and 3 it is seen that two stage configurations of CRS can improve the CO2 recovery for the same concentration of CO2 in the feed stream and if the same waste CO2 purity is required. It can also produce a CO2 recovery equal to single stage CRS at lower energy costs, as can be seen in figure 8. In the range of 70–90% CO2 recovery all configurations are within a variation of CO2 recovery of 10% (figure 6). Thus vapor–sequential CRS not only improves the CO2 recovery in the range of 70 to 90%, it also can decrease the energy penalty for the same CO2 recovery as compared to single stage CRS. For an estimation of the energy penalty of CRS, a simple model consisting of stoichiometric combustion of carbon with normal air (N2 , O2 ) was used, to predict flue gas concentration and flue gas mass flow. Real coal does not exist solely of carbon but also contains hydrogen, nitrogen, sulfur, oxygen moisture and small amounts of other components and its composition and energetic value varies widely over the different mining locations. Furthermore, coal is normally combusted with excess air. Therefore flue gas from coal–fired power plants also contains H2 O and O2 . Water is condensed during cooling, or can be removed prior to the CRS process. The oxygen left over from excess air combustion might influence the separation performance of CRS, since O2 behaves like N2 in the flue gas. The influence of O2 on separation performance should be checked for flue gas mixtures of N2 /O2 /CO2 . CRS in combination with another enrichment technology is most effective as a bulk separator as can also be concluded from other CO2 removal applications, i.g. sour gas [24] and coal gasification [25]. Even when CRS by itself cannot reach the required CO2 recovery it can be very cost effective to use CRS in tandem with a conventional technology. The energy consumption for the classic absorption process for example, increases rapidly with both the volume flow and CO2 concentration. CRS might be used to lower this energy penalty.
16
5
Conclusions • Condensed rotational separation, abbreviated as CRS, requires a CO2 enrichment step prior to the CRS process, in order to separate a considerable amount of CO2 at a sufficiently high CO2 purity. This enrichment can be realized by oxygen enriched coal combustion. It could also be realized by CO2 /N2 separation after combustion in the case of membranes. Application of CRS for CO2 capture can thus supplement other separation techniques to achieve possible lower overall energy costs, while guaranteeing the pipeline specification for transport and subsurface injection of CO2 . • All CO2 capture processes have to compress the sequestration stream to a supercritical transport and or injection pressure. In general this is a gas compression. CRS produces the CO2 in a liquid form which requires the liquid pumping to supercritical pressure. The energy penalty of CRS is found to be only slightly higher than gas compression. • By replacing gas compression with CRS in oxy–fuel combustion, the requirement for the CO2 content of the flue gas after combustion can be lowered. This allows for reduced demands on air tightness of the boiler and coal feed system, the fuel quality and the air separation.
References [1] J. Katzer, S. Ansolabehere, and J. Beer. The future of coal: An interdisciplinary mit study. Technical report, Massachusetts Institute of Technology, Massachusetts, USA, 2007. [2] Lars Strömberg, Jürgen Jacoby, Uwe Burchhardt, and Frank Kluger. Update on vattenfall’s 30 MWth oxyfuel pilot plant in schwarze pumpe’. Energy Procedia, 1, 2009. [3] Cengel Y. and Boles M. Thermodynamics: an engineering approach. McGraw-Hill, 5th ed edition, 2006. [4] Mougin P., Renaud X., and Elbaz G. Operational validation of the sprex´l process for bulk H2 S and mercaptans removal. The Gas Industry: Current & Future, 6, 2008. [5] Hart A. and Gnanedran N. Cryogenic CO2 capture in natural gas. Energy Procedia, 1, 2009. [6] Mart C.J., Valencia J.A., and Northrop P.S. Developing sour gas resources: controlled freeze zone technology with integrated acid gas management. Sour Oil & Gas Advanced Technology, 2010. [7] Clodic D., El Hitti R., Younes M., Bill A., and Casier F. CO2 capture by anti-sublimation, thermoeconomic process evaluation. National Conference on Carbon Sequestration, 2005. [8] Best C.J.J.M. de, Kemenade H.P. van, Brunner T., and Obernberger I. Particulate emission reduction in small-scale biomass combustion plants by a condensing heat exchanger. Energy and Fuels, 22, 2008. [9] Austrheim A. Experimental Characterization of High-Pressure Natural Gas Scrubbers. PhD thesis, University of Bergen, 2006. [10] Willems G.P., Golombok M., Tesselaar G., and Brouwers J.J.H. Condensed rotational separation of CO2 from natural gas. AIChE Journal, 56, 2008. [11] Strauss W. Industrial gas cleaning. Pergamon, 2nd ed edition, 1975.
17
[12] Campbell J.M. Gas conditioning and processing: The equipment modules. Norman OK: Campbell, vol. 2 edition, 1994. [13] Purchas D.B. Solid liquid separation technology. Croydon: Uplands, 1981. [14] Clift R. Inertial separators: basic principles in Gas cleaning in demanding applications. Seville JPK, 1997. [15] Svarovsky L. Hydrocyclones. 1984. [16] Brouwers J.J.H. Phase separation in centrifugal fields with emphasis on the rotational separator. Exp Therm Fluid Sci, 26:325–334, 2002. [17] Mondt E., Kemenade H.P. van, and Schook R. Operating performance of a naturally driven rotational particle separator. Chem. Eng. Techn., 29, 2006. [18] Willems G.P., Kroes. J.P., Golombok M., van Esch B.P.M., Kemenade H.P. van, and Brouwers J.J.H. Performance of a novel gas-liquid separator. march 2010. [19] Willems G.P. Condensed Rotational Cleaning of Natural Gas. PhD thesis, Eindhoven Univerity of Technology, 2009. [20] Wissen R.J.E. van. Centrifugal Separation for cleaning well gas streams. PhD thesis, Eindhoven Univerity of Technology, 2006. [21] Commission of the European communities. Limiting global climate change to 2 degress celsius: the way ahead for 2020 and beyond. 2007. [22] Jared Cliferno. Existing plants, emissions and capture-setting co2 program goals. Technical Report 2009/1366, National Energy Technology Laboratory, U.S., 2009. [23] Arthur Darde and Rajeev Prabhakar. Air separation and flue gas compression and purification units for oxy-coal combustion systems. Enery Procedia, (1):527–534, 2009. [24] J.J.H. Brouwers and Kemenade van H.P. Condensed rotational separation to upgrade sour gas. Sour Oil & Gas advanced Technology, pages 173–187, 2010. [25] Brouwers J.J.H. and Kemenade van H.P. Condensed rotational separation for co2 capture in coal gasification processes. 4th Int. Freiburg Conference on IGCC & XtL technologies, 2010.
Influence of pressure on dry reforming of methane over carbonaceous catalyst Bingmo Zhang, Yongfa Zhang*, Guojie Zhang, Fengbo Guo Key Laboratory of Coal Science and Technology of Shanxi Province and the Ministry of Education ,Taiyuan University of Technology , Taiyuan 030024, China Abstract:Reforming of methane by CO2 to syngas has been studied on carbonaceous catalysts on high pressure. The catalysts have been characterised by BET and FITR techniques. The catalytic activity and stability of catalysts are closely related to the pressure. It has been observed that the catalytic activity, CH4 and CO2 conversion decrease with the pressure increased. Some process methods, such as increasing the reaction temperature, prolonging reaction residence time and increasing CO2 and CH4 molar ratio, can improve catalysts stability. Moreover, the positive effect of carbonaceous catalysts oxygen-containing groups(C-O) on catalysts activity has been evidenced. The basic function of the carbonaceous materials surface area also seemed to increase H-abstraction of methane and CO2 adsorption. Keywords: Methane reforming, carbon dioxide, pressure, carbonaceous catalyst
1. Introduction Because the dry reforming of methane technology has effectively combine the methane using with carbon dioxide conversion together. The synthesis gas is not only suitable for Fischer-Tropsch synthesis, methanol synthesis and carbonyl synthesis raw material gas, but also can achieve carbon dioxide emission reductions. The dry reforming of methane technology to a synthesis gas has attracted academic and industrial interests. The extensive literature has reported the study over a wide range of catalysts such as group Ⅷ metals
[1-8]
. However, the major drawback of these catalysts is noble
metals catalysts high cost, nickel-based catalysts coking. In recent year, a possible alternative and a new kind of cheap catalyst is carbonaceous catalyst, which has been
*Corresponding author. Tel. : +86 0351 6018676
E-mail address: [email protected]
reported to have significant potential [8-11]. Technological economy has shown that process will be greatly reduced energy consumption if it can be carried out under a certain pressure(2~4Mpa). However, the CH4-CO2 reforming over carbonaceous catalysts have carried out under atmospheric pressure. In this paper, the effect of reaction pressure, temperature, CO2 and CH4 feed gas ratio, residence time on the reforming reaction in a high pressure reactor has presented.
2. Experimental 2.1.
Catalyst preparation Table 1 The proximate analysis and ultimate analysis of carbon catalyst proximate analysis/ w %, ad Ultimate analysis/ w %, daf
Sample M
A
V
C
H
N
S
O(diff)
DT Coal
3.10
12.20
29.00
87.70
4.96
1.27
0.42
5.36
Semi-coke
1.20
13.30
4.50
94.60
1.47
0.99
0.17
2.36
The carbonaceous catalysts used in the experiment were prepared by pyrolysis of coal at 1150℃ for 1.5h and crushing the catalyst mass to 30-60mesh-size particles
[8]
.
The Proximate analysis and ultimate analysis of carbonaceous catalysts were shown in Table 1. 2.2. 2.2.1.
Physico-chemical characterization BET
All samples were degassed at 573K for 1 h before measurements to remove impurities of the catalyst surface. The specific surface area of the catalysts was determined by nitrogen adsorption-desorption measurement at -196 ℃ in Tristar gas adsorption system. 2.2.2
FTIR
The oxygen-containing function group of the carbonaceous catalysts was determined by means of a VERTEX70 Fourier transformation infrared spectrometer. Scanning area was 4000cm-1 ~ 400cm-1. Resolution was 4cm-1.Wave number accuracy was 0.01cm-1. Scan time was16 s.
2.3.Catalytic activity The catalytic activity measurements were performed in a high pressure fixed bed reactor
[12]
. Before CH4-CO2reforming reaction, a certain carbonaceous catalyst was
reduced in 10% H2-90% N2 at 900 ℃ for 1h. Then a feed mixture consisting of CH4-CO2 was introduced into reactor. The reforming reactions were tested with a temperature range from 700 ℃ to 950 ℃ with the stabilization of 2 h at each temperature.
3.Results and discussion 3.1. Effect of pressure on catalytic performance 0.5Mpa 1.6Mpa 2.8Mpa
80
Conversion of CO2,%
Conversion of CH4,%
60
0.5Mpa 1.6Mpa 2.8Mpa
80 70 60
40
50
20
0
40
0
50
100
150
Time,min
200
250
30
0
50
100
150
200
250
Time,min
Fig. 1 Effect of pressure on conversion of methane and carbon dioxide
CH4-CO2 reforming is a volumetric expansion process. According to thermodynamic analysis, the reaction equilibrium conversion decreases with the increase of pressure. Effect of pressure on conversion of methane and dioxide carbon over carbonaceous catalyst, the residence time 6 s, reaction temperature 950℃, which is shown in Fig. 1. The pressure increasing, the conversion of methane and carbon dioxide decreased. When the reaction was stable, the pressure was 0.5, 1.6, 2.8 Mpa, the methane conversion was 38.7%, 29.0%, 26.7% and the conversion of carbon dioxide was 61.1%, 50.3%, 46.3%, respectively. In addition, it can be concluded that the reaction equilibrium time is influenced by pressure, that is, the pressure increasing, the reaction equilibrium time is short. This is mainly due to the partial pressure of the components increased, with the system pressure increasing, the contact time between
the gas molecules and catalysts shorten. 3.2. Effect of temperature on catalytic performance
Conversion of CO2, %
Conversion of CH4, %
60
40
20
0
80
900℃ 800℃ 700℃
80
0
50
100
150
200
250
Time, min
900℃ 800℃ 700℃
60
40
20
0
0
50
100
150
200
250
Time,min
Fig.2 Effect of temperature on conversion of methane and carbon dioxide
Figure 2 shows temperature effect on CH4 and CO2 conversion over carbonaceous catalyst at pressure of 2.8MPa for 240 min, the residence time 6 s. Carbon dioxide reforming of methane was a strong endothermic reaction. Increasing reaction temperature was in favor of the increase of methane and carbon dioxide transformation rate. However, methane decomposition was an endothermic reaction, with the temperature increasing, carbon deposition accelerated caused by methane decomposition. The test results showed that the conversion of methane and carbon dioxide was improved with the reaction temperature increasing. From Fig. 2 also seen, the conversion of carbon dioxide and methane was almost same in the temperature range of 700 to 800℃. When the temperature was up to 900℃, the conversion of carbon dioxide was obviously high than methane. Reaction temperature affected catalyst capability obviously, higher reaction temperature was in favor of CH4 conversions and carbon elimination by CO2. 3.3. Effect of feed ratio of gas on catalytic performance Feed ratio of gas has an important influence on the catalyst performance. Methane reforming reaction also can get high-quality synthesis gas by adjusting the ratio of feed gas for Fischer-Tropsch synthesis. Figure 3 shows feed ratio of gas effect on CH4 and CO2 conversion over carbonaceous catalyst at pressure of 1.6MPa for 240 min at 950℃.
CH4/CO2=1.8:1
CH4/CO2=1.8:1
CH4/CO2=1.4:1 CH4/CO2=1:1
CH4/CO2=1:1.5 CH4/CO2=1:2
60
40
20
0
0
50
100
150
Time, min
200
250
Conversion of CO2, %
Conversion of CH4, %
80
CH4/CO2=1.4:1
80
CH4/CO2=1:1
CH4/CO2=1:1.5 CH4/CO2=1:2
60
40
20
0
0
50
100
150
200
250
Time, min
Fig 3 Effect of feed ratio of gas on catalytic performance
As can be seen from the figure 3, conversion of methane increased with CH4/CO2 ratio decreasing. When the CH4/CO2 ratio was 1.8, methane conversion was about 3.1% after 240 minutes under. The CH4/CO2 ratio arrived at 1.0 and 0.5, methane conversion was up to 26.0% and 32.9%, increased by 22.9% and 29.8%, respectively. This was mainly due to methane partial pressure increased with CH4/CO2 ratio increasing. As s result, the surface concentration of the reaction intermediate (CH x) increased in catalysts surface and the surface residence time extended on catalysts surface
[13]
. Then the surface concentration of carbon increased, the gathering of
carbon atoms raises to cover catalysts surface catalytic active sites, the catalytic activity decreased. On the contrary, increasing the carbon dioxide partial pressure, the surface concentration of carbon dioxide increased, the gathering of carbon atoms weakened to cover catalysts surface catalytic active sites. As a result, it brought out an increase in the carbon elimination by CO2 and inhibited CH x dissociation on catalysts surface. As the conversion of carbon dioxide was increased, CH4/CO2 ratio decreased, while as CH4/CO2 ratio was less than 1.0 continuously, the conversion of carbon dioxide decreased instead. The two main reasons resulting in the problems may be the decreasing carbon dioxide conversion, one being an increase in absolute quantity (space velocity) in feed gas, relative conversion of carbon dioxide decreased, and the second being reaction formed carbon monoxide disproportionation to form carbon dioxide. Based on the two factors above, the conversion of carbon dioxide decreased
by CH4/CO2 ratio being less than 1. 3.4. Effect of residence time on catalytic performance 100
3s 6s 10s
80
3s 6s 10s
Conversion of CO2, %
Conversion of CH4, %
80
60
40
20
60
40
20
0
0
50
100
150
200
250
0
Time, min
50
100
150
200
250
Time, min
Fig 4 Effect of residence time on conversion of methane and dioxide
Figure 4 shows residence time effect on CH4 and CO2 conversion over carbonaceous catalyst at pressure of 1.6MPa, reaction temperature 950℃. With the prolonging of residence time, the conversion of methane and carbon dioxide was improved. When the residence time was 3 s, the conversion rate of methane and carbon dioxide was about 22.3% and 31.3%. Increasing the residence time to 10 s, the conversion of methane and carbon dioxide was up to 40.7% and 62.0%, respectively. 3.5. Catalysts Characterization The BET surface area of the carbonaceous catalysts is shown in table 2. The catalysts specific surface area, pore diameter and pore volume decrease when adding reaction pressure. The surface area of original sample is 94 m2/g, which is dropped to 87 m2/g, 79 m2/g and 72 m2/g after reaction (1.6 MPa, 2.8 MPa and 3.2 MPa), respectively. The pore diameter of original sample is 19 μm, which is reduced to 17 μm, 16 μm and 15 μm after reaction (1.6 MPa, 2.8 MPa and 3.2 MPa), respectively. The pore volume of original sample is 0.08 cm3/ g, which is lowered to 0.07 cm3/ g, 0.06 cm3/ g and 0.05 cm3/ g after reaction (1.6 MPa, 2.8 MPa and 3.2 MPa), respectively. It indicated that the carbonaceous materials surface area is attributed to enhancement of H-abstraction of methane and CO2 adsorption, speeding up coke gasification by CO2 and methane decomposition [8, 14].
Table 2 Characteristics of carbonaceous catalysts ABET/(m2/g) Vp/(cm3/g)
Sample original sample
94
samples after reaction(1.6MPa)
87
0.07
17
samples after reaction(2.8 MPa)
79
0.06
16
samples after reaction( 3.2 MPa)
72
0.05
15
-1
Wave number,cm
0.08
dp(μm) 19
Table 3 Infrared spectra of functional groups Functional groups Compounds
3200-3600
O-H
-OH
3030-2800
C-H
Alkanes
1600-1900
C=O
1330-1110
C-O
Carboxyl, carbonyl, aldehyde, ketone Phenols, ether alcohols, esters
870-700
aromaticring 0.8
1.6MPa
0.7
2.8MPa
0.6
Absorbance
0.5
3.2MPa origin sample
0.4
3.6MPa
0.3 0.2 0.1 0.0 500
1000
1500
2000
2500
3000
Wavenumber,cm
3500
4000
4500
-1
Fig. 7 FITR of catalysts
Table 3 shows the infrared spectra of functional groups. The catalysts FITR is shown in Fig. 7. From the FTIR spectra of catalysts can be seen, for original catalyst sample, four absorption peaks appear, and the peaks are located in the wavenumber of 3400cm-1, 1600cm-1, 1400cm-1 and
800cm-1, respectively. It indicates that the
original catalyst sample may contain O-H, C = O and C-O functional groups. After reaction, the absorption peak in 1400 cm-1 weakens, the higher the pressure of reaction the weaker of the peak, and even disappears, that is, oxygen-containing functional group of C-O consumed in reforming process. It indicated that the carbonaceous catalysts oxygen-containing groups(C-O) have a positive effect of on catalysts activity. It also has also been reported by other authors [14, 15].
4. Conclusions
The catalytic activity and stability of catalysts are closely related to the pressure. The catalytic activity, CH4 and CO2 conversion decrease with the pressure increased in CH4-CO2 reforming. Increasing the reaction temperature, prolonging reaction residence time and increasing CO2 and CH4 molar ratio, could restrain carbon deposition, increase methane and carbon dioxide conversion and improve carbonaceous catalysts stability. The carbonaceous materials surface area is attributed to enhancement of H-abstraction of methane and CO2 adsorption. As to the positive effect of the catalytic activity, carbonaceous catalysts oxygen-containing groups (C-O) may create active sites speeding up coke gasification by CO2.
Acknowledgements This work was supported by the National Basic Research Program of China (2005CB221202) and Shanxi Provincial Natural Science Foundation (2010011014-1).
References [1] S.Barama, C. Dupeyrat-Batiot, M.Capron, et al. Catalytic properties of Rh, Ni, Pd and Ce supported on Al-pillared montmorillonites in dry reforming of methane. Catalysis Today, 2009, 141: 385–392. [2] Zhigang Hao, Qingshan Zhu, Zheng Jiang. Characterization of aerogel Ni /Al2O3 catalysts and investigation on their stability for CH4-CO2 reforming in a fluidized bed. Fuel Processing Technology, 2009, 90: 113-121. [3] JunXia Wang, Yan Liu, TieXin Cheng, et al. Methane reforming with carbon dioxide to synthesis gas over Co-doped Ni-based magnetoplumbite catalysts. Applied Catalysis A: General, 2003, 250: 13–23. [4] K Nagaoka, K Takanabe, Ken-ichi Aika. Modification of Co/TiO2 for dry reforming of methane at 2MPa by Pt, Ru or Ni. Applied Catalysis A: General, 2004, 268: 151–158. [5] V A.Tsipouriari, X E.Verykios.Kinetic study of the catalytic reforming of methane with carbon dioxide to synthesis gas over Ni/La2O3 catalyst.Catalysis Today,
2001, 64: 83-90. [6] M Nagai, K Nakahira, Y Ozawa, et al. CO2 reforming of methane on Rh /Al2O3 catalyst. Chemical engineering science, 2007, 62: 4998~5000. [7] A I. Tsyganok, M Inaba, T Tsunoda, K Uchida. Rational design of Mg–Al mixed oxide-supported bimetallic catalystsfor dry reforming of methane. Applied Catalysis A: General, 2005 292: 328–343. [8] Guojie Zhang,Yue Dong, Meirong Feng, et al.CO2 reforming of CH4 in coke oven gas over coal char catalyst.Chemical Engineering Journal, 2010, 56(3) : 519-523 [9] Yongfa Zhang,A new technology of three products from coal based on a L T integrated apparatus,Proceeding of workshop on coal gasification for clean and secure energy for china, Beijing, 2003, 235-246. [10] Weidong Zhang, Yongfa Zhang. Study of CH4-CO2 reforming syngas with carbonaceous catalyst. Shanxi Energy and Conservation, 2009, 4: 34-37. [11] Meng Zhang, Guojie Zhang, Yongfa Zhang. Research technique of carbon deposite in CH4/CO2 reforming reaction process. Guangzhou Chemical Industry, 2009, 37(3): 58-60. [12] Yue Dong, Bingmu Zhang, Wenjun Zhao, et al. Research and development of a small – scale high temperature and high pressure reactor foe coal methanation. Coal conversion, 2010, 33(2): 64-67. [13] Jixiang Chen, Rijie Wang, Jiyan Zhang. Effeets of pressure on catalysis performance of catalyst for CH4 reforming with CO2. Chemical Reaction Engineering and Technology, 2005, 21(4): 365-369. [14] Guojie. Zhang, Yongfa. Zhang, Meng Zhang, Wei Zhao, Methane decomposition over carbonaceous catalyst and its kinetic characteristics. Modern Chemical Industry, 2009, 29(S1): 29-32. [15] Weidong Zhang, Yongfa Zhang. The catalytic effect of both oxygen-bearing functional group and ash in carbonaceous catalyst on CH4-CO2 reforming. Front. Chem. Eng. China, 2010, 4(2): 147–152.
PREDETERMINATION OF THE FAULT CROSSING THE UNDERGROUND COAL MINE GALERIES BY SEISMIC REFLECTION METHOD: AN APPLICATION AT A LONGWALL COAL MINE IN TURKEY G.G.U. ALDAS1*, A. CAN2 , B. KAYPAK 1, B. ECEVITOĞLU 1 1
Ankara University, Faculty of Engineering, Department of Geophysics, Ankara, Turkey
2
General Directorate of Mineral Research and Exploration, Ankara, Turkey
*
Corresponding author: [email protected] Tel: 90-312-2033395 Fax:90-312-2120071 In longwall mining operations, to provide continuous coal production, it is crucial to know the locations, types, dips, strikes and throw of the faults, before the operations. The aim of this study was to determine the existent faults and their important features in the longwall coal mining region, before starting the exploitation. Multi channel seismic reflection method, which is the most widely used and powerful geophysical method to view the tectonic structures, was used in the study. The method provides good results in determining the existent faults and their structures in the region. Seismic reflection, longwall mining, underground coal mining
INTRODUCTION This is a geophysical study conducted at a region where longwall coal mining operations are carried out. Longwall mining is an underground exploitation method used in fairly flatlying, thin, tabular deposits in which a long face is established across a panel between sets of entries and retreated or advanced by narrow cuts, aided by the complete caving of the roof or hanging wall. Because the exploitation openings are deliberately destroyed in the progress of mining, the caving class is truly unique. While the length of face is measured in hundreds of meters, the width of the working place is narrow, measured in meters. The longwall is kept open by a system of heavy-duty, powered, yielding supports that form a cantilever or umbrella of protection over the face. As a cut or slice is taken along the length of the wall, the supports retract, advance and reengage, allowing the roof to cave behind [1]. The existent tectonic structure of the region is very important for longwall mining. General trend in the longwall mining operations is to define the structure of the existent faults and joints while progressing in the mine galleries. The miners can observe the existent
tectonic only in the mine galleries. They cannot follow the faults in the whole panel. If the faults, encountered during the coal production, have considerably large throw, the miners should have to leave the existent gallery (because rotary cutting machine start to cut rocks instead of coal). This case is undesirable since the moving cost of the rotary cutting machine is very high (in terms of time and capital). Therefore, to provide continuous coal production, it is crucial to know the locations, types, dips, strikes and throw of the faults, previously. The aim of this study was to determine the existent faults and their important features in the longwall coal mining region, before starting the exploitation. The mine galleries (5mx5mx300m) were 300 m below the surface. Therefore, the target depth was 400 m. Length and width of the production panel was 3 km and 220 m, respectively. 80 cm cut or slice was taken along the length of the wall. In order to get image from 400 m depth, multi channel seismic reflection method was considered. In fact, the seismic reflection method is most widely used and powerful geophysical method to view the tectonic structures. In seismic reflection surveys, the travel times are measured of arrivals reflected from subsurface interfaces between media of different acoustic impedance [2-4]. GEOLOGY AND FIELD PATTERN OF THE STUDY AREA Figure 1 illustrates the 3D map of the underground coal production region. The line with letter HB (start of line) and HS (end of line) is the test profile (505 m). The blue line pieces are the faults crossing the mine galleries. The notations: PSJ.737, PSJ.738……, PSJ.743, and PSJ.744 are the drilling locations (for core sampling). In order to see the power of the reflection method, a test profile was designed such that it crossed five previously known faults. Figure 2 is the close view of the test profile. The numbers on the profile shows the CDP (common depth point) stations. The numbers at the end of the fault lines illustrate the throw amounts of the faults. „+‟ means upper block, „-„ means lower block of the fault. The numbers 1, 2, 3, 4 and 5 are the fault numbers, which are the same in the interpreted seismic stack section (Figure 5). The big elliptical points arranged in an order are the coal galleries.
Figure 1. 3D map showing the coal production sector.
Figure 2. The close view of the test profile in Figure 1.
At the seismic stack sections, the depth (m) was given in the form of time (ms). In order to make time to depth conversion and to make the correlation between the seismic events and real geological structures, synthetic seismograms should have to be produced. Figure 3 shows this kind of synthetic seismogram produced by using drill core data. It illustrates that, seismic
waves enter the coal layer at 0.256 s and get out at 0.259 s. The time spent in the coal layer is only 3 ms (see Table 1). The test profile was near the PSJ.741 drilling point. Table 1 demonstrates the parameters used in production of synthetic seismogram. The approximate velocities for the geological units provided by drills are found in the literature. The densities (d) and the seismic quality factors (q) are calculated by standard ampirical equations.
Figure 3. A synthetic seismogram produced by using PSJ.741 drill data.
Table 1. The parameters belong to the synthetic seismogram. i 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
t 0.043 0.094 0.119 0.136 0.162 0.192 0.205 0.214 0.231 0.245 0.256 0.257 0.257 0.258 0.258 0.259
r 0.05 0.26 -0.18 -0.37 0.14 0.11 -0.01 0.19 -0.02 -0.04 -0.29 0.03 -0.03 0.34 -0.34 0.34
h 52.00 66.00 49.00 26.00 21.00 30.00 15.00 11.50 27.10 20.90 16.70 0.40 0.20 0.50 0.55 1.15
v 2400 2600 4000 3000 1600 2000 2400 2350 3200 3100 2900 1800 1900 1800 3200 1800
d 2.17 2.21 2.46 2.29 1.96 2.07 2.17 2.16 2.33 2.31 2.27 2.02 2.04 2.02 2.33 2.02
q 62 73 172 97 28 43 62 59 110 103 91 35 39 35 110 35
Geological unit green claystone marl silisifiye limestone laminalı bitüm sandstone bitüm mudstone claystone tuff sandstone conglomerate volcanic sandstone lignite clayey-lignite lignite sandstone lignite
TECHNICAL EQUIPMENT AND FIELD PARAMETERS Seismic source is a Buffalo-Gun, simultaneously firing 36 Magnum cartridges [5]. Seismic Recording Instruments are provided by General Directorate of Disaster Affairs Earthquake Research Department. They consist two Seistronix RAS-24 modular units, 48 p-Geophones (28 Hz, total of 56 units), and cables. Off-End type field spread is used. The field parameters are given in Table 2.
Table 2. The field parameters Number of geophones Types of geophones Near offset Far offset Distance of geophones Distance of shotting Total profil length Sampling rate Recording time Number of shot
48 28Hz P-geophones 40 m 275 m 5m 10 m 505 m 1 ms 2s 24
SEISMIC DATA PROCESSING Our licensed W_Geosoft‟s Visual_SUNT v6.0 software package is used to process the seismic data. The processing sequences are [6-14]: 1- Static Correction: Static Correction is used to compensate for topography and lowvelocity surface layers effects. 2- Filter: Filters are used to eliminate high (such as wind noise) and low (such as industrial noises) frequency noises. First Band-Pass filtering implemented was in the range of 5-10-90-100 Hz. 3- AGC: AGC is used to compensate for the amplitude losses due to progressively increasing traveled-distances. Automated Gain Control time window implemented was 500 ms. 4- Mute: Mute is used to suppress First-breaks and Ground-rolls, leaving the reflections in the seismic data.
5- Geometry Definition: The shot and receiver coordinates for each shot were written to the seismic trace headers for subsequent data processing steps. 6- CDP Sort: Common-Depth-Point sorting of Shot-gather filed data should be done prior to Velocity Analysis. Data is sorted according to: First-Key: CDP, Second-Key: Offset. 7- Stack: Stacking provides the raw subsurface image. Sorted data is stacked following Velocity Analysis. 8- Filter: Noise introduced as a side-effect of Stacking Process was eliminated using a second Band-Pass filter in the range of 1-5-90-100 Hz. 9- Horizontal Smoothing: Horizontal Smoothing of a seismic section improves continuities of seismic horizons to ease the data interpretation. Trace-Mix was implemented with parameters 0.25-0.5-1.0-0.5-0.25. 10- Time-Cut: Time-Cut is used to trim the unnecessary portion of the seismic data. Recoding time was reduced to 1000 ms.
SEISMIC DATA INTERPRETATION The fault 3 and 5 can easily be seen in seismic stack section in Figure 4. The throw of the fault 5 was measured as 14.5 m in the mine gallery. The crushed zone of this strong fault leads to great decrease in signal amplitudes. Therefore, horizontal continuity was lost and a vertical corridor occurred in the section. The same case can be said for the fault number 3. Seismic horizons are determined in various colors in Figure 5. Since the sonic log is not available, it is very difficult to interpret which events on the synthetic seismogram belong to which geological units. Although the throws of the fault 1, 2 and 4 are small with respect to fault 3 and 5, they can easily be identified in the seismic section. The great negative reflection coefficient at the roof of the mine gallery (due to seismic wave travelling from rock to air) causes to reverse polarity in the seismic waves. Therefore, an upward movement of the traces occurred. The width of the mine gallery was 5 m. Therefore, the seismic wave, travelling 340 m from the surface, was diffracted from the gallery.
Figure 4. The seismic reflection section of the test profile (no interpretation). The horisontal axis is the distance and it is given in terms of CDP station numbers. The distance between CDP stations is 2.5 m. The vertical axis is two-way travel time.
Figure 5. Seismic reflection section of the test profile (with interpretation). The faults are shown by red lines. +- or -+ signs imply the fault throws and directions. The red numbers below the fault lines are the same with the numbers in Figure 2. The synthetic seismogram near the PSJ.741 drilling is placed on the seismic trace at CDP111 station. The crossing point of profile and mine gallery in Figure 2 is shown by orange arrow.
CONCLUSIONS Since the field topography was very rough and plant cover was very dense, the data acquisition step was very difficult. It was realized that, the seismic energy source Buffalo-Gun was very suitable for those kinds of hard-field work. 48- channel seismic equipment provided a good resolution to determine the fault. Therefore it is suggested that, this kind of seismic reflection study should be done at least 48-channel seismic equipment. As a summary, the seismic reflection method provides good results in determining the existent faults in the region. Therefore, especially in underground mining operations, this geophysical method increases the performance of the mine production and decrease the cost of man power and time.
REFERENCES 1- H.L. Hartman, “ Introductory Mining Engineering”, John Wiley&Sons, New York, 633,(1987). 2- M.E. Badley, “Practical Seismic Interpretation”, IHRDC, 266 p., (1985). 3- P. Kearey and M. Brooks “An Introduction to Geophysical Exploration”, Blackwell Scientific Publications, London, 254, (1991). 4- B.J. Evans, “ A Handbook for Seismic Data Acquisition”, SEG, 305 p., (1997). 5- L. Canyaran, and B. Ecevitoğlu B., “Yönlü sismik enerji kaynağı”, Türk Patent Enstitüsü, Ankara, Patent No. 2002/01203, (2002). 6- H.W. Waters, “ Reflection Seismology”, John Wiley and Sons, 453 p., (1981).
7- H.N. Al-Sadi, “ Seismic Exploration”, Birkhauser Verlag, 215 p., (1982). 8- P.M.Tucker and H.J. Yorston, “ Pitfalls in Seismic Interpretation”, SEG, 50 p., (1982) 9- R.E. Sheriff and L.P. Geldart, “Exploration Seismology: History, Theory, and Data Acquisition”, Cambridge University Press, 253 p., (1982). 10- R.E. Sheriff and L.P. Geldart, “Exploration Seismology: Data Processing and Interpretation”, Cambridge University Press, 221 p., (1983). 11- R.L. Sengbush, “ Seismic Exploration Methods”, IHRDC, 296 p, (1983). 12- O. Yilmaz, O. “ Seismic Data Processing”, SEG, 526 p., (1987). 13- C.S. Clay, “ Elementary Exploration Seismology”, Prentice-Hall Inc., 346 p., (1990).
14- D.G. Stone, “ Designing Seismic Surveys in Two and Three Dimensions”, SEG, 244 p., (1995).
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
GEOSCIENCE/COAL RESOURCES ALBERTA’S 2 TRILLION TONNES OF ‘UNRECOGNIZED’ COAL R.J. Richardson, P.Geol., President, CanZealand Geoscience Ltd. 250 Kepa Road, Mission Bay, Auckland 1071, New Zealand E-mail: [email protected]
Abstract: Historically and currently Canada is considered by the international energy industry to have minor coal reserves relative to those of the traditionally accepted major coal nations. Recent estimates by the World Energy Council (WEC) for proved recoverable reserves (that is, the tonnage of coal that has been proved by drilling etc. and is economically and technically extractable) has the United States at 238.308 Gt (gigatonnes or billion tonnes), Russian Federation at 157.010 Gt, China at 114.500 Gt, Australia at 76.200 Gt, India at 58.600 Gt while Canada is listed as having only 6.578 Gt; less than a 1% share of the world coal reserves and a little less than nations such as Poland (7.502 Gt), Brazil (7.059 Gt) and Columbia (6.814Gt). In Alberta (the Canadian Province having the largest coal resources) the Alberta Energy Resources Conservation Board ERCB) estimates the remaining established reserves (similar to WEC’s proved recoverable reserves category) of all types of coal in Alberta at December 31, 2008, to be 33.4 Gt. Of this amount, 22.7 Gt (or about 68 per cent) is considered recoverable by underground mining methods, and 10.8 Gt is recoverable by surface mining methods. In addition the ERCB recognize an ultimate potential of 620 Gt and ultimate in place coal resource of 2000 Gt (the Alberta Geological Survey using different methods put the estimate at a minimum of 2500 Gt). Alberta’s coal resources are vast and at 2000 Gt it is similar in scale to that of total coal resources of the United States. Coal data varies in quantity and quality from nation to nation and part of the definition of reserves involves economics and again changes markedly from place to place and from time to time. In addition there is geological complexity to consider in estimating reserves. Resource estimates can be even more wide ranging. Alberta does have a substantial amount of publically available geological data that can be used to establish coal resources. In addition to tens of thousands of shallow coal exploration holes and more than 17 thousand recent Coal Bed Methane (CBM) wells there are data from over 350 thousand oil and gas boreholes drilled in Alberta. The database grows by 15-20 thousand boreholes a year. Many of these oil and gas boreholes intercept deep coals and the coals that are readily indentified on geophysical logs. Industry geologists often use hundreds to thousands of these oil and gas boreholes to outline CBM plays. Almost all of the deeper coals are not extractable (economic) through traditional mining at present. However traditional coal mining is no longer the only option for coal. Thinking about coal has changed in recent years with new and improved technology for in-situ gasification paired with Carbon Capture and Storage (CCS) and with the current concerns of world depletion of oil and climate change. The recent international interest in in-situ gasification, surface gasification and associated energy technologies such as CTL, GTL, hydrogen and fuel cells has surfaced in a number of jurisdictions including Alberta.
Now with coal being in-situ mined up to 1400m depth in an Alberta coal gasification pilot project; Alberta’s huge coal endowment needs to be recognized.
TEMPERATURE AS A FACTOR AFFECTING ADSORPTION BEHAVIOR OF COAL TO LEAD (II) IONS Boleslav Taraba *, Petra Veselá, Roman Maršálek, Zuzana Navrátilová Ostrava University, Dept. of Chemistry, 30.dubna 22, 701 03 Ostrava 1, Czech Republic * E-mail: [email protected] Abstract Adsorption behavior of lead (II) ions on bituminous and subbituminous coals was investigated at temperatures 30, 60 and 80°C. Adsorption isotherms as well as adsorption kinetics were studied using batch adsorption experiments. In addition, flow calorimetric measurements were applied to direct determination of adsorption enthalpy. The shape of the adsorption isotherms indicated that the adsorption data could be tightly fitted by the Langmuir adsorption model. For both coal types, practically no change in adsorption capacity of the coals to Pb(II) ions was ascertained with increasing temperature. On the other hand, practically doublefold increase in the values of adsorption rate constant k was found when temperature rises from 30 to 60°C. The interaction of lead (II) ions with coal was then confirmed to be slightly exothermic process giving „net“ adsorption heat = 1 J/g for both coal samples.
1. Introduction The effect of temperature on adsorption of coals to gases like methane, carbon dioxide or oxygen is well established /1-2/. The adsorption is always an exothermic process and, thus, uptake of the physically adsorbed gas generally decreases with rise in temperature /1-2/. More complicated situation is for adsorption processes from aqueous solution where another types of interaction should be taken into account not occurring in gas-phase systems /3,4/. As a result, the effect of temperature on adsorption behaviour of compounds from the solutions is not quit elementary. In many cases, increase in the uptake with temperature rise has been reported /5-7/. Such a behaviour is in conflict with basic thermodynamic understanding of the adsorption process. In another cases, however, decrease in uptake of immobilized cations was found /8/. The paper tries to gain view on the adsorption behaviour of the subituminous and bituminous coals to lead (II) ions from solution with respect to the temperature effect.
2. Experimental 2.1 Coal samples Sample of subbituminous coal from the North Bohemian Coal Basin (denoted as SBC) and sample of oxidative altered bituminous coal from the Upper Silesian Coal Basin (denoted as OAC) were investigated. Elemental compositions and proximate analysis data of the coals are shown in Table 1.
Table 1: Proximate and ultimate analyses of the coal samples Sample Ash Stotal H (O+S)dif C
SBC OAC
dry basis (wt%,) 8.0 1.2 11.5 2.4
74.4 76.6
daf basis (wt%,) 6.5 18.1 4.1 17.5
N
1.0 1.8
Before measurements, grain size 0.2 –0.75 mm of the samples was leached with water, filtered and dried at 105°C. 2.2 Batch adsorption experiments and kinetics measurements Adsorption of Pb2+ ions (nitrate salt) from aqueous solution on the above samples was studied at 30, 60 and 80°C while kinetics at 30 and 60°C only. Experiments were conducted in the batch mode. About 0.5 g of dried sample was added to 50 mL of given solutions of different initial concentration co and shaken. For kinetics measurements, initial concentration of the Pb2+ ions was constant, co = 1 mmol/L. In the adsorption experiments, at least 7-day period of contact time was applied. In the end of the period, the pH value of each suspension was measured using a combination singlejunction pH electrode with Ag/AgCl reference cell. Then, the coal sample was removed by filtering through a paper filter. Metal concentration of filtered solutions was determined by means of the ICP optical emission spectrometry. All adsorption isotherms as well as kinetic curves were at least duplicated. 2.3 Calorimetric investigations Calorimetric investigations were performed using flow calorimeter FMC-4010 in the laboratory of MICROSCAL, Ltd. in London. Frontal mode was applied when water flow (3 ml/h) was changeover for flow of Pb(NO3)2 water solution at concentration 20 mmol/l. Weight of the coal sample for the measurements was about 0,12 g. 3. Results and discussion 3.1 Adsorption isotherms Figure 1 shows typical adsorption isotherm as obtained from the experimental data at 60°C for sample OAC.
-1
a [mmol.g ]
0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 0,0
1,0
2,0
3,0
4,0
5,0
6,0
7,0
8,0 -1
c [mmol.l ] Figure 1: Adsorption isotherm of lead (II) ions on sample OAC at 60°C The shape of isotherms indicates that the adsorption data could be fitted by the Langmuir adsorption model of monolayer coverage. In a linear form, the Langmuir equation is given by
c/a = c/am + 1/(am.K)
(1)
where a is the amount of the metallic ion adsorbed, c is an equilibrium concentration of the metallic ion in solution, K represents a monolayer binding constant and am is the monolayer adsorption capacity. Adsorption isotherms proved to be consistent with the Langmuir model as deduced from calculated r-square values close to 1. Basic parameters of Langmuir equation at studied temperatures were thus evaluated, see Table 2. Table 2: Parameters of Langmuir equation as evaluated from adsorption measurements am (mmol/g) Sample Temperature K SBC 30°C 0.087 + 0.001 1.7 + 0.2 60°C 0.086 + 0.001 3.3 + 0.3 2.8 + 0.5 80°C 0.086 + 0.001 OAC 30°C 0.66 + 0.05 31 + 15 60°C 0.65 + 0.03 52 + 13 80°C 0.69 + 0.02 25 + 18
Conformity of the adsorption data with Langmuir equation has often been reported /4,5/ indicating that Langmuir´s concept of monolayer surface coverage reflects immobilization of substances from solutions on coal surfaces. The most informative parameter in the Langmuir equation is am, providing information on adsorbed amount at monolayer surface coverage. From the Table 2, it is thus obvious that Pb2+ cations are immobilized on both coal to the same extent, irrespective on the temperature level. 3.2 Kinetic curves Within the three-hour interval of lead (II) adsorption, the rate of the process was found to follow a first-order reaction for both temperatures. In general form, it can be expressed as:
ln(c) = ln(co) – k.t
(2)
where c is the concentration of lead (II) at time t, co is the initial concentration of lead(II) and k is the rate constant at given temperature. The consistency of the adsorption kinetics with first-order reaction model was found to be very tight giving regression coefficients better than 0.98. Thus, from the slopes of the linear plots, values of the adsorption rate constant k could be calculated. Results are summarised in Table 3. Table 3: Rate constants as evaluated from kinetic measurements at 30 and 60°C Sample Temperature Rate konstant k (min-1) SBC 30°C (4,8 + 0,5)·10-4 60°C (8,4 + 2,5) ·10-4 OAC 30°C (3,2 + 0,7) ·10-3 60°C (5,7 + 1,5) ·10-3 The knowledge of the adsorption rate constants at different temperatures enables to determine value of the activation energy of lead (II) adsorption E, using Arrhenius equation. Thus, activation energies of 15.7 kJ/mol and 16.2 kJ/mol were found for sample of SBC and OAC, respectively. It is worth mentioned that such values of E are very close to that obtained by Kuhr et al. for cobalt adsorption on lignite /5/. 3.3 Adsorption enthalpy Using flow calorimetric approach it was possible to obtain “wonderful” calorimetric curves manifesting (i.o.) full reversibility of the adsorption/desorption cycle of lead (II) ions on coal surface, see Fig. 2. Reproducibility of the obtained adsorption heats was about 5%rel as deduced from triplicated measurements.
Figure 2: Adsorption and desorption cycle of Pb2+ ions on sample OAC, flow calorimeter FMC 4010, temperature of the measurements 25°C Heat effects of the „pure“ adsorption of the lead (II) on both studied coals were then evaluated to be 1 J/g. Thus, evident exothermicity of the immobilization process of lead (II) on the coals was approved. Acknowledgements Financial support of the GAAV project IAA 301870801 is gratefully appreciated. Autors thank management of Microscal, Ltd. (London) for helpful approach to enable calorimetric measurements in London and Dr. Mark Templer (Microscal, Ltd.) for efficient assistance during experimental data processing. References /1/ Yalcin, E., Durucan, S., , Methane capacities of Zonguldak coals and the factors affecting methane adsorption, Mining Science and Technology, v. 13, (1991) 215-222. /2/ Marsh, H., Heintz, E.A., Rodriguez-Reinoso, F.: Introduction to Carbon Technologies, Alicante Universidad, Spain, 1997. /3/ Radovic, L.R., C. Moreno-Castilla, and J. Rivera-Utrilla.. Carbon materials as adsorbents in aqueous solutions. Chem. Phys. Carbon (Ed. Ljubisa R.Radovic) 27, (2000) 227-405. /4/ Mattson, J.S., and H.B. Mark Jr.. Activated Carbon (Surface Chemistry and Adsorption from Solution). New York: Marcel Dekker, 1971 (ISBN 0-8247-1443-1). /5/ Kuhr, J.H., Robertson, J.D., Lafferty, C.J., Wong, A.S. and Stalnaker,N.D.: Ion exchange properties of a western kentucky low-rank coal. Energy and Fuels 11 (1997) 323-6. /6/ Qadeer, R. and Hanif,J.: Kinetics of zirconium ions adsorption on activated charcoal from aqueous solutions. Carbon 32 (1994) 1433-1439. /7/ El-Shafey, El., Cox, M., Pichugin, A.A., Appleton,Q.: Application of a carbon sorbent for the removal of cadmium and other heavy metal ions rom aqueous solution. Journal of Chemical Technology and Biotechnology 77 (2002) 429-436. /8/ Mohan,D., Gupta, V.K., Srivastava, S.K., Chander, S. Kinetics of mercury adsorption from wastewater using activated carbon derived from fertolizer waste. Colloids and Surfaces 177 (2001) 169-181.
Manuscript Not AVAILABLE
Authors.: Peter Fecko, Tana Kantorkova, Radomir Michniak,Lukas Koval, Alena Kasparkova [[email protected]] Title.: Bioflotation of coal This work deals with application of bioflotation on samples of brown and black coal. The bacteria Acidithiobacillus ferrooxidans werw used for the purposes and the effect of bacterial action (o-24 hours) on the results of flotation was observed. The acquired results imply that bacterial action manifested itself in a negative way on the black coal samples (drawn in the localities of Marcel mine and Zyglowice), in principle , there was a Loir concentrate yield and the ash and sulphur contents were higher. On the other hand, the results of bioflotation using brown coal samples drawn CSA Mine in Most brought positive results and i tis apparent that even short-term action of the bakteria, i.e. 30 minutes, is able to increase the concentrate yield and improve the concentrate quality, which means that both contents of ash and sulphur were lower.
SUITABILITY OF THE SULCIS COAL FOR CWS PREPARATION
Raimondo Ciccu, Professor., Dept. of GeoEngineering and Environmental Technologies – University of Cagliari Piazza D'Armi 19 – 09123 Cagliari – [email protected] Paolo Deiana, Dipartimento Energia, Sezione Impianti e Processi ENEA - Agenzia Nazionale per le Nuove Tecnologie, l'Energia e lo Sviluppo Economico Sostenibile - [email protected] Giovanni Mei, PhD Student, Dept. of GeoEngineering and Environmental Technologies – University of Cagliari Piazza D'Armi 19 – 09123 Cagliari – [email protected] Caterina Tilocca, Researcher, Dept. of GeoEngineering and Environmental Technologies – University of Cagliari Piazza D'Armi 19 – 09123 Cagliari – [email protected]
ABSTRACT: A Coal Water Slurry (CWS) containing about 65-70% solids by weight can be defined as a stable combustible mixture having a high heat power, in spite of the presence of water. A suitable particle size distribution accompanied by a small quantity of additives contribute to the obtainment of such properties enabling handling, trasportation, storage and combustion as a heavy fuel oil. The utilization of CWS offers a number of economic and environmental advantages concerning either the transport and storage operations (better management of coal stockpiles, lower intensity of traffic ….) and the final combustion (CWS can be the starting phase for the application of a number of clean coal technologies including gasification and hydrogen production) Aim of the research work dealt with in the paper is the study of the influence on suspension stability and rheologic characteristics (viscosity, thyxotropy) of various parameters such as: •
Particle size distribution
•
Type and dosage of a fluidising additive
•
Type and dosage of a stabilising agent
•
Proportion of coal in the mixture.
Optimum CWS should be characterised by a high heat value, good stability and a viscosity low enough for pipeline delivery. Results obtained through a systematic experimental programme, while confirming a better aptness of high.rank coals, have shown that also the sub-bitominous coal mined in the Sulcis coalfield is amenable to the preparation of CWS to be burned in the nearby power stations.
1. FOREWORD
Coal is yet again one of the most important sources of energy, since it covers about 26% of the world demand and generates up to 41% of electricity.1, with a contribution likely to raise in the future. In Italy too, in spite of an energy mix strongly unbalanced towards the use of natural gas, the employment of steam coal is growing, also because some oil-fired power stations are being converted to coal, while 13 already existing coal-fired plants are being restructured. The future of coal can be summarized by the sentence agreed upon by the most important Countries: Coal is an abundant resource in the world. It is imperative that we figure out a way to use coal as cleanly as possible 2, In compliance with this, the development of clean technologies for coal has been progressively promoted in the last years under the pressure of environment protection issues, attaining some important technical results in terms of pollution control, with the goal of reducing sulphur dioxide emissions by at least 80%, particulate matter by 75 % and nitrogen oxides by 60% while recovering 100 % of ash. Following this innovation progress, some important targets have been hit by the Italian power stations, witnessed by a 45% energy recovery from coal, compared with a European mean level of 39%. Since the last studies by the International Agency for Energy predict the attainment in Europe of an average efficiency of 42% by the year 2020, it can be said that Italy is 20 years ahead in the pursue of this goal. Among the Clean Coal Technologies (CCT), the Coal Water Slurry technology (CWS) can play an important role, owing to a number of advantages such as: the possibility of storing this kind of fuel without the problems posed by the management of coal storage heaps, the opportunity of delivering the coal over long distances via pipeline without the environmental impacts posed by airborne dust and by heavy traffic, and the chance to replace fuel oil in the power stations with minor technical adjustments. A CWS is a complex system where a number of parameters are involved concerning the definition of rheologic and flow characteristics of the mixture. The present work was aimed at assessing the feasibility of preparing CWS with the Sulcis coal, mined in the Nuraxi Figus district in Sardinia (Italy), classified as sub-bituminous with high sulphur content and thus considered unsuitable for that particular use. In the study, carried out at the DIGITA laboratories of the University of Cagliari in the frame of a co-operation project with ENEA, the National Agency for Energy and Environment, the results obtained with the “Sulcis” are compared with those obtained with a high-rank coal imported from Russia, used by the company Energy Coal for the preparation of CWS in its demonstration plant located in the industrial area of Oristano. The performance of CWS obtained with each of the two kinds of coal was investigated as a function of particle size distribution, kind and dosage of fluidizing and stabilizing agents.
1 2
World Coal Institute 2008 Energy Senate Confirmation Hearing January 13, 2009
2. EXPERIMENTAL PLAN 2.1 Coal samples The main characteristics of the two coals used for the tests are summarized in the following tables 1 and 2.
Table 1. Characteristics of the Sulcis coal determined with the LECO MAC 400 apparatus.
Table 2. Characteristics of the Russian coal made available by Energy Coal.
The heat value of the Sulcis coal (hand sorted sample) appears higher that that of the Russian coal (run-of-mine sample) due to a small content of ash-forming mineral matter.
2.2.Comminution Each coal sample was crushed in a jaw crusher and then dry ground in a rod mill in closed circuit with a control screen of 0.212 mm opening. Half of this product was further ground in similar way below 0.075 mm. The feed samples, “coarse” and “fine”, for CWS tests were obtained by blending in two different proportions (70:30 and 30:70, respectively) the two products of comminution. The corresponding particle size distribution are reported in figure 1. The histograms show a marked “bimodality” of the size distributions in compliance with the suggestions found in the literature in order to ensure better stability of the suspension with acceptable fluidity, taking also into account the energy content of the resulting CWS. Figure 1. Particle size distribution of the feed material for CWS preparation
2.2 Additives Basically two kinds of chemicals are used in the preparation of CWS: a fluidizing agent aiming at reducing the viscosity of the slurry for easier flow and injection into the boiler and a stabilizing agent for preventing settling along the delivery pipes and inside the storage reservoirs. The additives should be selected in order to:
• maximize the coal load (yield) of the slurry • provide better conditions for supply, handling and storage of CWS • meet the requirements for safety, health and environment • achieve economic advantages in the use of the fuel. The choice of the chemicals for the tests has been based on the findings of scientific research as well as on some considerations concerning their market price and availability. Following the above criteria, four agents have been selected: − Na–Esametaphosphate, a dispersing agent widely used in froth flotation for reducing the viscosity of dense pulps; −
Superwater®. a polyacrylamide soluble polymer that proved the best as a result of a broad study carried out at the
DIGITA’s Waterjet Laboratories in the frame of a European Project aimed at improving the performance of abrasive suspension jets. − Proxanol®, a copolymer of polythene and polypropylene ether behaving as a non-ionic surface active agent, used also in the medical field as softener and in general for providing “plasticity” to a variety of organic substances; − Rhodopol®, a polysaccharide-based substance having low viscosity, often supplied in the form of concentrated aqueous solution prepared using specific water-soluble anionic copolymers. The last two are currently employed in the preparation of CWS at the above mentioned Energy Coal demonstration plant. For the experimental tests each pair of chemicals (a stabilizer + a fluidizer) were added to the coal suspension in the form of solution in distilled water as follows: • Combination A (Proxanol + Rhodopol) • Combination B (Na–Esametaphosphate + Superwater). The additives were prepared in the form of solutions in distilled water that were dosed in the slurry in variable quantities with respect to coal as shown in table 3.
Table 3. Dosage of additives in the CWS as % of their solutions with respect to coal and in grams of dry substance per tonne of coal . Proxanol (0,3%) % g/t 0.65 455 0.96 672 1.27 889 1.61 1127 2.37 1659
Fluidizer Metaphosphate (1%) % g/t 2.00 200 3.06 306 4.19 419 6.60 660 11.55 1155
Stabilizer Rhodopol (0,02%) Superwater (0,3%) % g/t % g/t 4.00 8.00 0.27 8.1 6.04 12.08 0.42 12.6 8.19 16.38 0.58 17.4 10.44 20.88 0.90 27.0 15.06 30.12 1.56 46.8
2.3 Water In the experimental study distilled water in proportions variable from 37% to 50% was used for the coal suspensions. Since the importance of water quality for the performance of CWS is well known, the influence of the chemical composition of the water will be studied in a further development of the research.
2.4 Experimental procedure The experimental plan was conceived in order to study the features of CWS as a function of coal properties (reflected by its rank), with special reference to the Sulcis coal under variable experimental conditions. In particular the following aspects have been explored: • the particle size distribution • the water-to-coal ratio in the CWS • the type and dosage of a fluidizing agent • the type and dosage of a stabilizing agent • the stirring time. The structure of the testing plan is sketched in the block diagram of figure 2.
Figure 2. Structure of the testing plan
According to it, the influence of additives was investigated firstly at low density of the suspension. Then the density was increased by steps (each corresponding to a 5% increase of the proportion of coal in the mixture). Each series of tests was performed by simply increasing the additive dosage or adding new coal to the initial suspension, weighting 2 kg and having a coal proportion of 50%. On completion of each series, the procedure was repeated for a new sample of coal (different in type and size distribution).
At the end of every standard test, after about one hour stirring time, three samples were taken by means of a 50 cc syringe: one for the determination of sedimentation parameters, one for the measurement of rheologic properties and the third for checking purposes. Therefore for each of the four series of tests 10 triplets of samples were obtained: •
4 for the tests with increasing additive dosage at constant density
•
4 for the tests with increasing density while keeping the additive concentration at its maximum value
•
2 for the control tests with decreasing density (backwards path) again at constant peak additive concentration
After each sampling operation the density of the mixture was checked and, whenever necessary, a small quantity of water was added for compensating the evaporation losses. The three components of the system (water, coal and additives) were intimately mixed inside a stirring vessel having a capacity of 6 litres provided with a rotor fitting a vertical shaft to which a counter-rotating revolution motion is also applied by means of a cam (figure 3). Rotation velocity can be varied continuously. The following data have been obtained and processed: • settling curves for the evaluation of CWS stability • viscosity curves • thixotropy values Moreover, some considerations have been drawn concerning the heat value of the CWS. Figure 3. Mixing vessel
Figure 4. Viscometer
2.5 Data obtained 2.5.2
Viscosity curves
Rheologic parameters (viscosity and thixotropy) have been measured by means of a HAAKE RotoVisco 20 viscometer using the RheoWin 3 Software (figure 4). Viscosity has been obtained through the measurement of the tangential force per unit surface [Pa] as a function of the velocity gradient determined by the speed of the viscometer’s rotor from 5 to 200 s-1 over a total duration of the operation of 60 s. The asymptotic value of the curve has been assumed as representative of the fluidity of the mixture.
A typical curve of tangential tension and corresponding viscosity is shown in figure 5.
Figure 5 Curves of tangential tension and viscosity as a function of time with linearly increasing velocity gradient.
Figure 6 Curves of tangential tension and viscosity as a function of time after a complete up/down cycle of he velocity gradient.for thixotropy assessment
2.5.3 Determination of thixotropic characteristics Thixotropy can be defined as the capability of non-Newtonian fluids (pseudoplastic, Bingham, ….) to modify their viscosity when subject to shear actions or in the case of progressive agitation starting from quiet conditions. Under these circumstances a slurry can pass from a thick, almost solidified state to that of a fluid or, more generally, from gel to liquid. Therefore thixotropy reveals the inertia of a still suspension to flow freely when forced to move. Thixotropy has been determined by detecting the difference in the shear resistance in the two cases of increasing or decreasing velocity gradient dV/dr as a function of time, where V is the peripheral velocity of the viscometer’s rotor and r is the radial distance from the rotor’s rim. A measure of thixotropy is represented by the difference between the areas below each of the two branches of the curve, both over a base interval of 30 s. A typical curve giving the trend of shear resistance allowing the determination of thixotropy is shown in figure 6 together with viscosity as a function of time. 2.5.1 Sedimentation curves The stability conditions of CWS have been assessed through the observation of the displacement as a function of time of the meniscus separating the layer of clear water from the underlying settling suspension inside a graduated cylinder (figure 7). Settling velocity can be represented by the parameter t80 i.e. the time necessary for achieving 80% of the final volume of clear water in the sedimentation test. (figure 8).
Figure 7. Graduated cylinder for settling assessment
Figure 8. Typical sedimentation curve showing the t80 parameter
3. RESULTS AND DISCUSSION The observed values and tendencies allow to draw with enough clearness and reliability the most important considerations useful for the preparation of coal-water slurries. Based on experimental data the following aspects concerning the yield, as well as the stability and rheologic characteristics of the mixture under the simplified assumption of independency of the variables and parameters can be highlighted. 3.1 Influence of coal properties Generally speaking, the results obtained with the Sulcis” (younger, porous, oxidized, rich in volatile matters) compared to those with the Russian coal (higher in rank), put into evidence a better suitability of this latter, at least as far as stability is concerned, in agreement with the knowledge reported in the literature, in spite of its relatively high ash content, being it an untreated raw material (above 12% against the 7% of the hand picked Sulcis coal, actually averaging about 20% after gravity concentration of the run-of-mine). Table 3. Influence of size distribution and kind of additives on CWS stability for the two coals tested (lower values are for the higher additive dosage).
Additives Size Distribution Coal SULCIS RUSSIAN In fact:
FINAL VOLUME OF CLEAR WATER [%] Proxanol + Rhodopol Na-metaphosphate + Superwater Coarse Fine Coarse Fine 15 - 18 16 - 18
15 - 16 8 - 12
17 - 19 13 - 15
16 - 25 11 - 16
• the final volume of clear water in sedimentation tests is significantly lower for the Russian coal, as represented in table 3 giving the results at low density (coal content 50%) with varying kind and dosage of additives. Moreover the behaviour of the two coals depends on particle size distribution; • concerning the parameter t80 that describes concisely the kinetics of the settling process, it can be said that sedimentation is significantly slower in the case of the Russian coal for all the experimental conditions (table 4). Moreover sedimentation time is generally shorter in the case of coarser size distribution of suspended particles, as expected. Therefore, in agreement with the suggestions of the technical literature, the mixtures will be more stable if high-rank coals are used. • the viscosity of the slurry does not appear significantly influenced by the rank of the coal, at least in the range of low densities where it nears a level around 50 mPas for coarser size distributions and a 30 mPas for finer ones with both kinds of coal tested. Table 4. Influence of size distribution and kind of additives on settling time for the two coals tested
Additives Size Distribution Coal SULCIS RUSSIAN
SEDIMENTATION TIME t80 [h] Proxanol + Rhodopol Na-metaphosphate + Superwater Coarse Fine Coarse Fine 0.4 - 0.7 1.0 - 1.8
0.5 - 1.5 1.6 - 3.5
0.2 - 0.7 1.5 - 2.2
0.6 - 1.5 1.5 - 3.5
• similar considerations hold also for thixotropy: there is not any substantial difference in rheologic behaviour between the Sardinian and the Russian coals, since the average values for this parameter are basically the same at equal particle size distribution. It is trivial to remind that stability features of industrial CWS can be improved by means of a moderate stirring of the slurry inside the storage vessels. 3.2 Influence of slurry density As the density of the mixture is increased by adding new coal, the final volume of clear water above the settling suspension diminishes. The waiting time for the attainment of stationary conditions varied from case to case as witnessed by the parameter t80 but after about 6 hours an asymptotic level was closely approached and the suspension remained practically stable in all cases. In the tests with the combination of additives A (Proxanol + Rhodopol), starting from an initial value around 16-18% for the 50:50 mixtures, the final volume of clear water drops down to 2-4% (the higher value is for the coarser size distribution) in correspondence to the top density (about 63% of coal in the slurry). An inexplicable deviation from his rule has been observed for the Russian coal that exhibited a substantially stable performance (6-8%) even at low densities. A similar behaviour is registered also with the additive combination B (Metaphosphate + Superwater), although in presence of some sporadic divergence from the decreasing trend, probably due to unknown experimental anomalies. Concerning the sedimentation kinetics, the experimental curves put into evidence the presence of a minimum value for the parameter t80 corresponding to an intermediate density (around 55% coal), generally lower for the coarser size
distributions. A diverging behaviour from the general trend, difficult-to explain, was exhibited by the Russian coal “fine” with additive combination A and by the Sulcis coal “coarse” with additive combination B. Particularly important, for practical purposes, are the curves representing the viscosity of the slurry as a function of its density since they allow to single out the peak yield achievable for a given coal corresponding to a viscosity limit of about 1000 mPas, beyond which the fluid consistency is rapidly lost, making it impossible to handle the slurry as a pumpable liquid. Such limiting viscosity is reached with a yield different for the two kinds of coal: slightly higher for the “Sulcis” with respect to the Russian coal (a surprising outcome!). The extrapolation of the curves to 1000 mPas show also that the yield is lower for the finer size distribution (according to expectation). The better yield for the “Sulcis” could be explained by the lower ash content of the sample used for the tests as pointed out earlier. The curves at maximum additive dosage for the determination of thixotropy as a function of density reflect strictly those for viscosity and they almost overlap each other except for a different scale. 3.3 Influence of coal comminution Stability parameters are strongly influenced by the fineness of solids in the slurry in a sense that mixtures composed by coarsely ground coal are less stable and settle more quickly, the other conditions being the same. This fact is substantiated by the data reported in Tables 3 and 4 concerning the influence of coal characteristics. Also rheologic parameters of the slurry are affected by particle size distribution, given that the limiting value for viscosity as density increases is reached earlier in the case of finer comminution. Therefore it is worthless to force too much the grinding operation, since the advantage of a better stability are outbalanced by a sharp deterioration of fluidity and thence by a lower yield of the mixture, besides the increase of CWS preparation cost. The use of additives can be helpful in the attempt to improve the performance, although being unable to reverse the trend. 3.4 Influence of chemical additives. The influence of additives has been firstly studied in conditions of dilute suspensions (coal proportion: 50%) in the attempt to isolate the results from the concurrent influence of density and size distribution that could have “masked” the effect of additives. The following aspects can be underlined concerning the stability features of the suspension: •
the final volume of clear water does not appear to depend considerably on the concentration of additives within the
range studied: actually the curves obtained are almost flat and the expected reduction in solids/liquid segregation did not take place as dosage was increased. •
the trend of the kinetics parameter t80 as a function of additive dosage resulted to be poorly consistent for the different
experimental conditions, thus providing no contribution to a better understanding of the mechanism of additive action.
•
However if the attention is concentrated on the average values it emerges that the combination of additives A (Proxanol
+ Rhodopol) is slightly more advantageous than combination B (Metaphosphate + Superwater) for which sedimentation time is somewhat longer as already shown in Table 4. No significant contribution to a clear comprehension of the effect of additives was obtained by the consideration of the rheology parameters in conditions of dilute suspension. In fact: • the curves of viscosity as a function of additive dosage are almost flat with episodic and little significant deviations. However viscosity seems to be influenced by the different combination of additives as shown in table 5 giving the span of variation for viscosity within the range of additive concentration explored: the pair “Proxanol-Rhodopol” seems always more efficient in the control of viscosity than the pair “Metaphosphate-Superwater”. Table 5. Influence of size distribution and kind of additives on the CWS viscosity
Additives Size Distribution Coal SULCIS RUSSIAN •
VISCOSITY [mPa s] Proxanol + Rhodopol Na-metaphosphate + Superwater Coarse Fine Coarse Fine 25 - 30 35 - 40
40 - 50 45 - 55
30 - 32 36 - 40
55 – 65 58 - 62
•rather different considerations can be put forward concerning thixotropy for which the curves exhibit the presence of a
minimum value corresponding to the intermediate dosage of additives, followed by an increase at higher dosages; this behaviour is observed for both kinds of additives and for both types of coal as shown in table 6 where the increment of thixotropy starting from the minimum value is reported. •
however, while the maximum value is always observed for the higher dosage of additives, the minimum point shifts
rightwards for the Russian coal with respect to the “Sulcis”. The general trend of thixotropy against additive dosage seems to indicate a certain interaction between the fluidizing and the stabilizing agent (the effect of the first prevails at the lower concentrations of additives whereas the second overcomes at higher dosages). The fact that thixotropy increases while viscosity remains constant is certainly a positive aspect for the preparation and utilization of CWS. Table 6. Influence of particle size and kind of additives on CWS thixotropy (into brackets the relative increment with respect to the minimum value)
Additives Size Distribution Coal SULCIS RUSSIAN
THIXOTROPY [Pa s] min - Max Proxanol + Rhodopol Na-metaphosphate + Superwater Coarse Fine Coarse Fine 3.2 – 7.3 (1.28) 3.5 – 8.0 (0.45)
5.2 – 6.5 (0.25) 8.0 – 9.7 (0.21)
1.1 – 8.6 (6,19) 6.0 – 10.5 (0.75)
4.2 – 10.5 (1.5) 8.2 – 16.6 (1.0)
The table shows also that the increment in thixotropy is higher for the coarsely ground Sulcis coal which derives benefits from the addition of the additives, especially “Superwater”, gaining a substantial improvement in performance, albeit starting from a lower minimum value with respect to the Russian sample, that however always holds higher thixotropy values.
3.5 Effect of mixing time In the course of each series of experimental tests it emerged that: •
working at low density, the addition of a fluidizing agent had a poor effect on viscosity that remained practically constant while the dosage was increased
•
working at high density, the sampling operation with the syringe became more and more difficult as the cumulative stirring time elapsed.
It was retained likely that a prolonged stirring could have caused a progressive shifting of particle size distribution towards the fine range, thus increasing the viscosity of the slurry and upsetting the effect of the additive. In order to prove the correctness of this assumption, a series of long-duration tests has been carried out on a medium density mixture (60% solids) with the “coarse” sample of Sulcis coal in absence of additives. It came out that a prolonged stirring time even exceeding 10 hours did not produce any significant change of CWS rheology. The observed variations are very small and can be due to possible changes of experimental conditions (fluctuation of temperature, release of ions in the solution, ….). Also the variations in the final volume of clear water did not point out the onset of any phenomenon in some way imputable to the time (and energy) of mixing. In order to cast some light on this aspect, two samples were taken and analysed, at the beginning and after 11 hours of continuous stirring. Wet screening at 53 μm gave the results reported in table 7. Table 7. Variation of particle size after 12 h stirring time for a sample of Sulcis coal.
Size classes + 0.053 mm - 0.053 mm Total
Mass of size fraction [%] Initial sample Final sample 56.86 53.04 43.14 46.96 100.00 100.00
The small difference in weight between oversize and undersize does not fully elucidate the problem. Therefore it can reasonably be assumed that the observed results might be due also to a certain inertia in the interaction process between the additives and the suspension. It is also likely that some slow-evolution phenomena may occur, resulting in the liberation of inorganic or organic substances present in the coal and subsequent modification of the chemistry of the solution.
4. CONCLUSIONS
Experimental results show that it is possible to produce good CWS with the Sulcis coal, up to now considered unsuitable for this application. However additional research efforts must be done aiming at further increasing the yield of the slurry and thus improving the economic feasibility of the operation. The data obtained represent a sound starting base for the development of the technology through:
•
a broader choice of additives in order to single out the most appropriate combination of a fluidizing and a stabilizing
agent and optimizing the respective dosage •
a careful control of water quality and chemistry of the solution
•
a deeper insight into the comminution process (size distribution and operational procedure and conditions)
•
the assessment of the optimum degree of coal cleaning (ash content) and the study of the kind of washing operations
before the CWS preparation (presence of flotation reagents, dewatering, …..). Moreover the results obtained can encourage a feasibility study concerning the construction of a CWS plant near the Carbosulcis mine from which the fuel can be transported via pipeline to the nearby power station (about 3 km away) where it can be burned after minor modifications to the boiler. As an alternative, the CWS can be used for feeding a coal gasification plant often proposed for the environmentally friendly utilization of the high-sulphur Sulcis coal (5-6%).
ACKNOWLEDGEMENTS Work carried out in the frame of a project financially supported by ENEA REFERENCES
1. Agus, M., Carbini, P., Ciccu, R., Ghiani, M., Satta, F., Tilocca, C., “Influence of coal properties in surface-based separation processes”, Proc. International Coal Preparation Congress, Crakow 1994. 2. Agus, M., Carbini, P., Ciccu, R., Ghiani, M.: “Characterisation of some high-sulphur, low-rank coals, Processing and Utilization of High Sulphur Coals IV” Ed. Dugan, Quingley and Attia, Elsevier Sci. Publ. Amsterdam, 1991. 3. Aktaş Z., Woodburn E.T., “Effect of addition of surface active agent on the viscosity of a high concentration slurry of a low-rank British coal in water”, Fuel Processing Technology 62 (2000) 1–15. 4. Allardice D.J. , L.M. Clemow, G. Favas, W.R. Jackson, M. Marshall, R. Sakurovs, “The characterisation of different forms of water in low rank coals and some hydrothermally dried products”, Fuel 82 (2003) 661–667. 5. Ananth P. Chikkatur, “A Resource and Technology Assessment of Coal Utilization in India”, Kennedy School of Government, Harvard University, Cambridge, MA. 6. Anderson C.M., Musich M. A., Young B. C., Richter J. J., DeWall R. A., Ness R. 0., “Wiang Haeng Coal-Water Fuel Preparation and Gasification, Thailand - Task 39”, Topical Report July 1996 For U.S. Department of Energy Office of Fossil Energy Morgantown Energy Technology Center P.O. Box 880 Morgantown, West Virginia 26507-0880. 7. Atesok G. , Boylu F. ,Sirkeci A.A. ,Dincer H. , “The effect of coal properties on the viscosity of coal–water slurries”, Fuel 81 (2002) 1855–1858. 8. AA VV “The future of coal A interdiscilinary study” 2007 Massachusetts Institute of Technology (MIT). 9. Banford A.W., Aktas Z., “The effect of reagent addition strategy on the performance of coal flotation”, Minerals Engineering 17 (2004) 745–760. 10.Boylu F. , Ateşok G., Dinçer H., “The effect of carboxymethyl cellulose (CMC) on the stability of coal–water slurries”, Fuel 84 (2005) 315–319.
11.Boylu F., Dincer H., Atesok G., “Effect of coal particle size distribution, volume fraction and rank on the rheology of coal–water slurries”, Fuel Processing Technology 85 (2004) 241– 250. 12.Boylu F., Dinçer H., Ateşok G., “Effect of coal particle size distribution, volume fraction and rank on the rheology of coal – water slurries”, Fuel Processing Technology 85 (2004) 241–250. 13. Carbini P. , Ciccu R. , Ghiani M. ,Satta F. ,TiloccaC., “Flotation of coal fines using high-shear water jets”, Proc.13th International Coal Preparation Congress, Vol. I, Brisbane 1998. 14. Chen L., Wang Q., Ren Y., Wang Y., Duan Y: Rheology and wall slip effect of high concentrated coal water slurries, South East University, China. 15.De Lorenzi, L. Bevilacqua P.: “The influence of particle size distribution and non ionic surfactant on the rheology of coal water fuels produced using Iranian and Venezuelan coals”. International Journal of Coal Preparation and Utilization, 1939-2702, Volume 22, Issue 5, 2002, Pages 249 – 268 16.Dinçer H., Boylu F., Sirkeci A.A., Ateşok G., “The effect of chemicals on the viscosity and stability of coal water slurries”, International Journal of Mineral Processing 70 (2003) 41–51. 17.Fuerstenau D. W., “Rheology of Coal-Water Slurries Prepared by the HP Roll Mill Grinding of Coal”, Quarterly Technical Progress Report No. 4 June 1- August 31, 1993. 18.IEA Statistics, “Coal Information 2007 -with 2006 data”. 19.Laskowski J.S. , Parfitt G.D. , “Electrokinetics of coal–water suspensions, Interfacial Phenomena in Coal Technology”, Surfactant Science Series, New York, 1989. 20.Leonard J.W. “Coal Preparation”, American Institute of Mining, Metallurgical and Petroleum Engineers, New York 1979. 21.Marvin D. C., Frankiewicz T. C., “CWS Rheology: The Role of the Coal Particle”, Occidental Research Corporation 2100 S.E. Main Street Irvine , California 92713. 22.Papachristodoulou G., Trass O., “Coal slurry fuel technology”, Canada Journal of Chemical Engineering 65 (1987) 177–201. 23.Roh N.S., Shin D.H., Kim D.C., Kim J.D., “Rheological behavior of coal – water mixtures”, Fuel 74 (1220-1225) (1995) 1313–1318. 24.Tiwari K.K., Basu S.K. , Bit K.C. , Banerjee S. , Mishra K. , “High-concentration coal–water slurry from Indian coals using newly developed additives”, Fuel Processing Technology 85 (2003) 31–42. 25.Umar D.F., Usui H., Daulay B. , “Change of combustion characteristics of Indonesian low rank coal due to upgraded brown coal process”, Fuel Processing Technology 87 (2006) 1007–1011. 26. Vitolo S., Belli R., Mazzanti M. , Quattroni G. , “Rheology of coal –water mixutures containing petroleum coke”, Fuel 75 (1996) 259–261. 27.Zhaobing Guo, Ruo Feng, Youfei Zheng, Xiaoru Fu, “Improvement in properties of coal water slurry by combined use of new additive and ultrasonic irradiation”, Ultrasonics Sonochemistry 14 (2007) 583–588.
COMPARATIVE STUDY OF OIL AGGLOMERATION AND FLOTATION OF LOW GRADE COALS Feridun BOYLU, Fırat KARAKAŞ Istanbul Technical University, Faculty of Mines, Mineral Processing Engineering Department, 34469, Maslak, Istanbul, TURKEY [email protected]
ABSTRACT In this study, the collector, also called as the bridging oil, used in both oil agglomeration and flotation is introduced to test in the form of emulsion. Kerosene, Fuel oil and Diesel oil were used for both flotation and oil agglomeration as bridging oil or collector. In flotation experiments, MIBC, DTAB and n-Decanol were also used as frother, and emulsifiers respectively. In agglomeration tests, it was found that diesel oil was superior to kerosene and fuel oil according to the combustible recoveries and ash contents of the agglomerates. The emulsification of the bridging oils resulted in enhancement on combustible recoveries from 40 % to 75 %.and the reduction on the ash contents of original coal samples from 35 % to 19 % In flotation tests with emulsified diesel oils of 8000 g/t based on the coal weight, the clean coal products with approximately 15 % ash contents and 60-70 % combustible recoveries were obtained. Finally the oil agglomeration and the flotation of low grade coal were compared and it was found that the oil agglomeration methods should be announced as superior method to flotation for providing higher combustible recoveries and lower ash contents. Keywords: oil agglomeration, flotation, release test, lignite, floatation index.
INTRODUCTION
Nowadays, fine coal production increases continuously due to the new mechanization techniques on coal mining and the amount of fine coals have been getting higher in all steps of the coal processing such as excavation, transportation and handling depending on the coal nature. The fine coal particles usually contain large amount of mineral matter and the utilization of these coal fines required to washing processes. There are several fine coal cleaning techniques including gravity based separation, froth flotation and oil agglomeration techniques, although the last technique did not find any industrial application.
In the oil agglomeration or selective oil agglomeration techniques, addition of immiscible liquid in to coal water mixture causes better adhesion among coal particles by increasing capillary interfacial forces. The oil agglomeration process depends on difference between the surface properties of organic and inorganic particles. The carbonaceous constituent of fine coals can be agglomerated from the mixture in the presence of immiscible oils known as collecting or bridging liquids. Mineral-rich matter normally remains in the mixture and is rejected (Steedman and Krishman, 1987). In the literature, it is clearly indicated that the agglomeration characteristics of coals depend so many factors such as particle wetting characteristics (surface tension and three phase contact angle etc.), the dosage and type of bridging oils, coal rank, particle size (Capes et al, 1974; Slaghuis and Ferreira, 1987; Abdelrahman and Brookes, 1987; Kawashima et al, 1989; Petela et al, 1995; Wheelock et al, 1994; Shrauti and Arnold, 1995; Alonso et al, 1999; Darcovich et al, 1989; Garcia et al, 1998; Aktaş, 2002). When it is compared to other related methods it is clearly seen that oil agglomeration of coals has an advantage considering both upgrading and size enlargement of coal fines This method, however, can has an application only for the high grades coals, which give proper surface properties, since the low grade coals cannot response well and higher bridging oil concentrations are necessary. The flotation of low grades coals also shows the similar trend to oil agglomeration. The last technology, however, makes the oil agglomeration and flotation of low grade coals possible by the utilization of surfactants as emulsifiers and promoters (Laskowski and Yu, 2000; Abdelrahman and Brookes, 1987; Cebeci and Eroğlu, 1998; Capes et al, 1985; Baruah et al, 2000). Emulsifying of bridging oils or utilization of promoters reduces the bridging oil consumption in both flotation and oil agglomeration. In froth flotation, the hydrophobic coal particles are removed selectively from suspension by attachment to rising gas bubbles in water. For fine coals, this is one of the best methods to separate the carbonaceous materials from the mineral matter. The flotation characteristics depend on the similar parameters to agglomeration and generally the flotation characteristics is determined applying some conventional methods such as release and tree tests. Release test gives the theoretical floatation behavior (limits for the combustible recoveries and ash contents) and inherent ash contents of coal. In this study detailed agglomeration tests using different hydrocarbons and emulsifiers and release tests for the froth flotation were performed and these two methods were compared to each other.
EXPERIMENTAL Material and Method A low rank sub-bituminous coal (LS 20) provided by the Luscar Sterco mine, Alberta, Canada, was used in the tests. The sample was crushed and then pulverized below 150 µm (d90), and kept in plastic bags for flotation and agglomeration studies. The proximate and the X-ray analyses of the coal sample are given in Table 1 and Table 2 respectively. Table 1. Proximate analyses of the tested coal COAL Moisture, % Ash, %
LS-20
3.64
34.6
Volatile matter, % 27.48
Volatile matter, d.m.m.f., % 44.5
Fixed carbon d.m.m.f., % 55.5
Table 2. XRAY analyses of the mineral matter of LS-20 cal Quartz Muscovite Kaolinite Gypsum Calcite Total
Ideal Formula SiO2 KAl2.AlSi3O10(OH)2 Al2Si2O5(OH)4 CaSO4·2H2O CaCO3
Coal LS-20
42.2 32.3 16.9 3.0 5.6 100.0
In agglomeration and flotation tests; kerosene, fuel oil and diesel oil with their original and emulsified forms (as collector and bridging oil) and MIBC (methyl iso-butyl carbinol), 2-ethyl hexanol (ETHXNL) as frother were used. In emulsification process, Dodecyl trimethyl ammonium bromide (DTAB), n-decanol (DCNL) and octadecyl ammonium bromide (ODA) were used as emulsifying agents. Agglomeration test were carried out by using 1 liter cylindrical cell with 6 buffer parts and impeller. Prior to agitation in the agglomeration cell, the coal sample and bridging oil was blended at 10000 rpm for a minute in order to provide rapid agglomeration. After that, the pulps containing the agglomerates were poured into agglomeration cell in where they are agitated for different periods to enhance the agglomeration process by drainage of the impurities that can be found in agglomerates. Following the agglomeration process, the pulps were poured onto screen with the openings of -500 microns. The agglomerates accumulate on the top of the screen while impurities pass it; and so they were collected separately and were washed 3 times by using acetone and filter paper.
In froth flotation stage, release tests, were performed with 2L Galigher Lab flotation cell (Model LA-500) as in our earlier study (Boylu and Laskowski, 2008) using the procedure described by Dell (1953).
Results and Discussion Agglomeration tests In the first step of the agglomeration tests, the agglomeration time was investigated using only diesel oil at constant experimental conditions of 10 % diesel oil, 5 % solids ratio, 10000 rpm stirring periods for 60 seconds at neutral pH. 40 min agitation time (at 1000 rpm) was found as an optimum agglomeration time (Fig 1a). Further agglomeration tests were carried out using three kinds of hydrocarbons namely kerosene, diesel and fuel oil in different amounts to make a comparison among them (Figs 1b and 1c). As it is seen from Figs 1b and 1c, although the ash contents of the agglomerates were close to each other, ranged from 19 % to 21 %, the higher combustible recoveries were obtained using heavy bridging oils (diesel oil and fuel oil) when compared to light hydrocarbon of kerosene. That is, for the bridging oil concentrations of 10 %, the combustible recoveries were recorded as 84, 69 and 38 % for diesel oil, fuel oil and kerosene, respectively. This result is most likely about the binding mechanisms working depending on the bridging oil characteristics. It was thought that kerosene keep the coal particles in comparatively weaker agglomeration forms of pendular and funicular while diesel oil makes funicular and capillary forms of agglomeration which results in keeping the coal particles together more strictly when water spraying process to drain the noncarbonaceous minerals from the agglomerates. In addition, fuel oil which is heavier than the diesel oil does not agglomerate the coal particles as much as diesel oil. This situation is caused by structure of diesel oil that has much more bridging effect. Further agglomeration tests where the bridging oil was introduced as emulsified form were carried out using diesel oil. Emulsification process was performed as described in our early study (Boylu and Laskowski, 2006) using DTAB, ODA and DCNL. Figs 1d and 1e illustrate the effect of diesel oil emulsified with different emulsifiers. In these figures, DTAB was not included since the use of DTAB deteriorates the bridging effect significantly and thought not to be examined in detail. The use of emulsification process enhanced the agglomerating capacity (combustible recovery, see the fig 1d) of diesel oil in the range of 5-10 % for the oil dosages higher than 10 %. With emulsification, although no significant enhancement was observed on the quality of agglomerates,
the ash contents of tailings was increased to 75 % from 63-64 %. The emulsification process caused to thinning the bridging oil which resulted in decrease in oil consumption and higher selectivity since the decrease in bridging oil viscosity allowed to noncarbonaceous minerals to be drained easily. 90
27
(A) Comb. Rec.
25
Ash Cont. 86
23
84 21 82 19
80 78
17
76
15 0
10 20
30 40 50 60
Ash Cont. of Aggl., %
Comb. Recovery, %
88
70 80 90
Agglomeration Time, min. 100
35
(B)
(C)
33
80 70 60 50 40
Stirring time 40 minute Coal concentration 5 % Kerosene only Diesel oil only Fuel oil only
30 20 10 0 0
5
10
15
20
25
29 Stirring time 40 minute Coal conc. 5 %
27 25 23 21 19 17 15
30
0
5
Oil Dosage, %
10
15
20
25
30
Oil Dosage, %
100
80
(D)
90
(E)
70
80
Ash Contents, %
Comb. Recovery, %
Kerosene only Diesel oil only Fuel oil only
31
Ash Contents, %
Comb. Recoveries, %
90
70 60 50 40 30 20
Diesel oil only Diesel oil + 0.1 % ODA
10
Diesel oil + 0.5 % DCNL
60 Diesel oil only (C) Diesel oil only (T) Diesel oil + 0.1 % ODA (C) Diesel oil + 0.1 % ODA (T) Diesel oil + 0.5 % DCNL (C) Diesel oil + 0.5 % DCNL (T)
50 40 30 20 10 0
0 0
5
10
15
20
Oil Dosage, %
25
30
0
5
10
15
20
25
30
Oil Dosage, %
Fig 1. Agglomeration Test Results; effect of agglomeration time (a), effect of bridging oil type based on the combustible recovery (b) and ash content (c), effect of emulsifiers based on the combustible recovery (d) and ash contents (e).
Flotation (Release) Tests In flotation tests, the collector and frother dosages were examined to determine the optimum separation and release test conditions; the results were illustrated in Figs 2 a-e. As seen from figs 2 a and b, an increase on collector dosages resulted in better flotation and release test conditions especially when ETHXNL was used as frother. This is most likely because of the nature of used coal since its qualification degree is low (sub-bituminous). Sub-bituminous coals, because their surface are partly surrounded by oxygen containing functional groups, need much more neutral hydrocarbons to be floated when compared to bituminous coal. The floating ability of the used coal in low collector dosages in case of MIBC was attributed to the nature of the frother that some frothers exhibit collecting effect as well. When both collecting effect (see combustible recoveries) and selectivity at the collector dosages of 8000 g/t were considered, 2-ethyl hexanol was found to be superior as it is clear in Figs 2 c-e. Although the inherent ash contents were closer at around 11.5-12 % (see the interpretation of the curves at x axis), Fig 2e in where the comparison of two frothers based on the flotation index proposed by Mohanty et al (1999) were illustrated points out the advantage of 2-ethyl hexanol utilization.
Comparison of Agglomeration and Flotation Test Results of LS-20 Coal Flotation is one of the most important separation methods for fine/ultrafine coals. The flotation characteristics of coals are generally determined by using release test (alternatively tree tests) giving the theoretical cumulative combustible recovery and ash contents of clean (floated) products and also inherent ash contents. Although there is no any industrial practice, oil agglomeration is also one the best method for the fine coal separation. In this study agglomeration and flotation characteristics of LS-20 coal were determined at optimum conditions and compared in Fig 3. As seen from Fig 3, it has been found that, for clean coal production with high combustible recoveries, oil agglomeration was superior to the flotation method. That is, by using the flotation technique to produce the clean coal with ash contents lower than 15 %, yield the combustible recoveries in the range of 70-80 %, while the same quality clean coal products are enable to be produced with the combustible recovery of 90 % by using oil agglomeration. This advantage of oil agglomeration on combustible recoveries is getting much more important especially for the clean coal product with ash content of lower than 13%.
100
LS-20 Release Analyses @ 60 g/t MIBC
90 80
Cum. Comb. Recovery, %
Cum. Comb. Recovery, %
100
70 60 50 40 30 20 10
4000 g/t EDO 6000 g/t EDO 8000 g/t EDO
(A)
15 20 25 30 Cum. Ash Content, %
35
6000 g/t EDO
70
8000 g/t EDO
60 50 40 30
LS-20 Release Analyses @ 60 g/t ETHYLHEXANOL
20
(B) 10
100
(C)
90
80 70 LS-20 Release Analyses @ 8000 g/t EDO
60 50 40
20 g/t MIBC
30
40 g/t MIBC
20
60 g/t MIBC
10
15
20
25
30
35
Cum. Ash Content, %
Cum. Comb. Recovery, %
Cum. Comb. Recovery, %
80
0
10
90
4000 g/t EDO
10
0
100
90
(D)
80 70 LS-20 Release Analyses @8000 g/t EDO
60 50 40
20 g/t ETHYLHEXANOL
30
40 g/t ETHYLHEXANOL
20
60 g/t ETHYLHEXANOL
10
80 g/t MIBC
0
80 g/t ETHYLHEXANOL
0 10
15 20 25 30 Cum. Ash Content, %
35
10
15 20 25 30 Cum. Ash Content, %
35
60 LS-20 Release Analyses @8000 g/t EDO
58
Flotation Index
56 54
FI 1 ( MIBC)
52
FI 1 ( ETHXNL)
50
FI overall ( MIBC)
48
FI overall (ETHXNL)
46 44 42
(E)
40 0
20
40
60
80
100
Frother Dosage, g/t
Fig 2. Flotation (Release) Test Results; Effect of collector dosages for MIBC (a) and ETHXNL (b), Effect of frother dosages for MIBC (c) and ETHXNL (d), Comparison of the frothers (e)
100
Cum. Combustible Recovery, %
90 10-25 % Emulsified Diesel oil addition
80 70 60
20 g/t ETHXNL 40 g/t ETHXNL 60 g/t ETHXNL 80 g/t ETHXNL Diesel oil only EDO (0.25 % DCNL) EDO (0.50 % DCNL) EDO (1.00 % DCNL)
Release Tests
50 40
Aggl. Tests
30 20 10 0 10
15
20
25
30
35
40
Cum. Ash Content, % Fig 3. The Comparison of Agglomeration and Flotation Characteristics of LS-20 Coal
CONCLUSION
In this study, the comparison of oil agglomeration and flotation method was made for the low grade coal sample in lignite character considering the optimum process conditions for both methods. Oppositely to the general idea in which release test is thought the best method giving the theoretical separation characteristics, it was showed that oil agglomeration application yielded the clean coal products with higher combustible recoveries. It is just because that if the washing process (drainage the impurities from coal agglomerates) is applied appropriately, the application of agglomeration results in clear separation eliminating the bypasses. In addition, the use of higher hydrocarbon dosages resulted in higher combustible recoveries which are not suitable for flotation process since the higher collector dosages deteriorate the selectivity of separation. REFERENCES
Abdelrahman, A.A., Brookes, G.F., 1987. An economic assessment of using surfactants in cleaning coalby the oil agglomeration method, Energy Prog., 7, 1, 47-50. Aktaş, Z., 2002. Some Factors affecting spherical oil agglomeration performance of coal fines, International Journal of Mineral Processing, 65, 177-190. Alonso, M.I., Valdes, A.F., Tarazona, M.R.M., Garcia, A.B., 1999. Coal recovery from coal fines cleaning wastes by agglomeration with vegetable oils; effects of oil type, and concentration, Fuel, 78, 753-759. Baruah, M.K., Kotoky, P., Baruah, J., Bora, G.C., 2000. Cleaning of Indian coals by agglomeration with xylene and hexane. Separation and Purification Technology, 20, 235-241. Boylu, F., Laskowski, J.S., 2006. True flotation and Entrainment in Coal Flotation, XV International Coal Preparation Congress and Exhibition, China, 406-416. Capes, C.E., McIlhinney, A.E., Russel, D.S., Sirianni, A.F., 1974. Rejection of trace metals from coal during beneficiation by agglomeration, Environ. Sci. Technol., 8, 35-38. Capes, C. E., Coleman, R.D., Thayer, W.L., 1985. Treatment of crude petroleum/water emulsions by agglomeration methods using fine coal. Coal preparation, 1, 131-143. Cebeci, Y., Eroğlu, N., 1998. Determination of bridging liquid type in oil agglomeration of lignites. Fuel, 77 (5), 419424. Darcovich, K., Capes, C.E., Talbot, F.D.F., 1989. Surface characteristics of coal-oil agglomerates in the floc regime, Energy and Fuels, 3, 64-70. Dell, C.C., 1953. Release analysis, a new tool for ore dressing research. Recent developments in Mineral Dressing, Londoın, IMM, 75-84. Garcia, A.B., Tarazona, M.R.M., Vega, J.M.G., Wheelock, T.D., 1998. On the role of oil wetting in the cleaning of high rank coals by agglomeration, Fuel, 75, 7, 387-392. Kawashima, Y., Hande, T., Takeuchi, H., Niwa, T., Niwa, K., Sunada, H., Otsuka, A., 1989. Computer simulation of agglomerationby a two-dimensional random addition model-agglomeration kinetics and micromeritic properties of agglomerate accompanied by compaction process, Powder Technology, 57, 157163. Laskowski, J.S., Yu, Z., 2000. Oil agglomeration and its effect on beneficiation and filtration of low-rank/oxidized coal. International Journal of Mineral Processing, 58, 237-252. Mohanty, M.K., Honaker, R.Q., Govindarajan, B., 1999. Development of a characteristic flotation index for fine coal, International Journal of Mineral Processing, 55, 231-243. Petela, R., IOgnasiak, B., Pawlak, W., 1995. Selective agglomeration of coal: analysis of laboratory batch test results, Fuel, 74 (8), 1200-1210. Shrauti, S.M., Arnold, D.W., 1995. Recovery of waste fine coal by oil agglomeration, Fuel, 74 (3), 459-465. Slaghuis, J.H., Ferreira, L.C., 1987. Selective spherical agglomeration of coal. Fuel, 66, 1427-1430. Steedman, W.G., Krishnan, S.V., 1987. Oil agglomeration process for the treatment of fine coal. In: Mishra, S.K., Klimpell, R.R (Eds), Fine coal processing, Noyes Data/Noyes Publications, Park Ridge, NJ, 179-204. Weelock, T.D., Milana, G., Vettor, A., 1994. The role of air in oil agglomeration of coal at a moderate shear rate, Fuel, 73 (7), 1103-1107.
Briquetting Studies of Çanakkale-Çan Coals Presenting Author (Name, Title, Affiliation): Oğuz ALTUN, Mining Engineer, Contact Information: Mineral Research and Exploration Directorate in Turkey ([email protected])
Co-Authors: Ayşe ERDEM*, Akan GULMEZ*, Mining Engineers, Zafer GENCER*, Zeki OLGUN**, Chemical Engineers Contact Information: *Mineral Research and Exploration Directorate in Turkey, **Turkish Coal Enterprises
ABSTRACT Canakkale-Can coals, after being washed within 1.4 g/cm3 density, were studied and air channeled briquettes were prepared to be used for domestic consumption, in this study. Due to high sulfur content, Canakkale-Can coals cause to emit large amount of sulfur oxide gasses and resultantly, environmental pollutions. This property of the coals makes the usage of the coals hard as domestic fuels or conventional power plants’ fuels. It is obvious that, due to high moisture, volatile matter, ash and sulfur contents, direct usage of Canakkale-Can coals will cause environmental pollutions. Utilizing the coals domestically may only be possible with decreasing ash values and increasing calorific values with beneficiation processes, and also with some additions of lime. Moreover, making air channeled briquettes out of them, will make it possible to decrease their sulfur and smoke emissions to tolerable limits. In this study, chemical analysis and stove thermal efficiency test results of the briquettes, made of Canakkale-Can coals were presented. INTRODUCTION Due to comprising of high pyrite content, Canakkale-Can coals emit large amount of sulfur dioxide during combustion, therefore, causing pollution. This property of Canakkale-Can coals do not allow them to be used as domestic fuel and as fuel of fixed bed power plants. But, when used in fluid bed systems, sulfur emissions may be lowered and their environmental impacts were decreased. Due to high moisture, volatile matter, ash and sulfur content, it is obvious that as Canakkale-Can coals are used in house heating, they will inevitably produce lots of pollutions. The utilization of the coals at homes, may only be possible as upgrading of them which include decreasing of the ash content, increasing of the calorific values as well as adding some limes, meanwhile to put them in air channeled briquette forms, therefore to lower the ash and smoke emissions of them to certain tolerable limits.
EXPERIMENTAL STUDIES Briqutte production works As a result of washing Canakkale-Can coals with 1.4 g/cm3 density, the obtained product was put in air channeled briquette forms and they were aimed to be used as house heating fuels. The chemical analysis of the coals to be used in briquette Works is shown on Table 1. Table 1. Chemical analysis of the Canakkale-Can coal, used in the study Analysis Air Dried Sample Dried Sample 18.4 Moisture, % 11.6 14.2 Ash, % 34.1 41.7 Volatile Matter, % 35.9 44.0 Fixed Carbon, % 4.6 5.7 Combustible Sulfur, % 0.2 0.2 Sulfur in Ash, % 4.8 5.9 Total Sulfur, % 4414 5537 *LCV, kcal/kg * LCV (Lower Calorific Value) Coal used in briquettes contains 5.9% total sulfur, and a big share, about 5.7% of it is combustible. In order to hold sulfur with Calcium in ash as CaSO4 in ash, a mol of Calcium addition is stoichiometrically required. During combustion in stoves and boilers, heat profiles form along vertical and horizontal axis. When the ideal capturing temperature, 8150C, of sulfur with calcium is considered, some unexpected conditions such as partial deterioration of CaSO4 or unattained formation temperature to lower the capturing efficiency may be achieved, within high and low heat profiles which are formed in fixed bed combustion units. For this reason, it was indicated in the previous studies that the required addition ratio of lime to coal should be more than a mole for each sulfur mole. In the previous works, not to decrease the calorific value, sulfur amount was not increased more, since sulfur ratio in S/Ca ratio of the coals yields 1 mole in the standard with lime addition. In this study too, lime was added in a proportion that a mole of Ca was added for every combustible sulfur mole, to reduce sulfur ratio to proper levels of the standard. But, since combustible sulfur of the coal is considerably high, addition amount of lime was determined as 9%, considering to meet 1 mole combustible S/1 mole Ca ratio. This increase of lime also caused to increase of the binder (CMC) as well. The chemical analyses of the produced briquettes are shown on Table 2.
Figure 1. The Produced Briquettes Table 2. Chemical analysis of the briquettes, having air channels Analysis Air Dried Sample 9.9 Moisture, % 23.7 Ash, % 38.1 Volatile Matter, % 28.4 Fixed Carbon, % 1.5 Combustible Sulfur, % 2.8 Sulfur in Ash, % 4.3 Total Sulfur, % 4130 LCV, kcal/kg
Dried Sample 26.2 42.3 31.5 1.6 3.1 4.8 4646
Stove Heat Eficiency Works The produced air channeled briquettes were tested for the stove heat efficiency. Stove heat efficiency tests were conducted according to the TS4900 standard. The purpose of TS 4900 oven heat efficiency and capacity test is to keep the fuel type and quality as stable as it can be to determine heat efficiency and capacities the stoves in market, produced with different material and designs.
Table 3. Stove heat efficiency test results 1. Used fuel, (kg) 2. Burning period, (hour) 3. Burning speed of the fuel, (kg/hour) 4. Heat loss of flue gasses, qa (%) 5. Chemical heat loss of flue gasses, qb (%) 6. Heat loss of combustible slag particulates, qr (%) 7. Heat efficiency, (η) (%) 8. LCV, (kcal/kg) 9. Total obtained heat, (kcal/kg) 10. Smoke Concentration, (mg/m3) 11. Smoke emission period, (hour) 12. Flue gasses flowrate, (Nm3/hour) 13. Total smoke, (mg) 14. Discharged smoke value by obtained heat units, (mg/kcal) 15. Discharged Sulfur value, (g) 16. Discharged Sulfur value by burned fuel unit, (g sulfur/kg fuel) 17. Sulfur ratio mixed with flue gasses, (%, kg/kg) 18. Smoke emission value, (g smoke/kg coal)
5.54 1.83 3.03 41.09 2.31 0.26 56.35 4130 12877 771.11 0.67 45.00 23133.30 1.7965 84.99 15.34 1.53 11.46
CONCLUSIONS It is obvious that the prepared mixture of the briquettes would have an effect on the results received from the burning tests in the stoves. High amount of lime addition to briquettes had inevitably negative effects on combustion. The combustion period of the fuel and discharged smoke period delayed, the flue gasses heat loss and chemical loss increased. Naturally, this situation affected the emission amount and heat efficiency. However, with provided sulfur reduction, the other parameters such as smoke emission and heat efficiency were optimized. Thus, sulfur and smoke emission values were reduced to certain limit values which the standards stipulate. REFERENCES Gencer, Z, 2002. Production Of Smokeless Fuel From The Lignites By Air Channalled Briquetting Methot, MTA Report, Ankara TS 12055 Standard, April 1996. Coal Briquettes- For Household Heating TS 4900 EN 13240 Standard, April 2003. Room heaters fired by solid fuel- Requirements and test methods
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COAL SCIENCE DRYING KINETICS OF ÇANAKKALE – ÇAN LIGNITE Ufuk Gündüz Zafer, Department of Chem. Eng., Faculty of Engineering and Clean Energy Research and Application Center, Gazi University, Ankara, Turkey [email protected] Ö. Murat Dogan, Department of Chem. Eng., Faculty of Engineering, Gazi University, Ankara, Turkey Duygu Öztan, Department of Chem. Eng., Faculty of Engineering, Gazi University, Ankara, Turkey Zeki Olgun, Turkish Coal Enterprises (TKI), Ankara, Turkey Mustafa Ozdingis, Turkish Coal Enterprises (TKI), Ankara, Turkey Ömer Sezgin, Turkish Coal Enterprises (TKI), Ankara, Turkey Selahaddin Anac, Turkish Coal Enterprises (TKI), Ankara, Turkey Bekir Zühtü Uysal, Department of Chem. Eng., Faculty of Engineering, Gazi University, Ankara, Turkey
Abstract: In this research, the drying kinetics of Çanakkale - Çan lignite was investigated. Experiments were carried out in Denver IR 35 M model moisture analyzer. Drying experiments were performed at temperatures of 80, 100, 120, 140 and 160oC. Five different particle sizes (1.5, 2.6, 5.6, 10 and 30 mm) were studied.
The critical moisture of lignite and kinetic parameters of the drying process (ko,
Arrhenius equation pre-exponential factor and Ea, activation energy) were determined from the weight loss data as a function of drying time. The average value of critical moisture was found to be 0.23 kg water/kg dry lignite from experiments performed at different drying temperatures and with different particle sizes. Pre-exponential factor of the Arrhenius equation and activation energy were calculated as 0.517 s-1 and
25 381 kJ/kmol, respectively. It was observed that a 4 minute drying process was enough to
decrease the moisture content below 20 % in the range of particle sizes of 1.5 – 10 mm.
Innovative High Energy Efficiency Brown Coal Drying based on Self-Heat Recuperation Technology
Muhammad Aziz*, Chihiro Fushimi, Yasuki Kansha, Kazuhiro Mochidzuki, Shozo Kaneko, Atsushi Tsutsumi Institute of Industrial Science, The University of Tokyo *
Corresponding Author: Tel: +81-3-5452-6727; Fax: +81-3-5452-6728; e-mail: [email protected]
Abstract Brown coal drying based on Self-Heat Recuperation (SHR) technology which recovers effectively both latent and sensible heat is introduced to improve the drying energy efficiency. SHR based fluidized bed drier (SHR-FBD) with heat exchanger immersed inside the bed is adopted as the evaporator. To evaluate the drying efficiency of the proposed SHR-FBD system, a comparison to the available Mechanical Vapor Recompression (MVR) based FBD (MVR-FBD) system in relation to the fluidization velocity and bed aspect ratio in respect of the required energy input was conducted. From the results, the proposed SHR-FBD system was found to be able to drastically reduce the drying energy consumption in all evaluated fluidization velocities and bed aspect ratios. Furthermore, the proposed system can decrease the energy consumption to about 15% and 75% of that required in hot air and MVR drying systems. Keywords: brown coal, self-heat recuperation, energy saving, fluidized bed, fluidization velocity, bed aspect ratio 1. Introduction Long term future of coal as one of leading energy sources is predicated on its utilization as fuel and raw material for some chemical synthesis processes which being increased. As a low rank coal (LRC), brown coal has some advantageous characteristics of low mining cost (possibility for open-cut mining), high reactivity, and low sulfur content. Moreover, brown coal has a high content of volatile matter which make it easier to convert into gas and liquid products compared to other higher rank coals. Unfortunately, brown coal also has some disadvantages on high risk of spontaneous combustion and high inherent moisture content (MC) ranging from 45 to 66 wt% on wet basis (wb). This high MC results in lower calorific value (together with oxygen content), higher transportation cost, complex handing including loading and unloading, reduction in friability, etc. Therefore, brown coal is mostly required to be dried into the final MC below of 15% depending on its utilization. Dried coals could increase its calorific value as well as efficiency of power plants, decrease transportation cost, simplify loading and unloading, decrease power plant emission (GHG), etc. In addition, the efficiency of coke-oven in pre-heating and drying increases about 30-50% and 10-15% respectively when the dried coal is fed (Allardice, et al., 2004; Kudra and Mujumdar, 2009, Karthikeyan et al., 2009; Pikon and Mujumdar, 2009). Brown coal (and other LRCs) drying is quite different to bituminous coal drying. In brown coal, the existence of inherent and chemically bound moistures besides free moisture inside the particle leads to higher energy consumption to evaporate the water (Evans, 1973; Salmans et al., 2001; Sarunac et al., 2010). Water removal requires a large amount of energy input because of large latent heat required for water evaporation. Theoretically, assuming the ambient
1
temperature of 15 ºC, the energy required for water evaporation ranges from 2.48 to 2.57 MJ kg-evaporated water-1 depending on the wet bulb temperature (Brammer and Bridgwater, 1999). Furthermore, in hot air fluidized bed drying, heat consumption for water evaporation is typically in the range of 3.1 to 4.0 MJ kg-evaporated-water-1 (Karthikeyan et al., 2009, Pikon and Mujumdar, 2009). Moreover, brown coal dewatering in large power plant should be able to handle high mass flow rate (up to 200 t h-1) of raw brown coal and it must have high energy efficiency to avoid large amount of energy input. Hence, the development of drying technology with high energy efficiency has been emphasized in last few decades. To reduce the energy input required for drying, recently, vapor recompression/mechanical vapor recompression (VRC/MVR) based brown coal drying has been developed. Potter (1981) has a proposed a fluidized bed drying system utilizing superheated steam as the fluidizing/drying medium in a cascading multistage fluidized bed drier with heat recovery between successive beds. Furthermore, this system is further refined by RWE into WTA (German abbreviation standing for fluidized bed drying with internal waste heat utilization) (Karthikeyan et al., 2007). Unfortunately, this system does not utilize effectively the energy recovery of all the heat brought by the drying medium, the evaporated water and the dried products. To enhance more energy saving, recently, SHR technology which recovers both latent and sensible heat effectively has been developed (Kansha et al., 2009; Kansha et al., 2010; Fushimi et al., 2010). Fushimi et al. (2010) has developed a drying system based on SHR technology with a drier applying countercurrent type of heat exchanger utilizing air as drying medium and they presented a great possibility of SHR application to the drying process. However, they provided no concrete calculation for specific drier including its applicability and performance in brown coal drying. In this research, a self-heat recuperation based continuous fluidized bed drier (SHR-FBD) which is designed for brown coal is developed with the objective of improvement in overall energy efficiency. Furthermore, with the purpose of comparative study regarding the energy efficiency, the developed system is compared to the MVR-FBD system in some different fluidization condition regarding the ratio between the amount of evaporated water to the amount of fluidizing/drying medium. Specifically, this includes the effect of fluidization velocities and bed aspect ratio (i.e. the ratio of bed height to its side length) to the required total energy input. 2. Brown Coal Drying Characteristics The kinetics of coal drying depend on some important factors including coal origin and composition, drying temperature, particle size, residence time distribution (RTD) and permeability. Allardice and Evans (1971) has measured the vapor pressure isotherms for Victorian brown coal and concluded that the equilibrium MC depends strongly on the relative vapor pressure (i.e. the ratio of partial vapor pressure to saturated vapor pressure) or, in the case of drying with air, it is mainly governed by air relative humidity inside the drier. Furthermore, the heat of desorption which is required for brown coal drying can be approximated simply as the latent heat of water vaporization for its initial 80% of water amount and it is increasing gradually for remaining 20%. Numerically, considering the initial MC of the as-mined brown coal is 65 wt% wb, the heat of desorption is equal to the latent heat of water vaporization until drying to MC of about 37 wt% wb. However, it increases accordingly as the drying is progressing to a lower MC due to strong hydrogen bonding and capillary movement (diffusion) inside the particles. In the case of drying with superheated steam (SHS), drying to MC of 15 wt% wb requires a bed temperature of about 108 °C under atmospheric pressure. As the drying is continued to MC of 5 wt% wb, the required relative vapor pressure is about 0.2 corresponding to bed temperature of
2
about 150 °C. Different to drying with SHS, the bed temperature in drying with air is not constant and it changes following the mixing ratio of air-steam mixture. In case of fluidized bed drying, considering that the amount of water to be evaporated is constant, the bed temperature depends strongly to the fluidization velocity. Regarding drying of low rank coals with FBD, Calban (2006) has studied the effect of the bed height to the equilibrium MC and concluded that that there is no effect of bed height to the equilibrium MC of the low rank coals during drying. Drying employing SHS as drying medium is advantageous on lower risk of fire hazard or explosion due to oxidation. Unfortunately, SHS drying usually requires more complex drying system and maintenance. Air infiltration due to sample feeding, product discharging and piping leak must be avoided. Condensation at initial drying process before evaporation is inevitable in SHS drying. Superheated steam requires drying system which is resistant to high temperature/heat and corrosion and usually leading to higher equipment cost. In addition, large energy input is required to produce steam at every start up which means higher operating cost. On the other hand, drying with air as drying medium leads to lower initial equipment cost but usually it has higher risk of fire hazard and explosion than SHS drying. Also, in the conventional hot air drying, poor heat recovery method results in higher operating cost in the respect of energy which is required for drying. Hence, it is necessary to establish a drying method with lower risk of fire hazard or explosion but higher energy efficiency. 3. Self-Heat Recuperative Brown Coal Drying―Drying Concept A state-of-the-art SHR technology is developed based on the concept of exergy recuperation as it has been done in VRC/MVR technology previously with some further improvements (Kansha et al., 2009; Karthikeyan et. al., 2007; Ewers et al., 2003). The characteristics of SHR system is on the exergy elevation continued by its recuperation through heat exchange and heat pairing for each sensible and latent heat. In SHR technology, all of the energy (heat) involved in drying is recuperated and utilized as the energy (heat) source for the next drying process. This includes recuperation of sensible heat of the gas serving as drying medium, both sensible and latent heat of the evaporated water and sensible heat of the dried products. In addition, it also recovers the energy utilized in increasing the exergy level of the evaporated water and drying medium exhausted from the drier. Figure1 shows the basic schematic layout of SHR-FBD system developed for brown coal drying which utilizes air as fluidizing/drying medium. Drying is divided into three continuous stages, they are pre-heating, evaporation and superheating. Raw brown coal is heated initially in a pre-heating stage (drier 1a) to a certain temperature. Subsequently, the main drying process which is water evaporation is performed inside the FBD (drier 2). The internal heat exchangers which are fulfilled by the compressed mixture of fluidizing/drying medium and steam are immersed inside the FBD providing energy (heat) for water removal. Then, the exhausted mixture of fluidizing/drying medium and steam superheated (drier 3) and continuously it flows to the compressor for pressure elevation. Moreover, the compressed vapor having higher exergy level is circulated back and utilized as heat source for superheating (drier 3), evaporation (drier 2) and pre-heating (drier 1), in this order. In addition, the sensible heat of the hot dried brown coal is also recovered by the drying medium to improve the overall energy efficiency (drier 1b). By this method, all enthalpy-rich streams are utilized and the exergy is recovered and recuperated leading to minimization of total energy input. Initially, raw brown coal having particle size of up to 8 mm is ground in the grinder to fine size of 1-2 mm and subsequently it enters into the preheating stage to certain temperature. The preheated fine brown coal enters the second stage of evaporation inside FBD where the moisture is evaporated by the heat supplied by both fluidizing/drying
3
medium (direct heat transfer) and in-bed immersed heat exchangers (indirect heat transfer). The immersed heat exchangers increase the temperatures of both fluidizing/drying medium (air) and brown coal particle. As the result, it leads to increase of moisture-carrying capacity of the air accelerating the water removal from the brown coal. Here, the required heat for water removal inside FBD is provided mainly by the compressed air-steam mixture. This indirect heat transfer could decrease the risk of fire hazard and explosion which becomes a key issue in coal drying (Pikon and Mujumdar, 2009).
Fig.1. SHR-FBD system designed for brown coal utilizing air as fluidizing/drying medium 4. Process Calculation To evaluate the performance of the proposed SHR-FBD system in terms of energy efficiency, mass and energy balance is calculated using a commercial dynamic process simulator Pro/II (Invensys). The energy consumption required for drying in the proposed drying system is compared to that in the MVR-FBD system employing superheated steam (SHS) with the following assumptions: (1) FBD with internal heat exchanger is approximated consisting of a mixer, a heat exchanger and a separator, (2) complete mixing inside FBD, (3) the minimum approach temperature in FBD and other heat exchangers are 10 and 20 K respectively, (4) raw brown coal has an initial MC of 50 wt% wb, (5) drying is conducted to MC of 15 wt% wb, (6) the adiabatic efficiency of compressor, blower and expander are 80%, 80% and 90% respectively, (7) drying is performed under atmospheric condition, (8) Thermodynamic calculation is performed based on Soave-Redlich-Kwong (SRK) method, (9) no heat loss from the system, (10) no mass transport resistance in all phases and no pressure drop during heat exchange in each heat exchanger. Table 1 shows the properties of material and fluidized bed used in process calculation. In this study, with the purpose of evaluating the effect of the bed height to the drying energy efficiency, three different bed aspect ratios AR are set, they are aspect ratio of 1, 2 and 3. In this study, bed aspect ratio AR is defined as the ratio of the bed height Hb to the bed side length lb. One thing must be noticed is that the change of the bed aspect ratio AR simply means the change of the bed height Hb with bed side length lb is fixed. Coal amount and fluidization gas volume in each bed aspect ratio AR and fluidization velocity Uo are shown in Table 2.
4
Table 1 Properties of material and fluidized bed used in process calculation Properties
Value
Brown coal Initial MC [wt% wb] Average diameter ds [mm] Molecular weight [ - ] Coal composition [wt% db]
Notes and sources
50 1.5 384 C: 66.7, H: 4.7, S: 0.3, N: 0.6, O: 26, Ash: 1.7 51.1 900 1470 0.6
Volatile matter [wt% db] Bulk density ρs,bulk [kg m-3] Particle density ρs [kg m-3] Sphericity Φs [ - ] Fluidized bed Shape Side length lb [m] Bed aspect ratio AR [ - ]
Square 4 1, 2, 3
Kumagai et al. (1999) Li (2004)
Dry condition
Bed height Hb 4, 8, 12 m
Table 2 Coal amount and fluidization gas volume in each bed aspect ratio and fluidization velocity Bed aspect ratio AR [-] 1
2
3
Fluid. velocity U [ ×Umf ] 1 2 3 4 1 2 3 4 1 2 3 4
Wet brown coal mass [ ton h-1 ] 57.6
Evaporated water mass [ ton h-1 ] 23.7
115.2
47.4
172.8
71.2
Gas volume [ ×103 m3 h-1 ] 28.8 57.7 89.9 113.7 29.6 57.5 86.8 115.4 30.2 57.4 86.4 120.9
In this study, as the final MC is fixed on 15 wt% wb, the relative vapor pressure of the exhausted mixture of air and steam is quite below the saturated condition. Hence, the superheating stage could be omitted and the exhausted mixture of air and steam is directly compressed by the compressor before heat recuperation is performed in all drying stages. The minimum fluidization velocity Umf for brown coal is approximated with the equation recommended by Chitester (1984) for coarse particle with involving the correction factor for wet coal particles as it is proposed by Pata et al.(1988) and finally it can be summarized as follows:
[
U mf = μ g ⎧⎨ (28.7 ) + 0.0494 Ar ⎩ 2
]
1
2
⎧ ⎡ 7.413 × 10 −2 ⎤ ⎫ − 28.7 ⎫⎬ ⎨ ρ g ⎢1.182 d s − 5.65 × 10 − 4 + ⎥⎬ M 1.58 ⎭ ⎩ ⎣ ⎦⎭
(1)
Furthermore, the blower work Wbl is defined as the work to provide a pressure increase ΔPf which is required for fluidization and is calculated according to the equation proposed by Kunii and Levenspiel (1991) as follows: (3)
ΔPf = ΔPb + ΔPd
where ΔPb and ΔPf are pressure drops across the bed and distributor respectively and they could be calculated as follows (Kunii and Levenspiel, 1991):
ΔPb = (1 − ε mf )( ρ s − ρ g ) H b g
5
c
(4)
ΔPd = 0.4ΔPb
(5)
4.1 MVR-FBD employing SHS as fluidizing/drying medium Figure 2 shows the schematic process diagram of MVR-FBD system with SHS as fluidizing/drying medium as it has been suggested by Fitzpatrick and Lynch (1995). The exhausted steam from the drier is split into re-circulated and purged SHS which is equivalent to the water evaporated from the wet brown coal. The purged SHS is compressed by compressor CP and hence its temperature is elevated from Tv,1 to Tv,2. Then, it flows to HX2 (FBD) for latent heat exchange (condensation of compressed air-steam mixture and evaporation of water possessed by brown coal) and finally to HX1 for sensible heat exchange (brown coal pre-heating).
Fig.2. Schematic process of MVR-FBD system for brown coal with steam as fluidizing/drying medium The total energy input/duty in the optimized MVR-FBD system Wtot,MVR can be calculated as the sum of energy input to compressor Wcp and blower work Wbl which is expressed as follows:
Wtot ,MVR = Wcp + Wbl
(6)
4.2 SHR-FBD utilizing air as fluidizing/drying medium The schematic layout of the developed SHR-FBD system with air as fluidizing/drying medium is shown in Fig.3. Compared to the MVR-FBD system, some additional heat exchangers HX (HX3, HX4 and HX5), an expander EX, a flash FL and a condenser CD are installed in respect of heat/energy recovery. In SHR-FBD system, the compression work Wcp conducted by compressor CP is partly recovered by the expander EX as expander-work Wex. Moreover, the sensible heat of the dried brown coal is recuperated by the fluidizing/drying medium in HX3 before it is pre-heated by the compressed vapor. The total energy input/duty Wtot for SHR-FBD is defined as the sum of energy input to compressor Wcp, blower work Wbl and subtracted by the work gained by expander Wex and hence can be expressed as:
6
(7)
Wtot , SHR = Wcp + Wbl − Wex
Fig.3. Schematic process of SHR-FBD system for brown coal with air as fluidizing/drying medium 5. Results and Discussion Figure 4 shows the comparison of the energy input (total duty) required for drying in both optimized drying systems (Wtot,SHR and Wtot,MVR) in respect of different fluidization velocities Uo and bed aspect ratio AR. As comparison, Pikon and Mujumdar (2009) reported that in conventional hot air drying without heat recovery, the heat consumption in FBD is typically in the range of 3100 to 4000 kJ kg-evaporated-water-1. From Fig.4, it is clear that both drying systems (MVR-FBD and SHR-FBD) could decrease the total energy input significantly in all fluidization velocities Uo and bed aspect ratio AR. Numerically, the heat consumption required for drying to MC of 15 wt% wb in both MVR-FBD and SHR-FBD systems are about 730 and 550 kJ kg-evaporated-water-1. The proposed SHR-FBD system can decrease the required energy input to about 15% and 75% those required in conventional hot air FBD without heat recovery and MVR-FBD systems.
Total input energy Wtot [ MW ]
15
10
5 SHR, AR=1 SHR, AR=2 SHR, AR=3
0 0
1
2 3 4 Fluidization velocity Uo /Umf [ - ]
MVR, AR=1 MVR, AR=2 MVR, AR=3 5
Fig.4. Total energy input comparison Wtot of both SHR-FBD and MVR-FBD systems in relation with fluidization velocity and bed aspect ratio
7
It is very important to notice that in MVR-FBD system, the sensible heat brought by the dried brown coal is abandoned anyway without recovery because the temperature, furthermore the exergy level, of the dried brown coal is lower (or minimally same) compared to that of the fluidizing/drying medium (which is SHS). On the other hand, in the proposed SHR-FBD system, all of the heat including the sensible heat of the dried brown coal and the condensate can be recovered. Hence, higher energy efficiency can be earned from the SHR-FBD system employing air as fluidizing/drying medium. In addition, the compression work Wcp of the compressor CP is recovered as gained expander work Wex in SHR-FBD system but not in MVR-FBD system. 6. Conclusions A novel brown coal drying based of SHR technology employing fluidized bed as the main drier has been proposed and its performance regarding energy efficiency has been evaluated. From the result, it can be concluded that the proposed SHR-FBD system utilizing air as fluidizing/drying medium can reduce the drying energy consumption compared to conventional hot air drying or MVR-FBD system utilizing SHS as fluidizing/drying medium. Numerically, the developed SHR-FBD system can decrease the energy consumption to about 15% and 75% of that required in conventional hot air and MVR drying technologies. Hence, SHR-FBD system can be considered as the leading-edge of technology for coal drying concerning energy efficiency. Furthermore, drying in lower fluidization velocity seems to be preferable because it shows a higher energy efficiency compared to in higher fluidization velocity. Different to conventional hot air drying, in SHR based drying, the inlet air temperature is considerably low because the heat which is required for drying is provided indirectly through the in-bed immersed heat exchanger. Hence, it improves the feasibility of the SHR drying system to be applied in brown coal drying as this technology could decrease significantly the risk of fire hazard or explosion due to oxidation. Nomenclatures AR
bed aspect ratio, dimensionless
Ar
Archimedes number, dimensionless
c
conversion factor, 1 kg m N-1 s-2
d
particle diameter, m
g
gravity acceleration, m s-2
H
bed height, m
l
length, m
M
moisture content, wt% wb
p
pressure, kPa
po
saturated vapor pressure, kPa
T
temperature, K
U
fluidization velocity, m s-1
W
Work, W
8
Subscripts b
bed
bl
blower
cp
compressor
d
distributor
ex
expander
f
fluidization
g
gas
mf
minimum fluidization
s
brown coal sample particle
r
ratio
tot
total
v
vapor
Greek letters
ε
voidage, dimensionless
Φ
sphericity, dimensionless
ρ
density, kg m-3
μ
dynamic viscosity, Pa s
Acknowledgement The authors gratefully acknowledge to the New Energy and Industrial Technology Development Organization (NEDO), Japan, for the financial support to this research. References Allardice, D.J. and D.G., Evans (1971), The-brown coal/water system: part 2. Water sorption isotherms on bed-moist Yallourn brown coal, Fuel, Vol. 50, pp. 236-253. Allardice, D.J., A.L., Chaffee, W.R., Jackson, M., Marshall (2004), Water in brown coal and its removal, Chapter No 3 of Advances in the Science of Victorian Brown Coal/ed. C.Z. Li/, Elsevier, Oxford, 2004, pp. 85-133. Brammer, J.G. and Bridgwater, A.D. (1999), Drying technologies for an integrated gasification bio-energy plant, Renewable and Sustainable Energy Reviews, Vol. 3, pp. 243-289. Calban, T. (2006), The effects of bed height and initial moisture concentration on dying lignite in a batch fluidized bed, Energy Sources, Part A, Vol. 28, pp. 479-485. Chitester, D.C., R.M. Kornosky, L.S. Fan, J.P. Danko (1984), Characteristics of fluidization at high pressure, Chem. Eng. Sci., Vol. 39, pp. 253-261. Evans, D.G. (1973), The brown-coal/water system: part 4. Shrinkage on drying, Fuel, Vol. 52, pp. 186-190. Ewers, J., H. J. Klutz, W. Renzenbrink, G. Scheffknecht (2003), The development of pre-drying and BoA-plus technology, VGB Conference, March 19-20, Cologne. Fitzpatrick, J. and D. Lynch (1995), Thermodynamic analysis of the energy saving potential of superheated steam
9
drying, Trans. IChemE, Vol. 73(C), pp. 3-8. Fushimi, C., Y. Kansha, M. Aziz, K. Mochidzuki, S. Kaneko, A. Tsutsumi, K. Matsumoto, K. Yokohama, K. Kosaka, N. Kawamoto, K. Oura, Y. Yamaguchi, M. Kinoshita (2010), Novel drying process based on self-heat recuperation technology, Drying Technology, in press. Gibilaro, L.G and and P.N. Rowe (1974), A model for a segregating gas fluidized bed, Chemical Engineering Science, Vol. 29, pp. 1403-1412. Kansha, Y., N. Tsuru, K. Sato, C. Fushimi, A. Tsutsumi (2009), Self-heat recuperation technology for energy saving in chemical processes, Industrial and Engineering Chemistry Research, Vol. 48, pp. 7682-7686. Kansha, Y., N. Tsuru, C. Fushimi, K. Shimogawara, A. Tsutsumi (2010), An innovative modularity of heat circulation for fractional distillation, Chemical Engineering Science, Vol. 65, pp. 330-334. Karthikeyan, M., J. V. M. Kuma, C. S. Hoe (2007), Factors affecting quality of dried low-rank coals, Drying Technology, Vol. 25, pp. 1601-1611. Karthikeyan, M., W. Zhonghua, A.S. Mujumdar (2009), Low-rank coal drying technologies―Current status and new developments, Drying Technology, Vol. 27, pp. 403-415. Kudra, T. and A.S. Mujumdar (2009), Superheated steam drying, Chapter No 7 of Advanced Drying Technologies, 2nd edition, CRC Press, Taylor & Francis Group, Florida, pp. 89-122. Kumagai, H., T. Chiba, K. Nakamura, (1999), Change in physical and chemical characteristics of brown coal along with a progress of moisture release, American Chemical Society, Division of Fuel Chemistry, Vol. 44(3), pp. 633-636. Kunii, D. and O. Levenspiel (1991), Fluidization Engineering, second ed., Butterworth-Heinemann. Washington. Law, C.L. and A.S. Mujumdar (2009), Fluidized bed dryers, Chapter 8 of Handbook of Industrial Drying, 3rd edition/ed. A.S. Mujumdar/, Taylor & Francis Group, Florida, pp. 173-202. Li, C.Z. (2004), Introduction, Chapter 1 of Advances in the Science of Victorian Brown Coal/ed. C.Z. li/, Elsevier, Oxford, pp. 1-10. Mujumdar, A.S. (2009), Superheated steam drying, Chapter 8 of Handbook of Industrial Drying, 3rd edition/ed. A.S. Mujumdar/, Taylor & Francis Group, Florida, pp. 439-452. Pata, J., M. Carsky, M. Hartman, V. Vesely (1988), Minimum fluidization velocities of wet coal particles, Ind. Eng. Chem., Vol. 27, pp. 1493-1496 Pikon, J. and A.S. Mujumdar (2009), Drying of coal, Chapter 43 of Handbook of Industrial Drying, 3rd edition/ed. A.S. Mujumdar/, CRC Press, Taylor & Francis Group, New York, pp. 993-1018. Salmas, C.E., A.H. Tsetsekou, K.S. Hatzilyberis, G.P. Androutsopoulos (2001), Evolution of lignite mesopore structure during drying. Effect of temperature and heating time, Drying Technology, Vol. 19, pp. 35-64. Sarunac, N., E.K., Levy, M., Ness, C.W., Bullinger, J.P., Mathews, P.M., Halleck (2010), A novel fluidized bed drying and density segregation process for upgrading low-rank coals, Int. J. Coal Preparation and Utilization, Vol. 29, pp. 317-332.
10
Using Pyrolysis Tar to meet Fuel Specifications in Coal-to-Liquids Plants Jaco Schieke, Principal Process Engineer, Foster Wheeler Business Solutions Group, Reading, UK email: [email protected] Abstract: Indirect Coal-to-Liquids (CTL) facilities (i.e. liquids production via gasification and Fischer-Tropsch synthesis) are often sited adjacent to the coal mine feeding the facility, but these coal deposits are often found in remote locations; far from coastal refinery hubs where synergies with conventional refining operations can be exploited. The ability to produce a final saleable product as opposed to a blendstock is therefore an important consideration when designing a CTL facility. This opens up marketing opportunities in the immediate area of the facility and reduces the logistical cost of shipping the CTL product to a blending facility. There are several challenges to meeting final product specifications from a low temperature Fischer-Tropsch-based CTL facility. Hydrocracked Fischer-Tropsch diesel, although attractive from a cetane and sulphur perspective, has a lower density than conventional diesel. Fischer-Tropsch naphtha on the other hand is a very good chemicals feedstock, but typically has a lower octane than required. Foster Wheeler recently performed a number of feasibility studies on projects where there was a requirement to produce saleable product from the CTL facility. This article describes some of the key challenges and potential solutions to meeting this constraint. Through the integration of the refining needs with the coal and gasification technology selection, these challenges were addressed. By selecting the right combination of gasifier and coal, combined with the recovery of tar and oil by-products from the gasification process, the blendstock requirement to meet product specifications was substantially reduced.
Coal conversion into transportation liquids, commonly referred to as Coal-to-Liquids (CTL), through the conversion of syngas generated from coal gasification, is considered in a number of countries as a potential solution for coal-rich nations to reduce dependence on imported oil and refined products. This paper examines one of the key technical challenges in the development of a CTL facility; namely producing fuels meeting final market specifications, rather than producing blendstocks. Although a seemingly minor issue in the development of a CTL project, the inability to meet final market product specifications, in domestic or export markets, poses a substantial business risk to the viability of a CTL project. If a CTLderived product is unable to meet the final market specification, the viability of the project will be entirely dependent on availability of blendstock, which may increase project risk and logistical costs. Fuel specifications are generally determined by state or national governments. Although significant variation is still encountered in fuel specifications around the world, the drive for cleaner fuels, and pressure by engine manufacturers are leading to a convergence in specifications. The key global specifications are those used in the United States and the European Union. This paper compares typical low temperature Fischer-Tropsch (LTFT) derived CTL diesel against the Euro IV and Euro V diesel fuel specifications, and considers ways in which coal and gasification technology selection can play a role in addressing the shortfalls of CTL diesel. The paper is based on feasibility studies that Foster Wheeler recently completed on potential large-scale CTL facilities, in South Africa, China and the United States. In all cases, these facilities were mine-
mouth facilities at inland locations. All sites were remote, away from conventional refineries which would typically be the buyers of a blendstock product. Regional markets however existed, where a final product could be sold directly to the enduser. The table below presents CTL fuel properties typical of “straight-run” (SR) CTL diesel. This diesel is produced from LTFT synthesis, followed by hydrocracking of the synthetic waxes and hydrotreating of the lighter products. The Euro IV and Euro V fuel specifications under discussion are compared against this typical CTL product. Property
Unit
SR CTL
Euro IV
Euro V
Density
kg/m3 (min/max)
750-780
820/8451
820/8451
Cetane
Number
65-90
> 51
>51
Sulphur
ppm
<5
<50
<10
Notes 1) or Arctic diesel specification of 800/845 kg/m
3
It can be seen that CTL-derived diesel is an excellent diesel product and exceeds Euro IV and Euro V standards in all respects but density. The cetane number and sculpture specifications were well within the market specification, which makes LTFT diesel an ideal high cetane, low sulphur blendstock, particularly for conventional refiners that are unable to blend excess light cycle oil into their diesel pool. The low density of CTL-derived diesel does however pose a problem when targeting a final product. The lower density of CTL-derived diesel would not be a problem, if it were not for the consumer issues associated with the lower density. Due to the lower density, the volumetric energy content of CTL diesel is less than that of conventional diesel. Because fuel is sold on a volumetric basis, the fuel is therefore not compatible with the market specification. The consumer perceives this as reduced volumetric fuel efficiency when using CTL-derived fuels. Therefore, in order to produce a compatible diesel product, a CTL facility requires additional measures in order to meet market density specifications. This can take one of two forms: (i) blending with an external blendstock prior to the end user sale, or (ii) additional processing steps within the CTL facility. It is however not possible to increase straight-run CTL diesel density using current technology without incurring a significant cost and yield penalty. It should be noted that meeting the less stringent Arctic diesel specification of 800kg/m3 may be feasible, but this severely limits the potential market for CTL diesel. Figure 1 shows a typical indirect CTL facility, consisting of gasification, followed by acid gas removal and Fischer-Tropsch synthesis. The Fischer-Tropsch products then undergo mild hydrocracking followed by the hydrotreatment of olefins. This straight-run diesel can be used as a diesel blendstock, while the straight-run naphtha from the process is either upgraded to improve the octane through isomerisation or reforming or sold directly as a cracker feed to a chemical complex.
O2
Coal
Isomerisation Gasification
Gasoline
Raw syngas
Acid Gas Removal Clean syngas
LTFT Synthesis
LTFT Wax
CCR Reforming
SR CTL Naphtha
Hydrocracking/ treating
SR CTL Diesel
Diesel Blendstock LPG
Figure 1: Typical indirect Coal-to-Liquids Flow scheme The use of low-temperature gasification presents an alternative route to final product densification. Low-temperature gasification and, specifically, fixed-bed technologies, produces a substantial amount of tar and pyrolysis oil as part of the gasification process. This occurs due to the counter-current flow of synthesis gas and coal in these gasifiers, where the hot synthesis gas liberates volatile pyrolysis tar and oils from the coal without cracking. This happens in the upper part of the gasifier, where temperatures are sufficient to devolatilise the coal, but low enough to limit cracking. Where LTFT-derived diesel is a low density, high cetane material, pyrolysis products are high density, low cetane blendstocks. Density of pyrolysis products ranges upwards of 865kg/m3and therefore makes an ideal blendstock to increase the density of the LTFT diesel. The amount of pyrolysis product produced is heavily dependent on the coal feedstock rank and grade. Although the proximate analysis gives an indication of the volatile matter in the coal, the most reliable estimate of pyrolysis oil can be obtained from a Fischer assay of the feedstock. The Fischer assay measures the condensable liquid produced when heating a sample to 500°C for one hour in the absence of oxygen, a process very similar to conditions inside the gasifier. This analysis is useful when assessing different potential coal feedstocks and should be considered when targeting maximum pyrolysis oil production. In a typical CTL flow scheme using fixed-bed gasification, the pyrolysis product is recovered in two stages: coal tar is knocked out from the raw synthesis gas during initial quench, and gasification naphtha is recovered during the cooling of synthesis gas to the acid gas absorber feed temperature. The coal tar typically contains some of the coal dust entrained in the raw syngas after gasification. This has to be taken into consideration when selecting the refining route, as filtration and solids removal is difficult and adds another degree of complexity to the flowscheme. There are two main routes for upgrading the pyrolysis products: carbon rejection (e.g. coking) or hydrotreatment. Foster Wheeler’s work focused predominantly on the hydrotreating route, as this yields the maximum product from refining the tar. The flowscheme considered, shown in Figure 2, consisted of filtration, followed by fractionation and separate hydrotreating of the resulting tar distillate and naphtha streams. The residue was recycled back to the gasifiers, although coking was considered an attractive option to increase product yield. The flowscheme used conventional fixed-bed hydrotreating technologies, but the opportunity exists to hydrotreat the unfiltered feed in an ebullated-bed reactor. Hydrotreated distillate was blended with the straight-run CTL diesel, whilst naphtha was combined with the conventional LTFT naphtha and sent to a reformer. Other refinery options are discussed in more detail in Lamprecht (2010).
O2
Coal
Coal Tar Filtration & Distillation
Coal Tar
Gasification
Acid Gas Removal
Coal Tar Hydrotreating
Gasification Naphtha
Gasoline (and Reformate)
Isomerisation
Raw syngas
CCR Reforming Diesel (and Blendstock)
Clean syngas
LTFT Synthesis
LTFT Wax
Hydrocracking/ treating
SR CTL Diesel
LPG
Figure 2: Indirect Coal-to-Liquids Flow scheme incorporating Pyrolysis Tar and Oil The results from Foster Wheeler’s studies are presented in Figure 3 below. The final product slate from an indirect CTL facility is shown for the two cases considered: CTL with tar blending, and CTL without tar blending. The benefit of utilising the gasification pyrolysis products in the final fuel blend is evident. On-specification diesel production increases by 22 percentage points from 5% to 27% of the overall product. This equates to a 33% increase of saleable diesel in the diesel pool and a commensurate reduction in the blendstock requirement. The remainder of the diesel product will require further blending to meet the density specification, but the amount of blendstock required is reduced by one third. The study by Lamprecht (2010) has shown that all diesel could be produced to meet the specification if sufficient pyrolysis products are available in the facility. The quantity required is more than would be produced in a conventional CTL facility and other options would have to be considered to increase the pyrolysis tar yield.
Without Tar Blending
With Tar Blending 9%
10%
LPG
19%
39%
21%
RON95 Gasoline Reformate
63%
Diesel
3% 4%
5%
Diesel Blendstock
27%
Figure 3: CTL Facility Product Slate without and with Gasification Tar (and Oil) Blending An LTFT-based CTL facility typically aims to maximise diesel production, but it can be seen from the results that the balance of the products comprises about a third of the total products. The gasoline products indicated in Figure 3 (RON95 gasoline and reformate) arises from the further processing of the LTFT naphtha into gasoline. Although LTFT naphtha makes an excellent reforming feedstock for chemicals production, an isolated facility may again need to target onspecification gasoline production rather than naphtha export. LPG is produced in the facility as needed to meet the
specification for light ends in gasoline. The remainder of the paper considers the gasoline value chain and the ways in which pyrolysis products aid the production of gasoline. The straight chain, paraffinic LTFT-derived naphtha has a very low octane content and further refining is required to increase the octane of this stream. This is typically achieved through isomerisation of the C5/C6 fraction, and continuous catalyst regeneration (CCR) reforming of the heavier components. The severity of the CCR process is determined by the N2A (naphtha + 2 x aromatics) of the feed. The straight-run LTFT naphtha has a very low N2A, leading to a very high severity and low yields in the CCR reformer. The high aromaticity of the hydrotreated tar oil fraction makes an ideal blendstock to the CCR reformer feed, increasing the feed N2A and reducing the severity requirement. Additionally, the aromaticity of the pyrolysis products improves the octane of the straight-run LTFT naphtha. In the Foster Wheeler studies there was an excess of reformate from the CCR reformer, because the gasoline pool was already constrained by the aromatics limit. The gasoline pool typically consists of a combination of straight-run naphtha, isomerate and reformate. When blending only these components to meet a final gasoline specification, the facility is typically constrained by the benzene and/or aromatics specifications of the gasoline when targeting a Euro IV or Euro V specification. Although the pyrolysis products therefore help in increasing the octane, an octane booster may still be required in order to meet final product specifications. This could take the form of MTBE, TAME, ETBE or ethanol, depending on local regulations. Care must be taken, in selecting the octane booster, as the trade-off between vapour pressure (RVP), aromatics and octane remains a constraint in a refinery blending only reformate and isomerate. In conclusion, the Foster Wheeler studies demonstrated that the use of gasification pyrolysis products can aid in meeting final diesel specifications from an LTFT-based CTL facility, by providing a high density blendstock for the final diesel pool. In addition, the gasoline pool will also benefit from including pyrolysis products due to the increased octane and lower severity of the CCR reactor. The selection of a low temperature gasification technology, combined with the use of the resultant pyrolysis tars within the refinery, has the potential to greatly improve the fuel pool produced within a standalone CTL facility. References 1.
Lamprecht D, Nel R, Leckel D, Production of On-Specification Fuels in Coal-to-Liquid (CTL) Fischer-Tropsch Plants Based on Fixed-Bed Dry Bottom Coal Gasification, Energy Fuels, 2010, 24 (3), pp 1479–1486
OVERVIEW OF THE RENTECH PROCESS Belma Demirel and Harold Wright Rentech Inc., Rentech Energy Technology Center, 4150 East 60th Avenue, Commerce City, CO 80022, USA [email protected] Rentech®, which stands for Renewable Energy Technologies, envisions to be a global provider of clean energy solutions. Rentech develops projects to produce certified low carbon synthetic fuels and electric power from carbon‐containing materials such as biomass, waste and fossil resources. Rentech owns technologies that enable the production of such fuels and power when integrated with other technologies. Rentech strives to deliver solutions to its customers and licensees that will help enable countries to effectively address national energy security issues, rising long‐term demands for fuels and electric power and limited domestic petroleum refining capacity while lowering their emissions profiles. Rentech is pursuing the worldwide deployment of our patented and proprietary Rentech‐ SilvaGas biomass gasification technology (Figure 1) as well as our patented Rentech Process (Figure 2). Together with gasification and upgrading technologies, the Rentech Process can economically produce ultra‐clean synthetic fuels such as military and commercial jet fuels and ultra low sulfur diesel as well as specialty waxes and chemicals from a wide variety of waste, biomass and fossil resources. During our nearly 30‐year history, Rentech and our licensees have successfully applied the Rentech Process in facilities ranging in size from pilot scale to approximately 250 barrels per day of synthetic fuels and chemicals production. Our first renewable facility was built in 1992 and used landfill gas to produce alternative fuels. The Rentech‐SilvaGas biomass gasifier is one of only a handful of biomass gasifiers that have operated on a commercial scale. Today, we are working on waste‐to‐energy projects, including a facility in Rialto, California that would produce approximately 640 daily barrels of renewable synthetic fuels and 35 megawatts of renewable power. We are also working on commercial‐scale fossil projects, including a facility contemplating total production of 30,000 barrels per day of synthetic fuels and chemicals in Adams County, Mississippi.
Figure 2. Rentech Process. Figure 1. Rentech-SilvaGas Gasification Process.
Rentech‐SilvaGas Gasifier The patented Rentech‐SilvaGas Biomass Gasification Process produces syngas from a wide variety of biomass feedstocks (Figure 1). The life‐cycle carbon footprint of renewable fuels and power facilities using the Rentech‐SilvaGas gasifier coupled with the Rentech Fischer‐Tropsch Process for synthetic fuels are anticipated to be near zero. The renewable synthetic diesel and jet fuels produced at these facilities will meet all applicable fuels standards, be compatible with existing
engines and pipelines, and burn cleanly, with emissions of particulates and other regulated pollutants significantly lower than emissions of traditional fuels. With approximately $100 million in funding from sources including private investors and the U.S. Department of Energy (DOE), SilvaGas developed biomass gasification technology that has operated successfully on a commercial scale. A gasifier in Burlington, VT, built in partnership with the DOE, National Renewable Energy Laboratory (NREL) and Battelle Columbus Laboratory converted 400 dry tons per day of wood‐based biomass into synthesis gas used for power production. The SilvaGas gasifier operated on this scale between 1998 and 2001 during a series of operating campaigns. Prior to that, the SilvaGas process was tested at pilot scale during more than 22,000 hours of operation. Rentech acquired SilvaGas in June 2009. In addition to owning the SilvaGas biomass gasifier, Rentech has a 25% strategic investment in ClearFuels Technology Inc., a biomass gasification and project development company. ClearFuels owns a proprietary flexible biomass gasification technology platform that converts multiple rural cellulosic biomass feedstocks such as sugarcane bagasse and virgin wood waste into clean syngas suitable for integration with synthesis gas‐to‐liquids technologies (see Figure 3). ClearFuels has signed an exclusive worldwide license with Rentech for the use of Rentech's patented and proprietary Fischer‐Tropsch synthetic fuels technology for the production of renewable drop‐in fuels from sugarcane bagasse. ClearFuels has also signed a license with Rentech for the use of the Rentech Fischer‐Tropsch Process for the production of renewable synthetic fuels from virgin wood waste at up to twelve U.S.‐based projects to be developed by ClearFuels. To facilitate the development process, ClearFuels will build a 20 ton‐per‐ day biomass gasifier designed to produce syngas from bagasse, virgin wood waste and other cellulosic feedstocks at Rentech's Product Demonstration Unit (PDU) in Colorado. The gasifier will be integrated with Rentech's Fischer‐Tropsch Process and UOP's upgrading technology to produce renewable drop‐in synthetic jet and diesel fuel at demonstration scale. Rentech and ClearFuels have been selected to receive up to $23 million grant from the U.S. Figure 3. ClearFuels Gasification Process. Department of Energy (DOE) to construct the biomass gasifier at RETC. Over the course of 20 years, more than $20 million has been invested in the development of the technology underlying the highly efficient ClearFuels HEHTR technology platform. Between 1987 and 2004, Pearson Technologies, Inc. developed and tested the underlying biomass gasification technology at a 5 ton per day pilot scale facility in Mississippi, converting a wide range of feedstocks including bagasse, wood waste, rice straw and corn stover, into syngas during over 10,000 hours of operation. From 2005 to 2007, ClearFuels co‐funded additional testing of the gasification process for the conversion of bagasse into syngas. Multiple third parties, including Idaho National Laboratory and Hawaii Natural Energy Institute, have independently validated the results of the biomass gasification testing at the pilot facility. Rentech Process The Rentech Process is based on Fischer‐Tropsch chemistry. The Rentech Process as seen in Figure 2 is a proprietary technology that, together with gasification and upgrading technologies, converts synthesis gas from biomass and fossil resources into hydrocarbons that are subsequently processed and upgraded into ultra‐clean synthetic fuels and specialty waxes and chemicals. For nearly 30 years, Rentech has been perfecting this process with reproducible results. Much of the original technology has been improved by Rentech, including catalyst composition, reactor design
and design parameters of synthetic fuels and chemicals facilities. These enhancements have improved the quality and yields of products produced as well as the operating efficiency of the facilities. Rentech’s technology is environmentally sound and enables the production of ultra‐clean synthetic fuels, waxes, chemicals, and power from a wide variety of biomass and fossil feedstocks. The most critical component of the Rentech Process is our proprietary iron‐based Rentech catalyst, which drives the Rentech Process in our reactor. Rentech is the only United States based company that uses an iron‐based catalyst to produce synthetic transportation fuels and chemicals. There are multiple global opportunities for deploying Rentech's iron‐based catalyst as it performs well with a wide range of syngas compositions from a variety of feedstocks. In addition, the ingredients of our proprietary iron‐based catalyst are relatively inexpensive and more economical to use than competing catalysts. Rentech's catalyst is also non‐toxic and does not require hazardous disposal handling. The Rentech Process uses a slurry bubble column reactor, known as the Rentech Reactor, to turn the synthesis gas feed into useful products. Synthesis gas is fed into the bottom of the reactor where it is mixed with liquid wax and the solid catalyst. The catalyst and wax are collectively referred to as "slurry". This chemical reaction is exothermic (giving off heat) and steam is generated in the internals of the reactor to produce steam and to remove heat and control the reaction temperature. Product upgrading is the final step for the production of synthetic fuels and chemicals at a facility utilizing the Rentech Process. The Rentech reactor produces a wide array of hydrocarbons that are hydrocracked and/or hydrotreated to produce ultra‐clean synthetic fuels and specialty waxes and chemicals. Our preferred upgrading technology provider is UOP LLC, the premier refining process and technology company. Rentech and UOP, a Honeywell company, maintain a technology and marketing alliance to provide a one‐stop solution to developers of commercial synthetic fuels and chemicals facilities worldwide for synthesis gas conversion and product upgrading. Fuels produced from our process are among the most greenhouse gas friendly transportation fuels available. Our process produces biodegradable RenJet® and RenDiesel® transportation fuels that are cleaner burning than petroleum‐derived fuels and either meet or exceed all applicable fuels and environmental standards. In addition, with carbon capture and sequestration, the carbon dioxide emissions from the production of fossil‐based synthetic fuels from the Rentech Process are substantially lower than those generated in the production of petroleum‐derived fuels. If biomass is used as an input, our technologies can produce renewable diesel and jet fuels with the same qualities and characteristics noted above. Unlike some other alternative fuels, fuels produced from our process can be used and distributed without modifications to existing diesel or jet engines or pipelines. In addition, fuels produced from the Fischer‐Tropsch process, on which our technology is based, are currently the only synthetic fuel type certified for use by the United States Air Force and in commercial aircraft. Product Demonstration Unit Rentech’s Product Demonstration Unit (PDU) located in Commerce City, Colorado is the only operating synthetic transportation fuels facility in the United States (see Figure 4). This facility is designed to produce over 400 gallons per day of ultra‐clean synthetic jet fuel, aviation fuel, ultra‐low sulfur diesel, and specialty waxes and chemicals and is scalable for greater output. The PDU demonstrates the successful design, construction and operation of a fully‐integrated synthetic fuels and chemicals facility utilizing the Rentech Process. It represents the first time Rentech has operated an integrated facility. While operating the PDU, our catalyst demonstrated greater efficiency and yielded more product than predicted. In addition, we were able to produce a quality of syngas at the PDU that is typically created from solid feedstocks. We believe this demonstrates that the Rentech catalyst can successfully react with syngas streams from a wide variety of feedstocks including natural gas, biomass and fossil‐based resources. At the PDU, we demonstrated the operation of our proprietary catalyst wax separation system at scales above the laboratory and pilot scale. In addition,
significant, commercially relevant quantities of Rentech catalyst have been manufactured in cooperation with large commercial catalyst vendors to support the PDU operations. At the PDU, thousands of gallons of ultra‐clean synthetic fuels including military jet fuel, commercial Jet A and Jet A‐1 and ultra low sulfur diesel have been produced and samples of Rentech products have been shipped for testing to potential customers, including the United States Air Force. The PDU is designed at a scale that allows us to collect engineering and performance data necessary for commercial design application. Operations to date at the PDU have verified a number of those engineering and performance criteria and we expect future operations to validate the Figure 4. Rentech’s Product Demonstration remaining criteria for commercial design. Unit Rentech and ClearFuels Technology Inc. (ClearFuels) have been selected to receive up to $23 million grant from the U.S. Department of Energy (DOE) to construct a ClearFuels 20 ton‐per‐day biomass gasifier at RETC. The gasifier will be integrated with Rentech’s PDU for the production of renewable synthetic fuels from biomass. This joint demonstration of an integrated bio‐refinery will lead to the final design basis for commercial facilities that are expected to use the combined technologies. Rentech continues to advance our technology, with a goal of reducing operating and capital costs. We are also working on next generation technologies with significant possibilities to reduce cost and increase efficiency of the Rentech Process. Rialto Project Our proposed Rialto Renewable Energy Center (Rialto Project) will be located in Rialto, California (see Figure 5). The facility is designed to produce approximately 640 barrels per day of pure renewable synthetic fuels and export approximately 35 megawatts of renewable electric power. The renewable power is expected to qualify under California's Renewable Portfolio Standard (RPS) program, which requires utilities to increase the amount of electric power they sell from qualified renewable‐energy resources. The plant will be capable of providing enough electricity for approximately 30,000 homes. RenDiesel®, the renewable synthetic diesel to be produced at the facility, meets all applicable fuels standards, is compatible with existing engines and pipelines and burns cleanly, with emissions of particulates and other regulated pollutants significantly lower than the emissions from the combustion of CARB ultra‐low sulfur diesel. The project will use the Rentech‐ SilvaGas biomass gasification technology. Rentech's proprietary technology for the Figure 5. Satellite view of Rialto Project site. conditioning and clean‐up of syngas will provide the next critical link in the technology chain after gasification. The conditioned syngas will be converted by the Rentech Process in a commercial scale reactor to finished, ultra‐clean products such as synthetic diesel and naphtha using upgrading technologies under an alliance between Rentech and
UOP, a Honeywell Company. Renewable electric power will be produced at the facility by using conventional high‐efficiency gas turbine technology. The power is anticipated to be sold to local utilities under the California RPS program. The primary feedstock for the Rialto Project will be urban woody green waste such as yard clippings, for which Rentech is currently negotiating supply agreements. The location of the project will provide local green waste haulers with a cost‐effective alternative to increasingly scarce landfills for the disposal of woody green waste. Construction of the Rialto facility is expected to create approximately 250 jobs with about 70 full‐ time regular jobs during operation, based on the preliminary design work completed to date. We have signed an unprecedented multi‐year agreement to supply eight airlines with up to 1.5 million gallons per year of renewable synthetic diesel (RenDiesel®) from the Rialto Project for ground service equipment operations at Los Angeles International Airport (LAX). The initial purchasers under the agreement with Aircraft Service International Group (ASIG), the entity that provides fueling services to many airlines that operate at LAX, are Alaska Airlines, American Airlines, Continental Airlines, Delta Air Lines, Southwest Airlines, United Airlines, UPS Airlines and US Airways. Additional airline purchasers of RenDiesel can be added under the agreement with ASIG. Natchez Project Our proposed Rentech Strategic Fuels and Chemicals Complex, or Natchez Project, will be located in Adams County, Mississippi, near the city of Natchez (see Figure 6). The project design contemplates using petcoke and biomass feedstocks for large‐scale production of synthetic fuels and power. We have completed the feasibility study for this project and additional engineering work to finalize technology selection for the facility. We engaged CH2M Hill to serve as our environmental consultant to assist us with the filing of our permits for the Natchez Project.
Figure 6. Schematic of Natchez Project site. In December 2009, Rentech signed a non‐binding Memorandum of Understanding (“MOU”) with thirteen domestic and international passenger and cargo carriers, Air Canada, AirTran Airways, American Airlines, Atlas Air, Delta Air Lines, FedEx Express, JetBlue Airways, Lufthansa German Airlines, Mexicana Airlines, Polar Air Cargo, United Airlines, UPS Airlines and US Airways that contains terms that are anticipated to serve as the basis of a possible definitive purchase agreement by these carriers for the Natchez Project’s entire synthetic jet fuel production. We own an approximately 450‐acre property located adjacent to the Mississippi River near Natchez, Mississippi, which is the intended site of this facility. The site has access to multiple feedstocks and product distribution channels including rail, pipeline, barge and roads. Equally important, we believe the Natchez Project has the best carbon reduction solution of any planned commercial scale synthetic transportation fuels plant in the United States. We have contracted to sell all of the carbon dioxide to be captured at this proposed facility to Denbury Onshore, LLC, a wholly owned subsidiary of Denbury Resources Inc. (“Denbury”). Carbon dioxide
purchased under the long‐term contract would be used for enhanced oil recovery to produce otherwise unrecoverable oil at Denbury’s Cranfield oil field in Southwest Mississippi as well as at the company’s oil fields within the greater Gulf Coast area. Rentech Energy Midwest Corporation Rentech owns one of the few remaining U.S. ammonia fertilizer production facilities, Rentech Energy Midwest Corporation (REMC) (Figure 7). This facility has been operating for over 40 years and is one of the largest producers of ammonia fertilizer products in the upper Midwest, which is the largest corn‐growing region in the world. REMC provides nutrients for over 4 million acres of farmland, which equates to approximately 5% of 2009 corn acres. Our products help meet the increasing need for nitrogen to restore the fertility and productivity of the soil. In addition to helping to maximize land productivity, a portion of the corn grown by Midwestern farmers using our fertilizer products supports the ethanol industry. Thus, REMC also supports Rentech’s mission of helping America to reduce its dependency on foreign oil for transportation fuels. REMC sells over 600,000 tons of nitrogen products including anhydrous ammonia, liquid and Figure 7. Rentech Energy Midwest Corporation located in granular urea, nitrogen solutions East Dubuque, Illinois. (urea ammonium nitrate solution or UAN). Most of these products are sold to growers and other customers within a 150 mile radius of our plant. Our proximity to our customers provides REMC with a competitive advantage over imported fertilizer manufacturers. In addition, REMC sells nitric acid to industrial customers and carbon dioxide to the food and beverage industry. REMC, our wholly‐owned subsidiary, supplies domestically produced fertilizer to the farming industry, which not only helps farmers and the food industry, but also aids the biofuels industry that uses plants and crops as feedstocks.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 COAL TO LIQUID
Comparative evaluation of different coal to liquid process conditions via Fischer-Tropsch synthesis Serhat Gül*, Atilla Ersoz, Murat Baranak, Omer Faruk Gul, Fehmi Akgun TUBİTAK Marmara Research Centre, Energy Institute, Po. Box 21, Gebze, 41470 Kocaeli, Turkey *Corresponding Author. Tel: +90 262 677 27 91, E-mail: [email protected]
Abstract: The aim of this study is to investigate the optimum design of process configuration and operation parameters for coal to liquid (CTL) fuel production process by using an engineering simulation software tool Aspen HYSYS. The CTL process has several sub systems such as gasification, gas cleaning, gas conditioning and FT units. For gasification process of selected coal (Soma-Turkish lignite), Gibbs reaction model has been used obtaining thermodynamic equilibrium conditions. The set parameters considered in the simulations are the fuel feeding capacity, gasifier temperature and H2O/CO ratio of syngas. At the gas cleaning stage, there are 5 different gas cleaning units which are the removal of H2S unit, cracking/reforming unit for tar and methane, HCl removal unit, particulate removal and finally H2S guard unit. Gas conditioning section consists of water gas shift reactor in order to adjust the H2/CO ratio of syngas for obtaining the stoichiometric ratio for FT process by using equilibrium reactor, and CO2 removal unit where approximately 90% of the CO2 content of the gas is removed. At the end of gas conditioning section, syngas has to be compressed to the level of FT reactor requirements which is minimum 20 bara. FT process consists of reactor itself, off gas condenser unit and off gas turbine for electricity generation. Conversion reactor model is used for FT reactions with 20 bara operation pressure and 260°C operation temperature. The Anderson-SchulzFlorry (ASF) approach is used for determining the product distribution of the liquid fuel. The liquid product compose of CnH2n+2 molecules, where n is in the range of C1 to C30. 1.
Introduction
The three main components of energy for sustainable development are sufficient energy resources, technical and economical feasibility of the technologies and being environmental friendly. Global warming, low reserves of oil and corresponding increase in prices has led to the diversification of energy resources and a considerable change in energy technologies. Coal and biomass, which are more evenly distributed throughout the worldwide have gained strong interest due to their utilization in the production of clean second generation liquid and gaseous fuels besides power generation. One of the main technologies for second generation liquid fuel from solid fuels is Fischer-Tropsch (FT) synthesis of syngas that is produced with gasification process. In this study, the affect of different gasifier conditions on the performance of coal to liquid processes are mainly examined. In fact there are many operational parameters for gasification process and if all those parameters are considered experimentally, it would require the long time for realization. Even with using simulation program, it is hard to cover whole parameters. In order to make it possible, some important operational parameters are selected as variable, and the others are held constant as convenient to previous studies. Various pressure levels of gasifier that has been changed from 2 to 20 bara with 2 bar steps are examined in terms of its effect on performance of the process. Two different gasifier temperatures have been also examined as 900°C and 1400°C to represent the fluidized bed and the entrained bed conditions respectively. The influence of different gasification agents (air, O2, H2O and CO2, reticulated syngas/off gas) has been evaluated. In addition, CO2 removal options have also been investigated before and after the FT reactor. The results of simulation studies of the CTL process concerning liquid fuel efficiency, system electricity consumption rate and quantity/quality of the waste heat of whole system have been presented depending on various design, configuration and operational parameters. In addition to these comparative data, the other operational parameters and mass&energy balance of all units are also given for all conditions.
2.
Motivation and Objectives
This study is conducted for preliminary feasibility analysis of an ongoing project related to the technologies for the liquid fuel production from coal and biomass blend. Within the scope of this project, gasification, high temperature gas clean-up, gas conditioning, CO2 separation, synthesis gas to liquid fuels conversion via FT process, integrated power production technologies will be developed and demonstrated. GAS C L E A N IN G
COAL
O2
H2O ( STEAM)
G A S IF IC A T IO N
P a r tic u la te M a tte r R em oval
H 2S R em oval (S o r b e n t)
Tar and NH3 R em oval
A lk a li R em oval
F iltr a tio n
CO2
H 2S R em oval (P o lish in g )
G A S C O N D IT IO N IN G
STEAM REFORM ER
W ATER GAS S H IF T REACTOR
CO2 REM OVAL
HEAT
HCl R em o v a l
FT S Y N T H E S IS
GAS E N G IN E
L IQ U ID FUEL
E L E C T R IC IT Y
Figure 1. The concept of CTL Process The aim of this study is, to determine the effect of different operational parameters on the outputs of the CTL plant by using Aspen HYSYS simulation program. Mass and energy balance for all sub units have been established considering various operational parameters. Thus, gasifier performance and total CTL plant performance have been determined. 3.
Methodology 3.1. Inspected outputs
The investigated performance parameters of a CTL plants are; -
Liquid fuel efficiency, Electricity consumption/efficiency Heat load of sub-units Heat generation potential/heat quality Gasifier cold gas efficiency and product gas composition
Liquid fuel efficiency ( % ); is the ratio of the total LHV of the product liquid fuel to the total LHV of the input fuel. Electricity load ( % ); is the ratio of producing/consuming electricity load to the total LHV of the input fuel. Amount of waste heat ( % ); is the ratio of producing waste heat to the total LHV of the input fuel. Table1. Different Operational parameters of gasifier
Comparative evaluation of different operational parameters has been performed. There are several operational and design parameters which effect the performance of the CTL plant. In this study, mainly the effects of gasifier conditions such as temperature, pressure and agents on the performance of the CTL plant have been investigated. 3.2. Gasifier The first subsystem of the CTL process is gasification system. Gasification is chemical process where hydrocarbons are converted into combustible gas mixtures consisting of CO, H2, CH4, etc., in the presence of sub-stochiometric oxidizers such as air, oxygen and steam. In this study, thermodynamic equilibrium model is used for gasification reactions. Gibbs free energy minimization reactor is used for reaching the equilibrium conditions. At Gibbs reactor, there is no predefined reaction set. But all possible reaction can take place. Input data for gasifier are, the amount of fuel and fuel composition as original basis without ash content, the desired moisture content of the product gas in order to determine the steam feed, gasification agent, unconverted carbon ratios (assumed as %5 and %3 for fluidized bed and entrained bed conditions, respectively), heat losses of gasifier (%3 of total thermal load), and the desired pressure and temperature of the gasifier. By using these input data, Aspen-Hysys program solve the equilibrium conditions, such as, required steam and oxygen/air and product gas composition and other physical and chemical properties.
Figure 2. Aspen-Hysys process flow diagram of gasifier Soma Coal has been selected for all simulation studies. Analysis of the selected coal as fed basis is shown in Table 2. In simulation studies, ash content of the fuel is not defined as fuel component; instead just consider the thermal effect (heat loss) of the ash content. The thermal effect of the ash on the gasifier is %2 and %3 for fluidized bed and entrained bed, respectively. Table 2. Analysis of selected Turkish coal C (%)
H (%)
O (%)
45,2
2,84
11,5
Coal Analysis S N (%) (%) 1,16
0,94
Moisture (%) 13,75
Ash (%) 24,61
Lower heating value (LHV) of the selected fuel was measured as 3773 kcal/kg with bomb calorimeter. However, with the defined fuel composition, simulation program calculates the LHV of the fuel as 4376 kcal/kg. This difference appears because; Aspen-Hysys considers the fuels as composed of smallest molecules(C, H2, O2, N2, etc.). However in real situation, hydrogen and carbon are in the form of big hydrocarbon molecules and before the combustion process of coal, which is the part of the gasification process, those big hydrocarbon molecules have to be cracked into smaller molecules. Heat is needed for the cracking of those molecules and the LHV value of calorimeter measurements is always less than calculated values. In order to compensate the LHV values difference, some amount of heat has to be taken out from the gasifier at simulation studies.
Compressor has been used for air or oxygen compression to the gasifier, with the isentropic efficiency of %80 where the ambient temperature is assumed as 20 oC. If the oxygen is used as gasification agent, energy consumption for the production of pure oxygen (%95) is assumed as 800 kJ/kg oxygen. Steam has been fed as super heated vapor at 300 oC and at gasifier pressure. 3.3. Gas cleaning The second subsystem of the CTL process is gas cleaning system. Gas cleaning is required for the gas conditioning and FT process. In this study, there is no defined reaction mechanism for the removal of contaminants like tar, sulphur, HCl, etc. Instead of this, contaminants in the product gas is removed with assumed ratio and just the thermodynamic effect of gas cleaning reactors are considered with achieving desired temperature level before and after the reactors. Gas cleaning units do not have considerable effect on the performance of the CTL process with the efficiency point of view.
Figure 3. Aspen-Hysys process flow diagram of gas cleaning There are 5 different gas cleaning units. First unit is for H2S removal and assumed as %90 of the H2S is removed from product gas. With second gas cleaning unit, tar and methane content is cracked. At third unit, HCl is removed from product gas. Following units have been used for particulate and final H2S removal. 3.4. Gas conditioning The third subsystem of the CTL process is gas conditioning system. Gas conditioning systems is required for making product gas to be suitable for FT conditions. Gas conditioning section consists of water gas shift (WGS) reactor, CO2 removal unit and compressor unit in order to increase gas pressure up to FT pressure level. WGS reactor adjusts the H2/CO ratio for the requirements of the FT process. Equilibrium reactor model has been used as WGS reactor and the temperature at equilibrium condition is 450 oC. This temperature has also limited by the catalysts constraints. It is suitable to the operating range of the commercial WGS catalysts.
Figure 4. Aspen-Hysys process flow diagram of gas conditioning
CO2 removal unit takes place after the WGS unit. 90% of the CO2 content in the product gas is removed at this section. In the real process conditions, CO2 will be removed by using MEA-based absorption technology from syngas mixture. At the end of gas conditioning unit, there is a mix gas compressor to pressurize the product gas through the FT reactor. 3.5. FT System/Off-gas turbine The last subsystem of the CTL process is Fischer-Tropsch(FT) synthesis system. Fischer-Tropsch synthesis is a chemical process where H2 and CO reacts under high pressure in the presence of iron/cobalt catalyst. In the reactor, synthesis gas is converted into wax/liquid fuel. The stochiometry of the main reaction is shown at the following equation.
FT synthesis: (2n+1)H2 + nCO ↔ CnH2n+2 + nH2O + heat FT system consists of; FT reactor, off-gas condenser and off-gas turbine. In this study, conversion reactor model is used for FT synthesis reactions. The pressure and temperature of the FT reactor is defined as 20 bars and 260 oC, respectively. FT reactor CO conversion ratio is assumed as 80%.
10,000 9,000 8,000
% ü rü n d a ğ ılım ı
7,000 6,000 5,000 4,000 3,000 2,000 1,000 0,000 C1 C2 C 3 C 4 C5 C6 C7 C8 C 9 C10 C11 C 12 C13 C14 C15 C16 C17 C18 C 19 C20 C21 C22 C23 C 24 C 25 C26 C27 C28 C29 C 30 n ( C n H2n+2 )
Figure 5. Aspen-Hysys process flow diagram of FT process and ASF product distribution The product molar distribution of FT synthesis is defined with the Anderson-Schulz-Florry (ASF) product distribution approach. According to ASF approach, the selectivity of the reactor product stream is determined by α chain growth probability factor. Molar fraction of product hydrocarbons is presented with following equation.
Where N indicates the carbon number, α is the chain growth probability andMN is the molar fraction of linear products with carbon number N. In this study product distribution is assumed to be between C1 to C30 hydrocarbons and α is assumed as 0,87. At the end of CTL process, there is an off-gas turbine, in order to generate electricity.
4.
Results and Discussion
Base case for this simulation study is fluidized bed gasification by using oxygen and steam as gasification agent at the gasifier pressure of 2 bara and CO2 is removed from product gas before the FT reactor. Table 3. CTL plants outputs @ 2 bara gasifier pressure
Table 4. CTL plants outputs @ 20 bara gasifier pressure
At base conditions, liquid fuel efficiency, electricity load and net waste heat generation is, 42,1%, -3,0% and 41,9%, respectively. 4.1. Effect of gasifier type Fluidized bed gasification process under the conditions of 900 °C reactor temperature, and using oxygen and steam as gasification agent, liquid fuel efficiencies have been obtained as 42, 1% and 43, 4% for the gasifier pressure of 2 bara and 20 bara, respectively. However, entrained bed gasification process under the conditions of approximately 1400 °C, liquid fuel efficiencies were 38,6% and 39,3% for the gasifier pressure of 2 bara and 20 bara, respectively. Due to the higher temperature of the entrained bed-means more combustion reactions take place, CTL plant with the entrained bed gasifier has the lower performance than the plant that has the fluidized bed gasifier. There is also slight difference between the electricity load of fluidized bed and entrained bed conditions. This difference is caused by the oxygen consumption rates of gasifier types. Due to the higher oxygen consumption rate of entrained bed, electricity consumption at lower gasifier pressure is higher; electricity generation at higher gasifier pressure is lower for CTL process with entrained bed gasifier. 4.2. Effect of CO2 removal There are two options for CO2 removal process. Removal processes before and after the FT reactor options is investigated. There is considerable difference between two CO2 removal options. If CO2 is removed from product gas before the FT reactor, liquid fuel efficiency has been 42, 1% as base case. If CO2 removal takes place after the FT reactor, liquid fuel efficiency becomes 38, 3 %. This situation can be explained with the partial pressure difference at offgas condenser unit. If CO2 is removed before the FT reactor, partial pressure of gas phase hydrocarbon molecules increases and condensing process becomes easier which means the quality of offgas is decreasing while the quality of liquid fuel is increasing. According to simulation results, partial pressure of hydrocarbon molecules at offgas condenser unit for CO2 removal options of before and after the FT reactor are, 13% and 60%, respectively. 4.3. Effect of gasifier pressure Gasifier pressure also affects the composition of the product gas. While the operation pressure of gasifier rises from 2 bara to 20 bara, CO and CO2 concentration has not been changed, considerably. Due to the high pressure level of gasifier, CH4 concentration generally increases and thus H2 concentration decreases. However, the efficiency of the gasifier almost remains constant with the change of gasifier pressure.
Figure 6. Gas composition (O2+H2O)
Figure 7. Electricity Load (O2+H2O)
As shown in Figure 7, net electricity load (consumption/generation) of total CTL plant has been significantly affected by the gasifier pressure. There is a requirement of reaching FT pressure level which is assumed in this study as 20 bara and one way of reaching FT pressure is compressing everything at the beginning of the CTL process which means high pressure gasifier, that is more efficient but has more complexity, and the other way is compressing product gas just before the FT reactor which means low pressure gasifier but has less complexity. Due to the less volumetric flow of gasifier agent, compression work at the beginning of process is more efficient.
Figure 8. Liquid Fuel Efficiency (O2+H2O) Liquid fuel efficiency of CTL plant is slightly increasing with the increase of gasifier pressure as shown in Figure 8. Due to the higher compression work requirements before the gasifier for pressurized conditions, input energy has been increased and thus, energy content of the product gas has been increased. Because of the higher gasifier output, total CTL plant liquid fuel efficiency has been increased. 4.4. Effect of gasification agent There are 5 different gasification agents investigated. Base case of gasification agent is oxygen and steam mixture at a specified ratio. The other gasification agents are the mixture of steam and oxygen with; air, CO2, recycle syngas and recycle offgas. While the liquid fuel efficiency of the base case is 42,1% at 2 bara gasification pressure condition, the efficiency values decrease to 40,7%, 37,5% and 29,5% for gasification with recycle syngas, CO2 and air, respectively. Due to the extra thermal load effect, gasifier efficiency decrease and thus the liquid fuel efficiency decrease. For the case of using the mixture of oxygen, steam and offgas as gasification agent, the liquid fuel efficiency increases to 46, 5%. FT offgas recirculation to the gasifier tends to increase the conversion of gas to liquid and liquid fuel efficiency increases. 4.5. Energy balances One of the main objectives of this study is determining the mass&energy balance of the CTL system with specified conditions. Energy balances of the base case condition with the operation gasifier pressure of 2 bara and 20 bara, are shown
in Figure 9 and Figure 10, respectively. Inputs of energy balance are, coal or biomass which is defined as 100 units and auxiliary power consumption. Outputs of system are liquid fuel, offgas, waste heat and losses.
Figure 9. Energy balance of CTL plant @ 2 bara gasifier pressure
Figure 10. Energy balance of CTL plant @ 20 bara gasifier pressure
5.
Conclusion
In this study, the affect of different gasifier conditions on the performance of coal to liquid processes are mainly examined. For the sake of simplicity, some important operational parameters are selected as variable, and the others are held constant as convenient to previous studies. Various pressure levels of gasifier are examined in terms of its effect on performance of the process. Two different gasifier temperatures have been also examined as 900°C and 1400°C to represent the fluidized bed and the entrained bed conditions respectively. The influence of different gasification agents has been evaluated. In addition, CO2 removal options have also been investigated before and after the FT reactor. The results of simulation studies of the CTL process concerning liquid fuel efficiency, system electricity consumption rate and waste heat generation of whole system have been presented depending on various operational parameters. Gasifier type has considerable effect on the plant performance. Due to the higher temperature of the entrained bed-means more combustion reactions, CTL plant with the entrained bed gasifier has the lower performance than the plant that has the fluidized bed gasifier. Liquid fuel efficiency changed slightly with the change of gasifier operation pressure, but electricity load has been considerably changed by the gasifier pressure. While the gasifier pressure increase, electricity consumption has been decreased and electricity generation potential has been increased. Oxygen and steam mixture has been selected as base case for gasification agent. According to base case conditions, air gasification has the lowest liquid fuel efficiency and oxygen and FT offgas mixture gasification has the highest liquid fuel efficiency. There are three main outputs of the CTL plant which are; liquid fuel, electricity consumption/generation and waste heat generation. The liquid fuel efficiency varies from 29,5% to 48,7% with respect to process configuration and operational parameters. The electricity load is varied between -5,9% which means consumption of electricity and +4,1% which means there is a potential of net electricity generation. The net waste heat generation varies between 36,8% and 57,2% and the mean temperature of waste heat is varied between 330 °C and 450 °C. This means there is a potential of generating electricity by using steam cycle.
6.
Acknowledgment
This study has been realized within the “Liquid Fuel Production from Biomass and Coal Blends” project. TUBITAK is greatly acknowledged for the support of this project, the contract code 108G043, under the Support Programme for Research Projects of Public Institutions (1007). General Directorate of Electrical Power Resources Survey and Development Administration and Turkish Coal Enterprises Administration are acknowledged for being the customers of the project. Project partners are also acknowledged for their constructive comments and suggestions: Istanbul Technical University, Marmara University, UMDE Engineering, Contracting and Trading Co.Ltd. and HABAŞ Industrial and Medical Gases Production Industries Inc. 7.
References
[1] R. Domenichini, M. Gallio, A. Lazzaretto, Combined production of hydrogen and power from heavy oil gasification, pinch analysis, thermodynamic and economic evaluations, Energy 35, 2010 [2] Giovanilton F. Silva et al, Simulation and optimization of H2 production by auto thermal reforming of glycerol, 10th International Symposium on Process Systems Engineering, 2009 [3] M. Pe´ rez-Fortes et al, Conceptual model and evaluation of generated power and emissions in an IGCC plant, Energy 34, 2009 [4] Yong Heon Kim et al, A simulation study on gas-to-liquid (natural gas to Fischer-Tropsch synthetic fuel) process optimization, Chemical Engineering Journal 155, 2009 [5] Cesat Nieto et al, Simulation of IGCC technologies, influence of operational conditions(environmental and fuel gas production), Revista Energetica40, 2008 [6] Anthony Andrews, Jeffrey Logan, Fischer-Tropsch Fuels from Coal, Natural Gas, and Biomass: Background and Policy, CRC Report for Congress, 2008. [7] Thomas G. Kreutz, Eric D. Larson, Guangjian Liu, Robert H. Williams, Fischer-Tropsch Fuels from Coal and Biomass, 25th Annual International Pittsburgh Coal Conference, 2008. [8] Grabner M. et al, Numerical simulation of coal at circulating fluidized bed conditions, Fuel Process Technologies, 2007 [9] Smitha V. Nathen et al, Gasification of New Zealand coals, a comparative simulation study, Energy & Fuels 22, 2008 [10] Mar Pérez-Fortes et al, Enhanced modeling and integrated simulation of gasification and purification gas units to clean power production, 18th European Symposium on Computer Aided Process Engineering, 2008 [11] David A. Bowen, Techno-economic analysis of hydrogen production by gasification of biomass, Hydrogen, Fuel Cells, and Infrastructure Technologies, FY 2003 [12] Progress Report Greenhouse Gas Emissions Control By Oxygen Firing In Circulating Fluidized Bed Boilers; NETL And Document Control Bldg. 921, Us Department Of Energy, May 2003. [13] J. Andries et al, Pressurized fluidized bed Combustion and Gasification of coal using flue gas recirculation and oxygen injection, Energy Conv. Mgmt., 38, 1997.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Final Manuscript
PROGRAM TOPIC: COAL-DERIVED PRODUCTS TECHNOECONOMIC AND ENVIRONMENTAL LIFE CYCLE ANALYSIS OF COAL AND COAL/BIOMASS TO LIQUIDS FACILITIES Anastasia M Gandrik, R&D Engineer, Idaho National Laboratory PO Box 1625, Idaho Falls, ID 83412-3710, [email protected], 412-443-6550 Vivek P. Utgikar, Department of Chemical and Materials Engineering, University of Idaho 1776 Science Center Drive, Suite 306, Idaho Falls, ID 83404
Abstract A technical and economic evaluation as well as a life cycle emissions analysis of a synthetic fuels plant utilizing coal or a coal/biomass feedstock is presented in the following paper. The feedstock is first converted into synthesis gas (syngas) in a dry-fed, entrained flow gasification process. Syngas, comprised mainly of H2 and CO, is converted into higher hydrocarbons (diesel, naphtha, and liquefied petroleum gas) via the Fischer-Tropsch reaction in a slurry bubble column reactor. A detailed chemical process model was developed using Aspen Plus® process simulation software for the production of approximately 50,000 barrels per day of liquid products. Optimal design requirements and operating conditions were determined to maximize heat integration between different areas of the plant and the benefits of blending biomass with coal to reduce greenhouse gas emissions were studied through the simulation of the process. The Aspen model results were input into the economic model to determine the economic viability of the synthetic fuel facilities using standard evaluation methods. The life cycle emissions analysis presents the well-to-wheel greenhouse gas emissions for the coal and coal/biomass configurations. A comparison of the emissions of the synthetic diesel with conventional petroleum derived diesel is also presented. The following conclusions can be drawn from the technical, economic, and life cycle assessments performed: 1. Coal and coal/biomass based synthetic fuel production are technically feasible processes. However, incorporation of biomass co-feed causes the size of the facility to increase due to its lower heat content and higher moisture content. In addition, the amount of power available for export from the plant decreases with incorporation of biomass. 2. Both processes are economically feasible given today’s market conditions. However, incorporation of CO2 sequestration and biomass cause the rate of return to decrease slightly from coal based production. Overall, coal and coal/biomass based synthetic fuels will provide a highly desirable rate of return for major investors if fuel prices remain above $2.00/gallon (wholesale).
3. Integration of biomass co-feed and CO2 sequestration are both necessary to reduce life cycle, well-to-wheel, greenhouse gas emissions for a coal to synthetic fuels facility to levels below those for imported and/or baseline petroleum diesel. Introduction and Background Energy is considered the fundamental unit of prosperity by many economists. While only containing 4.6% of the world’s population, the U.S. consumes 21.1% of the world’s energy, of which only 15.1% is domestically supplied (EIA 2009). In 2008, the U.S. consumed 99.2 quadrillion BTUs of energy; petroleum products accounted for 37.1 quadrillion of these BTUs, of which 71% were used in the transportation sector. Overall, energy demand for transportation accounted for 28% of the U.S.’s total energy demand (EIA 2009). In 2008, the U.S. consumed 7.1 billion barrels of oil of which 66.3% or 4.7 billion barrels were imported. As developing nations, such as China and India, begin to surpass the U.S. in energy consumption, specifically for transportation use, increased global energy demand will cause petroleum prices to rise. Crude oil is a finite resource, with the majority of reserves among just a handful of nations. Furthermore, the U.S.’s oil well productivity, 18.6 barrels per day per well at its peak in 1972, has decreased to 10.9 barrels per day per well in 2000 (EIA 2009). Therefore, alternative technologies must be developed which increase U.S. energy security through domestic fuel production. One such way is the gasification of coal or coal/biomass blends to produce synthesis gas (H2 and CO) which can be used to generate a variety of higher value end products including liquid transportation fuels, chemical products such as fertilizers, process heat, and electric power. Coal is the U.S.’s most abundant fuel resource with over 400 billion tons of coal reserves of which over 200 billion tons are recoverable. In 2008 1.2 billion tons of coal was mined in the U.S., which corresponds to at least 200 more years of coal consumption based on current usage (EIA 2009). The only drawback of utilizing coal to produce synthetic fuels is that an appreciable amount of carbon dioxide, a major greenhouse gas (GHG), is produced and emitted in the process. Given the implications of climate change it is anticipated that legislation will be passed which will potentially tax carbon dioxide emissions in order mitigate current and future emissions. This can pose a problem for traditional coal to liquids (CTL) facilities given their large carbon footprints; however, incorporating CO2 capture technology and co-feeding biomass can reduce the overall GHG emissions from synthetically derived diesel to levels that meet or beat conventionally derived petroleum diesel. Reducing emissions from transportation fuels, even slightly, would greatly benefit the net carbon dioxide emissions from the U.S. as over 1/3rd of the total CO2 emissions are from the transportation sector (EIA 2009). The following paper addresses the technical, economic, and environmental impact of CTL and coal/biomass to liquids (CBTL) facilities. This is accomplished through a detailed technical and economic evaluation as well as a life cycle emissions analysis of a synthetic fuels plant utilizing coal or a coal/biomass feedstock. The plant produces approximately 50,000 barrels per day of liquid products, including diesel, naphtha, and liquefied petroleum gas (LPG). A detailed chemical process model was developed using Aspen Plus®, steady state process simulation software, to determine optimal design requirements and operating conditions to maximize heat integration between different areas of the plant and to study the benefits of blending biomass with coal to reduce GHG emissions. The models results are input into the economic model to determine the economic viability of the synthetic fuel facilities using standard evaluation methods. The lifecycle emissions analysis presents the well to wheel GHG emissions for the coal and coal/biomass configurations, specifically comparing the emissions of the synthetic diesel with conventional petroleum derived diesel.
Process Modeling Analysis The plant models for the CTL and CBTL processes were developed using Aspen Plus®. Aspen Plus is a steady-state chemical process simulator that includes extensive thermodynamic databases, built-in routines for common unit operations, and the ability to properly handle complex chemical feedstocks such as coal and biomass (Aspen Tech 2010). Because of the size and complexity of the processes modeled, the simulations were constructed using “hierarchy” blocks, a method for nesting one simulation within another simulation. In this fashion, submodels for each major plant section were constructed separately and then combined to represent the entire process. Significant emphasis in the models has been placed on heat integration between different parts of the plant. To facilitate energy tracking, Aspen’s “utility” blocks were used extensively. Two cases were identified for modeling: •
CTL process
•
CTL process with biomass co-feed, CBTL
For the CTL case, a generic Montana subbituminous coal was used as the feedstock (Higman and van der Burgt 2008). When biomass is co-fed, a generic hybrid poplar biomass was blended with the Montana coal (Energy Research Center of the Netherlands 2009). Table 1 presents the proximate and ultimate analyses of feedstocks used in the process and the higher heating values, all on a dry basis. The moisture content of the biomass was assumed to be 50% (Phillips, et al. 2007). Biomass is co-fed at 25 weight% after drying. Shell gasifiers have been operated with up to 30% biomass co-feed in Buggenum (Nuon 2005). Table 1 – Feedstock Proximate and Ultimate Analyses 1 Moisture Proximate Analysis, Dry Basis Moisture Fixed Carbon Volatile Matter Ash Ultimate Analysis, Dry Basis Ash Carbon Hydrogen Nitrogen Chlorine Sulfur Oxygen Higher Heating Value (BTU/lb)
Montana Subbituminous 10.5
Hybrid Poplar 50
10.5 48.72 38.77 12.51
50 12.5 84.8 2.7
12.51 66.84 4.9 1.49 0 1.22 13.04 11,961
2.7 50.2 6.06 0.6 0.01 0.02 40.41 8,178
A dry-fed, entrained-flow, slagging gasifier (similar to a Shell, Uhde, or Siemens design) was selected as the gasification technology for this evaluation with a capacity of 3,968 tons/day after drying. The number 1
All results are presented on a dry, weight-% basis, except the moisture content.
of gasifiers was set to produce at least 60 MMSCFH of syngas for feed to the FT reactors, any excess syngas was diverted to the gas turbines to produce electricity as the gasifiers were assumed to be operated at capacity. The liquid products produced include diesel, naphtha, and liquefied petroleum gas. The block flow diagram for the CTL/CBTL case is shown in Figure 1. The proposed process includes unit operations for air separation, coal milling and drying, coal gasification, syngas cleaning and conditioning, sulfur recovery, CO2 compression/liquefaction, FT synthesis, product upgrading and refining, power production, cooling towers, and water treatment. Each unit operation is briefly described below, with additional detail in the following sections. The same submodels were used for both the CTL and CBTL cases, only the resulting material flows changed.
Figure 1 – Block Flow Diagram for the CTL and CBTL Processes •
Air Separation – Oxygen is produced for the traditional CTL process via a standard cryogenic Linde type air separation unit (ASU) that utilizes two distillation columns and extensive heat exchange in a cold box. The oxygen product is used for gasification. In order to reduce the inert content in the synthesis gas, an O2 purity of 99.5% is specified.
•
Coal Milling & Drying (CMD) – Coal is pulverized to below 90 μm using a roller mill to ensure efficient gasification. Drying is accomplished simultaneously using a heated inert gas stream.
•
Gasification – Coal is gasified using Shell’s SCGP technology (entrained-flow, dry-fed, slagging, oxygen-blown, upflow gasifier). Heat is recovered in the membrane wall and in the syngas coolers. The syngas is further cooled by a water quench.
•
Syngas Cleaning & Conditioning – After gasification, a fraction of the syngas is passed through sour shift reactors and then remixed with the unshifted syngas to provide the optimal hydrogen to carbon monoxide ratio for FT synthesis. Elemental mercury is then captured in a mercury guard bed. The syngas is further treated for the removal of CO2, H2S, and COS, using the Rectisol process. A portion of the syngas is passed to a pressure swing absorption (PSA) unit where hydrogen is separated and diverted for product refining and upgrading.
•
Sulfur Plant – Sulfur recovery is based on the Claus process. Tail gas from the Claus unit is hydrogenated to convert the remaining sulfur species to H2S, and this stream is recycled to the Rectisol unit to maximize sulfur recovery.
•
CO2 Compression – Purified CO2 from the Rectisol process is compressed and liquefied prior to being pumped to the required pressure for or sequestration.
•
Fischer-Tropsch Synthesis – Syngas is converted to liquid synthetic crude in slurry bubble column reactors using a cobalt catalyst. Separations are performed to fractionate the product into light gas, crude naphtha, middle distillate, and molten wax.
•
Product Upgrade & Refining – The middle distillate product is hydrotreated to saturate olefinic bonds. The hydrotreated product is refined via a combination of pressurized and vacuum distillation into naphtha and diesel fuel products. The bottoms product from vacuum distillation and the molten wax stream are hydrocracked to improve overall yield of the diesel and naphtha fractions.
•
Power Production – Light gas from the FT and refining areas is used to fire gas turbines to produce electricity for the plant. To increase power production, a combined cycle is utilized. Hot exhaust from the gas turbine is routed to the heat recovery steam generator (HRSG) to produce superheated steam. This steam is used in conventional condensing turbines to produce additional power. To further maximize power production, saturated steam produced in the FT process and in other areas of the plant is fed to saturated steam turbines.
•
Cooling Towers – Conventional cooling towers were assumed; however, air cooling could potentially be used in certain areas of the plant to decrease water consumption.
•
Water Treatment – Water treatment is simplistically modeled in Aspen Plus, as wastewater contaminants are not modeled in detail at this point.
Process Modeling Results A summary of the modeling results for the cases analyzed is presented in Table 2 for the CTL and CBTL cases. Again, the number of gasifiers was set to produce at least 60 MMSCFH of syngas for feed to the FT reactors, any excess syngas was diverted to the gas turbines to produce electricity as the gasifiers were assumed to be operated at capacity. From the results below it is apparent that the number of gasifiers required is greater for the CBTL case. This is because the feed for the CBTL case has a higher moisture content than the CTL case and the biomass co-feed to the gasifier results in significantly more CO2 being produced in the gasifier. This is because more heat is required via the combustion reaction to vaporize the moisture fed to the gasifier, as the biomass cannot be dried to as low a level as the subbituminous coal without devolatilization occurring. As a result, to produce the feed requirement to fully load the FT reactors an additional gasifier is required. Furthermore, the power production decreases by 38% in the CBTL case, mainly due to the increased power load required to dry the biomass co-feed and the increased oxygen requirement in the gasifier. Both cases produce nominally the same amount of liquid fuel products. A high-level material and energy balance summary for the CTL and CBTL cases are graphically presented in Figure 4.
Figure 2 – CTL and CBTL Case Summary
Table 2 – CTL and CBTL Case Aspen Model Summary Results Inputs # Gasifiers Gasifier Capacity, After Drying (tonne/day) Raw Coal/Biomass Feedrate (ton/day) Coal (ton/day) Biomass (ton/day) Dry Feedrate, Coal to 6%, Wood to 30% (ton/day) Coal (ton/day) Biomass (ton/day) Utility Summary Net Power (MW) Power Generated Gas Turbine Condensing Steam Turbines Saturated Steam Turbines Power Consumed Coal Milling & Drying Air Separation Gasifier Island Gas Cleaning Claus Sulfur Reduction Power Block CO2 Compression/Liquefaction Fischer Tropsch Synthesis Refinery Refrigeration Cooling Towers Water Treatment Water Requirements Water Consumed (gpm) Water Consumed/Coal Feed (lb/lb) Water Consumed/bbl Product (bbl/bbl) Outputs Total Liquid Products (bbl/day) Diesel Naphtha LPG Carbon Balance Summary % Carbon to Liquid Fuel % Carbon to Slag and Flyash % Carbon to Sequestration or EOR % Carbon to HRSG Exhaust % Unaccounted Carbon
CTL
CBTL
7 3,600 29,173 29,173 0 27,776 27,776 0
8 3,600 36,115 25,005 11,110 31,744 23,808 7,936
325 1,090 380 236 475 765 15 320 28 154 0 5 7 149 15 11 24 16 21
123 999 313 203 482 876 42 347 66 160 0 5 7 161 15 11 24 16 22
21,431 4.41 14.4
21,822 3.63 14.7
50,922 35,844 12,947 2,131
50,846 35,789 12,920 2,137
30.3 0.4 53.0 16.2 0.2
29.7 0.4 55.5 14.2 0.2
Table 2 – CTL and CBTL Case Aspen Model Summary Results CO2 Summary Total CO2 Produced (ton/day) Emitted (ton/day) From Coal From Biomass Captured (ton/day) From Coal From Biomass % Capture
CTL
CBTL
43,821 10,332 10,332 0 33,489 33,489 0 76
44,937 9,240 6,276 2,964 35,697 24,424 11,273 79
Economic Analysis The economic viability of the CTL and CBTL processes was assessed using standard economic evaluation methods. The economics were evaluated for both options described in the previous sections. The total capital investment (TCI), based on the total equipment costs, annual revenues, and annual manufacturing costs were first calculated for the cases. The present worth of the annual cash flows (after taxes) were then calculated for the TCI, as well as the TCI at +/- 30%. The following assumptions were made for the economic analyses: •
The plant startup year is 2013
•
A construction period of three years is assumed
•
o
Plant construction begins in 2010
o
Percent capital invested for the fossil plant is 33% per year
Plant startup time is one year o
Operating costs are 85% of the total value during startup
o
Revenues are 60% of the total value during startup
•
The analysis period for the economic evaluation assumes an economic life of 30 years, excluding construction time (the model is built to accommodate 20 to 40 years)
•
An availability of 92% was assumed, the plants are assumed to operate 365 days a year, 24 hours per day
•
An inflation rate of 2.5% is assumed
•
Debt to equity ratios of 80%/20% and 55%/45% are calculated; however, results are only presented for 80%/20% debt to equity ratios as the first major CTL or CBTL facility would most probably have loan guarantees
•
o
The interest rate on debt is assumed to be 8%
o
The repayment term on the loan is assumed to be 15 years
The effective income tax rate is 38.9% o
State tax is 6%
o
Federal tax is 35%
•
MACRS depreciation is assumed, with a 15 year plant life
•
A CO2 tax of $0/ton to $100/ton CO2 emitted is investigated for both cases
The following sections describe the methods used to calculate the capital costs, annual revenues, annual manufacturing costs, and the resulting economic results. Capital Cost Estimation Equipment items for this study were not individually priced. Rather, cost estimates were based on scaled costs for major plant processes from published literature. Cost estimates were generated for coal preparation, the ASU, gasification, gas cleanup, FT synthesis, product refining and upgrading, gas turbines, steam turbines, the HRSG, and cooling towers. In some instances, several costs were averaged. Gas cleanup includes costs for water-gas-shift reactors, the Rectisol process, sulfur recovery, and CO2 compression/liquefaction. When sequestration is not assumed the cost of the final compressor stage is deducted from the CO2 compression capital cost. The estimate presented is a study (factored) estimate which has a probable error up to +/-30% (Perry and Green 2008). Estimation of Revenue Yearly revenues were estimated for all cases based on recent price data for the various products generated. Revenues were estimated for low, average, and high prices for diesel and naphtha. High prices correspond to prices from July 2008, low prices are from March 2009, and average prices were the average of the high and low values. Diesel prices were obtained from the Energy Information Administration (EIA 2010) and represent wholesale prices and do not include taxes. Naphtha prices were scaled based on diesel prices. Selling prices for LPG, electricity, slag, and sulfur were not varied in the study; this was a reasonable assumption since these prices historically follow the standard rate of inflation and do not vary widely during the year, or they only make up a small fraction of the yearly revenue. Revenues were also calculated to determine the necessary selling prices of diesel and naphtha to achieve a 12% rate of return. Estimation of Manufacturing Costs Manufacturing cost is the sum of direct and indirect manufacturing costs. Direct manufacturing costs for this project include the cost of raw materials, utilities, and operating labor and maintenance. Indirect manufacturing costs include estimates for the cost of overhead and insurance and taxes (Perry and Green 2008). Maintenance costs were assumed to be 3% of the total capital investment (Jones and Rujado 2006). Taxes and insurance were assumed to be 1.5% of the total capital investment, an overhead of 65% of the labor and maintenance costs was assumed, and royalties were assumed to be 1% of the coal cost or coal/biomass cost (Jones and Rujado 2006). The annual manufacturing costs include costs for sequestration, in the model an analysis was performed to assess the impact of sequestering or not sequestering CO2 on the economics. For the CBTL case emissions of CO2 from biomass are deducted from the total emissions. Economic Modeling Results Table 3 presents the results for an 80%/20% debt to equity ratio for the CTL and CBTL cases. Figure 3 depicts the associated IRR results for the CTL and CBTL cases at TCI. Table 4 presents the carbon tax results for both cases and Figure 4 depicts the carbon tax results.
Table 3 – CTL and CBTL IRR Results CTL
CTL with Sequestration
CBTL
TCI
TCI
TCI
IRR $/gal $5,517,744,563.05 10.3 $1.41 28.7 $2.72 43.5 $4.04 12.0 $1.08
IRR $/gal $5,517,744,563.05 4.2 $1.41 24.0 $2.72 39.6 $4.04 12.0 $1.08
IRR $/gal $6,179,873,910.62 4.7 $1.41 23.4 $2.72 38.3 $4.04 12.0 $1.08
CBTL with Sequestration TCI IRR $/gal $6,179,873,910.62 -4.1 $1.41 18.5 $2.72 34.1 $4.04 12.0 $1.08
Figure 3 – CTL and CBTL IRR Results at TCI From these results it is apparent that both the CTL and CBTL cases, with or without sequestration, are profitable at average fuel prices, and extremely profitable at high fuel selling prices. When sequestration is assumed both cases returns decrease by about 4% when fuel prices are fixed at the low, average, and high prices. When the IRR is fixed at 12%, the corresponding fuel price must increase by about $0.30 to $0.40 to provide the same rate of return if sequestration is assumed. This is due to the increase in capital and manufacturing costs as well as the decrease in revenues when sequestration is assumed. Implementing biomass co-feed also decreases the IRR by about 5% when compared to CTL. Again, this is due to increased capital and manufacturing costs and a decrease in revenues since export power decreases when biomass is co-fed. Given this decrease, if fuel prices fall below $2.00 per gallon the CBTL option could prove to have greater financial risk.
Table 4 – CTL and CBTL Carbon Tax Results Carbon Tax
CTL
$/ton CO2 0 25 50 75 100
$/gal 1.51 2.20 2.88 3.56 4.24
CTL with Sequestration $/gal 1.87 2.03 2.19 2.35 2.51
CBTL 1.87 2.03 2.19 2.35 2.51
CBTL with Sequestration $/gal 2.24 2.34 2.44 2.54 2.64
Figure 4 – CTL and CBTL Carbon Tax Results at TCI and 12% IRR From the carbon tax results it is apparent that sequestration is key to both cases remaining financially viable if a carbon tax is implemented on CO2 emissions. In fact, even at a $100/ton CO2 tax, when sequestration is assumed an IRR of 12% can be achieved for both cases for a selling price of diesel less than the average price assumed in this study. When sequestration is not assumed, the CBTL case financially outperforms the CTL case when the carbon tax exceeds about $45/ton CO2. When sequestration is assumed, it is apparent that the CBTL case would eventually financially outperform the CTL case; however, the carbon tax must be greater than $100/ton CO2. In addition, a sensitivity of the economic analysis data was produced to determine the impact of adding biomass co-feed, sequestration, etc. to the baseline CTL case, with respect to the average selling price of diesel. The results can be found in Figure 5. From these results it is evident that incorporation of
biomass does increase the selling price of diesel necessary for the same return; however, the corresponding price is less than the average diesel price used in this study.
Figure 5 – Economic Sensitivity Analysis Results Greenhouse Gas Emission Analysis This section presents a full life-cycle inventory or well-to-wheel (WTW) analysis of greenhouse gas emissions for the production of synthetic diesel using the CTL and CBTL processes described in the preceding sections. For this study, all results are scaled for the diesel, naphtha, LPG, and/or electricity products. This is accomplished by ratioing the lower heating values of the products along with the electricity, if produced in the plant, to determine the emissions assignment, or the percentage of the total energy content for the diesel, naphtha, LPG, and/or electricity product. LPG, naphtha, and diesel and all have similar heating values on a mass basis; thus, including the LPG and naphtha with the diesel product has no appreciable impact on overall WTW emissions. The emissions for the diesel product are converted to a gram per mile basis using a vehicle fuel economy of 25.8 miles per gallon. The fuel economy was adjusted to account for the heating value of the synthetic fuel versus traditional petroleum derived products (SAE 1999). The vehicle fuel economy represents the average mileage of a diesel powered SUV. The GHG emissions considered in this report include carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). Emissions for CH4 and N2O are converted into CO2 equivalents using their global warming potentials (GWP). CO2 equivalents are the amount of carbon dioxide by weight emitted into the atmosphere that would produce the same radiative force as a given weight of another radiatively active gas. The GWPs used in this report are referenced from the Intergovernmental Panel on Climate Change’s (IPCC) most recent climate study in 2006. The 100-year GWP for CH4 and N2O are 23 and 296,
respectively (IPCC 2006). The WTW analysis conducted for this study was based on the formal methodology presented by NETL in the Life-Cycle Greenhouse-Gas Emissions Inventory for FischerTropsch Fuels, and categorizes GHG emissions according to the following sources (NETL 2001): •
Resource extraction – GHG emissions for resource extraction are calculated for the two feeds considered in this study, coal and biomass. Coal extraction emissions include emissions from fuel usage associated with coal mining and coal bed methane. Biomass production emissions include emissions associated with biomass cultivation and harvesting as well as the credit for GHGs absorbed during biomass regrowth, which is directly related to the carbon content of the biomass.
•
Transportation of resources and products – All scenarios considered in this study include transportation of resources and products over large distances. The mode of transportation depends upon the location and destination of the products as well as the type of product being transported. For instance, dry materials being transported short distances would utilize trucks as the main mode of transportation, while dry materials being transported long distances near major rivers would take advantage of barge transportation. The modes of transportation were assumed based on the amount of product being transported, the product state, the distance transported, and the available transportation methods.
•
Conversion and refining of the product – GHG emissions are generated from several sources within the CTL and CBTL plants, including: emissions from importing power, upstream emissions associated with methanol use, Rectisol plant emissions, HRSG stack emissions, fired heater emissions, high pressure (HP) and low pressure (LP) flare systems, and fugitive emissions.
•
End use combustion of the product – Emissions for the end use combustion of the fuel were estimated from the carbon content of the synthetic diesel. It was assumed that all carbon present in the fuel is completely combusted to form CO2. Based on the fuel density, an output from the Aspen model, this would provide the emissions of CO2 per barrel of fuel.
Results from the WTW analysis for FT diesel were compared to WTW emissions for the U.S. baseline and average imported WTW emissions for conventional diesel fuel to determine the environmental impact of the synthetic fuels in comparison to standard petroleum fuels. The U.S. baseline and average imported WTW emissions for diesel were derived from a 2009 NETL refinery report (NETL 2009). Greenhouse Gas Emission Results A summary of the GHG results for the cases analyzed is presented in Table 5 for the CTL and CBTL diesel products. Fuel GHG emissions results are presented on a gram CO2 equivalent per barrel of diesel fuel (g CO2-eq/bbl) basis, a gram CO2 equivalent per LHV (g CO2-eq/MMBTU), and a gram CO2 equivalent per mile (g CO2-eq/mile). Fuel GHG emission results are presented in Figure 6.
Table 5 – Lifecycle GHG Emission Results for Diesel Product
gCO2-eq/bbl diesel Extraction and Production Transportation to Plant Conversion and Refining Transportation to Pump End Use Combustion Total Fuel Chain gCO2-eq/MMBTU diesel Extraction and Production Transportation to Plant Conversion and Refining Transportation to Pump End Use Combustion Total Fuel Chain gCO2-eq/mile Extraction and Production Transportation to Plant Conversion and Refining Transportation to Pump End Use Combustion Total Fuel Chain
CTL
CTL w/ Seq.
CBTL
CBTL w/ Seq.
Baseline Diesel
Imported Diesel
16,563 1,501 707,746 1,089 360,375 1,087,274
17,148 1,554 177,579 2,717 360,375 559,373
-159,083 3,978 771,902 1,089 360,375 978,260
-159,083 3,978 170,460 2,943 360,375 378,673
35,894 7,070 51,666 4,895 439,910 539,434
45,683 9,245 57,104 4,351 439,910 556,293
3,431 311 146,622 226 74,658 225,247
3,553 322 36,789 563 74,658 115,884
-32,956 824 159,909 226 74,656 202,659
-32,956 824 35,313 610 74,656 78,447
6,600 1,300 9,500 900 80,888 99,188
8,400 1,700 10,500 800 80,888 102,288
17 2 736 1 375 1,130
18 2 185 3 375 582
-165 4 802 1 375 1,017
-165 4 177 3 375 394
33 7 48 5 406 498
42 9 53 4 406 513
Figure 6 – Summary Lifecycle GHG Emission Results for Diesel Product
From the results presented in the tables and figures above, to reduce WTW GHG emissions to levels below imported and/or baseline conventional diesel incorporation of biomass and CO2 sequestration are both necessary. CTL WTW emissions are significantly higher than conventional diesel and even with incorporation of sequestration emissions are greater than conventional fuels. However, it is possible to reduce GHG emissions below conventional diesel with incorporation of biomass co-feed but sequestration would still be necessary. Conclusions The conversion processes of CTL and CBTL are both technically feasible processes. However, given the lower heat content and high moisture content of biomass versus coal, a slightly larger facility is required to produce the same volume of liquid fuel products. In addition, less electricity is produced in the CBTL plant for a number of reasons, including the large amount of steam required to dry the biomass for feed preparation. Both CTL and CBTL processes are economically feasible given today’s market conditions. However, incorporation of CO2 sequestration and biomass causes the rate of return to decrease slightly from the coal based production. Overall, coal and coal/biomass based synthetic fuels will provide a desirable rate of return for major investors if fuel prices remain above $2.00/gallon (wholesale). If a carbon tax is implemented based on CO2 emissions, it will be imperative that any synthetic fuels facility incorporate CO2 sequestration. A significant portion of the CO2 emitted could be discounted from the overall emissions since a portion of the CBTL feed is from a renewable energy resource, depending on Climate Change legislation. Integration of biomass co-feed and CO2 sequestration are both necessary to reduce WTW GHG emissions to levels below imported and/or baseline conventional diesel. Conventional CTL WTW emissions are significantly higher than conventional diesel and even with incorporation of sequestration emissions are greater than conventional fuels. It is possible to reduce GHG emissions below conventional diesel with incorporation of biomass, but sequestration is still necessary. Therefore, if there is policy enacted which legislates that synthetically produced diesel fuels must meet or beat current fuel WTW GHG emissions; incorporation of biomass and CO2 sequestration would be a necessity. Lifecycle emissions from the electricity available for sale from the CTL and CBTL processes are lower than the average U.S. Energy mix even without sequestration. Acronyms ASU
Air Separation Unit
CBTL Coal/Biomass to Liquids CMD Coal Milling and Drying CTL
Coal to Liquids
EIA
Energy Information Administration
GHG
Greenhouse Gas
GWP
Global Warming Potential
HP
High Pressure
HRSG Heat Recovery Steam Generator IPCC
Intergovernmental Panel on Climate Change
LP
Low Pressure
LPG
Liquefied Petroleum Gas
NETL National Energy Technology Laboratory PSA
Pressure Swing Absorption
SAE
Society of Automotive Engineers
TCI
Total Capital Investment
WTW Well to Wheel References
Aspen Tech. Aspen Plus version 2006. Burlington, 2010. Davis, Stacy C., Susan W. Diegel, and G. Boundy Robert. Tansportation Energy Data Book (Edition 2). ORNL-6984, Oak Ridge: ORNL, 2009. EIA. Annual Energy Review 2008. June 26, 2009. http://www.eia.doe.gov/emeu/aer/contents.html (accessed March 18, 2010). —. Petroleum Navigator. March 1, 2010. http://tonto.eia.doe.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=A233600002&f=M (accessed March 1, 2010). Energy Research Center of the Netherlands. ECN Phyllis the Composition of Biomass and Waste. July 5, 2009. http://www.ecn.nl/phyllis/ (accessed October 23, 2009). Higman, Christopher, and Maarten van der Burgt. Gasification (2nd Edition). Houston: Elsevier, 2008. IPCC. 2006 IPCC Guidelines for National Greenhouse Gas Inventories. IPCC National Greenhouse Gas Inventories Programme, Hayama: IGES, 2006. Jones, David S. J., and Peter R. Rujado. Handbook of Petroleum Processing . Dordrecht: Spring, 2006. NETL. An Evaluation of the Extraction, Tansport and Refining of Imported Crude Oils and the Impact of Life Cycle Greenhouse Gas Emissions. DOE/NETL-2009/1362, Morgantown: NETL, 2009. NETL. Baseline Technical and Economic Assessment of a Commercial Scale Fischer-Tropsch Liquids Facility. DOE/NETL-2007/1260, Morgantown: NETL, 2007. NETL. Life-Cycle Greenhouse-Gas Emissions Inventory for Fischer-Tropsch Fuels. Morgantown: NETL, 2001. NETL. Shell Gasifier IGCC Base Cases. PED-IGCC-98-002, Morgantown: NETL, 2000.
Nuon. "Update of Nuon Power Buggenum Plant Performance and Fuel Flexibility." IGCC 2005. Pittsburgh: IGCC, 2005. Perry, Robert H., and Don W. Green. Perry's Chemical Engineers' Handbook (8th Edition). New York: McGraw Hill, 2008. Peters, Max S., and Klaus D. Timmerhaus. Plant Design and Economics for Chemical Engineers (5th Edition). New York: McGraw Hill, 2002. Phillips, S., A. Aden, J. Jechura, D. Dayton, and T. Eggeman. Thermochemical Ethanol via Indirect Gasification and Mixed Alcohol Synthesis of Lignocellulosic Biomass. NREL/TP-510-41168, Golden: NREL, 2007. SAE. Emissions from Buses with DDC 6V92 Engines Using Synthetic Diesel Fuel. 199-01-1512, Dearborn: SAE International, 1999. U.S. Census Bureau. 2007 Economic Census. October 30, 2009. http://factfinder.census.gov (accessed November 16, 2009). U.S. DOT. Freight in America - A New National Picture. Washington, D.C.: Bureau of Trasportation Statistics, 2006. U.S. EPA. Inventory of U.S. Greenhouse Gas Emissions and Sinks. Washington D.C.: EPA, 2009.
CONVERSION OF WASTE BIOMASS TO TRANSPORTATION FUELS: ENERGY FOR THE FUTURE S R K Rao, S K Srivastava and Amalendu Sinha Central Institute of Mining & Fuel Research, Digwadih Campus P.O.F.R.I.; DHANBAD-828108, Jharkhand State
ABSTRACT The growing availability of economically competitive bio-based alternatives to petroleum can be attributed mainly to advances in the production and use of transportation fuels. The rising exhaust gas emission standards and the increasing demand for cleaner and cheap energy sources prompt oil companies to develop technologies such as the Gas to Liquid (GTL) technology which showed great potential. Parallel to the GTL Technology, strong research and development activities have started on the Biomass to Liquid (BTL) production process. This process is envisioned to be more carbon dioxide neutral than using fossil fuels because the primary energy source is renewable plant matter. In India, wild variety of Babool wood is available in huge quantity, which is practically a waste biomass, not being used for any other purpose. First time in India, a 50 liter/day capacity pilot plant has been built up wherein firstly dried Babool wood is gasified in a 2100 Kg/hour feed gasifier, giving rise to real gases containing mainly Hydrogen 19.1%, Carbon Monoxide 17.2%, Nitrogen 50.4% along with other gases viz.methane, carbon dioxide and oxygen. From the gaseous mixture thus produced, moisture and oxygen have been removed. Additional hydrogen has been added to the product gas to make the carbon monoxide to hydrogen ratio of 1:2. This synthetic gas was then converted into transportation fuels in Fisher-Tropsch Synthesis process using catalyst in a fixed bed reactor at the desired temperature. Typical products obtained are: straight run gasoline 48-50%, Jet Fuel 21-25%, Diesel Fuel 22-25% and soft (Pharmaceutical Grade) Wax 1-10%.
1
Introduction: During World War II over a million gasifiers were built for the civilian sector while the military used up all the gasoline. Now that world oil supplies are being depleted and global warming is perceived as a threat to our environment, there is renewed interest in gas from BIOMASS. "Woodgas" is the name for the various gases that can be easily made from wood or biomass. Various forms are: Synthesis gas- typically 40% hydrogen, 40% carbon monoxide, 3% methane, and 17% Carbon dioxide, Producer gas- made by gasifying biomass with air (and therefore containing ~50% nitrogen); pyrolysis gas, similar to synthesis gas, but include lots of water and tar as also accompanied by production of 10-30% charcoal. Biomass is one of the most promising renewable energy sources showing great variety through thermochemical processing routes.Through Biomass gasification via Fischer Tropsch (FT) synthesis route, transportation fuels can be produced to replace existing gasoline and diesel. In the course of FT synthesis numerous products are formed, which consequently influence fuel characteristics.
Biomass resources include any organic matter available on a renewable basis, including dedicated energy crops and trees, agricultural food and feed crops, agricultural crop wastes and residues, wood wastes and
2
residues, aquatic plants, animal wastes, municipal wastes, and other waste materials. Material handling, collection logistics and infrastructure are important aspects of the biomass resource supply chain.
BABOOL USED AT CIMFR
In CIMFR, various that using BABOOL
research works has resulted wood as the BIOMASS gives
many useful liquid
fuel products.
Acacia Arabica popularly known as 'Babool' is originated from the Sind province of Pakistan. It occurs wild in the Indian subcontinent & tropical Africa. It is also cultivated for its bark. The leaves, the bark, the pods & the gum of the tree have medicinal virtues. It is medium sized tree with blackish-grey to brown fissured bark. The reference literatures define the vegetable gum, as a group of plant product resembling carbohydrates. They are composed of carbon, hydrogen and oxygen with small quantities of mineral matter constituents and sometimes a little nitrogen and small quantities of tannin may also be present in some gums. But the
3
gums contain more than these that can not be detected by modern analytical methods. They are not aware of its chemical constituents and properties but well aware of its unique medicinal properties and utilities.
PRODUCTS AND THEIR USES. PRODUCTS
USES
CH4, C2-C4 hydrocarbon.
Industrial and Household.
Naptha – C5-C10 (low octane number) Middle distillate c9-c20
Gasoline Blending Jet and diesel fuel
Medium gas oil C16-C25
Marine diesel fuel
Heavy gas oil C22+
Lube oil, waxes
4
Synthesis gas production The initial reactants (synthesis gases) used in the Fischer-Tropsch process are hydrogen gas (H2) and carbon monoxide (CO). These chemicals are usually produced by one of two methods: 1. The partial combustion of a hydrocarbon:
5
CnH(2n+2) + ½ nO2 → (n+1)H2 + nCO
When n=1 (methane), the equation becomes :2CH4 + O2 → 4H2 + 2CO 2. The gasification of coal, biomass, or natural gas: CHx + H2O → (1+0.5x)H2 + CO The value of "x" depends on the type of fuel. For example, natural gas has a greater hydrogen content (from x=4 to x≈2.5) than coal (x<1). The energy needed for this endothermic reaction is usually provided by the (exothermic) combustion of the hydrocarbon source with oxygen. The mixture of carbon monoxide and hydrogen is called synthesis gas or syngas. The resulting hydrocarbon products are refined to produce the desired synthetic fuel. The carbon dioxide and carbon monoxide is generated by partial oxidation of coal and wood-based fuels. The utility of the process is primarily in its role in producing fluid hydrocarbons from a solid feedstock, such as coal or solid carbon-containing wastes of various types. Non-oxidative pyrolysis of the solid material produces syngas which can be used directly as a fuel without being taken through Fischer-Tropsch transformations. If liquid petroleum-like fuel, lubricant, or wax is required, the Fischer-Tropsch process can be applied.
6
Biomass conversion Processes
RENEWABLE ENERGY IN USE
7
BIOENERGY CONVERSION
PROCESS CONDITIONS Generally, the Fischer-Tropsch process is operated in the temperature range of 150-300°C (302-572°F). Higher temperatures lead to faster reactions and higher conversion rates, but also tend to favour methane production. As a result the temperature is usually maintained at the low to middle part of the range. Increasing the pressure leads to higher conversion rates and also favours formation of long-chained alkanes both of which are desirable. Typical pressures are in the range of one to several tens of atmospheres. Chemically, even higher pressures would be favorable, but the benefits may not justify the additional costs of high-pressure equipment. A variety of synthesis gas compositions can be used. For cobalt-based catalysts the optimal H2:CO ratio is around 1.8-2.1. Iron-based catalysts promote the water-gas-shift reaction and thus can tolerate significantly lower ratios. This can be important for synthesis gas derived from coal or biomass, which tend to have relatively low H2:CO ratios (<1). Predominant source of world’s fuel and chemical is petroleum crude oil and estimated reserves of methane & coal exceeds that of crude oil by factors of about 1.5 and 25, respectively. Production of syngas from methane or coal and conversion of the syngas to a range of fuels and chemicals could become increasingly
8
of interest as the reserves of crude oil are depleted and/or the price of crude rises. The industrial application of the FT synthesis started in Germany and by 1938 there were nine plants in operation having a combined capacity of about 660 X 103 t per year [1]. These plants ceased to operate after the World War II; however interest in the FT process remained because at this stage there was the persistent perception that the reserves of crude oil were very limited. Based on syngas produced from methane, an FT plant with a capacity of 360 X 103 t per year was built and operated in Brownsville, TX, during the 1950s but a sharp increase in the price of methane caused the plant to be shut down. During the same time period, based on the world-wide prediction of increasing crude oil prices an FT plant based on coal came on stream in Sasolburg, South Africa. Even before construction of this plant was completed, however, the huge oil fields of the Middle East were discovered and consequently the predicted rise in the price of crude oil did not materialize and interest in the FT process all but disappeared. The economic viability of the FT synthesis depends on the price of crude oil and as it has varied considerably over last three decades (Fig. 1), the decision to construct an FT plant is always associated with risk. The oil crises of the mid 1970s prompted Sasol to construct two much larger coal-based FT complexes and this time things worked out better as the two plants came on-line in 1980 and 1982 when the price of crude exceeded US$ 30 per barrel. At that stage the combined capacity of the three Sasol plants was about 6000 X 103 t per year. Based on methane, the Mossgas plant in South Africa and the Shell plant at Bintuli, Malaysia, came on stream in 1992 and 1993, respectively. At that stage, however, the price of crude had dropped to about US$ 15 per barrel and so the timing of these two plants, with capacities of about 1000 X 103 and 500 X 103 t per year, respectively, was not as fortunate as had been the case for the two new Sasol plants. Crude oil prices continued to decline to about US$ 10 per barrel, but in 1999 there was a dramatic increase and at the time of writing the price is again above US$ 30 per barrel. Environmental/political considerations are also factors. Thus, for instance, certain large new oil fields are associated with high levels of natural gas the flaring of which is prohibited. In order to recover the crude oil the construction of FT plants to convert the methane to oil is being considered. The FTS is a catalyzed chemical reaction in which carbon monoxide and hydrogen are converted into liquid hydrocarbons of various forms. Typical catalysts used are based on iron and cobalt. The principal purpose of this process is to produce a synthetic petroleum substitute for use as synthetic lubrication oil or as synthetic fuel. FTS involves three process steps: (1) the generation of the synthesis gas, a mixture of CO and H2; (2) the conversion of the syngas to a mixture of hydrocarbon and oxygenate products; (3) hydroprocessing of synthesis products to produce transportation fuel products, today with emphasis on diesel.
9
Fig 1: Change in Global Crude Oil Price
A number of approaches are available for the conversion of synthesis gas. Two catalysts—iron and cobalt based— have been widely used and these are the only ones receiving serious consideration today. The major difference between these two catalyst types is the formation of the oxygen containing products: with cobalt the oxygen in the syngas is rejected as water and with iron the oxygen is rejected as carbon dioxide. Since cobalt catalysts do not have water-gas-shift (WGS) activity, the feed gas must have a H2/CO ratio of about 2.0–2.3. Iron based catalysts have significant WGS activity so that much lower H2/CO ratio feed gas can be utilized. Thus, the perception has developed that the iron catalyst is preferable for syngas with a low (0.5–1.3) H2/CO ratio, typical of those produced from coal as the carbon source, and that the cobalt based catalyst is preferred for high (2.0+) H2/CO ratios, typical of those produced from natural gas. Since natural gas usage is the primary driving force for FTS today, cobalt has been receiving much attention. However, utilization of coal indicates that iron catalysts should not be automatically eliminated from consideration for syngas derived from natural gas. The Central Institute of Mining & Fuel Research, Dhanbad being the premier Coal Research organization in the country, has been working on FTS for last five decades. Initial research on FTS was aimed for liquid fuel synthesis from coal derived syngas and iron based catalysts was studied for this purpose. However, during last one decade, attention has been focused towards the utilization of vast natural gas reserve of India hence cobalt based catalysts are studied for conversion of natural gas derived syngas for synthesis of liquid hydrocarbon or wax (Fig 2).
10
Fig 2: Pathway of conversion of natural gas to liquid fuels through FTS
Reaction Chemistry of FTS The original Fischer-Tropsch process is described by the following chemical equation:
The initial reactants in the above reaction (i.e., CO and H2) can be produced by other reactions such as the partial combustion of methane (in the case of gas to liquids applications):
OR by the gassification of coal or biomass:
The mixture of carbon monoxide and hydrogen is called synthesis gas or syngas. The resulting hydrocarbon products are refined to produce the desired synthetic fuel. Chain growth pathways and possible secondary reactions of olefins in the FT-Synthesis are shown in Fig 3.
11
Fig 3: Chain growth pathways and possible secondary reactions of olefins in the FT-Synthesis
Fischer Tropsch Kinetics
F-T chemistry is often modelled by a simple power law equation: -rCO = Ae-E/RT pH2mpCOn where: rCO = rate of conversion of CO pH2, pCO = partial pressure of H2 and CO, respectively. In general m is positive (~ 1.0) and n is negative (-0.1 – -0.35). Energy of activation is generally found in the range of 20 –25 kcal/gmol.
International Status Since 1955, F-T synthesis is being exploited commercially by SASOL in South Africa. Sasol is a synfuel technology supplier established to provide petroleum products in coal rich but oil poor South Africa. The firm has built a series of FT coal to oil plants, and is one of the world’s most experienced synthetic fuels organizations and now marketing a Natural Gas To Oil Technology. It has developed the world’s largest synthetic fuels project, the Mossgas complex at Mossel Bay in South Africa that was commissioned in 1993. To increase the proportions of higher molecular weight hydrocarbons, Sasol has modified its Arge reactor to operate at higher pressures. Sasol has commercialized four reactor types with the Slurry Phase Distillate process being the most recent. Its products are more olefinic than those from the fixed bed
12
reactors and are hydrogenated to straight chain paraffins. The slurry phase distillate converts natural gas into liquid fuels, most notably superior quality diesel. Another technology uses the Sasol Advanced Synthol (SAS) reactor to produce mainly light olefins and gasoline fractions. Sasol has developed high performance cobalt based and iron based catalysts for these processes. The company claims a single module or sasol slurry phase distillate plant which converts 100 MM scfd of natural gas into 10,000 barrels/day of liquid transport fuels can be built at a capital cost of US $ 25,000 including utilities, off-site facilities and infrastructure units. Due to steeply growing market demand and for the utilization of large natural gas reserves many new players has came in GTL market which are now licensing their GTL technologies. SASOL, SHELL & SYNTROLEUM are the major licensing companies and other companies like Exxon, Statoil, Rentech are also spreading their business worldwide. Shell has carried out R&D since the late 1940s on the conversion of natural gas, leading to the development of Shell Middle Distillate Synthesis (SMDS) process, a modified FT-process. This process aims on maximizing the yield of middle distillates. First commercial SMDS project started in 1993 at Bintulu-Malayasia. In this process natural gas is first partially oxidized into syngas (CO + H2) in Shell gasifier. This syngas is then converted into heavy paraffins in a fixed bed multi-tubular reactor. This heavy paraffin is then selectively hydrocracked & isomerized for getting the middle distillates. The product obtained by this process is free from contaminants, sulphur, aromatics and heavy metals.Syntroleum Corporation is the developer, user and licensor of the Syntroleum Process, which is a proprietary process for converting natural gas (or synthesis gas from coal) into synthetic liquid hydrocarbons. The Syntroleum Process produces synthetic liquid hydrocarbons, also known as synthetic crude oil, that are virtually free of contaminants that are normally found in products made from conventional crude oil. These synthetic liquid hydrocarbons can be further processed into higher margin products through conventional refining techniques, including Syntroleum's proprietary Synfining Process. These products include: Ultra-clean liquid fuels for use in internal combustion engines and fuel cells; specialty products such as synthetic lubricants, process oils, waxes, etc. Costs to produce these products using GTL technology are increasingly competitive with conventional process technologies. Moreover, the ultra-clean properties of GTL fuels meet or exceed the new and proposed environmental requirements that will soon go into effect for the U.S. and Europe. Key advantages to the Syntroleum Process include its use of air in the conversion process, which is economically competitive and inherently safer than the requirement for pure oxygen in other GTL technologies; and Syntroleum's proprietary catalysts, which enhance conversion efficiency of the catalytic reaction. These features help to reduce capital and operating costs of plants, and permit smaller plant sizes, including mobile plants that can be mounted on barges for offshore service. Research indicates that the
13
Syntroleum Process can be economically applied in plant sizes from less than 20,000 to over 100,000 barrels per day. RENTECH proposed electrical power and FT fuels project in West Australia as announced October 29, 1999. Rentech Inc. is the developer and licensor of a patented and proprietary Fischer-Tropsch, gas-to-liquids process, for conversion of synthesis gas made from natural gas, industrial off-gas, or solid or liquid carbon-bearing materials into high-value fuels, products and chemicals. These products include cleaner burning, sulfur- and aromatic-free diesel fuel, naphtha, waxes and fuel for fuel cells. Preliminary project sizing is for a 200-megawatt electrical power plant and a 6,000-barrel per day Fischer-Tropsch plant. It is expected that Texaco's gasification technology will be used to gasify the lignite to product a syngas consisting of hydrogen and carbon monoxide. The syngas will in turn be used to produce electricity and feed the Rentech-designed Fischer-Tropsch reactor. With its large gas reserves, Norway’s STATOIL has been developing catalysts and process reactors for the FTprocess to produce middle distillate from natural gas. This process employs a three-phase slurry type reactor in which syngas is fed to suspension of catalyst particles in a hydrocarbon slurry, which is a product of the process itself. The process continues to be challenged by catalyst performance and the ability to continuously extract the liquid product. EXXON has developed a commercial FT system from natural gas feed stock. They claim for the slurry reactor and proprietary catalyst systems result in high productivity and selectivity along with significant economy of scale benefits. Exxon employs a three-step process: fluid bed synthesis gas generation by catalytic partial oxidation; slurry phase FT-synthesis; and fixed bed product upgrade by hydro isomersation. The process can be adjusted to produce a range of products. More recently they have developed a new chemical method based FT-process to synthesize diesel from natural gas. Exxon’s better catalysts and improved oxygen–extraction technology have reduced the capital cost of the process and is actively marketing the process internationally. One GTL-Microsystems of UK claims to have developed a process
-GTL Microsystems modular Process
Intensification (PI) reactors for Steam Methane Reforming (SMR) and Fischer-Tropsch process, by using micro reactor technology. This has reduced size of about one tenth of conventional GTL plants, which uses a unique metallic substrate catalyst with high conversion of C5+ hydrocarbons. This process has low capital cost and superior economics, compact in design, high efficiency and no compressed air, oxygen or water is required for its operation. Another company VELOCYS of USA is also claiming for the micro channel reactor technology application to FT synthesis reactions. A group of Japanese companies intent on making gas-to-liquids (GTL) fuels a commercially viable overseas business by 2011 plans to start building a trial GTL plant in Niigata. This facility is targeted to begin operation in 2008 with a daily output capacity of about 500 barrels per day of synthetic fuels and chemicals using natural gas. 14
The partners include Nippon Oil, Nippon Steel, Japan Petroleum Exploration, Inpex, Cosmo Oil, Chiyoda and the government-owned Japan Oil, Gas and Metals National Corp (JOGMEC), along with support from the Ministry for Economy, Trade and Industry (METI). JOGMEC has been researching GTL technology since 1998, with an emphasis on the efficient production of the syngas used in the Fischer-Tropsch (FT) process that actually produces the end products. Working in partnership with Chiyoda (which developed the catalyst), JOGMEC is developing a GTL process that eliminates the need for three units usually found the syngas section of a conventional GTL operation, thereby potentially significantly reducing the cost. There shall be no CO2 removal unit, oxygen plant and syn gas conditioning unit. The JOGMEC process has only been tested in very low production pilots (7 bpd). This trial is designed to help perfect in support of building an overseas plant (likely in Indonesia) with a production capacity of 15,000 barrels per day. The major GTL Technology overviews are shown in Table-1. It is evident that continuous developmental work on F-T synthesis is being carried out since the beginning of this century to develop alternative source of liquid fuels for particular situation and to meet the need of coming decades. It is found that fundamental work F-T process development is being carried out in Universities & Institutes and application part is being carrying out by Multinational & Large Companies.
CIMFR Status The research work on F-T synthesis in India is considered to have begun within two decades of the invention of the process in 1913 in Germany. The first Indian research paper was published in 1935. Since then, investigational studies on F-T synthesis is continuing. Most of these studies were to examine the effectiveness of catalyst and obtain data on adsorption kinetics and mechanism aspects of F-T synthesis. These studies were initially started at Universities of Bengal and followed up by IISC (Bangalore), CIMFR (Dhanbad), IIT (Kharagpur). In the early 1960s, IIT (Kharagpur) put up a pilot plant of considerable size (product capacity: 100 gals per day) and operated the same. This was by far the biggest F-T synthesis unit in the country. However, this unit used fixed bed reactor of Pre-World War II period, was not suitable for engineering developmental studies. CIMFR, which started investigations on F-T synthesis from 1952 put up a fluidized bed reactor of maximum catalyst capacity of 1.2 liter by the end of 1950s and operated the same for about a decade. By the end of 1960s, studies on FT synthesis practically stopped at IIT (Kharagpur) and IISC (Bangalore). Studies at CIMFR, however, continued on low-key basis till the middle of 1970s. In the early 1970s, when the oil crisis occurred, interest on F-T synthesis was revived and IICT (Hyderabad), IIT (Bombay) took up projects on F-T synthesis. In the early 1980s, IIT (Madras) also started working on F-T synthesis. From 1976, CIMFR started working on the design of scaled up F-T synthesis unit with fixed bed reactor system to install the same and study the catalysts developed at CIMFR. 15
Since the above studies were taken up institution wise, the contributions to the science of F-T synthesis were diverse in nature. During the early part of
Table-1: The major GTL Technology overviews. Sr. No.
Company
O2 Plant
Syn Production
gas FT-Synthesis
(Location)
(Catalyst/ Source) (Catalyst source)
1.
2.
GTLMicrosystems (UK) VELOCYS
None
None
(USA)
Steam-Methane- Micro Reforming in PI reactor Reactor
Channel
(Microsystems)
(Microsystems)
SMR
Micro reactor
Channel
(Velocys) (Velocys)
3.
JOGMEC
None
Tubular reformer Slurry (Chiyoda) (Co/NSC)
Reqd.
Auto thermal Slurry reformer (Co/Sasol) (Topsoe)
Reqd.
POX (Shell)
None
Auto thermal Slurry bed reformer (Co/Syntroleum) (Syntroleum)
Reqd.
Auto thermal Slurry bed reformer (Co/ExxonMobil) (ExxonMobil)
Reqd.
CPOX (Conoco)
Reqd.
Compact reformer Slurry (BP) (Co/BP)
(Japan) 4.
SASOL (SA)
5.
SHELL
Fixed (Co/Shell)
(Malayasia) 6.
SYNTROLEUM (USA)
7.
Exxon Mobil (USA)
8.
Conoco (USA)
9.
BP (USA)
Slurry (Co/Conoco)
16
bed
bed
bed
bed
bed
investigation, nickel, cobalt and iron based catalysts were studied. These studies were aimed to examine the yield of liquid hydrocarbons, activity aspect of the catalyst, role of promoters as well as basic studies like adsorption of CO and H2 over these catalysts. After the Second World War, emphasis was put to develop iron catalyst for F-T synthesis as iron had some advantages over cobalt and nickel. From mid-50s, investigations, therefore, concentrated on the development of iron based F-T synthesis catalysts. Considerable amount of studies had been carried out metals of group I, II, III, IV, VI and VII of periodic table as promoters for iron catalysts. At IISc (Bangalore), investigation to understand the mechanism of FTS reactions were carried out using carbon balance method, in addition to chemisorption studies on cobalt and iron catalysts. Such carbon balance method helped to understand the product distribution pattern and reaction mechanism in respect of intermediate complexes had also been studied.
At IIT (Kharagpur), Fe based F-T catalyst was tested in the pilot plant and sufficient product was obtained for characterization. However, only one catalyst was studies and because of using obsolete type of reactor, the plant was not suitable for engineering development studies. Further, bench scale slurry phase reactor system was studied for some time. At IIT (Bombay), some studies on the preparation technique and nature of supported mono and bimetallic catalysts using metals of group VIII had been carried out. It was reported that these catalysts, however, did not produce higher hydrocarbons during synthesis. At IIT (Madras), some studies on the basic aspects of F-T catalysts was done using modern instrumental technique. These studies were related to the role of support and promoter in F-T catalysts and also mode of carbon monoxide activation and adsorption. At IICT (Hyderabad), some studies on F-T catalysts have been carried out recently. Different methods of preparations of supported F-T catalysts and characterization of these catalysts have been carried out and studied. Synthesis over multi-metallic catalyst system has been reported. At CFRI, extensive studies on iron catalysts from basic and applied angles had been carried out. Three different types of F-T synthesis reactors, viz., fixed bed; fluidized bed and slurry phase had been tested. The emphasis was mainly on fluidized and fixed bed type of reactors. Fluidized bed reactor system has been found to be suitable for olefins, gasoline and chemicals like alcohols and ketones production. The fixed bed reactor is suitable for paraffinic and high molecular weight hydrocarbons, like middle distillate and waxy products. The studies in lab scale slurry phase reactor showed production of low molecular weight (C1-C4) gas phase hydrocarbons. In the field of catalysis, catalysts preparation technique, role of promoter and changes in catalyst phases during their use have been studied. Based on such studies, CIMFR has developed catalyst for the production of liquid fuels with
17
high yield of middle distillate from synthesis gas. The results of iron based FT catalysts studied at CIMFR are presented in Table-2.
Table-2: Previous GTL Runs of Iron catalyst in Liquid Fuels Section Catalyst Name
H2/CO
Pressure
Rxn. Temp.
GHSV
Run time (h)
Kg/cm2 o
Yield Aqs / Wax
C
Oil /
FT/CS- 23/151
1.561.70
15-15.5
200-205
450-460
129
108.2/ 148.5 /18.69
FT/CS-14/77/167
1.70
15-16
200-215
450
119
134 /84 /---
USA Catalyst
1.70
15-16
220
450
48
8 / 4/-----
FT/CS/15/79
1.70
15-16
200
300
134
153 /93/----
FT/CS- 23/151
1.561.70
15-16
200-205
450-460
53
137.5 /56.5/ -
FT/CS/23/B3/181
1.55
10-13
205-210
450
76
41 / 51/---
The laboratory scale studies of catalysts were translated into scale-up studies. For this purpose, Fixed Bed Tubular Reactor (Catalyst Capacity 150 ml) and Process Development Unit (Catalyst Capacity 3.20 liter) have been installed at CIMFR. The promising catalysts developed earlier were undergone large scale testing in the Process Development Unit. Most of the FTS studies were performed under the in-house projects of CIMFR but in 1998, a project titled “Fischer-Tropsch (F-T) Synthesis: Development of Catalysts for Wax Production” was sponsored by Dept. of Science and Technology, Govt. of India under the Indo-Russian collaboration scheme (ILTP).This study was done in the temperature range of 180 – 220oC. It iwas observed that although pressure increases the selectivity for higher alkanes, due to limitations by the material construction of the STR, it was kept at 1.2 – 1.25 MPa. The work on the conversion of syngas to wax was initiated with iron catalyst. Three iron-based catalysts – two supported on clay
18
having different proportion of promoter and the third with silica support were prepared by precipitation technique. One Fe-based clay supported catalyst showed appreciable wax yield which was tested for 208 h at GHSV 450h -1 and 1.2 MPa of pressure. At 205oC and H2/CO ratio of 1.7, wax yield was 15% whereas for H2/CO ratio of 1.6 and 210oC, wax yield was 32%. The wax yield for the Fe-catalysts was low so cobalt-based catalysts, which are known for its selectivity towards high molecular weight compounds, were taken up. Two cobalt-based catalysts (CIMFR/C1 and CIMFR/C2) which were prepared differently and showed maximum activity, were compared with one Russian catalyst (cobalt-based). The wax yield for CIMFR/C1 was 18.3 wt% of C5+ fraction at 195oC, H2/CO =1.7 and GHSV~550h-1 and the other catalyst CIMFR/C2 showed maximum wax selectivity 58.3 wt% C5+ fraction at 1.2 MPa, 200oC, GHSV~ 650h-1 and feed H2/CO = 2.0. Analysis of the wax component by simulated distillation showed the maximum carbon number of wax was 62. Very recently extensive investigations have been carried out on alumina supported Cobalt catalyst under the Taskforce Project of CSIR, India, titled, “Development of Gas to Liquid Technology for Fischer-Tropsch &
DME Fuels”. Three different formulations of Cobalt based catalyst (FT-Cat12, FT-Cat12A and FT-Cat14) for middle distillate synthesis have been prepared, characterized and its activity evaluated. All the catalysts have been tested at 500-1000 h-1 GHSV’s, 20 bar reaction pressure, H2/CO ratio 2.0, and at 220-2500C reaction temperature. FT-Cat 12, FT-Cat 12A and FT-Cat 14 have been tested for 125 Hrs, 122 Hrs and 78.5 Hrs of continuous operation respectively. FT-Cat 12, FT-Cat 12A and FT-Cat 14 have yielded 207 ml, 147 ml and 101 ml of C 5+ Hydrocarbon Liquid respectively, which was characterized in GC-FID-SIMDIST. The SIMDIST results show the yield of naphtha (b.p up to 1400C) varies from 4-25 %, middle distillate (b.p 141-3600C) varies from 61-83 % and waxy oil (> 3600C) varies from 8-33 % of total liquid product. The CO conversion varies from 73-95%. The selectivity of C5+ hydrocarbon at 500 h-1 GHSV of FT-Cat 12, FT-Cat 12A and FT-Cat 14 were 77, 72 and 66% respectively. It has been observed that CO conversion increases and C5+ hydrocarbon liquid selectivity decreases with increase in space velocity of feed. The C5+ Hydrocarbon Liquid selectivity is maximum at low space velocities, which may be due to larger residence time. Also, as per SIMDIST result, IBP of liquid hydrocarbons of product varies from 70-75 0C and FBP varies from 400450 0C. As per the yield and selectivity pattern of the developed catalyst system it can be emphasized that these may be successful formulations leading to commercialization for converting the huge Natural Gas Reserves of our country and adding up in the current oil pool. First time in India, 50 liter/day capacity pilot plant has been set up wherein dried Babool wood was gasified in a 2100 kg/hour feed gasifier, giving rise to real gases containing mainly Hydrogen 19.1%, Carbon Monoxide 17.2%, Nitrogen 50.4% along with other gases viz. methane, carbon dioxide and oxygen. The gaseous mixture thus produced is cleaned so as to free from moisture and oxygen. Additional Hydrogen has been added to the product gas to make the ratio of Carbon Dioxide to Hydrogen was 1:2. This synthesis gas was then converted into transportation fuels in Fisher-Tropsch Process using catalyst in a fixed bed reactor at the desired temperature. Typical Products
19
obtained were: Straight run Gasoline 48-50%, Jet Fuel 21-25% and Soft Wax (Pharmaceutical Grade) 1-10%. Reproducible results were obtained. This process was technically established but its economic feasibility, catalyst life etc. are yet to be ascertained. Further, in the SSRC Subcommittee, a project titled, “Development of Coal to Liquid Technology” has been cleared. For the CTL technology it has been targeted to develop suitable catalysts and system/process for converting the coal derived syngas into liquid hydrocarbons specifically the middle distillate which is a mixture of diesel and kerosene.
Conclusion:
In order to avert climate change due to atmospheric green house accumulations, global emissions ought to be significantly reduced. The use of renewable fuels offers the potential to rectify this problem as carbon cycles can be closed and additional carbon dioxide emissions avoided. Currently Biomass is being used to produce power and heat. As transportation accounts for a significant share of CO2 emissions the use of biomass should be extended beyond heat and power production. Through biomass gasification transport fuels such as gasoline and diesel can be produced to replace existing transport fuels. While the latter represents proven technologies, the production of liquid fuels via Fischer Tropsch synthesis has received growing interest.
Different process concepts for the production of Fischer Tropsch transportations fuels from Biomass were designed & evaluated economically & ecologically. The considered concepts were a small scale plant, a stand alone large scale plant & a large scale plant directly connected to petroleum refinery. The essential advantage of the small scale plant was the nearness to the producing area of the biomass which avoided long transport routes. The gasification systems of these plants are presented, as well as their gas cleaning and conditioning concepts Fischer Tropsch Synthesis, reactors, product separation & upgrading & off - gas utilization. The economic analysis showed that under the given condition the transportation fuels produced by this method from the small scale plant were the most expensive while both types of large scale plant could produce transportation fuels in a more economic manner. The ecological analysis demonstrated that the biomass derived fuel could reduce the output of Greenhouse gases drastically. References: 1. Middle Oil distillates from Fischer-Tropsch process; V.A.Krishna Murthy, A.N.Basu, N.G. Basak, A. Lahiri; Journal of Scientific & Industrial research: 21A (7), 338 (1962). 2. Conversion of coal to oil by Fischer –Tropsch process; Samiran Basu, V.A.Krishna Murthy; Seminar proceedings, Need for coal based chemical Industries, The Inst. of Engineers, 5-6 Dec. 1981, p 48-55.
20
3. Semi Pilot Scale investigations of FT synthesis for conversion of syn gas to Middle Distillates and Wax.; Samiran Basu, G.C.Nandi, S. K.Mitra; Chemical Engineering World, 29, (1), 123-125, 1994. 4. Design, synthesis, and use of cobalt-based Fischer-Tropsch synthesis catalysts; Enrique Iglesia; Applied Catalysis A: General 161 (1997) 59-78. 5. The Fischer–Tropsch process: 1950–2000; Mark E. Dry, Catalysis Today 71 (2002) 227–241. 6. GTL Technology-Challenges and Opportunities in Catalysis; P. Samuel; Bulletin of The Catalysis society of India, (2), 2003, 82-99. 7. Fischer-Tropsch synthesis: relationship between iron catalyst composition and process variables Burtron H. Davis; Catalysis Today 84 (2003) 83–98.
21
THE EFFECT OF COAL COMPOSITION ON IGNITION AND FLAME STABILITY IN CO-AXIAL OXY-FUEL TURBULENT DIFFUSION FLAMES
Dadmehr M. Rezaei, Ph.D. Candidate, Department Of Chemical Engineering, University Of Utah Email: [email protected] Yuegui Zhou, Associate Professor, Shanghai Jiao Tong University, Shanghai, China, Email: [email protected] Jingwei Zhang, Ph.D., Department of Chemical Engineering, University of Utah Email: [email protected] Kerry E. Kelly, Research Associate, Department of Chemical Engineering, University Of Utah Email: [email protected] Eric G. Eddings, Professor, Department of Chemical Engineering, University Of Utah Email: [email protected] Jost O.L. Wendt, Presidential Professor, Department of Chemical Engineering, University Of Utah Email: [email protected]
Abstract Past research on flame stability and stand-off distance under oxy-coal combustion conditions has used a 100kW pulverized coal test rig with a co-axial turbulent diffusion burner, and has been described at previous Pittsburgh Coal Conferences. These studies were for one specific coal, namely a Utah Bituminous Coal. The purpose of the research described in this paper is to extend the previous work and to explore how coal composition changes affect the following dependencies that control flame stand-off distance and flame ignition, namely: 1) the effect of partial pressure of oxygen (PO2) in the primary stream with differing preheat temperatures in the secondary stream; and 2) the effect of PO2 in the secondary stream with zero O2 in the primary stream. The results of this new study were designed to extend previously obtained knowledge on effects of secondary preheat temperature, turbulent mixing, PO2 in various streams, from one single coal to other coals of differing compositions. This paper, therefore, explores the effects of coal composition on ignition in oxy-coal, coaxial, turbulent diffusion flames. In this research, the stability and stand-off distance of the flame were studied for the following three types of coal: Utah Skyline Bituminous, Illinois #6 Bituminous, and a Powder River Basin (Black Thunder) coal. To this end we investigated: 1) the effect of PO2 in the primary stream, 2) the effect of PO2 in the secondary stream, and 3) the effect of preheat temperature in the secondary stream, on flame stand-off distance, using the same photo-imaging methodology described elsewhere. The results of the ignition and flame stability analysis for these three coals under oxy-firing conditions are compared, and the effects of coal composition are elucidated.
1. Introduction The increase of green house gases, and particularly carbon dioxide, for energy production has resulted in new technologies with lower emissions of NOx, and SOx. These techniques are capable of complying with carbon dioxide capture and sequestration. Oxy-Fuel Combustion technology has been suggested as the shorter remedy for the conventional power plants.
In Oxy-Coal combustion, coal burns with pure oxygen instead of air, and the combustion gases are diluted using the recycled flue gas (RFG) which essentially contains CO2 and H2O. The gas mixture has higher emissivity and subsequently greater heat transfer, at the same time a lower volume. Previous study at the University of Utah on Utah Skyline Bituminous in 100 kW vertical oxy-fuel combustor explained the effect of PO2 in the primary stream at different preheat temperature in the secondary stream [1]. Also, this work provided information on the results of increasing of PO2 in the secondary stream without O2 in the primary stream, and flame stability. The result of the study was presented in the 2009 Pittsburgh Coal conference. However the influence of coal composition which is one of the key criteria of combustibility on this set of experiments is not well known. This research explores the effect of coal composition on combustion behavior in the similar oxy-fuel operating conditions. The present study was conducted to compare the ignition behavior of three coals representing a spectrum of rank and coal composition in oxy-coal combustion by 100 kW pilot scale combustor with a co-axial turbulent diffusion burner.
2. Experimental 2.1. Coal selection and sample preparation Three types of coal with different rank have been used in this study. The ultimate, approximate, and ash analysis of the prepared coals is provided in Table 1 and Table 2 where coals are listed by increasing rank as determined by carbon, and volatile matter. In order to see the effect of moisture clearly the PRB (Black Thunder) was chosen as one of the samples.
Table 1 LOD
Ash
C
H
N
S
O (by diff.)
Volatile matter
Fixed carbon
HHV BTU/lb
Utah Skyline
3.18
8.83
70.60
5.41
1.42
0.53
13.21
38.60
49.39
12606
Illinois #6
9.65
7.99
64.67
5.59
1.12
3.98
16.65
36.78
45.58
11598
PRB (Black Thunder)
23.69
4.94
53.72
6.22
0.78
0.23
34.11
33.36
38.01
9078
Table 2 Ti
Al
Ca
Fe
Mg
Mn
P
K
Si
Na
S
Element as
Al2O3
Cao
Fe2O3
MgO
MnO
P2O5
K2O
SiO2
Na2O
SO3
TiO2
Utah Skyline Illinois #6
14.52
6.11
5.09
1.39
0.02
0.59
0.57
60.89
1.41
2.33
0.88
17.66
1.87
14.57
0.98
0.02
0.11
2.26
49.28
1.51
2.22
0.85
PRB (Black Thunder)
14.78
22.19
5.20
5.17
0.01
1.07
0.35
30.46
1.94
8.83
1.30
Three samples of coals have been pulverized and prepared for steady feeding to the combustor. Since that size of the coal particle is one of the indices of the coal combustion, we tried to have the same coal particle size. The coal particle size is smaller than 150 µm. The particle size distribution has been calculated and summarized in Fig 1. According to data shown in the Fig.1, the mass average diameter was calculated 62 µm for Utah Skyline coal particle, and 68.5 µm for the Illinois #6 coal particle. As it is obvious, the size and quantity of both Illinois #6 and Utah Skyline coals are similar; therefore, the study is able to focus only on the effect of coal composition.
50
50
45
45
Utah Skyline coal particle size distribution
40
35
30 25 20
30
25
20
15
15
10
10
5
5
0
0
20
40
60
80
100
120
Coal particle size (µm)
140
160
180
Illinois #6 coal particle size distribution
40
Quantity(%)
Quantity (%)
35
200
0
0
20
40
60
80
100
120
140
160
180
200
Coal particle size (µm)
Fig. 1.Particle size distribution of Illinois #6 bituminous, and Utah Skyline bituminous coals
2.2. Combustion furnace The Oxy fuel Combustor (OFC) consists of three sections, namely: 1) Burner zone or the main chamber 2) Radiant zone which is located in the bottom of the main chamber 3) The convective zone benefits from the heat exchangers to resemble the industrial furnace conditions. The dimension of the Burner zone is 0.61m I.D, 0.91m O.D. and, 1.22m as the height of the burner zone. The chamber has been insulated 76 mm thickness of the Fiberboard that is able to tolerate the temperature of 1700°K. The temperature of the furnace can be monitored by three high temperature resistant type K thermocouples that are located along the height of the chamber. Also, to have an optical access to the flame regarding optical diagnostic, four quartz windows have been provided on the quadrants of the cylindrical chamber The chamber of the OFC has been equipped with 24 ceramic electric heaters that have been arranged in three rows. These 840 watt heaters allow us to control the wall temperature of the chamber in an accurate way. In this study the wall temperature was kept at 1850 °F as one of the parameters which has an important role on ignitibility of coal. 2.3 Gas analyzers, and OPTO22 Gas probe is located at the end of convective zone of the OFC. This probe gives us the ability to monitor the exhaust gas composition during the combustion process using oxygen, carbon dioxide, NOx, SOx analyzers. The furnace is controlled and monitored by the OPTO22 control system. All the data from the furnace and analyzers have been plotted, and saved on charts during the experiment. This system helped us to have a more accurate results and control on the OFC. 2.4 Burner and feeder The schematic of the burner drawn by Solidworks is shown in the Fig. 2. The burner consists of 1) Primary stream 2) Secondary stream. The primary line is located in the center of the burner. Carrier gas which is a mixture of O2 and CO2 carry the gas through this line and spray the coal into the chamber. The secondary stream contains mixture O2 and CO2 and flows around the primary jet. Also the secondary stream is equipped with a heat exchanger that gives us the ability to preheat the secondary flow up to 700°K. The flows and temperatures of both streams are accurately automated and under control using the OPTO22 control system. Having a steady coal feeding system is one of the most effective indices on the flame stability data. In this study it was decided to use a twin-screw coal feeder with an eductor. In the eductor system, the primary stream gases get mixed with coal fed by the screw feeder. It lets us have a steady stream of pulverized coal feeding through the burner.
Fig. 2. The schematic of OFC burner
2.5 Methodology of flame stability measurement In this research, it was determined to indicate the flame stability based on the flame stand-off distance concept. Flame stand-off distance is defined as the distance between the tip of the burner and visible part of the flame. In order to measure stand-off distance, a high speed camera was used to capture the flame images. It is obvious that because of turbulent nature of the co-axial flame, fluctuations are present. The camera operating conditions chosen in this set of experiments are 8.3 ms exposure time and 30 fps. For every combustion condition, 6000 images were taken. For this exposure time, it was found out that there are stand-off distance fluctuations that need to be considered in the method. It was suggested to include all the fluctuations in the flame images by the Probability Density Function (PDF). In this paper, Flame stand-off distance plots are shown based on the PDF. 2.6 Combustion Operating Conditions In order to have a reasonable comparison to investigate the effect of coal composition on ignition and flame stability, the experiment jet aerodynamics parameters showed in Table 3 and Table 5 were kept constant for each type of coal. However to keep the total Stoichiometric Ratio (SR), it was necessary to change the coal feeding rate based on the SR. Table 3 Constant key parameters constant Parameters Total S.R.
Quantity 1.15
Pri. stream temp
305 °K
Pri. stream vel.
6.3 (m/s)
Sec. stream vel. & temp. Wall temp.
1283°K a
Overall O2 VF. a
Refer to Table 5
40%
Volume fraction
Table 4 Coal feeding rate based on SR. Coal type
Feeding rate based on SR. (Kg/hr)
Ratio of Coal weigh/ Carrier Gas volume (Kg/m3)
Utah Skyline Bituminous
4.84
1.266
Illinois # Bituminous
5.26
1.376
PRB ( Black Thunder)
8.07
2.111
Table 5 Combustion aerodynamic operating conditions for both Utah Skyline and Illinois #6 coal Experiment
Kg/hr 15.91 16.27 16.60 16.92 17.35
Average Vel. m/s 13.2 13.2 13.2 13.3 13.3
Residence time ms
Pri.a PO2
Pri. Vel.b
Sec.c Vel.
Sec.Td
Pri. O2
Pri. CO2
Sec. O2
Sec. Co2
m/s 6.3 6.4 6.3 6.4 6.3
m/s 14.9 14.9 14.9 14.9 14.9
K 489 489 489 489 489
Kg/hr 0.00 0.27 0.50 0.73 1.04
Kg/hr 6.84 6.48 6.16 5.87 5.40
Kg/hr 11.05 10.80 10.58 10.33 10.01
A A A A A
1 2 3 4 5
M.F.e 0.000 0.055 0.101 0.146 0.210
B B B B B
1 2 3 4 5
0.000 0.055 0.101 0.146 0.210
6.3 6.4 6.3 6.4 6.3
16.6 16.6 16.6 16.6 16.6
544 544 544 544 544
0.00 0.27 0.50 0.73 1.04
6.84 6.48 6.16 5.87 5.40
11.05 10.80 10.58 10.33 10.01
15.91 16.27 16.60 16.92 17.35
14.5 14.6 14.6 14.6 14.6
83 82 82 82 82
C C C C C C C C
1 2 3 4 5 6 7 8
0 0 0 0 0 0 0 0
6.3 6.3 6.3 6.3 6.3 6.3 6.3 6.3
14.9 15.3 15.6 15.9 16.3 17.8 18.6 19.5
489 489 489 489 489 489 489 489
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
6.84 6.84 6.84 6.84 6.84 6.84 6.85 6.85
11.05 11.52 12.02 12.53 13.03 15.34 16.60 17.96
15.98 15.98 15.98 15.98 15.98 15.98 15.97 15.97
13.2 13.5 13.8 14.1 14.4 15.7 16.5 17.3
91 89 87 85 84 76 73 69
91 91 91 91 90
a
Primary stream Velocity c Secondary stream d Temperature e Mole fraction b
In this study there are three sets of experiments. Experiments A, and experiments B essentially are determined to research the effect of oxygen in the primary stream. The second benefit of this experiment is that we are able to investigate the role of turbulent diffusion mixing in the burner jet by increasing the oxygen in the primary stream. The overall oxygen concentration is kept at 40%; however, in each case the amount of oxygen gets deducted from the secondary stream, and added to the primary stream. The flow rates in both experiments A and experiments B are identical. In order to evaluate the effect of secondary stream preheat temperature on the combustibility of the coals, two temperatures were determined. For the experiments A the secondary stream temperature was kept at 489°K, and for the experiments B of the experiment it was at 544°K. The experiments C were considered to see the importance of overall oxygen concentration in the furnace. In this case the primary stream oxygen concentration is zero; however, the amount of oxygen in the secondary stream increases. The overall oxygen increases until the attached flame is obtained. The secondary stream temperature in the experiments C is kept at 489°K.
3. Results 3.1 PRB (Black Thunder) In order to measure the flame stand-off distance, it is necessary to have a visible flame. In PRB case, due to high moisture composition of the coal, visible flame was never observed. Thus, we could not measure stand-off distance of this type of coal. To have combustion, each coal particle first needs to be dried. The second step is the coal pyrolysis that causes the ignition of the coal particle, and then propagation of the flame happens. For the PRB case, the residence time was not sufficiently long enough for the coal particles with 23.69% moisture to be dried. Therefore, moisture has a significant effect on the coal ignitibility.
3.2 Combustion of Utah Skyline and Illinois #6 coals at 489°K preheat secondary stream temperature and Overall 40% Oxygen concentration The results of the experiments A are shown in the PDF plots for Illinois #6 and Utah Skyline coals. It is noticeable that at these conditions, we obtained an attached flame from Illinois #6 coal at PO2=0.144 in the primary stream. However, for Utah skyline coal, the attached flame was obtained at PO2=0.207 in the Primary stream.
0.12
Illinois #6, Preheat T=489 K, Primary PO2=0, Overall PO2=40%
Utah Skyline, Preheat T=489 K, Primary PO2=0, Overall PO2=40%
0.11
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Standoff Distance (cm)
Utah Skyline, Preheat T=489 K, Primary PO2=0.054, Overall PO2=40% 0.14 0.13 0.12
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 10
15
20
25
30
0.07 0.06 0.05 0.04 0.03 0.02 0.01 5
10
Utah Skyline, Preheat T=489 K, Primary PO2=0.099, Overall PO2=40%
Probability Density (1/cm)
Probability Density (1/cm)
0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Standoff Distance (cm)
0.11
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 10
15
20
25
Standoff Distance (cm)
30
35
15
20
25
30
35
Illinois #6, Preheat T=489 K, Primary PO2=0.144, Overall PO2=40%
Utah Skyline, Preheat T=489 K, Primary PO2=0.144, Overall PO2=40%
5
25
Standoff Distance (cm)
0.12
0 0
20
Illinois #6, Preheat T=489 K, Primary PO2=0.099, Overall PO2=40%
0.1 0.09
15
15
Standoff Distance (cm)
0.11
10
35
0.1 0.09 0.08
0 0
35
0.12
5
30
0.11
Standoff Distance (cm)
0 0
25
Illinois #6, Preheat T=489 K, Primary PO2=0.054, Overall PO2=40% 0.15
0.11
5
20
Standoff Distance (cm)
0.12
0 0
15
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
Standoff Distance (cm)
30
35
Illinois #6, Preheat T=489, Primary PO2=0.207, Overall PO2=40%
Utah Skyline, Preheat T=489 K, Primary PO2=0.207, Overall PO2=40% 0.12 0.11
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
35
30
25
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
25
20
15
10
5
Standoff Distance (cm)
30
35
Standoff Distance (cm)
Fig. 3. Comparison of PDF of stand of distance for Utah Skyline (Left column) and Illinois #6 (Right column) coals at 489°K preheat temperature.
3.3 Combustion of Utah Skyline and Illinois #6 coals at 544°K preheat secondary stream temperature and Overall 40% Oxygen concentration This set of experiments was selected to evaluate the effect of secondary stream temperature on the combustibility of the coals. The temperature of the secondary was kept at 544°K during the experiment. The results of this set are provided in Fig. 4. According to the results, for Utah Skyline, a semi-attached flame was obtained at PO2=0.054 and it was fully attached at PO2=0.099 in the primary stream. However, Illinois #6 coal had an attached flame at PO2=0.144. Also, it is noticeable that the results of the Illinois #6 coal at 544°K preheat temperature did not change considerably compared to the results at 489°K of the secondary stream preheat temperature. Illinois #6, Preheat T=544 K, Primary PO2=0, Overall PO2=40%
Utah Skyline, Preheat T=544 K, Primary PO2=0, Overall PO2=40% 0.12 0.11
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Utah Skyline, Preheat T=544 K, Primary PO2=0.054, Overall PO2=40% 0.11
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 5
10
15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Standoff Distance (cm)
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 10
15
20
25
Standoff Distance (cm)
30
35
15
20
25
30
35
Illinois #6, Preheat T=544 K, Primary PO2=0.099, Overall PO2=40%
Utah Skyline, Preheat T=544 K, Primary PO2=0.099, overall PO2=40% 0.12
5
25
Standoff Distance (cm)
0.11
0 0
20
Illinois #6, Preheat T=544 K, Primary PO2=0.054, Overall PO2=40%
0.12
0 0
15
Standoff Distance (cm)
Standoff Distance (cm)
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
Standoff Distance (cm)
30
35
Illinois #6, Preheat T=544 K, Primary PO2=0.144, Overall PO2=40%
Utah Skyline, Preheat T=544 K, Primary PO2=0.144, Overall PO2=40% 0.12 0.11
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
Standoff Distance (cm)
Utah Skyline, Preheat T=544 K, Primary PO2=0.207, Overall PO2=40%
Illinois #6 , Preheat T=544 K, Primary PO2=0.207, Overall PO2=40%
0.12 0.11
Probability Density (1/cm)
Probability Density (1/cm)
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
35
30
25
20
15
10
5
Standoff Distance (cm)
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Standoff Distance (cm)
15
20
25
30
35
Standoff Distance (cm)
Fig. 4. Comparison of PDF of stand of distance for Utah Skyline (Left column) and Illinois #6 (Right column) coals at 544°K preheat temperature
3.4 Combustion of Utah Skyline and Illinois #6 coals at 489°K secondary stream temperature with increasing of overall oxygen concentration in the secondary stream In this set of experiment, the secondary stream preheat temperature was kept at 489°K. However, to see the effect of overall oxygen on the combustion, the amount of oxygen in the secondary stream was increased. It is important to know that in this test, it was tried to keep the primary streams aerodynamics constant, in order to be able to study only the influence of the overall oxygen concentration. Therefore, the amount of oxygen in the primary steam was zero. The results of this test are shown in Fig.5 in PDF form.
Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=40%
0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
Probability Density (1/cm)
Probability Density (1/cm)
Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=40% 0.15 0.14 0.13 0.12 0.11 0.1
5
10
15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Probability Density (1/cm)
Probability Density (1/cm)
0.13 0.12 0.11
5
10
15
20
25
Standoff Distance (cm)
20
25
30
35
Illinois #6, Increasing Secondary O2, Primary PO2=0, Overall PO2=41%
Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=41% 0.15 0.14
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
15
Standoff Distance (cm)
Standoff Distance (cm)
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
Standoff Distance (cm)
30
35
Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=42%
Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=42%
0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
Probability Density (1/cm)
Probability Density (1/cm)
0.15 0.14
5
10
15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Probability Density (1/cm)
Probability Density (1/cm)
0.13 0.12 0.11
5
10
15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Standoff Distance (cm)
10
15
20
25
30
35
Probability Density (1/cm)
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
Standoff Distance (cm)
Probability Density (1/cm) 15
20
25
30
35
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
20
25
30
35
15
20
25
30
35
15
20
25
30
35
Standoff Distance (cm)
Standoff Distance (cm)
Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=50%
Probability Density (1/cm)
Probability Density (1/cm)
0.13 0.12 0.11
10
15
Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=48%
Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=48%
5
35
Standoff Distance (cm)
0.15 0.14
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
30
Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=44%
Probability Density (1/cm) 5
25
Standoff Distance (cm)
Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=44% 0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
20
Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=43%
Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=43% 0.15 0.14
0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
15
Standoff Distance (cm)
Standoff Distance (cm)
0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
Standoff Distance (cm)
30
35
Probability Density (1/cm)
Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=52% 0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0
5
10
15
20
25
30
35
Standoff Distance (cm)
Fig. 5. Comparison of PDF of stand of distance for Utah Skyline (Left column) and Illinois #6 (Right column) coals at 489°K preheat secondary temperature with increasing of overall oxygen concentration
3.5 TGA Analysis In order to have information regarding devolatilization, and structure of Illinois #6 and Utah Skyline coal, we decide to perform TGA analysis. The TGA test carried in both Oxygen and Nitrogen environments. The temperature ramp of the TGA was 20 °K/min. Also the amount of the purging gas was chosen 100 ml/min. In the nitrogen environment, there is ability to see the weight loss of the coal due to drying and devolatilization in an inert gas. However, in Air we have the chance to explorer the influence the oxygen on the pyrolysis as well. The TGA plots are provided in the Fig 6 and Fig 7.
0.4
100 ––––––– Illinois #6 coal ( pure N2 ) – – – – Utah Skyline coal ( pure N2) 447.6°C
90 0.3
Weight (%)
80 0.2 70
Deriv. Weight (%/°C)
464.3°C
0.1
73.8°C
60
50
0
200
400
600
Temperature (°C)
0.0 1000
800
Universal V4.7A TA Instruments
Fig. 6. TGA analysis for Utah Skyline and Illinois #6 coal in Nitrogen environment
0.4
100 – – – – Utah Skyline coal (Air) ––––––– Illinois #6 coal (Air)
448.1°C
458.1°C
80
0.3
60
0.2
Weight (%)
251.3°C 295.7°C 78.3°C
40
0.1
0.0
20
0
Deriv. Weight (%/°C)
364.3°C
0
200
400
600
-0.1 1000
800
Temperature (°C)
Universal V4.7A TA Instruments
Fig. 7. TGA analysis of the Utah Skyline and Illinois #6 coal in Air environment
4. Discussion According to the results in Fig 3, by increasing the oxygen in the primary stream from the secondary stream, the turbulent mixing develops, and the flame stability increases. Better flame stability implies a better ignition. The first criterion of ignition is the amount of volatile matter; however, it is believed that more correct way to determine the combustibility is the index of Fuel Ratio (FR) which is the ratio of (fixed carbon content / volatile matter content) [2,3,4]. General trend is observed that the higher fuel ratio correspond with less carbon burnout. The PDF data agree with the fuel ratio. However, the difference is not that significant. Therefore, in order to justify the results, it was decided to use the TGA tests. The TGA plots and data are presented in Fig 6 and Fig 7. The fuel ratio of Utah Skyline coal and Illinois #6 is provided in the Table.6. Table 6 Fuel ratio of coals Coal type
Fuel ratio
Utah Skyline Bituminous
1.28
Illinois # Bituminous
1.24
TGA plot in Fig. 6 and Fig. 7 show the weight loss of both Utah Skyline and Illinois #6 coals in nitrogen environment. The first peak indicates the moisture lost, and the second peak manifest the devolatilization of both coals. As it is obvious, pyrolysis of the Illinois #6 coal occurs at lower temperature than Utah Skyline. The TGA experiment was carried in air environment as well. Results are exhibited in Fig.7. There are three peaks in the plots. It is believed the first peak is for moisture loss, the second peak corresponds with oxidization of aliphatic structure of the coals. Aliphatic compounds have weaker bonds with smaller activation energy; therefore, they get released at lower temperatures. The pre-devolatilization of the aliphatic compounds of coals boosts heating the particle. The third peak which is largest peak of the plot shows the major devolatilization of the coal. This devolatilization is due to the oxidation of the aromatic compounds of coals. Looking more accurately to large devolatilization peak, we can see a left shoulder peak. This peak is also an indication of existence of more volatile compounds in Illinois #6 coal
than Utah Skyline coal. The result of TGA is a reliable evidence for the two types of coals combustion behaviors in Experiment A and B. Illinois #6 pyrolysis occurs at lower temperature than Utah Skyline; therefore, it is seen that in Experiment A carried out at 489°K secondary stream preheat temperature, flame stability of Illinois #6 develops by increasing the primary stream oxygen concentration. However this effect is not significant for Utah Skyline coal. Experiment B was performed at 544°K secondary stream preheat temperature. It is notable that only 55°K increasing temperature of the secondary stream stimulates the devolatilization of Utah Skyline. Comparing Fig.3 and Fig.4, It is shown Utah coal flame stability increases tremendously. However, Illinois #6 is already able to be pyrolyzed at lower temperatures, and increasing the temperature will not have any more considerable influence on the ignitibility of the Illinois #6 coal. As it is shown in Fig.3 and Fig.4, PDF’s of the experiments B-4 and A-4 look the same. In experiment C, the amount of overall oxygen was increased to able to obtain an attached flame. The amount of oxygen in the secondary stream was increased in each case of experiment C; however, the concentration of oxygen in the primary stream was kept at zero. Fig.5 shows the influence of increasing of overall oxygen is more important for Utah Skyline than Illinois #6. The Utah coal flame has some indication of an attached flame at 42% overall O 2 concentration, and it is fully attached at 44% overall concentration of O2. However; there was not any sign of an attached flame for the Illinois #6 coal until the concentration of O2 was increased to 48% and even at 52% of overall O2, a fully attached flame was not witnessed. According to the approximate analysis of both types of coal, the moisture content of Utah Skyline it 3.18% and for Illinois coal is 9.65% which is three times larger. It is important to note that when the overall oxygen concentration increases, the velocity of the burner jet in the secondary stream rises as well. Therefore, the residence time decreases. The values of both velocity and residence for each case of experiment C has been calculated and tabulated in Table5. The lack of sufficient residence time for the high moisture content coal retards the rate of both heat transfer and mass transfer for drying and devolatilization of the coal particles. So the flame stability of Illinois #6 coal lowers compared to Utah skyline coal, even though O2 concentration was increased.
5. Conclusion Studies carried out in the pilot scale Oxy fuel combustion and TGA on three types of coal with different rank reveals the important role of moisture content and volatile matter on the ignitibility of the coal particles. The existence of the moisture in the coal particle retards the heat transfer and mass transfer for the devolatilization; therefore, the ignitibility of the particles decreases. Higher residence time for high moisture content coals requires to be considered in the burner design. Fuel ratio is a good index to estimate the combustibility of the coal; However, It is not an accurate way to anticipate the combustibility of the coals. It is important to note that coal structure can have a significant effect on the ignition of the coal particle in the oxy-coal combustion. The details of the chemical structure of three coals from the Argonne Premium Coal Sample Bank (Wyodak, Blind Canyon, and Illinois #6) which are quite similar to the three coals used in this study (Black Thunder, Utah Skyline, and Illinois #6) suggest that the differences in the chemical structures of these coals could be the source of the variability in the pyrolysis/combustion properties of these coals [5]. TGA analysis is a suitable method to investigate the devolatilization of the coal for the prediction of the combustion behavior [6]. Secondary stream preheat temperature has an important influence on the flame stability of the coals. However, it is valuable to know that increasing the temperature higher than required temperature for the devolatilization of the coal particle will not have any considerable development on the ignitability of the coal.
6. Acknowledgment This material is based upon work supported by the Department of Energy under Award Number DE-NT0005015. Also; hereby, I would like to express my appreciation for the advices from Prof. Terry Ring, and Prof. Ronald Pugmire. Help in running the experiment, and preparing the data were provided by Ryan Okerlund, Colby Ashcraft, and Charles German.
References: [1] Jingwei Zhang, Kerry E Kelly, Eric G Eddings, Jost O.L Wendt. Ignition in 40kW co-axial turbulent diffusion oxy-coal jet flame, Proc. Combust. Inst. (2010), doi: 10.1016/j.proci.2010.06.106 ( In press) [2] N.Oka. T. Murayama, H. Matsouka, S. yamada, S. Shinozaki, M. Shibaoka, C.G. Thomas. The influence of Rank and maceral composition on the ignition and char burnout of pulverized coal. Fuel Processing technolog, 15 (1987) 213-224 [3] S.Su, J.H. Pohl, D. Holcombe, J.A. Hart. A Proposed maceral index to predict combustion behavior of coal. Fuel, 80 (2001) 699-706 [4] S. Pregermain. Rank and maceral effect on coal combustion characteristic. Fuel Processing technology, 20 (1988) 297-306 [5] M. S. Solum, R. J. Pugmire, D. M. Grant. 13C Solid-State NMR of Argonne Premium Coals, Energy & Fuel, 3 (1989), 187-193. [6] S.Su, J.H. Pohl, D. Holcombe, J.A. Hart. Techniques to determine ignition, flame stability and burnout of blended coals in p.f. power station boilers. Progress in Energy and Combustion Science, 27 (2001) 75-98
Study on the In-furnace Desulfurization in Oxy-fuel Combustion Using Drop Tube Furnace with Limestone Hyung-Keun Lee, Wook Choi, Hang-Dae Jo, Won-Kil Choi, Korea Institute of Energy Research Greenhouse Gas Research Center, 71-2, Jang-dong, Yuseong-gu, Daejeon 305-343, Korea Phone: +82-42-860-3647, Fax: +82-42-860-3134, e-mail: [email protected] Sang-In Keel Korea Institute of Machinery & Materials 171, Jang-dong, Yuseong-gu, Daejeon 305-343, Korea
ABSTRACT Oxy-fuel combustion uses high-purity oxygen as combustion oxidant instead of air used in conventional air combustion to produce pure CO2 stream as combustion products for easy separation and storage of CO2. Oxy-fuel combustion with many advantages like high combustion efficiency, low flue gas flow rate and low NOx emission has emerged as a promising CCS technology for coal combustion facilities. In this study, the effects of limestone types and characteristics, reaction temperature, Ca/S molar ratio, the concentrations of CO2, O2, SO2 on SO2 removal efficiency and decomposition of CaSO4 were investigated in a drop tube furnace under typical oxy-fuel combustion conditions represented by high concentrations of CO2 and SO2 formed by gas recirculation to control furnace combustion temperature. SO2 removal efficiency increased with reaction temperature, but over around 1250 ℃ decreased with reaction temperature due to promoted decomposition of CaSO4 formed by sulfation reaction. And SO2 removal efficiency increased with SO2 concentrations, because the increased SO2 concentrations suppressed the decomposition of CaSO4. The increased SO2 removal efficiency by increased CO2 and O2 concentrations showed that SO2 removal by limestone is mainly done by the direct sulfation reaction under oxy-fuel combustion conditions. Also, it was proved experimentally that the increased concentrations of CO2 and O2 have inhibited the decomposition of CaSO4.
1. INTRODUCTION Oxy-fuel combustion uses high-purity oxygen as combustion oxidant instead of air used in conventional air combustion to produce pure CO2 stream as combustion products for easy separation and storage of CO2. Oxy-fuel combustion with many advantages like high combustion efficiency, low flue gas flow rate and low NOx emission has emerged as a promising CCS technology for coal combustion facilities.(1,2)
Most coal has sulfur compounds that exist in various forms, so the combustion process for coal inevitably produces and emits sulfur oxides (SOx). Moreover, in oxy-fuel combustion, SO2 concentration increases about three times as high as that of conventional air combustion system owing to the flue gas recirculation for controlling the combustion temperature. Such a high concentration of SO2 in oxy-fuel combustion system is very harmful for the combustion system itself, CO2 transportation system and storage facilities due to corrosive potential, therefore it should be treated.(3,4) In-furnace desulfurization can be applied to high temperature region like upper part of oxy-combustion furnace as dry desulfurization process to reduce SOx emission using solid particle alkali reagent like limestone. In general, it is known that the desulfurization efficiencies of in-furnace desulfurization for conventional air combustion system are lowered by the decomposition of CaSO4 which is product of sulfation reaction between alkali reagent and SOx in combustion flue gases. On the other hand, much higher concentrations of CO2, SO2 in oxy-fuel combustion conditions would affect the decomposition behaviors of CaSO4, so higher SOx removal efficiency can be expected. Hao Liu et al. explained that enriched SO2 inside the furnace owing to flue gas recirculation under oxy-fuel combustion conditions suppressed the decomposition of CaSO4. And also CO2-rich atmosphere accelerates CaSO4 decomposition in the absence of oxygen.(5,6) Higher CO2 concentration formed in oxy-fuel combustion conditions enhance the sulfation reaction of SO2 with less calcination, so-called direct sulfation. (7) In this study, in-furnace desulfurization reactions for oxy-fuel combustion were simulated using DTF (Drop Tube Furnace). Using DTF system, the experiments were performed to investigate the effects of limestone type and characteristics, the experimental conditions including reaction temperature, the concentrations of CO2, O2, SO2 on the decomposition of CaSO4 and SO2 removal efficiencies. 2. EXPERIMENTAL 2.1 Drop tube furnace system and experimental conditions The experiments were performed using DTF system as shown in Fig.1 which consists of several parts: a reactor, a limestone feeder, gas supplying system, a reaction products collecting system and gas analyzing system. The reactor was drop tube furnace type and had physical dimensions of 500 mm length, 500 mm width, and 1200 mm height. In the center of DTF, a ceramic tube was located, which was heated electrically and had physical dimensions of 1200 mm height, 50 mm diameter and 5mm thickness. The temperatures of the reactor were controlled by three divided sections in the range of 600 - 1400 ℃ and were measured by thermocouples. Reaction gases used in these experiments were O2, CO2, SO2 to simulate the typical flue gas compositions of oxy-fuel combustion. The exact compositions of reaction gases were controlled by MFCs (Mass Flow Controller) for each gas. The simulated reaction gases flowed into the reactor at the top of the ceramic tube after passing the gas preheater to reduce the temperature difference between gases and the reactor. Reacted gases pass through the sintered metal filter equipped in the collecting system and flow to the gas analyzing system consisting of gas sample conditioner and infrared type gas analyzer. Gas concentrations of reacted gases were measured using gas analyzer (Siemens, Model: Ultramat 23) after removing the moisture and the dust using gas sample conditioner (Model: WE-GSC4P). Solid particles were captured by collecting system at the
exit of the reactor. Table.1 shows the experimental conditions for in-furnace desulfurization tests by limestone and the decomposition characteristics of CaSO4.
Fig.1 Schematic diagram of drop tube furnace type experimental system.
Several type of limestone were collected from various regional mines and characterized. We selected four of them depending on the grade and characteristics for these experiments. The selected two types of limestones were labeled KS and UR1. KS was group label for 2 limestones labeled KS1and KS2 sampled at different times from same mine. Table.2 shows the chemical composition and the average particle size data of the limestone samples used in these experiments. The reagent grade anhydrous CaSO4 with 10.92 ㎛ of the average particle size and 99% of purity manufactured by Alfa Aesar company was used for these experiments.
Table.1 Experimental conditions for oxy-fuel combustion. Experimental Condition
Desulfurization
CaSO 4 decomposition
Total gas flow rate (L/min)
5 ~ 20
8.0
Reaction temperature (℃) Ca/S molar ratio
1000 ~ 1350 1~4
1000 ~ 1350
20 ~ 80 0 ~ 20 1800 ~ 3600 0~40
20 ~ 80 0.5 ~ 20 0 ~ 3000
Gas Concentration CO2 (%) O2 (%) SO2 (ppm) H2O (%)
Table.2 The chemical composition and the average particle size data of the limestones Compositions Limestones
KS
Al2O3 Fe2O3 (%) (%)
CaO (%)
MgO Ig-loss Purity Crystal size (%) (%) (%) (㎛)
Mean diameter (㎛)
KS1
0.09
0.12
0.04
55.70
0.26
42.50
99.5
50~120
19.69
KS3
0.97
0.14
0.23
54.90
0.37
42.60
98.0
50~120
23.17
0.59
0.30
0.076 54.82
0.83
43.35
97.9
20~50
9.86
UR1 *
SiO2 (%)
Purity data were calculated CaCO3 contents based on CaO contents of limestones.
3. RESULTS and DISCUSSION 3.1 Desulfurization SO2 is absorbed through sulfation reaction by the calcined limestone in the in-furnace desulfurization. Since the calcination of limestone is an endothermic reaction and is enhanced at high temperatures, the reaction temperature is a very important factor in the desulfurization reaction. It is important to identify the type of limestone favorable desulfurization reaction and favorable conditions to improve the performance and the utilization of limestone experimentally. Fig.2 shows the effects of the reaction temperature and the type of limestones on SO2 removal efficiency. The experiments were carried out at the conditions of the reaction temperature of 1000-1350 ℃, the inlet gas concentrations of CO2, O2 and SO2 were 80%, 20%, 2400ppm, respectively. Ca/S ratio was fixed at 2.0. SO2 removal efficiencies increase with the reaction temperature for most of the limestone samples until the reaction temperature approach to 1200 ℃. Over 1200 ℃, SO2 removal efficiency is found to be reduced, because of the sintering and the clogging of the active surfaces and pores of the absorbents and the decomposition of CaSO4 due to temperature rise. Fig. 3 shows the effects of SO2 concentration and Ca/S ratio on SO2 removal efficiency for KS3 limestone. The experiments were carried out at the conditions of the reaction temperature of 1200 ℃. The inlet SO2 concentrations were varied in the range of 2400~3600ppm, and Ca/S ratio were varied in the range of 1~4. The results showed that SO2 removal efficiencies increased with SO2 concentration and Ca/S ratio. Fig. 4 shows the effects of CO2 concentration changes on SO2 removal efficiency using KS3 limestone. The experiments were carried out at the conditions of the reaction temperature of 1200 ℃, 2400ppm of the inlet SO2 concentrations and 20% of O2. CO2 concentrations were varied in the range of 40~80%, and Ca/S ratio were varied in the range of 1~4. The results showed that SO2 removal efficiency increased with the increase of CO2 concentrations from 40% to 80% due to the enhancement of direct sulfation over the reduction of active surface area of the calcined limestone by increased CO2 concentration. Fig. 5 shows the effects of O2 concentrations on SO2 removal efficiency for KS1 limestone. The experimental conditions were 1200 ℃ of the reaction temperature,
2400ppm of the inlet SO2 concentrations and 80% of CO2 concentration. O2 concentrations were varied in the range of 0.2~20%. In the case of very low O2 concentrations, SO2 removal efficiency showed very low value. SO2 removal efficiency increased rapidly with O2 concentrations over this point, but the increase rate of SO2 removal efficiency decreased over 5 % of O2 concentration. From these results, it was found that oxygen was needed for the direct reaction of SO2 with CaCO3 forming CaSO4, so-called direct sulfation, as the following equation. Therefore the higher concentration of oxygen entering the reactor promote the formation of CaSO4 and finally increase SO2 removal efficiency. CaCO3 + SO2 +1/2 O2 → CaSO4 + CO2
70
35
SO2 removal efficiency (%)
30
SO2 removal efficiency (%)
KS1 KS3 UR1
25
20
15
10
65
SO2=2400ppm
60
SO2=3000ppm
55
SO2=3600ppm
50 45 40 35 30 25 20 15 10
5
5
0 950
0
1000
1050
1100
1150
1200
1250
1300
1350
0
1400
1
2
Temperature℃
Fig. 2 Effect of the reaction temperature and limestone type on SO2 removal efficiency.
(Qg=8
L/min,
CO2=80%,
O2=20%,
3
4
5
Ca/S
SO2=2400ppm,
Fig. 3 Effect of Ca/S ratio and SO2 concentration on SO2 removal efficiency. (Qg=8 L/min, CO2=80%, O2=20%, 1200℃, KS3)
Ca/S=2)
20
55
CO2=80%
50
CO2=60%
18
CO2=40%
SO2 removal efficiency (%)
SO2 removal efficiency (%)
60
45 40 35 30 25 20 15
16 14 12 10 8 6 4
10 2
5
0
0 0
1
2
3
4
5
Fig. 4 Effect of Ca/S ratio and CO2 concentration on SO2 removal efficiency. 1200℃,
O2=20%,
5
10
15
20
25
O 2 concentration (%)
Ca/S
(Qg=8L/min, KS3).
0
SO2=2400ppm,
Fig. 5 Effect of O2 concentration on SO2 removal efficiency. (Qg=8L/min, 1200℃, Ca/S=2, KS1)
CO2=80%,
SO2=2400ppm,
3.2 CaSO4 decomposition In in-furnace desulfurization reaction for the air combustion, the desulfurization efficiency was lowered by the decomposition of CaSO4, the product of the sulfation reaction, in the high temperature region. On the other hand, the higher CO2, SO2 concentrations kept in oxy-fuel combustion suppress the decomposition of CaSO4. Fig. 6 shows the decomposition rate of CaSO4 according to the reaction temperature. The experiments were carried out at the conditions of the reaction temperature of 10001350 ℃. The inlet gas concentrations of CO2, O2 were 80%, 20%, respectively. The feed rate of CaSO4 was 1.2 g/min. The results show that the decomposition rate of CaSO4 increase with the reaction temperature and the decomposition reaction is activated rapidly above 1250 ℃. Fig. 7 shows the decomposition rate of CaSO4 according to SO2 concentration and the reaction temperature. In this figure, the decomposition rate of CaSO4 decreased with SO2 concentration and increased with the reaction temperature. For the high temperature conditions, the decomposition rates of CaSO4 were significantly reduced with increasing SO2 concentrations. In addition, it was found that there were no remarkable differences of the decomposition rate of CaSO4 with the reaction temperature in high SO2 concentration. [6] From these results, it was found that the high SO2 concentration conditions offered by the flue gas recirculation in oxy-fuel combustion inhibited the decomposition of CaSO4, therefore, higher SO2 removal efficiency could be maintained. 0.06
0.14
Decomposotion rate of CaSO4 (1/s)
Decomposotion rate of CaSO4 (1/s)
0.16
0.12 0.10 0.08 0.06 0.04 0.02
1200℃ 1250℃
0.05
0.04
0.03
0.02
0.01
0.00
0.00 -0.01
950
1000
1050
1100
1150
1200
1250
1300
1350
0
1400
Temperature(℃)
500
1000
1500
2000
2500
3000
3500
SO2 concentration (ppm)
Fig. 6 Effect of temperature and on the decomposition rate of CaSO4. (8L/min, CO2=80%, O2=20%, CaSO4=1.2g/min)
Fig. 7 Effect of reaction temperature and SO2 concentration on the decomposition rate of CaSO4. (8L/min, CO2=80%, O2=20% CaSO4=1.2g/min)
4. Conclusions In this study, the effects of limestone types and characteristics, reaction temperature, Ca/S molar ratio, the concentrations of CO2, O2, SO2 on SO2 removal efficiency and decomposition of CaSO4 were investigated in a drop tube furnace under typical oxy-fuel combustion conditions represented by high concentrations of CO2 and SO2 formed by gas recirculation to control furnace combustion temperature. 1) SO2 removal efficiencies increase with the reaction temperature until the reaction temperature approach to 1200 ℃. Above 1250 ℃, SO2 removal efficiency is found to be
reduced, because of the sintering and the clogging of the active surfaces and pores of the absorbents and the decomposition of CaSO4 due to temperature rise. 2) SO2 removal efficiencies increased with Ca/S ratio and the concentration of SO2. It has been proven experimentally that the increased concentration of SO2 contributed the inhibition of the decomposition of CaSO4. 3) The increase of SO2 removal efficiencies with CO2 and O2 concentrations demonstrated that the direct sulfation reaction was primary reaction for SO2 absorption under oxy-fuel combustion conditions. In addition, it has been proven experimentally that the increased concentrations of CO2 and O2 inhibited the decomposition of CaSO4. 4) The SO2 concentration has larger impact on the decomposition rate of CaSO4, especially in high temperature conditions. 5) From these results, it was found that the decomposition of CaSO4 had larger impact on the desulfurization efficiency in applying the in-furnace desulfurization for the oxy-fuel combustion and the major factors affecting the decomposition of CaSO4 were SO2 concentration and the reaction temperature rather than CO2 concentration and O2 concentration. References
1. IPCC Fourth Assessment Report(AR4), (2007). 2. B.J.P Buhre, L.K. Elliott, “Oxy-fuel combustion technology for coal-fired power generation", Progress in Energy and Combustion Science, 31, 283-307(2005). 3. Yewen Tan, Eric Croiset, “Combustion characteristics of coal in a mixture of oxygen and recycled flue gas", Fuel. 85, 507-512(2006). 4. Molburg JC, Doctor RD, “CO2 capture from PC boilers with O2-firing"(2001). 5. Hao Liu, Ken Okazaki, “Drastic SOx Removal and Influences of Various Factors in O2/CO2 Pulverized Coal Combustion System", Energy & Fuel, 15, 403-412(2001). 6. Hao Liu, et al. "Decomposition behavior and mechanism of calcium sulfate under the condition of O2/CO2 pulverized coal combusion" Chem, Eng, Comm, 199-214(2001). 7. Jun Cheng, Junhu Zhou, “Sulfur removal at high temperature during coal combustion in furnaces: a review", Progress in Energy and Combustion Science, 29, 381-405(2003) .
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: 1. COMBUSTION: NOVEL TECHNOLOGIES (OXYFUEL, CHEMICAL LOOPING, ETC) EXACT TITLE OF PAPER: EFFECTS OF COAL COMPOSITION ON IGNITION LOSS, SOOT, AND ULTRAFINE PARTICLE DISTRIBUTION IN AIR AND OXY-FIRED COAL FLAMES William J. Morris, Ph.D Graduate Student, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected] Dunxi Yu, Visiting Research Associate, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected] Jost O.L. Wendt, Presidential Professor, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected]
Abstract: A 100kW maximum design down-fired laboratory combustor was used to determine effects of switching from air to oxy firing on soot, unburned carbon and ultrafine particle emissions from practical pulverized coal flames. Of interest here were potential practical effects of substitution of the N2 in air by CO2 in practical pulverized coal flames. Therefore, the focus is on effects of using oncethrough CO2, simulating cleaned flue gas recycle with all contaminants removed. Three coals, a western bituminous, PRB sub bituminous, and a high sulfur eastern bituminous, were fired at 36.6kW in a) air, b) 27% O2/ 73%CO2, c) 32% O2/68%CO2, respectively. Tests were conducted at (nominally) 3%, 2%, 1% and 0% O2 in the exhaust. For each condition, particulate samples were iso-kinetically withdrawn far from the radiant zone, and analyzed using a scanning mobility particle sizer (SMPS) for ultra-fine particles, a photo-acoustic analyzer (PA) for “black carbon”, and a total sample loss on ignition (LOI) method for unburned carbon in ash. Data suggest that at low stoichiometric ratios ultra-fine particles consist primarily of black carbon, which is produced in lesser amounts under oxy-fired conditions than under air-fired, even when adiabatic flame temperatures are matched. For the three coals, significant differences in the ultrafine particle distributions were noted indicating that particles formed in this region are affected by coal rank, moisture content, and sulfur content. However, significant changes in mineral matter vaporization were not observed unless the flames were hotter. These and other results are interpreted in the light of available mechanisms.
High Speed Video Analysis Of Oxycoal Combustion In 40kw Coaxial Turbulent Diffusion Flames Terry A. Ring, Professor, University of Utah Institute for Clean and Secure Energy Chemical Engineering Department 50 S. Central Campus Drive, MEB 3290 University of Utah Salt Lake City, UT 84112 Phone: 801-581-6915, Fax: 801-585-9291 [email protected]
Jingwei Zhang, Husam el Gendy, Jost O.L. Wendt, Kerry Kelly and Eric G. Eddings Institute for Clean and Secure Energy Chemical Engineering Department University of Utah Salt Lake City, UT 84112 Phone:801-581-6915 , Fax: 801-585-5705 [email protected], [email protected],, [email protected], [email protected], [email protected]
Abstract: A 61 cm diameter, down-flow, axial-flame combustion system utilizing various ratios of oxygen to CO2 for combustion of various types of coal has been studied. In companion work flame stability and the standoff distance between the burner and the point of flame ignition has been determined using low-speed video analysis of the flame. A large number of images were analyzed and probability density functions (PDFs) for flame detachment has been determined. In addition, high speed video analysis of the flame has also been performed. These images, performed at 3,000 f/s and with a shutter speed of 1/500,000 s, show hot coal particles less than 100 microns in diameter and flamelets of hot soot generated by eddies of volatiles reacting with oxygen that range in size from several hundred microns in size to centimeters in size and have temperatures that range from 1600 K to 2300 K. The size and shape of the flamelets are analyzed giving fractal shapes for the larger structures. Finally, frequency analysis of the video images were performed giving a Fourier transform power spectra with resonance characteristics associated with the frequency at which coal particles and flamelets pass by, and power spectrum decay that is characteristic of isotropic turbulence. Introduction: The objective of this experimental program is to gather data that will allow comparison to high-fidelity coal combustion simulations. These computational models provide high temporal and spatial information for temperature, pressure and gas and coal and soot particle velocities, as well as chemical species of various types, in a highly turbulent environment. These models are expensive to run taking multiple days to weeks on 1,000 cpu computer clusters. The challenge to experimentalists is to provide experimental data of a similar highfidelity nature for comparison with these simulations. In this work, we provide a first view of some highfidelity experimental data – high temporal and spatial resolution soot temperature data. A challenge to both experimentalists and simulators is how to condense the massive amount of high-fidelity data into a form that can be used for direct comparisons between experiment and theory. Here we suggest several methods including: population distribution functions (PDF) and Fourier transform power spectra of temporal data for direct comparison. Experimental Setup A 61 cm inside diameter, 1.219 m tall down-flow, axial-flame oxy coal combustion system utilizing various ratios of oxygen to CO2 for combustion has been used, see Figure 1. The experimental apparatus is
1
described in detail in a companion paper in this conference1. The walls in the main burner zone, 76 mm thick, are insulated with Insboard 2600 ceramic board material, and the downstream sections are insulated with a castable refractory material. In the main burner zone, twenty four 840 W electrical heaters are placed within the Insboard material to maintain the wall temperature at 1301±10 K using K-type themocouples (Omega) and individual controllers. The heaters enable the 100 kW system to approximate the near-burner conditions of larger boilers in the field. Illinois bituminous coal with the analysis shown in Table 1 has been used for combustion after it was pulverized and classified to 80% minus 200 mesh (74 μm). The coal was burned at a nominal flow rate of 5.26 kg/hr with a nominal 15% excess oxygen, and the incoming oxidant had an oxygen mole fraction of 40%, the balance being CO2 delivered from a liquid storage tank. The burner consisted of coannular flow of coal and transport gas in the primary stream through a 15.8 mm ID tube and oxidizer gas only in the secondary flow through the annular region between the OD (21.34 mm) of the primary and the ID (35.05 mm) of the secondary tube. The primary flow temperature was 305±2 K and that of the secondary was 544±2 K. For these experiments, the mole fraction of oxygen in the primary flow was varied from 0 to 0.209 keeping the total oxygen mole fraction for both the primary and secondary the same at 0.40. The coal was metered with a screw feeder into the primary gas stream. To keep the primary flow at a constant so that the feed rate of the coal could be kept constant, additional CO2 flow was added to maintain the velocity of the primary constant at 6.338 m/s. The secondary flow compensates for variations in the O2 content of the primary, but since the flow of the secondary is much larger (typically 3:1), it only varies by a small amount keeping the ratio of gas velocity between the secondary and the primary gas flows at a constant of 2.4. The specific experimental conditions used in this work are given in Table 2. These experiments are a subset of those performed in the companion paper1. The results of the companion paper suggest that the conditions under study here show detached (A3), partially attached (A4) and fully attached (A6) flame ignition behavior.
C o a l
Primary
Secondary
Figure 1 Schematic of Down-fired Coal Combustion System with view of flame in top-most windows indicated by vertical blue rectangles.
1
Rezaei, D.M., Eddings, E.G., Kelly, K.E., Zhang, J-W, Shou, Y-G and Wendt, J.O.L., “The Effect of Coal Composition on Ignition and Flame Stability in Co-axial Oxy-fuel Turbulent Diffusion Flames,” Proceedings Pitts burgh Coal Conference, Istanbul, Turkey, October 11-14, 2010.
2
Table 1 Analysis of Illinois Bituminous Coal LOD@ Ash@ C H N S O Volatile Fixed HHV 105ºC 750 ºC Matter Carbon (BTU/lb) 9.65% 7.99% 64.67% 5.59% 1.12% 3.98% 16.65% 36.78% 45.58 11598 Ash Composition (wt %) Al2O3 CaO Fe2O3 MgO MnO P2O5 K2O SiO2 Na2O SO3 TiO2 17.66% 1.87% 14.57% 0.98% 0.02% 0.11% 2.26% 49.28% 1.51% 2.22% 0.85% Table 2 Experimental Conditions for Oxy Combustion Experiments. The primary flow temperature was 305±2 K, secondary temperature is 544±5 K and wall temperature is 1301±10 K and the coal (80% -200 mesh Illinois Bituminous) had a flow rate of 5.26±0.3 kg/h and the total oxygen mole fraction was always 0.401 corresponding to a stoichiometric ratio of 1.15. Primary PO2 Primary Secondary Primary O2 Primary Secondary O2 Secondary Velocity Velocity CO2 CO2 m/s kg/s kg/s kg/s kg/s A Mole fraction. m/s 3
0.100
6.338
14.908
1.39E-04
1.71E-03
2.94E-03
4.61E-03
4
0.146
6.338
14.909
2.02E+04
1.63E-03
2.87E-03
4.70E-03
6
0.209
6.338
14.908
2.90E-04
1.50E-03
2.78E-03
4.82E-03
High Speed Video Analysis Video analysis was performed with a Photron high-speed, 1024x1024 pixels, black and white camera [Itronx Imaging Technology, Westlake Village, CA] at 3,000 f/s and a shutter speed of 1/500,000 of a second. This very fast shutter speed essentially stops the turbulent action of the flame and prevents saturation of any of the pixels in the field of view allowing full resolution of the image. A 70 - 300mm macro zoom lens [Tamron USA, Inc., 10 Austin Blvd., Commack, NY 11725] set at 300 mm (macro), an f-stop of 4 and a focal length of 1.3 m was used. The depth of field for these settings was ~3 cm. At these settings a 5 cm x 5cm field of view at the center line of the combustion chamber was in focus. At these camera settings each pixel observes the light from a 50 μm x 50 μm section of the camera’s view. The camera was positioned to observe at the center line of the burner two different axial locations below the tip of the burner – one 7.6 to 12.6 cm and the other 30.6 to 35.6 cm from the burner tip. The intensity of light captured is converted to an electrical current by the amplifier attached to the photo detector. Each pixel in this camera can detect a range in the number of electrons generated by the photo detector. These electrons pass through an on-pixel amplifier and into the pixel’s resistor producing a voltage that is measured –the pixel intensity scales linearly with the measured voltage and the number of electrons generated in the photo detector. The spectral resolution for the camera is between 400 nm and 1,000 nm with various quantum efficiencies at each wavelength. Using the spectral resolution obtained from the camera manufacturer under confidentiality agreement, the pixel intensity can be scaled albeit nonlinearly with temperature using Planck’s black body radiation equation assuming an emissivity of 1.0. Work done by Lou, et. al.2 using two-color pyrometry of similar flames has shown that the emissivity is typically 0.99 to 1.0. Spectroscopic measurements for visible wavelengths fit to Planck’s grey body radiation model gave a flame emissivity of 0.989. As a result, the assumption of black body radiation (emissivity of 1.0) is not a bad assumption. The resulting pixel calibration curve is given in Figure 2. In this figure the equilibrium flame temperature [2468 K] for Illinois coal in this atmosphere was also plotted. Typically 512 frames of video are taken covering a time frame of 0.17 s and creating 0.5 Gb of data for each experimental run.
2
Lou, C. Zhou, H-C, Yu, P-F, Jiang, Z-W., “Measurements of the flame emissivity and radiative properties of particulate medium in pulverized coal fired boiler furnaces by image processing of visible radiation,” Proceedings of the Combustion Institute, 31(2007), 2771-78.
3
300
Pixel Intensity
240 180 120 60 0 500
3
1×10
3
1.5×10
3
2×10
3
2.5×10
Temperature (K) Figure 2 Temperature Calibration based upon Pixel intensity for 1/500,000 s shutter speed and specific camera setup. Green half circle is the equilibrium flame temperature (2468 K) for Illinois coal at stoichiometric conditions. Measurements were made at an axial distance from the burner of 7.6 cm and 30.1 cm corresponding to the top and bottom of the first window indicated by the yellow circles in Figure 1. The windows were cleaned with compressed air just prior to taking high-speed video images, and the images were recorded after the disturbance due to cleaning cleared the combustion system.
a b c Figure 3 Three Frames of 3,000 f/s High-Speed Video for a 5 cm x 5 cm view at 30.1 to 35.1 cm down stream from the burner tip for experimental conditions A6. Results Viewing with a traditional video camera at 30 f/s, we find that the image of the flame, while not static, is rather constant in its color and intensity, see single color frame in Figure 1. An example of three frames from the 3,000 f/s high speed video for a 5 cm x 5 cm viewing area at 30.1 to 35.1 cm downstream from the burner tip, where the flame is well established for experimental conditions A6 is shown in Figure 3. The high speed image is far from constant in intensity but consists of discrete wispy bright areas of various shapes and roughly circular spots of light. The circular spots of light are from less than 100 μm to 300 μm in size, with the smaller ones emitting less light and are likely to be coal particles or clusters of coal particles (larger ones) heated to radiating conditions. The whispy bright areas cover a broad range of sizes from 100’s of microns to centimeters in length. Given the bright edges of these curved shapes, we can surmise that reaction proceeds at the interface between a turbulent eddy containing volatiles generated from the heated coal, with another turbulent eddy rich in oxygen. The reaction between coal volatiles and oxygen forms soot which is most likely responsible for the light emissions observed. Depending upon the local mixing conditions, the fuel to oxygen ratio will vary
4
significantly in these flamelets giving rise to various flamelet temperatures. Observing the individual images in a sequence, the bright zones tend to get hot over a period of 2 to 10 images or about 2 ms and cool over a period of 15 to 25 images or about 7 ms. Flame velocity can also be assessed using several individual frames of this high speed video. Image Analysis Analyzing all 512 of the images taken at one time using the temperature calibration in Figure 2, converting pixel intensity to temperature, we can determine the temperature distribution for any image. An example of a temperature distribution plot is shown in Figure 4 where the vertical axis is temperature. As we can see the temperature profile ranges from 1900 K to 2200 K which is lower than the equilibrium flame temperature (2468 K) for the coal at stoichiometric conditions. During heat up the coal undergoes drying then devolatilization. The mass released is initially water, followed by coal volatiles, which consist of varying sizes of hydrocarbon molecules that are released as the temperature of the coal particle increases. When a turbulent eddy containing coal volatiles collides with eddies having a range of oxygen concentrations, then the adiabatic flamelet temperature for those various collisions will vary and the amount of soot produced will vary. If the local mole fraction of oxygen due to eddy mixing is 0.2 and not that of the overall mole oxygen fraction of 0.4 then the equilibrium flame temperature decreases to ~1900 K. In addition, after the flamelet has reacted the soot produced will cool back down to the reaction-increased gas temperature as time proceeds. It is for this reason that we have such a broad distribution of flamelet temperatures in Figure 4.
T(K) 2246 2134 2049 1912 0
Figure 4 Temperature distributions for the image shown in Figure 3a. Taking the temperatures as in Figure 4 for all locations in all 512 frames of video taken, a temperature probability distribution function (PDF) can be developed for each experimental condition. Temperature PDFs will allow direct comparison between high-fidelity experiments and high-fidelity simulations. The temperature PDFs for all three experimental conditions for 7.6 to 12.6 cm (Top) and 30.1 to 35.1 cm (Bottom) from the burner tip are shown in Figure 5. These results show that with a detached flame, experimental conditions A3, there are slightly higher temperatures near the top of the flame. With this detached flame there is significant pre-mixing of coal with oxidant that, once ignited, burns very hot with significant heat release relative to the downstream region due to the combustion of coal volatiles. There is also a very high-temperature tail with the bottom PDF, but these higher temperatures are much more dispersed as evidenced by the lower probability magnitude. This high temperature tail is likely a result of the onset of char oxidation, with the dispersed higher temperatures associated with glowing char particles. For the partially attached flames (A4) shown in Figure 5B, the PDFs show essentially the same temperature distribution for the top and bottom with only the magnitude of the PDFs being different in these two locations; e.g., more of the frames are filled with bright areas further from the burner tip. For the partially attached flame there is less premixing before ignition, which slows down the
5
initial heat release and makes it more uniform. For the fully attached flame (A6) shown in Figure 5C, the PDFs show a slight decrease in temperature as it moves away from the burner tip from top to bottom. In the fully attached flames, stable ignition occurs near the burner tip which provides significant heat release in the early portion of the flame as a portion of the fuel is consumed with the oxidant available at that location. With continued entrainment of oxidant along the length of the flame, additional fuel is consumed to provide continual heat release. Some heat loss to the furnace walls and observation window provides for limited cooling of the combustion gases along the length of the flame. The cooling in Figure 5C is from 1840 K to 1815 K over the 23.5 cm axial distance between the top location and the bottom location. Another means to correlate a large amount of either experimental or simulation data is to perform frequency analysis using Fourier transform power spectra of temperature data from individual pixel locations. The 512 frames from the high speed camera operating at 3,000 f/s correspond to information over a 0.17 sec time interval analyzed. The resulting power spectrum probes frequencies from 12 Hz to 3 kHz. These singlepixel Fourier transform power spectrums are rather noisy so we have chosen to average the power spectrums from 128 individual pixels spread horizontally across the center of the image. These averaged Fourier transform power spectra are shown in Figure 6. In Figure 6A we see the results from the top view 7.6-12.6 cm from the burner tip and the bottom view 30.1-35.1 cm from the burner tip for experimental conditions A3. These results show that the power is low for the top view and significantly higher for the bottom which is consistent with the increase temperature away from the burner when the flame is detached. The bottom view shows a power spectrum that has a -5/3 power law slope (dashed black line) which is consistent with isotropic turbulence. Using the primary and secondary gas velocity we find that objects will pass through the view of the camera at 125 Hz and 293 Hz based upon the primary or secondary gas velocity, respectively. In the red (top) line in Figure 6A we see a peak in the power spectrum at 120 Hz. When the gases are heated to 1800 K (see Figure 5A) from 544 K the velocities will increase by a factor of 3.3 increasing the frequency with which object will pass through the view of the camera. In the blue (bottom view) curve in Figure 6A we see peaks in the power spectrum at both 130 Hz and at 410 Hz. In Figure 6B we see the results from the top view 7.6-12.6 cm from the burner tip and the bottom view 30.1-35.1 cm from the burner tip for experimental conditions A4. With this partially attached flame, the power spectrum for the top view decays according to the -5/3 power law above 200 Hz, while that for the bottom view does so only for a limited range of frequency, 200 Hz to 800 Hz. The high frequency (> 1 kHz) results show essentially a constant power for the bottom view which suggests that the newly reacting material with its higher temperatures is responsible for these high frequency results. In Figure 6C we see the results from the top view 7.6-12.6 cm from the burner tip and the bottom view 30.1-35.1 cm from the burner tip for experimental conditions A6. The top view shows higher power than the bottom view for these experimental conditions. Both views show high frequency power law slopes of -5/3 indicating isotropic turbulence. Based upon these very different Fourier transform power spectra, we can clearly see the differences between detached, partially attached and fully attached turbulent flames.
6
A3-PDF Data for all Images 0.0001 0.00009 0.00008 0.00007
PD F
0.00006 0.00005 0.00004 0.00003 0.00002 0.00001 0 1000
1200
1400
1600
1800
2000
2200
2400
2000
2200
2400
2000
2200
2400
Temperature (K) Bottom -2
Top -6
5A) Data from Experimental Conditions A3 – Detached Flame A4 - PDF Data all Images 0.00003
0.000025
P D F
0.00002
0.000015
0.00001
0.000005
0 1000
1200
1400
1600
1800
Temperature (K) Bottome -13
Top -11
5B) Data from Experimental Conditions A4 – Partially-Attached Flame. A6 PDF Data all Images 0.00005 0.000045 0.00004 0.000035
PD F
0.00003 0.000025 0.00002 0.000015 0.00001 0.000005 0 1000
1200
1400
1600
1800
Temperature (K) Bottom -14
Top -18
5C) Data from Experimental Conditions A6 – Fully-Attached Flame. Figure 5 Temperature PDFs for all experimental conditions given in Table 2. Figure 5A corresponds to experimental conditions A3, Figure 5B corresponds to experimental conditions A4 and Figure 5C corresponds to experimental conditions A6. Top corresponds to a view from 7.6 to 12.6 cm from burner tip and Bottom corresponds to a view from 30.1 to 35.1 cm from burner tip.
7
Power
1 ×10
7
1 ×10
6
1 ×10
5
1 ×10
4
1 ×10
3
Fourier Tran s form Power Spectrum
100 10 10
100
1 ×10
3
1 ×10
4
Fre quency (Hz )
Power
6A) Data from Experimental Conditions A3, Red line – Top, Blue line – Bottom 1×10
7
1×10
6
1×10
5
1×10
4
1×10
3
100 10
Fourier Tran sform Power Spectrum
100
1×10
3
1×10
4
Frequency (Hz)
Power
6B) Data from Experimental Conditions A4, Red line – Top, Blue line – Bottom 1×10
7
1×10
6
1×10
5
1×10
4
1×10
3
Fourier Tran sform Power Spectrum
100 10 10
100
1×10
3
1×10
4
Frequency (Hz)
6C) Data from Experimental Conditions A6, Red line – Top, Blue line – Bottom Figure 6 Fourier Transform Power Spectra for Experimental Conditions A3, A4 and A6. Red line – Top View 7.6 to 12.6 cm from burner tip, Blue line – Bottom View 30.1 to 35.1 cm from burner tip. Dashed black lines correspond to a power law slope of -5/3 corresponding to isotropic turbulence.
8
Conclusions High speed B&W video has been used to measure the soot temperature in an oxy-coal combustion system operating at 40 mole % oxygen. The video images have been converted to temperature using Plank’s black body radiation equation assuming an emissivity of 1.0. The temperature profiles show that individual coal particles and clusters of coal particles are present as well as clouds of soot produced by the reaction of oxygen with volatiles generated from the heating of coal. Experimental conditions consist of differing amounts of oxygen in the primary burner gas flow containing the coal. The range of experimental conditions provides for detached, partially attached and fully attached flames. From the high-speed video data temperature PDFs and Fourier transform power spectra were determined. With detached flames the temperature PDF is initially of low, near constant temperature and increases in breadth to higher temperatures moving away from the burner and the power spectrum increased from low power that is essentially constant with frequency to one that follows the -5/3 power law decay law corresponding to isotropic turbulence further away from the burner tip. With the attached flame the temperature PDF is high in temperature, relative to the detached flame, near the burner tip and decreases in temperature and narrows in breadth moving away from the burner tip. The power spectrum for an attached flame only shows agreement with the -5/3 power law decay law corresponding to isotropic turbulence at higher frequencies. Partially attached flames show no decrease in temperature moving away from the burner tip and isotropic turbulence near the burner tip decaying away to an essentially constant power spectrum away from the burner. Thus the various flame characteristics - detached, partially attached and fully attached – are easily identified with a combination of temperature PDFs and Fourier transform power spectra measurements. Acknowledgement This material is based upon work supported by the Department of Energy under Award Number DENT0005015. Help in running the experiments was provided by Dadmehr Rezaei and Dr. Yuegui Zhou.
9
Comparison of the mathematical model of pulverized coal burnout with results gained from experimental tests on drop tube Ing. Radim PALUSKA, listopadu 15/2172, 708 [email protected]
33
Energy Research Center of VSB-Technical University of Ostrava, 17. Ostrava-Poruba, Czech Republic. tel. (+420) 597323845, E-mail:
Ing. Marian BOJKO, Ph.D, Department of Hydromechanics and Hydraulic Equipment, Faculty of mechanical Engineering, VŠB-Technical University of Ostrava, tř. 17. listopadu 15, 708 33 Ostrava-Poruba, tel. (+420) 597 324 385, e-mail [email protected]
1. ABSTRACT
In connection with construction of new supercritical power plants which burn pulverized lignite coal it was started the research of kinetic parameters of coal reserves in Czech Republic. Experimental facility and methodology of pulverized coal termokinetic properties determination with use of mathematical modelling is described in the paper. Thermokinetic properties as a mean for better understanding the nature of combustion process can be determined by experiment using the Drop Tube Test Facility (DTTF) described further in the text. DTTF provides conditions occurring in pulverized coal fired boiler by emulated oxygen concentration, temperature and velocity of reaction gas. The DTTF presented in the paper was build recently at the Energy Research Center. Experimental data acquired from DTTF are planned to be used in mathematical modelling using the code Fluent. The paper describes in detail main differences between used Fluent models of particles combustible fraction and reaction rate. Program Fluent can define different mathematical models of volatile evolution (devolatilization model) and char combustion (surface combustion model) to simulate coal combustion. The single kinetic rate devolatilization model assumes that the rate of devolatilization is first-order and the kinetic/diffusion-limited rate model assumes that the surface reaction rate is determined either by kinetics or by a diffusion rate. Temperature field and distribution of species mass fraction is evaluated for comparing with experimental tests. Results from adjusted mathematical model should provide closer information about combustion process in real operation. 2. INTRODUCTION
Combustion of pulverized coal in large-scale boilers is a complicated process, which is bound with a lot of partial reactions depending on many factors. Main influences on combustion progress are temperature of reaction gas, concentration of oxygen in environment of a particle and then surface area and composition of burned coal. Burning proceeds in several stages. In the first stage, a particle warm-up occurs, then, volatile fraction evolution and combustion proceeds and char (fixed carbon) included in a particle burns as a final stage. Study of the last stage of particle combustion is the main theme of this contribution. For this purpose, a unique experimental stand DTTF has been built (see Fig. 1) which provides relevant data for a mathematical model. 3. EXPERIMENTAL FACILITIES AND INSTRUMENTATION
In order to achieve the most precise setting of boundary conditions of the whole fuel burn-out process, a complete system has been built. This system consists of reaction gas preparation, heating and keeping the temperature on the required level within the reaction chamber. At the end of sample trajectory, an instant cooling and sampling section is situated (see Fig. 1). The system is divided into 5 sections. Major component parts are automatically controlled by PLC unit via visualization on PC station. Parameters of the DTTF are given in Table 1.
Table 1: Parameters of ERC Drop tube test facility Temperature of reaction gas 600 - 1200 °C O2 concentration in reaction gas 0 – 21 %vol. Reaction gas velocity 1 - 4 m.s-1 Vertically hinged, electrically heated 4,800 mm long metal reaction chamber with inner diameter of 66 mm is made of Kanthal APM (Fe75Cr20Al5). This material is able to resist high temperatures in an oxidation environment for a long time. Eight batching holes (with a distance 500 mm one from another) are situated on sides of the drop tube in which a water cooled batching unit charges a fuel sample. The distance between these batching holes and the sampling point combined with velocity of reaction gas gives residence time of a pulverized coal particle in the reaction chamber. For more information see [2] and [8].
Fig. 1. ERC drop tube and schematic view of whole DTTF gas system (I. reaction gas preparation to desired oxygen concentration and flow rate II. Heating of reaction gas mixture to desired temperature III. Reaction chamber IV. Fuel sample batching into the reaction chamber V. Sampling with simultaneous cooling of a sample to temperature below 50°C)
The experimental tests were performed with pulverized black coal with main properties shown in Table 2. Table 2: Proximate analysis of the coal (as fired) Proximate analysis (by weight) Moisture, H0 0.5 % 0 Ash, A 26.6 % 0 Volatiles, V 28.5 % 0,fix Fixed carbon, C 44.4 % This coal was sieved on ASTM sieves with the aim to obtain 80 - 90μm granulometry.
Particle-size distributions were measured using a laser diffractometer Malvern Mastersizer as shown in Fig. 2.
Fig. 2. Size distribution of the coal sample used in the tests
4. TESTS
Approx. 5 g of a sample is continuously batched into the drop tube for the period of 15 minutes long single test. A small sample quantity and extended batching duration are used in order to reduce substantially the effect of boundary conditions of reaction environment by contributing intrinsic heat power of the reaction fuel. After passing through the reaction zone, the sample is quenched using liquid nitrogen to temperature below 50°C and sampled on an ashless filter. Then, the sample is analyzed and the unburnt U fraction (quantity of unburnt fixed carbon) is determined according to equation (1) in dependence of residence time in the reaction chamber. m3 m1 m4 (1) U 100 [%] m4 m2 100 m4 m2 A
These values serve to draw the burn-out curve that characterizes fuel with given granulometry (particle dimensions) for the set conditions, ergo temperature and oxygen concentration in the reaction chamber. These results can be also used to set (specify) a mathematical model of particle burnout, as described in the following chapter. 5. MATHEMATICAL MODEL OF PULVERISED COAL COMBUSTION IN THE DROP TUBE
Problems of flow in the drop tube can be characterized as multidimensional compressible flow of gaseous species (oxygen, carbon dioxide, nitrogen, water vapor and volatile) and flow of solid particles (pulverized coal) in the drop tube including heat transfer between gaseous phase and discretion phase (solid particles), devolatilization of volatile and combustion of pulverized coal char. Move of solid particles (the discrete phase of pulverized coal) is modeled as multiphase flow in a mathematical model based on the Lagrangian discrete phase. In the Lagrangian model, the gaseous phase is considered a continuum (continuous phase) and solid coal particles as a dispersed phase. The resulting defined mathematical model presents a system of partial differential equations that were solved by the ANSYS Fluent 12.0 program which is based on the finite volume method where the computational area (in this case the drop tube) is filled with elements of finite volumes (hexahedral, tetrahedral) in which numerical simulation of the mathematical model defined and described above is performed.
6. COMPUTATIONAL GRID OF THE DROP TUBE, BOUNDARY CONDITIONS, PHYSICAL PROPERTIES
The computational grid of the real drop tube has been constructed by the GAMBIT 2.4.26 code according to dimensions of the experimental equipment. A scheme of geometry is shown in Fig. 3. including the specification of types of the boundary conditions defined on the individual boundaries of the model. T=927°C (temperature of wall) stěny) Outlet from drop tube
Inlet of pulverised coal (diameter 5mm)
Inlet to drop tube O2+CO2+N2 Qm=0.004kg/s T=927°C
66
600 4900
Fig. 3. Scheme of geometry of the drop tube for construction of the computational grid including the type of boundary conditions
On the base of the scheme, the gaseous mixture conditions at the inlet into the drop tube (composition, temperature, mass flow rate) are evident. The final model of the drop tube is a 3D model of dimensions shown in Fig. 3.Then, inlet of pulverized coal particles through the probe of ds=5mm outlet diameter can be seen. Pulverized coal mass flow rate (Qm=5.5555e-06 kg.s-1) into the reaction area of the drop tube is measured continuously. Composition of pulverized coal is shown in Tab. 1. In Fig.5, the actual computational grid in the longitudinal section view along the axis of the drop tube in a detail of pulverized coal inlet is shown. Compaction of the computational grid in the area of coal particle inlet and in the predicted area of pulverized coal solid particle flow in the longitudinal section view can be seen in Fig. 4.
Fig. 4 Computational grid in the area of pulverized coal inlet into the reaction area
Physical properties of gaseous mixture and pulverized coal are defined in the next step. Computation of density is defined by ideal gas law (2) due to consideration of compressible flow of gaseous mixture (O2, CO2, N2, H2O).
pabs
Yi (2) i M i The mixture’s specific heat capacity is defined as mass fraction average of the pure species heat capacities (3). c p Yi c p ,i i (3) RT
Viscosity and thermal conductivity of mixture are defined on the basis of kinetic theory for ideal gas law. Physical properties of pulverized coal are defined in Table 3. Table 3: Physical properties of pulverized coal 1400 kg.m-3 cp 1680 J.kg−1.K−1 λ 0.0454 W.m−1.K−1 7. MODEL OF CHAR BURNOUT (KINETIC/DIFFUSION SURFACE REACTION RATE MODEL)
The kinetic/diffusion-limited rate model (The Kinetic/Diffusion Surface Reaction Rate Model) was defined for numerical simulation of burnout of pulverized coal char where the reaction rate of char burnout is defined by kinetics or by diffusion rate (4). dm p dt
Ap
RTYox Do r M ox
(4)
Do r
where Ap=π*d2 is the surface area of a spherical particle. Kinetic rate r is defined by Arrhenius expression E / RTp (5) r Ce Constants C (pre-exponential factor) and E (activation energy for the reaction) are determining parameters of char burnout process according to the above defined model. Activation energy defines minimum of energy needed to start the char burnout process and the pre-exponential factor influences the burnout process velocity. The particle heat balance during surface reaction is defined by the following equation mpc p
dTp
hAp T Tp f h
dm p
H reac (6) dt dt For coal combustion, if the char burnout product is CO2, then ƒh=0.3. The final 3D mathematical model of gaseous species flow with pulverized coal flow, except of a char burnout model, contains a model of volatile devolatilization and a water vaporization model. Detailed characterization and definition of the devolatilization model is introduced in [7]. 8. „INTRINSIC“ MODEL
„kinetic/diffusion-limited rate model“ is in detail described in0. This model assumes, that rate of char consumption (combustion) is defined by kinetics or diffusivity.
D0 C1 where C1 5 10 constant r.
T
T / 2
0.75
p
(7)
dp
12
je diffusion constant. Consumption of char by kinetics is defined by r C2 e
E1 / RT p
(9)
where C2 0.00035 is pre-exponential factor and E1 7.4 10 is Activation energy. 7
Rate of fixed carbon consumption (and thus of particle mass) is defined by combination of kinetics and diffusion on basics of next equation:
dm p dt
Ap pOX
D0 R D0 R
(10)
„Intrinsic“ model involves influence of reaction rate kinetics and effect of volume diffusion in regards to equation (4). Diffusion constant is defined by equation (2), but reaction rate constant r is defined with equation (5): d R p p Ag ki (11) 6 where factor considers effect of diffusion on particle pores and thus porosity, ki is intrinsic reaction rate defined by Arrhenius: E / RT ki Ai e i p (12) Input pre-defined parameters of intrinsic model in ANSYS Fluent 12.1 code are in table 2 with adapted porosity value. Table 2: Input pre-defined parameters of „Intrinsic“ model Diffusivity constant C1
5*10-12
Pre-exponential factor A1
0.030198
Activation energy E1
1.794*108
Porosity
0.2
Mean diameter of pore rp
6*10-8
surface area of a particle Ap
300000
Pores crimp
1.414214
Mathematical modeling goes from set parameters of reaction gas in reaction chamber of DTTF. Next chapters describe 2 main tested conditions. During the evaluation of numeric simulation of combustion gave discrepancy in defined volatiles content (28,5%) with respect to measurement results (see fig. 4). Accordingly, it was changed volatiles content to 10% and reached conformity with experimental results (fig. 5).
Fig. 5 – Comparison of particle burnout for 2 different volatiles content
The same results give sample in case of grain size d=55m (fig. 6).
Fig. 6 – Comparison of particle burnout for 2 different volatiles content with granulometry of d=55m
Modeling of particle (d=55m) burnout with reaction gas concentration of oxygen 2%vol. and temperature 1000°C. For numeric simulation was used „Intrinsic“ model with modification of pore r 3 109 m mean diameter p . Volatiles content is 10% and porosity 0.2 . Particle burnout of in comparison to experimental measurement is presented in figure 7.
Fig. 7 – Comparison of particle burnout for adapted „Intrinsic“ model
9. CONCLUSION
This paper describes new experimental facility which serves for combustion tests of pulverized coal. Methodology and experiments in conditions (temperature and oxygen concentration of reaction gas) similar to combustion process give relevant data for mathematical model of pulverized coal burnout in drop tube. Mathematical model differentiates devolatilization process from burnout of fixed carbon apparently on course of curves, which can be considerably changed by modification of kinetic constants as preexponential factor and activation energy. This contribution gives first results gained from mathematical modeling of combustion of pulverized black coal. There are described two models of fixed carbon burnout and comparison of model with real experiments. Results are different with volatile fraction content which lead to modification of mathematical model definition.
[1] Carpenter A.M., Skorupska N.M.: Coal Combustion - Analysis and Testing, IEACR/64, IEA Coal Research, London, 1993. [2] Žídek M., Horák J., Paluska R.: Experiences with building of drop tube. Int. Conf. „Power engineering and environment 2008“, Ostrava, 11th to 12th September 2008, ISBN 978-80-2481832-0. [3] FLUENT: Fluent 12.0 User’s Guide, Fluent Inc. 2007. [4] Baum M.M., Street P.J.: Predicting the Combustion Behavior of Coal Particles. Combust. Sci. Tech., 3(5): p. 231-243, 1971. [5] Badzioch S., Hawksley P.G.W.: Kinetics of Thermal Decomposition of Pulverized Coal Particles. Ind. Eng. Chem. Process Design and Development, p. 521-530, 1970. [6] Field M. A.: Rate of Combustion of Size-Graded Fractions of Char from a Low Rank Coal between 1200 K and 2000 K. Combustion and Flame, p.237-252, 1969. [7] Sahajwall V., Eghlimi A., Farell K.: Numerical Simulation of Pulverized Coal Combustion. International Conference on CFD in Mineral Metal Processing and Power Generation CSIRO. 1997, p. 197-204. [8] Paluska R., Bojko M., Horák J.: Determination of pulverized coal thermokinetic properties with use of mathematical modeling Transactions of the VŠB - Technical University of Ostrava, Mechanical Series, issue 2009 ISSN: 1210-0471, p. 157-166. [9] PALUSKA R., BOJKO M., HORÁK J.: Evaluation of thermokinetic parameters influence on pulverized coal burnout in drop tube. Rynek Energii nr. 2, 2010. ISSN: 1425-5960, p. 113-117
OPTIMIZATION OF CANADIAN PETROLEUM COKE, COAL AND FLUXING AGENT BLENDS VIA SLAG VISCOSITY MEASUREMENTS AND MODELS Marc A. Duchesne, PhD candidate, University of Ottawa 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 9470287 Alexander Y. Ilyushechkin, Research Scientist, CSIRO Energy Technology, Technology Court, Pullenvale, QLD, 4069 AUSTRALIA [email protected] (617) 33274187 Arturo Macchi, Associate Professor, University of Ottawa 161 Louis Pasteur St., Room A409, Ottawa, Ontario, CANADA, [email protected], (613) 562-5800 ext. 6939 E.J. Anthony, FBC and Gasification Group Leader, CanmetENERGY 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 996-2868
Abstract The slagging behavior of petroleum coke must be known to determine suitable feedstock blends for entrained-flow slagging gasification. To increase the amount of slag formed and maintain a low viscosity, petroleum coke may be blended with coal and/or a fluxing agent such as limestone or dolomite. Viscosity measurements were performed for various blends of artificial Genesee coal ash, Suncor petroleum coke ash, limestone and dolomite in a neutral gas atmosphere. Adding petcoke to the coal provided a moderate reduction in viscosity, while limestone and dolomite additions were very effective for viscosity reduction. FactSage phase equilibrium predictions and quenched sample analysis via SEM and EPMA were used to link solids formation to changes in the viscosity-temperature relation. Slag blends without limestone or dolomite showed glassytype behaviour, while those with limestone or dolomite showed crystalline-type behaviour. Predictions from several slag viscosity models were compared to measured values. The viscosity model which provided the most accurate predictions was utilized for optimization of fluxing agent addition to various petcoke and coal blends. Keywords: Slag, Viscosity, Coal, Petcoke
1 Introduction Production of petroleum coke (petcoke), a by-product of the oil refining industry, has been and is expected to continue increasing [1]. Major factors driving this increase include the rising demand for transport fuels, the use of heavier crude oils and new environmental regulations pushing for reduced waste and highly refined fuels. Due to petcoke’s high heating value and low cost, there is much interest in its use as a primary fuel or in a coal-petcoke blend. Although various technologies can utilize petcoke as a fuel for power generation, gasification may provide lower gaseous emissions and less solid waste [2]. Another advantage of gasification for petcoke conversion is the trapping of hazardous metals such as vanadium and nickel in non-leachable slag. Furthermore, gasification may be used to provide steam and hydrogen in addition to power generation. This may be a huge benefit, particularly if the gasifier is located near the petcokeproducing oil refinery which requires steam, hydrogen and power [3]. The majority of gasifiers in operation today are of the entrained-flow type. In this type of gasifier, most of the inorganic component of the fuel (ash) is partially or fully melted, sticks to the reactor wall and flows to the bottom as slag. To determine the suitability of a fuel and appropriate operating conditions for gasification, it is important to know its slagging properties. A slag which is too viscous may accumulate on the reactor wall till the reactor is plugged, ceasing operation. As a rule of thumb, slag viscosity should not exceed 25 Pa⋅s at the slag tapping temperature [4,5]. If a fuel’s slag is too viscous, a fluxing agent such as limestone or dolomite may be added, and/or the fuel may be blended with another fuel. Blending fuels may also be necessary if there are issues with fuel availability, flame stability or ash content. In the present study, the viscosities of artificial Genesee coal ash, reduced Suncor petcoke ash, limestone and dolomite blends were measured. Observed and predicted phases in the blends were investigated to determine a correlation, if any, between solids precipitation and changes in viscosity. Finally, a carefully selected slag viscosity model was applied to different scenarios involving blends of the studied samples.
2 Experimental 2.1 Sample preparation Artificial Genesee coal ash, reduced Suncor petcoke ash, limestone and dolomite compositions are shown in Table 1. These compositions exclude components which represent less than 1 wt% of the ash. Artificial Genesee coal ash was prepared by mixing laboratory or analytical grade Al2O3, CaO, Fe2O3, K2CO3, MgO, Na2CO3, S, SiO2 and TiO2 powders. Composition was based on an ash analysis of real Genesee coal. Petcoke
ash was prepared by repeatedly heating Suncor petcoke up to 800°C in air. The ash mass was recorded after each heating cycle. This process was continued until the mass change between cycles was less than 2 wt%. The petcoke ash was then reduced by heating to 1000ºC in the presence of syngas (10% CO2, 25% H2, 50% CO and balance N2). The purpose of reduction prior to slag viscosity measurements is to avoid unwanted reactions involving vanadium in the melted slag. Limestone and dolomite were used as provided. Table 1. Ash compositions (in wt%, excluding components representing <1 wt%)
SiO2 Al2O3 Fe2O3 TiO2 CaO MgO SO3 Na2O V2O5
Coal ash
Petcoke ash
Limestone ash
Dolomite ash
58.7 23.7 4.6 0.0 5.7 1.6 3.4 2.4 0.0
37.9 16.8 7.1 1.3 17.9 4.3 8.3 1.5 5.0
2.0 0.0 0.0 0.0 96.8 1.2 0.0 0.0 0.0
2.9 0.0 1.2 0.0 56.9 39.0 0.0 0.0 0.0
2.2 Rheology Slag viscosity measurements were made using a Haake high temperature viscometer using a rotational bob [6]. Ashes were placed into a molybdenum crucible and heated to 1520-1550ºC in a neutral atmosphere (N2). To ensure the absence of oxygen, the crucible was placed in a sacrificial graphite liner. Viscosity measurements were performed using a molybdenum rotating bob connected to a viscometer head. Measurements were conducted at incremental temperature steps within the temperature range 1200-1550oC. The temperature cooling step for each set of measurements was of 25ºC with at least 30 min intervals to equilibrate temperatures and compositions inside the crucible.
2.3 SEM/EPMA . Quenched slag samples were collected by dipping a molybdenum rod in the heated sample, removing it and then cooling it in water. Quenched samples were analyzed using Scanning Electron Microscopy (SEM) in backscattering mode. Energy Dispersive X-ray Spectroscopy (EDS) was undertaken to identify the elemental composition of selected
regions of the slag. Electron Probe Microanalysis (EPMA, JEOL8200) was used to identify the compositions of solid and liquid phases in quenched samples.
2.4 FactSage FactSage predicts equilibrium solid-liquid-gas phases and compositions based on Gibbs energy minimization [7]. Gibbs energy is calculated from optimized models with parameters based on empirical data with various compositions, temperatures and pressures. This information is contained within compound and solution databases which must be carefully selected prior to equilibrium calculations. The FactSage 6.1 Equilib module was utilized with the FACT53 and FToxid databases. Oxide components representing 1 wt% or more of the samples were included in the calculations. 1 g of N2 per 100 g of sample was also included. All possible gas compounds, solid compounds and solutions (excluding solutions containing titanium) were considered.
3 Results and Discussion 3.1 Viscosity of coal ash blended with petcoke ash and flux Viscosity measurements for six samples were performed (Figure 1). Sample C is coal ash. Sample CL is coal ash with 25 g of limestone per 100 g of coal ash. Sample CD is coal ash with 25 g of dolomite per 100 g of coal ash. Sample CP is coal ash with 25 g of petcoke ash per 100 g of coal ash. Sample CLP is coal ash with 25 g of limestone and 25 g of petcoke ash per 100 g of coal ash. Sample CDP is coal ash with 25 g of dolomite and 25 g of petcoke ash per 100 g of coal ash. The limestone and dolomite were not ashed prior to the viscosity measurement procedure. Since Genesee coal contains approximately 20 wt% ash, 25 g of limestone/dolomite per 100 g of coal ash is roughly equivalent to 5 g of limestone/dolomite per 100 g of coal. Similarly, considering Suncor petcoke contains approximately 5 wt% ash, 25 g of petcoke ash per 100 g of coal ash is roughly equivalent to 100 g of petcoke per 100 g of coal
Figure 1. Viscosity of ash blends.
Sample C has a viscosity well above 25 Pa·s for the entire temperature range tested. This was not a surprise as its SiO2 content is high and there were slag plugging problems when Genesee coal was used in a pilot-scale entrained-flow gasifier [8]. Both limestone and dolomite are effective for viscosity reduction of artificial Genesee coal ash. CaO and MgO, the main components of limestone and dolomite, are known network modifiers in silicate melt structures and hence are expected to reduce slag viscosity [9]. Dolomite has a slightly greater viscosity-reducing effect. This was also expected as dolomite has a higher MgO content and it has been shown that MgO has a greater effect than CaO [4,10,11]. Adding petcoke ash also reduces the viscosity, but does not have as much of an effect as limestone or dolomite. This result was more difficult to predict as, compared to other slag components, the effect of vanadium on slag viscosity is not widely reported in literature. Nonetheless, it has been shown to increase the viscosity of Jangsung coal ash in a previous study [10]. However, in the present study, the low network former (such as SiO2) and high network modifier (such as CaO and MgO) content of the petcoke ash may dominate over vanadium’s viscosity-increasing effect. Even with limestone or dolomite already added to the coal ash, petcoke ash addition slightly decreases the resulting slag’s viscosity. The temperature of critical viscosity (Tcv) is a parameter used to describe crystalline-type slags (as opposed to glassy-type slags which do not have a Tcv). It is the temperature below which there is an abrupt increase of the viscosity’s dependence on temperature [9]. It is often assumed that below this temperature, slag viscosity is affected
by the presence of solid crystals. There is no obvious Tcv for samples C or CP in the temperature range studied. Sample CL has an approximate Tcv at 1325ºC, sample CD has one at 1275ºC, sample CLP has one at 1300ºC and sample CDP has one at 1275ºC.
3.2 FactSage and SEM/EPMA phase analysis To determine whether the Tcv’s corresponded to the onset of solids precipitating, FactSage predictions were made for the temperature ranges of the viscosity measurements (Figure 2). All mass fractions given in Figure 2 exclude the gas phase. In addition to FactSage predictions, quenched slag samples were analysed using SEM (Figure 3) and EPMA (Table 2). Note that some small white inclusions were observed by SEM for most samples. When they were big enough for EPMA, the compositions obtained were metal-sulphur compounds. These inclusions were not considered in the following analysis due to their relative scarcity. For sample C, FactSage predicts that upon cooling, mullite starts to precipitate at 1425ºC and reaches 4.4 wt% at 1375ºC (Figure 2a). However, SEM/EPMA revealed only liquid down to 1375ºC (Figure 3a). For sample CL, FactSage predicts that upon cooling, anorthite should start to precipitate around 1375ºC and go up to 18.0 wt% at 1300ºC (Figure 2b). Again, SEM/EPMA did not detect any solids down to 1300ºC. For sample CD, FactSage predicts that upon cooling, anorthite should begin to precipitate around 1300ºC and reach 12.4 wt% at 1250ºC (Figure 2c). Formation of cordierite (1.2 wt%) is also predicted at 1250ºC. SEM/EPMA reveals only a liquid phase at 1350ºC, but some solids (anorthite) at 1250ºC (Figure 3b). For sample CP, FactSage does not predict any precipitation down to 1375ºC (Figure 2d). SEM/EPMA did not detect any solids at 1440ºC or 1400ºC, but did detect a slight amount of an aluminum compound at 1375ºC (Figure 3c). For sample CLP, FactSage predicts 5.9 wt% anorthite at 1350ºC, and 19.wt% anorthite at 1275ºC (Figure 2e). SEM/EPMA does see solids (anorthite) at these temperatures, with more present at 1275ºC (Figure 3d,e). For sample CDP, FactSage predicts only a liquid phase at 1325ºC, 10.0 wt% anorthite at 1275ºC and 14.8 wt% anorthite at 1250ºC (Figure 2f). Increasing solids (anorthite) were detected by SEM/EPMA from 1325ºC to 1250ºC (Figure 3f-h). There are thus some discrepancies between the FactSage predictions and SEM/EPMA observations. Perhaps, when low amounts of solids are formed, they may not be detected by SEM/EPMA due to sample heterogeneity.
1
0.8 0.6
Liquid slag
0.4
Mullite
0.2 0 1375
0.8
Liquid slag
0.6 Anorthite
0.4 0.2 0
1425
1475
1300
1525
Temperature (°C)
a)
Mass fraction
Mass fraction
1
1350
1400
1450
1500
1550
Temperature (°C)
b)
0.8
Liquid slag
0.6
Anorthite
0.4
Cordierite
0.2
1
Mass fraction
Mass fraction
1 0.8 0.6
0.2 0 1375
0 1250 1300 1350 1400 1450 1500 1550
1475
d)
c)
1
0.8
Liquid slag
0.6 Anorthite
0.4 0.2 0 1275
1325
1375
1425
Mass fraction
1
Mass fraction
1425
Temperature (°C)
Temperature (°C)
0.8 0.6
Liquid slag
0.4
Anorthite
0.2 0 1250
1475
Temperature (°C)
e)
Liquid slag
0.4
f)
1300
1350
1400
1450
1500
Temperature (°C)
Figure 2. FactSage phase predictions. (a) Sample C. (b) Sample CL. (c) Sample CD. (d) Sample CP. (e) Sample CLP. (e) Sample CDP.
a)
b)
c)
d)
e)
f)
g)
h)
Figure 3. SEM images. (a) Sample C at 1375ºC. (b) Sample CD at 1250ºC. (c) Sample CP at 1375ºC. (d) Sample CLP at 1350ºC. (e) Sample CLP at 1275ºC. (f) Sample CDP at 1325ºC. (g) Sample CDP at 1275ºC. (h) Sample CDP at 1250ºC.
Table 2. Composition of phases detected by SEM/EPMA Sample
Temperature (°C) [form]
SiO2 (wt%)
Al2O3 (wt%)
CaO (wt%)
Fe2O3 (wt%)
MgO (wt%)
K 2O (wt%)
Na2O (wt%)
V 2O 3 (wt%)
C C CL CL CL CD CD CD CP CP CP CP CLP CLP CLP CLP CLP CLP CDP CDP CDP CDP CDP CDP
1475 [liquid] 1375 [liquid] 1400 [liquid] 1350 [liquid] 1300 [liquid] 1350 [liquid] 1250 [liquid] 1250 [solid] 1440 [liquid] 1400 [liquid] 1375 [liquid] 1375 [solid] 1350 [liquid] 1350 [solid] 1300 [liquid] 1300 [solid] 1275 [liquid] 1275 [solid] 1325 [liquid] 1325 [solid] 1275 [liquid] 1275 [solid] 1250 [liquid] 1250 [solid]
59.58 60.20 53.40 53.29 53.29 51.01 50.77 49.54 59.14 59.46 59.66 0.02 52.13 53.07 51.76 49.06 52.11 48.29 51.49 50.38 51.83 50.21 51.64 49.50
23.16 23.10 20.39 20.37 20.28 22.16 22.20 31.84 22.98 22.64 22.84 100.32 18.95 42.57 20.12 32.03 18.78 32.09 21.55 32.66 18.79 32.27 17.70 32.82
4.54 4.50 16.79 16.81 16.85 11.93 11.93 15.41 7.86 7.87 7.74 0.02 16.97 0.71 16.95 15.77 17.12 15.59 12.81 15.80 11.98 15.99 11.76 16.23
5.47 5.42 3.99 4.00 3.96 4.51 4.59 0.24 4.01 4.00 4.14 0.08 3.73 0.10 3.84 0.19 4.55 0.22 5.24 0.21 6.18 0.18 6.79 0.20
1.60 1.61 1.56 1.55 1.55 6.28 6.39 0.34 2.21 2.25 2.24 0.00 2.31 0.03 2.04 0.19 2.38 0.18 6.17 0.34 7.64 0.34 8.31 0.28
0.65 0.66 0.58 0.58 0.57 0.61 0.61 0.12 0.78 0.76 0.74 0.01 0.83 1.46 0.71 0.14 0.77 0.14 0.73 0.13 0.86 0.11 0.91 0.11
1.83 1.86 3.12 3.12 3.11 3.41 3.39 3.40 6.41 6.68 6.48 0.00 3.22 2.10 3.07 3.10 3.14 3.18 6.73 5.92 6.68 6.02 6.19 5.95
0.14 0.12 0.06 0.19 0.17 0.05 0.29 0.00 1.11 1.06 1.05 0.00 1.60 0.00 1.03 0.12 1.19 0.35 0.92 0.10 0.99 0.16 1.44 0.16
For a given sample, higher viscosities are observed with higher mass fractions of solids (as determined by FactSage and estimated visually by SEM). However, at a given temperature it is not necessarily the sample with the greatest solid fraction which has the highest viscosity. For instance, C and CP reach a viscosity well above 150 Pa·s when very little or no solids are observed and predicted. Other samples have a viscosity well below 150 Pa·s despite the presence of solids (as determined by FactSage and SEM). Also, at 1250ºC, CDP has a greater solids fraction than CD (as determined by FactSage and estimated visually by SEM), but a lower viscosity. Hence the solids fraction is not a good indicator of absolute viscosity. Tcv is sometimes assumed to be the temperature below which viscosity is affected by the presence of precipitated solids, but above which viscosity is not affected by the presence of precipitated solids [9]. C and CP may be considered glassy-type slags as the viscosity increases sharply without the presence of solids. The approximate Tcv for the other slags are all at temperatures with 0.08-0.16 wt% solids (according to FactSage predictions). Solids were observed by SEM at, or near, the Tcv of all these samples, except for sample CL. As a result, these slags could be classified as crystalline slags. The
presence of solids does not seem to have a great impact on the viscosity value, but may have an impact on the viscosity-temperature dependence and flow type (Newtonian vs. non-Newtonian). Accordingly, the temperature ranges within which measurements could be performed below each Tcv were very small as the observed shear stress-shear rate relationship was no longer linear (a sign of non-Newtonian behaviour).
3.3 Viscosity predictions Five slag viscosity models were tested for performance with the viscosity data collected in this study (Table 3). The average absolute logarithmic error (AALE) was used as a performance indicator [11]. Since the Kalmanovitch-Frank model provides the lowest error (see Table 3), it was subsequently used to evaluate the effectiveness of dolomite as a viscosity reducer in three scenarios involving Genesee coal and Suncor petcoke. Table 3. Average absolute logarithmic error of viscosity model predictions
Model
AALE
Urbain [9] Riboud [9] Kalmanovitch-Frank [9] BBHLW [4] ANN [11]
0.300 0.257 0.131 0.468 0.232
In the first scenario (Figure 4), Genesee coal is the main fuel and some petcoke is added (5 g of petcoke ash per 100 g of Genesee coal ash, roughly equivalent to 20 g of petcoke per 100 g of coal). Dolomite is varied from 0 to 50 g per 100 g of coal ash. Assuming a slag tapping temperature of 1450ºC and maximum viscosity of 25 Pa·s, approximately 20 g of dolomite per 100 g of coal ash (or 4 g of dolomite per 100 g of coal) are required. In the second scenario (Figure 5), Genesee coal and petcoke are used in roughly the same proportion (25 g of petcoke ash per 100 g of Genesee coal ash, roughly equivalent to 100 g of petcoke per 100 g of coal). Dolomite is varied from 0 to 50 g per 100 g of coal ash. Assuming a slag tapping temperature of 1450ºC and maximum viscosity of 25 Pa·s, approximately 10 g of dolomite per 100 g of coal ash (or 2 g of dolomite per 100 g of coal) are required.
In the third scenario (Figure 6), only petcoke is used as fuel. Dolomite is varied from 0 to 50 g per 100 g of petcoke ash. Assuming a slag tapping temperature of 1450ºC and maximum viscosity of 25 Pa·s, no dolomite is required. However, caution is advised when applying these predictions. The vanadium content of the pure petcoke ash is much greater than the vanadium content in the previous prediction scenarios and viscosity measurements. The Kalmanovitch-Frank viscosity model does not account for vanadium and the actual viscosity may be much higher. Also, before running a gasifier with pure petcoke, other factors such as flame stability and minimum ash requirements should be considered.
Figure 4. Viscosity predictions for Genesee coal ash with 5 g of petcoke per 100 g of coal ash and a varying amount of dolomite.
Figure 5. Viscosity predictions for Genesee coal ash with 25 g of petcoke per 100 g of coal ash and a varying amount of dolomite.
Figure 6. Viscosity predictions for petcoke ash with a varying amount of dolomite.
4 Conclusion Viscosity measurements were performed with blends of artificial Genesee coal ash, reduced Suncor petcoke ash, limestone and dolomite. Both limestone and dolomite are effective slag viscosity reducers. The addition of petcoke ash also reduces the viscosity of the coal ash, but not to the same extent. All tested samples containing limestone or dolomite can be categorized as crystalline-type slags, while those without display glassytype behaviour. FactSage predictions and SEM/EPMA observations do not agree for all samples and temperatures, but in general, the Tcv of crystalline slags corresponded with the formation of precipitates. A slag viscosity model was used to provide recommendations as to the optimal dolomite fluxing ratio for three Genesee coal-Suncor petcoke blending scenarios. Based on the viscosity measurements and predictions, various Genesee coal-Suncor petcoke blends are suitable for gasification but may require some limestone or dolomite. However, to properly assess the suitability of pure petcoke or blends containing a petcoke with greater vanadium content, more viscosity measurements should be conducted as current slag viscosity and phase equilibrium models do not consider vanadium.
Acknowledgements The authors are grateful to the Natural Sciences and Engineering Research Council of Canada for financial assistance.
References [1] L. Roskill Information Services, The Economics of Petroleum Coke, 5th Edition, (2007). [2] E. Furimsky, Gasification of oil sand coke: Review, Fuel Process Technol. 56 (1998) 263-290. [3] E. Furimsky, Gasification in Petroleum Refinery of 21st Century, Oil and Gas Science and Technology. 54 (1999) 597-618. [4] G.J. Browning, G.W. Bryant, H.J. Hurst, J.A. Lucas, T.F. Wall, An empirical method for the prediction of coal ash slag viscosity, Energy and Fuels. 17 (2003) 731-737. [5] B.C. Folkedahl, H.H. Schobert, Effects of atmosphere on viscosity of selected bituminous and low-rank coal ash slags, Energy and Fuels. 19 (2005) 208-215. [6] H.J. Hurst, F. Novak, J.H. Patterson, Viscosity measurements and empirical predictions for some model gasifier slags, Fuel. 78 (1999) 439-444.
[7] C.W. Bale, A.D. Pelton, W.T. Thompson, G. Eriksson, K. Hack, P. Chartrand, et al., FactSage (http://www.factsage.com), (2009). [8] A. Cousins, D.J. McCalden, R.W. Hughes, D.Y. Lu, E.J. Anthony, Entrained-flow gasifier fuel blending studies at pilot scale, Can. J. Chem. Eng. 86 (2008) 335-346. [9] S. Vargas, F.J. Frandsen, K. Dam-Johansen, Rheological properties of hightemperature melts of coal ashes and other silicates, Progress in Energy and Combustion Science. 27 (2001) 237-429. [10] W. Park, M.S. Oh, Slagging of petroleum coke ash using Korean anthracites, J. Ind. Eng. Chem. 14 (2008) 350-356. [11] M.A. Duchesne, A. Macchi, D.Y. Lu, R.W. Hughes, D. McCalden, E.J. Anthony, Artificial neural network model to predict slag viscosity over a broad range of temperatures and slag compositions, Fuel Process Technol. 91 (2010) 831-836.
EFFECT OF DENSE MEDIUM SEPARATION OF A SOUTH AFRICAN COAL SOURCE ON SLAG-LIQUID FORMATION: AN EXPERIMENTAL AND FACTSAGE APPROACH JC van Dyk∗, FB Waanders** *Sasol Technology, R&D Division, Syngas and Coal Technologies, P. O. Box 1, Sasolburg, 1947, South Africa, Tel: +27 16 9604375; Fax; +27 11 5224806, [email protected] **School of Chemical and Minerals Engineering, North West-University, Potchefstroom 2520, South Africa, Tel: +27 18 299 1994; Fax; +27 18 299 1535; e-mail: [email protected]
The AFT of coals and coal blends is one of the parameters currently widely used in coal marketing and utilization to assess coal quality, ash fusibility and melting characteristics, as well as to predict the melting behaviour of the coal ash in coal conversion processes. The AFT of a coal source gives an indication of the extent to which ash agglomeration and ash clinkering are likely to occur within the gasifier. It has been demonstrated that ash flow temperature can be correlated with FACT equilibrium calculations. The principle aim of this investigation is to quantify the effect of dense medium separation (DMS) and the change in mineral composition on slag-liquid formation during fixed bed gasification, by using amongst others, experimental results derived from AFT analyses and FACT equilibrium calculations. The results indicated that dense medium separation of coal has a significant effect on slag-liquid formation and the associated mineral matter transformations during gasification. Results indicated that the amount of anorthite increased with decreasing relative density and that the amount of slag-liquid present at 1250oC during gasification decreased with decreasing relative density. The higher concentration of CaO seems to result in a higher amount of anorthite formation at specific operating temperatures. 1.
INTRODUCTION
Coal sources vary substantially in terms of chemical and physical properties. These properties directly impact on gasifier behaviour, and more specifically on the ash fusion temperature (AFT) and slag-liquid formation. The AFT of coals and coal blends is one of the parameters currently widely used in coal marketing and utilization to assess coal quality, ash fusibility and melting characteristics, as well as to predict the melting behaviour of the coal ash in coal conversion processes [Jak, 2002]. Due to the large variation in coal properties from various sources, detail coal and feedstock characteristics are essential to predict gasification performance when a specific coal source is to be gasified, either for fixed bed gasification or entrained flow gasification purposes. One property that specifically gives detail information on the suitability of a coal source for gasification purposes is the AFT. The AFT of a coal source gives an indication of the extent to which ash agglomeration and ash clinkering are likely to occur within the gasifier [Van Dyk, et. al., 2001, and Alpern, et. al., 1984 and Seggiani, 1999]. Ash clinkering inside a fixed bed gasifier can cause channel burning, pressure drop problems and unstable gasifier operation, whereas in entrained flow gasification technologies, flux addition and viscosity are critical parameters to have a detailed understanding of mineral matter transformations [Van Dyk, et. al., 2001].
∗
Corresponding author. Tel.: +27 16 960 4375; fax.: +27 11 522 4806; e-mail: [email protected]
Although the standard AFT analysis is currently used as the only predictive tool for ash fusion temperature determination of coal, literature studies have shown that this may not represent the actual flow temperature of certain minerals and mineral phases [Alpern, et. al., 1984]. Iron (Fe), for example, in a specific phase can slag at temperatures as low as 700oC and then solidify again. This is not reflected by a standard AFT analysis. In AFT analyses the softening and flow (melting or slagging) behaviour of ash as it is heated through the temperature ranges to a specified temperature are measured [Carpenter, 2002], which in this study is 1600oC under oxidizing conditions. A cone of ash is prepared by the standard ashing procedure as used for a proximate analysis and then heated at a controlled rate in an oxidizing atmosphere to simulate the gasification environment in the ash bed [Carpenter, 2002]. The four temperatures at which the melting process proceeds are illustrated in Figure 1:
A
B
C
D
E
(A) Original cone before heating (B) Initial deformation temperature where first rounding of cone tip is taking place (IDT) (C) Softening or sphere temperature where the cone height = cone width (ST) (D) Hemispherical temperature where cone height = ½ cone width (HT) (E) Fluid or flow temperature where cone height = 1.6mm (FT) FIGURE 1 CHANGES OF CONE SHAPE DURING ASH FLOW TRANSFORMATION [CARPENTER, 2002 AND SLEGEIR, ET. AL., 1988] (ASTM D1857)
Although AFT tests are widely employed, they do not always predict the AFT behaviour accurately. Two ashes which have similar fusion characteristics can have significant different melting behaviours [Carpenter, 2002]. Advantages of the AFT test include: • Widely employed • Standardized • Inexpensive • Capable of automation Concerns against the standard AFT test are the following [Carpenter, 2002]: • Subjective - based on observations rather than measurements • Poor reproducibility. The correlations and approach described by Slegeir, et. al. [1988] routinely provide results that are comparable to the ASTM reproducibility limits, but that mean deviations for the average and expert systems of about 100oF (38oC) are attainable. • The IDT is not the temperature at which melting begins. • AFT’s are measured over short periods, whereas deposits typically accumulated for hours are formed during cooling
2
Although the standard AFT test is currently used as the only predictive tool for measuring the AFT of coal, Alpern (1984) has shown that this may not represent the actual flow temperature of certain minerals and mineral phases. Various authors, such as Seggiani [1999] and Alpern et. al. [1984], have reported and expressed the fusibility of the ash as a function of the content of the eight principal oxides frequently found in coal ash, i.e. SiO2, Al2O3, TiO2, Fe2O3, CaO, MgO, Na2O and K2O. The acid/base ratio is the most frequently used parameter for correlating ash fusibility with coal ash composition. Slegeir, et. al. (1988) highlighted the fact that coal ash fusibility characteristics are difficult to determine precisely, partly because coal ash contains many components and does not have a sharp melting point like a pure compound. Correlations between AFT data and ash composition indicates that, although the regression approaches are more complicated, in many cases they are no more accurate than the approaches based on the acid-base formalism. Correlations are thus not obtained with single elements, but as an interaction between different ash elements [Alpern, et. al., 1984 and Seggiani, 1999]. Fluxing components in the ash, specifically Ca, Mg and Fe, provide a fair indication of the expected ash fusion behaviour. A Ca and/or Fe-rich coal source, according to AFT analyses, normally has a lower AFT due to the fluxing properties of the Ca and Fe minerals. Collet (2002) believes that sodium in the coals, is mainly responsible for ash clinkering, and thus is the limit of sodium content at which clinkering can be avoided. 2.
PURPOSE OF THIS STUDY
Coal preparation, like screening, beneficiation, destoning or any other process will change the coal properties, i.e. ash content, flow temperature, mineral composition, slag-liquid formation, etc. During a study by Jak (2002), it was concluded that the phase equilibrium approach, as an additional tool, can be successfully applied to the understanding of mineral transformations and liquid characterization of coal mineral matter. It has been demonstrated that AFT’s can be correlated with FACT equilibrium calculations, where Fact is a powerful predictive tool, which can be applied to a wide range of coal utilization technologies [Jak, 2002]. However, specific phenomena do not always follow the AFT trends and normal trends. Kaolinite, for example, is a common clay mineral with high concentrations of Si and Al that can react with other elements, like sodium which can be a fluxing agent to form products that melt at significantly higher temperatures [Ross, et. al., 2003]. Taken the results from literature into account, the principal aim of this investigation is to quantify the effect of coal preparation or manipulation and the subsequent change in the mineral composition on slag-liquid formation during gasification, by using amongst others, experimental results and FACT equilibrium calculations. 3.
EXPERIMENTAL AND MODELLING APPROACH
In order to illustrate the hypothesis that coal preparation will have a significant effect on coal composition of the product, a typical dense medium separation (DMS) process was experimentally simulated as an example. A South-African Highveld bituminous coal source was used as starting material. 3.1
Coal characterization and DMS results
The proximate analyses of the sample used in this study are given in Table 1.
3
TABLE 1 PROXIMATE ANALYSIS (MASS %) OF COAL SAMPLES Inherent Volatile Fixed carbon Ash moisture content 4.0 21.4 46.3 28.4
Beneficiation of the coal sample was simulated by means of float/sink analyses. The sample was washed at various relative densities from RD=1.4 to 2.1 and each individual fraction was then characterized. Coal washing was performed according to the ISO 7936 standard, where a 25mm to 0.5mm size fraction sample was prepared by crushing and screening steps. Individual particle size fractions >25mm were crushed to –25mm and screened at 0.5mm before washing. 3.2
FACTSAGE
FactSage modelling provides the opportunity to calculate and manipulate phase diagrams, but has been established mainly in the field of complex chemical equilibrium and process simulations. For example, with FactSage it is possible to access both Fact (slag, matte, salt, ceramic and aqueous) and alloy databases, import and export streams and mixtures and also import ChemSage data files. Another advantage of FactSage is that it can also handle carbon reactions together with the minerals present in such a sample, whilst varying the gas composition and atmosphere. GTT Technologies are the developers and databases administrators of the thermodynamic software and packages, such as FactSage (www.factsage.com). FactSage is the fusion of two well-known software packages in the computational thermochemistry – Fact-Win and ChemSage. The Fact development which already started in the late 70’s, is today the largest thermochemical package and database available. FactSage modelling supply insight into specific mineral interactions, slag formation and liquid temperatures of mineral compositions. The specific value for Sasol in this study will be that these thermochemistry models can be used to analyze equilibrium conditions for reactions occurring between inorganic and/or organic materials, as well as providing insight into the mineral formation and slag formation speciation. The database will assist in understanding, as well as predicting, what might happen with specific coal and mineral sources inside the gasification process. In a study conducted by the University of Queensland [Jak and Hayes, 2002] the focus was on the relationship between ash fusion temperature and FACT liquidus calculations. It has been demonstrated that the application of FACT computer software and database for the prediction of AFT is successful. Thermo-equilibrium simulations also supply detailed insight into the slagging behavior and associated changes in mineral composition during a gasification process even in the presence of the organic components. The focus of this study will be on slag-liquid formation and mineral changes as function of temperature for any gasification process. The critical temperature zone for various gasification processes are given in Figure 2-4.
4
Operating zone
FIGURE 2 FIXED BED GASIFICATION [Holt, 2006]
Operating zone
FIGURE 3 FLUIDIZED BED GASIFICATION [Holt, 2006]
5
Operating zone
Operating zone
FIGURE 4 ENTRAINED FLOW GASIFICATION [Higman and Van der Burgt, 2007]
4.
RESULTS
The experimental results based on dense medium separation, as well as the FACT equilibrium results will be discussed herewith. 4.1
Effect of dense medium separation on mineral composition and AFT
The coal composition of the original coal sample, together with the 5 prepared fractions by means of dense medium separation are given in Table 2. This data will also be used for modelling purposes in the FactSage thermodynamic equilibrium calculations.
6
TABLE 2 CHARACTERISTICS OF PREPARED FLOAT FRACTIONS ORIGINAL Description COAL F2.1 F1.95 F1.8 Yields 100.0 90.3 86.8 84.4 PROXIMATE ANALYSES (mass % Inherent H2O 4.0 4.17 4.22 4.25 Ash 28.4 23.1 21.9 21.2 Volatile matter 21.4 22.3 22.6 22.7 Fixed carbon 46.3 50.4 51.3 51.9 ASH COMPOSITION (mass %) SiO2 46.8 45.5 44.9 44.5 Al2O3 25.5 25.9 25.9 26 Fe2O3 2.9 2.61 2.58 2.5 P2O5 1.2 1.32 1.35 1.38 TiO2 1.5 1.49 1.47 1.47 CaO 10.8 11.5 11.8 12 MgO 3.3 3.61 3.71 3.78 K2O 0.7 0.69 0.68 0.67 Na2O 0.6 0.61 0.63 0.64 SO3 5.6 5.86 6.03 6.12 LOI* 1.0 0.97 0.97 0.96 * Loss of ignition ** F1.4 to F2.1 denotes the different DMS fractions.
F1.6 63.4
F1.4 12.6
4.41 17.4 23.8 54.4
4.6 9.5 28.6 57.3
42.2 25.8 2.22 1.61 1.47 13.3 4.07 0.62 0.71 6.99 1.04
36.8 25.8 2.3 2.4 1.9 14.4 3.5 0.6 1.0 9.6 1.6
From Table 2 it is clear that significant differences in ash content, as well as ash composition, were obtained by dense medium separation. These characteristics (proximate and ash composition) of the fractions will be used individually to quantify the slag-liquid formation and differences based on FactSage modelling. The individual fractions will be treated as individual coal sources as if gasified individually per prepared fraction. The wash curve of the coal sample, as well as the ash content of the coal, is given in Figure 5.
7
90 Cummulative yield (mass %)
23.1
Yield (mass %)
21.9
21.2
Ash content (mass %)
90
87
84
80
25
20
17.4
70 60
15
63
50 9.5
40
10
30 20
5
10
13
0
Cummulative ash content (mass %)
100
0 F1.4
F1.6
F1.8
F1.95
F2.1
Relative density FIGURE 5 CUMMULATIVE YIELD AND ASH CONTENT OF THE COAL SAMPLE
The yields above RD=1.8 were relative high (>80%), where after the yield decreased significantly towards washing at lower relative densities. The ash content decreased from 21.9% at a RD=1.95 to as low as 9.5% at a RD=1.4. Figure 6 depicts the changes in mineral composition with dense medium eparation, i.e. changes in Si, Ca and percentage basic components. 48
21
46
19
44
17
42
15
40
13
38
11
36
9
34
SiO2 CaO
7
% SiO 2 (cummulative mass %)
% CaO and % Basic components (cummulative mass %)
23
32
% Basic components 5
30 F1.4
F1.6
F1.8
F1.95
F2.1
Relative density
FIGURE 6 EFFECT OF DENSE MEDIUM SEPARATION ON Si, Al, Ca AND Fe CONTENTS
8
From Figure 6 the following observations can be highlighted: • The SiO2 concentration decreases with decreasing relative density, implying and confirming the removal of high density rock fragments or extraneous particles with destoning. • The percentage basic components (Ca, Fe, Mg and Na) increased with decreasing relative density. A similar trend is observed for CaO, which contributes more than 50% of the total basic component. Ca-particles are inherent to the coal structure and destoning only removes a limited percentage of Ca, actually increasing the Ca-content in relation to other elements.
1600
o
Ash flow ( C) (oC) Measured ash temperature flow temperature
An ash flow temperature versus % basic component (Ca, Mg, Na, K and Fe) graph is given in Figure 7. Based on these data points for each relative density fraction, a clear decrease in the flow temperature is visible with increasing % basic components up to ±20%, after which the flow temperature is starting to increase. More detail discussions and evaluation will follow with the FactSage modelling results.
1550 1500 1450 1400 1350 0
5
10
15
20
25
% Basic components (Ca, Na, Mg, K and Fe) FIGURE 7 CHANGE IN FLOW TEMPERATURE VERSUS % BASIC COMPONENTS
This correlation is also confirmed by work done by Microbeam Technologies (2003) on Sasol’s coal blend indicating that the ash composition (oxide elemental composition) has an acceptable fit to the ash flow prediction curve (Figure 8).
9
___ Actual ----- Target 2900 T250 (°F) Ash flow temperature
2800 2700 2
2600
y = 1.1914x - 87.066x + 3867 2 R = 0.9489
2500 2400 2300 2200 2100 2000 10
20
30 40 % Basic components percent basic
50
60
FIGURE 8 ASH MELTING TEMPERATURE PREDITION CURVE [MICROBEAM TECHNOLOGIES, 2003]
Taking into account the results discussed above, it can be concluded that significant differences in mineral composition can be obtained by dense medium separation. 4.2
Effect of dense medium separation on slag-liquid formation based on FactSage thermodynamic equilibrium calculations
The amount of slag-liquid formation and associated anorthite (CaAl2Si2O8) formation, at a randomly selected temperature of 1250oC as determined by thermo-equilibrium simulation, is given in Table 3 and graphically illustrated in Figure 9. A temperature of 1250oC was selected to obtain partial slag-liquid formation, and simultaneously have crystalline material, i.e. anorthite, present in the mixture. To determine the correlations between slag-liquid formation (based on FactSage simulation) of coal ash and different mineral ratios, the following ratios from literature [Slegeir, et. al. 1988, Seggiani, 1999 and Collet, 2002] were decided on and calculated for use in this study, where the ash composition is expressed as mass percentages: % Base = Na2O + K2O + CaO + MgO + Fe2O3 % Acid = SiO2 + Al2O3 + TiO2 Silica value = SiO2 / (SiO2 + Fe2O3 + CaO + MgO) Dolomite ratio = (CaO + MgO) / (Fe2O3 + CaO + MgO + K2O + Na2O) Acidity = (SiO2 + Al2O3) / (Fe2O3 + CaO + MgO + Na2O + K2O) Base-acid ratio = % Base / % Acid
(1) (2) (3) (4) (5)
The calculated values, according to equations 1 to 5, together with the slag-liquid and anorthite formation, are given in Table 3.
10
TABLE 3 EXPERIMENTAL COAL AND COAL ASH RATIOS AND FACTSAGE THERMODYNAMIC EQUILIBRIUM OUTPUTS (ANORTHITE AND SLAG-LIQUID FORMATION AT 1250oC) Original Description coal F2.1 F1.95 F1.8 F1.6 F1.4 % Base 18.4 19.0 19.4 19.6 20.9 21.8 % Acid 73.8 72.9 72.3 72.0 69.5 64.5 Silica ratio 0.7 0.7 0.7 0.7 0.7 0.6 Dolomite ratio 0.8 0.8 0.8 0.8 0.8 0.8 Acidity 3.9 3.8 3.7 3.6 3.3 2.9 Acid/Base ratio 4.0 3.8 3.7 3.7 3.3 3.0 % Anorthite* 30.5 33.1 33.5 33.9 35.3 36.9 % Slag-liquid* 58.1 55.4 54.6 53.9 50.5 43.8
* FactSage results 70
% Anorthite* % Slag-liquid*
60
Mass %
50
40
30
20
10
0 F1.4
F1.6
F1.8
F1.95
F2.1
Original coal
Relative density
FIGURE 9 ANORTHITE AND SLAG-LIQUID FORMATION BASED ON FACTSAGE THERM-EQUILIBRIUM CALCULATIONS
From Table 3 and Figure 9 it is clear that the amount of anorthite increased with decreasing relative density and that the amount of slag-liquid present at 1250oC decreased with decreasing relative density. In order to explain these phenomena, a detailed look at the mineral composition (Table 2) as well as mineral ratios (Table 3) will have to be made. From Figure 10 a direct correlation can be made between anorthite formation and the CaO content of the prepared fractions. The detail mechanism of anorthite formation is published by Van, Waanders and Hack (2008).
11
40 y = 1.6x + 14.4 R2 = 0.9
38
Mass % Anorthite
36 34 32 30 28 26 24 22 20 10
11
12
13
14
15
Mass % CaO
FIGURE 10 CaO-CONTENT VERSUS ANORTHITE FORMATION
The higher concentration of CaO seems to result in a higher amount of anorthite formation. The anorthite formed as a product between the SiO2, Al2O3 and Ca-containing species. Take note that anorthite is a crystallized product from the slag formed at lower temperatures. Dense medium separation of this coal sample also supported the findings of another study on oxygen capture tendencies [Van Dyk and Waanders, 2008] of minerals during gasification. From that study, and in support of this study and findings, the following can be concluded: • The coal fractions with the highest concentration of CaO and acidic components (Al2O3 and SiO2) resulted in the highest percentage of Ca-Al-Si minerals (CaAl2Si2O8 - anorthite plus CaAl4Si2O10(OH)2 margarite) formation. In this study the float fraction at RD=1.4 yields these minerals. • The free-SiO2 in the mineral structure of coal sources resulted in forming minerals containing Mg, Na or Ca such as KAl3Si3O10(OH)2 (muscovite), Mg5Al2Si3O10(OH)8 (clinochlore), or other high oxygen molecule-capturing mineral compounds. Thus, if the free-SiO2 is decreased, i.e. by an increase in anorthite formation as with the float fraction at RD=1.4, the concentration of Si-oxygen capture compounds were then relatively low, with a high concentration of anorthite forming, as in the prepared fraction at RD=1.4, from the present investigation. Another output from FactSage that was studied for each prepared fraction was the temperature predicted by the FactSage thermodynamic equilibrium calculation where 100% of the mineral content will be slag-liquid. This information is given in Table 4 and compared to the flow temperature as obtained by the AFT analyses. TABLE 4 FLOW TEMPERATURE AND SLAG-LIQUID TEMPERATURE OF MINERALS FLOAT 2.1 FLOAT 1.95 FLOAT 1.8 Flow temperature obtained by AFT analyses (oC) 1550 1440 1410 FACTSAGE 100% slag-liquid 1565 1565 1550 prediction (oC) Difference between AFT analyses and prediction (oC) 15 125 140
12
FLOAT 1.6
FLOAT 1.4
1400
1410
1580
1600
180
190
From Table 4, a correlation between the difference in actual and predicted temperatures and the CaO content can be obtained, which is shown in Figure 10.
Difference between flow temperature (AFT) and slag-liquid (Factsage) simulation ( o C)
250 y = -10.6x2 + 302.1x - 1959.9 R2 = 1.0
200
150
100
50
0 10.0
10.5
11.0
11.5
12.0
12.5
13.0
13.5
14.0
14.5
15.0
Mass % CaO
FIGURE 10 CORRELATION BETWEEN THE DIFFERENCE IN FACTSAGE FLOW TEMPERATURE PREDICTION AND MEASURED FLOW TEMPERATURE VERSUS CaO CONTENT
From Figure 10 it can be seen that for CaO contents of 11-15% the predicted slag-liquid temperature from FACT is comparable with the standard flow temperature according to an AFT analysis. The difference of less than 10oC is within the experimental error of an AFT analyses of ±30oC. However, the increasing difference with increasing CaO content should be further discussed. According to Jak (2002) a different behaviour between predicted liquid temperatures from FACT and the AFT were also observed in other studies for higher Ca and Fe containing coal sources. It was found that high CaO containing ash cones melt and change shape at higher temperatures relative to the rest of the coal sources.
13
FIGURE 11 CaO CONTENT VERSUS SLAG-LIQUID FORMATION [JAK, 2002]
From Figure 11 it can be seen that the so-called turning point is also reported at about 10% CaO. For the coal used in this study (Figure 12), a decrease in slag-liquid formation for increasing CaO content was also observed.
70.0
Slag-liquid (mass %)
65.0 60.0 55.0 50.0 45.0 40.0 35.0 30.0 10
12
14
16
18
20
CaO content (mass%) FIGURE 12 SLAG-LIQUID FORMATION AT 1250oC versus CaO CONTENT
In an AFT test the material was found to be heterogeneous and consisted of particles with various compositions. As the temperature increases different diffusion processes take place within the ash cones. The rates of reaction, dissolution and homogenisation strongly depend on whether the intermediate phase is solid or liquid [Jak, 2002]. The presence of any liquid phase or low melting mineral significantly increases the dissolution since diffusion in the liquid is much
14
faster. The formation of the liquid and solid phases is directly related to the compositional and temperatures stability ranges of the phases that in turn are described by the phase equilibrium science [Jak, 2002]. These phenomena can be explained, as stated by Jak (2002), that the formation of the intermediate stable phase (anorthite) in the high CaO ash slags makes dissolution and melting rates slower and moves slower towards equilibrium. The higher CaOcontaining coal sources (such as the fractions prepared at lower relative densities), contain higher concentrations of anorthite (CaAl2Si2O8) and the proportion of the liquid lower, as also observed and confirmed in this study. In high CaO content coal sources, the Al-Si-Ca particles causes delay in the formation of slag-liquid. 6.
CONCLUSIONS
The amount of anorthite increased with decreasing relative density and the amount of slagliquid present at 1250oC during gasification decreased with decreasing relative density. The higher concentration of CaO seems to result in a higher amount of anorthite formation. The anorthite formed as a product between the SiO2, Al2O3 and Ca-containing species. Thus, a minimum amount of fluxing agent, which enhances slag-liquid formation and anorthite crystallization and a high concentration of acid component SiO2, which suppresses slag-liquid formation, are then present, to form a slag-liquid material. Dense medium separation of this coal source also supported the findings on the oxygen capture tendencies of minerals during gasification where the coal float fractions at RD=1.4 yielded the highest concentration of CaO and acidic components (Al2O3 and SiO2) resulting in the highest percentage of Ca-Al-Si minerals (CaAl2Si2O8 - anorthite plus CaAl4Si2O10(OH)2 margarite) formation.The free-SiO2 in the mineral structure of coal sources resulted then in forming minerals containing Mg, Na or Ca to form new mineral compounds such as KAl3Si3O10(OH)2 (muscovite), Mg5Al2Si3O10(OH)8 (clinochlore), or other high oxygen molecule-capture mineral compounds. Thus, if the free-SiO2 is decreased or not present after gasification, i.e. by an increase in anorthite formation as with the float fraction at RD=1.4, the concentration of Sioxygen capture compounds were then relatively low, with a high concentration of anorthite forming, as in the prepared fraction at RD=1.4, from the present investigation. For a CaO content of less than 10%, the predicted slag-liquid temperature from FACT calculations is comparable with the standard flow temperature according to an AFT analyses, with a difference of less than 10oC, but for an increasing CaO content the difference increased. The formation of the liquid and solid phases is directly related to the compositional and temperature stability ranges of the phases that in turn are described by the phase equilibrium science. These phenomena can be explained by the fact that the formation of the intermediate stable phase (anorthite) in the high CaO ash slags makes dissolution and melting rates slower and moves slower towards equilibrium. The higher CaO-containing coal sources (such as the fractions prepared at lower relative densities), contain higher concentrations of anorthite (CaAl2Si2O8) and the proportion of the liquid lower, as also observed and confirmed in this study. In high CaO content coal sources, the Al-Si-Ca particles causes delay in the formation of slag-liquid. 7.
REFERENCES
Alpern, B., Nahuys, J., Martinez, L., Mineral matter in ash and non-washable coals – Its influence on chemical properties, In Symposium on Gondwana Coals, Lisbon – proceedings and papers, vol. 70 1984, p299-317 Comun. Serv. Geol. Portugal, 1984, t. 70, fasc. 2, pp. 299-317, 1988.
15
Bale, C.W., Chartrand, P., Degterov, S.A., Eriksson, G., Hack, K., Manfoud, R. B., Melancon, J., Pelton, A.D. and Peterson, S., FactSage Thermochemical Software and Databses, GTT-Technologies, Germany, Calphad, 2002, 26, p. 189-228. Collet, A.G., Matching gasifiers to coal, IEA Clean Coal Centre, 2002, p.1-64. Gray, V.R., Prediction of ash fusion temperature from ash composition for some New Zealand coals, Fuel 66 (1987), p.1230-1239. Higman, C. and Van der Burgt, M., Gasification, Gulf Professional Publishing, Amsterdam, 2007. Holt, N. Gasification technology status – December 2006, Electric Power Research Institute, Palo Alto, 2006 IEA Clean Coal Centre, Coal quality assessment – the validity of empirical tests, September 2002. Jak, E. and Hayes, P.C., Applications of the new F*A*C*T database to the prediction of melting behaviour of coal mineral matter, Coorperative Research Centre for Black Coal Utilisation, Pyrometallurgy Research Group, The University of Queensland, Australia, p.1-9, 2002. Jak, E., Prediction of coal ash fusion temperatures with the FACT thermodynamic computor package, Fuel, 81, 2002, p. 1655-1668. Keyser, M.J., Personal communication, Sasol Technology, [email protected], 2006. Keyser, M.J., van Dyk, J.C., 17th International Pittsburgh Coal Conference, 2000, Pittsburgh, USA, Full Scale Sasol/Lurgi Fixed Bed Test Gasifier Project: Experimental Design and Test Results. Keyser, M.J., van Dyk, J.C., Coetzer, R.L.C., Wagner, N.J., 8th Coal Science and Technology Conference of the Fossil Fuel Foundation of Africa, South Africa, 15-17 October 2002, Full Scale Sasol/Lurgi Fixed Bed Test Gasifier Project: Impact of particle size distribution, destoning and gasifier operating conditions on sulphur production. MICROBEAM TECHNOLOGIES, INC., www.microbeam.com, 2003. Ross, D.P., Kosminski, A. and Agnew, J.B., Reactions between sodium and silicon minerals during gasification of low-rank coal, 12th International Conference on Coal Science, 2003, Australia, p. 1-9. Seggiani, M., Empirical correlations of the ash fusion temperatures and temperature of critical viscosity for coal and biomass ashes, Fuel 78 1999, p. 1121-1125. Slegeir, W.A., Singletary, J.H. and Kohut, J.F., Application of a microcomputor to the determination of coal ash fusibility characteristics, Journal of Coal Quality, 1988, Volume 7, Number 2, p. 48-54. Van Dyk, J.C., Keyser, M.J., van Zyl, J.W., Suitability of feedstocks for the Sasol-Lurgi Fixed Bed Dry Bottom Gasification Process, GTC Conference, San Francisco, USA, October 2001. Van Dyk, J.C., Melzer, S. and Sobiecki, A., Mineral matter transformations during Sasol-Lurgi fixed bed dry bottom gasification – utilization of HT-XRD and FactSage modelling, Minerals Engineering 19 (2006), 1126-1135. Van Dyk, J.C., PhD Thesis – Manipulation of gasification coal feed in order to increase the ash fusion temperature of the coal to operate the gasifiers at higher temperatures, North West University, 2006. Van Dyk, J.C., Waanders, F.B. and van Heerden, J.H.P. Quantification of oxygen capture in mineral matter during gasification, Fuel 87, 2008, p2735-2744.
16
2010 Pittsburgh Coal Conference October 11-14, 2010
SHAPING SLAG FLOW IN AN ENTRAINED FLOW GASIFIER: NUMERICAL SIMULATION AND PHYSICAL EXPERIMENTS Randy Pummill, Graduate Research Assistant, University of Utah 50 Central Campus Dr Rm 3290 Salt Lake City, UT 84112 [email protected] Gabriel Hansen, Research Assistant, University of Utah 50 Central Campus Dr Rm 3290 Salt Lake City, UT 84112 [email protected] Dr. Kevin J. Whitty, Associate Professor, University of Utah 50 Central Campus Dr Rm 3290 Salt Lake City, UT 84112 [email protected]
INTRODUCTION The aggressive environment inside a high temperature slagging gasification reactor can make getting reliable data from the reactor difficult. Any probes used to obtain data are subject to the extreme temperatures and reducing environment of the reactor and will have a very short lifetime. One way to overcome this limitation is to use a non-invasive measuring device such as a tunable diode laser. By firing a tunable laser through the reactor, data such as temperature and major species concentrations can be obtained [1]. In order to function properly, the laser would need a clear line of sight across the reactor. As coal is converted, much of the mineral matter present in the coal remains as ash. In a gasification environment, the high temperatures cause the ash to become molten slag. The slag accumulates on the walls of the reactor and runs downward. This slag flow would interfere with the line of sight necessary for the laser measurements. It is the goal of this paper to demonstrate a possible solution that would divert the slag flow around sight ports in the reactor so that a clear line of sight is maintained during operation of the gasifier. Using a physical diverter attached to the refractory above the sight port would be inefficient, as the aggressive environment would destroy any such device in short time and the device would need to be replaced frequently. Instead, it is thought that by employing a focused stream of purge gas the slag can be diverted around the sight port. Experiments simulating flow of high viscosity fluid around sample ports and a numerical model were developed in order to
simulate and model the slag flow inside the reactor. The results of these experiments are presented here. NUMERICAL MODEL Theory There have been previous successful attempts to model slag inside a gasifier [2] [3]. In the model developed by Bockelie et al., the reactor was divided into two-dimensional vertical strips and solved simultaneously. These strips were then patched together into a pseudo-three-dimensional result. However, this approach assumed a smooth wall and so for the problem of modeling fluid diversion, it was necessary to take a different approach. To solve this problem numerically, a two-dimensional solution of the Navier-Stokes equations was formulated. The equations used in the solution were the standard continuity and momentum balance equations; also, since the thickness of the film is important in this case, the equations needed to include the oft-excluded z term. Figure (1) is minimal schematic of the model domain and also clarifies the coordinate system used. In this domain, the slag flow moves down a plane. It impacts and obstruction to simulate the fluid diverted by the purge stream and then is free to flow out the bottom of the domain.
1
The projection method was employed on a staggered grid to solve this problem. A staggered grid offsets the variables so that no two variables occupy the same grid point, as illustrated below in Figure (2). Here the x and y velocities are represented by the squares and triangles and the pressure is represented by the circles. The fluid thickness was calculated at the pressure nodes and then linearly interpolated to the x and y-direction velocity nodes. Red lines are used to mark the domain boundaries. The staggered grid separates the pressure nodes from the boundaries, which results in easier conversion of the pressure matrix while solving the problem.
Figure 1: Cartoon of the simulation domain which also shows the orientation of the coordinate system used.
Assuming constant density, the continuity equation for this system can be written as Figure 2: Illustration of a staggered grid. The squares mark grid points where the velocity in the x-direction is calculated. The triangles represent the velocity in the y-direction and the circles represent the pressure calculation point. The red grid lines indicate the boundary.
where Rd is the mass flux due to slag deposition. Assuming Newtonian fluid behavior, the momentum equation for the x-direction can be written as
The steps taken to solve this problem will be discussed in detail. The algorithm used to solve the problem is: 1. 2. 3. 4.
Additional simplifications can be made by removing the dimensional dependence of the equations. All velocities are scaled by g/w (gravitational constant divided by the length of one side of the domain), distances are scaled by w, and so on. This is a common approach to this kind of problem [4] and results in
where all the variables have been appropriately scaled and
5. 6. 7.
Set initial velocity and film thickness Impose boundary conditions Set initial matrix values for new time step Solve for intermediate velocity (v*) while neglecting pressure effects Iteratively solve the Poisson equation to determine the pressure gradient Use pressure to calculate the velocity of the next time step Perform a mass balance to calculate film thickness of next time step
The first step is only done once and the other steps are repeated for the requested number of time steps. Two matrices are initialized, one to represent the staggered grid and the other
2
to represent the fluid thickness. It was proposed set the fluid thickness in the empty grid locations, but this would require more interpolation calculations to be performed. A separate matrix is used to track the fluid thickness, as it is thought this will increase overall accuracy. For the second step, the top boundary is a Dirchlet condition set to have zero x and y-direction velocity. This represents the top wall of the reactor. All other boundaries, including the pressure boundaries enforced in Step (5), are Neumann boundary conditions, also with a value of zero. This allows the fluid to freely flow out the bottom or sides of the domain. For the tests reported here, the flow of the system was entirely gravity-driven. For step (4), the motion equation (for the x-direction) is discretized as
and
where
which is quite simple compared to some of the previous equations. The last step, Step (7), is to perform a mass balance on each pressure node to calculate the new fluid thickness. This is based on the continuity equation,
where ω is the Successive Over Relaxation (SOR) constant. Solving this equation gives the pressure field for the next time step. Tests in which the SOR constant changed from 1.00 to 2.00 in order to minimize the number of iterations required to find a solution to the pressure matrix. The optimal value of was found to be 1.93 for this problem. The next step, Step (6), is to use the pressure field to solve for the new velocity field. This is done (for the x-direction) using
and
and, finally,
A second similar equation is used for the y-direction. For step (5), the Poisson equation is solved iteratively using a Gauss-Seidel with Successive Over-Relaxation (GSSOR) method [5]. The discretized form of this step is solved as
where
These values are then interpolated linearly to the x and yvelocity nodes. The time step used was determined by a Von Nuemmann analysis [6]. The system was found to be stable for
Results and Discussion The code was run for two scenarios: a simple layer of slag flowing under gravitational pull without any obstruction and the same scenario with a rectangular obstruction. The major parameter of the simulation is the Reynolds number. By setting the Reynolds number of the system, most of the other parameters are also set. For the cases presented here the Reynolds number was set to 20. A one meter square plane was set as the domain with 25 grid point in both the x and y directions. The time step was then set at 1/20 of a second as per Equation (14). For these tests, the fluid was considered to be isothermal and of uniform viscosity and density. The first simple case was run for both a short amount of time (500 time steps) and a long time (5000 time steps). For
3
this simple flow case, the velocity field after 500 time steps is shown in Figure (3).
Figure 3: Velocity field for the short-timed case. Notice that the velocity at the top of the domain is zero and the quickly accelerates to a constant velocity distribution in the y-direction.
At the top of the domain the velocity is zero as per the boundary condition. The velocity begins to increase in the ydirection and reaches a constant gradient in the y-direction. Even in this short-timed case, a defined line is apparent where the fluid thickness is changing. This becomes obvious in Fig (4), which shows the fluid thickness for the simple, short-timed case. Looking at the two figures together, it can be seen that the areas with less fluid move more slowly than the thicker areas. This was the expected result. The plots make the difference look large, but examining the scales will show that the change in thickness is quite small (~3 mm).
For the long-timed case, the results are similar. As would be expected, the velocity and thickness distributions are even more pronounced. Figure (5) shows the velocity distribution for this case.
Figure 5: Velocity distribution for the long-timed case. Here the velocity gradient is even more pronounced than in the previous case.
The velocity field shows that as the fluid leaves the top section of the domain the velocity slows. Figure (6) shows the thickness distribution for this case. Again, the velocity and thickness distributions mirror each other. The fluid has continued to move down the wall and is exiting the bottom of the domain. The total thickness change is greater for this case, around 12mm.
Figure 6: Fluid thickness distribution for the long-timed case. The fluid has continued flowing downward. The fluid at the top boundary is lower than that observed in the first case; barely above the initial condition despite the constant mass deposition. Figure 4: Fluid thickness distribution for the short-timed case. The fluid flow is moving in the direction of gravity, as expected.
The fluid at the top boundary in the second case is about 3mm less thick than the fluid at the top of the first case. Even with the constant mass deposition across the domain, the fluid
4
has begun to recede. The fluid in the level part of the domain is much higher than in the first case, due to the flow of the fluid and the mass deposition. In both of these cases, the velocity in the x-direction was negligible. Since this is a gravity-driven flow, it was expected that there would be no x-direction velocity. For the second case with the rectangular obstruction, the code was run for 500 time steps. The velocity field is shown in Figure (7). The boundaries of the obstruction are shown with red lines.
meaning in the real gasifier. A heated system employing real slag running down refractory face was considered, but it was rejected due to the difficulty of keeping the slag above its temperature of critical viscosity through the entire flow circuit. Instead, a system was designed to try and simulate the most important aspects of the gasifier system. It was decided to test the effects of fluid viscosity and surface tension and the effect of the purge gas density on the ability of the gas to divert the flow around the sight port. For the physical experiments, a section of refractory was cast in the same geometry present in the entrained flow gasifier located at the University of Utah’s Industrial Combustion and Gasification Research Facility. The gasifier has an inner diameter of eight inches in the reaction section with 1.5 inch sight ports. The refractory was cast so that it also had these same characteristics. The cast piece was 3 feet in height. A photograph of the setup is shown in Figure (9).
Figure 7: Velocity field of the fluid with a rectangular obstruction. The obstruction is shown as a red box. The fluid moves around the obstruction as expected.
The fluid thickness distribution is shown in Figure (8). It can be seen that some fluid is predicted to build up against the leading edge of the obstruction and the fluid thickness drops behind the trailing edge of the obstruction. This makes intuitive sense and matches observations of the real experimental system.
Figure 8: Fluid thickness distribution for the obstructed case. The fluid builds up against the obstruction and thins in its wake.
PHYSICAL MODELING Experimental Setup The first consideration of designing the physical experiments was to ensure that the data collected would have
Figure 9: Photo of the experimental apparatus. The refractory is curved with an eight inch diameter and has a 1.5 inch hole to act as a sight port. A wooden frame supports the refractory in an upright position.
5
Water and silicone oils of four different viscosities were obtained to simulate the slag flow. Molten slag has a wide viscosity range that is heavily temperature dependent. Below a certain temperature, called the temperature of critical viscosity or Tcv, the slag behaves as a crystalline solid. Above the T cv the slag behaves approximately as a Newtonian fluid [7]. In order to flow freely enough to be tapped, the slag must have a viscosity of less than 25000 cP, with an optimal viscosity of 15,000 cP [8]. In order to include this range in the experiments, the oil viscosities chosen were 1, 100, 5000 and 30,000 cSt. Silicone oil was chosen as it is easy to obtain, has greatly modifiable viscosity, and its other physical properties, such as surface tension, are well known. Water was chosen as it has a high surface tension. By comparing the results for water with the results for the 1cSt oil, inference could be made concerning the effect of surface tension on the results. The fluid was pumped up to the top of the refractory section and allowed to flow downward. A piece of metal tubing with 1/8” holes drilled in it was used to distribute the oil at the top of the refractory. The 30,000 cSt oil also required the three additional white tubes (visible at the top of the refractory in Figure 1) in order to achieve a fairly uniform distribution of oil across the refractory face. The fluid thickness was approximately 1/8 inch for the water, and the 1, 100 and 5,000 cSt oils; the 30,000 cSt oil had an approximate thickness of 3/16 inch. The pump used was a Hydra-Cell membrane pump, which is capable of pumping very high viscosity fluids. A red laser was fixed inside the sight port of the refractory and directed at a piece of white paper. Having the paper made it easy to determine if the silicone oil was interfering with the line of sight as any oil interfering with the line of sight would cause the laser point to blur. To divert the slag around the sight port, a purge stream of inert gas was used. Eleven different devices were constructed to deliver the purge gas to the port. These included a simple setup allowing the gas to travel through the sight port diffusely and using tubes to focus the gas stream at the silicone oil at different angles and with different tube diameters. A table listing the purge devices is shown in Table (1). Figure (10) shows the last purge device, numbered (11) in the table, installed in the refractory section.
Table 1: Table of devices used to deliver the purge gas to the sight port. A short description of each device is also given.
Tube Size
Description
1
None
Diffuse flow through sight port
2
1/4"
Tube at the top of the hole, shooting forward
3
3/8"
Tube at the top of the hole, shooting forward
4
3/8"
5
1/2"
6
1/4"
7
3/8"
8
3/8"
9
1/2"
10
3/8"
Flattened tube, shot horizontally from top of hole Flattened tube, shot horizontally from top of hole Tube along bottom of port, angled to shoot upward Tube along bottom of port, angled to shoot upward Flattened tube along bottom of port, angled to shoot upward Flattened tube along bottom of port, angled to shoot upward Tube in center, laser shot down center of tube
11
1/4"
Tube in center of port, angled to shoot upward
#
Figure 10: Close up photograph showing device #11 installed in the refractory face. The tube can be seen protruding in the sight port, angled up towards the top edge of the port. Purge gas pushed through this tube served to divert the flow around the hole.
6
The purge gas flow rate was controlled using an adjustable rotameter. The flow was started at zero flow and increased manually over time. Both compressed air and carbon dioxide were used as purge gases in separate tests so that any relationship between good fluid diversion and gas density could be observed. Additionally, in order to simulate the convective forces present inside the reactor, a pair of nozzles was used to flow air downward across flowing oil. This permitted the collection of additional qualitative data regarding the stability of the flow diversion with convection occurring. Results and Discussion During the course of testing it was found that four behaviors were displayed by the silicone oil as it came into contact with the purge stream. These behaviors were: 1. 2. 3. 4.
Passing over the stream with little or no diversion or disturbance to the fluid flow. Breaking into droplets upon contact with the purge stream, causing line of sight interference. Diverting around the purge stream in an unstable flow pattern. Diverting around the purge stream while maintaining a stable flow pattern.
Figure (11) shows a photograph of 30,000 cSt fluid being diverted around the sight port by purge device (7) with a purge flow rate of 3.6 pounds per hour.
ridge at the point where the purge stream impacts the fluid is similar to the behavior predicted by the computational model.
For each purge gas, gas delivery device and oil viscosity, the regimes of purge flow rates for each of these behaviors was noted. As the 30,000 cSt fluid is the most relevant to slag behavior studies, the results for each nozzle and the 30,000 cSt fluid are shown below in Table (2). The results for the other fluids are not given in this paper, but follow a very similar pattern. In the table, the red squares indicate that for the corresponding mass flow rate and purge device, the fluid did not diverge, as per the first behavior regime. An orange square indicates that the settings produced an unstable diversion with regular observed line of sight interference. Yellow squares indicate that for the corresponding purge gas flow rate and nozzle combination, the diversion was sufficient, but the flow was unstable and could possibly have intermittent blockage of the line of sight. These two colors cover the second and third behavior regimes listed above. Lastly, the green squares indicate that the fluid diversion was stable and sufficient for a clear line of sight; which behavior falls under the fourth regime. The flows were exposed to a parallel gas flow to simulate the flow of the reacting gases past the slag. The flow rate was about 0.5 feet per second, which is similar to that calculated to exist in the actual reactor. This parallel flow often resulted in stabilizing the flow patterns in the fluid flow, though it did sometimes cause instability with some of the devices. Flows that reacted well to the parallel gas flow are marked with an asterisk in the table. Over the course of the experiments, it was found that aiming the purge stream for the apex of the circular sight port gave the best results for diversion. The table shows that device (11) worked very well at low flow rates and maintained a stable diversion even when exposed to the parallel gas flow. Devices (2), (6) and (8) also performed well, but device (11) provided the widest stable range of diversion.
Figure 11: 30,000 cSt fluid being diverted around the sight port using purge device (7). The purge flow rate is 3.6 pounds per hour. Notice the
7
Table 2: Data for the 30,000 cSt fluid. Again, device design #11 gave the best results.
Purge Flow Rate [lb/hr] 0.0 0.6 1.2 1.8 2.4 3.0 3.6 4.2 4.8 5.4 6.0 6.6 7.2 7.8 8.4 9.0 9.6 10.2 10.8 11.4 12.0 12.6 13.2 13.8 14.4 15.0 15.6 16.2 16.8 17.4 18.0
Purge Device 1
2
3
* * * *
4
* * * *
5
6
7
* * * * * * *
* * *
* * * * * * * * * * *
* * * * * * * * * * * * * * * * * * * * *
* * = Most stable at these flow rates when simulating convection
8
8
9
* * * * * *
10
11
* * * * * *
Water has a viscosity of 1 cSt and a relatively high surface tension of 72.0 dyne/cm. The results of water in the experimental system were compared with those of 1 cSt silicone oil, which has a surface tension of 17.2 dyne/cm. For these tests, the fluid film with a flow rate of 5.8 liters per minute was purged with dry air using device (11). Divergence with water was achieved with a purge rate of 2.23.4 lbs/hr while the oil showed divergence with a purge rate of 1.9-2.3 lbs/hr. The higher surface tension of the water held the fluid together at higher purge rates than the oil. The oil would form a mist at lower purge rates than the water. However, the water also required greater purge rates to diverge, also due to the higher surface tension. Most slag has a surface tension near 350 dyne/cm [9], it is expected that the purge rates required to divert the slag flow will be much greater than that required for oil of a similar viscosity.
CONCLUSIONS AND FUTURE WORK From the work presented here, it appears that it is possible to divert slag flow around a sight port in an entrained flow gasifier. These results will be used to design a purge system for use in a real gasification environment. The model developed for this paper was able to give good predictions for fluid behavior. The flow predications of the model were consistent with those found with the physical experiments. However, this method also has several weaknesses. Complex geometry was difficult to implement and including a purge flow perpendicular to the 2D computational field seems impossible to accurately implement. The model also currently lacks any sort of temperature prediction, which is very important in predicting slag properties. A three dimensional approach with an energy balance will need to be developed. The physical experiments provided a better understanding of how slag in a gasifier will behave when contacted with a purge stream. Diversion of the fluid depends on the mass of the purge gas use, not the volumetric flow rate. Also, the higher viscosity fluids require higher purge flow rates to be diverted, but are also more stable when subjected to parallel gas flows. Finally, fluids with higher surface tension require greater purge flow rates to get a good diversion.
[3] Seggiani, M. (1998). Modelling and Simulation of Time Varying Slag Flow in Prenflo Entrianed-flow Gasifier. Fuel , 1611-1621. [4] Bird, R. B., Stewart, W. E., & Lightfoot, E. N. (2002). Transport Phenomena. Hoboken: John Wiley & Sons. [5] Tannehill, J. C., et al., (1997). Computational Fluid Mechanics and Heat Transfer. Philidelphia: Taylor & Francis. [6] Ferziger, J. H., & Perić, M. (2002). Computational Methods for Fluid Dynamics. New York: Springer. [7] Reid, W. T., & Cohen, P. (1944). Factors Affecting the Thickness of Coal-Ash Slag on Furnace-Wall Tubes. Transactions of the ASME , 685-690. [8] Song, W., Tang, L., Zhu, X., Wu, Y., Zhu, Z., & Koyama, S. (2009). Flow Properties and Rheology of Slag from Coal Gasification. Fuel . [9] Hanao, M., et al., (2007). Evalutation of Surface Tension of Molten Slag in Multi-Component Systems. ISIJ International , 935-939. ACKNOWLEDGEMENTS This work was supported by a U.S. Department of Energy Cooperative Agreement via Stanford University, DEFE0001180. The author also wishes to thank Dave Wagner of the University of Utah Industrial Combustion and Gasification Research Facility, for his help with this project.
REFERENCES [1] Jeffries, J., et al., (2009). Tunable Laser Diode Absorption Temperature Measurements in a Fluidized Bed Gasifier. Pittsburgh Coal Conference Proceedings. Pittsburgh: PCC. [2] Bockelie, M., et al. (2004). A Computational Workbench Environment for Virtual Power Plant Simulation, Appendix F: Entrained Flow Gasifier - Slagging Wall Model. Salt Lake City: Reaction Engineering International.
9
INFLUENCE OF GASIFICATION CONDITIONS ON THE PROPERTIES OF FLY ASH IN A BENCH‐SCALE OPPOSED MULTI‐BURNER GASIFIER Qinghua Guo, Guangsuo Yu, Fuchen Wang, Zhenghua Dai (Key Laboratory of Coal Gasification, Ministry of Education, East China University of Science and Technology, Shanghai, 200237, China) Abstract: Entrained‐flow coal gasification offers a high efficiency and low pollutant emission way in coal’s utilization. Opposed Multi‐Burner (OMB) entrained‐flow coal gasification technology is now widely used. During the coal gasification process, fly ash and slag are the main solid by‐products. In this study, experimental work has been carried out to characterize the fly ash particles which were generated in the bench‐scale OMB gasifier. The composition of fly ash was determined by SEM and EDS. The particles size distribution was measured by Malvern particle size analyzer. The influencing factors of the fly ash particles properties with coal water slurry gasification, such as gasifier operation temperature and O/C ratio, were considered during the gasification experiments. The results show that the shape of the ash particles has irregular shape, and C in the fly ash is the major content element. O/C ratio and operation temperature affect the shape of particles and the particle size distribution significantly. At a fixed operation condition, the concentration of unburned carbon decreases along the gasifier. At various O/C ratios, fly ash particle size distribution had a bimodal distribution. The investigation on the characteristics of fly ash will be beneficial to the operation and optimization of the OMB entrained‐flow gasifier. Keywords: fly ash; gasification; particle size distribution
1. Introduction Coal gasification technology, especially high pressure and large scale entrained flow coal gasification, is the leading and key technology of coal cleaning utilization. A new type gasifier named as opposed multi-burner (OMB) gasifier has been widely used in China[1]. At the present time, the largest coal handling capacity of OMB commercial gasifier is 2000t/d, the design pressure involve both 4.0MPa and 6.5MPa. Fly ash and slag are the main by-products during the coal gasification process[2]. Ahmaruzzaman[3] reviewed the utilization of fly ash in construction, removal of organic compounds, heavy metals, dyes, and zeolite synthesis which can help a great deal in the reduction of environmental pollution. Fly ash is easy enriched with toxic heavy metal (i.e. Pb and Cd) due to its large specific surface area and subsequently damages to human health and pollutes the environment[4]. Many researchers have been focused on the properties of particulate matter and its formation mechanism during coal gasification and combustion [5-7].Wu Tao et al.[8] investigated particulate matter characteristics and formation mechanism during coal gasification in an entrained-flow gasifier. Hiromi Shirai et al.[9] studied particulate matter properties during pulverized coal combustion. Stanislav et al.[10] summarized the methods that used for characterization of the phase, mineral and chemical composition of fly ash from coal-fired power stations. However, the literature about fly ash particular matter properties at different operation conditions during coal water slurry gasification is rare. In this paper, the effect factors of operation temperature and oxygen carbon atom ratio on properties of fly ash particulate matter have been analyzed in a laboratory scale OMB gasifier by using Scanning Electron Microscopy (SEM), Energy Dispersive x-ray Spectrum (EDS) and Malvern Particle Size Analyzer.
2. Experimental work 2.1 Experimental setup A schematic diagram of the laboratory-scale opposed multi-burner entrained-flow gasifier is shown in Fig. 1. The gasifier that made by stainless-steel includes gasification chamber and quench chamber two parts. The inner diameter and length of the gasification chamber composed of a 15mm thick cast refractory wall are 300 and 2200mm, respectively. The refractory wall, wrapped with 235mm fiber blanket to reduce the heat transfer. The stainless-steel column shell of gasification chamber is 800mm in diameter and 2500mm in height, and the quench chamber is 500mm in diameter and 2000mm in height. Oxygen and coal water slurry (CWS) were fed in the gasifier through four symmetric opposed dual-channel burners. Oxygen is fed into the outer channel of burners from oxygen cylinder and measured by gas mass flowmeters. Diesel oil and CWS are fed into the inner channel of burners by the gear pumps and single-screw pumps respectively. High relative velocities between oxygen and fuel in the central region of gaisifer provide good conditions for active diffusion and convection at the droplet surface, which together with the high temperature results in fast combustion and gasification reactions. Then, the raw syngas flows down to the quench chamber, where it is cooled and partially scrubbed by water, and finally flows to the downstream processes. Nitrogen is fed from nitrogen cylinder to clean the inner channel of burners at the end of the experiment.
Fig.1
Schematic diagram of the experimental setup
The analysis of the coal sample which is used in the experiments is show in Table 1. The coal ash fusion temperature is 1255℃. Coal was milled into various particle sizes, with water and additive added, the concentration of CWS is about 61% during the CWS pulping process. Before feeding the coal water slurry, the gasifier was preheated to the reaction temperature (>1300℃) with diesel oil fuel. The main experimental conditions in this study were to control the oxygen carbon atom ratio at 0.9, 1.0, 1.1. Each burner’s CWS flow rate varified from 9kg/h to 12 kg/h, oxygen flow rate from 4Nm3/h to 5.6Nm3/h.
Table 1 Proximate analysis M A V FC
4.04 8.38 32.56 55.02
Analysis of experimental coal sample (mass%) Ultimate analysis C H O N S
71.66 3.87 10.49 0.94 0.80
Ash analysis SiO2 CaO SO3 Al2O3 Fe2O3 MgO
34.56 20.16 16.77 14.40 8.69 1.87
K2O Na2O TiO2 SrO MnO BaO
1.15 0.70 0.68 0.45 0.23 0.11
2.2 Fly ash collection and analysis Fly ash sampling system included a high-temperature water jacket sampling tube, a filter, a flow meter and a pump. Sampling places were located at the thermocouple ports along the gasifier. In this work, the main sampling position was 3# which was closed to the slag outlet of the gasification chamber. Particle size distribution was analyzed by using Malvern Particle size analyzer. Fly ash samples were pretreated in an ultrasonic cleaning machine in order to make particles dispersed uniformly in the bottle, after formed a uniform suspension solution, then took part of it to analysis using Malvern Particle Size analyzer. Morphology and element distribution were analyzed by SEM and EDS respectively. Suspension solution was firstly filtered and then dried for two hours under condition of 105°C before SEM and EDS analysis.
3. Results and discussion 3.1 Effect of operation temperatures 3.1.1
SEM Images
SEM images of fly ash at 1200℃(below ash fusion temperature) and 1350℃(above ash fusion temperature) are presented in Fig. 2 and highlight important structural and morphological features. Ash generated at high temperature was found to be much finer than ash generated at low temperature due to the differences in porous structure. It can be observed that the fly ash has porous structure both in 1200℃ and 1350℃. Temperatures affect the morphologies of particles significantly, the number of fine particles in 1200℃ less than that in 1350℃. Furthermore, more volatile element vaporized at higher temperature, and then condensed to form more spherical particles. On the contrary, there were less spherical particles at the operation temperature of 1200℃.
(a) 1200℃ (b) 1350℃ Fig. 2 SEM results at different temperatures
Surface element distribution at different temperatures was illustrated in the photos. It can be seen that the content of S, Fe and Na at the temperature of 1350℃ were significantly higher than that at 1200℃. Volatile element such as Na and S were more significantly vaporized, and there was a reducing atmosphere in the gasifier, iron was reduced to metal iron which is more volatility, which will result in a higher content of these elements on the surface of particles.
3.1.2 Particle size distribution The fly ash particle size distribution at different temperatures was shown in Fig. 3. Comparison with the experimental coal particles, the particles size became smaller when they reacted in the gasifier. The peak of the particle size distribution shift to left position significantly and the fraction of fine particles increased with the temperature increasing. Temperatures not only affect the ratio of particle size but also the peak of particle size distribution. The research has shown that temperature affect fine particle formation more significantly under a condition of higher oxygen content. Temperature played an important role in the fine particles formation process. More fine particles were formed at higher temperature than that formed at lower temperature. The results agreed with the previous work well. 5
Vol. %
4
1350℃ 1200℃ Experimental CWS
3 2 1 0
0.1
1
10
100
1000
Paticle size /mm
Fig. 3
Particle size distribution at different temperatures
3.2 Effect of oxygen carbon ratio 3.2.1 SEM Images Morphology of fly ash at different oxygen carbon ratio was shown in Fig. 4. When the oxygen carbon ratio is 0.9, coarse particles were mainly irregular and with less spherical shape. When the oxygen carbon ratio increased to 1.0, the number of spherical particles increased relatively and most fine spherical particle exist as single. For the case of oxygen carbon ratio is 1.1, fine particles mainly exist as agglomerate status. Conversation of coal particles increased as oxygen content increased, at the same time, diameter of particles decreased and content of vaporized mineral increased, subsequently formed more spherical particles. So we can conclude that with the oxygen carbon ratio increased, the diameter of coal particles decreased, resulting higher possibility for included mineral coalescence. And finally leading to more fine particles aggregated.
3000
3000 C 2500
2500
2000
2000
3000
C 2500
1500 1000 500
O
Al
Si
Ca 0 0.0
1.5
1500 1000 500
S 3.0 4.5 Energy (keV)
Fe 6.0
7.5
9.0
Intensity (a.u.)
Intensity (a.u.)
Intensity (a.u.)
C
0 0.0
O
Si Al S 1.5
2000 1500 1000 500
Ca 3.0
Fe
4.5 6.0 Energy (keV)
7.5
9.0
0 0.0
O
Al
Si
S Ca
1.5
3.0
Fe
4.5 6.0 Energy (keV)
7.5
9.0
(a) O/C=0.9 (b) O/C=1.0 (c) O/C=1.1 Fig. 4 SEM and EDS results at different oxygen carbon ratios The EDS results of fly ash surface element distribution at different oxygen carbon ratio were also shown in Fig.4. It can be seen that, carbon is the dominant element in the gasification fly ash. Oxygen carbon ratio was significantly affected particle carbon content and slightly affected content of elements such as S, Fe, Al and Si. Carbon content in the fly ash decreased gradually with oxygen carbon ratio increased. At the oxygen carbon ratio increased from 0.9 to 1.1, the carbon content decreased from 72.2% to 59.6%. The carbon content at various sampling ports had been tested at a fixed oxygen carbon ratio (O/C=1.0). The results show that, fly ash carbon content in position 1# is the highest 78.6%, positions 2# and 3# carbon content are 50.3% and 48.8% respectively. The concentration of unburned carbon decreases along the gasifier because of the residence time of fly ash increased with the high temperature reaction when it flows down to the slag outlet.
3.2.2 Particle size distribution Particle size distribution at different oxygen carbon ratio was shown in Fig. 5. The results indicate that, the fly ash had a bimodal distribution having two peaks. However, the curves of particle size distribution were obviously difference with various oxygen carbon ratios. More fine particles formed under the oxygen carbon ratio of 1.0 (more particles were ranged in size of 10-100µm, less particles were larger than 100µm), and particles formed under the oxygen carbon ratio of 1.1 were mainly coarse particles. For the oxygen carbon ratio of 0.9, particle size distribution was between the two other cases of distribution. It could be interpreted as low carbon conversion and more residual carbon in particles when the oxygen carbon ratio was 0.9. For the case oxygen carbon ratio is 1.1, besides the highest carbon conversion, many minerals were vaporized and condensed to formed fine particle aggregation. Furthermore, included mineral was more likely to coalescence under this condition.
5 4
O/C=0.9 O/C=1.0 O/C=1.1
Vol. /%
3 2 1 0 0.01
0.1
1
10
100
1000
Particle size /μm
Fig. 5
Particle size distribution at different oxygen carbon ratios
4. Conclusion Effect of temperature and oxygen carbon ratio on fly ash properties in a laboratory-scale OMB coal water slurry gasifier was investigated. The main conclusions are as follows: Fly ash particles in the gasifier performed irregular shape and carbon is the main element. Operation temperatures affect the morphology, surface element distribution and size distribution of fly ash particles. Higher temperature is beneficial to form spherical particles and increase content of element such as S, Fe, Al etc. Oxygen carbon ratio mainly affected carbon content of fly ash, carbon content decreased as the oxygen carbon ratio increased. At a fixed oxygen carbon ratio, the concentration of unburned carbon decreases along the gasifier. Fly ash particles had a bimodal distribution. The peaks of fine particles were almost coincident, and the peaks of coarse particles changed with various oxygen carbon ratios.
Acknowledgement This work is financially supported by National Key State Basic Research Development Program of China (973 Program, 2010CB227006), National Nature Science Foundation of China (20876048).
References [1] Fuchen Wang. Progress on the Large-scale and High-efficiency Entrained Flow Coal Gasification Technology[J]. China Basic Science, 2008, 10(3): 4-13. [2] R. H. Matjie, C. Van Alphen, Mineralogical features of size and density fraction in Sasol coal gasification ash, South Africa and potential by-products[J]. Fuel, 2008, 87(10):1439-1445. [3] M. Ahmaruzzaman. A review on the utilization of fly ash[J]. Progress in Energy and combustion Science. 2010, 36(3): 327-363. [4] Chow J C, Watson J G. Review of PM2.5 and PM10 apportionment for fossil fuel combustion and other sources by the chemical mass balance receptor model[J]. Journal of Energy and Fuels, 2002, 16(2): 222-260.
[5] M. Aineto, A. Acosta, J. Ma. Rincón, M. Romero. Thermal expansion of slag and fly ash from coal gasification in IGCC power plant[J]. Fuel, 2006, 85(16): 2352-2358. [6] L. Bartoňová, Z. Klika, D.A. Spears. Characterization of unburned carbon from ash after bituminous coal and lignite combustion in CFBs[J]. Fuel, 2007, 86(3): 455-463. [7] N.J. Wagner, R. H. Matjie, J.H. Slaghuis, et al. Characterization of unburned carbon present in coarse gasification ash[J]. Fuel, 2008, 87(5): 683-691. [8] T. Wu, M. Gong, E. Lester, et al. Characterisation of residual carbon from entrained-flow coal water slurry gasifiers[J]. Fuel, 2007, 86(7): 972-982. [9] Hiromi Shirai, Hirofumi Tsuji, Michitaka Ikeda, et al. Influence of combustion conditions and coal properties on physical properties of fly ash generated from pulverized coal combustion[J]. Energy and Fuel, 2009, 23(7): 3406-3411. [10] Stanislav V. Vassilev, Christina G. Vassileva. Methods for Characterization of Composition of Fly Ashes from Coal-Fired Power Stations: A Critical Overview[J]. Energy & Fuels, 2005, 19(3): 1084-1098.
GTI’S SAMPLING AND ANALYSIS SYSTEMS FOR GAS STREAMS OF GASIFICATION AND DOWNSTREAM PROCESSES Osman Mehmet Akpolat, Gas Technology Institute Des Plaines, IL, U.S.A. Tanya S. Tickel, Gas Technology Institute, U.S.A. Rachid B. Slimane, Gas Technology Institute, U.S.A. Chun W. Choi, Gas Technology Institute, U.S.A.
Abstract: Since early 2004, GTI has been operating its state-of-the-art pilot-scale gasification facility, the Henry R. Linden Flex Fuel Testing Facility (FFTF), to investigate the renewed interest in various gasification technologies. The use of FFTF in the recent years has been a key point in GTI’s focus on the development and commercialization of various programs ranging from research in evaluation of selected coal and biomass feedstock gasification characteristics, to programs where performance evaluation of downstream syngas processing units take the front stage. Syngas analysis of different constituents is important at different process locations. During these research programs, GTI has developed and implemented innovative analytical approaches to meet the challenging task of characterizing the composition of process streams. In addition to measuring major gas species, techniques are applied to analyze select contaminant species both upstream of cleaning, at higher concentrations and downstream of cleaning at trace levels. These systems ensure precise and accurate data that provide for the performance evaluation of the complete process. With every completed project on the FFTF, GTI has optimized these systems further and continues to develop flexible, scalable and reproducible analytical solutions for sampling and analysis for gas streams of gasification and downstream processes. In this paper, these systems are described to illustrate the effectiveness of these innovative systems. The main objective of the discussion is to illustrate the various unique analysis equipment capability of GTI. During the development process many analysis instruments, such as gas chromatographs (GC), mass spectrometers (MS and GC/MS), Fourier transform infrared spectrometers (FTIR) and multiple wet chemistry methods, that are commonly found in much less challenging laboratory environments were adapted to be used in a pilot plant environment. Multiple disciplines of chemistry, chemical engineering and mechanical engineering were merged together to interface with some of the most challenging sampling locations to condition and then deliver representative samples to delicate analysis equipment that were expected to operate almost autonomously. This paper shows how GTI has developed its analytical solutions capability and how such innovative systems can add significant value to any research that is on the pilot scale where gas composition analysis is critical.
Introduction: Over the last 6 years, GTI has been operating FFTF for various clients to evaluate the performance of different gasification technologies. The common challenge between all of these projects has been collecting and reporting
precise and accurate data. The collected data must be representative of the process conditions in order for the evaluation of the process to be meaningful. Over the years GTI has developed and optimized sampling and analysis solutions to successfully combine some of the most challenging sampling conditions in gasification processes, to some of the most delicate analytical equipment usually reserved for chemical laboratory use. These systems can be separated to three parts: • Sample probe • Sample conditioning • Sample analysis
PROCESS
ANALYZER
Process
Sample Probe
Sample Conditioning
Sample Analysis
Figure 1. Syngas Sampling System Components
This paper will focus on only the sample analysis portion, and will give detailed descriptions of the equipment used and how they are setup. Further information on sample probes and conditioning can be obtained by contacting coauthors Tanya Tickel and Rachid Slimane. The methods developed by GTI will be discussed grouped by compounds of interest. Each group will describe the equipment used and modification & customization performed to operate them almost autonomously. Almost all of the analytical equipment used in the FFTF has been designed for chemical laboratory use. GTI has successfully integrated these delicate analytical equipments to process environment and has developed its analytical solutions capability over many years of testing at pilot plant conditions. This has produced precise and accurate data that has been very valuable to pilot scale research where gas composition analysis is critical.
Methods: Major Syngas Species The first group of compounds of interest is the major syngas species. This group contains H2, CO, CO2, CH4, N2, H2S and COS. During initial stages of operation, O2 is also analyzed to monitor the combustion process; however it is not considered one of the major species analyzed during gasification. The concentrations of these compounds can change from a few percents by volume to 50% (v) depending on the process and the stages of the process.
Over the years GTI has used multiple Varian CP-4900 Micro Gas Chromatographs with thermal conductivity detectors (TCD) to measure the major species. These Micro GC’s were very efficient in analyzing the gas streams and produced very accurate and precise data when compared to spot samples taken and analyzed by GTI’s Chemical Research Services department. The typical detection limit for compounds of interest was about 10ppmv to 20ppmv depending on the compound of interest. The micro GC’s were configured with dual channels. Each channel is configured to analyze a portion of the compounds of interest. This resulted in a typical micro GC analysis of less than two minutes. The micro GC’s had dual carrier gas capability and used He and Ar as carrier gases. Channel one used Ar carrier gas and analyzed for H2, O2, N2, CH4 and CO. This channel used a Molsieve 5A column with a heated and back flushed injector. The back flush option allowed the unwanted compounds or compounds that would elute too late to be flushed out of the pre column and not injected to the analytical column. Restriction
System Pressure
Pre-column
Analytical Column
Detector
Pressure Regulator Pressure Point
Injector
Backflush Vent
Figure 2. Varian Micro GC Back Flush Option Flow Diagram Channel two used He carrier gas and analyzed for CH4, CO2, H2S, COS. It used a PorapakQ column and it also had a heated and back flushed injector. In this channel it was possible to analyze acetylene, ethylene and ethane but calibration for these compounds were not done. During syngas sampling the concentrations of these compounds were observed to be below detection limit and they were not reported in the major gas species group.
Figure 3. Varian CP-4900 Micro GC During regular testing, micro GCs at multiple sampling locations would run continuously to determine syngas gas composition. This online measurement would produce about 700 samples a day that was recorded and charted using an in house developed Excel macro package. The operator would start sampling and then the data was logged on an Excel spreadsheet and charted during the pilot plant testing.
The micro GC method and calibration was tested before and after each run. Four calibration gas mixtures covering the concentration range for each major gas species reported were used. Also on channel two, He was injected to record a “blank baseline”. This was later subtracted from each analysis on that channel to flatten the baseline and have more accurate integration of the compound peaks. During long term pauses in operation an intermediate calibaration check would be done. Later this was omitted since the calibration check before and after each pilot plant test showed very little change in the response factors. Using Molsieve column in the micro GC made that channel very susceptible to moisture. The accumulated water would reduce the effective length of the column by occupying the active sites on the column. This resulted in retention time shifting, and caused peaks to disappear from the charts. Since the peak would shift outside the retention time window, it would not be identified and therefore the Excel macros would not report the concentration. To prevent this, the micro GCs were kept at high temperature “bakeout” modes which restored the column to full efficiency. For long test periods a dual system micro GC with two of each channel was acquired. When a pair of channels was in use the second pair was in bakeout. This minimized the time without any data due to column regeneration. In addition to micro GCs, other equipment such as FTIR and GC-FID were used in early test campaigns to cross check results. The variations in measured syngas concentrations between reported values from different instruments were about 2%. Later due to its ease and unattended operation micro GCs were used as the primary online major species analysis instrument. Spot samples were still used to verify the gas concentrations that were reported. The spot samples were taken either daily or once every steady state period in the early pilot plant runs. Water Content of Syngas Water content of the syngas was measured by various methods. GTI’s experience during other projects has been that water is one of the most difficult compounds to measure accurately. Most available moisture analyzers do not work at gasifier sample point conditions, and conditioning gas streams to work for these analyzers usually removes most of the water. A typical syngas sample right after the gasifier can be 1600°F, 70 psig and can contain around 20% water. Temperature of the sample stream makes it a very challenging sample point. GTI’s first approach was to dilute the sample gas with an inert gas. When this diluted gas was cooled and depressurized, any condensable compounds that were present in the sample stream would stay in gas phase without condensation. The resulting sample stream was then analyzed using an IMACC process FTIR with a specially developed method to match the syngas sample gas concentrations. This data was then logged and charted with the major gas species data from the micro GCs. A fairly extensive custom macro was developed to automate this process and chart the resulting data using Microsoft Excel. In addition to this dilution approach, a gravimetric moisture collection method was developed. The hot gas stream was sampled through water absorbent material and the dry gas exiting the absorbent was measured with a dry gas meter. Water mass collected over time was divided by the volume of sampled gas to determine the water concentration. This method proved reliable however resulted in very rapid saturation of the absorbent material due to high water content of the syngas. The average sampling period for a gravimetric water sample was around one hour. The syngas flow rate was set at around one liter per minute. During initial method development it was observed that water collected during sampling was too much for efficient use of the absorbent material. A water condensation vessel in an ice bath was installed to collect most of the water before the sample gas contacted the absorbent material. This proved to be very efficient with minimal variation in reported values. The sampling time was adjusted to achieve the desired accuracy. After multiple pilot plant runs, it was decided that only gravimetric water sampling was satisfactory to report the water content of the sample gas. The period of sampling was set to every 3-4 hours with a sample collection time of around one hour.
In addition to these methods, another process water analyzer was installed in process line exiting the gasifier and tested. This was a moisture analyzer commonly used at boiler stack moisture measurement. This instrument was found to be quite hard to implement in the syngas stream right after the gasifier due to the location’s harsh environment. The probe for the analyzer was hard to keep clean. Dust and particulates accumulated on and around the probe preventing proper sampling. Heavy hydrocarbons & light tars Initial studies in heavy hydrocarbon and light tar analysis started with projects that were performed to characterize the syngas composition during biomass gasification. Initially a standard HP 5890 GC with a flame ionization detector (FID) was used. A non polar DB1 analytical column was used to monitor benzene, toluene, ethyl benzene, o,m,p-xylenes (BTEX) and naphthalene. Sample conditioning systems were developed to transfer sample gas from high temperature sample probes in the process lines to the analyzer in the analytical area. The GC was customized in GTI to inject directly to column. The HP 5890 was integrated to Varian chromatography software and data collected was merged with major gas species results using the GTI custom macros. The results were reported in Microsoft Excel.
Figure 4. HP 5890 GC with Flame Ionization Detector After successfully monitoring BTEX and naphthalene, a detailed characterization of biomass syngas was done with an Agilent 6890 GC and an Agilent 5973 Mass Selective Detector (MSD). This analyzer was assembled by Wasson Ece and used a proprietary Wasson 2318 analytical column (KC1 phase was similar to DB1) to separate 23 compounds. The analyzer was calibrated at the vendor. The syngas sample was delivered to the analyzer using GTI’s sample conditioning systems and injected at 200°C into the GC using a 1 ml sample loop and a 6 port injection valve. The detection limit was 0.01 ppmv.
Figure 5. Agilent 6890 GC with Mass Selective Detector The Agilent 6890 GC used Agilent Chemstation MSD software to collect data. Each run was about 1 hour and the resulting analysis was integrated to the major gas species using the custom GTI macros. The monitored compounds were: Table 1. Compounds Analyzed by Agilent GC-MSD 1. Methanethiol 2. Carbon Disulfide 3. Ethanethiol 4. Benzene 5. Thiophene 6. Toluene 7. Ethylbenzene 8. p-Xylene 9. m-Xylene 10. o-Xylene 11. Naphthalene 12. Benzothiophene 13. 2-Methylnaphthalene 14. 1-Methylnaphthalene 15. o-Cresol 16. Phenol 17. p-Cresol 18. m-Cresol 19. Acenaphthene 20. Acenaphthylene 21. Fluorene 22. Phenanthrene 23. Anthracene The same GC was also used with a flame ionization detector (FID) to measure light hydrocarbons. Using 1 ml sample injection and a set of Wasson Ece 3018+2378 analytical columns (KC153 and KC5 phases) the following light hydrocarbons were analyzed: Table 2. Compounds Analyzed by Agilent GC-FID 1. Methane 2. Ethane 3. Ethylene 4. Propane
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
Propylene Acetylene Isobutane Propadiene n-Butane t-2-Butene 1-Butene Isobutylene c-2-Butene Neopentane Isopentane Methyl Acetylene
After initial characterization of heavy hydrocarbons and light tars, it was observed that about 95% of these compounds were single or double aromatic ring compounds. For later pilot plant testing a simpler to operate analyzer was investigated. It was decided that an Agilent 7890 GC with dual FID will be able to sample two streams and analyze them at the same time. This would allow for online comparison of hydrocarbons before and after a process point. The syngas sample was conditioned and delivered to the analyzer at 200°C and injected using 1 mL sample loops and Agilent sampling valves. The sampling valves were connected to Agilent GS-GasPro columns. The analyzer was calibrated to monitor benzene, toluene and naphthalene. It was used in multiple biomass gasification campaigns and provided online benzene, toluene and naphthalene concentrations.
Figure 6. Agilent 7890 GC with Dual Flame Ionization Detectors In addition to online analysis equipment, GTI has used impingers to collect and dissolve hydrocarbons in syngas samples. The solutions that the compounds are captured were analyzed using GC-FID by GTI Chemical Research Services after the test campaign. This would generate the largest number of compounds and has the lowest detection limits. A typical syngas sample capture would require one to two hours of sample flow. This would flow enough sample gas to satisfy the detection limits of the laboratory instruments. This period was modified when capturing samples with suspected high concentrations of hydrocarbons to prevent overloading the laboratory equipment. Low concentration sulfur & other contaminants One of the earlier projects for FFTF involved H2S removal to trace levels in the syngas. To evaluate the performance of this process, a highly specialized and sensitive sulfur analyzer was ordered. This Multi Purpose Sulfur Analyzer (MPSA) was custom made by Varian using a Varian 3800 GC and a Varian Pulsed Flame Photometric Detector. The standard detection limit of PFPD was further improved by using a proprietary cold trap. The MPSA used
multiple sample loops depending on the sample concentration. The operator would select the proper method to run and the MPSA would select proper valve positions for the transfer of that sample loop to the cold trap. Then the cold trap would be blocked and heated very rapidly evaporating all the sample that had been collected. Finally this sample would be injected into a Varian CP-sil5 non polar column and analyzed by PFPD.
Figure 7. Varian 3800 GC based Multi Purpose Sulfur Analyzer This analyzer was extremely sensitive. When the largest sample loop was used the detection limit was about 1ppt H2S. At the same time using the smallest loop samples containing over 1000ppm H2S were analyzed without reconfiguration. However the operational requirements for the analyzer were high and it took about one hour for one sample to be captured and analyzed. Varian MPSA delivered trace level H2S measurements during pilot plant testing. But later it was replaced by an Agilent 6890 GC with a Sievers 355 Sulfur Chemiluminescence Detector (SCD). During most of the pilot plant testing, this GC-SCD was used as the sulfur containing compound analyzer. Both coal and biomass gasification testing used this instrument to provide online data at 30 minute intervals. The GC used a 1 ml sample injection loop and a Wasson Ece 2332 analytical column (KC68 phase) to identify 15 sulfur compounds. Standard detection limits were at 0.005 ppmv. Agilent Chemstation software was used to integrate the results with the major gas species using custom macros developed in GTI. Final data was represented as Microsoft Excel charts. Table 3. Compounds Analyzed by Agilent GC-SCD 1. Hydrogen Sulfide 2. Carbonyl Sulfide 3. Methanethiol 4. Ethanethiol 5. Dimethyl Sulfide 6. Carbon Sulfide 7. 2-Proanethiol 8. 2-Methyl-2-Propanethiol 9. 1-Propanethiol 10. Thiophene 11. Diethyl Sulfide 12. 1-Butanethiol 13. Diethyl Disulfide 14. Benzenethiol 15. Benzothiophene
During some pilot plant testing the H2S and COS concentrations were high enough to use micro GCs instead of these specialized sulfur analyzers. The detection limit when a micro GC was used increased to about 20 ppmv due to the limitations of the micro GC. Miscellaneous In addition to process monitoring at gasifier or downstream locations to generate project reports, GTI has used different analyzers for process monitoring only. One of these analyzers was a Rosemount NDIR that monitored H2, CO, CH4 and O2. This analyzer was mainly used to monitor the change from combustion to gasification during the initial startup stages. The sample conditioning systems were the same as the rest of the analyzers. The readout from this analyzer was connected to the plant control system to give operators and indication of some of the major species during the test campaign. Another such analyzer was a Honeywell lead acetate tape H2S analyzer. This was located between two sulfur guard beds in the pilot plant. These guard beds were in series and the analyzer sampled after the first guard bed. The purpose of this analyzer was to determine if the first sulfur guard bed was used up. When the first guard bed was used up, the operators would make a decision to stop or continue the test based on test plan, predicted life of the second bed etc.
Discussion & Conclusion: Henry R. Linden Flex Fuel Testing Facility has been a state-of-the-art pilot scale gasification facility for GTI to investigate the renewed interest in various gasification technologies. Since 2004, with the addition of FFTF to GTI facilities, GTI has been focusing on the development and commercialization of various programs ranging from research in evaluation of selected coal and biomass feedstock gasification characteristics, to programs where performance evaluation of downstream syngas processing units take the front stage. During these research programs, GTI has developed and implemented innovative analytical approaches to meet the challenging task of characterizing the composition of process streams. In this paper, we have tried to describe these systems in an effort to illustrate the various unique analysis equipment capability of GTI. Representative syngas analysis at multiple process locations is challenging and at the same time necessary to evaluate the process properly. GTI’s analytical systems ensure precise and accurate data that provide for the performance evaluation of the complete process. With every completed project on the FFTF, GTI has optimized these systems further and continues to develop flexible, scalable and reproducible analytical solutions for sampling and analysis for gas streams of gasification and downstream processes. This paper shows how GTI has developed its analytical solutions capability over the several years at FFTF pilot scale testing facility and how real world experience modified and focused these systems to be more efficient with each completed test campaign. Multiple iterations of sample conditioning systems have been developed and modified over the different projects to improve and increase efficiency and design. Together with the developments in sample conditioning many instruments that very originally designed to be operated in much less challenging laboratory environments were adapted to be used in a pilot plant environment. Some were modified and kept such as the micro GCs and some more specialized equipment now wait for new projects where they can add significant value to project deliverables. In addition to measuring major gas species, different techniques were applied to analyze select contaminant species both upstream of cleaning at higher concentrations and downstream of cleaning at trace levels. Operating FFTF through multiple projects, GTI has been able to understand how to sample and analyze gas compositions in pilot plant gasifier operations and was able to use highly specialized equipment to determine the most dominant areas to focus during plant operations.
As pilot scale testing continues at FFTF, GTI has been and will be investigating and evaluating new process monitoring equipment that might replace the modified chemical laboratory analyzers that are currently in use with highly sensitive process analyzers that are being developed now.
Acknowledgement: The author would like to acknowledge co-authors Tanya S. Tickel, Rachid B. Slimane and Chun W. Choi and the GTI FFTF Analytical Team for their hard work and contributions.
Manuscript Not AVAILABLE
DESIGN OF COMMINUTION UNIT FOR THE GASIFICATION PILOT PLANT
N. ACARKAN, G. ONAL, A. A. SIRKECI, G. ATESOK [email protected]
ABSTRACT
The comminution unit for the gasification pilot plant to be constructed for Turkish Coal Enterprises has been designed at the Department of Minerals Processing of Istanbul Technical University.
It has been designed that the comminution unit will be fed with coal below 100 mm and grind it under 100 m whose moisture content should be 1% maximum prior to be fed to the gasification unit. The capacity has been chosen to be 1800 kg/h. The run of mine coal should have 18% moisture, 25% ash and 38% volatile matter at maximum.
The primary crushing in the unit will be accomplished using a hammer crusher below -20 mm. Followed by crushing the moisture of the product will be decreased less than 10% from about 18% of initial value. The crushed and partially dried product will be stored in a silo before feeding to a vertical mill. Coal will be ground below 100 µm in the vertical mill while dropping the moisture content under 1%.
The project for a commiution process has been designed according to the information given above.
27th PCC, Oct. 11-14, 2010, Istanbul, Turkey
Preparation of coal water mixture with high concentration from low rank coals and lignite by dry fine coal with optimum particle distribution Feng He1 and Baoqing Li2 1
Yulin Western Coal Technology Research Center, Shan’Xi, Shenmu,719300 China
2
State Key Lab of Coal Conversion, Institute of Coal Chemistry, CAS, Taiyuan, 030001 China
Abstract: A new technology for preparation of coal water slurry (CWS) with high concentration from low rank coals and lignite by dry fine coal with optimum particle distribution was developed. According to coal characteristics the micronized coal was achieved by pre-drying and dry grinding of coal using several modified mills to obtain various fineness coals. CWS powder, which can prepare high concentration CWS, was produced by mixing different ratio of various fineness coals in optimum condition. CWS with concentration of up to 68% of eligible slurry using low rank coals has been produced by this technology in the industrial CWS powder plant with capacity of 750,000 t/a. 63% CWS using lignite as raw material has also been produced. Compared with conventional wet process for CWS preparation, the new technology provides less power consumption of about 5 kWh/t CWS and a half amount of additives. The reduction of the relative cost is about 8 CHYuan/t CWS with obvious economic and social benefits. This new technology will greatly promote the clean conversion and utilization of low rank coals and lignite. Keywords:coal water slurry; low rank coal; lignite; coal gasification
Introduction Development and promotion of clean coal technology to China's implementation of energy saving and sustainable development are of great significance. Coal water slurry (CWS), as a new coal-based liquid fuels due to its energy saving, substitution of oil, high-efficiency combustion, clean and easy-to-pump spray, etc., has been listed as one of clean coal technology and of national encouraging and developing key industries . Gasification technology is a core technology of clean coal technologies, which is leading technology for the development of coal-based chemicals (ammonia, methanol, dimethyl ether, alkylene, etc.), coal-based liquid fuel, multi-generation systems, hydrogen and other key process. Most modern gasification technology uses entrain bed slag gasifier including CWS and dry feeding gasification. Thus, the low rank coals with low ash melting point such as Shenmu coal is needed as raw material. CWS feeding GE (Texaco) gasification and multiple nozzle CWS gasification technology developed by the East China University of Science and Technology have been widely used in domestic coal-based ammonia and methanol production. However, Shenmu coal is featured by poor slurryability of CWS (the highest concentration of 60%) due to its high inherent moisture and surface area with high content of polar functional groups. Thus, the energy consumption and costs are increased,affecting the economic benefits. Meanwhile, lignite is an abundant resource in China, but its conversion and use are a major problem because it is also characterized by poor slurryability of CWS (about 50%) and difficult to prepare CWS with high concentration. To enhance the CWS concentration using low-rank coals, many institutes, universities and
27th PCC, Oct. 11-14, 2010, Istanbul, Turkey
companies did a lot of effort over the years, but with little success. Yulin Western Coal Research Center broke the traditional wet technology of CWS preparation and developed a new technology named dry fine coal with optimum particle distribution for CWS preparation with high concentration using low rank coals and lignite. More than 68% and 63% of eligible CWS can be produced using Shenmu coal and Inner Mongolia Baorixile lignite as raw materials, respectively. This technology will improve the economic and social benefits in the coal-based methanol and chemical enterprises by reducing energy consumption and amount of additives, and greatly promote the conversion and utilization of low rank coals and lignite especially their CWS gasification and related chemical synthesis. This article describes the development of the new technology of CWS preparation. The energy consumption and related cost are compared with the conventional technology.
1. Situation of CWS Preparation CWS is normally prepared by wet technology, namely coal, water and additives are added in a ball mill to produce CWS. The process is simple and CWS with 65% slurryability can be produced using the wet technology for most of coals by adjusting the distribution of various balls in the ball mill. However, it showed poor adaptability for low rank coals and lignite. The conventional technology is not suitable for coals with high inherent moisture and less grindability, since it is difficult to increase grain fineness to achieve the optimum particle size distribution (PSD) accompanied by preparation of high concentration CWS. The Raymond Mill machine is available for dry CWS preparation. However, the design of the classifier is unreasonable. All the leaves in the classifier are in strict accordance with the radial layout, which results in the high residence for air flow when the classifier is run at high speed. Large amount of qualified powders cannot be exited, leading to the blockage and a decrease in the processing capacity of the mill. The situation is more sever for raw materials with relatively high moisture such as lignite. The lower speed of air flow makes difficulty to loose the materials and to raise the particle fineness for the optimization of PSD. Meanwhile, the existing Raymond Mill machine depends mainly on the blower for transportation of materials. It brings a problem that air flow inside the mill must keep strict balance, otherwise, it will easily lead to positive pressure, resulting in powder leakage with environmental pollution. 2. Development of CWS preparation technology with high concentration using low rank
coals and lignite by dry fine coal with optimum PSD. After the analysis of existing CWS preparation technology combined with the features of low rank coals and lignite, we modified the Raymond machine and developed a new technology named dry fine coal with optimum PSD for CWS preparation with high concentration using low rank coals and lignite. A semi-industrial plant with annual output of 200,000 tons of dry CWS product has been established in 2009, and run and operated smoothly. After reformation, the capacity of the plant has been reached to 750,000 tons of dry CWS product per year. About 20,000 tons of dry CWS product has been produced and used for preparation of 69% CWS with viscosity of 800 mPa.s (100s-1) using Shenmu coal. The schematic sheet is shown in Fig. 1.
27th PCC, Oct. 11-14, 2010, Istanbul, Turkey
Raw coal bunker Hammer 0-8mm PLC feeding system Ultra-fine mill
Mid fine mill
Mid coarse mi
Coarse mill
Bunker
Bunker
Bunker
Bunker
Elec. scale
Elec. scale
Elec. scale
Elec. scale
Dry fine product bunker 30%additive jar
CWS storage tank
CWS Mix-stiring tank CWS trasport
Fig. 1 Schematic sheet of dry technology for high concentration CWS preparation Raw coal (without pre-drying for lignite with humidity <30% in analytical base) is ground to 0-8mm. Then, coal is further grounded by several sets of the improved mills to various finenesses and yields according to the optimum PSD based on coal used and the requirement of CWS concentration. After mixing of the powders with required finenesses and yields, a preset amount of additives are sprayed to the mixed powders to produce the dry coal fines as either product for packaging and transportation or intermediate for further CWS production by adding water. Table 1 lists the analytical data of properties of five Shenmu low rank coals and one Inner Mongolia lignite, and their CWS.
Table 1
Analytical data of properties of 6 coal samples and their CWS
Item
Shenmu
Shenmu
Shenmu
Shenmu
Shenmu
Hewan
Shangyushu
Erdaomao
Yujialiang
Huojitu
Baorixile lignite
Humidity
Mad/%
8.96
9.26
10.53
10.60
11.67
28.66
Ash
Aad/%
6.76
8.08
3.62
4.32
3.67
6.35
Volatile
Vad/%
31.35
31.95
33.62
25.91
34.50
35.60
S
St/%
0.16
0.19
0.21
0.17
0.44
0.42
5998
5968
6134
5994
5933
3926
6734
6715
6892
6765
6714
4778
Heating value
Qnet,ad /Kcal.Kg-1 Qgr,ad
Ash melting points
ST/℃
1251
1384
1237
1126
1255
1162
CWS conc.
%
68.0
68.5
68.0
69.5
70.0
63.00
27th PCC, Oct. 11-14, 2010, Istanbul, Turkey
CWS viscosity
mPa.s
712
800
805
823
768
1126
%(CWS)
0.3
0.3
0.25
0.25
0.2
1.0
>4 个月
>4 个月
>4 个月
>4 个月
>4 个月
>4 个月
Pseudoplastic
Pseudoplastic
Pseudoplastic
Pseudoplastic
Pseudoplastic
Pseudoplastic
Amount of additive CWS stability
No hard precipitation
CWS rheological ability
Table 1 shows that high concentration CWS with good rheological ability and stability with less amount addition of additive has been obtained from Shemu coal and lignite. By using this new technology, higher than 68% of CWS was produced for Shenmu low rank coal with moisture of 4.16%-6.85% in analytical base and 63% for Inner Mongolia Baorixile lignite with moisture of 15.76%. 2.1 Technical Features 1)With several self-improved dry mills, the average particle size of coal powder is less than 10 microns (typically 40-50 microns by conventional wet process) with the best PSD and the highest accumulation density. Thus, the quality of CWS is improved and the adaptability of different coals is expanded. High concentration CWS can be prepared using low rank coals and lignite. 2)CWS can be prepared for coals with higher inherent moisture even in the frozen season due to its high grinding efficiency. Thus, there is no frozen and agglomeration during transportation and storage because the dry fine powder contains 12% of humidity. 3)The fine powder containing less humidity as the product is packed by bag to avoid the specialized tanker for CWS transportation and its transport weight is also greatly reduced, leading to a significant decrease in transportation costs 4)CWS is prepared by end-user in situ and, thus, the stability of CWS is not an important factor. It is suitable for some coals that cannot produce CWS with high stability. The cost using additives can also be reduced. 5)CWS powder preparation is more suitable for North China due to coal-rich but water shortage situation. 2.2 Technological Innovations 1)Being different from the conventional wet technology, a new technology of high concentration CWS preparation using low rank coals and lignite, named dry fine coal with optimum PSD, is developed. Higher than 68% of qualified CWS is produced for Shenmu low rank coal and 63% for Inner Mongolia lignite with less energy consumption and amount of additives. The production costs are greatly reduced and the adaptability of different coals is expanded. It will also promote the conversion and utilization of low rank coals and lignite. 2)The design of the existing classifier in the mill is improved with operation under negative pressure to improve the adaptability of raw material, the powder quality and the processing capacity of the mill, and to eliminate environmental pollution. 3)An anti-static device with recovery of ultrafine coal under negative pressure is developed to ensure that the temperature rising caused by the friction between the internal grinding plate and cylinder during operation of the dry-mill can be controlled to less than 20 ℃. 95 % of the hot air was sucked away under negative pressure, and thus there is no dust leakage. 4)Multi-dry mills are used to produce pulverized coal with different yields and fineness to achieve the required optimum PSD related to coal used. 2.3 Comparison of energy consumption and relative costs of different CWS preparation
27th PCC, Oct. 11-14, 2010, Istanbul, Turkey
processes The Comparison of energy consumption and relative costs of different CWS preparation processes are shown in Table 2. Table 2 Comparison of energy consumption and relative cost between dry and wet CWS preparation technologies Electricity Technology
Yuan/ kWh
Yuan/t t/tCWS Yuan/t
Yuan/ tCWS
t/tCWS Yuan/t
Yuan/ Total cost tCWS
76%
20
0.46
9.2
0.003
4000
12
0.002
2000
4
25.20
Shenhua
73%
18
0.46 8.28
0.002
4000
8
0.002
2000
4
20.28
68%
18
0.46 8.28
0.010
4000
40
0.002
2000
4
52.28
63%
18
0.46 8.28
0.004
4000
16
0.002
2000
4
28.28
66%
18
0.46 8.28
0.003
4000
12
0.002
2000
4
24.28
63%
18
0.46 8.28
0.005
4000
20
0.002
2000
4
32.28
59%
23
0.46 10.58
18
28.58
59%
22
0.46 10.12
18
28.12
58%
23
0.46 10.58
18
28.58
59%
23
0.46 10.58
18
28.58
(Hailaer)
Dry
kWh/t
Additive-II
Shenhua
Lignite
Lignite (Hailaer) Lignite (yunnan) Lignite (Baodou) Shanghai (Shenmu) Shenmu Ch.Eng.
Wet
Conc.
Additive-I
Yulin Mining (Yulin) Hualu Hensheng (Shenhua)
It is found from Table 2 that compared with CWS wet process, the energy consumption can be decreased by 5 kWh /t CWS with half amount of additive addition using dry technology. Thus, the relative cost can be reduced about 8 CHYuan / tCWS. At the same time, due to the increase in the concentration of CWS, coal and oxygen consumption per unit product in gasification will be reduced and the effective components in synthesis gas will be increased to enhance the efficiency of coal gasification, which will also reduce the energy consumption and costs, and increase the social and economic benefits.
3 Conclusions A new technology of high concentration CWS preparation using low rank coals and lignite, named dry fine coal with optimum PSD, was developed. A demonstration plant with annual output capacity of 750,000 tons of CWS dry powder was established and operated. By this technology, higher than 68% of CWS was produced for Shenmu low rank coal with moisture of 4.16%-6.85% in analytical base and 63% for Inner Mongolia Baorixile lignite with
27th PCC, Oct. 11-14, 2010, Istanbul, Turkey
moisture of 15.76%. Also, energy consumption and amount of additives were reduced with decrease of production costs. The development of the technology resolves the problem of high concentration CWS preparation using western China rich coal resources of high quality low rank coals and lignite. The use of this technology can greatly enhance the economic and social benefits for coal chemical industry using CWS gasification. It also opens up broad prospects for the clean conversion and utilization of lignite.
References 1. Jiashan Wu, Qingjun Ji. Effect of properties of Shenmu coal on characterization of its coal water slurry. Coal Conversion, 1992, 15, 69-75 2. Chenggong Sun, Jiashan Wu, Baoqing Li. Characterization of the rheological behaviour of thermally upgraded low-rank coal water slurry. I Effect of coal upgrading temperature. J Fuel Chem & Technol., 1996, 24, 131-136 3. Xianrong Ye, Dinping Liu, Qizhong Chen, Hongming Cai. Effect of particle size distribution on concentration and viscosity of coal water slurry of blending coals. Coal Conversion, 2008, 31, 28-30
PROGRAM TOPIC: GASIFICATION FLUIDISED BED CO-GASIFICATION OF COAL AND BIOMASS UNDER OXYFUEL CONDITIONS
Nicolas Spiegl, Nigel Paterson, Cesar Berrueco and Marcos Millan* Department of Chemical Engineering, Imperial College London, London SW7 2AZ, UK *Presenting and Corresponding Author Contact Information: Department of Chemical Engineering, Imperial College London, London SW7 2AZ, UK Phone: +44(0)20 7594 1633 [email protected]
ABSTRACT Fluidised bed gasifiers are the preferred option to utilise low value coal, biomass and waste. However, fluidised bed gasifiers are traditionally air rather than oxygen-blown to avoid high temperatures in the gasifier leading to ash melting and loss of fluidisation. Therefore the flue gas of a possible FB-IGCC plant would be diluted by nitrogen, making expensive N2-CO2 separation technology necessary for subsequent capture and storage of the CO2 (CCS). To overcome this disadvantage, an oxy-fuel process is proposed, where the bed is fluidised with recycled flue gas (mainly CO2) and oxygen. A laboratory scale fluidised bed gasifier capable to operate up to 1000°C and 20 bar was set up to study the implications of oxy-fuel firing on flue gas composition and overall operability of the gasifier. High carbon conversions were achieved at 950°C with CO2 as gasification agent. At that temperature, the addition of O2 generated some agglomeration in the bed, leading to lower conversions. At lower temperatures, O2 addition increased carbon conversion, but not the heating value of the gases. The addition of steam enabled the gasifier to operate at lower temperatures without decreasing conversion and to increase the hydrogen content of the fuel gas. Pressure on the other hand produced a decrease in carbon conversion, mainly above 10 bars, due to a larger formation of secondary char with the subsequent loss in reactivity.
These results show that oxy-fuel firing of a fluidised bed gasifier could be a promising route to avoid N2 dilution of the fuel gas and enable integration of fluidised bed gasification with CCS technology.
INTRODUCTION Entrained flow gasifiers currently represent the leading technology for coal gasification due to their high throughput rates and conversions. However, in view of the increasing interest in gasification of low grade coals as well as co-gasification of coal with biomass and waste, fluidised/spouted bed gasifiers present some advantages. Firstly, they can process a broad range of fuels, including high ash and high sulphur coals. Secondly, they have less stringent needs regarding fuel preparation, making them suitable for processing biomass and certain types of waste. In addition, fluidised beds have a high tolerance to load changes, which enables them to process feedstocks with changing properties.
Air-blown fluidised bed gasification processes are not particularly well-suited for CO2 capture and storage, as combustion of the fuel gas produces a stream of CO2 diluted with large volumes of N2, whose separation is costly and complex. In order to make this process compatible with CO2 capture, it is proposed in this work the use of an O2-fired gasifier with the injection of CO2 and/or steam as both gasification agents and diluent needed to replace the role of N2 in avoiding high temperatures of the bed, which would cause ash melting and agglomeration and eventually a loss of fluidization.
Replacing N2 with CO2 and steam will affect the gasification performance and the operation of the gasifier. The main changes are an increase in the thermal capacity of the reactor, the tendency to increasing agglomeration of coal/char in atmospheres with a high partial pressure of CO2 and the reactive nature of CO2 and steam in gasification reactions in comparison with an inert gas as nitrogen, which will change the chemistry of the system. This paper focuses on the impact of the latter on the gasification of coal and coal-biomass mixtures.
EXPERIMENTAL
An existing laboratory scale fluidized bed gasifier was adapted to operate with a continuous feed under oxy-fuel conditions. A diagram of the system is shown in Figure 1. The modifications made to this reactor have been described in detail elsewhere [1]. Briefly, the reactor is made of Incoloy 800HT, 34 mm (i.d.), 504 mm (tall), which acts as both pressure shell and resistance heater. The reactor shell is heated by a high current, low voltage ac power supply, with copper electrodes connecting the reactor body to a transformer. A quartz glass liner contains the bed, protecting the reactor shell from corrosion and avoiding sample contamination by the metal shell. The feedstock used in this study was a German lignite.
T1
T2
E1 K1
V2 F3
F3 F1
F2
V9 BFA
V6
V5
F5
P2
Z1
X1 V8
I1
V7 A1
A2
CO CO2
A3
A4
A5
A6
H2
CH4
O2
CO
V13 Ventilation
I2
V4
V3
V13
MFC1
V10 R1 H1 MFC2 G1
P1 P1
O1 S1
V11
T3
V12
D1
T4 C1
U1
C2
C3
C4
T5 V1
U2
C1 C2 C3 C4 C5 C6 P1/2 T1-5 U1,2 P1 E1 K1 F1,2 F3-5
CO2 N2 20% O2/CO2 Air Calibration Gas Propane Pressure Transducer Thermo Couple Heating Tape HPLC Pump Extraction Hood Incinerator Absorber Filter
H1 D1 O1 MFC1,2 I1,2 S1 G1 R1 Z1 X1 A1-6 V9 V6 V13 BFA
Feed Hoper DC Motor Rotary Valve Mass Flow Controller Ice/Water/Salt Bath Spout Line Glass Liner Reactor Shell Tar Trap Gas Meter Online Gas Analysis H2S Measurement Point Pressure Let Down Valve Pressure Relieve Valve Back flow arrestor
Figure 1. Flowsheet of the continuous feed high pressure fluidised bed gasifier.
C5
C6
The tars were collected in a wire mesh-packed trap that uses ice/water/salt mixtures to decrease the temperature of the gases to less than 20°C. The composition of the tar-free gases was then determined by means of online analysers based on infra-red (CO, CO2, CH4), thermal conductivity (H2) and paramagnetic properties (O2).
RESULTS AND DISCUSSION 1) Effect of temperature and O/C ratio. Experiments at 750°C, 850°C and 950°C using different O2/CO2 mixtures as inlet gas were carried out at atmospheric pressure. The molar ratios between inlet O and C in the feed used in these experiments were: 0, 0.1 and 0.2, which correspond to O2 molar fractions in the gas phase of 0%, 5% and 10% respectively. The total inlet flowrate to the reactor was kept constant at 1.76 NL/min and the resulting superficial velocity was 0.16 - 0.18 m/s.
Figure 2. Carbon Conversion (left) and LHV of the fuel gas (right) as a function of the O/C ratio for gasification experiments at three different temperatures ( 950°C, □ 850°C, ∆ 750°C).
Carbon conversion markedly increased with temperature in the range 750-950ºC. At 950ºC, the conversion reached 85% (Figure 2, left). Although there was virtually no residual carbon in the bed, entrainment of fines in the outlet gases prevented the system from reaching higher conversions. The LHV of the gases increased steadily with temperature as CO2 became more reactive and reached a value of 10 MJ/m3 for gasification at 950ºC.
For the lower two temperatures used in this study, the conversion increased with the addition of oxygen, as a result of the larger extent of char combustion. At 750°C it can be estimated that approximately 60-70% of the injected O2 reacted to produce CO2, while the rest formed CO. This increased conversion of fuel carbon to CO resulted in an increase in the overall energy conversion from 31 – 36% and in the heating value of the fuel gas from 5.1 to 5.7 MJ/m3 (Figure 2, right). At 850ºC, the LHV of the gases stayed constant at 8MJ/m3 despite the enhanced carbon conversion, as the addition of O2 did not substantially change the concentrations in the gas phase (Figure 3).
Figure 3. Oxy-fuel gasification of German lignite at different temperatures and O/C ratios – Fuel gas composition (□ CO2, CO, ∆ H2, × CH4).
An altogether different trend occurred at 950ºC, where a drop in carbon conversion was observed with increasing O/C ratio. This was found to be an effect of bed material agglomeration. Analysis of the bed material recovered at different experimental times revealed that at 950ºC with 0.1 and 0.2 O/C ratios, large agglomerates had formed in the bed. These agglomerates built up in the cone of the bed around the injection spout (Figure 4), growing firstly upwards along the spout and then in width. De-fluidisation of the bottom part of the bed and related decrease in heat transfer from the walls explains the decrease in temperature detected by the bottom thermocouple. Due to the high spout velocity
(approximately 4 times bed velocity) the agglomerating material did not block the injection of solid material.
Figure 4. Bed material agglomerates produced by the use of oxygen in experiments at 950°C.
Agglomeration of silica bed material in fluidised bed reactors is a well known phenomenon. Na and K from the lignite ash together with the Si from the sand can form low melting silicates which act like a glue between the sand particles [2, 3]. The build up of agglomeration is directly related to the availability of free oxygen. No agglomeration was detected for experiments with pure CO2 and agglomerated particles built up around the spout, where levels of oxygen was at its highest. The temperature in the spout region where oxygen reacts with char and volatiles was likely to be significantly higher than the overall bed temperature and the main reason for the agglomeration. The build-up of static material in the cone of the bed prevents the recirculation of the bed material by entrainment in the spout gas. Therefore, the advantage of the spouted bed design in increasing mixing is lost, which in turn affects conversion. However, the increasing amount of agglomerates in the bed did not have an adverse effect on the overall stability of the gasification process.
2) Effect of CO2/C ratio Experiments at atmospheric pressure were carried out at 750°C, 850°C and 950°C using different CO2/C ratios. The resulting carbon conversions and fuel gas LHV are summarised in Figure 5. All experiments were carried out with 100% CO2 as fluidising gas. Different CO2/C ratios were achieved by changing the amount of CO2 injected into the reactor and keeping the fuel feeding rate constant. This changes the superficial velocity but separate experiments established that changes in superficial velocity do not significantly alter conversions under these conditions.
Figure 5. Carbon Conversion (left) and LHV of the fuel gas (right) as a function of the CO2/C ratio for gasification experiments at three different temperatures ( 950°C, □ 850°C, ∆ 750°C).
The effect of increasing CO2/C ratio was sharpest at 850ºC with a clear increase in carbon conversion from to 60% to 85%. Conversions at 950ºC were independent from the flowrate of CO2 as they had already reached a maximum level due to loss of fines as discussed above. On the other hand at 750ºC, the increase in carbon conversion was less pronounced as CO2 is less reactive at this temperature. By contrast, at 850°C, the char-CO2 reaction seems to be fast enough to be influenced by the partial pressure of CO2. At a CO2/C ratio of 1.8, 85% conversion was reached, the same as that obtained at 950°C. However, it was found that an increase in inlet CO2 dilutes the fuel gas and decreases its heating value at all temperatures. These results show that it would be possible to decrease operating temperature to 850°C and still achieve high carbon conversion by increasing the CO2 to C ratio. The LHV of the gas however drops, from 8 to around 7MJ/m3 in these experiments.
3) Injection of Steam and co-feeding of biomass. Different CO2/steam ratios (0, 0.3, 1 and 2.3) were used in gasification experiments at atmospheric pressure with GL as fuel. No O2 was added in this set of experiments. The total flow rate entering the reactor was kept constant at 2.64 NL/min and the superficial velocity was kept in the range 0.24 - 0.27 m/s. The resulting fuel gas compositions are presented in Figure 6.
Figure 61. Oxy-fuel gasification of German lignite at different temperatures and Steam/CO2 ratios ( CO, □CO2, ∆ H2, × CH4).
With increasing steam/CO2 ratio, the char-CO2 reaction is partially replaced with charsteam reaction. This produced an increase in the H2 concentration and a decrease in the CO concentration in the fuel gas. As shown in Figure 7, replacing 25% of CO2 by steam increased the carbon conversion at 750°C and 850°C while at 950°C no change occurred. Similarly to what was found in the case of CO2/C ratio, at 950°C an upper limit in carbon conversion was achieved, independently of steam/CO2 ratio. By adding 25% steam at 850°C, carbon conversion was increased to a level comparable to that of experiments carried out at 950°C. Further increases in steam to CO2 ratio did not change the carbon conversion at 850°C, although the hydrogen production continued increasing.
Figure 7. Carbon Conversion as a function of the steam/CO2 ratio for gasification experiments at three different temperatures ( 950°C, □ 850°C, ∆ 750°C).
Experiments run with a partial substitution of the German lignite by olive bagasse showed this as a way to increase the carbon conversion without negatively affecting the gasification performance. For example, when the gasifier is operated at 850°C with GL the carbon conversion is 75%. In order to increase the carbon conversion either temperature, the CO2/C ratio or the steam concentration would have to be increased. All this would significantly reduce the efficiency of the gasifier. However replacing 30% of the fuel with OB would increase the carbon conversion to nearly 100%, while maintaining the gasification efficiency. A synergistic effect between coal and biomass was observed. In addition, co-gasification is a potential pathway to utilise waste for power generation and to reduce the CO2 emissions of the plant.
4) Effect of Pressure Experiments were also performed at pressures up to 20bar. Carbon conversion decreased significantly with increasing pressure mainly above 10 bar. These results may be a combination of a slight decrease in primary volatile release, as established in a separate experiment using a Wire Mesh Reactor, and char deactivation due condensation of tars on the char particles and the formation of secondary char. The increasing fuel feeding rate and the subsequently higher particle density in the injection region of the FBR were identified as main reason for the increased coverage of char particles with secondary char. Similar trends to those described for atmospheric pressure operation were obtained regarding variation in carbon conversion with temperature, CO2 to C ratio and steam addition.
Figure 8. Carbon Conversion as a function of pressures for gasification experiments at different conditions (□ 850⁰C- CO2, 850⁰C -O2/ CO2, ○ 850⁰C-steam/O2/CO2, ∆ 950⁰C-O2/CO2).
The fuel gas compositions are shown in Figure 9. With increasing pressure the concentration of CO2 increased (60 – 70%) while that of CO diminished (30 – 20%). At the same time less fuel-hydrogen was converted to H2 (43 – 30%) and more to H2O. The concentration
of
CH4
in
the
fuel
gas
increased
slightly
with
pressure.
Figure 9. Oxy-fuel gasification of GL at 850C with 3.7% O2/CO2 (O/C = 0.1) at different pressures: Fuel gas composition (□CO2, CO, ∆ H2, × CH4).
CONCLUSIONS A laboratory scale fluidised bed gasifier was used to study the gasification performance of a German lignite under O2/CO2/steam–blown conditions. A high carbon conversion, only limited by loss of fines entrained in the fuel gas, was achieved when the gasifier was operated at 950°C. To achieve conversions closer to 100%, recirculation of fines would be necessary. Similar carbon conversions can be reached at 850°C if either the inlet CO2 to C in the coal ratio is increased or if a fraction of the inlet CO2 is replaced with steam. At the lower temperatures (750 and 850°C), conversion can also be enhanced by using a higher O2/CO2 ratio in the inlet. At these temperatures, addition of O2 slightly increases the heating value of the gases. However, increasing the O2 content of the inlet gases at 950°C produced a drop rather than an increase in carbon conversion due to agglomeration of bed material in the zone near the spout.
Using a higher flowrate of CO2 produces a decrease in LHV. The increasing addition of steam generates a linear increase in the H2 fraction in the fuel gas, even after the carbon
conversion has reached its maximum. This is accompanied by a decrease in CO as predicted from the water-gas shift reaction equilibrium.
Increasing pressure above 10 bars produced a drop in carbon conversion due to a drop in char reactivity. It is thought that this is due to deposition of tars on the char surface leading to the formation of secondary char, which is enhanced by the higher coal throughputs used at higher pressures. Overall, these results show oxy-fuel firing of a fluidised bed gasifier as a promising route to avoid N2 dilution of the fuel gas and enable integration of fluidised bed gasification with CCS technology.
REFERENCES [1] Spiegl, Sivena, Lorente, Paterson, and Millan, “Investigation of the Oxy-fuel Gasification of Coal in a Laboratory-Scale Spouted-Bed Reactor: Reactor Modifications and Initial Results”, Energy Fuels 2010, 24, 5281–5288.
[2] Kikuchi, Suzuki, Mochizuki, Endo, Imai, and Tanji, “Ash Agglomeration Gasification of Coal in a Spouted Bed Reactor”, Fuel 1985, 64, 368-372.
[3] Bartels, Lin, Nijenhuis, Kapteijn, Ommen, “Agglomeration in Fluidized Beds at High Temperatures: Mechanisms, Detection and Prevention”, Progress in Energy and Combustion Science, 2008, 34, 633-666.
THE 27TH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE. ISTANBUL, 2010.
1
Co-Gasification of Footwear Leather Waste and High Ash Coal: A Thermodynamic Analysis Rodolfo Rodrigues∗ , Nilson R. Marcilio∗ , Jorge O. Trierweiler∗ , Marcelo Godinho† and Adriene M. S. Pereira‡ ∗ Federal University of Rio Grande do Sul (UFRGS), Department of Chemical Engineering. Porto Alegre (RS), Brazil. E-mail: [email protected], [email protected], [email protected] † University of Caxias do Sul (UCS), Department of Chemical Engineering. Caxias do Sul (RS), Brazil. E-mail: [email protected] ‡ Pontifical Catholic University of Rio Grande do Sul (PUC-RS), Department of Chemical Engineering. Porto Alegre (RS), Brazil. E-mail: [email protected]
Abstract—The leather and shoe industry is one of the sectors generates more wastes at same time that has a high polluting potential. In Brazil, the majority of these wastes are disposed of in landfills and less than 5% are recycled. It corresponds to 62.5% of about 190,000 tons per year of hazardous wastes have been generated there. Since 1997 the Laboratory of Residues Processing (LPR) at the UFRGS has been developing practical projects on thermal treatment of leather wastes (biomass) based on the combined gasification-combustion technology for cogeneration. This work evaluates the co-gasification of leather wastes together a fraction of local coal over the fuel gas (syngas). The coal is subbituminous and high ash content (nearly 50%). A thermodynamic equilibrium model is applied for analysis of the co-gasification process. A sensitivity analysis of biomass cogasification related to usual operational parameters (air, steam and blending ratios) is done. That allows to identify efficiencies of ∼64% for middle air demands (∼60% of stoichiometric demand) at any leather-coal blending. Whereas an efficiency of ∼75% should be reached with steam-to-carbon ratios higher than 1.5. From future studies the model can be used to evaluated the formation of N and S compounds in the flue gas. Index Terms—Co-gasification, footwear leather waste, high ash coal, equilibrium model.
I. I NTRODUCTION The leather and shoe industry is one of the sectors generates more wastes at same time that has a high polluting potential. This environmental problem is proven the danger of waste containing chromium, derived from the salt utilized to tan hides. In Brazil, the majority of these wastes are disposed of in landfills and less than 5% are recycled [1]. Southern region locates most of those industries since more than 50% of Brazilian leather is produced in the State of Rio Grande do Sul (RS). It corresponds to 62.5% of about 190 thousand tons per year of hazardous wastes have been generated there [2]. Studies from Godinho et al. [1] shown nearly 50% wastes generate by those industrial sectors are footwear leather residues. The products contain chromium source from tanning substances used in hide processing. The Brazilian regulation [3] classifies chromium compounds as hazardous components in waste materials due to carcinogenic affect under trivalent state — Cr(VI) [4]. These residues contain high quantities of volatile matter and low ash content, while their heating value
is high sufficiently to enables combustion under auto-thermal conditions [1]. Since 1997 the Laboratory of Residues Processing (LPR) at the UFRGS has been developing practical projects on thermal treatment of leather wastes. In 2003 a semi-pilot unit based on the combined gasification-combustion technology has started to operate and results to 350 kW of thermal generating. From new investments the thermal capacity of the unit increases to 600 kW that starts to operation on the second half of 2010. The unit will produce power from the heat generated in the combustion. With the gasification-combustion technology a reducing environment followed by an oxidative environment decreases the possibility of formation of critical components such as Cr(VI) and PCDD/F [1], [5]. Otherwise leather waste as biomass makes the gasification into attractive method of treatment in result a neutral carbon emission. The Brazilian coal mines are placed in the South and the power plants are installed near these mines [6]. Rio Grande do Sul has 89.25% of coal reserves from 7 billion tons total of Brazil [7], [8]. The coal is subbituminous at Candiota and Le˜ao-Buti´a, and high volatile bituminous at Santa Terezinha [9]. Besides that, all of them are high ash content (about 48.5%) typical of Brazilian coals. It is low nitrogen content (0.61%) and relativity high sulfur content (2.15%). And it is poor in carbon (25.1%) compared to worldwide coals (65.7– 85.8%) [6] that limits the thermal yield. As a biomass, the leather wastes are high nitrogen content (12.42%) and they have a sulfur composition (1.83%) around the sulfur content of coal. Its calorific value (18.45 MJ/kg) is 37.7% higher than coal calorific value (13.4 MJ/kg) which should be justified for a higher H/C ratio (17.3% against to 10.4% of coal). The proximate and ultimate analyses and the high heating value of the footwear leather wastes and a mean coal sample are given in Table I. In literature a few number of studies on the thermochemical processing of leather wastes are represent and all of them use leather alone [1], [10]–[14]. This work evaluate the possibility of co-gasifying leather wastes with a fraction of local coal to take advantage of combining the features of each feedstock alone. Low-grade coals are usually difficult to gasify [15], but
2
THE 27TH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE. ISTANBUL, 2010.
TABLE I C HARACTERIZATION OF THE FEEDSTOCKS CONSIDERED IN THIS STUDY. Leather waste1
Coal2
Proximate analysis (%wt) Moisture 12.4 11.0 Volatile matter 67.7 20.3 Fixed carbon 14.8 25.1 Ash 5.1 42.7 Ultimate analysis (%wt, d.b.) C 49.31 33.39 H 8.52 3.47 O 24.70 16.68 N 12.42 0.61 S 1.83 2.15 Cl 0.45 <0.007 Cr 2.77 – HHV (kJ/kg, d.b.) 18,448 13,386 LHV (kJ/kg, d.b.)* 16,729 12,686 1 Ref. [1]. 2 Mean values from Candiota, Le˜ ao-Buti´a and Santa Terezinha coalfields [9]. *Estimated value
mixing it with leather wastes could equilibrate effects of ash and sulphur contents, making more viable the gasification of these coals. For that the main operational parameters: air-fuel ratio, steam-fuel ratio, and leather-coal blending are analysis regarding to heating value increasing in release gases.
III. R ESULTS AND D ISCUSSION This study applies a thermodynamic equilibrium model to evaluate co-gasifying operations. Equilibrium models are valuable because they predict the thermodynamic limits of the gasification reaction system. However, it remains important to realize that the main assumptions behind this approach may not always be valid for practical gasifiers [20]. At literature we can discern two distinct equilibrium approaches: stoichiometric and non-stoichiometric [21]. Although equivalent in essence, the stoichiometric approach applies the equilibrium constants from related chemical reactions [22]–[24], while the non-stoichiometric approach minimizes the Gibbs free energy subject to mass balance and nonnegative constraints [25]–[27]. We assume a non-stoichiometric approach for this study. In general, the main hypothesis considers that gasification reaction rates are fast enough and residence time is sufficiently long to reach the equilibrium state. Moreover the model also assumes: • • •
II. T HE C O -G ASIFICATION T ECHNOLOGY •
Since the coal is one of major source of energy it is predicted that will continue to play an important role in the world energy demands. Although as a fossil resource, that should take into account atmosphere emissions. According to Hern´andez et al. [16], the co-gasification is a thermochemical process that involves the joint conversion of 2 or more carbon-based fuels (one of them of fossil origin) into a gaseous fraction with a usable heating value. This promising technology is becoming increasingly important since it allows one to use coal in a cleaner way. Fermoso et al. [17] consider the co-gasification which a bridge between the energy production systems based on fossil fuels and those based on renewable energy sources. It could contribute to reducing CO2 emissions and the dependency on fossil fuels. The addition of biomass to coal leads to a reduction of both CO2 emissions and operating problems associated with the N and S coal content. The higher reactivity of biomass compared to coal [16] promotes to an improvement of the gasification process without unstabilize the system operation. Co-gasification might mitigate the effect of short term variations in feedstock quality, or buffer the system [18]. Therefore it can contribute to the introduction of biomass as a gasification fuel on the commercial scale, since it could be used in existing coal facilities [16]. Finally, we should say one of the co-gasification issues, that is increasingly raising interest, is the existence of synergies among coal and biomass regarding thermochemical processes [16]. Some authors [17], [19] suggest such effects will occur is dependent upon certain gasification conditions such as feedstock type, operation conditions, reactor type, etc.
•
• •
The feedstocks are a combinations of C, H, O, N, S, and Cl elements. A multiphase formulation consisting of a two-phase misture: gas and solid-phases. The solid carbon fraction produced (char) is represented by graphite [25]. There is not a tar fraction so that it is assimilated to solid carbon. Ash fraction is considered to be chemically inert and only evaluate to thermal effects. Its composition is assumed as only Cr2 O3 , SiO2 , Al2 O3 , Fe2 O3 , TiO2 , P2 O5 , CaO, Na2 O, K2 O, MgO (Table II). The gasifier is often regarded as a perfectly insulated apparatus, i.e. an adiabatic system. The gasifier is perfect mixing and uniform temperature.
The model considers 95 chemical species, 83 for the gaseous phase and 12 for the solid one (shown in Table III). All thermodynamic data applied on this study are from Burcat and Ruscic [28]. Burcat’s thermodynamic database includes ideal gas thermodynamic data in polynomial form, for over 1,300 chemical species used in combustion and air pollution. For this analysis 61 species of Baratieri’s study [25] were chosen considering together 24 Cl-Cr compounds to improve the prediction capability regards leather wastes [29] and also 11 metal oxides for represent the ash fraction. The feedstocks are assumed as a substitution fuel formula Ca Hb Oc Nd Se Clf Ashg starting from the ultimate and proximate analyses (Table I). As nonconventional compounds, the fuels (leather waste and coal) must have the standard enthalpis of formation estimated from its higher heating value (HHV ). So that a stoichiometric combustion reactions (normalised in C) is considered as below: CHb Oc Nd Se Clf Ashg + 1 + −→ CO2 +
b 2 H2 O(l)
+
d 2 N2
b 4
−
c 2
+e−
f 2
O2
+ eSO2 + f ClO + +gAsh + HHV
(1)
R. RODRIGUES et al.: CO-GASIFICATION OF LEATHER WASTE AND HIGH ASH COAL
3
TABLE II A SH CHARACTERIZATION OF THE FEEDSTOCKS CONSIDERED IN THIS STUDY.
Fuel Cr2 O3 SiO2 Al2 O3 Leather waste1 55.91 23.58 7.35 Coal2 NA 29.54 9.02 1 Ref. [1]. 2 Ref. [9]. NA: not available.
Concentration Fe2 O3 TiO2 2.59 1.53 4.21 0.36
(wt%) P2 O5 1.59 0.03
CaO 1.20 1.02
Na2 O 1.08 0.16
K2 O 0.79 0.70
MgO NA 0.41
TABLE III L IST OF CHEMICAL SPECIES CONSIDERED BY THE EQUILIBRIUM MODEL Phase Gas
Group Inorganic carbon compounds Hydrocarbons
Other organic compounds Oxygen compounds Hydrogen compounds Nitrogen compounds
Sulfur compounds Chlorine compounds Chromium compounds
Solid
Carbon Chromium compounds Other compounds (ash)
Compounds C(g), CO, CO2 CH, CH2 , CH3 , CH4 , C2 H, C2 H2 , C2 H3 , C2 H4 , C2 H5 , C2 H6 , C3 H7 , C3 H8 , C6 H6 , C10 H8 , C12 H10 CH2 O, CH2 OH, CH3 O, CH3 OH, HCCO, CH2 CO, HCCOH, CH2 CHO, CH3 CHO O, O2 H, H2 , OH, H2 O, HO2 , H2 O2 , HCO N, N2 , NH, NH2 , NH3 , NNH, NO, NO2 , N2 O, HNO, CN, HCN, H2 CN, HCNN, HCNO, HOCN, HNCO, NCO S, SO2 , SO3 , H2 S, COS, CS2 Cl, Cl2 , HCl, ClONO2 , ClO, ClO2 , ClO3 , Cl2 O, Cl2 O2 , Cl2 O7 Cr(g), CrO, CrO2 , CrO3 , CrOOH, CrO2 (OH), CrO2 (OH)2 , CrCl, CrCl6 , CrOCl, CrOCl2 , CrO2 Cl, CrO2 Cl2 C(s) Cr(s), Cr2 O3 SiO2 , Al2 O3 , Fe2 O3 , TiO2 , P2 O5 CaO, Na2 O, K2 O, MgO
(a) Leather waste
And from Ref. [30] we define the standard enthalpy of ◦ formation (∆Hf,fuel ) as follows: ◦ ◦ ◦ ◦ ∆Hf,fuel = ∆Hf,CO + 2b ∆Hf,H + e∆Hf,SO + 2 2 O(l) 2 ◦ ◦ +f ∆Hf,ClO + g∆Hf,Ash + HHV ·Mwfuel
(2)
The simulations have been carried out through a python code using Cantera software libraries [31]. Cantera is a suite of oriented-object package for problems involving chemical kinetics, thermodynamics, and transport processes. The model applied in this study uses a VCS algorithm (multiphase method) implemented in Cantera, that finds the composition minimizing the total Gibbs energy of an ideal mixture [32]. That algorithm searches among a set of states that satisfies the element constraints for the one state that satisfies the conditions of chemical equilibrium. The equilibrium temperature for each operational condition is estimated in an external iterative calculation such as in Caton’s application [27]. Figure 1 shows a pyrolysis test for both feedstocks separately (Figs. 1a and 1b). This case assumes the termochemical processing of fuels under absence of oxygen (gasifying agent) in a large range of temperature (200–1,100◦ C). The test results shows that only 6 species is present in a significant fraction besides of a residual solid carbon fraction, since at
(b) Coal Fig. 1. Single pyrolysis tests at 1 atm for leather waste (a) and coal (b), where dotted lines represent the solid carbon fraction (char).
high temperature even that products are mainly in the gas phase. Analysing the figures, it can be seen that CH4 and CO2 formation are favored at lower temperatures whereas CO and H2 are favored at higher temperatures (>800◦ C). That range also corresponds to maximum conversion of solid carbon (dotted lines), reaching values of 19.6% for leather waste and 29.5% for coal. A sensitivity analysis of biomass co-gasification related to usual operational parameters is done varying leather-coal blending. The operational parameters are the air supplied (25◦ C and 1 atm) and the steam supplied (200◦ C and 1
4
THE 27TH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE. ISTANBUL, 2010.
atm) contents. That allows to identify the performance of the co-processing of biomass-coal blends regarding syngas heat values. The feed air flow rate is related to the feedstock by a parameter called equivalence ratio (φ, Eq. 3), which indicates the oxygen used relative to that required for complete combustion (stoichiometric oxygen). In a similar way, the steam flow rate is quantified by a steam-to-carbon ratio (stm, Eq. 4): φ=
nO2 1+
b 4
−
stm =
c 2
+e−
f 2
(3) nfuel
nH2 O(vap) nC
(4)
The responses to parametric changes were observed in output parameters: higher heating value (HHV ) of the energetic gas product (CO, H2 , and CH4 ) (Eq. 5) and cold gas efficiency (ηcg ) (Eq. 6), according to the equilibrium model. These output parameters are important to quantify the performance of the process.
(a) Coal with no ash
◦ ◦ HHVproducts = nCO ∆Hf,CO − ∆Hf,CO + 2 ◦ ◦ ◦ +nCH4 ∆Hf,CH4 − ∆Hf,CO2 − 2∆Hf,H2 O(l) − ◦ −nH2 ∆Hf,H 2 O(l)
ηcg =
HHVproducts HHVfuel ·Mwfuel
(5) (6)
As first analysis of gasifying process it is presented a comparative between a coal sample with no ash fraction (Fig. 2a) and a full coal sample (Fig. 2b). Since 42.7%wt of coal is ash, the decreasing of efficiency could be seen. Both cases show the peak of maximum efficiency at point where the solid carbon fraction is fully consumed. It happens at φ = 0.67 for efficiency of 68.5% regards to coal with no ash and φ = 0.57 for efficiency of 66.0% to full coal. In absolute terms they correspond to 4.04 Nm3air /kgcoal and 2.52 Nm3air /kgcoal , respectively, that is, a higher ash amount means a higher air demand for approximately the same efficiency because of the higher initial carbon fraction. Figures 3 and 4 compare the main output parameters of gasifying process for both samples. Leather waste (dashed lines) has been a higher efficiency (69.15%) for a lower temperature (1,472◦ C) and air-to-fuel ratio (φ = 0.54). These values are close to those simulated in previous study (φ = 0.34–0.53) using a stoichiometric approach [24]. Coal (dotted lines), in its turn, reaches the maximum efficiency of 66% at 1,537◦ C (φ = 0.57). Notwithstanding those similar temperature profiles (Fig. 3), coal (dotted lines) and leather (dashed lines) reach the peak by means of different absolute air demands, respectively, 5.68 Nm3air /kgleather and 4.40 Nm3air /kgcoal Last point of analysis suggests the study of co-gasifying process involving leather waste and coal. In Figure 5, the efficiencies are calculated to different operational conditions. Figs. 5a to 5f show a set of inlet steam amount (steam-tocarbon ratio) covering a range of air-to-fuel ratio (φ = 0.4–1.0). Here it should be noted that each equivalence ratio is related
(b) Coal with ash Fig. 2. Gasification of coal at 1 atm considering (a) with no ash content and (b) with ash content, where dotted lines represent the cold gas efficiencies.
to the specific blend composition. The evaluation of estimated data suggests that a minimum efficiency of ∼64% should be attained with φ = 0.58–0.63 at any leather-coal blending. However, an increasing of efficiency in +6% (70% total) can be taken with a 57.5%–32.5%wt leather-coal blend (or richer in leather content) and lower air-to-fuel ratio <0.56 by a steamto-carbon ratio at least 1.0 (Fig. 5c). An additional of efficiency in 5% (75% total) should be attained just increasing steamto-carbon ratio >2.0 and lower air demands <0.45 to any blends (Fig. 5e). Higher steam-to-carbon ratios do not cause any significant gain of efficiency (Fig. 5f). IV. C ONCLUSION A two-phase equilibrium model has been presented in this study. The model has been applied to predict important bases to measure the performance of a specific system. The analysis of a co-gasifying process of leather-coal feedstocks could be done evaluating operational parameters: air-to-fuel and steamto-fuel ratio. The single thermochemical processing of fuels
R. RODRIGUES et al.: CO-GASIFICATION OF LEATHER WASTE AND HIGH ASH COAL
5
R EFERENCES
Fig. 3. Equilibrium temperature for gasifying process of single leather waste and coal at 1 atm, where dashed line represents the leather waste and dotted line represents the coal
Fig. 4. Cold gas efficiency for gasifying process of single leather waste and coal at 1 atm, where dashed line represents the leather waste and dotted line represents the coal
shows higher efficiencies of 69.15% and 66%, respectively, to leather waste and coal. Whereas an efficiency of ∼75% should be reached with steam-to-carbon ratios higher than 1.5 to any leather-coal blends at lower air demands (φ < 0.45). And efficiencies of ∼64% could already be reached with any blends and middle air demands (φ = 0.58–0.63). Next study steps include additional analyses regards to atmosphere emissions (formation of N and S compounds) that is also important into improvement of process performance in an environmental point of view. ACKNOWLEDGMENT The study received financial support from the Brazilian Research Foundation (CNPq) under project 551386/2010-0 and also from the Brazilian Network of Coal (http://www. ufrgs.br/rede carvao).
[1] M. Godinho, N. R. Marcilio, A. C. F. Vilela, L. Masotti, and C. B. Martins, “Gasification and combustion of the footwear leather wastes,” J. Am. Leather. Chem. As., vol. 102, no. 6, pp. 182–190, 2007. [2] FEPAM, “Report on the generation of industrial solid wastes in Rio Grande do Sul,” Foundation for Environmental Protection in Rio Grande do Sul (FEPAM), Porto Alegre, Brazil, Tech. Rep., 2003. [3] ABNT, NBR 10,004: Solid waste – Classification, Std., 2004. [4] K. Kolomazn´ık, M. Adamek, I. Andel, and M. Uhlirova, “Leather waste — Potential threat to human health, and a new technology of its treatment,” J. Hazard. Mater., vol. 160, no. 2–3, pp. 514–520, 2008. [5] M. Godinho, N. R. Marcilio, L. Masotti, C. B. Martins, D. E. Ritter, and B. M. Wenzel, “Formation of PCDD and PCDF in the thermal treatment of footwear leather wastes,” J. Hazard. Mater., vol. 167, no. 1–3, pp. 1100–1105, 2009. [6] F. S. Depoi, D. Pozebon, and W. D. Kalkreuth, “Chemical characterization of feed coals and combustion-by-products from Brazilian power plants,” Int. J. Coal. Geol., vol. 76, no. 3, pp. 227–236, 2008. [7] ANEEL, “Atlas electrical energy of Brazil,” National Agency of Electrical Energy (ANEEL), Bras´ılia, Brazil, Tech. Rep., 2008. [8] EPE, “Brazilian energy balance 2009, year 2008,” Energy Research Company (EPE), Rio de Janeiro, Brazil, Tech. Rep., 2009. [9] W. Kalkreuth, M. Holz, M. Kern, G. Machado, A. Mexias, M. Silva, J. Willett, R. Finkelman, and H. Burger, “Petrology and chemistry of Permian coals from the Paran´a basin: 1. Santa Terezinha, Le˜ao-Buti´a and Candiota coalfields, Rio Grande do Sul, Brazil,” Int. J. Coal. Geol., vol. 68, no. 1–2, pp. 79–116, 2006. [10] R. Font, J. A. Caballero, M. M. Esperanza, and A. Fullana, “Pyrolytic products from tannery wastes,” J. Anal. Appl. Pyrol., vol. 49, no. 1–2, pp. 243–256, 1999. [11] A. Bahillo, L. Armesto, A. Cabanillas, and J. Otero, “Thermal valorization of footwear leather wastes in bubbling fluidized bed combustion.” Waste. Manage., vol. 24, pp. 935–944, 2004. [12] J. Yanik, S. Ebale, A. Kruse, M. Saglam, and M. Y¨uksel, “Biomass gasification in supercritical water: Part 1. Effect of the nature of biomass,” Fuel, vol. 86, no. 15, pp. 2410–2415, 2007. [13] O. Yilmaz, I. C. Kantarli, M. Y¨uksel, M. Saglam, and J. Yanik, “Conversion of leather wastes to useful products,” Resour. Conserv. Recy., vol. 49, no. 4, pp. 436–448, 2007. [14] S. Swarnalatha, A. G. Kumar, S. Tandaiah, and G. Sekaran, “Efficient and safe disposal of chrome shavings discharged from leather industry using thermal combustion,” J. Chem. Technol. Biot., vol. 84, no. 5, pp. 751–760, 2009. [15] R. N. Andr´e, F. Pinto, C. Franco, M. Dias, I. Gulyurtlu, M. A. A. Matos, and I. Cabrita, “Fluidised bed co-gasification of coal and olive oil industry wastes,” Fuel, vol. 84, no. 12–13, pp. 1635–1644, 2005. [16] J. J. Hern´andez, G. Aranda-Almansa, and C. Serrano, “Co-gasification of biomass wastes and coal–coke blends in an entrained flow gasifier: An experimental study,” Energ. Fuel., vol. 24, no. 4, pp. 2479–2488, 2010. [17] J. Fermoso, B. Arias, M. Gil, M. Plaza, C. Pevida, J. J. Pis, and F. Rubiera, “Co-gasification of different rank coals with biomass and petroleum coke in a high-pressure reactor for H2 -rich gas production,” Bioresource. Technol., vol. 101, no. 9, pp. 3230–3235, 2010. [18] B. Ricketts, R. Hotchkiss, B. Livingston, and M. Hall, “Technology status review of waste/biomass co-gasification with coal,” in 5th Eur. Gasif. Conf., Noordwijk, The Netherlands, April 2002. [19] K. Sj¨ostr¨om, G. Chen, Q. Yu, C. Brage, and C. Ros´en, “Promoted reactivity of char in co-gasification of biomass and coal: Synergies in the thermochemical process,” Fuel, vol. 78, no. 10, pp. 1189–1194, 1999. [20] M. J. Prins, K. J. Ptasinski, and F. J. J. G. Janssen, “From coal to biomass gasification: Comparison of thermodynamic efficiency,” Energy, vol. 32, no. 7, pp. 1248–1259, 2007. [21] X. Li, J. R. Grace, A. P. Watkinson, C. J. Lim, and A. Erg¨udenler, “Equilibrium modeling of gasification: A free energy minimization approach and its application to a circulating fluidized bed coal gasifier,” Fuel, vol. 80, no. 2, pp. 195–207, 2001. [22] A. Melgar, J. F. P´erez, H. Laget, and A. Horillo, “Thermochemical equilibrium modelling of a gasifying process,” Energ. Convers. Manage., vol. 48, no. 1, pp. 59–67, 2007. [23] A. K. Sharma, “Equilibrium modeling of global reduction reactions for a downdraft (biomass) gasifier,” Energ. Convers. Manage., vol. 49, no. 4, pp. 832–842, 2008. [24] R. Rodrigues, A. R. Secchi, N. R. Marcilio, and M. Godinho, “Modeling of biomass gasification applied to a combined gasifier-combustor unit:
6
[25]
[26] [27]
[28]
[29]
[30] [31] [32]
THE 27TH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE. ISTANBUL, 2010.
Equilibrium and kinetic approaches,” in 10th Int. Symp. Proc. Syst. Eng., vol. 27A, 2009, pp. 657–662. M. Baratieri, P. Baggio, L. Fiori, and M. Grigiante, “Biomass as an energy source: Thermodynamic constraints on the performance of the conversion process,” Bioresource. Technol., vol. 99, no. 15, pp. 7063– 7073, 2008. P. Baggio, M. Baratieri, L. Fiori, M. Grigiante, D. Avi, and P. Tosi, “Experimental and modeling analysis of a batch gasification/pyrolysis reactor,” Energ. Convers. Manage., vol. 50, no. 6, pp. 1426–1435, 2009. P. A. Caton, M. A. Carr, S. S. Kim, and M. J. Beautyman, “Energy recovery from waste food by combustion or gasification with the potential for regenerative dehydration: A case study,” Energ. Convers. Manage., vol. 51, no. 6, pp. 1157–1169, 2010. A. Burcat and B. Ruscic, “Third millennium ideal gas and condensed phase thermochemical database for combustion with updates from active thermochemical tables,” Chemistry Division, Argonne National Laboratory, Argonne, Illinois, Report TAE 960 Table 4, 2005. [Online]. Available: ftp://ftp.technion.ac.il/pub/supported/aetdd/thermodynamics O. E. Kashireninov and A. Fontijn, “Modeling of chromium combustion in incineration: Thermochemistry of Cr-C-H-Cl combustion in air and selection of key reactions,” Combust. Flame., vol. 113, no. 4, pp. 498– 506, 1998. M. L. Souza-Santos, Solid fuels combustion and gasification: Modeling, simulation, and equipment operation. New York: Marcel Dekker, 2004. D. Goodwin, Cantera: An object-oriented software toolkit for chemical kinetics, thermodynamics, and transport processes, Caltech, Pasadena, 2009. [Online]. Available: http://code.google.com/p/cantera W. R. Smith and R. W. Missen, Chemical reaction equilibrium analysis: Theory and algorithm. John Wiley & Sons, 1982.
(e) At stm = 2.0
(d) At stm = 1.5
(f) At stm = 6.0
(c) At stm = 1.0
Fig. 5. Cold gas efficiency of the gasifying process of leather-coal blends in range φ = 0.4–1.0 for different steam-to-carbon ratios (stm = 0, 0.5, 1.0, 1.5, 2.0, and 6.0) at 1 atm. In the leather-coal blending, 0 means only leather waste and 1 means only coal.
(b) At stm = 0.5
(a) At stm = 0
R. RODRIGUES et al.: CO-GASIFICATION OF LEATHER WASTE AND HIGH ASH COAL 7
Mineralogical, Petrographic and Geochemical Features of the Achlada and Mavropigi Lignite Deposits, NW Macedonia, Greece Nikolaos Koukouzas1, Stavros P. Kalaitzidis2, Colin R. Ward3, Zhongsheng Li3 1. Centre for Research and Technology Hellas, Institute for Solid Fuels Technology and Applications, Mesogeion Ave. 357-359, GR-15231 Halandri, Athens, Greece 2. Geological Services, BHP Billiton Mitsubishi Alliance, Central Queensland Office, Peak Downs Mine, Moranbah, 4744, Australia 3. School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, 2052, Australia
ABSTRACT The Achlada and Mavropigi lignite deposits in northern Greece provide the main coal source for the next generation of Greek power plants. A comparative characterization of these two lignite deposits is presented, covering the coal rank and the features of the maceral components, based on detailed coal petrography, and the mineralogical and geochemical features of the coals and their ashes, based on XRF and XRD analysis. The data are used to interpret the palaeoenvironments of the lignite beds, and identify factors that may affect their combustion behavior. The Achlada lignite, with ash yields >30 wt % (db), contains a higher proportion of inorganic material than the Mavropigi deposit (ash values <30 wt %, db). Petrographic data indicate significant fragmentation of the organic matter in the Achlada lignite, attributed to accumulation mostly of soft herbaceous tissues combined with mechanical destruction of the tissues during transportation. Deposition took place in a fluvial environment, mainly in a reed-marsh setting but with minor peat accumulation under forest swamp conditions. The associated mineral matter is dominated by illite, smectite and I/S with lesser proportions of kaolinite and quartz. The calcium sulphate mineral bassanite, derived from interaction of organically-associated Ca and S during oxidation of the maceral components, is also found in the low-temperature ash (LTA) of the Achlada materials. The Mavropigi lignite was developed in a lacustrine environment, in which peat-forming periods alternated with deposition of shell-rich calcareous sediment during widespread flooding events. The nature of the organic matter suggests formation in a reed-marsh environment, interrupted by short dry periods and the formation of inertinite macerals. Greater proportions of calcium sulphates are developed in the LTA than from the Achlada lignite, consistent with a lower proportion of detrital input to the paleomire and a greater overall proportion of organically-associated Ca. Kaolinite, thought to be mainly authigenic, is also more abundant, consistent with a lower rate of detrital input in the less-contaminated lacustrine setting. The different nature of the inorganic matter in the two deposits needs to be taken into account in optimizing their utilization for power production. A preliminary assessment indicates that the Achlada lignite may have more favorable slagging and fouling properties than the Mavropigi lignite, although experimental studies are required for more definitive conclusions to be reached. INTRODUCTION Lignite combustion provides almost 60% of the electric power used in Greece, drawing on resources in more than 60 sedimentary basins across the country (Kavouridis and Koukouzas, 2008). The Achlada and Mavropigi deposits, located in the Ptolemais-Florina Graben (Figure 1) in the north1
west of the country, are two fields that have been recently developed to maintain the nation’s lignite supplies for power generation. These two deposits are of different geological age and were formed in somewhat different depositional environments, leading to differences in petrographical constituents and mineral matter characteristics that need to be recognised for effective operation of the power plants to which they are supplied.
Figure 1: Location of the Achlada and Mavropigi lignite deposits The Achlada deposit, in the Florina Basin, is part of the Middle Miocene Komnina lignite sequence and is dominated by xylite-rich lignite. The lignites in this succession were developed mainly in abandoned channels and on the floodplain of a meandering river system (Oikinomopoulos et al., 2007). The Mavropigi deposit, in the adjacent but younger Ptolemais Basin, is part of the Early Pliocene Ptolemais lignite sequence and is represented mainly by matrix lignite material (Koukouzas, 2007). The Mavropigi lignites were developed in a lacustrine environment that occupied the broader area of the Ptolemais Basin during Pliocene times (Antoniadis et al., 2006; Koukouzas et al., 2010a). SAMPLING AND ANALYSIS PROGRAM Five lignite samples and two inorganic sediment samples were collected from a mine exposure to represent the section worked in the Achlada deposit (Figure 2), and a further five lignite samples plus a selection of associated inorganic samples from an open-cut mine to represent the Mavropigi deposit. Further details of the sections studied are given by Koukouzas et al. (2009, 2010b). 2
The lignite samples were crushed to < 200 μm, and moisture (105-110°C) and ash yield (815°C) were determined using ASTM procedures. Volatile matter was determined using a TGA-2000 automatic multiple sample thermo-gravimetric analyzer, heating the material at 930°C. Gross calorific values were determined using a Parr 2040 calorimeter. Ultimate analysis was carried out using LECO C-H-N 1000 and LECO CS elemental analyzers. Major element concentrations in the lignite ashes and the inorganic sediments were determined from fused borosilicate disks using wavelength-dispersive X-ray fluorescence (XRF) spectrometry techniques.
AR1 MR1
MR2 MR3 AR2 MR4 MR5
MR6 MR7
MR8
Figure 2: Distribution of samples from Achlada and Mavropigi deposits Polished blocks of each lignite, 30 mm in diameter, were subjected to maceral analysis by point counting (500 counts) using both white reflected light and blue-light excitation, with the macerals described following the Stopes-Heerlen terminology (Taylor et al., 1998; ICCP, 2001, Sýkorová et al., 2005). The mean reflectance of ulminite in each sample was determined according to ASTM D2798-04 (ASTM, 2004). 3
Representative portions of each lignite sample were subjected to low-temperature oxygen-plasma ashing as outlined in Australian Standard 1038, Part 22 (Standards Australia, 2000). The mass percentage of low-temperature ash (LTA) was determined in each case. Each LTA and each non-coal sample was analysed by X-ray powder diffraction, with quantitative analyses of mineral phases using the Rietveld-based Siroquant™ software system (Taylor, 1991; Ward et al., 2001). The clay fraction (< 2 μm effective diameter) of each LTA sample and inorganic sediment was concentrated by ultrasonic dispersion in water with sodium hexametaphosphate (Calgon) and subsequent settling, and the nature of the clay minerals further investigated by oriented-aggregate XRD after glycol and heat treatment (Hardy and Tucker, 1988). CHEMICAL AND PETROGRAPIC PROPERTIES As indicated in Table 1, the Achlada lignite samples have higher ash yields (dry basis) than the Mavropigi materials. The Mavropigi lignites, however, have higher (as-received) moisture contents, due in part to the lesser proportion of admixed mineral material. The gross calorific value (GCV) of the Achlada lignite on a dry basis is lower than that of the Mavropigi samples, but this trend reverses if the data are converted to a dry, ash free basis. Mean reflectance of ulminite for both the Achlada and Mavropigi deposits ranges between 0.18 and 0.29%. Based on a combination of properties (Koukouzas et al., 2010a), the samples studied would be classed as medium to high ash low-rank C (or lignite C) under the International Standard for classification of coals (ISO, 2005). Table 1: Lithology, proximate analysis, calorific value and ultimate analysis data for lignite samples (after Koukouzas et al., 2010a) Sample
Lithotype
Achlada A1 matrix, MM-rich lignite A2 xylite-rich lignite A3 carbonaceous sediment A4 matrix, MM-rich lignite A5 matrix lignite Mavropigi M1 matrix lignite M2 matrix lignite M3 matrix lignite M4 matrix lignite M5 matrix lignite
M wt% (ar)
Ash wt% (db)
VM wt% (db)
Stot wt% (db)
20.9 31.9 29.3 32.1 41.3
41.9 41.4 59.6 52.4 38.0
37.7 37.2 29.6 35.5 33.6
56.7 54.5 62.7 54.8 54.9
21.1 34.8 12.9 24.6 30.6
50.4 44.2 52.5 48.8 47.7
GCV MJ/kg (db)
GCV MJ/kg (daf)
1.1 1.1 1.5 2.8 2.4
15.3 16.2 9.9 11.3 16.2
26.4 27.7 24.5 23.8 26.2
2.1 2.2 1.1 1.5 2.7
17.8 13.9 18.5 15.7 12.3
22.6 21.3 21.2 20.8 17.7
Huminite is the dominant maceral group in the lignite samples (Table 2). Detrohuminite is the most abundant huminite subgroup in the majority of the samples, suggesting significant mechanical degradation of the organic matter. Sample A2, however, a xylite-rich lithotype in the Achlada deposit, is dominated by telohuminite. Inertinite, mainly in the form of inertodetrinite but with some fusinite and semifusinite, is more abundant in the Mavropigi lignites than in the Achlada samples. The liptinite contents of both lignites are similar, with liptodetrinite, alginite and in some cases sporinite as the dominant types in most of the samples studied. Resinite is the dominant liptinite component in the xylitic lignite (A2) of the Achlada deposit. Based on the tissue preservation and gelification indices (Diessel, 1992; Kalaitzidis et al., 2004) deposition appears to have been mainly in a reed-marsh environment; only the A2 and M3 layers 4
show evidence of deposition under forest swamp conditions (Koukouzas et al., 2010a). Short dry periods caused aerobic exposure of the peat and the formation of inertinite macerals, especially in the Mavropigi deposit. Table 2: Maceral composition (vol. %, dry, mineral free) and random reflectance (Rm %) of lignite samples (after Koukouzas et al., 2010a) Sample Rm % Maceral Subgroup / Group Telohuminite Detrohuminite Gelohuminite Total Huminite Total Inertinite Total Liptinite TPI c GI d a : matrix lignite b : xylite-rich lignite
A1a 0.18
A2b 0.18
A3a 0.26
Achlada A4a A5a 0.28 0.24
M1a 0.25
M2a 0.20
M3a 0.29
Mavropigi M4a M5a 0.23 0.23
63.4 21.2 6.0 90.6 5.2 4.2 2.5 3.2
23.6 60.4 4.0 88.0 5.2 6.8 0.4 0.6
vol. %, mineral-free basis 17.6 73.8 16.8 14.0 21.4 74.2 11.0 64.0 62.4 65.4 1.2 10.4 6.0 12.8 5.2 93.0 95.2 86.8 89.2 92.0 1.0 1.0 2.0 2.0 0.6 6.0 3.8 11.2 8.8 7.4 0.2 6.0 0.3 0.3 0.3 0.1 2.1 0.6 0.7 1.0 c : Tissue preservation index d : Gelification index
24.0 48.8 8.8 81.6 6.0 12.4 0.5 1.2
24.4 42.4 7.4 74.2 14.2 11.6 0.6 0.6
29.0 41.2 12.2 82.4 6.0 11.6 0.7 1.1
MINERAL MATTER CHARACTERISTICS The ashes of the Achlada lignites (Table 3) have higher proportions of SiO2, Al2O3 and K2O than the ashes of the Mavropigi materials. The Mavropigi lignite ashes, by contrast, have relatively high proportions of CaO, and also higher proportions of SO3 retained in the laboratory ash materials. Table 3: Chemical composition of lignite ash samples, prepared at 815 °C (wt %) Sample SiO2 Al2O3 Fe2O3 TiO2 P2O5 CaO MgO Na2O K2O SO3
A1 53.1 29.1 5.21 1.02 0.15 3.15 2.68 0.08 3.05 1.9
A2 51.2 28.3 5.68 1.01 0.19 4.32 2.70 0.03 3.06 2.8
A3 53.8 29.9 5.81 0.93 0.13 2.34 3.02 0.07 3.49 1.2
Achlada A4 A5 47.2 51.7 26.5 28.6 6.45 5.78 0.92 1.04 0.25 0.26 7.28 3.48 3.08 2.92 0.06 0.03 3.08 3.27 5.7 2.5
M1 37.4 21.2 6.19 1.14 0.19 19.90 1.99 0.09 1.19 10.1
M2 45.0 20.8 6.70 1.04 0.17 13.80 3.33 0.11 1.05 7.6
Mavropigi M3 M4 17.2 40.1 9.2 17.7 7.64 6.51 0.40 0.94 0.18 0.30 40.30 21.00 7.52 4.53 0.23 0.20 0.59 1.14 11.5 7.5
X-ray diffraction studies (Table 4) indicate that the LTA of the Achlada lignites is dominated by illite and kaolinite, with minor proportions of quartz. Oriented-aggregate XRD data (Koukouzas et al., 2009) show that the LTAs of the Achlada lignites also contain significant proportions of smectite and irregular illite/smectite. The remainder of the LTA isolated from the Achlada samples is represented by bassanite (CaSO4.½H2O), probably formed from interaction of organically-associated Ca and S as the macerals were oxidised during the plasma-ashing process.
5
The LTA of the Mavropigi lignites is dominated by bassanite, or in some cases a more hydrated gypsum (CaSO4.2H2O) component. Small proportions of gypsum are also indicated by XRD analysis of the raw coal for some samples, without any low-temperature ashing process (Koukouzas et al., 2010b). This suggests that some of the Ca sulphate in the Mavropigi LTAs may represent gypsum that was present in the lignite before ashing, formed either as authigenic deposits prior to sampling or by crystallization of Ca and SO4 ions from the pore waters of the lignite during airdrying of the lignite samples. Table 4: Mineralogy of LTA isolated from lignite samples (wt %) Mineral Quartz Kaolinite Illite and I/S Chlorite Feldspar (?) Pyrite Bassanite Gypsum Whewellite
A1 5.0 46.8 40.8
A2 4.5 38.3 45.4
A3 3.7 47.1 41.7
Achlada A4 A5 4.1 7.8 41.5 44.6 37.4 38.0
M1 5.9 25.3 28.1
M2 14.6 31.6 26.5
40.6
0.5 26.7
M3 4.6 3.3
9.3 7.4
11.8
7.5
17.1
9.6
4.1 53.2 25.5
Mavropigi M4 M5 8.5 6.4 23.1 32.6 20.4 19.8 5.0 2.6 26.0 22.1
13.1 20.5
In addition to gypsum and/or bassanite, the LTA of one Mavropigi sample (M3) contains the calcium oxalate mineral whewellite (CaC2O4.H2O). The more hydrous calcium oxalate mineral weddellite (CaC2O4.2H2O) is also noted by Koukouzas et al. (2010b) from XRD studies of the raw (un-ashed) coal of this coal sample. Ca oxalates are known in other Greek coals (e.g. Kalaitzidis and Christanis, 2004), and their occurrence in the Mavropigi material probably represents an authigenic accumulation in the relevant (telohuminite-rich) lignite bed. Apart from the bassanite/gypsum and (where present) whewellite, the LTAs of the Mavropigi lignites contain kaolinite, illite, smectite and in some cases chlorite (Table 4), along with lesser but significant proportions of quartz. A trace of pyrite in one case and a small proportion of feldspar in another are also indicated. Oriented-aggregate XRD studies (Koukouzas et al., 2010b) indicate that, in contrast to the Achlada samples, kaolinite is the most abundant component of the clay fraction of the Mavropigi lignites, with illite and smectite being present in lesser and roughly equal proportions. The interbedded sediments associated with the Achlada lignites (Table 5) contain feldspar, mica and chlorite, as well as a greater proportion of quartz and a lesser proportion of kaolinite than the LTAs of the Achlada lignite samples. Most of the sediments associated with the Mavropigi lignites, however, are dominated by calcite, together with aragonite in some cases, and contain only minor proportions of illite plus traces of quartz, kaolinite and pyrite. The two exceptions (MR6 and MR8), taken from near the bottom of the seam (Figure 2), contain abundant feldspar, along with quartz and mica in one instance, and a mixture of clay mineral components. RELATION TO DEPOSITIONAL ENVIRONMENT Previous studies have shown that the Achalda lignite was formed in a fluviatile environment, mainly in the abandoned channels of a meandering river system during flooding periods (Oikonomopoulos et al., 2007). Petrographic data (Koukouzas et al. (2010a) indicate significant fragmentation of the organic matter; this is attributed to the mainly herbaceous nature of the peat-forming plants, combined with mechanical destruction of the soft plant tissues during short-term transportation. 6
Deposition took place under moderate to high water-table conditions, which resulted in the accumulation of significant proportions of inorganic material in the form of detrital silicates, mainly illite, smectite, I/S and quartz, within the palaeomire . The Mavropigi lignite deposit, by contrast, was developed in a lacustrine environment that occupied the broader area of the Ptolemais Basin during the Pliocene (e.g. Anastopoulos and Koukouzas, 1972; Kalaitzidis et al., 2004). The lithological column (Figure 2), combined with the calcareous nature of the associated sediments (Table 5), suggests an alternation of peat-forming periods with extensive flooding events and deposition of carbonate-rich sediments with abundant shell fragments. As with the Achlada deposit, the fragmented nature of the organic matter suggests formation mainly from herbaceous plant species, although the M3 layer shows derivation from mixed herbaceous and arboreal vegetation. Table 5: Mineralogy of associated rock samples (wt %) Mineral Quartz Albite Anorthite Mica Kaolinite Illite Exp clays Chlorite Calcite Aragonite Dolomite Siderite Pyrite Rutile
AR1 11.1 3.4
AR2 12.0 7.2
20.0 17.4 37.8
28.6 13.2 32.5
9.3
5.5
MR1 0.7
MR2
MR3 1.7
MR4
MR5
MR6 23.9 26.3 3.7 18.5 22.2 1.4 3.2 0.5
0.5 5.9
6.3
1.3 8.2
0.4 6.9
0.7 8.0
92.6
92.9
88.3
88.3 3.4
84.2 6.0
0.5
0.2 0.4
0.4
0.7 0.2
MR7 0.2
0.4 4.0
89.1 5.7
0.6
1.2 0.3
0.3
MR8 22.2 12.0 2.6 10.2 14.8 18.2 10.3 7.3
0.9
0.3 0.8
The greater proportions of calcium sulphate and oxalate minerals in the LTA of the Mavropigi lignites, relative to the LTA of the Achlada samples, are consistent with a lower proportion of detrital input to the Mavropigi paleomire and a greater proportion of Ca associated with the more abundant organic matter. The greater abundance of kaolinite, thought to be authigenic, in the clay fraction of the Mavropigi lignites is also consistent with a lower detrital input. Such differences may reflect the wider development of less-contaminated lacustrine environments, compared to the fluvial mires that formed the coals in the Achlada succession, where frequent flooding of the peat mire by overbank sediments seems to have taken place (Oikonomopoulous et al., 2007). SIGNIFICANCE FOR UTILISATION The technological significance of the petrographic and mineralogical properties of the coals in the two deposits is discussed more fully by Koukouzas et al. (2010a). The predominance of huminite and particularly of detrohuminite (mainly attrinite) in the lignites, for example, may contribute to a relatively low grinding energy requirement. Deviation might occur, however, for the telohuminiterich materials such as samples A2 and M3, in which more abundant telohuminite (especially ligninrich ulminite) could require more grinding energy. The high proportion of inorganic matter in the coals should also be taken into consideration, together with the nature of the inorganic fraction as expressed by the ash chemistry and the mineral matter studies. Consideration of the ash chemistry suggests that the Achlada lignite may have more 7
favorable slagging and fouling properties than the Mavropigi lignite (Koukouzas et al., 2010a), although experimental studies are required for more solid conclusions to be reached. Although abundant in the associated rocks, crystalline calcite does not occur in the actual lignite beds of the Mavropigi deposit. While some Ca may occur as crystalline sulphates (gypsum) or oxalates (weddellite) in the raw coal, or Ca in the pore waters may produce those minerals on drying, most of the relatively abundant Ca in this lignite is probably associated with the organic matter. Such material tends to be more reactive during combustion than Ca in crystalline form (Falcone and Schobert, 1986). Mine products that incorporate some of the calcareous strata along with the lignite, however, may contain separate calcite phases that react in a different way, and may also have higher carbonate CO2 contents. ACKNOWLEDGEMENTS Thanks are expressed to the General Secretariat for Research and Technology Hellas (Greece) for financial support of the project, and to Evangelia Mpali and Dimitra Papanikolaou of the Institute for Solid Fuels Technology and Applications for assistance with the research program. Thanks are also expressed to CSIRO Energy Technology, especially to David French and Owen Farrell, for assistance with XRD analysis and for high-temperature ash production, and to Irene Wainwright of the UNSW Analytical Centre for provision of the XRF analyses. REFERENCES American Society for Testing and Materials (ASTM), 2004. D2798. Standard Test Method for Microscopical Determination of the Vitrinite Reflectance of Coal. In: 2004 Annual book of ASTM Standards, section 5, Petroleum products, lubricants, and fossil fuels, 05-06, 301-305. Antoniadis P., Mavridou, E., Papazisimou, S., Christanis, K., Gentzis, T., 2006. Palaeoenvironmental conditions of the Mavropigi lignites, Ptolemais Basin, Greece: A petrological study. Energ Source Part A. 28 (4) 311-327. Diessel, C.F.K., 1992. Coal-bearing Depositional Systems. Springer Verlag, Berlin. Falcone, S.K., Schobert, H.H., 1986. Mineral transformations during ashing of selected low-rank coals. In: Vorres, K.S. (Ed.), Mineral Matter and Ash in Coal. American Chemical Society, Washington, DC, 114-127. Hardy, R.G., Tucker, M.E., 1988. X-ray powder diffraction of sediments. In: Tucker, M.E. (Ed.), Techniques in Sedimentology. Blackwell, Oxford, UK, pp. 191– 228. International Committee for Coal and Organic Petrology (ICCP), 2001. The new inertinite classification (ICCP System 1994). Fuel 80, 459-471. International Organisation for Standardisation (ISO), 2005. Classification of coals. International Standard ISO 11760:2005(E), International Organisation for Standardisation, Geneva, Switzerland, 9 pp. Kalaitzidis, S., Christanis, K., 2004. Mineralogical composition of solid residues derived from wet and thermal oxidation of peat. Mineral Wealth, 132, 7-16 [in Greek, with English abstract]. Kalaitzidis, S., Bouzinos, A., Papazisimou, S., Christanis, K., 2004. A short-term establishment of forest fen habitat during Pliocene lignite formation in the Ptolemais basin, NW Macedonia, Greece. Int. J. Coal Geol., 57/3-4, 243-263.
8
Kavouridis, K., Koukouzas, N., 2008. Coal and sustainable energy supply challenges and barriers. Energ. Policy 36, 693-703. Koukouzas, N., 2007. Mineralogy and geochemistry of diatomite associated with lignite seams in the Komnina Lignite Basin, Ptolemais, Northern, Greece, Int. J. Coal Geol. 71, 276 – 286. Koukouzas, N., Ward, C.R., Papanikolaou, D., Li, Z., 2009. Quantitative evaluation of minerals in lignites and intra-seam sediments from the Achlada Basin, northern Greece. Energ. Fuel. 23, 21692175. Koukouzas, N., Kalaitzidis, S.P. and Ward, C.R., 2010. Organic petrographical, mineralogical and geochemical features of the Achlada and Mavropigi lignite deposits, NW Macedonia, Greece. International Journal of Coal Geology (in press). Koukouzas, N., Ward, C.R., Li, Z., 2010b. Mineralogy of lignites and associated strata in the Mavropigi field of the Ptolemais Basin, northern Greece. Int. J. Coal Geol. 81, 182-190. Oikonomopoulos, I., Perraki, Th., Kaouras, G., Antoniadis, P., 2007. Mineralogical study of inorganic intercalated seams at Achlada lignite deposits (NW Greece). Proceedings, 9th International Geological Congress, Geological Society of Greece, XXXX, 906-917. Standards Australia, 2000. Higher rank coal - mineral matter and water of constitution. Australian Standard 1038.22, pp. 20. Sýkorová, I., Pickel, W., Christanis, K., Wolf, M., Taylor, G.H. and Flores, D., 2005. Classification of huminite – ICCP System 1994. Int. J. Coal. Geol., 62, 85-106. Taylor, G.H., Teichmüller, M., Davis, A., Diessel, C.F.K., Littke, R., Robert, P., 1998, Organic Petrology. Gebrüder Borntraeger, Berlin-Stuttgart. Taylor, J. C., 1991. Computer Programs for Standardless Quantitative Analysis of Minerals Using the Full Powder Diffraction Profile. Powder Diffr. 6, 2-9.Sýkorová et al., 2005 Ward, C.R., Matulis, C.E., Taylor, J.C., Dale, L.S., 2001. Quantification of mineral matter in Argonne Premium Coals using interactive Rietveld-based X-ray diffraction. Int. J. Coal Geol. 46, 6782.
9
X-Ray computer tomography on coal particles
Patrick J. Masset 1,a, Heikki Suhonen b
a)
Freiberg University of Mining and Technology
Centre for Innovation Competence Virtuhcon – Group “Multiphase Systems” Fuchsmühlenweg 9, Reiche Zeche D-09599 Freiberg, Germany
b)
European Synchrotron Radiation Facility
BP 220, F-38043 Grenoble Cedex 9, France
Abstract Computer tomography was used to investigate the local microstructure of coal particles. For German lignite grade coal four different areas have been observed. The shape, size and distribution of pores and minerals have been evidenced and it provides quantitative data for further modelling of coal gasification of single particle. In addition, 3D-representation of particles was achieved using the images of the single slices recorded along the tomographic axis.
1
Corresponding author. E-mail: [email protected]
Tel/Fax: +49 3731 39 4810/4555
Introduction Coal is used as feedstock in gasification processes. It is known to be one of the world’s most chemically complex natural mixtures. The coal composition and morphology depends on the environmental conditions during its formation, its geographical origin and the mining site. The main physical components of coal include the coal matrix (organic components), the mineral matter (inorganic constituents) and the void-space (volume occupied by pores and fractures). The pores and the fractures are the important flow pathways for the volatiles and also for the environmental gas during the gasification of coal particles in gas flowing conditions. The permeability may be reduced when the void-space is filled with minerals. All these parameters influence significantly the coal reactivity. Therefore, the quantification of the volume and spatial repartition of the pores, fractures and mineral matter in coals with a high resolution (100 nm) represents a primary and fundamental requirement for the modelling of the coal particle gasification. Three methods are usually used to characterize the pore and the fracture in coals: Hg porosimetry, BET analysis and microscopic observations. However, none of them is well suited for evaluating the internal coal heterogeneity of the particles in three dimensions (3D). The results may lead to misleading information on the pores and fractures. In addition, these techniques are destructive to the sample. X-ray computed tomography (CT) is a non destructive technique, which was used successfully to provide quantitative detection of internal structure of rocks in 3D. Recently, CT techniques have shown good prospects for application in coal petrology. Petro-physical researchers have shown the feasibility in differentiating pores, fractures and minerals from the coal matrix [1]. CT is a truly 3D technique, which may provide quantitative detection of internal structure of coal particles for different gasification stages. X-ray phase contrast images with a resolution close to 100 nm were recorded. The reconstruction of the 3D pictures of coal particles provides an estimation of the pore size and their distribution. In addition, due to the difference of density between the coal matrix and the mineral matter, the location and shape of mineral
aggregates may be estimated. Chemical and physical analyses complete the morphological analysis. These data are used to support modelling of coal particle transformation.
1. Experimental 1.1 Materials For the tomography experiments a gasification grade German Rhenish Lignite coal was used. It was dried and milled in a mortar. It was sieved using Tyler screen series to recover coal particles in the size range 100-200μm. Single particles were selected to be investigated and stuck on a glass pin.
1.2 Apparatus Images were recorded at ESRF beamline ID22-NI. The x-rays were focused with dynamically bent multilayer coated KB mirror pair giving a spot size of 100 nm (FWHM) and a bandwidth of 2% at 17.5 keV mean energy. The x-ray focus was used as a secondary source for x-ray projection microscopy (Figure 1) giving the magnification of M = (z1+z2)/z1 [2,3]. X-rays were converted to visible light by an LSO:Tb scintillator [4] which was coupled to a CCD chip by a microscope giving the physical pixel size of 0.96 µm. In the present case z1+z2=525 mm and z1=32.7 mm gave the effective pixel size of 60 nm. X-ray propagation gave strong phase contrast fringes, and a phase retrieval algorithm using images recorded at 4 different distances z1 was applied to recover the x-ray phase change in the sample [5].
Figure 1: Schematic picture of the x-ray imaging setup. Focused x-rays come from the left, and the sample is placed downstream of the focus at position z1. Magnification by a geometrical factor M = (z1+z2)/z1 is obtained. Four different distances z1 are used (while keeping z1+z2 constant) to obtain sufficient data for phase retrieval.
2. Results and discussion 2.1 Coal characterisation The chemical analysis was carried out on a representative sample (100 g) in order to avoid effect of inhomogeneity of the coal sample. Table 1 and Table 2 summarise the results of the ultimate and proximate analyses of the coal sample performed according standard procedures. The values obtained are classical for a lignite type coal.
Element
Composition / mass %
C
55.76
H
4
O
18.62
N
0.68
S
1.02
H2O
12
Ash
7.92
Total
100
Table 1: Composition of German Rhenish Lignite gasification grade coal.
Amount / mass % Fixed Carbon
35.44
Volatile Matter
44.64
Water
12
Ash
7.92
Total
100
Table 2: Properties of German Rhenish Lignite gasification grade coal.
2.2 Computer tomography analysis For a single particle a volume of 1500 x 1500 x 1500 voxels was recorded at voxel size of (60 nm)3. The general features of the morphology through a single slice are detailed as shown in Figure 2. Four regions of different density can be indentified. The darkest part in the picture is ascribed to dense elements in the sample (referred as point 1). It looks to have some type of crystalline shape. These could be impurities of some metal inside the coal particle (minerals). The almost white colour regions are voids (labelled as point 2). Then two main levels of gray
were observed, a darker one and a pale gray, identified as points 3 and 4 in Figure 2, respectively. These could be two different phases or one pure carbon and the other one carbon combined with another element. Another possibility is that the lighter region contains very small pores which are not yet resolved in the images, and then the apparent average density is smaller than in the other region without these pores.
Figure 2. Typical computer tomography image obtained on a coal particle. Different types of regions are labelled with numbers as explained in the text.
Figure 3 shows single slices at different locations along the tomographic axis at intervals of 4.5 µm. It gives an overview of the size and shape changes of the pores and also the random distribution of minerals. The pores exhibit mainly two shapes. The first type looks like crevices progressing more or less inside the core of the particle whereas the second type is spherical. The first type of voids may offer some pathways inside the coal particle for the gas evolution during the gasification process. These pores are regularly located along or around
the dark gray particles. On the other hand, the spherical pores are less numerous. They offer well defined contours whereas the interface with the matrix of crevice type pores is tortuous. In the dark gray areas some small spherical pores were observed and are located in colony type arrangement. This feature is important for the evolution of the pore size during the gasification process: do the pores remain individual or do they collapse to form larger pores. These preliminary results provide a good estimation of the shape and the size of pores in coal particles. In addition, the distribution and the volume fraction can be derived from the picture analysis. The information represents a valuable source of data on the local microstructure of coal particles for the modelling of gasification conditions of single particles. Figure 4 shows the evolution of the shape of the observed dark particles which were ascribed to minerals. Pictures were taken at an interval of 10 slices which represents a distance of 0.6 µm. They exhibit either irregular shapes like interconnected needles (size close to 10 µm) surrounded by pores or some small nodules (size close to 1 µm) in the pale gray areas. From the individual slice pictures a 3D-representation of the needle shape was obtained (Figure 5) and it exhibits a high tortuosity. In addition, the picture evidences that some the needle like particles offer a high specific surface due to their complex shape and should be highly reactive with their environment. In addition, the presence of some smaller mineral particles at this particle was observed and it may be due to fragmentation during the coal particle formation. The size of the observed nodules dispersed in the matrix is very small and less than 1 µm. Due to their small size they may be easily disperse in the environment during the pyrolysis step and act as catalyst with the gas phase. At intermediate temperatures (500-800 °C) these nodules or fragments may be released from the carbon matrix during the gasification process and may stick and react with the molten slag. These measurements provide a set of data about the repartition, size of minerals in coal particles. Due to their small size the classical law of thermodynamics may be unusable to describe their properties and the specific surface should be included in the expression of the free Gibbs energy to evaluate correctly the thermodynamic properties of the minerals in gasification processes.
Figure 3. Internal structure of a coal particle at different locations along the tomographic axis (from left to right and top to bottom).
Figure 4. The shape of a mineral particle located in the coal at several different slices (from the left to right and top to bottom).
Figure 5. 3D-representation of the observed needle type mineral in the coal particle.
3. Conclusions This work shows that computer tomography is a well suited technique to investigate the local microstructure of coal particles. For German lignite grade coal four different areas have been observed. The shape, size and distribution of pores and minerals have been assessed and it provides quantitative data for further modelling of coal gasification of single particle. However, it has been shown that the structure of the coal matrix may need further study to investigate fine pores in the carbon matrix.
Acknowledgements The financial support from BMBF (Federal Ministry of Education and Research) and SMWK (Saxon State Ministry for Higher Education, Research and the Fine Arts) is gratefully acknowledged.
References [1] C.Ö. Karacan et al., J. Coal Geol. 72, 209 (2007). [2] Mokso R., Cloetens P., Maire E., Ludwig W., and Buffiere J.-Y., Appl. Phys. Lett., 90, 144104 (2007). [3] P. Bleuet et al., Rev. Sci. Instr. 80, 056101 (2009) [4] T. Martin et al., IEEE Trans. Nucl. Sci. 56, 1412 (2009) [5] P. Cloetens et al., Appl. Phys. Lett. 75, 2912 (1999)
1
Mineralogy, Geochemistry, and Petrography of Upper Permian Bituminous and Carboniferous Anthracite Coals from Xuanwei County, Eastern Yunnan Province, China Harvey E. Belkin U.S. Geological Survey, 956 National Center, Reston, VA 20192 USA [email protected] James C. Hower and Jordan W. Drew University of Kentucky Center for Applied Energy Research, 2540 Research Park Drive, Lexington, KY 40511 USA [email protected]; [email protected] Linwei Tian School of Public Health and Primary Care, Chinese University of Hong Kong, Rm 412, Prince of Wales Hospital, Shatin-NT, Hong Kong, People's Republic of China [email protected]
INTRODUCTION: Certain communes in Xuanwei County, eastern Yunnan Province, have some of the highest lung cancer mortality in China. Age-adjusted mortality rates (per 100,000) for Chinese males and females nationwide are 6.8 and 3.2, respectively (Mumford et al., 1987). The rates for Yunnan Province are males 4.3 and females 1.5. In Xuanwei County, the rates are males 27.7 and females 25.3 and in the three communes with the highest mortality, the rates are males 118.0 and females 125.6 (Mumford et al., 1987). This exceptionally high rate and the similarity between the male and female rate was unusual as the females in this locality are essentially non-smokers. Early workers attributed the high incidence of lung cancer to domestic combustion of locally mined coal in houses with un-vented stoves (Chapman et al., 1988). The communes using smoky coal (bituminous) accounted for more than 90% of the lung cancer cases for both men and women; whereas those communes using smokeless coal (anthracite) had a lower incidence of lung cancer. The relationship of polycyclic aromatic hydrocarbon (PAH) generation with combustion and lung cancer was first thought to be the disease etiology, but more recent work suggests that the presence of nano-sized quartz particles may be a major factor (Tian et al., 2008). Recent studies of similar-age coals have identified the chlorite, chamosite, as cell-fillings (Dai and Chou 2007; Dai et al. 2008), but conclude that it is not a causal factor in lung cancer. Examination of twelve bituminous coals of the Upper Permian Xuanwei Formation (correlative with the Longtan Formation) that represent smoky coal is the main focus of our study. We also have examined three Carboniferous anthracite (smokeless) coals of the Weining Formation. In this report, we present preliminary data on organic petrography, vitrinite reflectance, proximate and ultimate analysis, and major-, minor- and trace-element chemistry. Scanning electron microscopy and electron microprobe wave-length dispersive spectrometry have been used to characterize the mineral chemistry and mode of occurrence of trace elements.
2 The bituminous coals and related clastic sedimentary rock members of the Xuanwei Formation are a fresh-water lake-swamp facies inboard of the more marine-influenced Longtan Formation of Guizhou Province. In eastern Yunnan Province, the Xuanwei Formation was deposited on the Permian Emeishan basalts (Peng et al., 2005; He et al., 2007).
Figure 1: Location map showing the bituminous (solid circle), anthracite (solid square) coal mines, and identifying code in Xuanwei County, Yunnan Province, People's Republic of China. Xuanwei County is a county-level municipality in Qujing
3 Prefecture and is divided up into various communes; the Xuanwei metropolitan area is shown. Table 1 gives details of the mines and samples. SAMPLES AND METHODS: Representative splits of coal collected as part of an earlier study (Mumford et al., 1987) were prepared as 2.5 cm pellets and 1 cm blocks and polished for optical and electron beam petrography and analysis. Additional coal splits were powdered and submitted for major, minor, and trace element analysis to Johnson Space Center, Houston, TX, Activation Laboratories, Ontario, Canada, and the U.S. Geological Survey, Denver, CO. Proximate, ultimate, and calorific value were determined by Commercial Testing & Engineering Co., Lombard, IL (Mumford et al., 1987). Figure 1 shows the sample locations and Table 1 gives the sample and mine names. Petrographic studies were conducted on polished resin-bound particulate pellets using Leitz microscopes, with reflected-light, oil-immersion techniques and a final magnification of 500x, at the Center for Applied Energy Research, Lexington, KY. Scanning electron microscope backscattered electron imaging (SEM-BSE) and energy dispersive analysis (SEM-EDS) and electron microprobe analysis (EPMA) were conducted at the U.S. Geological Survey, Reston, VA. Quantitative electron microprobe analyses of major and minor elements were obtained with a JEOL JXA-8900R five spectrometer, fully-automated electron microprobe using wavelengthdispersive X-ray spectrometry. Analyses of the mineral phases were made at 15 or 20 keV accelerating voltage,10 or 30 nA probe current, and counting times on peak of 10 to 120 s, using a 1-, 5-, and 10-µm diameter probe spot for sulfides, layered silicates (chlorite, kaolinite), and carbonate, respectively. Natural and synthetic standard reference materials were used. [Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.] RESULTS: Organic Petrography The coals have ranks (Table 1) ranging from high-volatile A bituminous (Hongchong) to semianthracite (Reshui, Luoshui, and Xize). Vitrinite reflectance (Rmax%) ranges from 0.98 to 1.74% in the bituminous coals and from 2.26 to 2.77% in the anthracite coals (Table 1). Maceral analyses show that the bituminous coals are predominantly inertinite (collotelinite, fusinite, semifusinite, and suberinite) with lesser amounts of vitrinite (telinite, vitrodetrinite, micrinite, macrinite) and minor liptinite. A summary of the maceral content is given in Table 1. A number of Chinese researchers have noted the presence of a different form of liptinite, termed “barkinite” in the Chinese literature (Sun, 2003; Sun, 2005; Sun and Horsfield, 2005). Despite their attempts to place “barkinite” in a distinct maceral category, it has not been recognized as a distinct maceral by the International Committee for Coal and Organic Petrology. Overall, it has the appearance of the liptinite macerals cutinite and/or suberinite (Hower et al., 2007). Figures 2A & 2B show some examples of “barkinite” from the high volatile A /medium volatile bituminous Taoyuan and Yangchang coals. The “barkinite” has a higher reflectance than the examples in Hower et al. (2007) due to the higher rank of the coals in this study. Fusinite is
4 present as relatively massive forms (Figures 2C & 2D) and with calcite-filled cell lumens and calcite cementation of broken fusinite (Figure 2E). Anthracite coals are mainly vitrinite (collotelinite, semifusinite, fusinite and vitrodetrinite) that exceeds inertinite in abundance and the content of mineral matter is higher than in most of the bituminous coals (Table 1). Table 1: Sample name, mine name, mine code, vitrinite reflectance (%), rank, and maceral summary (%). Sample
Mine name
Mine code
vitrinite reflectance (%)
Rank
Rmax
s.d.
Rrandom
s.d.
maceral summary (%) vitrinite
inertinite
liptinite
mineral matter
L-1-17-H
Hongchong
HC
0.98
0.07
0.92
0.07
hvAb
28.2
41.0
2.2
28.6
L-2-5-G
Guanyun
LB
1.09
0.06
1.01
0.07
mvb
31.0
50.0
1.6
17.4
L-3-9-T
Taoyuan
TY
1.12
0.05
1.06
0.07
mvb
21.2
59.8
8.6
10.4
T-4-1-Y
Yangchang
YC
1.11
0.09
0.98
0.08
mvb
28.8
49.4
11.4
10.4
T-5-10-T
Tianba
TB
1.12
0.08
1.06
0.09
mvb
22.6
54.2
14.4
8.8
B-6-4-B
Baoshan
BS
1.15
0.08
0.97
0.07
mvb
34.2
36.2
2.0
27.6
L-7-8-L
Longli
LL
1.15
0.09
1.00
0.11
mvb
41.2
44.8
2.0
12.0
T-8-5-J
Jiupu
JP
1.24
0.05
1.18
0.06
mvb
28.2
46.6
7.0
18.2
Y-9-7-L
Lanba
LN
1.42
0.06
1.32
0.07
mvb/lvb
28.2
52.6
0.2
19.0
Y-10-4-H
Heping
HP
1.20
0.10
1.11
0.12
mvb
46.2
42.6
0.0
11.2
Y-11-4-F
Fangmaping
FM
1.17
0.12
1.05
0.12
mvb
26.8
59.6
1.8
11.8
W-12-4-W
Wenxing
WX
1.74
0.09
1.57
0.11
lvb
56.0
40.8
0.0
3.2
S-13-4-R
Reshui
RS
2.77
0.13
2.49
0.24
sa
56.8
18.6
0.0
24.6
S-14-7-L
Luoshui
LS
2.26
0.10
2.14
0.13
sa
22.6
8.2
0.0
69.2
S-15-10-X
Xize
XZ
2.57
0.18
2.36
0.17
sa
39.0
23.4
0.0
37.6
s.d. = 1 sigma standard deviation hvAb = high volatile A bituminous, mvb = medium volatile bituminous lvb = low volatile bituminous, sa = semi-anthracite
Mineralogy Mineral matter in the bituminous coals, especially authigenic cell-fillings, shows a wide variety of mineral species and textures. Common cell-fillings are chlorite, kaolinite, quartz, and calcite (Figures 3A, B, C, D) and less commonly TiO2, chalcopyrite (CuFeS2), and clausthalite (PbSe); pyrite is uncommon (Figure 3B). The chlorite species is chamosite with Mg/(Mg+Fe) that ranges from 0.08 to 0.32 and MnO less than 0.2 wt.%. Calcite is low MgO (< 1 wt%) and has total MnO and FeO contents < 6 wt.%. Kaolinite and various sulfides are stoichiometric. Detrital mineralogy in the bituminous coals consists of quartz, zircon, apatite, barite, TiO2, ilmenite, monazite; fractures are filled by quartz, calcite, clay, and iron sulfide. Mineral-growth textures in the authigenic cell-fillings of the bituminous coals define the passage of fluids of variable
5 composition that were mostly silica and carbonate saturated, and in some cases, Fe-rich. Both the Fe-rich siliceous fluids and the organic coal constituents were low sulfur and did not form
Figure 2: Reflected-light photomicrographs of macerals. A. "Barkinite" with collotelinite and inertodetrinite (not labeled); sample L-3-9-T Taoyuan, B. "Barkinite" through center of image, with inertinite; sample T-4-1-Y Yangchang, C. Fusinite (brighter areas) with (possible) corpogelinite (darker cell fillings); sample L11-4-F Fangmaping, D. Fusinite; sample Y-9-7-L Lanba, E. Fusinite (f) and secretinite (s), with macrinite (m), collotelinite (c), and calcite cement within fusinite/secretinite area; sample Y-10-4H Heping.
6 composition that were mostly silica and carbonate saturated, and in some cases, Fe-rich. Both the Fe-rich siliceous fluids and the organic coal constituents were low sulfur and did not form abundant iron sulfides. Deformation and cross-cutting textures suggests that the introduction of cell-filling fluids was accompanied by moderate compression of some parts of the coal (Figure 3D). Anthracite detrital mineral matter consists mainly of clay and sulfides, with quartz, zircon, TiO2, Fe-oxide, pyrite, and calcite; pyrite framboids also are present.
Figure 3. BSE-SEM of bituminous coal showing various cellfilling textures and mineralogy. A. Calcite (c) and kaolinite (K) cell-fillings; sample T-5-10-T, B. Calcite cell-fillings with uncommon pyrite crystals (arrows); sample L-2-5-G, C. Chlorite (chl) and quartz (qtz) cell-fillings; sample L-2-5-G, D. Deformed cells filled with chlorite and quartz. sample L-2-5-G.
7 Geochemistry The bituminous coals on an as-received basis are lower sulfur (0.1 to 0.4 wt.%) and have ash yields of 13 to 32 wt.%, compared to anthracite coals which have higher sulfur (3 to 5 wt.%) and somewhat higher ash yields of 30 to 38 wt.%. A summary of the proximate and ultimate data is given in Table 2.
Table 2: Proximate and ultimate data summary. Sample
Moisture Ash 750 °C Volatile matter
Fixed carbon
Hydrogen
Carbon
Nitrogen
Sulfur Oxygen
Calorific value
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
Btu/lb kilocalorie/kg
L-1-17-H
0.83
28.8
20.7
50.5
3.35
63.4
1.02
0.34
3.09
10803
2722
L-2-5-G
0.73
28.0
21.3
50.7
3.17
63.5
1.03
0.29
4.07
10940
2757
L-3-9-T
1.08
26.9
19.7
52.4
3.23
64.8
1.05
0.43
3.36
10967
2764
T-4-1-Y
0.89
13.4
26.5
60.1
4.08
76.7
1.49
0.13
4.17
13371
3369
T-5-10-T
1.03
16.3
27.7
56.0
4.14
73.5
1.46
0.16
4.50
12827
3232
B-6-4-B
0.94
21.4
22.8
55.8
3.57
69.8
1.23
0.20
3.81
11995
3023
L-7-8-L
0.90
21.6
21.6
56.8
3.48
69.4
1.23
0.16
4.06
12071
3042
T-8-5-J
0.58
32.0
19.7
48.4
3.15
60.6
1.03
0.12
3.08
10515
2650
Y-9-7-L
0.82
16.3
19.5
64.3
3.74
74.9
1.24
0.23
3.60
13039
3286
Y-10-4-H
0.82
31.1
15.0
53.9
2.93
61.4
0.96
0.12
3.42
10456
2635
Y-11-4-F
0.75
20.8
22.9
56.3
3.59
70.8
1.11
0.16
3.59
12188
3071
W-12-4-W
0.71
17.7
15.1
67.3
3.58
74.8
1.11
0.13
2.76
12721
3206
S-13-4-R
2.03
30.1
14.9
55.0
2.49
58.4
0.69
2.86
5.42
9468
2386
S-14-7-L
0.91
35.2
22.3
42.4
2.54
53.6
0.64
5.38
2.60
8793
2216
S-15-10-X
6.38
37.5
20.3
42.2
2.18
47.5
0.58
3.80
8.42
7803
1966
Major- and minor-elements analyzed and reported on an as-determined ash basis (Table 3) reflect the variable mineralogy in both the bituminous and anthracite coals. The presence of calcite as cell-fillings is shown by CaO values generally greater than 10 wt.%; whereas, the amount of iron (reported as Fe2O3) is a measure of the chlorite content, as pyrite is uncommon in the bituminous coals. The sub-equal abundance of silica and alumina in the anthracite coals is
8 consistent with the preponderance of clay as mineral matter. Also, this is the case for bituminous coal sample W-12-4-W.
Table 3: Major and minor elements on an as-determined ash-basis. Sample
SiO2
TiO2
Al2O3
Fe2O3*
MgO
CaO
Na2O
K2O
P2O5
SO3
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
wt.%
L-1-17-H
62.6
0.8
15.5
16.4
1.1
2.3
0.1
0.3
<0.02
0.9
L-2-5-G
65.0
0.5
10.6
17.7
1.0
3.7
0.1
0.1
0.2
1.0
L-3-9-T
58.7
0.5
11.1
15.4
1.2
10.5
0.1
0.2
0.1
2.2
T-4-1-Y
62.2
1.3
13.7
8.9
1.2
9.5
0.1
0.2
0.1
2.9
T-5-10-T
64.2
0.8
20.2
5.6
0.9
6.4
0.1
0.3
<0.02
1.4
B-6-4-B
58.8
0.4
19.4
7.0
1.4
10.4
0.1
0.3
<0.02
2.1
L-7-8-L
73.1
0.5
11.7
11.4
0.9
1.8
0.1
0.2
0.0
0.3
T-8-5-J
73.5
0.6
9.8
11.1
0.9
2.8
0.1
0.2
0.1
1.0
Y-9-7-L
75.1
0.7
12.6
8.0
0.9
1.7
0.1
0.2
0.1
0.7
Y-10-4-H
66.7
0.7
11.6
5.4
0.6
12.9
0.1
0.2
<0.02
1.8
Y-11-4-F
62.0
0.5
11.1
13.9
1.0
10.1
0.1
0.2
<0.02
1.1
W-12-4-W
48.9
1.3
42.2
5.5
0.4
0.1
0.2
0.9
0.0
0.4
S-13-4-R
47.7
1.5
42.9
6.2
0.4
<0.02
0.2
0.6
<0.02
0.5
S-14-7-L
47.4
1.9
43.7
5.3
0.5
0.1
0.1
0.5
0.2
0.3
S-15-10-X
44.4
1.3
34.3
18.7
0.2
<0.02
0.1
0.6
0.1
0.2
* = total iron as Fe3+ < = below detection limit at that value
Minor- and trace-elements in both the anthracite and bituminous coals were determined in order to explore the possibility that high concentrations of trace elements deleterious to human health were a causal factor for the high lung cancer rates in Xuanwei County. We calculated the geometric mean for both types of coals, but we are more interested in the bituminous coals as they are used in the areas of high lung cancer incidence. Table 4 shows that the bituminous coal geometric means compared to the geometric means for U.S. coal, none of the elements are in excess concentration, i.e., greater than one order of magnitude, except manganese. Some elements known to be hazardous to human health related to domestic combustion of coal (Finkelman et al., 1999), are low, such as As and Hg. Some potentially hazardous elements, such as Be, Cr, Ni, Pb, Se, and Th whose bituminous geometric means exceed that of U.S. coal (Finkelman 1993) are higher in the anthracite coals. Communes burning anthracite coals do not have an excess lung cancer rate suggesting that these elements are not a casual factor in those areas burning bituminous coal. Manganese in the bituminous coals, due to its presence in calcite, is the one hazardous air pollutant element (HAPS) that has an high abundance relative to the U.S. mean.
9 Table 4: Minor and trace-element data on a whole coal, asdetermined basis and a comparison with average U.S. coal. Sample
As
Ba
Be
Co
Cr
Cu
Ge
Hg
Mn
Ni
Pb
Sb
Se
Th
U
V
Y
Zn
Zr
ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm L-1-17-H
1.3 41.7
0.9 23.2 26.5 68.1
1.5 0.04 201 35.1 27.9
0.8
9.2
6.8
2.8 102 11.7 54.3 135
L-2-5-G
1.5 15.8
1.0 28.8 14.8 26.8
1.9 0.14 258 28.0 17.0
0.6
6.7
3.7
3.1
36 24.3 63.4
39
L-3-9-T
0.9 11.6
1.4 21.3 11.2 25.8
1.9 0.04 372 18.8 16.3
0.6
5.7
4.4
1.2
32 14.7 48.3
32
T-4-1-Y
0.9 23.4
0.9 18.4 15.3 21.1
3.6 0.05 304 11.7 10.3
0.5
4.4
4.2
0.9
35
7.1 39.9
51
T-5-10-T
0.7 17.8
1.2
3.1 0.06 188
9.7 10.1
0.5
4.3
5.7
1.4
22
9.9 14.0
41
B-6-4-B
1.7 38.1
2.0 20.2 10.0 32.9
3.6
0.3 634 24.2 15.8
0.7
5.4 11.5
2.8
62 16.8 47.9
61
L-7-8-L
0.5 10.1
1.4 24.3 11.4 30.2
1.1 0.03 107 28.3 15.0
0.9
5.6
3.2
0.9
29
9.0 49.4
31
T-8-5-J
0.9 20.6
1.3 22.0 14.5 32.8
1.7 0.05 244 26.3 17.8
0.5
5.8
4.4
2.3
58 14.9 40.1
52
Y-9-7-L
0.8 27.0
1.6 21.3 16.3 69.7
2.1 0.04
65 18.4 12.4
0.7
5.1
2.8
1.2 179
9.9 34.0
60
Y-10-4-H
0.9 22.8
2.2 18.5 16.4 43.5
2.0 0.05 404 14.2 11.6
0.2
6.5
2.3
0.5 148
9.2 37.5
36
Y-11-4-F
1.4 27.9
2.6 24.3 17.4 48.6
2.9 0.06 512 22.3 16.5
0.6
8.3
4.1
1.1
89 13.2 60.7
35
W-12-4-W
0.6 15.0
2.5 17.3
9.4 21.7
3.2 0.03
22
4.2
0.2
3.7
3.6
0.8
35 30.0
85
S-13-4-R
4.4 16.8
1.7
6.2 52.7 21.9
1.0 0.11
14 17.4 15.7
0.5
4.6
9.8
4.5
93 20.8 11.3 123
S-14-7-L
7.5 40.2
3.0 14.1 66.5 21.4
1.8 0.14
23 47.1 20.0
0.8
3.8 22.5
4.1
98 22.2 17.4 237
S-15-10-X
14.6 13.0
2.9 16.2 68.3 55.8
1.6 0.19
8 57.8 46.5
1.4
5.4 18.3
7.3 124 41.6 19.2 145
0.9 20.8
1.5
2.2 0.06 204 18.9 13.4
0.5
5.7
1.3
7.8 20.6
2.5 11.2 62.1 29.7
1.4
13 36.1 24.4
0.8
4.6 15.9
5.1 105 26.8 15.6 162
6.5
1.3
0.6 0.10
19
0.6
1.8
1.1
geometric mean bituminous geometric mean anthracite U.S. coal*
93
9.1
7.1 10.3
20 13.4 31.8
3.7
10
12
0.1
9.5
9
5
4.3
1.7
9.2
55 13.0 36.8
17
6.6
13
50
19
* = U.S. coal geometric mean data from Finkelman (1993)
DISCUSSION and CONCLUDING STATEMENT: The goal of this preliminary report was to see if the coals burned in those communes with the high incidence of lung cancer were noticeably different from coal from the same formation burned in communes with a lower incidence of lung cancer or different from the anthracite used by communes in the southwest part of Xuanwei County. To date, from the standpoint of our study of inorganic chemistry, mineralogy and organic petrology, we see no significant differences among the studied coals from different communes of variable lung cancer incidence. Manganese is the only HAPS element significantly enriched above average U.S. coal concentration (Finkelman, 1993). Although lung toxicity associated with Mn has been noted (Crossgrove and Zheng, 2004), the major toxic response of Mn overexposure is to the central nervous system. We believe that if Mn is a problem in those homes using coal for domestic combustion, neurotoxicity would have been reported.
10 Tian et al. (2008) suggest that nanoquartz emitted from coal combustion is the cause of the high incidence of lung cancer. Coal from mines HC, LB, and TY (Table 1) are used in communes with high lung cancer rates. Our preliminary observations do not show any significant difference related to quartz between these coals and those in areas of low lung cancer rates, such as coal from mines BS, TB, and WX and the anthracite coals (Table 1). Further work on these coals will involve low-temperature ashing technique and the examination of the residue with microbeam techniques in order to assess potential differences in quartz particle size and shape. REFERENCES: Chapman, R. S., Mumford, J. L., Harris, D. B., He., Z. Z., Jiang, W. Z., and Yang, R. D., 1988. The epidemiology of lung cancer in Xuan Wei, China: current progress, issues, and research strategies. Archives of Environmental Health, 43: 180-185. Crossgrove, J., and Zheng, W., 2004. Manganese toxicity upon overexposure. NMR in Biomedicine, 17: 544-553. Dai, S., and Chou, C-L., 2007. Occurrence and origin of minerals in a chamosite-bearing coal of Late Permian age, Zhaotong, Yunnan, China. American Mineralogist, 92: 1253-1261. Dai, S., Tian, L., Chou, C-L., Zhou, Y., Zhang, M., Zhao, L., Wang, J., Yang, Z., Cao, H., and Ren, D., 2008. Mineralogical and compositional characteristics of Late Permian coals from an area of high lung cancer rate in Xuan Wei, Yunnan, China: Occurrence and origin of quartz and chamosite. International Journal of Coal Geology,76, 318-327. Finkelman, R.B., 1993. Trace and minor elements in coal. In.: Organic Geochemistry, eds. Engel, M.H. and Macko, S.A., Plenum Press, New York, p. 593-607. Finkelman, R. B., Belkin, H. E., and Zheng, B., 1999. Health impacts of domestic coal use in China. Proceedings of the National Academy USA, 96: 3427-3431. He, B. Xu, Y-G., Huang, X-L., Luo, Z-Y., Shi, Y-R., Yang, Qi-J., and Yu, S-Y., 2007. Age and duration of the Emeishan flood volcanism, SW China: Geochemistry and SHRIMP Zircon U-Pb dating of silicic ignimbrites, post-volcanic Xuanwei Formation and clay tuff at the Chaotian section. Earth and Planetary Science Letters, 255: 306-323. Hower, J. C., Suárez-Ruiz, I., Mastalerz, M., and Cook, A. C., 2007. The investigation of chemical structure of coal macerals via transmitted-light FT-IR microscopy by X. Sun. Spectrochimica Acta. Part A: Molecular and Biomolecular Spectroscopy, 67(5): 1433-1437. Mumford, J. L., He, X. Z., Chapman, R. S., Cao, S. R., Harris, D. B., Li, X. M., Xian, Y. L., Jiang, W. Z., Xu, C. W., Chuang, J. C., Wilson, W. E., and Cooke, M., 1987. Lung cancer and indoor air pollution in Xuan Wei, China. Science, 235: 217-220.
11 Peng, Y., Zhang, S., Yu, T., Yang, F., Gao, Y., and Shi, G.R., 2005. High-resolution terrestrial Permian-Triassic eventostratigraphic boundary in western Guizhou and eastern Yunnan, southwestern China. Palaeogeography, Palaeoclimatology, Palaeoecology, 215:285-295. Sun, Y.Z., 2003. Petrologic and geochemical characteristics of “Barkinite” from the Dahe mine, Guizhou Province, China. International Journal of Coal Geology, 56: 269–276. Sun, X.G., 2005. The investigation of chemical structure of coal macerals via transmitted-light FT-IR microscopy. Spectrochimica Acta Part A 62: 557–564. Sun, Y.Z., and Horsfield, B., 2005. Comparison of the geochemical characteristics of “barkinite”and other macerals from the Dahe mine, South China. Energy Exploration and Exploitation 23 (6): 475–494. Tian, L., Dai, S., Wang, J., Huang, Y., Ho, S. C., Zhou, Y., Lucas, D., and Koshland, C. P., 2008. Nanoquartz in Late Permian C1 coal and the high incidence of female lung cancer in the Pearl River origin area: a retrospective cohort study. BMC Public Health, 8, 398, doi: 10.1186/14712458-8-398.
Manuscript Not AVAILABLE
PYROLYSIS RESIDUE FROM WASTE MATERIALS IN BLACK COAL FLOTATION Authors:Peter FEČKO, Alena KAŠPÁRKOVÁ, Vlastimil KŘÍŽ, Josip IŠEK,Tien PHAM DUC Address: VSB – Technical University of Ostrava, Faculty of Mining and Geology 17. listopadu 15, 708 33 Ostrava-Poruba [email protected]
Abstract: The paper deals with verification of floatability of classical collector Montanol 551 and pyrolysis oils, which were obtained through pyrolysis of waste, namely mixed plastics, tyres and waste rubber in combination with black coal from Lazy Mine, in black coal flotation. Black coal from ČSA OKD, a.s. coal preparation plant was used for flotation tests. The results imply that it is possible to produce collectors from waste materials which may be applied in the flotation of black coal. Keywords: flotation, black coal, pyrolysis, plastic waste, tyre and rubber
Studies of a Multi Gravity Separator (MGS) to Produce Clean Coal from Turkish Lignite and Hard Coal Fine Tailings
Selçuk ÖZGEN, M. Fatih CAN, Eyüp SABAH*, Department of Mining Engineering, Afyon Kocatepe University, 03200 Afyonkarahisar, Turkey
ABSTRACT The multi-gravity separator (MGS) is a novel piece of equipment for the separation of fine and ultra-fine minerals. This study was conducted to evaluate the effects of different process variables on the performance of the Multi Gravity Separator (MGS) for beneficiation of Turkish lignite and hard coal fine tailings to recover ultra-fine coal. The main minerals of Tunçbilek lignite tailings are kaolinite, illite and mica. The dominant minerals of Zonguldak hard coal tailings are chlorite, quartz, mica, calcite, pyrite, and amorphous materials. Various operating and design conditions of MGS such as drum speed, tilt angle, shaking amplitude, wash water rate, feed rate and pulp solid ratio were investigated. A hydrocyclone was used for pre-enrichment with the MGS. Operation parameters of the hydrocyclone, namely feed solids, inlet pressure, vortex finder and apex diameters were investigated. The results showed that clean coal was obtainable with 22.83% ash, 5.696 kCal/kg calorific value and recovery of 49.32% from lignite has 66.21% ash and 1.835 kCal/kg calorific value, with 6.98% ash, 7.214 kCal/kg calorific value and recovery of 85.61% from a hard coal has 28.41% ash and 5863 kCal/kg calorific value by this two-stage concentration process.
Key words: Fine lignite coal tailings, Waste processing, Hydrocyclone, Multi Gravity Separator (MGS)
*
Corresponding author: Prof. Dr. Eyüp Sabah
Afyon Kocatepe University Department of Mining Engineering 03200 AFYONKARAHİSAR e-mail: [email protected] Phone: +90-272-228-1423 / 1320 Fax: +90-272-228-1422
INTRODUCTION Coal is the single most abundant source of energy available in most nations around the word [Duong et al., 2000]. In Turkey, coal washing plants use gravity separation and beneficiate upper 500 µm coal and finer parts are discharged as tailings. In most coal washeries, around 20% of total run-of-mine coal size is found to be (under) minus 500 µm. The increased use of highly mechanized mining methodologies to enhance productivity has been the major cause for the generation of large quantities of coal fines [Majumder et al., 2007a]. So, not only great deal of material is economically lost, but also coal areas encounter serious environmental issues.
The processing of the finer coal fraction is relatively difficult due to the high processing cost, low process recovery and high moisture content of the product. Fine-coal beneficiation practice in the past involved the use of most of the commercially viable concentration processes either individual or in combination [Rao and Bandopadhyay, 1992; Majumder et al., 2007b; Majumder and Barnwal, 2008]. The equipment used in the beneficiation of coal fines consists of jigs, dense-medium cyclones, spirals, shaking tables and flotation.
Recently, a new gravity based processor, the Multi Gravity Separator (MGS), has appeared on the market with an operating principle that seems to be very promising for processing fine particle [Traore et al., 1995; Majumder and Barnwal, 2006; Aslan, 2007]. The MGS
concentrates particles based on the combined effects of centrifugal acceleration and forces acting on a conventional table. MGS may be visualized as a cylindrical version of a conventional shaking table [Majumder et al., 2007a; Chan et al., 1991; Udaya Bhaskar et al., 2002; Menéndez et al., 2007]. This device was developed for the selective separation of fine and ultra-fine particles mostly based on the differences in their densities. The use of centrifugal forces in a MGS enhances the relative settling velocity differential between the particles differing in size and density [Majumder et al., 2007a, Majumder and Barnwal, 2006], and the additional shearing force created by the shaking motion of the drum enhances the particle separation process [Chan et al., 1991a]. A detailed description of the MGS is given by Venkatraman et al. [1995].
The early applications of the unit for concentrating heavy minerals like tin, tungsten, tantalum, chromites and celestite have been reported elsewhere [Traore et al., 1995; Udaya Bhaskar et al., 2002; Chan et al., 1991b; Chan et al., 1989; Tucker et al., 1991; Turner and Hallewell, 1993; Clemente et al., 1993; Uçbaş and Özdağ, 1994; Burt et al., 1995; Aslan and Canbazoğlu, 1996; Udaya Bhaskar et al., 1999; Aslan and Kaya, 2009]. Recently, research studies have increased on its use in fine coal processing [Göktepe et al., 1996; Rubiera et al., 1997; Aslan et al., 1999; Çiçek et al., 2002; Engin et al., 2006; Aslan, 2007].
In this study, the applicability and modeling of fine gravity concentration technique for recovering fine coal from the gravity tailings included different inorganic matters of two different coal concentrators of Turkey is investigated using a laboratory scale MGS. The first samples are taken from Tunçbilek coal preparation plant lignite tailings, the second samples are taken from Zonguldak Catalagzi coal preparation plant hard coal tailings.
2. MATERIALS AND METHODS 2.1 Materials Lignite coal tailings obtained from Tunçbilek Coal Preparation Plant of G.L.I of Turkish Coal Enterprises (Kütahya-Turkey). The hardcoal slurry tailings sample used in these experiments was obtained from Çatalağzı Coal Preparation Plant of Turkish Hardcoal Enterprises (Zonguldak-Turkey). The samples were taken from slurry waste with standard of TS ISO 5667-10.
2.2 Methods 2.2.1. Characterization Tests Types of minerals phases were identified by the X-ray diffraction (XRD) technique (RigakuGiger Flex). X-ray fluorescence (XRF) Spectro IQ was used for chemical analysis. The particle size distribution of the sample was measured a using Mastersizer 2000. The ash and sulfur values were analyzed according to ISO 1171 and ISO 351, respectively. The calorific value was determined based on ISO 1928.
2.2.2. Classification and Concentration Tests Classification tests were performed using a hydrocyclone. The aim of this test was to separate the clay minerals from the coal. The effect from varying the operating parameters for the hydrocyclone such as vortex diameter, apex diameter, feed inlet pressure, and suspension solid ratio were investigated. Since large amounts of fine and ultra fine particles in the tailings cause a low cut point, a small diameter hydrocyclone (44 mm) was selected for using the experiments. Underflow products was concentrated a Multi-Gravity separator (Richard Mozley Ltd., Model C900). The effects of varying operating parameters such as drum speed,
tilt angle, shaking amplitude, wash water rate, pulp feed rate and pulp solid ratio was investigated as well. For each MGS test, 2000 g of the dry coal sample was used. When changing a parameter in each test, the other parameters were kept constant and the optimization results were used in the other tests. The shaking frequency of the MGS was fixed at 4.9 cps for all the experiments.
To obtain the required feed solids, measured quantities of solids and water were mixed in the slurry tank. The MGS variables were adjusted at the required levels. The feed slurry was pumped into the MGS drum at the required flow rate using the peristaltic pump while the MGS was in operation. Samples from the clean coal and tailing streams were collected at steady-state condition. The samples were filtered, dried and analyzed for ash content and combustible recovery.
3. RESULTS AND DISCUSSION 3.1 Characteristics of Tailings The mineralogical analysis of the tailings indicates that most of the minerals are chlorite, quartz, mica, calcite, pyrite, and amorphous materials for hard coal and kaolinite, smectite, illite quartz, feldspar, siderite, dolomite and pyrite for Tunçbilek coal tailings. The chemical analysis for the sample is presented in Table 1. As seen in Table 1, LOI of the sample is about 37% and 74% which means that the amount of organic material in the tailings was significant when compared to the amount of inorganic impurities for Tunçbilek and Hardcoal, respectively.
Table 1. Chemical analysis of original tailings. Composition SiO2 Al2O3 Fe2O3 MgO CaO Na2O K2O TiO2 P2O5 MnO LOI
Weight (%) Tunçbilek Hardcoal 35.53 13.79 13.34 6.86 5.89 1.65 3.68 0.66 1.53 1.05 <0.11 0.13 1.23 1.1 0.50 0.29 0.10 0.04 0.13 0.03 37.60 74.30
Table 2 presents the ash contents of the sample with respect to particle size. Table 2 shows that the ash content of the sample increased with a decrease in particle size. The qualitative and quantitative analysis results of tailings are presented in Table 3.
Table 2. Ash values of original coal wastes. Particle Size (µm) +425 -425+300 -300+106 -106+75 -75+38 -38 Total
Ash Content (%) Tunçbilek Hardcoal 46.88 10.71 46.79 12.63 46.70 17.00 51.13 32,39 56.71 39.29 77.43 43.19 66.21 28.41
Table 3. Main properties of tailings. Tailings
Parameters 3
Specific gravity, gr/cm Solids content, % Particle size interval, m Percentage of clay size particles Percentage of silt size particles Percentage of sand size particles Mineral content
Total ash content, % Calorific value, kcal/kg
Tunçbilek 1.88 7.20 -500 37 59 4 Kaolinite, smectite, illite quartz, feldspar, siderite, dolomite and pyrite 66.43 1835
Hardcoal 1.53 5.11 -1000 4 20 76 Chlorite, quartz, mica, calcite, pyrite and amorphous materials 28.41 5863
3.2 Studies on MGS with Tunçbilek and Hardcoal Tailings The variables and results regarding the obtained MGS final beneficiation tests with the Tunçbilek and Hardcoal tailings were given respectively in Table 4-5. In these tests regression analysis was performed to determine the relationship between 6 variables and 2 response functions.
From the experimental results listed in Table 4-5 and the response functions representing the ash content of clean coal and the recovery of clean coal could be expressed as functions of the pulp solid ratio (s), drum speed (r), tilt angle (˚), shaking amplitude (sh), wash water rate (w), feed rate (f).
Table 4. Level of variable and MGS test results for Tunçbilek coal tailings.
Exp. No 1 2 3 4 5 6 7 8 9 10 11 12 13
x1 (s), % 15 15 15 15 15 15 15 10 20 15 15 15 15
Level of variables x4 x2 (r), x3 (a), x5 (w), (sh), rpm ˚ Lpm mm 229 2 10 1 229 2 10 3 229 2 10 5 229 0 10 3 229 4 10 3 170 2 10 3 201 2 10 3 201 2 10 3 201 2 10 3 201 2 15 3 201 2 20 3 201 2 10 3 201 2 10 3
Observe results x6 (f), Lpm
Ash Content, %
Recovery, %
1 1 1 1 1 1 1 1 1 1 1 2 3
19.21 20.86 26.44 16.75 25.81 37.68 19.72 32.48 27.19 31.47 36.18 22.83 26.37
23,98 41,72 56,12 29,12 38,31 78,57 44,38 70,28 41,34 63,22 74,18 49,32 47,79
Table 5. Level of variable and MGS test results for Hardcoal coal tailings. Level of veriables Exp . No 1 2 3 4 5 6 7 8 9 10 11 12 13
x1 (s), %
x2 (r), rpm
x3 (a), ˚
x4 (sh), mm
10 15 25 15 15 15 15 15 15 15 15 15 15
201 201 201 170 229 201 201 201 201 201 201 201 201
2 2 2 2 2 0 4 2 2 2 2 2 2
10 10 10 10 10 10 10 15 20 10 10 10 10
Observe results x5 (w), Lpm 3 3 3 3 3 3 3 3 3 1 5 3 3
x6 (f), Lpm
Ash Content, %
Recovery, %
1 1 1 1 1 1 1 1 1 1 1 3 5
7.92 6.98 8.00 8.06 7.17 7.90 7.34 8.50 9.08 7.12 8.54 7.74 7.56
73.23 61.73 64.90 87.00 26.63 60.43 65.64 84.10 82.00 56.76 77.26 58.91 55.44
Regression analysis was chosen to determine the relationship between the response functions (ash value and recovery of the clean coal) and operating conditions for the MGS (pulp solid ratio, drum speed, tilt angle, shaking amplitude, wash water rate, feed rate). For ash content of the clean coal AshTunçbilek = 65.9-0.529*x1-0.233*x2+2.27*x3+0.824*x4+1.81*x5-1.53*x6
(1)
AshHardcoal = 8.19+0.017*x1-0.015*x2-0.14*x3+0.146*x4+0.355*x5-0.014*x6
(2)
For Recovery of the clean coal RecoveryTunçbilek = 181-2.89*x1-0,633*x2+2.30*x3+1.91*x4+8.04*x5-3.79*x6
(3)
RecoveryHardcoal= 233.59-0.3*x1-1.01*x2+1.11*x3+2.23*x4+5.13*x5-2.25*x6
(4)
where AshTunçbilek and RecoveryTunçbilek is ash value and recovery (in %) of the clean coal of the Tunçbilek, AshHardcoal and RecoveryHardcoal is ash value and recovery (in%) of the clean coal of the Hardcoal, respectively. X1 is pulp solid ratio, X2 is drum speed, X3 is tilt angle, X4 is shaking amplitude, X5 is wash water rate, and X6 is feed rate.
The response factors at any regime in the interval can be calculated from these equations. The observed and the predicted values obtained using the models are provided in Figs. 1 and 2 (the R2 values were indicated above). The predicted values and the observed data points, indicating a very good fit (R2 value of 0.807 and 0.845 for ash content of the clean coal, R2 value of 0.944 and 0.807 for recovery of the clean coal on MGS wit Tunçbilek and Hardcoal, respectively) of the response equations.
Predicted Ash Content, %
Tunçbilek Coal Hard Coal y = 0.8069x + 5.1665 R² = 0.8072
y = 0.8448x + 1.2175 R² = 0.8447
Observed Ash Content, %
Figure 1. Predicted versus observed response of ash content on MGS with Tunçbilek and Hardcoal tailings.
Tunçbilek Coal Predicted Recovery, %
Hard Coal
y = 0.8838x + 7.6939 R² = 0.8849
y = 0.9438x + 2.673 R² = 0.9438
Observed Recovery, %
Figure 2. Predicted versus observed response of recovery on MGS with Tunçbilek and Hardcoal tailings.
In the final beneficiation tests that performed in MGS, optimum result is dependent test number 7 when took in consideration the ash contents and recovery of clean coal for Tunçbilek. The ash content of the clean coal is 19.72% and the coal recovery 44.38%.
However, the coal that has the lowest ash content with 16.75% in test number 4; the highest recovery with 78.57% in test number 6 were obtained. In the final beneficiation tests with the Hardcoal tailings that performed in MGS, optimum result is dependent test number 2 when took in consideration the ash contents of clean coal and recovery of coal. The ash content of the clean coal is 6.98% and the coal recovery 61.73%. Namely at test 2 the lowest ash content of coal with 6.98% and at test 4 the highest recovery with 87.00% were obtained.
4. CONCLUSIONS The results presented in this article taking care study on tailings at Tunçbilek Lignite coal tailings and Zonguldak hardcoal tailings shows that the application of a hydrocyclone followed by a Mozley separator as a reasonable method for the concentration of fine coal containing large amount of clay and sulfur bearing minerals at a relatively fine size than coal particles.
The concentration tests for the clean coal showed that it is economically possible to recover the Tunçbilek and Hardcoal tailings containing clay minerals. The separation and concentration studies were performed using two stages. In the first stage, clay minerals were removed from the hardcoal using a hydrocyclone. After classification with hydrocyclone, ash content of Tunçbilek coal decreased from 66.21% to 45.87%, where hard coal ash content decreased from 28.41% to 19.59%.
After the hydrocyclone separation tests, the underflow product was concentrated with MGS. The experimental results with the MGS show that the optimum parameter for the separation is 201 rpm of drum speed, 2◦ of tilt angle, 1 Lpm of feed rate, 3 Lpm of wash water rate, 4.9 cps shaking frequency, 10mm of shaking amplitude and 15% of solid weight for Tunçbilek and
Hardcoal. A coal sample with 19.72% and 6.98% ash content and net calorific value of 5696 kCal/kg and 7214 kCal/kg were obtained with a weight recovery of 44.38% and 61.73%, respectively.
References Aslan, N., Canbazoğlu, M., 1996, Processing of thickener underflow from celestite concentrator by Multi Gravity Seperator, In: Changing Scopes in Mineral Processing, Turkey, 103-106. Aslan, N., Canbazoğlu, M., Ulusoy, U., 1999, An investigation of washability characteristics of lignites from Yeniçubuk-Gemerek districts by MGS, In: 16th Mining Congress of Turkey, 321-326. Aslan, N., 2007, Application of response surface methodology and central composite rotatable design for modeling the influence of some operating variables of a Multi-Gravity Separator for coal cleaning, Fuel, 86, 769-776. Aslan, N., Kaya, H., 2009, Beneficiation of chromite concentration waste by multi-gravity separator and high-intensity Induced-roll magnetic separator, The Arabian Journal for Science and Engineering, 34 (2B), 285-297. Burt, R.O., Korinek, G., Young, S.R., Deveau, C., 1995, Ultrafine tantalum recovery strategies, Mineral Engineering, 8, 859-870. Chan, S.K., Mozley, R.H., Childs, G.J.C, 1991a, The Multi-Gravity Seperator (MGS)-A mine scale machine, Richard Mozley Limited, Redruth, Cornwall, UK; 20 p. Chan, S.K., Mozley, R.H., Childs, G.J.C, 1991b, Extended trials with the high tonnage multigravity separator, Mineral Engineering, 4(3-4), 489-496. Chan, S.K., Mozley, R.H., Childs, G.J.C., 1989, The multi-gravity separator (MGS)-a mine scale machine. In: Dowd PA (Ed.). Symposium, Mineral Processing in the United Kingdom. IMM, 107-123. Clemente, D., Newling, P., Betalho de Sousa, A., Lejeune, G., Barber, S.P., Tuker, P., 1993, Reprocessing slimes tailings from a tungsten mine, Mineral Engineering, 6, 831-839. Çiçek, T., Cöcen, İ., Engin, V.T., Demir, Ş., 2002, Recovery of fine coal from tailings of coal washing plants. In: IXth Int Min Pro Symposium, Turkey, 198-199. Duong, C., Choung, J., Xu, Z., Szymanski, J., 2000, A novel process for recovering clean coal and water from coal tailings, Mineral Engineering, 13, 173-181. Engin, V.T., Güler, G., Cöcen, İ., Çiçek, T., 2006, Recovery of coal from fine tailings of a coal washing plant using centrifugal gravity seperators, In: XXIII. International Mineral Processing Congress, Turkey, 1204-1207. Göktepe, F., Pooley, F.D., Williams, K.P., Wise, G.J., Trillo-Soto, R., 1996, Coal desulphurisation with the Mozley Multi-Gravity Separator, In: Kemal M, Arslan V, Akar A, Canbazoğlu M (Ed.). Chenging Scopes in Mineral Processing, Balkema, 97-101. Majumder, A.K., Barnwal, J.P., 2006, Modeling of enhanced gravity concentrators-present status, Mineral Processing & Metallurgy Review, 27, 1-26. Majumder, A.K., Bhoi, K.S., Barnwa,l, J.P.,2007a, Multi-gravity separator: an alternate gravity concentrator to process coal fines, Mineral Processing & Metallurgy Review, 24, 133-138. Majumder, A.K., Tiwari, V., Barnwal, J.P., 2007b, Separatıon characterıstıcs of coal Fınes ın a knelson concentrator-a Hydrodynamıc approach, Coal Preparation, 27, 126-137. Menéndez, M., Gent, M., Toraño, J., Diego, I., 2007, Optimization of multigravity separation for recovery of ultrafine coal, Mineral Processing & Metallurgy Review, 24, 253-263. Rao, L.S., Bandopadhyay, P., 1992, Application of a mozley mineral separator for treatment of coal washery rejects, International Journal of Mineral Processing, 36, 137-150. Rubiera, F., Hall, S.T., Shah, C.L., 1997, Sulfur removal by fine coal cleaning processes, Fuel, 76 (13), 1187-1194. Traore, A., Conil, P., Houot, R., Save, M., 1995, An evaluation of the Mozley MGS for fine particle gravity separation, Mineral Engineering, 8, 767-778.
Tucker, P., Chan, B.S.K., Mozley, R.H., Childs, G.J.C., 1991, Modelling the multi-gravity separator, XVII International Mineral Processing Congress, 77–89. Turkish Standard, TS ISO 5667-10, 2002, Water quality-sampling-Part 10: Guidance on sampling of waste waters, Turkish Standards Institution, 8 p. Turner, J.W.G., Hallewell, M.P., 1993, Process improvements for fine casseterite recovery at wheal jane. Mineral Engineering, 6, 817-829. Uçbaş, Y., Özdağ, H., 1994, Relationships between shake frequency and amplitude in the concentration of chromite fines by Multi-Gravity Separator, In: Proceedings of the 5 th International Mineral Processing Symposium, Turkey, 71-76. Udaya Bhaskar, K., Barnwal, J.P., Rao, T.C., Venugopal, R., 1999, Multigravity separator to enrich heavy minerals from a lead flotation concentrate, Mineral Processing & Metallurgy Processing, 16 (1), 61-64. Udaya Bhaskar, K., Govindrajan, B., Barnwal, J.P., Venugopal, R., Jakhu, M.R., Rao, T.C., 2002, Performance and modeling studies of an MGS for graphite rejection in a lead concentrate, International Journal of Mineral Processing, 67, 59-70. Venkatraman, P., Luttrell, G.H., Yoon, R.H., Knoll, F.S., Mankosa, M.J., 1995, Fine coal cleaning using the Multi-Gravity Separator. In: Kavatra SK (Ed.). High Efficiency Coal Preparation: An International Symposium, Colorado, 109-117.
Petrographic Characterisation of Beneficiated Material of Tailings from Soma Işıklar Derekoy Coal Washing Plant (TURKEY), by Multi Gravity Seperator (MGS) Presenting Author (Name, Title, Affiliation): Selami TOPRAK, Assoc. Prof. in Geology, Contact Information: Mineral Research and Exploration Directorate in Turkey ([email protected]; [email protected]) Co-Authors: Ayşe ERDEM*, Akan GULMEZ*, Oguz ALTUN*, Sarper ALYILDIZ**, Mining Engineers Contact Information: *Mineral Research and Exploration Directorate in Turkey, **Turkish Coal Enterprises
ABSTRACT In this study, coal tailing samples taken from Soma Isıklar Washing Plant belonging to Aegean Lignite Enterprise of Turkish Lignite Authority were analyzed petrographically and the results were used in the ore beneficiation studies. Representative sample was taken from determined tailing pond and with aid of a mechanical mixer working by electricity; a homogeneous mixture of the material was obtained and reduced to a working amount in a laboratory. Samples were obtained for analysis (mineralogic-petrographic, chemical and sieve analysis) and wet screening. The tailing is composed of coal and associating gangue minerals. In order to determine the types, percentages and liberation degree and their average sizes, mineralogic and petrographic analyses were conducted to the samples. It was determined that the tailing is composed of 55% of mineral matter which contains of clay, calcite and pyrite minerals. And the coal part seems to contain about 45% of the material and is composed mostly of Huminite macerals. According to evaluation of the mineralogic and petrographic, chemical and sieve analyses, the samples with -0.5+0.3 mm and -0.3+0.02 mm sizes were determined as suitable for beneficiation studies by MGS and -0.02 mm size were determined as to be worked by flotation methods. The sizes of different pyrites were determined and their sizes seemed to be too tiny for beneficiation. MGS was used to upgrade and increase the calorific values of the coal wastes. After the cleaning process, the ash content of the concentrate lowered up to 16.45%, total sulfur to 1.26% and calorific value to 5335 kcal/kg.
1. Introduction This study is a part of a Project, supported by The Scientific and Technological Research Council of Turkey (“TUBITAK”). Soma is located in the mid-western part of Turkey. It is one of the major coal and energy production center of the nation. With 633 million tons of coal, it holds almost 25% of the country’s total reserve. 66 % of the coal production is consumed in Soma’s Thermic Power Plant and about 34 % of it is used in domestic consumption, mostly heating. Annually about 4.5 million of coals is produced in Soma region. The coal is produced from either underground or open pit areas and from various
locations, but mostly by state owned coal establishment (“Turkish Lignite Authority”) (www.tki.gov.tr).
Figure 1. Turkish Coal Basins and some important Lines (From: Toprak, 2009). Coal in Soma has two different coalification ranks; mostly sub-bituminous, but also some lignites as well. The lignite is directly sent to the thermic power plant. The sub bituminous coal is processed, some sent to the thermic power plant and some supplied to the market as briquettes or upgraded forms. Lignite in Soma is younger (Pliocene aged) coal and the sub-bituminous coal is much older (Miocene) and of much better quality coal. The average thickness of sub-bituminous coal is about 12 meters. The studied coal is related with the sub-bituminous coals in the region.
2. Coal and The Applied Processes The sub bituminous coal seam, which has the best quality and is of the thickest in the region, is inclined with a slight angle as seen on Figure 2. The coal is associated with clastics and carbonate materials but sometimes with volcanics as well. The area was surrounded by young volcanics at South, caused the coal got higher ranks and show high reflectance measurements at various places in the region, sometimes even showing more than 1% Rmax values. The tectonics and volcanic activities have affected the coals. The region, during deposition, was suitable for plant accumulation to form thick coal beds, then buried with sediments, as seen on Figure 2 and 3. On Table 1, the chemical properties of the material can be seen. The ash content of the samples are very high (over 50%) and the calorific values of them are very low (in the order of 2000 levels). After drying and upgrading the sample, the calorific values get more than twice higher values.
Figure 2. The N-S Directioned X-Section of the region and the coal seams (Toprak, 2009).
Figure 3. Removing the overburden for open operation of the thickest coal seam in the area.
Table 1. The chemical analysis of the taken samples from the washery. Analysis Air Dried Dry Moisture (%) 5,43 Ash (%) 50.45 53.21 Volatile Matter (%) 28.95 30.38 Fixed Carbon (%) 15.17 16.41 Combustible Sulfur (%) 0.17 0.12 Sulfur in ash (%) 0.77 0.78 Total Sulfur (%) 0.89 0.90 Low Calorific Value (Kcal/kg) 2318 2361 Upper Calorific Value (Kcal/kg) 2460 2496 Figure 4. A view of Soma Thermic Power Plant and a part of the town.
Soma region (Figure 4) is a small valley, surrounded by coal bearing mountains. The washery is one of the biggest in Europe (Figure 5). The samples were taken from the discharge of the pool. About 1.5 ton of representative sample was taken from determined tailing pond. In a laboratory, a homogeneous mixture of the material was obtained and reduced to a working amount, with aid of a mechanical mixer working by electricity. In order to gain data enlighten further studies, 100 kg of the samples were obtained for analysis (mineralogic-petrographic, chemical and sieve analysis) and 400 kg of the samples for Russell type wet screening were obtained as classified. As known, the tailing is composed of coal and associating gangue minerals. In order to determine the types, percentages and liberation degree and their average sizes, mineralogic and petrographic analyses were conducted to the samples. For this, the samples were put into pellet forms, and then polished to be observed with incident light by Leitz MPV-SP coal microscope. It was determined that the tailing is composed of about 55% of mineral matter which contains of clay, calcite and
pyrite minerals. And the coal part seems to contain about 45% of the material and is composed mostly of Huminite (Vitrinite) macerals (more than 75 % of the original coals). Liptinites and inertinites are considerably very rare (less than 4 %).
Figure 5. An Outlook of the Washery. Pyrites in the coals are less than 2% and the grain size of pyrites very are small (few microns at most) but the size of their forms change between 5 to 20 microns, but averagely between 515 microns in size (Figure 6 and 7). The individual grains are so small and disseminated all over coal itself. Therefore, it is considerably hard to get rid of them easily. As separation process advances, it is expected to get higher amount of pyrites from the coals. But unfortunately, the pyrite content does not correspond with the decreasing figure of the other impurities. It keeps almost constant and remains locked within the coal itself as small grains. It can be said that it behaves with the coal itself in ore processing. Since pyrite grains are so small and distributed all over coal, it is not surprising to have similar ratio as it gets finer. According to evaluation of the mineralogic and petrographic, chemical and sieve analyses, the samples with -0.5+0.3 mm and -0.3+0.02 mm sizes were determined as suitable for beneficiation studies by MGS and -0.02 mm size were determined as to be worked by flotation methods. The sizes of different pyrites were determined for beneficiation but the decreasing of the sulfur value seemed very difficult, due to their sizes (1-15 microns) and the adopted technology for beneficiation. In Figure 7, the microscopic view of the inorganic types and Figure 8, 9 SEM pictures of the Pyrites are clearly seen. The framboidal shape (Figure 6 and 7), and idiomorphic crystal fillings of the cavities in the coals are clearly seen. It is rather difficult to get rid of such small grains, if they are individually distributed in the coals which are common for the coals.
Figure 6. The microscopic view of the material.
Figure 7. A SEM (Scanning Electron Microscope) Picture of Coals and Framboidal (circular forms) and tiny pyrites
Figure 8. A SEM Picture of the Tiny Crystal forms of Pyrites in Coals.
MGS (Figure 10) was used to upgrade and increase the calorific values of the coal wastes. When -0.5+0.3 mm and -0.3+0.020 mm size fraction sized concentrates were combined, MGS was chosen as the best method to upgrade and increase the calorific values of the coal tailings. Of the MGS feeding material (except for -0.020 mm size fraction), 67.43% (as weight) of coal concentrate, 16.45% of ash, 1.26% of total sulfur and about 5335 kcal/kg calorific value were obtained (Table 2). In the cleaning coal Project of Soma Işıklar Dereköy Washery to obtain clean coal from the thickener entry the material (under 0.500 mm), clean coal was successfully produced with MGS, from the material coarser than 0.020 mm. But, the same success might not be achieved for the material finer than 0.020 mm. Therefore, about 50% of the material which comprise of lower calorific value of 1073 kcal/kg, ash content of 67.85 % and average size changing between 8 – 10 microns, were tried to enrich by using Jameson flotation cell. According to the experimental results, ash content as well as calorific values of the collected materials are considerably low. As the lowest calorific value and combustion efficiency were considered, (concentrate amount to feeding is 13.8 %), about a ton of clean coal can be produced out of about 7.2 tons of the tailing. In order to this, 144 kg of collecting agent (Kerosene in this case) is required. This figure makes the last process not economical to conduct at all.
Figure 10. The Laboratory Scale MGS instrument and working condition. Table 2. Total balance according to the entry. Products
MGS -0.5+0.3 (Concentrate) MGS -0.5+0.3 (Tailing) MGS -3+0.020 (Concentrate) MGS -0.3+0.020 (Tailing) -0.020 ENTRY
Weight
Content Calorific Value, Kcal/kg
Total S,%
Distribution % Calorific Ash, Total Value, % S,% Kcal/kg
Combustion Efficiency
%
*Ash, %
9.18
19.25
5084
1.27
4.07
21.59
14.47
17.49
2.80
75.20
630
1.66
3.66
0.62
2.18
1.38
24.37
15.40
5429
1.25
8.44
63.40
38.17
50.49
13.47
79.00
430
1.25
17.82
2.86
15.53
6.35
50.17
67.85
1073
0.56
65.97
11.53
29.65
24.30
100.0
52.31
2404
0.92
100.0
100.0
100.0
100.00
*:dry bases. 3. Conclusion The tailing is composed of coal and associating gangue minerals; about 50 % of mineral matter is clay, calcite and pyrite minerals. And the coal part seems to contain about 45% of organics which is mostly Huminite. MGS has a very powerful effect on upgrading the coals. In ore processing experiments, as ash content of the two sizes seem to have a decrease, total sulfur content of the material does not seem to show any changes at all. Though a small change of sulfur content of the material was seen in small fractions, pyrite liberation of the materials was not clearly observed at all.
4. References Toprak, S, 2009. Petrographic Properties of Major Coal Seams in Turkey and Their Formation, International Journal of Coal Geology, Elsevier, Vol. 78, 4, 263-275. www.tki.gov.tr (web pages of Turkish Coal Enterprises).
Soma Region’s Coals Washing at Dereköy Washery with Working 800 tph and It’s Washing Performance Evaluation Sarper ALYILDIZ., Sevgi GÜRKAN., Serap TUNCER Directorate of the Aegean Lignite Establishment, Atatürk St. N:111 Soma/Manisa, Türkiye ABSTRACT: Coal is one of the major energy sources in most nations like Türkiye. The washing of coal is an important industry in Turkish coal mining. For this reason, run of mine (R.O.M.) coal must be washing. The TKI (Directorate of the Turkish Coal Enterprises) administration came to the decision to set up a coal washing plant to produce high quality coal according to the demands of the industry and the thermal power plant. In 2006 a coal washing plant was set up by a private company (Çiftay Şti.) by means of build-operate model for washing 20 Million tonnes R.O.M. coal. This plant consists of heavy medium drum, heavy medium cyclone and spiral concentrators with a washing capacity of 4 million tonnes. With two parallel circuits each working with 400 tph capacity, 800 tonnes of raw coal can be washed. 17.6 million tonnes of run of mine coal have been washed by this company until August 2010 in Soma Region. The product and calorific values of tailings are given. Minimal coal losses to reject were achieved.
INTRODUCTION Soma is a town with a population of 70.000 people by Manisa in the western part of Türkiye in the Aegean zone. It is a mining and energy production city. Its economy depends on coal and electricity. Soma is the one of the biggest Türkiye’s lignite production region. In Soma Region, T.K.İ has higher than 639 million tones of coal reserve.
Figure 1. Soma – Manisa The Aegean Lignite Establishment has 193 million tones of reserves at open pit mine, 446 million tones at underground mine totaling over 639 million tonnes of coal reserve. The Aegean Lignite Establishment sells the coal to Thermal Power Station (1.044 MW) and individual domestic buyers. Administration of General Directorate of Turkish Coal has decided to establish a modern washing unit to clean low quality coals (Figure 2) and provide increasing demand of high quality coals. Float sink tests of various pits showed that quantity and quality of the products can be changed with major quality improvements. Various coal process flow sheets were developed with simulations (Anaç., 2006). After this study, Çiftay Company made a contract with the Aegean Lignite Establishment for washing of 20 million tonnes of raw coals with 4 million tonnes capacity, each year. In this paper, new coal washing plant flow
sheet and 2006–2010 production results and performance data will be presented in the Targets and Performances section.
Figure 2. The quality of Turkish lignites Direct usable high quality coal reserves are limited in the world, like Türkiye (Figure 2). These coals (low quality coals) must be washed. DEREKÖY COAL WASHING PLANT DESIGNING CRITERIAS The most common method, nowadays, for upgrading coal is based on density difference between coal and mineral material. ROM coal size characteristics must be known very well for designing plants to insure specific circuits are not overloaded. Taking, representative samples from raw coal mining sections are very important. The samples, from underground and open pits which will be fed to washing plant, were taken as in the standard methods. ROM coal size variation curve graphics are given in Figure 3 for each mining section. 100
100
80 Güney Kısrakdere
70
Eynez-open pit
60
Eynez-underground
50
Işıklar A
40
Işıklar ED Sarıkaya
30
Karanlıkdere
20 10 0 1.2
1.4
1.6
1.8
Cumulative Undersize (%)
The amount of float product(%)
Güney Kısrakdere
90
90
Eynez-Open Pit
80
Işıklar ED
70
Eynez-Underground
60
Işıklar A Sarıkaya
50
Karanlıkdere
40 30 20 10
2
0
Density (g/cm 3)
0.1
1
10
100
1000
Particle Size (mm)
Figure 3. Raw coal size variation and Washability curve ROM coal washability curves plotted (Figure 3). Based on this data the washing plant was designed. Targeted for clean coal and expected tailings ash levels at the washing plant are shown on Table 1.
Table 1. Targeted limit values of clean coal and tailings Products and Tailings + 18 mm 10-18 mm 0.5-10 mm 0.1-0.5 mm Middling 0.5-18 mm Tailing +18 mm Tailing
% Ash ( dry basis) 12-23 10-22 9-21 14-40 ▬ Min. 67 Min. 67
% Moisture ▬ ▬ Max. 22 Max. 42 Max. 22 ▬ ▬
DESCRIPTION OF THE PLANT The coal washing plants consist of two identical circuits. Each of them has 400 tph capacities. Dereköy coal washing plant general view is shown on Figure 4. Dereköy Coal Washing Plant
Figure 4. A view of Dereköy washing plant The plant can be investigated in five main sections, namely. Feeding, Coarse size coal washing circuit, Fine size coal washing circuit spiral circuit and Magnetite recovery. Run of mine coal (ROM) is fed from 600×300 mm grate. It is separated by screening from 150 mm screen. Oversize material of the screen is sent to the crusher. ROM coal is fed to intermediate bunker after combining with undersize material of screen. The 150 mm size material in the intermediate bunker is fed to two identical and parallel circuits. -150+18 mm sized material is fed to the heavy medium drum which runs with two different densities. First compartment density is ~1.67 gr/cm3. Floated material of the first compartment is taken as product. The sink is fed to second compartment which has working density as ~1.85 gr/cm3. Floating product (middling) of the second compartment was sent to Thermal Power Plant unit B. The sink product of the second compartment is taken as rejects. Products were conveyed to drain and rinse screen for washing and recovery of magnetite. High-density and diluted medium are collected at different tanks. Magnetite separator is fed with the diluted medium to remove water from the system and clean it. Flowsheet of coarse coal washing circuit is shown on Figure 5. Under the 18 mm screen’ material is passed trough the 0.5 mm dewatering screen. And then coal is fed to heavy medium cyclone. First cyclone works with low density more or less 1.60 gr/cm3 and floating material is taken from first cyclone as a clean coal. The material which sinking in the heavy medium part is fed to the second cyclone and then it is separated as a middling and tailing. Then 0.5-18 mm coal is separated into the 10 mm classification screen as a 0.5 -10 mm and 10-18 mm for different marketable demand. 0.5-10 mm clean coal is dried by centrifuge. Finally all products are sent product silos. Flowsheet of fine size coal washing circuit is shown on Figure 6. Three Product Heavy Medium Drum -150+18
Prewashing Screen
Prewashing Screen
-0,5
Middling Product
Drain and rinse screen Tailing stock area
Drain and rinse screen
10-18 Clean Coal
Spiral Circuit
0,5-10
Dryer Middling silo
Heavy medium Tank
2. Heavy Me Cyclon
0,5-18
Coarse Tailing
+18 piece
Drain and rinse screen
1. Heavy Medium Cyclone
Heavy Medium Container
Dilute medium Tank
Figure 5. Coarse size coal washing circuit
Figure 6. Fine size coal washing circuit
Undersize of the 0.5 mm dewatering screen is fed to the hydrocyclone. The cyclone over flow is passed trough the thickener when bottom flow is passed to the spirals. Three products can be obtained from the spiral. Tailing is taken from inner spiral. Clean coal is taken from outer spiral. Coal is sent to sieve bend and then this coal is given to the basket type dryer and the coal is passed the silos. Tailing is passed trough the dewatering screen. After this process it is thrown out to the tailing area. Flowsheet of fine spiral circuit is shown on Figure 7. Prewashing Screen 0,5-18
Dewatering Screen
Thickener -0,5
Hydrocyclone Spirals Sieve Bend
Dryer 0.1-0.5 mm Clean Coal Silos
Dewatering Screen
Spiral Tailing
Figure 7. Spiral circuit All of the system’s screen undersize filtrates are collected dilute medium tank. Dilute medium is fed to three magnetic separator units. Heavy medium is obtained from magnetic separators. While this material is feeding heavy medium tank, dirty water is fed very dilute medium tank. Very dilute medium is pumped to a cyclone. Over flow of the cyclone is sent to rinsing screen. Bottom flow is sent to separator. Dilute medium is gotten from the single magnetic separator and sent to dilute medium tank. Dirty water is fed to rinsing screen. Flowsheet of magnetite recovery is shown on Figure 8. Magnetite cyclones Head Box
Single separator
Systemize Dilute medium Container
Very Dilute medium Container
Heavy medium Container
Figure 8. Magnetite recovery circuit
Ultra Fine tailings treatment In Türkiye, the ultra fine tailings of lignite coal processing plants are sent in most cases to the tailings ponds without any treatment. For this reason, a new project has been conducted to reuse the coals, in the coal tailings ponds. The subject of the project is to beneficiate coal from coal tailing ponds with Multi Gravity Separator (Laboratory/Pilot–MGS) equipment. Rig for the model test is given on Figure 9. Multi Gravity Separator model is C900.
MGS Laboratory Working Mechanism
Figure 9. MGS equipment (C900) Old washery and the new coal washing plant (Dereköy) samples were used in the Multi Gravity Separator tests. Project’s last data showed that it is possible to recover coals economically from the coal tailing ponds. Laboratory analysis shows that clean coal quality is good enough for the marketing. TARGETS AND PERFORMANCE The first unit of the plant was started in February 2006, the second, in April 2006. The plant has reached o full capacity, and easily reached to 4 million tones per year as shown on Figure 10.
2010 July
Figure 10. Annual averages tonnes for washed coal Over 17 million tones of raw coal, was washed from the beginning of operation to 2010 July, at Dereköy coal washing plant. The production target was determined according to environmental standards and the quality of coals preferred for Thermal Power Plants. Production target of Dereköy washing unit is classified as below; 1. To produce -150+18 mm sized clean coal with having minimum 4.000 kCal/kg calorific values, 2. To produce -18+0.5 mm sized clean coal with having minimum 4.000 kCal/kg calorific values, 3. To produce middling for Power Plant unit B with having between 2.000-3.000 kCal/kg calorific values, 4. To produce middling for Power Plant unit A with having between 3.000-3.500 kCal/kg calorific values, 5. To produce tailings having heating value lower than 1.000 kCal/kg
L.H.V. (Kcal/kg)
5,000
Run of mine coal
4,000
+18 mm clean coal 10-18 mm clean coal
3,000
05-10 mm clean coal
2,000
Middiling product
1,000
Schist 01-05 mm clean coal JUNE
APRIL
FEBRUARY
OCTOBER
DECEMBER
JUNE
AUGUST
APRIL
FEBRUARY
OCTOBER
DECEMBER
JUNE
AUGUST
APRIL
FEBRUARY
OCTOBER
DECEMBER
JUNE
AUGUST
APRIL
FEBRUARY
OCTOBER
DECEMBER
JUNE
AUGUST
APRIL
FEBRUARY
0
Figure 11. Monthly average calorific values belonging to 200-2007-2008-2009-2010 years of raw coal, clean coal and schist The plant’s washing density varies between 1.63-1.72 gr/cm3 on the low density circuits and between 1.75-1.90 gr/cm3 on the high density circuits. The density range is related to raw coals property. Annual average ash and net calorific values of raw coal and washed coals are shown on Table 2. (Ash as percent, calorific value as kCal/kg). As seen from the graphics and the tables, all products are of adequate quality as required. Table 2. Annual average ash and net calorific values of raw coal for washed coal Years
Run of mine coal (tonnes)
ROM COAL Ash
+18 mm clean 10-18 mm clean coal coal
0.5-10 mm clean coal
Middling
Tailing
0.1-0.5 mm Clean coal
Calorie
Ash
Calorie
Ash
Calorie
Ash
Calorie
Ash
Calorie
Ash
Calorie
Ash
Calorie
2006
2,816,589.39 42.42
2543
15.41
4452
13.68
4528
12.08
4432
43.40
2348
65.99
467
23.65
3180
2007
4,060,494.00 42.84
2368
16.12
4445
14.03
4601
12.42
4538
39.28
2364
59.19
661
22.69
3239
2008
4,203,615.94 43.34
2467
16.29
4533
14.48
4646
12.23
4657
43.73
2430
63.76
559
23.33
3231
2009
4,191,520.00 43.91
2380
16.02
4631
13.45
4778
11.31
4738
41.17
2532
65.17
458
21.55
3740
2010 JULY
2,373,130.00 43.75
2,373
15.60
4,592
13.47
4,724
11.87
4,644
41.86
2,463
63.96
457
22.02
3,537
CONCLUSIONS In the Dereköy washing plant, density based separating devices were selected to upgrade low quality lignite coals in the Soma region of General Directorate of Turkish Coal The plant has delivered the expected quality at greater than projected tonnage levels. The project has been a great success. The calorific value of discarded tailings at 520 kCal/kg value, is considerably better than target of 1.000 kCal/kg at the time of establishment of the plant. 17.645.349,33 tonnes of coal has been washed until July 2010. It is planned that 2.3 million tones or coal is to be washed up to the year 2011 January. REFERENCES Alyildiz S.I., Ulu A., 2010 Soma Region’s Coals Washing at Dereköy Coal Washery and Performance Evaluation, 2010 ICPC, Lexington – Kentucky- USA Anaç, S., Mucuk, K., Ersun, E., Ergün, L. and Gülsoy, Ö., 2006. Studies to design a new washing plant for TKİ’s Soma region coals, Proc. of 21st IMPC, İstanbul (accepted for presentation). Gurkan S., Alyildiz S.I., Binol S., Coal Washing Plant With A Capacity Of 800 Tph In TKI-Soma Region, 2010 ICPC, Lexington – Kentucky - USA
Manuscript Not AVAILABLE
Topic: 5. Coal-Derived Products DEVELOPMENT OF HYDROGEN TRANSPORT MEMBRANES FOR SEPARATING HYDROGEN FROM COAL GASIFICATION STREAM U. (Balu) Balachandran ([email protected]; 630-252-4250) T. H. Lee ([email protected]; 630-252-5374) C. Y. Park ([email protected]; 630-252-3823) Y. Lu ([email protected]; 630-252-4193) S. E. Dorris ([email protected]; 630-252-5084) Energy Systems Division Argonne National Laboratory Argonne, IL 60439-4838
Hydrogen, the fuel of choice for both electric power and transportation sectors, can be produced from fossil and renewable resources by various technologies. Because it is produced in gas streams with numerous components, purification is a critical step in its production. Argonne National Laboratory is developing dense cermet (i.e., ceramic-metal composite) hydrogen transport membranes (HTMs) for separating hydrogen from coal gasification streams. Hydrogen separation with Argonne's HTMs yields high purity hydrogen, thereby eliminating the need for post-separation purification steps. HTMs were prepared by standard ceramic fabrication techniques, and their hydrogen permeation rate, or flux, was measured in the range of 400-900°C. A cermet membrane (thickness ≈18 m) on a porous support structure gave a maximum hydrogen flux of ≈52 cm3 (STP)/min-cm2 at 900°C in tests using 100% H2 at ambient pressure as the feed gas. We also measured the hydrogen flux through a thicker membrane (≈150 m) at 400°C using H2, CO, CO2, H2O, and He as feed gas at ≈200 psig. Because good chemical stability is critical for HTMs, due to the corrosive nature of product streams from coal gasification, we evaluated the effect of various contaminants on the chemical stability of cermet membranes. Hydrogen sulfide (H2S), a particularly corrosive contaminant, impedes hydrogen permeation through cermet membranes by reacting with them to form palladium sulfide (Pd 4S). To evaluate the chemical stability of membranes, the Pd/Pd4S phase boundary was determined in the temperature range ≈400-700°C in tests using various feed gases that contained 10-73% H2 and ≈8-400 ppm H2S. The Pd-containing cermets are stable between about 430°C and 680°C in gas stream containing 73% H2 with between about 60 and 400 ppm H2S. When the gas contains only 10% H2, the membrane is stable for H2S concentrations between about 8 and 50 ppm. We assessed the effect of syngas components on the Pd/Pd4S phase boundary by locating the phase boundary in feed gas that contained CH4, CO2, and CO in addition to 400 ppm H2S. The present status of membrane development at Argonne and the challenges involved in bringing this technology to fruition will be presented in this talk.
Work supported by the U.S. Department of Energy, Office of Fossil Energy, National Energy Technology Laboratory’s Hydrogen and Fuels Program, under Contract DE-AC02-06CH11357.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Completed Manuscript
PROGRAM TOPIC: COAL-DERIVED PRODUCTS HTGR-INTEGRATED COAL TO LIQUIDS PRODUCTION ANALYSIS 1 Anastasia M Gandrik, R&D Engineer, Idaho National Laboratory PO Box 1625, Idaho Falls, ID 83412-3710, [email protected], 412-443-6550 Rick A. Wood, R&D Engineer, Idaho National Laboratory PO Box 1625, Idaho Falls, ID 83412-3710, [email protected]
Abstract As part of the Department of Energy’s (DOE) Idaho National Laboratory (INL) nuclear energy development mission, the INL is leading a program to develop and design a high temperature gas-cooled reactor (HTGR), which has been selected as the base design for the Next Generation Nuclear Plant (NGNP). Because an HTGR operates at a higher temperature, it can provide higher temperature process heat, more closely matched to chemical process temperatures, than a conventional light water reactor. Integrating HTGRs into conventional industrial processes would increase U.S. energy security and potentially reduce greenhouse gas emissions (GHG), particularly CO2. This paper focuses on the integration of HTGRs into a coal to liquids (CTL) process, for the production of synthetic diesel fuel, naphtha, and liquefied petroleum gas (LPG). The plant models for the CTL processes were developed using Aspen Plus. The models were constructed with plant production capacity set at 50,000 barrels per day of liquid products. Analysis of the conventional CTL case indicated a potential need for hydrogen supplementation from high temperature steam electrolysis (HTSE), with heat and power supplied by the HTGR. By supplementing the process with an external hydrogen source, the need to “shift” the syngas using conventional water-gas shift reactors was eliminated. HTGR electrical power generation efficiency was set at 40%, a reactor size of 600 MWth was specified, and it was assumed that heat in the form of hot helium could be delivered at a maximum temperature of 700°C to the processes. Results from the Aspen Plus model were used to perform a preliminary economic analysis and a life cycle emissions assessment. The following conclusions were drawn when evaluating the HTGR-integrated CTL process against the conventional process:
1
•
When compared to conventional CTL production, HTGR integration decreases coal consumption by 65% using electrolysis and nuclear power as the hydrogen source. In addition, HTGR integration decreases CO2 emissions by up to 95% when compared to conventional CTL.
•
In order to support a 50,000 barrel per day CTL facility, 11 HTGRs (600 MWth each) are required.
Work supported by the U.S. Department of Energy, Office of Nuclear Energy (NE), under DOE Idaho Operations Office Contract DE-AC07-05ID14517.
•
The preliminary economic assessment indicates that the incorporation of HTGRs and the associated HTSEs impacts the expected return on investment, when compared to conventional CTL with or without sequestration. However, in a carbon constrained scenario, where CO2 emissions are taxed and sequestration is not an option, a reasonable CO2 tax would equate the economics of the HTGR-integrated CTL case with the conventional CTL case. The economic results are preliminary, as they do not include economies of scale for multiple HTGRs and are based on an uncertain reactor cost estimate. Refinement of the HTGR cost estimate is currently underway.
•
To reduce well to wheel (WTW) GHG emissions below baseline (U.S. crude mix) or imported crude derived diesel, integration of an HTGR is necessary. WTW GHG emissions decrease 8% below baseline crude with HTGR-integrated CTL. Even with CO2 sequestration, conventional CTL WTW GHG emissions are 24% higher than baseline crude emissions.
•
Current efforts are underway to investigate the incorporation of nuclear integrated steam methane reforming for the production of hydrogen, in place of HTSE. This could potentially reduce the number of HTGRs required for the process.
1.
Introduction
Under direction from the DOE, the Next NGNP project’s mission is to develop, design, construct, and operate a prototype plant to generate electricity, heat, and/or produce hydrogen. The prototype plant is based on HTGR technology. An HTGR produces and transfers energy in the form of high-temperature process heat. It differs from a third-generation light-water reactor by using helium instead of water as a coolant, graphite instead of water as the moderator, and tristructural-isotropic (TRISO) fuel instead of metal-clad fuel. With these features, an HTGR is capable of operating at higher temperatures, from 750 to 950°C, which offers a broader range of temperatures applicable to industrial processes and higher thermal efficiencies than are achievable with the lower operating temperatures of light water reactors. The capability of the HTGR for producing high-temperature process heat offers numerous practical and economic advantages, including: •
Reducing carbon dioxide and other greenhouse gas emissions by replacing the process heat derived from burning fossil fuels in a wide range of chemical and petrochemical processes, with the added capability for co-generation of electricity and hydrogen
•
Generating electricity at higher efficiencies than are possible with current nuclear power generation technology
•
Providing a secure, long-term domestic energy supply and reducing the dependence on unreliable or unstable sources for energy
•
Producing synthetic transportation fuels with lower life-cycle, WTW GHG emissions than fuels derived from conventional synthetic transportation fuel production processes and similar or lower WTW GHG emissions as fuels that are refined from crude oil
•
Producing energy at a stable long-term cost that is relatively unaffected by volatile fossil fuel prices and a potential carbon tax, a price set on carbon dioxide emissions
•
Extending the availability of limited natural resources for more valuable uses, such as using coal for use as a petrochemical feedstock
•
Providing benefits to the U.S. economy, including more near-term jobs and increased long-term employment associated with the construction of multiple plants and a reinvigorated heavy manufacturing sector.
The HTGR produces process heat (high-temperature helium), electricity, and/or hydrogen. A summary of these products and a brief description is shown in Table 1. For this study the HTGR outlet temperature is assumed to be 750°C, this reflects the initial HTGR design and assumes a more conservative outlet temperature, eventually temperatures of 950°C are anticipated. Additionally, a 50°C temperature approach is assumed between the primary and secondary helium loops. As a result, the helium stream available for heat exchange is assumed to be at 700°C. In conventional chemical processes heat is generated by the combustion of fossil fuels such as coal or natural gas, resulting in significant emissions of greenhouse gases, such as carbon dioxide. Heat, electricity, or hydrogen produced in an HTGR could be used to supply process heat, electricity, or hydrogen to conventional chemical processes without generating any GHGs. The use of an HTGR to supply process heat, electricity, or hydrogen to conventional processes is referred to as an HTGR -integrated process.
Table 1. Assumed outputs of the HTGR.
The following paper presents potential integration opportunities for coal to liquids production. First, an overview of the process modeling performed for the conventional and HTGR-integrated CTL is presented, followed by the modeling results, specifically the impact of HTGR integration on the material and energy balance. Next, a brief overview of the preliminary economic model is presented, followed by preliminary results for CTL and HTGR-integrated CTL. Finally, the method for calculating greenhouse gas emissions is discussed along with the WTW GHG emission results. 2.
Process Modeling Overview
The plant models for the CTL processes were developed using Aspen Plus. Because of the size and complexity of the processes modeled, the simulations were constructed using “hierarchy” blocks, a method for nesting one simulation within another simulation. In this fashion, submodels for each major plant section were constructed separately and then combined to represent the entire process. Rigorous submodels of the nuclear power cycle and HTSE have not yet been integrated; this integration is planned for the near future. Significant emphasis in the models has been placed on heat integration between different parts of the plant. To facilitate energy tracking, Aspen’s “utility” blocks were used extensively. Two cases were identified for modeling: •
Conventional CTL process
•
HTGR-integrated CTL process
A generic Illinois #6 coal was used as the feedstock. A dry-fed, entrained-flow, slagging gasifier (similar to a Shell, Uhde, or Siemens design) was selected as the gasification technology for this evaluation. Capacities for the coal cases were set to produce 50,000 barrels per day of liquid products. The general model descriptions are presented below. 2.1
Conventional CTL
The block flow diagram for the conventional CTL case is shown in Figure 1. The proposed process includes unit operations for air separation, coal milling and drying, coal gasification, syngas cleaning and conditioning, sulfur recovery, CO2 compression/liquefaction, FT synthesis, product upgrading and refining, power production, cooling towers, and water treatment. Each unit operation is briefly described below.
Figure 1. Block flow diagram for the conventional CTL process. •
Air Separation – Oxygen is produced via a standard cryogenic Linde type air separation unit (ASU) that utilizes two distillation columns and extensive heat exchange in a cold box. The oxygen product is used for gasification. In order to reduce the inert content in the synthesis gas, an O2 purity of 99.5% is specified. It should be noted that lower oxygen purity could be specified, such as 95%; however, the high purity oxygen is desired to minimize diluent nitrogen in the fuel synthesis loops. The nitrogen co-product from the ASU can be used for coal drying and transport, and as an inert gas to be used throughout the plant. The waste stream from the ASU is an O2-enriched air stream. A portion of the enriched air stream is used as feed to the Claus unit in place of air.
•
Coal Milling & Drying – Coal is pulverized to below 90 μm using a roller mill to ensure efficient gasification. Drying is accomplished simultaneously using a heated inert gas stream, nitrogen from the ASU. The gas stream removes evaporated water as it sweeps the pulverized coal through an internal classifier for collection in a baghouse. The Illinois coal is dried to 6% moisture. Nitrogen is also used as transport gas for the coal from the baghouse to the lock hoppers. Pressurized carbon dioxide, from the Rectisol unit, is then used to transport the dry, sized coal to the gasifier.
•
Gasification – The dry coal is gasified at 2,800°F in a dry-fed, entrained-flow gasifier. Although some heat is recovered in the membrane wall of the gasifier, the majority of the heat recovery is accomplished downstream of the gasifier in the syngas coolers, which generate steam used in other areas of the process. The syngas is further cooled by a water quench. A portion of the quenched syngas is returned to the top of the gasifier to cool the particle-laden gas to below the ash softening point.
•
Syngas Cleaning & Conditioning – After gasification, a fraction of the syngas is passed through sour shift reactors and then remixed with unshifted syngas to provide the optimal H2:CO ratio to the FT reactors which utilize a cobalt catalyst; a ratio of approximately two to one (H2:CO). Steam is added to the syngas stream to maintain the water concentration necessary for the water gas shift reaction. Elemental mercury is then captured in a mercury guard bed. The syngas is further treated in an absorber with refrigerated methanol which acts as a physical solvent for the removal of CO2, H2S, and COS, the selective Rectisol process. Because of the extreme sulfur intolerance of the Fischer-Tropsch (FT) catalyst, guard beds are included as an added measure of protection against poisoning. A portion of syngas is sent to a pressure swing absorption unit (PSA), where a pure hydrogen stream is produced for use in the refinery, for hydrocracking and hydrotreating, and the sulfur reduction unit, to reduce sulfur compounds to H2S.
•
Sulfur Plant (Claus) and Tailgas Sulfur Reduction – Sulfur recovery is based on the Claus process. Tail gas from the Claus unit is hydrogenated over a catalyst to convert the remaining sulfur species to H2S, and this stream is recycled to the Rectisol unit to maximize sulfur recovery. A small stream of clean syngas is used to fire and preheat the feed gas to the sulfur reduction unit.
•
CO2 Compression – Carbon dioxide is removed from the syngas in the Rectisol process. By properly designing the solvent regeneration scheme, a pure stream of CO2 is produced. The resulting stream is then compressed, along with the CO2 recycle from coal transport, and liquefied prior to being pumped to the required pressure for use in enhanced oil recovery or sequestration.
•
Fischer-Tropsch Synthesis – Syngas is converted to liquid synthetic crude in a slurry bubble column reactor utilizing a cobalt catalyst. Steam is generated from the exothermic FT reactions. The resulting product is primarily paraffinic, but also contains some olefins and oxygenates. The product distribution is estimated using a modified version of the Anderson Schulz Flory (ASF) distribution. Modifications are required to the classical ASF distribution to better match actual performance of FT catalysts, especially for carbon numbers between one and four. Carbon chain length in the product stream varies from one (methane) to more than 100; hence, separations are performed to fractionate the product into light gas, crude naphtha, middle distillate, and molten wax. To improve conversion a light gas recycle is implemented. Currently a single-stage slurry bubble column reactor is modeled; however, a two-stage reactor may improve conversion and reduce the amount of light gas recycled.
•
Product Upgrade & Refining – The middle distillate product is hydrotreated to saturate olefinic bonds. The hydrotreated product is refined via a combination of pressurized and vacuum distillation into naphtha and diesel fuel products. The bottoms product from vacuum distillation and the molten wax stream are hydrocracked to improve overall yield of the diesel and naphtha fractions. At present, no attempt is made to refine the naphtha fraction.
•
Power Production – Light gas from FT synthesis and refining areas is used to fire gas turbines to produce electricity. To increase power production, a combined cycle is utilized. Hot exhaust from the gas turbine is routed to the heat recovery steam generator (HRSG) to produce superheated steam. This steam is used in conventional condensing turbines to produce additional power. To further maximize power production, steam generated throughout the plant is sent to the power production block where they are passed through three saturated steam turbines. The
condensed steam from the turbine outlets are mixed with condensate return from the plant and makeup water is added to provide the necessary flow to the boiler feedwater pumps. Steam is added to the deaerator to achieve the appropriate temperature. Other minor unit operations that were included in the process models are:
2.2
•
Cooling Towers – Conventional cooling towers are modeled in Aspen Plus using literature data. Air cooling could potentially be used in certain areas of the plant to decrease water consumption; however, for simplicity cooling water only was assumed.
•
Water Treatment – Water treatment is simplistically modeled in Aspen Plus using a variety of separation blocks. Aspen transfer blocks are used to reconcile water in and out flows from various parts of the plant, allowing reconciliation of the entire plant water balance. HTGR-integrated CTL
The block flow diagram for the HTGR-integrated CTL case is shown in Figure 2. The proposed process includes the same unit operations as the conventional process with the following exceptions: the cryogenic ASU and water gas shift reactors are replaced by high-temperature steam electrolysis to provide oxygen and hydrogen for the process.
Figure 2. Block flow diagram for the HTGR-integrated CTL process. While developing the HTGR-integrated case, opportunities for heat integration between the nuclear, electrolysis, and fossil plants were also evaluated; however, very few opportunities were identified. The primary reason for this conclusion was that the fossil plant produced an excess of heat that could provide for the heat requirements within the fossil portion of the plant. In a few instances (notably product refining), it was believed that HTGR heat could displace burning of light gas to reduce overall plant greenhouse gas emissions. However, the modeling analysis indicated that light gas recycle would lead to unacceptable buildup of inert gases in the process. Hence, it was deemed practical to use this gas as fuel rather than develop complex schemes to separate inerts from the light gas. An opportunity to use heat from the gasifier as topping heat for the electrolysis unit was identified. However, use of this heat would require the exchanger to be constructed utilizing exotic materials to guard against metal dusting by carbon formed from the Boudouard reaction. To avoid this complication, syngas is fired to provide topping heat; however, this does increase CO2 emissions to the atmosphere. As HTGR technology matures and reactor outlet temperatures increase, the nuclear reactors may be able to
supply electrolysis topping heat of 800°C. However, because of the upper limit of 700°C deliverable heat assumed in this study, supplying topping heat from syngas firing to the electrolyzers is an attractive means of increasing electrolyzer efficiency. With the ASU and water gas shift reactors removed from the flowsheet, an unexpected result was observed. A shortage of inert gas for use in coal drying, transport, and feeding was created. To overcome this issue, air was selected for use in coal drying and transport, rather than nitrogen. Each unit operation in the HTGR-integrated CTL flowsheet is briefly described below. Because the majority of unit operations remain unchanged from the conventional CTL flowsheet, emphasis is placed on differences in configuration between the two cases. •
Electrolysis – Water is converted to hydrogen and oxygen utilizing high temperature electrolysis units. Hot helium is used to convert the water to steam and raise the temperature, while heat recuperated from the fired heater is used to provide topping heat to raise the steam temperature to 1,472°F for electrolysis. Electricity is provided to the HTSE unit from the HTGR. The oxygen generated is used for gasification and air enrichment for the Claus and sulfur reduction units, the hydrogen is used to adjust the hydrogen to carbon monoxide ratio for the FT reactor, in place of sour shift reactors.
•
Coal Milling & Drying – Coal milling and drying for the HTGR-integrated case is similar to the conventional case. However, because nitrogen is not readily available in this scenario, coal drying is accomplished using air. Air is also used as transport gas for the pulverized coal. Although air is used industrially for coal drying and transport, it introduces additional flammability issues compared to using an inert gas for this purpose, and therefore must be closely controlled. Transport of coal into the gasifier is accomplished using CO2 recovered from the Rectisol unit. The air for drying is heated using heat recovered throughout the process.
•
Gasification – Gasification for the HTGR-integrated case is similar to the conventional case. However, because hydrogen is supplied externally from the electrolyzers rather than shifting the syngas, the gasification island throughput is reduced to 35% of the conventional design to produce the same amount of liquid fuel product.
•
Syngas Cleaning & Conditioning – Syngas cleaning is greatly simplified for the HTGR-integrated case, because the water gas shift reactors are eliminated. Hydrogen from the electrolyzers is added to the syngas to achieve the optimal H2:CO ratio for the cobalt FT catalyst. Rectisol capacity and utility usage are reduced by more than half in the HTGR-integrated case as compared to the conventional case.
•
Sulfur Plant – The Claus and sulfur reduction plants for the HTGR-integrated case are similar to those in the conventional case. However, as with the gasification island, the required capacity of these units is less than half that of the conventional case configuration.
•
CO2 Compression – CO2 compression for the HTGR-integrated case is similar to CO2 compression in the conventional case. However, when the shift converters are eliminated, required capacity and utility usage are reduced. Additionally, the last stage of compression is removed, as all CO2 is recycled to the gasifier to increase carbon conversion to the liquid product.
•
Fischer-Tropsch Synthesis – The FT synthesis plant remains unchanged between the conventional and HTGR-integrated cases.
•
Product Upgrade & Refining – The product refining and upgrading process in the HTGRintegrated case remains almost unchanged from the process in the conventional case. However, the duty required for topping heat in the electrolyzers is supplied by the fired heater block in the
upgrading section. This fired heater would likely be located in the HTSE section of the plant; this location was used to simplify the model.
3.
•
Power Production – Power production in the HTGR-integrated case changes because the gas turbine system is removed, since the light gases are recycled to the gasification island. As a result there is no longer hot tailgas to superheat steam to use in the condensing steam turbines. Only the saturated turbines remain, being fed the steam generated throughout the plant. Due to size reductions in some portions of the plant, the capacity of the steam system in the HTGR-integrated case is less than half of the conventional case.
•
Cooling Towers – The cooling water system requirements are similar for both cases. Cooling water requirements for the HTGR are not included in this analysis.
•
Water Treatment – The water treatment system in the HTGR-integrated case is similar to the conventional case. No further comparison will be made on water treatment between the two cases until the water treatment scenarios have been tuned. Modeling Results
Analysis of the conventional CTL case indicated a potential need for hydrogen supplementation from HTSE. By supplementing the process with an external hydrogen source, the need to “shift” the syngas using conventional water-gas shift reactors was eliminated. The primary benefit of this change is a reduction in greenhouse gas emissions from the process. It was also determined that the conventional CTL case produced heat beyond what was needed to support demands of the plant. Based on these observations, an HTGR-integrated model was developed which focuses primarily on integrating nuclear hydrogen rather than nuclear heat. A summary of the modeling results for all cases is presented in Table 2. A high-level material and energy balance summary for each case is graphically presented in Figure 3.
Figure 3. CTL modeling case material balance summary.
Table 2. CTL modeling case study results.
Inputs Coal Feed rate (ton/day) % Carbon to Liquid Product # HTGRs (600 MWt) Outputs Total Liquid Products (bbl/day) Diesel Naphtha LPG Utility Summary Total Power (MW) Power Consumed Electrolyzers ASU Coal Milling and Drying Gasification and Gas Cleanup CO2 Compression/Liquefaction Fischer Tropsch & Refining Processes Refrigeration Cooling Tower Water Treatment Power Generated Gas Turbine Condensing Turbines Saturated Turbines Water Requirements 2 Water Consumed (gpm) Water Consumed/lb Feed (lb/lb) Water Consumed/bbl Product (bbl/bbl) CO2 Summary Total CO2 Produced (ton/day) Emitted Capturable HTGR Integration Summary Electricity (MW) HTSE Balance of Plant Electrolysis Heat (MMBTU/hr) From Nuclear Plant From Fired Heater Electrolysis Products Total Hydrogen (ton/day) Total Oxygen (ton/day) Used in Plant (ton/day) Excess (ton/day) 2
Does not include water requirements for the HTGR.
Conventional CTL
HTGR-integrated CTL
26,911 31.8% N/A
9,520 90.0% 10.87
49,992 35,455 12,571 1,976
49,998 35,007 12,189 2,802
252.9 -719.6 N/A -301 -13.8 -166 -140.6 -35.1 -23.9 -14.5 -18.7 972.5 299.8 184.4 488.3
-2,323.6 -2,727 -2,525.5 N/A -9.7 -78.8 -19.5 -41.1 -24.8 -10.7 -14.6 403.4 N/A N/A 403.4
19,696.9 4.40 13.5
15,425.7 9.73 10.6
40,008 8,793 31,215
1,874 1,874 N/A
N/A N/A N/A N/A N/A N/A
-2,323.6 -2,525.5 201.9 2,600.8 2,442.3 158.5
N/A N/A N/A N/A
1,966 15,490 9,076 6,414
4.
Economic Modeling
The economic viability of the CTL processes was assessed using standard economic evaluation methods. For all cases the fuel price necessary for an internal rate return (IRR) of 12% was calculated, as well as the fuel price for a range of carbon taxes at a 12% IRR. The following table outlines the assumptions that were made for the economic analyses:
Table 3. Economic modeling assumptions.
The economics were evaluated for the conventional and HTGR-integrated options described in the previous sections. Given the preliminary nature of the reactor cost estimate, the details of the economic modeling performed are not presented. A summary of the economic analysis trends are presented below for the carbon tax scenarios. The results will be updated when the reactor cost estimate is complete. 5.
Economic Modeling Results
The cost estimate for the HTGR used in the economic analysis is uncertain and does not currently include economies of scale for multiple HTGRs. As a result only the trends of the economic results are presented, specifically as a function of a tax on CO2 emissions. These results will be updated when the HTGR cost estimate update is complete. It is anticipated that this will improve the economic viability of the HTGR-integrated cases. Figure 4 presents a summary of the economic modeling results.
Figure 4. Diesel price as a function of CO2 tax. 6.
GHG Modeling
This section presents a full life-cycle inventory or well-to-wheel analysis of greenhouse gas emissions for the production of synthetic diesel using the conventional and HTGR-integrated CTL processes described in the preceding sections. For this study, all results are scaled for the diesel, naphtha, LPG, and/or electricity products. This is accomplished by ratioing the lower heating values of the products along with the electricity, if produced in the plant, to determine the emissions assignment, or the percentage of the total energy content for the diesel, naphtha, LPG, and/or electricity product. The emissions for the diesel product are converted to a gram per mile basis using a vehicle fuel economy of 25.8 miles per gallon. The fuel economy was adjusted to account for the heating value of the synthetic fuel versus traditional petroleum derived products. The vehicle fuel economy represents the average mileage of a diesel powered SUV. The GHG emissions considered include carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). Emissions for CH4 and N2O are converted into CO2 equivalents using their global warming potentials (GWP). CO2 equivalents are the amount of carbon dioxide by weight emitted into the atmosphere that would produce the same radiative force as a given weight of another radiatively active gas. The WTW analysis conducted for this study was based on the formal methodology presented by the National Technology Engineering Laboratory, in the “Life-Cycle Greenhouse-Gas Emissions Inventory for Fischer-Tropsch Fuels,” and categorizes GHG emissions according to the following sources: •
Resource Extraction and Production – GHG emissions for resource extraction are calculated for the coal consumed in the CTL processes. Coal extraction emissions include emissions from fuel usage associated with coal mining and coal bed methane. According to the Energy
Information Administration (EIA) in 2007 approximately 82% of the coal mined in Illinois was mined using underground mining methods, the remainder was surface mined. Fuel usage per ton of coal mined for both surface and underground mining were calculated based on the most recent U.S. Census data available. Based on this census data, power, coal, diesel, residual fuel oil, natural gas, and gasoline usage for mining activities were calculated. The associated CO2 emissions were calculated based on the lower heating values (LHV) and carbon contents of the various fuel types, for power the emissions for the average U.S. energy mixed were assumed. Emissions for CH4 and N2O were calculated assuming either mobile or stationary combustion emission factors from the 2006 Intergovernmental Panel on Climate Change (IPCC) report. Emissions for mining support activities were calculated in a similar fashion. Finally, coal bed methane emissions are calculated for the methane released during Illinois mining operations based on Environmental Protection Agency data. •
Transportation and Distribution – All scenarios considered include transportation of resources and products over large distances. The mode of transportation depends upon the location and destination of the products as well as the type of product being transported. For instance, dry materials being transported short distances would utilize trucks as the main mode of transportation, while dry materials being transported long distances would take advantage of rail transportation. The modes of transportation were assumed based on the amount of product being transported, the product state, the distance transported, and the available transportation methods. The emissions associated with the various transportation methods include the combustion of fuel necessary for the transportation (or electricity use) as well as the upstream emissions associated with producing the fuel or electricity.
•
Conversion and Refining – GHG emissions are generated from several sources within the conventional and HTGR CTL plants, including: emissions from importing power, emissions associated with nuclear power use, upstream emissions associated with methanol use, emissions from coal milling and drying, Rectisol plant emissions, HRSG stack emissions, fired heater emissions, high pressure and low pressure flare systems, and fugitive emissions. Fugitive emissions are emissions from leaking equipment, such as valves and pumps, storage tanks, and wastewater treatment facilities.
•
End Use Combustion – Emissions for the end use combustion of the fuel were estimated from the carbon content of the synthetic diesel. It was assumed that all carbon present in the fuel is completely combusted to form CO2. Based on the fuel density, this would provide the emissions of CO2 per barrel of fuel. Emissions for CH4 and N2O were added based on IPCC guidelines for mobile combustion sources.
Results from the WTW analysis for FT diesel were compared to WTW emissions for the U.S. baseline and average imported WTW emissions for conventional diesel fuel to determine the environmental impact of the synthetic fuels in comparison to standard petroleum fuels. 7.
GHG Modeling Results
A summary of the GHG results for the cases analyzed is presented in Table 4 for conventional and HTGR CTL diesel. GHG emissions results are presented on a gram CO2 equivalent per barrel of diesel fuel (g CO2-eq/bbl) basis, a gram CO2 equivalent per LHV (g CO2-eq/MMBTU) basis, and a gram CO2 equivalent per mile (g CO2-eq/mile) basis. GHG emissions results are presented in Figure 5 for the CTL diesel cases.
Table 4. CTL WTW GHG results.
gCO2-eq/bbl diesel Extraction and Production Transportation to Plant Conversion and Refining Transportation to Pump End Use Combustion Total Fuel Chain gCO2-eq/MMBTU diesel Extraction and Production Transportation to Plant Conversion and Refining Transportation to Pump End Use Combustion Total Fuel Chain gCO2-eq/mile Extraction and Production Transportation to Plant Conversion and Refining Transportation to Pump End Use Combustion Total Fuel Chain
CTL
CTL w/ Seq
HTGR CTL
Baseline Diesel
Imported Diesel
40,450 1,446 674,538 4,357 360,375 1,081,164
41,856 1,496 158,177 5,940 360,375 567,843
16,018 572 71,725 4,357 360,375 453,047
35,894 7,070 51,666 4,895 439,910 539,434
45,683 9,245 57,104 4,351 439,910 556,293
8,379 299 139,724 903 74,648 223,953
8,670 310 32,765 1,230 74,648 117,623
3,318 119 14,856 902 74,642 93,836
6,600 1,300 9,500 900 80,888 99,188
8,400 1,700 10,500 800 80,888 102,288
42 2 701 5 375 1,124
44 2 164 6 375 590
17 1 75 5 375 471
33 7 48 5 406 498
42 9 53 4 406 513
Figure 5. CTL WTW GHG results.
8.
CTL Conclusions
Results from the HTGR-integrated CTL case indicate that integration of nuclear hydrogen can drastically improve carbon utilization and reduce GHG emissions: •
Coal consumption is decreased by 65% using electrolysis and nuclear power as the hydrogen source.
•
Integrating nuclear power and HTSE decreases CO2 emissions from the plant:
•
•
If carbon capture and sequestration are assumed for the conventional configuration, CO2 emissions decrease by 79% when electrolysis and nuclear power are utilized.
•
If carbon capture and sequestration are not assumed for the conventional configuration, CO2 emissions decrease by 95% when electrolysis and nuclear power are utilized.
In order to support a 50,000 barrel per day HTGR-integrated CTL facility, 11 HTGRs (600 MWth each) are required.
Economically, the incorporation of HTGRs and the associated HTSEs impacts the expected return on investment, when compared to conventional CTL with or without sequestration: •
When the HTGR capital cost is decreased by 30%, the target reactor cost, the HTGR-integrated case become more economically competitive.
•
In a carbon constrained scenario, where CO2 emissions are taxed and sequestration is not an option, a reasonable CO2 tax would equate the economics of the HTGR-integrated CTL case with the conventional CTL case.
•
The economic results are preliminary, as they do not include economies of scale for multiple HTGRs and are based on an uncertain reactor cost estimate. Refinement of the HTGR cost estimate is currently underway
Integration of the HTGR reduces WTW GHG emissions to levels below imported and/or baseline conventional diesel. •
Conventional CTL WTW emissions are significantly higher than conventional diesel. Even with incorporation of sequestration, conventional CTL WTW emissions are greater than conventional fuels.
•
HTGR integration reduces WTW GHG emissions of coal based synthetic fuels below conventional fuel, without CO2 sequestration.
•
If future policy mandates that synthetically produced diesel fuels must meet or beat current fuel WTW GHG emissions; HTGR integration provides a viable solution.
9.
Future Work
The following items should be performed in the future to further refine the process and economic modeling performed for the CTL cases: •
Refine the HTGR capital cost estimate and include the refined HTGR capital cost in the economic model •
This will potentially reduce the cost of multiple module scenarios, including HTGRintegrated CTL, which would drastically decrease the carbon tax required to achieve breakeven with the conventional process
•
Consider independent owner/operator arrangements and financing of the fossil, HTSE, and HTGR processes
•
Investigate the impact of increasing the reactor outlet temperature to 950°C, which will improve power cycle efficiency and eliminate the need to fire syngas for topping heat in the HTSE •
•
The number of reactors required for HTGR-integrated CTL would be reduced, which will decrease the carbon tax required to achieve break-even with the conventional process.
Investigate the incorporation of nuclear integrated steam methane reforming for the production of hydrogen, in place of HTSE, and eliminate the need to for HTSE •
The number of reactors required for nuclear-integrated CTL could potentially be reduced, which will again decrease the carbon tax required to achieve break-even with the conventional process.
10.
Acronym List
ASF
Anderson Schulz Flory
ASU
Air Separation Unit
CTL
Coal to Liquids
DOE
Department of Energy
EIA
Energy Information Administration
FT
Fischer-Tropsch
GHG
Greenhouse Gas
GWP
Global Warming Potential
HRSG Heat Recovery Steam Generator HTGR High Temperature Gas Reactor HTSE High Temperature Steam Electrolysis INL
Idaho National Laboratory
IPCC
Intergovernmental Panel on Climate Change
IRR
Internal Rate of Return
LHV
Lower Heating Value
LPG
Liquefied Petroleum Gas
NETL National Energy Technology Laboratory NGNP Next Generation Nuclear Plant PSA
Pressure Swing Absorption
TRISO Tristructural-Isotropic WTW Well to Wheel
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: 5. COAL-DERIVED PRODUCTS EXACT TITLE OF PAPER: METHANE PRODUCTION FROM COAL, COAL-BIOMASS MIXTURES Arun SK Raju, Director of Research, Viresco Energy LLC 1401 Research Park Dr., Suite 400, Riverside, CA – 92507 [email protected] Ph: 951-236-7777, Fax: 951-784-7287 Christophe Capelli, Research Assistant, Viresco Energy LLC 1401 Research Park Dr., Suite 400, Riverside, CA – 92507 Ph: 951-784-7238, Fax: 951-784-7287
Abstract: Viresco Energy has developed an innovative conversion technology that can produce methane (Synthetic Natural Gas) from carbonaceous feedstocks such as coal, biomass, municipal and green waste and biosolids. This technology uses the Steam Hydrogasification Reaction (SHR) based gasification process invented at the College of Engineering – Center for Environmental Research & Technology (CE-CERT) at the University of California Riverside (UCR). SHR gasification uses steam and hydrogen to convert feedstocks into high energy content product gases under relatively modest temperatures and pressures. This gasification process has several advantages over conventional partial oxidation or air blown gasifiers. The process configuration for methane production involves the utilization of the innovative SHR gasifier with a shift reactor. The slurry made of the carbonaceous feed and water, along with the recycled hydrogen is fed to the SHR gasifier. The SHR generates a high methane content product gas that is subjected to warm gas cleanup in order to remove contaminants such as sulfur. The clean product gas is then fed into a water gas shift reactor. In the shift reactor, the CO present in the clean product gas reacts with the steam to produce H 2. Methane This product gas is then cooled down and H2 is separated for recycle to the SHR as feed. The recycle hydrogen stream eliminates the hydrogen supply problem. The final product gas contains high quantity of methane. Gasification experiments have been conducted using lignite coal and lignite-biomass mixtures. The experimental work was conducted in a pressurized kiln type continuous flow steam hydrogasifier. The product gas was passed through a Non-Dispersive Infrared Sensor (NDIR) and a Gas Chromatograph (GC). After each test, the un-reacted char and ash were collected and weighed. The experiments were conducted over a temperature range of 650 to 850 C. All the experiments were conducted at 150 psi operating pressure and at a 2:1 ratio of H2O to feed mass ratio. The experimental data including the carbon conversion and the gas composition information will be presented. Aspen Plus based simulations have been performed in order to evaluate the impact of parameters on the process efficiency. The simulation results will also be presented.
1
Methanation of Syngas over Coral Reef-Ni/Alumina Catalysts 1
Shengli Ma,1,2 Yisheng Tan,1 and Yizhuo Han*1 State Key Laboratory of Coal Conversion, Institute of Coal Chemisty, Chinese Academy of Sciences, Taiyuan 030001, P. R. China 2 Graduate University of the Chinese Academy of Sciences, Beijing 100049, P. R. China (*E-mail: [email protected])
Coral reef-Ni/alumina catalysts prepared by coprecipitation method was used for methanation of syngas. The Ni/Al2O3 catalysts were tested in a continuous flow type fixed-bed reactor. The experimental results show that the Ni/Al2O3 catalyst calcined at 673 K exhibited better activity than those calcined at 573 K and 773 K. Under the reaction conditions of H2/CO(molar ratio)=3:1, 593 K, atmospheric pressure and 2500 h-1, CO conversion and CH4 selectivity reached to 98.8% and 86.9%, respectively. The structure and properties of fresh and used catalysts were analyzed by SEM, XRD and H2-TPR and BET techniques.
As the distribution and consumption of coal resources are highly uncoordinated and the consumption proportion of natural gas increases year by year, coal-to-natural gas technology is being spotlighted greatly in China.1,2 In comparison with coal, natural gas has more advantages, such as high calorific value, complete combustion, smoke-free and slag-free. The main component of natural gas is methane, which can be synthesized by relatively abundant coal resources in China. The application of methane can be beneficial to both pollutant disposal and environmental protection, and provides clean fuel for urban residents and industry. Methanation products of coal-to-syngas can decrease transport costs to associate with the natural gas transmission pipeline. By now, Ni- and Ru- based catalysts have exhibited very high activity for selective methanation of CO commonly. 3,4 Because of cheapness of Ni metal, methanation of CO was often carried out in H2-rich gases over Ni-based catalysts prepared by impregnation method, dry-mixing method or coprecipitation method.5-7 However, the activities of Ni-based catalysts decrease by carbon deposition, sintering, Ni loss and poisioning.8 Many researchers have attempted to solve these negative factors.9,10 In our previous work, the Ni/Al2O3 catalyst prepared by coprecipitation method has higher BET surface area than grind-mixing method and impreganation method.11 In this study, we further report the catalytic performance and characterization of Ni/Al2O3 catalysts prepared by coprecipitation method and calcined at different temperatures for methanation of syngas. Catalysts were prepared as follows: adequate nickel acetate and aluminium nitrate were dissolved in ethylene glycol (EG), keep the temperature at 393 K for 30 min. Then the 0.2 mol·L-1 Na2CO3 aqueous solution was slowly added to the Ni2+-Al3+-EG solution. The precipitates of Ni(OH)2 and Al(OH)3 were aged in the liquid for 1 h. After being filtered and washed thoroughly with distilled water, the precipitates were dried at 393 K for 8 h and calcined in air at the desired temperatures (573 K-773 K) for 4 h. 15wt% Ni/Al2O3 was shaped and broken into 20-40 mesh catalysts for reserve.
Depending on the calcination temperature, the obtained catalyst was designed as Ni/Al2O3-T, where T refers to the temperature of calcinations (in K). Methanation of syngas was tested in a continuous flow fixed-bed reactor. The catalyst (1.0 ml) was diluted with ground quartz to prevent the over-heating of the catalyst due to the exothermic reaction. The catalyst was reduced in a flow of H2 (20 ml·min-1) for 1 h at 673 K before the reaction. CH4 and other reaction products were analyzed using a GC-9A gas chromatograph (Shimadzu Co., Japan) equipped with a GDX403 column and flame ionization detector. H2, CO, CH4 and CO2 were monitored using a GC-4000A gas chromatograph (East & West Analytical Instruments, Inc., China) with a TDX-01 column and thermal conductivity detector. The composition of reaction products was obtained by CH 4 normalization method. SEM micrographs were taken by a NOVA NANO SEM 430 scanning electron microscope. X-ray diffraction (XRD) patterns were measured on a Bruker Advanced X-Ray Solution/D8-Advanced using Cu Kα radiation. Temperature programmed reduction (TPR) was carried out in a fixed-bed reactor system equipped with a gas chromatograph. The catalyst (0.10 g) was pretreated at 673 K under an Ar flow (34 ml·min-1) for 2 h and cooled to 313 K. Then the TPR spectra were recorded at a temperature rising rate of 5.5 K·min-1 from 313 K to 973 K under 10vt% H2/Ar flow. Surface areas of the samples were measured by the nitrogen adsorption method at 77.35 K using a TriStar 3000 machine. Total pore volume was analyzed by single point adsorption. Table 1. Methanation of syngas over 15wt% Ni/Al2O3 catalysts CO
Selectivity (C-mol%)
Catalyst conversion (C-mol%)
CH4
CO2
Other H.C.
Ni/Al2O3-573
97.4
85.7
13.7
0.6
Ni/Al2O3-673
98.8
86.9
12.7
0.3
Ni/Al2O3-773
69.1
88.1
9.6
2.3
Reaction conditions: T=593 K, atmospheric pressure, t=20 h, GHSV=2500 h-1.
Table 1 shows the catalytic performances of Ni/Al2O3 catalysts for methantion of syngas. Ni/Al2O3-673 catalyst exhibits higher CO conversion and CH4 selectivity than Ni/Al2O3-573. As the calcination temperature rises from 673 K to 773 K, CH4 selectivity increases slightly, but CO conversion decreases from 98.8% to 69.1%. CO2 selectivity decreases in the order of Ni/Al2O3-573 > Ni/Al2O3-673 > Ni/Al2O3-773. CO2 is produced by water-gas shift (WGS) from CO and H2O in the process of methanation of syngas. So the intensity of WGS can be reflected by CO2 selectivity, and the intensity order is Ni/Al2O3-573 > Ni/Al2O3-673 >
Ni/Al2O3-773. For CH4 selectivity the order is just the reverse. It indicates that methanation competes with WGS. However, CO conversion of Ni/Al2O3-673 is the highest, which also indicates that the two kinds of reactions cooperate with each other. On the whole, Ni/Al2O3-673 exhibits the best activity. b
a
c
K. It indicates that the samples calcined at high temperature are difficult to be reduced. Ni/Al2O3-573 pretreated at 610 K appears the characteristic peak of NiO at about 600 K, while Ni/Al2O3-673 and Ni/Al2O3-773 pretreated at 673 K do not present NiO peak but only NiAl2O4 peak. It further indicates that the pahse changes have occurred in temperature rising process of 610 K-673 K. Table 2. Structure properties of Ni/Al2O3–T catalysts The fresh catalyst
Figure 1. SEM images of Ni/Al2O3 catalysts. a: Ni/Al2O3-573, b: Ni/Al2O3-673, c: Ni/Al2O3-773
The SEM images of Ni/Al2O3 catalysts calcined at different temperatures are presented in Figure 1. The morphologies of Ni/Al2O3 catalysts are coral reef shapes. It is found that more dropwells and pores are developed on the surface of catalysts as the calcination temperature increase from 573 K to 773 K. ¡ô NiAl 2O4 ¡î HAlO 2 ¡ô
¡ô
B
¡ô c
¡î ¡î
¡î
¡î
¡ô NiAl 2O4
¡ô
¡ô
Intensity
Intensity
A
¡ô
b
¡î
c b a
a 10
20
30
40
50
2¦È/
60
70
80
¡£
10
20
30
40
50
60
70
80
2¦È/¡ã
Figure 2. XRD patterns of Ni/Al2O3 catalysts. A: the fresh catalyst; B: after reaction 20 h a: Ni/Al2O3-573, b: Ni/Al2O3-673, c: Ni/Al2O3-773
Figure 2 shows the XRD patterns of Ni/Al2O3 catalysts. From Figure 2-A, the peaks of NiAl2O4 are not found on Ni/Al2O3-573. With the increase of calcination temperature, NiAl2O4 peaks appear, and the intensities become stronger. It indicates that NiAl2O4 was not formed under 573 K, few was formed at 673 K and a bit more was formed above 773 K, this is agreement with the phase temperature of NiAl2O4.12 From Figure 2-B, there is no peak of Ni but NiAl2O4 easily found after 20 h reaction, and the peak intensity increases with the calcination temperature. It may be caused by the agglomeration of NiAl2O4. Besides, it may occur the phase changes in the reduction process of the catalyst.
Intensity
a ¡ü NiAl2O4
400
600
800
1000
Figure 3. H2-TPR profiles of Ni/Al2O3 catalysts. a: Ni/Al2O3-573, b: Ni/Al2O3-673, c: Ni/Al2O3-773
Figure 3 compares the H2-TPR profiles of Ni/Al2O3 catalysts calcined at different temperature. H2-TPR can be used to investigate the reducibility of the catalyst.13 As can be seen from Figure 3, the reduction peak of NiO lies at around 600 K.14 The peaks of NiAl2O4 are easily found at 750 K-850
673 K
773 K
573 K
673 K
773 K
265.6
252.0
237.3
210.6
225.4
219.4
0.276
0.301
0.327
0.325
0.322
0.312
4.164
4.786
5.518
6.178
5.720
5.680
The work is financially supported by Independent Research Subject (2008BWZ005) by Ministry of Science and Technology of China.
References 1 2 3 4 5 6 7 8 9 10 11
14 15
T /K
573 K
Structure properties of Ni/Al2O3-T catalysts are listed in Table 2. With the calcination temperatures increased from 573 K to 773 K, the structure properties vary differently before and after reaction. After 20 h reaction, BET surface area of Ni/Al2O3-673 is the highest. This could partly explain why Ni/Al2O3-673 exhibits the better catalytic activity in accordance with the results shown in Table 1. Thus, high BET surface area, as one of the factors, helps to increase the activity of the sample and promotes methanation of syngas.15
12 13
c b
NiO ¡ý
BET surface area A (m2/g) Total pore volume v (cm3/g) Average pore diameter d (nm) d=.4v/A
After 20 h reaction
X. Q. Mao, X. R. Guo, Y. G. Chang, Y. D. Peng, Energy Policy 2005, 33, 307. W. Z. Zhao, H. J. Wang, K. Qian, Petrol. Develop. 2009, 36, 280. S. Takenaka, T. Shimizu, K. Otsuka, Inter. J. Hydro. Energy 2004, 29, 1065. P. Panagiotopoulou, D. I. Kondarides, X. E. Verykios, Appl. Catal. 2008, 344, 45. L. T. Luo, S. J. Li, G. F. Deng, J. Fuel. Chem. Tech. 2001, 29, 302. L. Xu, J. H. Yu, J. Chem. Fert. Ind. 2005, 32, 13. W. S. Zhang, W. Dai, X. L. Wang, H. Y. Wang, Petrochem. Technol. 2005, 34, 115. Y. Y. Ji, W. Z. Li, H. Y. Xu, Y. X. Chen, C. Y. Yu, F. Y. Yu, Chin. J. Catal. 2000, 21, 415. T. H. Wu, S. B. Li, Q. G. Yan, Chin. J. Catal. 2001, 22, 501. Y. Lu, Y. Liu, S. K. Shen, J. Catal. 1998, 177, 386. S. L. Ma, Y. S. Tan, Q. D. Zhang, Y. Z. Han, Nat. Gas Chem. Ind. 2009, 34, 1. Y. X. Han, Petrochem. Technol. 1980, 12, 698. R. J. Shang, W. Z. Sun, Y. Y. Wang, G. Q. Jin, X. Y. Guo, Catal. Commun. 2008, 9, 2103. Y. Li, X. W. Xie. J. L. Liu, M. Cai, J. Rogers, W. J. Shen, Chem. Eng. J. 2008, 136, 398. Q. D. Zhang, Y. S. Tan, C. H. Yang, Y. Z. Han, J. Mol. Catal. A: Chem. 2006, 263, 149.
Manuscript Not AVAILABLE
SULPHUR CAPTURE UNDER FLUIDISED BED COMBUSTION CONDITIONS USING COAL ASHES AS SORBENTS Rufaro Kaitano*a, Dursman Mchabea, Raymond C Eversona and Hein W J P Neomagusa, a
Energy Systems Research Group, School of Chemical and Minerals Engineering, North-West University, Potchefstroom Campus, Private Bag X6001,Potchefstroom 2520, South Africa * Corresponding author: Tel: + 27 18 299 1664 Fax: + 27 18 299 1535 E-mail:[email protected]
Keywords: Coal ash; Sulphur retention and capture; Mineralogy; Desulphurisation kinetics _________________________________________________________________________________ ABSTRACT An investigation to determine whether coal ash can be used for reduction of sulphur emission from coal combustion was carried out. The experimental work involved the determination of sulphur retention and capturing capacities of three ashes derived from typical South African low grade coals. The coals under investigation had ash content which varied between 37 and 46 wt %, reactive calcium oxide between 1.48 and 4.13 wt % and total sulphur (organic and inorganic) within the range 0.70 to 1.90 wt. %. The sulphur capture experiments were carried out under isothermal (750ºC and 900ºC) and atmospheric pressure conditions. The simulation to assess the desulphurization potential of the ashes was carried out with a typical flue gas mixture composed of 3000 ppm SO2, 8% CO2, 8% O2 concentration and balance N2. Reactivity measurements were carried out with a Thermogravimetric analyser. A detailed examination of phase transformation of the calcium containing minerals during combustion and desulphurization was executed in order to identify and quantify reactive phases for the evaluation of sulphur retention and desulphurization properties of the coal ashes. Mineral analyses of coal and ash samples were done by QEMSCAN. The transformation of a large fraction of the calcium bearing minerals to sulphates is evident with total sulphur retention of between 55.3% and 66.9 % of the total sulphur present in the parent coal. The transformation of all the other minerals is also evident with the formation of calcium containing non-crystalline phases with no sulphur capturing properties. The results showed that approximately 40% of the reactive calcium was converted within the first 90 minutes of reaction at 900oC. The shrinking core model with diffusion through the product layer as the controlling mechanism was found to describe the reaction.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 PROGRAM TOPIC: COMBUSTION EXACT TITLE OF PAPER: CO2 REDUCTION POTENTIAL AND CO-COMBUSTION POSSIBILITIES OF THE FBC-BOILERS ON THE CZECH CONDITIONS. Presenting Author Dagmar Juchelkova, prof. Ph.d. Contact Information VSB-Technical University of Ostrava, 17. listopadu 15, Ostrava, Czech Republic, [email protected] Helena. Raclavska, Jiri Bilik, Pavlina Pustějovská Contact Information VSB-Technical University of Ostrava, 17. listopadu 15, Ostrava, Czech Republic, [email protected]
Abstract: At present the task of minimizing carbon dioxide emissions in relation to its influence on environment belongs to the priorities of EU research activities. For achieving the best possible results it is necessary to focus attention on information concerning input materials character study of production as well as manufacturing processes and subsequent returning the products back to environment (anthroposphere). The problem is very extensive and covers many fields. Problem of CO2 reduction is one from the EU priority in longtime context (2010 and further). The aim of research in our University is large scale experiments in the fluidized-bed boilers. The experiments are carried out for Czech brown coal, wood, sewage sludge and wastes including analyses and recommendations for optimal thermal utilization and minimizing CO2 and harmful emissions. The next step is thermal analyses of coal, alternative fuel- wood pellets and sewage sludge from treatment plant. From the results of experiments it is clear that alternative fuels can be used in the large fluidized-bed boilers in the Czech Republic.
INTRODUCTION The transformation processes in the Czech Republic and especially in the North Moravia Region bring with also the necessity of CO2 emissions reduction. The promotion of sustainable development enabling competitiveness and employment necessitate decoupling economic growth from environmental degradation. Minimizing the environmental impact of the cost-effective production and use of energy in Europe will help preserve the ecosystem by reducing emission at local and global levels and by increasing the share of new and renewable energy sources in the energy system. It will also have socio-economic impacts by enhancing the capability of European industry to compete in the world markets, helping to secure employment and promoting social cohesion with less favored regions. The situation can you see from left and below figures. The change of energy production can be taken from figure below. Very important is also the structure of the industrial and public structure in the production of CO2.
CO-COMBUSTION OF VARIOUS BIOWASTES WITH A HIGH-SULFUR TURKISH LIGNITE IN A CIRCULATING FLUIDIZED BED COMBUSTOR Murat Varola, Mustafa Can Çelebib, Hayati Olgunc, Aysel T. Atimtaya*, Hüsnü Ataküld, Alper Unluc, Berrin Bayc, Ufuk Kayahanc a
Middle East Technical University, Department of Environmental Engineering, 06531 Ankara, Turkey b Istanbul Technical University, Energy Institute, Maslak 34469, Istanbul, Turkey c TUBITAK-MRC, Energy Institute, P.O.Box 21, Gebze 41470, Kocaeli, Turkey d Istanbul Technical University, Department of Chemical Engineering, Maslak 34469, Istanbul, Turkey *Corresponding author: [email protected]
Abstract— In this study, combustion and co-combustion of two biomasses and a Turkish lignite coal was carried out in a circulating fluidized bed combustor. A lignite coal which has high sulfur content and two biomasses were used in the experiments. The biomasses were hazelnut shells and woodchips. The combustion system consists of a circulating fluidized bed combustor column, a fuel feeding system, electrical heaters, and two cyclones. Its thermal capacity is 30 kW. The combustor column has an inside diameter of 108 mm and a height of 6 m. The temperatures along the column are observed with thermocouples located at specific heights. The temperature of the column is kept at about 850 oC during the combustion experiments. The pressure drops along the combustor column, cyclone, downcomer, and loopseal are continuously measured and observed in order to determine the solid mass flux within the combustor. A series of combustion tests for each fuel and mixture was performed in order to investigate the effect of excess air ratio on the flue gas emissions. During the combustion experiments, CO 2, CO, O2, NO, and SO2, emissions in the flue gas was continuously measured by ABB-AO2000 Advanced Optima continuous gas analyzer and recorded. The results of the experiments showed that as the biomass ratio in the fuel mixture increases for cocombustion, the combustion takes place more in the freeboard of the main column. Therefore, the maximum temperatures are seen in the freeboard rather than in the bed. Also the CO emissions increase as the biomass percentage increases in the fuel. Biomass fuels have high CO emission which indicates that a secondary air addition is required for the system. Secondary air injection into the freeboard may be a good solution to decrease the CO and also hydrocarbon emissions and to increase the combustion efficiency.
Keywords-component; Circulating fluidized bed (CFB), combustion of coal in a CFB, emissions, biomass and coal cocombustion
I.
INTRODUCTION
Nowadays, energy sector has been faced with two crucial problems: Installation of power plants with huge capacities and global warming threat due to carbon dioxide emissions from these power plants. Increasing population together with the demands for better welfare of human beings pose a big challenge to be overcome in the foreseeable future. Energy production based on fossil fuels and other anthropogenic greenhouse gas emissions represent 65% of the global greenhouse emissions [1]. The electricity production capacity that needs to be installed in order to meet the future energy demand would necessitate the installation of as much new power generation plants as was installed the last century [2]. Due to its abundance in several parts of the world, coal is expected to be the dominant fuel in the future power generation mix. According to the World Coal Institute, coal accounts for 41.5% of the global electricity production and 26.5% of the primary energy consumption in 2007 [3]. Despite being abundant and proven as one of the most reliable power generation feedstock, coal combustion has several adverse effects on human health and environment. From the global warming and sustainable energy production perspective clean coal combustion turns out to be very important solution for meeting world’s growing energy demand. Among the combustion technologies, circulating fluidized bed combustion (CFBC) is considered as a flexible technology, which makes it possible to burn fuels of widely varying quality. Besides different coal qualities including high sulphur or high ash coals, low grade fuels such as various types of biomasses and waste derived fuels have become highly interesting as feed stocks to the CFBC. Many installations also burn mixtures of different fuels [4]. Fluidized bed technology has many other advantages in addition to fuel flexibility as compared to pulverized coal combustion technology. High heat transfer rate is one of the other advantages. Thus, heat exchangers used in the fluidized beds require smaller surface areas. In situ SO 2 removal in the combustor is possible by using limestone. Also due to lower combustion temperature (850-870 oC) in the combustor column the thermal NOx formation is minimized. It has a compact design as compared to pulverized coal technology because of lower surface area and volume requirement per unit energy produced. Combustion of mixtures of different fuels, particularly biomass materials with coal, in a fluidized bed combustor is a promising application to produce cleaner energy as well as to dispose waste products of biomass resources. In addition to that, combustion of biomass resources as a supplementary fuel for coal-fired power plants has an advantage of saving money for the fuel cost. Moreover, co-combustion of biomass with coal is also an effective way to reduce CO2 emissions [5]. Youssef, et al. [6] conducted the combustion of four different kinds of biomass in a circulating fluidized bed with 145 mm inner diameter and 2 m height. Different excess air ratios were utilized. SO 2 and CO emissions were measured. Desroches-Ducarne, et al. [7] investigated the emissions of NO, N 2O, HCl, SO2 and CO during the combustion of coal and municipal refuse mixtures in a circulating fluidized bed boiler. Cliffe and Patumsawad [8] studied the co-combustion of olive cake with coal. They used fluidized bed combustor to research the feasibility of using olive cake as an energy source. Armesto, et al. [9] investigated the cocombustion of coal and olive oil industry residues in fluidized bed. Two different Spanish coals, lignite and anthracite were used for the study. Topal, et al. [10] investigated the olive cake combustion in a CFB. They used a small scale circulating fluidized bed of 125 mm diameter and 1800 mm height. The results obtained from that combustor were used to compare them with the coal combustion. Permchart and Kouprianov [11] made an experimental study about the combustion of three different biomass sources in a single fluidized bed combustor (FBC). Sawdust, rice husk and per-dried sugar cane bagasse were used as a biomass fuel. Gayan, et al. [12] analyzed the circulating fluidized bed combustion performance burning coal and biomass together. Two kinds of coal and a forestry residue (pine bark) were used in experimental studies. Two CFB pilot plants with the power 0.1 and 0.3 MWth, respectively were used. Atimtay and Topal [13] studied the co-combustion of olive cake with lignite coal in a CFB. Experiments were conducted in the same bed used in the previous work by Topal, et al. [10]. Shimizu and Toyono [14] made a research to investigate NO x and N2O emissions during combustion and co-combustion of dried sewage sludge with coal in a bench scale circulating fluidized bed combustor and they compared the results with the results of their previous study. Xie, et al. [15] investigated SO2, NO, and N2O emissions during the combustion of biomass and coal in a 30 kW bench scale circulating fluidized bed combustor. Varol and Atimtay [16] investigated combustion performances and emission characteristics of olive cake and coal in a bubbling fluidized bed. Flue gas concentrations of O2, CO, SO2, NOx, and total hydrocarbons (CmHn) were measured during combustion experiments. Liu, et al. [17] studied the co-firing of oil sludge and coal-water slurry in an internal circulating fluidized bed reactor that is 9 m in height with a rectangular cross section of 3.8 m *4.2 m. CO, SO 2 and NOx emissions were turned out to be very promising for the disposal of the sludge. Jiang ,et al. [18] investigated the combustion of oil-shale and co-firing of high sulfur coal and oil-shale. In this study, a laboratory scale circulating fluidized bed combustion system is designed and manufactured. The preliminary results of this system are presented here. It is the objective to improve the
combustion performance and to investigate the emission characteristics of combustion and co-combustion of lignite coal and biomass in the circulating fluidized bed combustor. After this study is completed, an industrial demo plant for producing energy using a fluidized bed boiler will be set up. The CFB technology to be applied in this demo plant will produce 750 kWth, using coal/biomass as fuels. The Circulating Fluidized Bed facility to be assembled will consist of a circulating fluidized bed combustor coupled with an external boiler which will produce saturated steam at 24 bars. The CFBC coupled to the boiler will have a capacity to burn about 300-350 kg/h feedstock. II.
MATERIALS AND METHODS
A laboratory scale CFBC system was designed and installed for small scale tests before going into pilot scale studies. The thermal capacity of the laboratory scale CFBC is 30 kW. The experimental setup consists of a circulating fluidized bed combustor column, a fuel feeding system, electrical heaters, and two cyclones. The combustor column has an inside diameter of 108 mm and a height of 6 m. It consists of 8 modules. Six electrical heaters are used along the combustor column in order to heat the system during the start-up period. Following the fuel feeding electrical heaters are shut down and the desired temperature is maintained by combustion. Additionally, two electrical heaters are installed on the downcomer to be utilized in case of necessity. The combustor is connected to the first cyclone with a pipe which has a nominal diameter (ND) of 32 mm. The solids separated from the flue gas on the cyclone are fed back to the combustor column through a return leg. The fuel feeding system consists of two hoppers, two screw feeders, two electrical motors and a compressed air connection. The fluidizing air is preheated by six primary air preheaters. There are six ports for the secondary air injection along the column. The temperatures along the column are measured with thermocouples located at specific heights. The temperature of the column is kept at 850 oC during the combustion experiments. The pressure drops along the combustor column; cyclone, return leg, and L valve are continuously measured. 18 electrical heaters are used in the system. 6 of them are 2.5 kW in power and used to heat the primary air to the ignition temperature of the fuels. 8 of them are 4.5 kW in power and located on the combustor column to keep the combustor temperature at about 850 oC which is the operating temperature of fluidized bed combustor. Two of them are 4 kW in power and used on the downcomer in order to heat the particles returning to the main column. Last two heaters are 2.5 kW in power and are used to heat the secondary air given to the combustor column at specific locations. During the experiments, CO, SO2, NO, and O2 emissions were continuously measured with the ABB/AO2000 flue gas analyzer. Calibration of the flue gas analyzer was done with certified calibration gases. Figure 1 shows the experimental system, Figure 2 shows a photograph of the full system, and Figure 3 shows the temperature and pressure measurement points on the system.
Figure 1. Schematic view of laboratory scale experimental system
Figure 2. Photograph of the laboratory scale experimental system
Figure 3. Temperature and pressure measurement points
A. Fuels and Bed Materials Hot tests were carried out with Orhaneli lignite coal as the main fuel. Besides, hazelnut shell and wood chips are used as the auxiliary fuel. The fuel particles were between 1-2 mm. The proximate and ultimate analyses of fuels were carried out by the Analysis Laboratories of the Turkish Scientific and Technical Research Council (TUBITAK) and they are presented in Table 1.
Table 1. Characteristics of fuels
Proximate Analysis (dry basis)
Ultimate Analysis (dry basis)
III.
Ash, wt% Volatile Matter, wt% Fixed Carbon, wt% Lower Heating Value, kcal/kg Higher Heating Value, kcal/kg C, wt% H, wt% N, wt% O, wt% S (Total), %
Orhaneli Lignite 17.23 47.31 35.46 4992
Hazelnut shell 1.73 73.54 24.73 5215
Wood Chips 1.36 80.48 18.16 4369
5281
5518
4667
60.03 3.15 0.56 16.98 2.05
56.34 5.35 0.51 36.06 0.01
53.72 6.93 0.29 37.70 0
RESULTS AND DISCUSSION
A. Combustion Experiments 1) Temperature Distribution In Figure 4, temperature distribution along the combustor for seven different excess air ratios can be seen. Throughout the experiment emissions were taken for seven different excess air ratios and the time averaged combustor profile is displayed. The maximum temperature is observed at about 240 mm just above the distributor plate, whereas the minimum is observed at the exit of the combustor. The seven different excess air ratios are indicated below the graph.
Height above the distributor plate [mm]
Temperature Profile Along the Combustor for Orhaneli Lignite
Temperature [ oC] 1,66
1,70
1,65
1,61
1,34
1,39
1,18
Figure 4. Temperature profile along the combustor for Orhaneli Lignite combustion
Figures 5 and 6 shows the temperature profile for wood chips and hazelnut shell combustion. As can be seen in Figures 5 and 6, biomass combustion showed a little different temperature profile along the combustor.
On the contrary to coal combustion, the maximum temperature appears more above the dense phase than in the case for coal. The maximum temperature is observed at about 320 mm just above the distributor plate This could be attributed to the higher volatile matter and lower fixed carbon content of biomass as compared to coal. This fact has also been mentioned in the literature by several authors [19-21]. Besides this fact, the gap between the dense phase and freeboard temperatures is closed in biomass combustion since high volatile matter content of biomass is oxidized in higher parts of the combustor whereas; coal is oxidized just above the distributor plate. Out of these three fuels, hazelnut shell combustion provides more uniform temperature profile throughout the axial direction as can be seen from Figure 6. Temperature profile for the case of co-firing of biomass and coal lied in between the profiles that of coal and biomass combustion individually.
Height above the distributor plate [mm]
Temperature Profile Along the Combustor for Woodchips Combustion
Temperature [ oC] 1,50
1,40
1,21
1,30
1,61
1,71
1,98
2,18
Figure 5. Temperature profile along the combustor for wood chips combustion
2,23
Height above the distributor plate [mm]
Temperature Profile Along the Combustor for Hazelnut Shell Combustion
Temperature [ oC] 1,15
1,23
1,34
1,47
1,55
1,60
1,80
1,73
1,89
Figure 6. Temperature profile along the combustor for hazelnut shell combustion
2) Emissions All the emission data presented in this section were taken for the dense phase temperature of 85050 °C. Combustion tests Figure 7 demonstrates the emissions coming from coal combustion for various excess air ratios. As it is clear from the figure, SO2 emissions fluctuated a little according to the change in operating parameters, namely superficial velocity and excess air ratio. NO and CO emissions followed a more stable path. CO emissions are closely related with residence time, O2 percentage in combustion chamber, mixing of air and fuel particles as well as temperature of the combustion chamber [15,23]. Since the superficial velocity was kept constant throughout the emission measurements, the only operating parameter that could affect the NO x emissions is the excess air ratio. Excess air ratio was adjusted by changing the fuel feed rate. As a general tendency as the excess air ratio increased, NOx emissions are found to decrease if the other parameters are kept constant [23]. This can be attributed to the decrease in mean bed temperature since NO x emissions are favored at higher mean bed temperatures. However, as the excess air ratio is increased, NO showed a rising tendency as can be seen in Figure 7. Declining CO concentration could be the reason for this, since reduction of NOx depends on the presence of char particles and CO concentration. It is a known fact that as CO is present in the combustion chamber NOx is reduced to molecular nitrogen [15,23].
Temperature [oC]
Emissions (6% O2) [mg/Nm3]
Emissions vs Excess Air Ratio for Orhaneli Lignite Combustion
CO
Excess Air [-] NO SO2RatioTemperature
Figure 7. SO2, CO, NO emissions for different excess air ratios (100% by wt. Orhaneli Lignite) In Figure 7, CO emissions showed a decrease with increasing excess air ratio as expected. [23]. In terms of SO2 emissions, emissions increased as the dense phase temperature increased which is a well known fact that has been mentioned in previous studies [24,25]. In Figure 8, emissions from combustion of woodchips are presented. As the fuel did not contain elemental sulfur, SO2 emission is zero. On the other hand, high volatile matter and low fixed carbon content resulted in higher CO emissions. However, it could be justified from the figure that by optimizing the process conditions the CO emissions could be lowered to acceptable levels. Increasing excess air ratio from 1.3 to 1.7 significant reduction in CO emissions was detected. Since the dense phase temperature remained almost the same as well as the superficial velocity, the char combustion rate constant can be assumed to be the same. Thus, decline in CO emissions can be attributed to oxidation of volatile matter for higher O2 concentrations.
Temperature [ oC]
Emissions (6% O2) [mg/Nm3]
Emissions vs Excess Air Ratio for Woodchips Combustion
Excess Air Ratio [-] CO
NO
SO2
Temperature
Figure 8. SO2, CO, NO emissions for different excess air ratios (100% by wt. Woodchips)
Emissions vs Excess Air Ratio for Hazelnut Shell Combustion
Temperature [ oC]
Emissions (6% O2) [mg/Nm3]
In Figure 9, the effect of excess air ratio on flue gas emissions for the combustion of hazelnut shell is presented.
Excess Air Ratio [-] CO
NO
SO2
Temperature
Figure 9. SO2, CO, NO emissions for different excess air ratios (100% by wt. Hazelnut shell)
Emissions vs Excess Air Ratio for (90% by wt.)Orhaneli Lignite Woodchips Combustion (%10 by wt.)
Temperature [oC]
Emissions (6% O2) [mg/Nm3]
Co-combustion tests Figure 10 shows SO2, CO, NO emissions for different excess air ratios for the fuel mixture of 90% by wt. Orhaneli Lignite and 10% by wt. woodchips. From Figure 10, it is seen that biomass addition increased CO emissions from 2000 to about 3000 while making contribution to a decrease in SO2 emission as compared to coal combustion. Besides this, NO x emissions did not increase as the dense phase temperature increased. However, this is not contradicting with the above comments made for Figure 7 since NO x reduction mechanism is directly proportional with presence of char and CO concentration [23,25]. As the CO concentration is increased, char level is also increased and these two have most probably played great role in the reduction of NO x resulting in lower emissions. SO2 emissions are lower as compared to Orhaneli Lignite combustion for the same dense phase temperatures because of zero sulfur in the woodchips.
Excess Air Ratio [-] CO
NO
SO2
Temperature
Figure 10. SO2, CO, NO emissions for different excess air ratios (90% by wt. Orhaneli Lignite -10% by wt. Woodchips)
Temperature [ oC]
Emissions (6% O2) [mg/Nm3]
Emissions vs Excess Air Ratio for (70% by wt.) Orhaneli Lignite (30% by wt.) Woodchips Combustion
Excess Air Ratio [-] CO
NO
SO2
Temperature
Figure 11. SO2, CO, NO emissions for different excess air ratios (70% by wt. Orhaneli Lignite -30% by wt. Woodchips)
Figure 11 shows SO2, CO, NO emissions for different excess air ratios for the fuel mixture of 70% by wt. Orhaneli Lignite and 30% by wt. Woodchips. In Figure 11, SO2 emissions continued to decrease for the same dense phase temperatures as compared to Figures 7 and 10 where biomass addition was lower. For CO concentrations it is obvious that increasing excess air ratio had a vital role in decreasing the emissions. There are three exceptions where CO concentration has shown increase despite the increase in excess air ratio. Temperature of dense phase occurred to be greater for these three points in Figure 11. Although temperature rise is expected to increase combustion efficiency therefore resulting in lower CO emissions, this effect could not be seen for these three points. The reason for this contradiction with CO emission increase for higher dense phase temperatures might be, lack of turbulence in the bed so that fuel-air contact could not have been provided in the way required. In contrast to this, it is obvious that NOx emissions have undergone a decrease in spite of the increased temperature. As discussed earlier, higher CO and char level had role in NOx reduction. The change of SO2, CO, NO emissions for different excess air ratios for the fuel mixture of 70% by wt. Orhaneli Lignite and 30% by wt. Woodchips is given in Figure 12.
Temperature [ oC]
Emissions (6% O2) [mg/Nm3]
Emissions vs Excess Air Ratio for (50% by wt.) Orhaneli Lignite (50% by wt.) Woodchips Combustion
Excess Air Ratio [-] CO
NO
SO2
Temperature
Figure 12. SO2, CO, NO emissions for different excess air ratios (50% by wt. Orhaneli Lignite -50% by wt.Woodchips)
Emissions vs Excess Air Ratio for (90% by wt.) Orhaneli Lignite (10% by wt.) Hazelnut Shell Combustion
Temperature [oC]
Emissions (6% O2) [mg/Nm3]
The discussions for Figure 11 are almost valid for Figure 12. As the biomass share in the fuel mixture increased SO2 emissions decreased where CO emissions showed an increase. However, as a general trend introducing more air into the combustor helped to decrease CO emissions. Figures 13-15 represent the emission data for co-firing of coal and hazelnut shell.
Excess Air Ratio [-] CO
NO
SO2
Temperature
Figure 13. SO2, CO, NO emissions for different excess air ratios (90% by wt. Orhaneli Lignite -10% by wt. Hazelnut shell)
Emissions vs Excess Air Ratio for (70% by wt.) Orhaneli Lignite (30% by wt.) Hazelnut Shell Combustion
Temperature [ oC]
Emissions (6% O2) [mg/Nm3]
Lignite co-firing with hazelnut shell has decreased the SO2 emission as expected. Compared to Figure 11, the SO2 and CO concentration band has been broadened which might mean that more stable combustion process has been carried out. The situation becomes clear in case of examining Figures 14 and 15. Lignite-hazelnut shell co-firing has resulted in more uniform temperature and emission profiles. Especially for Figures 14 and 15, increasing the excess air ratio yielded lower CO concentrations in a more stabilized way compared to lignitewoodchips co-firing. The source of the increase in SO2 concentrations in Figures 14 and 15 is the increasing dense phase temperature. It turns out that, NOx reduction is most probably accomplished by increasing excess air ratio for hazelnut shell-lignite co-firing as expressed in the first part of the results section.
Excess Air Ratio [-] CO
NO
SO2
Temperature
Emissions vs Excess Air Ratio for (50% by wt.) Orhaneli Lignite (50% by wt.) Hazelnut Shell Combustion Temperature [ oC]
Emissions (6% O2) [mg/Nm3]
Figure 14. SO2, CO, NO emissions for different excess air ratios (70% by wt. Orhaneli Lignite -30% by wt. Hazelnut shell)
Excess Air Ratio [-] CO
NO
SO2
Temperature
Figure 15. SO2, CO, NO emissions for different excess air ratios (50% by wt. Orhaneli Lignite -50% by wt. Hazelnut shell)
Hazelnut shell combustion yielded lower CO emissions compared to woodchips combustion. This could be attributed to the lower fixed carbon and higher volatile matter of woodchips compared to hazelnut shell. The effect of biomass addition on CO, SO2, and NO emissions is presented in Figure 16 and Figure 17. Thermal power produced during the combustion was selected as the basis in order to make a meaningful comparison between emission performances of different fuel combinations. The system has been operated under two thermal powers namely, 13 and 22 kW. The superficial velocity was between 2 and 3.2 m/s as the excess air ratio was changed. The conditions at which mimimum emissions obtained were selected in order to compare the flue gas emissions at the same thermal power. In Figure 16 and Figure 17, the change in emissions is presented as the share of woodchips and hazelnut shell in fuel mixture increases. In both case, as the percentage of biomass in fuel mixture increases, CO concentrations decrease. While CO emissions are about 1200 mg/Nm3 for the combustion of Orhaneli lignite, they gradually increase by the addition of biomass. CO emissions reach up to 5200 mg/Nm3 for the case of woodchips and 3000 mg/Nm3 for the case of hazelnut shell. High CO emissions measured for the combustion of woodchips and hazelnut shell may be explained by the higher volatile matter (VM) content of biomasses (75.2% for woodchips and 65% for hazelnut shell) than that of lignite (33%). Since VM of woodchips is higher than VM of hazelnut shell, CO emissions obtained from woodchips combustion is higher than that obtained from the combustion of hazelnut shell. Biomass combustion mainly takes place at the freeboard region leaving the system in a short period of time. CO formed in the freeboard may not find so much chance to meet with oxygen and finally may leave the combustor not totally oxidized. Therefore, it seems necessary to introduce secondary air to the freeboard region in order to completely oxidize the incomplete combustion products.
Emissions of NO [6% O2
Emissions of SO2, CO [6% O2, mg/Nm3]
Effect of Woodchips Feeding Rate on Emissions with Orhaneli Lignite Co-firing
Share of Woodchips in Fuel Mixture (% by wt) SO2
CO
NO
Figure 16. Effect of the share of Wood Chips on flue gas emissions
Emissions of NO [6% O2 mg/Nm3]
Emissions of SO2, CO [6% O2, mg/Nm3]
Effect of Hazelnut Shell Feeding Rate on Emissions with Orhaneli Lignite Co-firing
Share of Hazelnut Shell (% by wt) SO2 CO NO Figure 17. Effect of the share of Hazelnut shell on flue gas emissions For the combustion of Orhaneli lignite, SO2 emissions were measured about 4000 mg/Nm3. Since both biomasses have almost zero sulphur content, SO2 emissions decrease as the share of biomass in the fuel mixture increases. No SO2 emission is measured for combustion of woodchips and hazelnut shell. The emission limit for SO2 in Turkish regulation is 2000 mg/Nm3 [26]. It is necessary to add limestone to the combustor in order to capture SO2 in the form of CaSO4 to bring the emissions to the level of the limits. In case of coal combustion, NO emissions are about 200 mg/Nm3. They are about 190 and 150 mg/Nm3 for the combustion of woodchips and hazelnut shell, respectively. NO emissions tend to decrease by burning biomass with coal. Minimum emissions were obtained for the co-combustion tests containing 50% by wt. biomass and 50% by wt. coal. When biomass resources are burned alone, NO emissions are observed to be higher than the NO emissions obtained for co-combustion tests. Giving a portion of combustion air to the freeboard region by creating a reducing atmosphere in the dense phase may help to decrease NO emissions. It is intended to operate the system at this condition for further works. IV.
CONCLUSION
The results of this study have shown that the co-combustion of lignite coal with woodchips and hazelnut shells in a circulating fluidized bed combustor can be performed with minimum emissions. The SO2 emissions are above the standards since no sorbent is used to absorb SO2 emissions. With the introduction of secondary air, especially the CO emissions can be decreased considerably. Biomass addition (woodchips and hazelnut shells) decreased the SO2 emissions as the percentage of biomass increased in the fuel mixture and CO emissions showed the opposite trend as it is expected. There is no doubt that in terms of emissions the system requires an optimization in operating conditions. However, for the first results of combustion tests it can be concluded that results are promising and several combinations can be applied to obtain the optimum emission performance for future work. V. [1] [2] [3] [4] [5]
REFERENCES
“CO2 emissionf from coal combustion”, IEA Highlight 2009, available on http://www.iea.org/co2highlights/co2highlights.pdf Lior N., “The ECOS 2009 World Energy Panel: An introduction to the Panel and to the present (2009) situation in sustainable energy development”,Energy, 2010, in press, corrected proof (doi:10.1016/j.energy.2010.05.036) World Coal Institute, available on http://www.worldcoal.org/resources/coal-statistics/ Hupa M., “Ash behavior in fluidized bed combustion-recent research highlights”, Paper presented at the 9th International Conference on Circulating Fluidized Beds, 2008, Hamburg, Germany. Hupa M., “Interaction of fuels in co-firing in FBC”, Fuel, 2005, 84, 1312-1319.
[6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]
Youssef M.A., et.al., “Experimental Study on Egyptian Biomass Combustion in Circulating Fluidized Bed”, Applied Energy, 2009, 86, 2644-2650. Desroches-Ducarne E., Eric M., Gerard M., Lucien D., “Co-combustion of coal and municipal solid waste in a circulating fluidized bed”, Fuel, 1998, 77, 1311-1315. Cliffe K.R., Patumsawad S., “Co-combustion of waste from olive oil production with coal in a fluidised bed”, Waste Management, 2001, 21, 49-53. Armesto L., et. Al., “Co-combustion of coal and olive oil industry residues in fluidized bed”, Fuel, 2003, 82, 993-1000. Topal H., Atimtay T.A., Durmaz A., “Olive cake combustion in a circulating fluidized bed”, Fuel, 2003, 82, 1049-1056. Permchart W., Kouprianov V.I., ”Emission performance and combustion efficiency of a conical fluidized bed combustor firing various biomass fuels”, Bioresource Technology, 2004, 92, 83-91. Gayan P., et. al., “Circulating fluidised bed co-combustion of coal and biomass”, Fuel, 2004, 83, 277-286. Atimtay T. A., Topal H., “Co-combustion of olive cake with lignite coal in a circulating fluidized bed”, Fuel, 2004, 83, 859-867. Shimizu T., Toyono M., “Emissions of NOx and N2O during co-combustion of dried sewage sludge with coal in a circulating fluidized bed combustor”, Fuel, 2007, 86, 2308-2315. Xie J, Yang X, Zhang L, Ding T., Song W.,Lin W., “Emissions of SO2, NO and N2O in a circulating fluidized bed combustor during co-firing coal and biomass”, Journal of Environmental Sciences, 2007, 19, 109-116. Varol M., Atimtay A. “Combustion of olive cake and coal in a bubbling fluidized bed with secondary air injection”, Fuel, 2007, 86, 1430-1438. Liu J., et. al., “Co-firing of oil sludge with coal–water slurry in an industrial internal circulating fluidized bed boiler”, Journal of Hazardous Materials, 2009, 167, 817-823. Jiang X.M., et.,al.,”Progress and recent utilziaiton trends in combustion of Chinese oil shale”, Progress in Energy and Combustion Science, 2007,33 552-579. Madhiyanan et.al., “Co-combustion of rice husk with coal in a cyclonic fluidized bed combustor”, Fuel, 2009, 88, 132-138. Suksankraisorn et. al., “Co-combustion of municipal solid wastes and Thai Lignite in a fluidized bed”, Energy Conversion and Management, 2004, 45, 947-962. Sathitruangsak et.al., “Rice husk co-firing with coal in a short combustion chamber fluidized bed combustor”, Fuel, 2009, 88, 13941402. Tsai et.al., “Co-firing of paper mill sludge and coal in an industrial circulating fluidized bed boiler”, Waste Management, 2002, 22, 439-442. Gungor A., “Simulation of NOx emission in circulating fluidized bed burning low grade fuels”, Energy&Fuels, 2009, 23, 2475-2481. Adanez J. Et. al., “Modeling of sulfur retention in circulating fluidized bed combustors”, Fuel, 1996, 75, (3), 262-270. Zhao J. Et. al., “Gas concentration profiles and NOx formation in circulating fluidized bed combustion”, Fuel, 1997, 76 (9), 853-860. RCAPIS, 2009. Turkish Regulation for Control of Air Pollution from Industrial Sources, Republic of Turkey Ministry of Environment and Forestry.
CO-COMBUSTION PERFORMANCE OF OIL SHALE AND BIOMASS FUELS Emre Özgür, Research Assistant, Middle East Technical University-Petroleum and Natural Gas Engineering Department Middle East Technical University, Petroleum and Natural Gas Engineering Department, Ankara-TURKEY 06531 Sharon Falcone Miller, Research Associate, EMS Energy Institute, Penn State University Bruce G. Miller, Associate Director, EMS Energy Institute, Penn State University EMS Energy Institute, Penn State University, C211 Coal Utilization Laboratory, University Park, PA-USA 16802 Mustafa Verşan Kök, Professor, Middle East Technical University-Petroleum and Natural Gas Engineering Department Middle East Technical University, Petroleum and Natural Gas Engineering Department, Ankara-TURKEY 06531
Abstract In this study, the thermal analysis of co-combustion of biomass fuels and oil shale were investigated. The objective was to assess the effect of the biomass on combustion performance when blended with oil shale. Thermogravimetric analysis and differential scanning calorimetry were used to analyze the samples. The biomass samples studied were hazelnut shell, wheat bran, poplar, and miscanthus. Co-combustion of blends were performed at different biomass proportions (10, 20, 50 % by wt.). 1.0 Introduction Worldwide, there are concerns about limited fossil energy sources in some countries and volatile energy prices. In addition there are increasing concerns about climate change, which is resulting in carbon management legislation being considered throughout the world. The use of agricultural wastes and non-conventional low-quality fossil fuels can assist in addressing these concerns, especially for those countries having such resources. In different regions of the world there are many agricultural wastes and, in some areas, there are significant oil shale reserves. The combustion of oil shale is problematic due to its higher ignition temperature, high ash content, and high greenhouse gases emission rates. On the other hand, biomass has very low ash
content, is carbon neutral, usually contains low sulfur content and has high reactivity. Blending biomass with oil shale for co-firing may be a good method to utilize oil shale taking into account combustion and environmental concerns. Based on the results of a literature survey, there are many published articles on the thermal analysis of fossil fuels (coal, lignite, oil shale), biomass fuels, and blends of coal/lignite and biomass fuels. However, no published work on the co-combustion of biomass and oil shale was found. Therefore, this study was performed to generate oil shale/biomass combustion data. In this research, the combustion behavior of several biomass feedstocks and oil shale were investigated as sole fuels. Then, the blends of biomass and oil shale were investigated to see the effect of blending on co-combustion performance. The biomass samples used in the experiments are hazelnut shell, wheat bran, poplar and miscanthus. Hazelnut shell and wheat bran are important agricultural residues for Turkey; poplar and miscanthus are potentially energy crops to be used in Turkey because of its suitable climate conditions. 1.1 Oil Shale Combustion The world fossil energy sources exhibit volatility in both availability and price; hence, nonconventional resources are becoming more popular. Oil shale is one of the non-conventional resources that could be used as an energy source in two different ways. One method to utilize the oil shale is through retorting. In this method, the process converts kerogen in the oil shale into synthetic crude oil by pyrolysis, or destructive distillation. The resulting oil can be used as a fuel after some processing. Another alternative is the direct combustion of oil shale (Lee, 1996).
The direct combustion of oil shale to obtain energy is gaining popularity. At present, about 69 % (or 11 millions of tonnes) of world oil shale production is used for the generation of electricity and heat. The utilization of oil shale for retorting, cement production and other uses form the alternative applications. As of 2008, the countries producing electricity based on direct combustion of oil shale are, with their generating capacities, Estonia (2,967 MW), Israel (12.5 MW), China (12 MW), and Germany (9.9 MW) (Brendow, 2003). In Estonia, the utilization of oil shale in pulverized combustion systems is common. Fluidized-bed combustion systems are used in China, Germany and Israel.
1.2 Biomass Combustion There are more than 3,000 biomass power plants worldwide with a capacity of more than 40,000 MW (WEC, 2007). Their generating capacities are between 1-80 MW due to the limited availability of biomass fuels within an economical milling and shipping distance (Bauen et al, 2004). The utilized biomass fuels are agricultural residues and energy crops. 1.3 Cofiring Cofiring is the simultaneous combustion of two different fuel types in the same combustion system. One of the advantages of cofiring is that an existing plant can be used to burn a second fuel, which, depending on the fuel type and combustion system, can result in little or no additional cost for the new fuel. Cofiring with biomass can be more environmentally friendly due to the reduction of greenhouse gases emitted from the carbon neutral fuel in the blend. Cofiring can also be used to improve the combustion of fuels with low energy content by blending with fuels having higher energy content (Loo and Koppejan, 2008).
In the last decade, the utilization of biomass in co-firing systems has shown rapid development. Co-firing technology is presently performed at commercial scales in the USA, Australia, Japan, many European Countries, and a number of other countries. There are more than 150 coal-fired power plants in which biomass is cofired (Loo and Koppejan, 2008). The generating capacities of these power plants are between 50-700 MW and the majority are pulverized coal systems. Cofiring in existing pulverized combustion systems will generally require minor modifications or additions to fuel handling, storage, and feed systems (Masuda et al, 2006). In our work, the cofiring of biomass and oil shale was investigated. Combustion of oil shale alone is a problematic process due to its high ash content, high emissions of unwanted gaseous such as CO2, SO2 during combustion and high ignition temperature. On the other hand, biomass has very low ash content, emits a low amount of unwanted gases and has high reactivity. Therefore blending biomass with oil shale could enhance oil shale utilization by addressing the combustion and environmental concerns.
2.0 Experimental Methodology 2.1 Analysis of Feedstocks The proximate analysis of the oil shale and biomass samples were determined by using a Thermogravimetric Analyzer (TGA). The samples were heated to 110 oC at a 50 oC/min heating rate and the samples were held at this temperature for 120 minutes to determine the moisture content. Then, the temperature was ramped to 950 oC at a heating rate of 50 oC/min. At 950 oC, the gas environment was switched to air from nitrogen. At this temperature, the samples were held 120 minutes isothermally to perform full combustion. Proximate analysis results of each fuels is given in Table 1.
Ultimate analysis was also performed to obtain more information about each fuel in LECO TruSpec CHN analyzer. Because of a small amount of sample to work with, sulfur (and hence oxygen by differences) was not determined. Sulfur contents of these feedstocks are typically low (Loo and Koppejan, 2008). The results are given in Table 2.
Table 1. Proximate analysis of each fuels (wt. %) Fuel Type Oil Shale Hazelnut Shell Miscanthus Poplar Wheat Bran
Moisture 7.0 1.5 2.0 1.0 8.0
Volatile Matter 8.4 69.5 80.5 74.0 65.0
Fixed Carbon 0.1 28.9 16.0 24.0 24.1
Ash 84.5 0.1 1.5 1.0 2.9
Table 2. Ultimate analyis of each fuels Fuel Type Oil Shale Hazelnut Shell Miscanthus Poplar Wheat Bran
Carbon Cont., wt.% 3.75 50.5 45.2 46.6 43.2
Hydrogen Cont., wt. % 1.05 5.63 5.28 5.71 6.19
Nitrogen Cont., wt. % 0.22 0.18 0.01 0.45 2.58
2.2 Combustion Tests Combustion tests were performed in the TGA and differential scanning calorimeter (DSC). For the combustion tests in the TGA, a similar procedure was performed. The atmosphere was air. The samples were heated to 110 oC at a 50 oC/min heating rate and held at 110 oC for 10 minutes.
Then, the temperature was ramped to 950 oC at a heating rate of 50
o
C/min. At 950 oC, the
samples were held 5 minutes isothermally to complete the combustion. It was observed that the devolatilized organic matter quantity before combustion decreases with increasing heating rate based on TGA and DSC output data, so the heating rate of 50 oC/min was chosen as a reasonable heating rate to prevent the loss of organic matter prior to combustion.
For the combustion experiments in the Differential Scanning Calorimeter, the procedure was that the samples were heated to 110 oC from ambient temperature at 50 oC/min heating rate and were held 10 minutes isothermally at 110 oC. Then the samples were heated at a rate of 50 oC/min to 600 oC. At this temperature, the system was held isothermal for 5 mins to ensure complete combustion. The atmosphere used was air.
In all the runs in the TGA and DSC, the weights of the samples were between 2 mg and 4 mg. Small samples and open pans were preferred in order to promote diffusion of air during the test. The particle sizes of the samples used in the experiments are in the range of 0.2 mm to 0.5 mm. TGA and DSC were used for the characterization of fuels and blends. TGA and DSC are useful thermoanalytical equipment to provide the burning profile for determining the relative combustion characteristics of a fuel. They provide information on oxidation rates from ignition to completion of burning. They are especially useful for the designers which can select the type and size of furnace, the type and arrangement of burners, and type of surface adjacent to the burners. They also help the designer determine the amount of ignition stabilizing fuel required for pulverized firing because of parallels between the size of samples used in equipment and required fuel size for pulverized firing (Wagoner and Winegart, 1973). Thermogravimetric analyzer monitors mass flow into a sample or out of a sample with increasing temperature. It could give information about the moisture, volatile, fixed carbon and ash content of the sample, providing proximate analysis. From TGA curves, activation energies, reaction mechanisms, kinetics and thermodynamics of the chemical reactions can be obtained. The model of TGA was Perkin Elmer TGA 7 in the study. Differential scanning calorimeter measures the difference between the heat flows from the sample and reference as a function of temperature or time by the help of the sensors. DSC is
capable of enabling qualitative and quantitative analysis. Onset ignition temperature, burn-out temperature, reaction intervals, exothermic and endothermic regions could be identified. The model of DSC was Mettler Toledo DSC1 heat-flux type in the study. 3.0 Results & Discussion 3.1 Onset Ignition Temperature Onset ignition temperature is defined as the temperature where exothermic reactions begin. Because the ignition temperature of solid fuels cannot be determined exactly since they depend upon operating conditions, the onset ignition temperatures are reported as the first point of the exothermic region in DSC curves. The onset ignition temperatures of each fuel type are given in Table-3. Addition of biomass noticeably improves the combustion of oil shale by lowering the ignition temperature.
3.2 Burnout Temperature Burnout temperature is defined as the temperature where combustion or exothermic reaction ends. It can affect the length of a furnace required to burn the fuel completely and the location of heat recovery equipment. Any changes in this temperature may adversely affect the combustion efficiency. The burnout temperatures of each fuel type are given in Table 3. The addition of biomass increases the burnout temperature of the blend, however after 50% biomass addition there is no affect or very low.
The similar tabulation of ignition and burnout temperatures of blended fuels were made by Muthuraman et al to see the change of temperatures with blend ratios (Muthuraman et al, 2010).
3.3 Activation Energy Activation energies of each fuels in the combustion was calculated to see the change based on Coats & Redfern kinetic method. When the sample size is small, and with an excess air supply, the process of the combustion reaction is independent of the concentration of oxygen and it is therefore reasonable to assume that oxidation can be described by first order kinetics method so first order kinetic was assumed. The data is also presented in Figure 1 for observation of trend.
Table 3. Onset ignition temperatures and burn-out temperatures of each fuel type Ignition Temperature, oC 271 288 303 305 300 306 311 325 295 296 321 325 295 296 308 310 325
Fuel type Hazelnut shell 50% hazelnut shell 20% hazelnut shell 10% hazelnut shell Miscanthus 50% miscanthus 20% miscanthus 10% miscanthus Poplar 50% poplar 20% poplar 10% poplar Wheat bran 50% wheat bran 20% wheat bran 10% wheat bran Oil shale
Burnout Temperature, oC 539 537 528 519 567 548 537 492 530 530 500 450 600 600 544 537 425
140
120
E a, kj/m ol
100
80
60
40
20
he ts
%
10
Ha
ze
ln u
ln u 50
%
ha
ze
ze
l ln ut l wh s he 20 e a ll % t wh br a 50 ea n % t wh br a ea n tb W h e ran at 10 b ra n % p 20 opla % r p 50 opla % r po pl ar 10 % Po m pl a 2 0 is c an r % th m u s 5 0 is c an % th m u is c s an th M i s c us an th us
ll
ll
he ts
he ts
ln u %
20
10
%
ha
ha
ze
O
il S
ha
le
0
Figure 1. Activation energies of each fuel type
Ea, kj/mol 31.20 35.8 57.7 74.8 57.6 66.1 86.9 112.7 42.1 61.1 71.9 76.6 45.3 49.9 52.6 64.3 125
Thermogravimetric Analyzer Curves & Differential Scanning Calorimeter Curves In Figure 2, the thermogravimetric curves of oil shale, biomass fuels and their blends are given. The differential scanning calorimeter curves are given in Figure 3. The additive behavior of biomass fuels on oil shales can be seen clearly in figures. It is clear that the combustion of biomasses are highly exothermic compared to oil shales based on DSC curves.
Figure 2a
Figure 2b
Figure 2c
Figure 2d Figure 2. TGA curves of parent fuels and blends. (a) hazelnut shell. (b) miscanthus. (c) poplar. (d) wheat bran
Figure 3a
Figure 3b
Figure 3c
Figure 3d Figure – 3. DSC curves of parent fuels and blends. (a) hazelnut shell. (b) miscanthus. (c) poplar. (d) wheat bran Biomass Model Compounds Biomass fuels are chemically complex polymeric lignocellulosic materials. They are mainly composed of cellulose, hemicelluloses, and lignin which cellulose and hemicellulose are tightly bound to the lignin. These model compounds form the physical and chemical properties of the biomass fuels based on their arrangements in the fuel. To find out the possible relations between model compounds and combustion behavior of biomass fuels, model compounds were analyzed. The onset ignition temperatures and burnout temperatures of biomass model compounds are given in Table 4, which were obtained from DSC with TGA assistance. The contents of biomass model compounds in each biomass fuel are given in Table 5, which was obtained from Phyllis biomass database (Phyllis, 2009). There are numerous types of model compounds, in our study the tested model compounds are CF11 fibrous cellulose powder, xylan (hemicellulose) from birchwood, and loblolly pine lignin.
Table 4. Onset ignition temperatures and burn-out temperatures of biomass model compounds Model Compound Cellulose Hemicellulose Lignin
Ignition Temperature, oC 390 257 285
Burnout Temperature, oC 590 516 612
Table 5. Lignocellulosic contents of biomass fuels Biomass Fuels Hazelnut Shell Miscanthus Poplar Wheat Bran
Cellulose Cont., wt.% 25.9 44.7 40 12
Hemicellulose Cont., wt.% 29.9 29.6 23 36
Lignin Cont., wt.% 42.5 21 20 3
Statistical Analysis To explain possible relations, statistical software Minitab was used. The relation of ignition temperature and burnout temperature with volatile matter contents and CHN contents were investigated for five parent fuels and each blend together. Later, the effect of model compound contents with volatile matter contents and CHN contents on only biomass fuels and each blend was investigated because oil shale doesn’t include model compounds. The results are given in Table 6. First two rows are for all fuels; last two rows are for only biomass fuels in the table. The results are tabulated in p-values and R2. The lower p-values indicate statistical significance for independent parameters. All p-values are lower than the alpha level (acceptable level for analysis) of 0.05 which shows a good statistical significance for independent parameters. R-squares (R2) values, coefficient of determination, show the relationship between dependent and independent parameters. In general, the higher the R2, the better the model fits the data. It is obtained that the ignition temperature of fuels is dependent on carbon content, hydrogen content and nitrogen content. As the increase of carbon content and nitrogen content in a fuel, the ignition temperature decreases. Hydrogen has reducing effect on ignition temperature. On the other hand, no strong correlation was observed for burnout temperature. Volatile matter content and nitrogen content has increasing effect on burnout temperature but the degree of effect is not definite.
For biomass fuels, it is observed that ignition temperature of biomass fuels is controlled by volatile matter content, cellulose content and lignin content. The increase in the content of cellulose leads to higher ignition temperature, however the increase in the volatile matter and lignin content lead to lower ignition temperature. For burnout temperature, no strong correlation was observed again. However, it is observed that carbon content and hemicellulose content are the most deterministic parameters for burnout temperature.
Table 6. Statistical relations between physical properties of fuels Dependent Parameter Ignition Temperature
Independent Parameter Carbon Content Hydrogen Content Nitrogen Content
P-values 0.006 0.011 0.014
R2 78.1 All fuels n=17
Burnout Temperature
Volatile Matter Content Nitrogen Content
0.014 0.030
56.0
Ignition Temperature
Volatile Matter Content Cellulose Content Lignin Content
0.000 0.001 0.001
86.6
Carbon Content Hemicellulose Content
0.018 0.002
64.9
Burnout Temperature
Biomass + blends n=16
Conclusions Direct combustion of solid biomass and oil shale is one of the practical ways of utilization of these solid fuels. Co-combustion of oil shale and biomass fuel types were investigated by using TGA and DSC methods at different biomass proportions. TGA and DSC can simulate the combustion conditions in industrial pulverized combustion of oil shales for this size of particles. Solid fuel combustion is mostly governed by its ignition temperature and burnout temperature and therefore the entire combustion process is based on these parameters, moreover these parameters are crucial in the design of burning equipment. So these parameters will give a representative idea about the co-combustion process. The blending of oil shale with biomass improves the ignition temperature. Because biomass contains high oxygen volatile content, it is easier to start the combustion.
The biomass and oil shale can be cofired in existing pulverized combustion coal systems. Since the combustion of biomasses is highly exothermic, biomass fuels could serve as a fine heating fuel. The selection of a biomass fuel for cofiring systems can be done based on the fuel’s model compound content since combustion behavior of biomass depends on the proportion of three model compounds; cellulose, hemicellulose, and lignin. Biomass fuels having high volatile matter content and low cellulose content are good options for cofiring with oil shale. An operation range between 10% and 20% of biomass proportion in weight percentage is a good option for co-combustion due to enough volatile content to maintain stability in combustion and low amount of biomass requirement.
Abbreviations: CHN Carbon, Hydrogen, Nitrogen DSC
Differential Scanning Calorimeter
MW
Megawatt
mm
millimeter
mg
milligram
min
minute
R2
coefficient of determination
TGA Thermogravimetric Analyzer
References: Bauen, A., Woods, J., Hailes, R., A Biomass Blueprint to Meet 15% of OECD Electricity Demand by 2020, Report by ICEPT and E4TECT, 2004
Brendow, K., `Global Oil Shale Issues and Perspectives- Synthesis of the Symposium on Oil Shale Held in Tallinn (Estonia) on 18 and 19 November 2002’, Oil Shale, 2003, Vol. 20, No. 1 pp.81-92 Lee, S., Alternative Fuels, Taylor & Francis, 1996 Loo, S.V. and Koppejan, J., The Handbook of Biomass Combustion of Co-firing, Earthscan, 2008 Masuda, H., Higashitani, K., Yoshida, H., Powder Technology Handbook, Taylor & Francis, 2006
Muthuraman, M., Namioka, T., Yoshikawa, K., Characteristics of co-combustion and kinetic study on hydrothermally treated municipal solid waste with different rank coals: A thermogravimetric analysis, Applied Energy, Vol. 87, Iss. 1, Pg. 141-148, 2010 Phyllis biomass database, retrieved from http://www.ecn.nl/phyllis on 11.15.2009 Wagoner, C.L. and Winegart, E.C., Further Development of the Burning Profile, Journal of Engineering for Power, Vol: 95 Iss: 2 Pg: 119-123, 1973 World Energy Council, Survey of Energy Resources – Biomass for Electricity Generation, 2007
INFLUENCE OF STEAM ON THE RELEASE OF ALKALI METAL, CHLORINE, AND SULPHUR SPECIES DURING HIGH TEMPERATURE GASIFICATION OF LIGNITE Marc Bläsing, Michael Müller
Institute of Energy Research (IEF-2) Leo-Brandt-Str. 1, 52425 Jülich, Germany Phone: +49-2461-61-1574, Fax: +49-2461-61-3699, E-mail: [email protected], E-mail: [email protected] Keywords: Lignite, Gasification, Molecular beam mass spectrometry, Release, Alkali metal, Sulphur, Chlorine species
1. INTRODUCTION Lignite is a major energy source for the production of electricity in North Rhine-Westphalia and Germany (26%) [1]. Instead of using lignite in conventional low-efficient steam cycles, the utilisation of lignite in high-efficient combined gas and steam turbine cycles can help to increase the efficiency and lead to a lower demand of fuel and therefore lower emission of CO2. One high efficient technology is the combined gas and steam cycle with integrated high temperature coal gasification (IGCC) [2]. However, combined cycle power systems require a reliable hot gas cleaning to cut the efficiency loss caused by state of the art gas cleaning systems which prevent downstream parts of the plant from hot corrosion caused by alkali metal, sulphur, and chlorine species [3]. Basic investigations on the release of these species form a crucial element in designing control measures and to develop gas cleaning strategies for coal gasification systems. Together with other experimental approaches a large body of useful data has already been delivered in this area, but the underlying reaction mechanisms are not yet sufficiently understood. Because alkali metal, sulphur, and chlorine species are in the focus of fouling, slagging and corrosion of special interest, a detailed understanding of the release mechanisms is needed [4]. Therefore, the release of inorganic species of four different
Rhenish lignite was investigated during lab-scale gasification experiments. The MolecularBeam-Mass-Spectrometry-technique used for hot gas analysis is well established and is able to detect and differentiate key chemical species released during gasification of coal [5-8].
2. EXPERIMENTAL 2.1 Fuel Preparation Samples from four Rhenish lignite were collected and prepared for analysis. The lignite were dried and ground in a mill to a particle size < 100 µm. After this treatment the lignite were stored under dry conditions at room temperature. The chemical analysis is given in Table 1. The lignite were chosen for investigation of a broad range of different coals. Table 1. Chemical composition of the lignite samples (as received; mass.-%)
C H O N S Cl Al Fe Ca Mg K Na Si
HKN-S65.8 4.81 28.1 0.78 0.205 0.01 0.034 0.25 1.0 0.37 0.02 0.22 0.01
HKN-S+ 65.8 4.98 26.6 0.84 0.508 0.025 0.034 0.25 1.2 0.44 0.023 0.22 0.023
HKS 62.0 4.89 28.0 0.69 0.365 0.023 0.12 0.48 1.4 0.48 0.024 0.22 0.72
HKT 57.3 4.18 27.5 0.75 0.478 0.011 1.5 0.28 1.3 0.47 0.085 0.23 3.6
2.2 Experimental Setup The release of alkali metal, sulphur, and chlorine vapour species during gasification experiments was monitored on-line and in-situ by molecular beam mass spectrometry (MBMS). For the determination of the gas composition of the product gas an electrical heated flow channel reactor was coupled to the MBMS. Further information of the apparatus and its application in research on alkali metal, sulphur, and chlorine species release during coal gasification and combustion are described elsewhere [5-8]. A schematic of the experimental setup is given in Figure 1.
Figure 1. Experimental setup A typical experimental run consisted of the following steps. At the start of the experiment a platinum sample boat loaded with 50 mg of lignite was inserted into the air cooled end of the heated flow channel. A gas flow of 3.7 l/min He and 0.3 l/min O2 corresponding to 92.5% He and 7.5% O2 was fed into the reactor to simulate a gasification like environment. A nebuliser was used to provide the gas flow with a steady moisture content of 2.5v%. The sample boat was placed in the 1400°C hot reaction zone by a horizontally displaceable alumina rod. All parts downstream the reaction zone were kept over the condensation point of the alkali metal, sulphur, and chlorine species of interest. The hot reaction products flowed to the end of the reactor where they entered the MBMS through a nozzle with 0.3mm in diameter. The MBMS consists of three differentially pumped chambers. Due to immediately supersonic free jet expansion of the hot gas into the first high vacuum chamber (10-2 mbar), the species are cooled down below room temperature in microseconds, attain free molecular flow and therefore, form a molecular beam [8]. The core of the free jet expansion is extracted by a conical skimmer of 1 mm diameter and directed into the third chamber. There, a hot filament emits electrons with an electron energy of 50 eV and an emission current of 1 mA. Every 10-410-3 molecule is ionised by electron impact. After passing the deflector, the ions are filtered in a quadrupole mass analyser and detected by an off axis electron multiplier. The amplified signal is recorded by a computer and software package as a function of time and mass-tocharge ratio. In order to be able to monitor the gasification process with sufficient temporal resolution, 10 scans per second were acquired. To ensure reproducibility, six samples were measured. In preliminary measurements mass spectra from 10 to 150 amu were scanned. Due to the high temperature of 1400°C most of the volatile organic matter was effectively cracked even despite the short residence time of 0.1-0.2 s [5-7]. NaOH is without a doubt a key specie,
however 40NaOH+ is not easily detectable by MBMS, because it has the same mass to charge ratio as 40Ar+. This results in an overlapping of the two peaks. Argon is always present in the gas phase to a small extent because the oxygen which has been used for the experiments is slightly contaminated with Argon due to its production process. Therefore, the monitored species were 23Na+, 34H2S+, 36HCl+, 39K+/39NaO+, 56KOH+, 58NaCl+, 64SO2+, and 74KCl+.
3. RESULTS AND DISCUSSION The gasification of a coal particle can be divided in three distinct phases: the pyrolysis phase, the char reactions phase, and the ash reactions phase. The pyrolysis and the char reactions phase are of great interest for the investigation of the release behaviour of sodium, potassium, sulphur, and chlorine species as shown in Figure 2. The main amount of the release occurred during the pyrolysis phase with high intensity. The quantification of the data was done by normalisation of the peak area during pyrolysis phase to the 34O2+-signal of the first 20s of the experimental run, when the steady oxygen concentration leaded to a steady signal. The averaged normalised peak areas of the lignite under investigation are depicted in Figure 4. Key alkali metal species detected by MBMS were 23Na+, 39K+/39NaO+, 56KOH+, 58NaCl+, and 74
KCl+. Tsuyoshi and Ichiro reported the release of metallic sodium during combustion of coal
[9]. However,
23
Na+ detected by MBMS is most likely a product of fragmentation of sodium
species, e.g. NaCl and NaOH. The mass to charge ratio 39 is occupied by 39K+, most likely a fragment of KCl and other Kspecies, and 39
39
NaO+, most likely a fragment of NaOH [7]. The normalised peak areas of
K+/39NaO+ were strongly increasing for all lignite under investigation under the influence of
steam. HKN-S- and HKN-S+ showed an increase of factor 5.2 – 36.9. HKS and HKT showed an increase of factor 46.7 – 68.8. However, both the increase of the released amount and the variance of the results are high and must be carefully discussed. During the experiments without steam the formation of sodium chlorine and therefore the amount of chlorine is highly important for the release of sodium. Further, the amount of chlorine is an order of magnitude smaller as the amount of sodium regarding to the chemical analysis of the lignite in Table 1. Therefore, the release of NaCl is limited by the amount of chlorine. During the experiments with steam highly volatile NaOH is formed which resulted in an enhanced volatilisation of sodium. Complementary explanations of the increase of
39
K+/39NaO+ are given in the
following. First of all, the fragmentation of KOH can lead to an increase of the m/z = 39 signal due to fragmentation of KOH and detection of
39
K+. However, the increase of
56
KOH+ is
about an order of magnitude lower than the increase of the signal of m/z = 39. Secondly, Wei et al. proposed a reaction mechanism for the transformation of Na bound to the organic coal matter into sodium silicate [10]. In the proposed mechanism the formation of intermediate NaOH under the influence of steam and the formation of Na2CO3 plays an important role. In analogy to our experiments intermediate NaOH could have led to an increased signal of m/z = 39 (39NaO+). Additionally, Kosminski et al. reported the formation of intermediate NaOH and Na2CO3 and the enhanced formation of liquid sodium carbonate and disilicate under the influence of steam [11,12]. In result the enhanced release of 39K+/39NaO+ under the influence of steam is an indicator for the correctness of the above mentioned assumptions and proposed reaction mechanism. 23
The released amount of steam. The increase of
23
Na+ and
58
NaCl+ were increasing during the experiments with
Na+ is likely the result of the increased amount of vapour sodium
species and the fragmentation process in the MBMS. The increase of 58NaCl+ is likely a result of the equilibrium NaOH/NaCl. An increased amount of NaOH leads to an increased amount of NaCl due to the reaction balance. However, one must keep in mind that the amount of chlorine is limited as mentioned above. 56
KOH+ was detected in significant amount only during experiments with steam (Figure 2 and
3). The normalised peak areas of 56KOH+ were increasing during the experiments with steam with a factor of about 1.4 – 4.3. An explanation for the increase is the enhanced formation of highly volatile potassium hydroxide under the influence of steam. During the experiments without steam the amount of chlorine is the limiting factor of the release of potassium due to its “carrier function” during the release of alkali metal. The released amount of
74
KCl+
increased as well under the influence of steam (factor of about 1.1 – 3.6) as shown in Figure 3. The increased formation of KCl is likely a result of the equilibrium KOH/KCl as mentioned for NaOH/NaCl. The release of 36HCl+ is constant for HKN-S- and slightly decreasing for HKN-S+, HKS, and HKT during the experiments with steam. A likely explanation is the reaction of NaOH and/or KOH with HCl and the formation of alkali metal chloride. This is in agreement with the slightly increased formation of NaCl and KCl which is a result of the reaction balance as mentioned above. Key sulphur species detected during experiments with and without steam are 64
SO2+. The amount of
34
34
H2S+ and
H2S+ was increasing under the influence of steam with a factor of
about 1.2 – 2.1. Further, the amount of 64SO2+ was slightly increasing during the experiments
with steam, e.g. with a factor of about 1.5 for the lignite HKS. A likely explanation is that an enhanced formation of alkali metal hydroxide and possibly alkali earth metal hydroxide lead to a reduced formation of metal sulphide species. Therefore, more sulphur is released in form of H2S and SO2.
Figure 2. Intensity-Time-Profiles for several species released during gasification experiments with the lignite HKN-S- at 1400°C in He/7.5%O2 with and without water steam. Two phases of thermal conversion can be observed: pyrolysis (I) and char reactions (II).
Figure 3. The averaged, normalised peak areas for the species 34H2S+, 36HCl+, 39K+/39NaO+, KOH+, 58NaCl+, 64SO2+, and 74KCl+ released during gasification experiments with lignite.
56
4. CONCLUSIONS In this work the influence of steam on the release of Na-, K-, S-, and Cl-species during gasification of four Rhenish lignite at 1400°C was investigated. Hot gas analysis was carried out by molecular beam mass spectrometry. Key chemical species detected during gasification experiments with and without steam are 58
NaCl+,
64
SO2+,
74
23
Na+,
34
H2S+,
36
HCl+,
39
K+/39NaO+,
KCl+. In principal all species under investigation, except
36
56
KOH+,
HCl+, showed
an increasing release under the influence of steam. Especially, the amount of the alkali hydroxide specie
56
KOH+ was significant enhanced. Further, there are indications that the
release mechanism including the formation of intermediate NaOH and Na2CO3 plays a major role in the release mechanisms of alkali metals as proposed by several authors.
ACKNOWLEDGMENT The work described in this paper has been done in the framework of the HotVeGas-EM project supported by Bundesministerium für Wirtschaft und Technologie (FKZ 0327773C). The coals were kindly supplied by RWE Power AG.
REFERENCES [1] Bundesministerium für Wirtschaft und Technologie: Energie in Deutschland - Trends und Hintergründe zur Energieversorgung in Deutschland, Berlin, 2009. [2] J.M. Beer, High efficiency electric power generation, The environmental role, Progress in Energy and Combustion Science 33 (2007) 107–134. [3] M. Müller, D. Pavone, M. Rieger, R. Abraham, Hot Fuel Gas Cleaning in IGCC at Gasification Temperature, Fourth International Conference on Clean Coal Technologies, Dresden, Germany May 2009. [4] E. Raask, Mineral impurities in coal combustion, Hemisphere Publishing Corporation, 1985. [5] M. Bläsing, M. Müller, Mass spectrometric investigations on the release of inorganic species during gasification and combustion of German hard coals, Combustion and Flame 157 (2010) 1374-1381. [6] M. Bläsing, M. Müller, Mass spectrometric investigations on the release of inorganic species during gasification and combustion of Rhenish lignite, Fuel 89 (2010) 2417–2424. [7] M. Bläsing, T. Melchior, M. Müller, Influence of the temperature on the release of inorganic species during high-temperature gasification of hard coal, Energy & Fuels (2010), DOI: 10.1021/ef100398z. [8] K.J. Wolf, Untersuchungen zur Freisetzung und Einbindung von Alkalien bei der reduzierenden Druckwirbelschichtverbrennung. PhD, RWTH Aachen, Aachen, 2003. [9] T. Tsuyoshi, N. Ichiro, Emission control of sodium compounds and their formation mechanisms during coal combustion, Proceedings of the Combustion Institute 31 (2007) 2863–2870. [10] X. Wei, J. Huang, T. Liu, Y. Fang, Y. Wang, Transformation of Alkali Metals during Pyrolysis and Gasification of a Lignite, Energy & Fuels 22 (2008) 1840–1844. [11] A. Kosminski, D.P, Ross, J.B. Agnew, Reactions between sodium and silica during gasification of low-rank coal, Fuel Processing Technology 87 (2006) 1037 – 1049. [12] A. Kosminski, D.P, Ross, J.B. Agnew, Transformations of sodium during gasification of low-rank coal, Fuel Processing Technology 87 (2006) 943 – 952.
Kinetics of Char and Catalyzed Char Gasification under High H2 and Steam Partial Pressure Katsuhiro Nakayama*+, Shiying Lin**, and Yoshizo Suzuki* *National Institute of Advanced Industrial Science and Technology 16-1 Onogawa, Tsukuba, Ibaraki 305-8569, Japan ** Japan Coal Energy Center 3-14-10 Mita, Minato-ku, Tokyo 108-0073, Japan
ABSTRACT Effects of high H2 and H2O (steam) partial pressure (H2 up to 1.2 MPa, H2O up to 2 MPa) on coal char and Ca loaded coal char gasification rate were investigated by using thermogravimetric apparatus with low temperature range (923 K to 1123 K).
It is found that gasification rate was quickly decrease with
PH2/PH2O increase, but gasification rate of Ca loaded coal char was much higher than that of coal char gasification, and the reduction of gasification rate by the increase of H2 partial pressure for Ca loaded coal char gasification was smaller than that for char gasification.
Effect of H2O and H2 partial pressure on char
and Ca loaded coal char gasification rate were analyzed by applying L-H mechanism. Temperature effects on gasification rate under high H2 and H2O pressure shown much difference for char and Ca loaded char. Gasification rate of Ca loaded coal char was about 4 times higher than that of char gasification rate at 1123 K, but it was 26 times higher at comparatively low temperature 923 K. The activation energy for char gasification, E was 234 [kJ/mol], and for Ca loaded coal char, E was smaller as 138 [kJ/mol]. INTRODUCTION Hydrogen is expected to serve as a clean source of energy, and its use will thus protect the environment, which has been polluted by human energy consumption for industrial and domestic uses. The International Association for Hydrogen Energy has advanced the view that, through the use of fuel cell technology, hydrogen can be used to produce power more efficiently than through traditional methods, and serve as clean fuel for transportation. However, hydrogen is a secondary energy and must be produced using some other energy source, such as nature and fossil energies. 1
Therefore, the development of highly
efficient technology for H2 production is important. The HyPr-RING method (Hydrogen Production by Reaction Integrated Novel Gasfication) is a highly efficient method of producing H2 from coal in a single reactor (gasifier)2.
Lime (CaO) was used to
attempt completely absorb CO2 in a gasifier. Once the CO2 is removed, CO is completely converted to H2 in the gasifier, and product gas containing high hydrogen content is produced in a single step +
Corresponding author, Katsuhiro Nakayama, Mail address: Clean fuel research group, AIST (west), 16-1 Onogawa, Tsukuba 305-8569, Japan, Tel: +81-298-8223; Fax: +81-298-8209; E-mail: [email protected]
1
in the gasifier (Eq.(1)).
C (char ) 2H 2O CaO CaCO3 2H 2
(1) Lin et al have reported the production of 80% H2, along with a small amount of methane (17%), in a 3
single reactor (3.0MPa, 923 K), the concentrations of CO and CO2 in the product gas were less than 3 % in comparatively low gasification temperature (as 923 K).
Since hydrogen partial pressure in gasifier is much
higher than that in a traditional gasifier, it is important to clear the coal (char) steam gasification behavior under high H2 and steam partial pressure. Effect of hydrogen presence on char steam gasification (Eq.(2)) have been studied by many previous studies4-11.
C (char ) H 2O CO H 2
(2)
Matsumoto and Kaida7 studied char catalyzed steam gasification with H2 partial pressure from 0.0025 to 0.095 MPa, and reported that gasification reactivity decrease with increase H2 partial pressure. Kajita et al10 studied steam gasification of biomass chars with or without H2 presence (H2 partial pressure from 0 to 0.02 MPa), and reported the noncatalytic and catalytic reactions obeyed respective Langmuir-Hinshelwood mechanisms that involved inhibition by H2. However, char gasification behavior under high H2 partial pressure (up to 1.5MPa) as well as steam partial pressure (up to 2.0 MPa) have not experimentally verified yet. In this study, we measured char gasification rate under high H2 and steam partial pressure until total pressure 3.0 MPa by using a high pressure thermogravimetric apparatus. Effect of H2 partial pressure on char steam gasification rate was examined, and kinetic reaction rate was established and compared with results of previous studies.
In order to improve char steam gasification rate under low temperature high
hydrogen partial pressure, catalytic gasification with a Ca loaded coal char was also measured and compared with result of the char gasification. THERMODYNAMIC STUDY 2.1 Variation of equilibrium constant of reactions (1) and (2) with temperature Equilibrium constants, KC, of reactions (1) and (2) with temperature variation were calculated from the Gibbs energy change, ΔG, for reactions (1) and (2) is shown in Eq. (3).
HSC Chemistry 4.0 software was
12
used in calculation based on thermodynamic data .
KC e
G RT
(3)
Since reaction (1) is a gas mole equal reaction, equilibrium constant, KC(1), can be defined by 2
K C (1)
PH 2 , PH O 2
(4)
and reaction (2) is a gas mole increase reaction, KC(2) can be defined by
PCO PH 2 K C ( 2) PH O 2 2
.
(5)
where R is gas constant [J·mol-1·K-1], T is reaction temperature [K] and PH2, PH2O and PCO are partial pressure for H2, H2O(steam) and CO, respectively [atm] Figure 1a and 1b shows calculation results of which KC(1)=(PH2/PH2O)2 and KC(2)=(PCOPH2/PH2O) varied with temperature, respectively.
From figure 1a, it can been seen that, the equilibrium constant KC(1)
increases with temperature decrease. Such as at 1073 K, the KC(1)=(PH2/PH2O)2=36, means the reaction (1) may occur under atmosphere that H2 is 6 times to H2O (steam), and when temperature reduce to 923K, the KC(1) increases to 144, induces reaction (1) can occur under atmosphere that H2 is 12 times to H2O(steam). The tendency of equilibrium constant of reaction (1) with temperature indicates clearly that, the reaction (1) is favorable to low temperature. Figure 1b shows equilibrium constant of reaction (2) as a function of temperature.
At 1073 K,
KC(2)= (PCOPH2/PH2O)=7, means reaction (2) can occur under H2 is 2.6 times to H2O (steam), but when temperature reduce to 923 K, KC(2) decreased to 0.6, means the reaction (2) can’t occur when H2 is larger than 0.8 times to H2O (steam).
This indicates that, the reaction (2) is not favorable to low temperature, is
favorable to high temperature.
Kc=(PH2/PH2O) /PH2O [atm/atm]
Kc=(PH2/PH2O) 2 [atm/atm]
1000
C+2H2 O+CaO=CaCO3 +2H2
1000
PH2 /PH2O =12 100 PH2 /PH2O =6
(a) 10 923 K
1073 K 1 0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
100 10 PCOPH2 /PH2O =7 1 0.1
PCOPH2 /PH2O =0.6 1073 K
0.01 0.6
0.7
103 /T [1/K]
0.8
0.9
1
923 K 1.1
1.2
1.3
103 /T [1/K]
Figure 1 Plot of Equilibrium constant, KC vs. reciprocal of temperature, 103/T; (a) reactions (1) is favorable to low temperature, and (b) reaction (2) is favorable to high temperature.
3.
EXPERIMENT
3.1 Sample preparation Table 1 shows the proximate and the ultimate analyses for parent coal. Table 2 shows the properties of char and Ca loaded coal char which were used in char gasification rate measurement. Ion exchange method was used to prepare Ca loaded coal, and calcium loaded in coal was about 5 wt%. The char and Ca loaded coal char were made by pyrolyzing coal and Ca load coal in fluidized bed at 1123 K under N2 atmosphere about 10 minutes. respectively.
Carbon contents of the char and Ca loaded coal char are 97 wt% and 81 wt%,
The char and Ca loaded coal char were crushed and sieved to approximately 0.125-0.25 mm
for the gasification rate measurements.
3
1.4
Table 1
Properties of the parent coal
Proximate anal. Wt% Vol. Mois. FC Ash 42.3 8.8 44.1 4.8 Table 2
Ultimate anal. Wt%, daf C H N O 70.44 4.90 0.35 24.30
Properties of char and Ca loaded coal char FC[-] 0.97 0.81
Char Ca loaded coal char
Fash[-] 0.03 0.19
FCaO in char[-] 0.17
3.2 Thermogravimetric apparatus Figure 2 shows a schematic diagram of high pressure thermogravimetric apparatus used to measure the char gasification rate. The thermogravimetric apparatus is mainly consisted with three sections. A balance room made by stainless steel is designed up to pressure of 5.0 MPa. A balance bar is set in the balance room.
Moving of balance bar with weigh change of sample is captured by a laser displacement
sensor, and be recorded in a computer. A reactor is made by stainless pipe with 20 mm inner diameter and 500 mm in length. Between the reactor and the balance room is a cooling section which constructed by double pipes. Inside pipe is gas quid tube (Din 6mm), transfers the purge gas (N2) from balance room to reactor. Outside pipe is cooled by N2 gas for intercepting the heat conduction from reactor to balance room. Sample basket (882mm) is made by platinum/rhodium, mesh (200 mesh), and hanged up on the balance bar with platinum/rhodium wire, where pass-through the inside pipe of the cooling section into reactor.
High pressure (~5 MPa) steam is generated by forcing water through multiples stainless coils,
which are housed in an electricity furnace (1073 K). electricity heater until into reactor.
Steam line is warmed not less about 723 K by
High pressure H2 is supplied by a cylinder. Laser displacement sensor Balance room Balance bar P Cooling gas Spring
Purge gas N2
HSteam 2 /steam
Gas guide tube
Reactor Sample pan
Thermo Couple
Thermo Couple
Furnace P Gas
Gas/Liquid separator
Pressure regulator Water
Figure 2 schematic diagram of high pressure thermogravimetric apparatus
4
3.3 Experimental procedure When char or Ca loaded coal char sample about 0.1 g was dispersedly placed in sample basket and set in the reactor, the reactor pressure was raised by N2, and the reactor temperature was raised by electricity furnace with heating rate about 25 K/min to reaction conditions. Purge gas of about 0.5 l/min N2 was supplied during whole measurement.
After reactor pressure and temperature were stabilized, steam as well
as hydrogen were injected into reactor to start gasification rate measurement. While, reaction weight change of sample was continuously measured and recorded into a computer. Steam supply was adjusted to about 1 l/min for char gasification. Figure 3a and 3b show the variations of temperature, pressure and weight change during gasification rate measurement. Compared with weight change of char gasification which was monotonously decrease after steam injecting, the weigh change of Ca loaded char gasification shown a sharp increase in the initial gasification time, and then monotonously decreased with gasification time.
The sharp increase of weight change is caused by the CaO in char reacted with CO2 to form CaCO3
(Eq.(6)) . CaO+CO2→CaCO3 N2
(6)
H2/steam/N2
(a) Adaro char Temperature
W0 Variation of Temperature and Pressure, and weight change
Weight
Wt Pressure
t0 (b) Adaro Ca loaded coal char
Weight
Wr
Temperature
W0 Wt
Pressure
t0 Time
Figure 3 The variations of sample weight, temperature and pressure during measurement. The carbon conversions of char and Ca loaded coal char gasification with reaction time were defined by using equations (7) and (8), respectively.
X
X
W0 Wt W0 FC
(7)
12 FCaO ) Wt 56 W0 FC
(Wr W0
5
(8)
where, X [-] is carbon conversion. W0 [kg] and Wt [kg] are the initial weight and the weight at time t [s]. Wr [kg] is maximum weight in measurement [kg].
FC [-] and FCaO [-] are the carbon contents in char and
CaO content in Ca loaded coal char. Initial rate of char gasification at X=0.1 was defined by equation (9).
RX 0.1
dx dt (1 X )
[s-1]
;
(9)
X 0.1
4. RESULTS and DISCUSSION 4.1 Effect of high H2 and steam partial pressure on char and Ca loaded coal char gasification rate Profiles of carbon conversion versus reaction time for char gasification and Ca loaded coal char gasification with various PH2/PH2O ratios are shown in Figure 4. Temperature and total pressure of the experiments in Figure 4 are 1023K and 3.0MPa.
H2 partial pressure is a range from 0 to 1.2 MPa
corresponding steam partial pressure is from 2 to 1.2 MPa.
It can be seen that, gasification rate for both of
char gasification and Ca loaded coal char gasification, are quickly decrease with P H2/PH2O ratio increase. For example, with H2 absence (PH2/PH2O=0), char gasification converted about 40% carbon with 30 minutes. With H2 partial pressure increase, reaction rate shows decrease, such as at PH2/PH2O=0.2, carbon was converted about 20 % with 30 minutes, and at PH2/PH2O=1, carbon was converted only 6 % with 30 minutes. H2 presence strongly retarded char gasification rate. Results of Ca loaded coal char gasification also shown the rate retardation tendency by H2 presence, such as about 86%, 67% and 55% carbon were converted with 30 minutes at PH2/PH2O=0, 0.2 and 1.0, respectively. However, gasiafication rate of Ca loaded coal char were much higher than that of char gasification, and the reduction of gasification rate by the increase of H2 partial pressure for Ca loaded coal char gasification was smaller than that for char gasification. Such as when PH2/PH2O ratio increased from 0 to 1, the carbon conversion at 30 minutes reduced to 0.64 times for Ca loaded coal char, and reduced to 0.15 times for char gasification. T: 1023 K; P: 3.0 MPa (N2 0.5 l/min, steam 1 l/min]
Carbon conversion, X [-]
1
PH2 /PH2O=0 P /PH2O=0.2 HH2 2 /Steam=0.2 PH2 /PH2O=1.0 PH2 /PH2O=0 PH2 /PH2O=0.2 PH2 /PH2O=1.0
Ca loaded coal Char
0.8 0.6 0.4
Char
0.2 0
0
10
20
30 40 50 60 Time [min] Figure 4 Carbon conversions of char and Ca loaded char gasification under various PH2/PH2O ratio.
6
4.2 Expression of gasification rate variation with H2 and H2O partial pressure Variation of RX=0.1 with H2O and H2 partial pressure were analyzed applying L-H mechanism associated with adsorption onto the active carbon sites and it’s desorption. This mechanism gives the overall rate of reaction as follows
RX 0.1
k a PH 2O 1 kb PH 2O kc PH 2
.
[s-1]
(10)
[s]
(11)
where, ka, kb and kc are coefficient. The reciprocal of Eq. (10) is shown as Eq.(11)
1 RX 0.1
kc PH 2 k 1 . b k a PH 2O k a k a PH 2O
When PH2/PH2O=0 (with H2 absence), Eq.(11) can be re-wrote into Eq.(12)
1 RX 0.1 C
where
1 C , k a PH 2O
kb . ka
[s]
[s]
(12)
(13)
Eq.(12) also can be re-wrote into Eq.(14) to show kc is function with respect to PH2/PH2O ratio,
kc
k
a
/ RX 0.1 1 / PH 2O kb ( PH 2 / PH 2O )
,
[MPa-1]
(14)
4.2.1 Obtain ka, kb and kc from gasification rate ka in Eq.(12) and kb in Eq.(13) were examined by plotting of the reciprocal of RX=0.1 (1/RX=0.1) versus the reciprocal of PH2O (1/PH2O). Figure 5 shows the plot for char gasification, and Figure 6 shows the plot for Ca loaded coal char gasification.
It can see that, relationship of (1/R) to (1/PH2O) for both char and Ca
loaded coal char gasification were shown linearly in Figure 5 and Figure 6. The linear relationships show applicability of Eq. (12) to the results of gasification rate measured. By using any two points i, ii on the line, ka and kb can be given by Eq.(15) –Eq.(17).
ka
C
1 / PH 2O ,i 1 / PH 2O ,ii 1 / RX 0.1,i 1 / RX 0.1,ii
1 RX 0.1,i
;
1 k a PH 2O ,i
kb C k a
[MPa-1s-1]
(15)
(16) (17)
Accordingly, for char gasification, ka and kb are obtained from Figure 5 as ka=1.25×10-3 ,
[MPa-1s-1]
C=1.85×103 , [s]
7
(18) (19)
kb=2.32. [MPa-1],
(20)
and for Ca loaded coal char gasification, ka and kb are obtained from Figure 6 as ka=0.74×10-3 , [MPa-1s-1]
(21)
C=1.05×103 , [s]
(22)
kb=0.014. [MPa-1]
(23)
1/Rx=0.1 [103s]
8 ka=1.25x10-3 [ MPa -1 s-1 ]
Char 1023 K, PH2 =0
6
kb =2.32 [MPa -1 ]
4 (i) (ii)
2
0 0
1
2
3
4
5
1/PH2O [MPa-1]
Figure 5 A plot of 1/RX=0.1 vs 1/PH2O for char gasification (1123 K, PH2=0).
1/Rx=0.1 [103s]
2
Adaro Ca char 1023 K, PH2 =0
ka=7.14x10-4 [MPa -1 s-1 ] kb =0.014 [MPa -1 ]
1
(i) (ii)
0 0.4
0.5
0.6
0.7
0.8
1/PH2O
[MPa-1]
0.9
1
Figure 6 A plot of 1/RX=0.1 vs 1/PH2O for Ca loaded coal char gasification (1123 K, PH2=0).
kC in Eq.(14) was examined by plotting of the term, (ka/RX=0.1-1/PH2O-kb), against the ratio of PH2/PH2O. Plot for char gasification is shown in Figure 7 and for Ca loaded coal char gasification is shown in Figure 8. The linear relationships show applicability of Eq. (14) to the measurement gasification rate.
8
Using any points on this line, kc can be given from Figure 7 for char gasification kc=33.33, [MPa-1]
(23)
and from Figure 8 for Ca loaded coal char gasification kc=0.66. [MPa-1]
(24)
ka/Rx=0.1-1/PH2O-kb
50 k c =[k a/R-1/PH2O -k b]/[PH2 /PH2O ] =33.33 [MPa-1 ]
Adaro char 1023 K
40 30 20 10 0 0
0.2
0.4
0.6
0.8
1
1.2
PH2 /PH2O [-]
Figure 7
Plot of term [ka/RX=0.1-1/PH2O-kb] vs. PH2/PH2O for char gasification (1023 K).
ka/Rx=0.1-1/PH2O-kb
1 k c =[k a/R-1/PH2O -k b]/[PH2 /PH2O ] =0.66 [MPa-1 ]
Adaro Ca char 1023 K
0.8 0.6 0.4 0.2 0 0
0.2
0.4
0.6
0.8
1
1.2
PH2 /PH2O [-]
Figure 8
Plot of term [ka/RX=0.1-1/PH2O-kb] vs. PH2/PH2O for Ca loaded coal char gasification (1023 K).
Table 3 shows parameters of ka, kb and kc, and the activation energy for char and Ca loaded coal char gasification. Table 3
Activation energy E and parameters of ka, kb and kc. E [kJ/mol/K]
Ka[10-3MPa-1s-1]
Kb [MPa-1]
Kc[MPa-1]
Char gasification
288
1.25
2.32
33.33
Ca loaded coal char gasification
214
0.74
0.014
0.66
9
Figure 9 shows initial gasification rate calculated by using Eq.(10) with parameters of ka, kb and kc shown in Table 3 for both char and Ca loaded coal char gasificaiton. With H2 absence, gasification rate for both char and Ca loaded coal char are increase with PH2O increase. The increase of gasification rate by increase PH2O for Ca loaded coal char is faster than the increase gasification rate for char gasification. Ca loaded in char my improve H2 (steam) adsorption on char surface area, consequently improve gasification. It can also see in figure 9 that, with PH2/PH2O ratio increase, gasification rate for both of char and Ca loaded coal char are decrease, such as the experimental result shown in Figure 4. However, decrease of rate by the increase PH2/PH2O ratio for Ca loaded coal char is slower than that for char gasification. It is considered that, H2O (steam) adsorption improvement by Ca which loaded in char may protect H2 adsorption on the active carbon site, consequently reduce rate retardation by H2.
Initial gasification rate [s-1]
0.01 Ca loaded coal char gasification rate Char gasification rate 0.001
0.0001
PH2O [MPa]
Figure 9
0.5
1.0
PH2/PH2O [-]
1.5 0.0
1.5 0.33
0.67
1.0
Initial gasification rate calculated by using Eq.(10) wtih parameters of ka, kb and kc.
4.3 Temperature effect on the char and Ca loaded coal char gasification rate To investigate the temperature effect on gasification rate under high H2 and H2O (steam) partial pressure, char and Ca loaded coal char were gasified under PH2/PH2O=1 with various temperature. Figure 10 shows results of carbon conversion vs. gasification.
From this figure, it can be seen that, both
gasfication rate for char and Ca loaded coal char increased with the increase temperature, and at same temperature, gasification rate for Ca loaded char was much higher than that of char gasification.
10
P: 3.0 MPa; P H2 /PH2O =1
Carbon conversion, X [-]
1 0.8
1123 K
1073 K 1023 K 973 K
0.6 Char
0.4 0.2 0
Figure 10
1073 K 1023 K 973 K 923 K
Ca loaded coal Char
0
10
20
30 40 Time [min]
50
60
Carbon conversion for char and Ca loaded coal char with variation of temperature.
Arrhenius plot of gasification rate for both char and Ca loaded coal char under various P H2/PH2O ratios are shown in Figure 11. At PH2/PH2O=1, it can be seen that, gasifiacaiton rate of Ca loaded coal char was about 4 times higher than that of char gasification rate at 1123 K, but it was 26 times high at comparatively low temperature 923 K.
The activation energy was obtained from rate date under PH2/PH2O=1.
The
activation energy for char gasification, E was 234 [kJ/mol], but for Ca loaded coal char, E was smaller as 138 [kJ/mol]. It was considered that Ca loaded coal char have more good reactivity at low temperature than char gasification. The experimental data also support the thermodynamic study which was described in Figure 1.
Char at P H2 /PH2O =0 Char at P H2 /PH2O =0.2 Char at P H2 /PH2O =1 Ca loaded coal char at P H2 /PH2O =0 Ca loaded coal char at P H2 /PH2O =0.2 Ca loaded coal char at P H2 /PH2O =1
4
-Ln RX=0.1 [s-1]
5 6 7
E=138 kJ/mol
8 9 10 11
E=234 kJ/mol
3.0 MPa
12 0.85
0.90
0.95
1.0
1.1
1000/T [1/K]
Figure 11
Arrhenius plot of reaction rate, RX=0.1.
11
1.1
5. CONCLUSION Coal char and Ca loaded coal char were gasified under high H2 and H2O (steam) partial pressure (H2 from up to 1.2 MPa, H2O from up to 2 MPa) and gasification rate was measured. Gasification rate for both of char gasification and Ca loaded coal char gasification, were quickly decreased with PH2/PH2O increase, but gasification rate of Ca loaded coal char was much higher than that of char gasification, and the reduction of gasification rate by the increase of H2 partial pressure for Ca loaded coal char gasification was slow than that for char gasification. Effect of H2O and H2 partial pressure on char and Ca loaded coal char gasification rate were analyzed applying L-H mechanism as reaction
RX 0.1
k a PH 2O 1 kb PH 2O kc PH 2
.
Ka, kb and kc were obtained from experimental rate results and simulated gasification rate under wide range of H2 and H2O partial pressure. Temperature effect on gasification rate shown much difference for char and Ca loaded char. Gasification rate of Ca loaded coal char was about 4 times higher than that of char gasification rate at 1123 K, but it was 26 times high at comparatively low temperature 923 K.
The activation energy for char
gasification, E was 234 [kJ/mol], but for Ca loaded coal char, E was smaller as 138 [kJ/mol]. It can be considered that Ca loaded coal char have more good reactivity at low temperature than char gasification.
ACKNOWLEDGMENT This work was carried out under the frameworks of the project of the high performance coal gasification project for the next generation. The authors give their great thanks to NEDO for their financial support.
REFERENCES 1
Bisio, A. and Boots S. “Encyclopedia of Energy technology and the Environment”, Volume 3,
2 3
John Wiley & Sons, Inc., New York, 1995 Lin, S. Y., Suzuki Y., Hatano H. and Harada M. Kagaku Kogaku Ronbunshu, 1999, 27(4), 520 Lin, S. Y., Suzuki Y., Hatano H. and Harada M., Fuel, 2002, 81, 2079-2085
4 Yang, R.T., Duan, R.Z. Carbon 1985, 23, 325 5 Huttinger, K.J. Carbon 1988, 26, 79 6 Lussier, M. G.; Zhang, Z.; Miller, D. J. Carbon 1998, 36, 1361 7 Matsumoto, S. and Kaida, Y., “Rate retardation by hydrogen in the calcium catalyzed steam gasification”, Fuel, 1992, 71, 467-468. 8 Wei, L., Xu, S., Zhang, L., Liu, C., Zhu. H. and Liu S., “Steam gasification of biomass for hydrogen-rich gas in a free-fall reactor”, Hydrogen Energy, 2007, 32, 24-31 9 Messenbock, R. C., Dugwell, D. R. and Kandiyoti, R., “CO2 and steam-gasification in a high pressure wire-mesh reactor: the reactivity of Daw Mill coal and combustion reactivity of its chars”, Fuel 1999, 78, 781-793 10 Kajita, M., Kimura, T., Norinaga, K., Li, C-Z., Hayashi, J., “Catalytic and Noncatalytic Mechanisms in Steam Gasification of Char from the Pyrolysis of Biomass”, Energy & Fuels, 2010, 24, 108-116 11. Weeda, M., Abcouwer, H.H., Kapteijn, F. and Moulijn, J.A., “Steam gasification kinetics and burn-off behavior for a bituminous coal derived char in the presence of H2”, Fuel Processing Technology, 1993, 36, 235-242.
12
12
Barin, I. “Thermochemical data of pure substances,” VCH., Verlagsgesellschaft mbH, D-6940 weinbeim (Federal Republic of Germany), 1989
13
MODELING OF STEAM GASIFIER IN DUAL FLUIDIZED BED GASIFICATION Toshiyuki SUDA* , Zhihong LIU, Makoto TAKAFUJI and Masahiro NARUKAWA Research Laboratory IHI Corporation 1, Shin-nakahara-cho, Isogo-ku, Yokohama 235-8501 JAPAN * [email protected]
tel : +81-45-759-2867
ABSTRACT Dual Fluidized Bed Gasification (DFBG) is one of the promising technology to produce high calorific and hydrogen rich syngas without using pure oxygen. It is a combination of steam blown gasifier and air blown combustor, and several configuration of dual fluidized bed (selection of bubbling fluidized bed or high speed riser) has been proposed in the world. Up to know, this technology has been used in gasification of biomass or waste to use syngas as a fuel of power generation or feed stock of synthetic fuel. Although it is preferable to apply this technology to low rank coal gasification, designing of the gasifier is one of the difficult issue because of the difference of gasification behavior between biomass and coal. Basically, coal has less volatile and more char than biomass, which means that larger residence time in the gasifier is necessary to achieve enough carbon conversion. Therefore, the modeling of fluidized bed steam gasifier is performed to predict the performance of the gasifier for low rank coal. The model is made by the combination of residence time distribution (RTD) of coal in bubbling fluidized bed with the fundamental rate equation of steam gasification. RTD of coal is calculated from the empirical model of particle movement in the fluidized bed including mixing and segregation effect. Rate equation of steam gasification is derived from the experiment using small scale fluidized bed steam gasifier. The gasification experiment using pilot scale DFBG is also performed, and the result of carbon conversion is compared with the prediction by the modeling work. From the result, although there is a certain difference between the experimental and predicted value, both value showed the same tendency for the effect of temperature, steam flow rate, size of the gasifier, which shows that the model basically can be used to predict the performance of the steam gasifier in DFBG for low rank coal. Keywords : Gasification, Combustion, Lignite coal, Fluidized bed, Modeling
1.Introduction .
The importance of coal gasification technology is increasing in the world due to the
rising cost of oil and natural gas. High volatile fuels like lignite coal and biomass are suitable for gasification because of its higher reactivity [1], and considering its huge reserve, much simple, cheap and efficient gasifier is needed for the practical application of the system. Circulating fluidized bed gasifier with dual reactor (DFBG : Dual Fluidized Bed Gasifier) is suitable for this purpose because it can produce synthesis gas with high calorific value without using high temperature, pure oxygen or pressurized condition. A novel gasification technology, Twin IHI Gasifier (TIGARR) has been developed for the production of synthesis gas and power generation from low rank coal. This gasifier is a combination of bubbling fluidized bed (BFB) gasifier and circulating fluidized bed (CFB) combustor, as shown in figure 1. Fuel is first fed into the bubbling fluidized bed gasiifer where pyrolysis and steam gasification of char occurs. Remaining char is transported into the combustor where char is combusted with air and heats up the circulating bed materials.
Circulating bed material from the combustor supplies the heat
needed for gasification reaction in the gasifier. Steam is mainly used as a gasification agent and reforming medium in the gasifier. The advantage of this gasifier is that synthesis gas with higher calorific value (less contaminated by nitrogen) can be obtained without using pure oxygen because combustor and gasifier is separated each other. Also, as part of the char is used in combustion, there is no need to gasify all of the char whose gasification reaction rate is rather smaller than the rate of combustion or pyrolysis reaction. DFBG has been mainly used as a biomass gasifier, as reviewed in the paper [2]. It is well known that biomass has a larger volatile matter amount compared with that of coal, and so gasification reaction strongly depends on pyrolysis (devolatilization) reaction. Pyrolysis reaction is a self-decomposition reaction by the heat, whose reaction rate is rather faster than the steam gasification reaction and does not strongly depends on the gasification condition. In contrast, coal has a larger amount of fixed carbon, and so the steam gasification reaction of char is rather important than pyrolysis reaction even for the low rank coal gasification. Steam gasification reaction rate is rather slower than the combustion reaction, and strongly depends on the gasification condition like temperature, steam partial pressure and so on. One of the reason why bubbling fluidized bed gasifier is chosen in figure 1 is to achieve higher residence time in the gasifier in order to obtained higher carbon conversion [3]. Therefore, it is important to know the basic behavior of the gasifier and to predict the performance in order to achieve much higher conversion of coal and to optimize the designing of the gasifier. In this paper, the prediction of gasifier carbon conversion using simple semi-empirical model is shown. The predicted value is compared with the experimental data obtained from the
pilot-scale prototype gasifier.
Fig.1 Overview of TIGARR gasifier
2. Modeling for the prediction of carbon conversion in the gasifier 2.1 Basic concept Carbon conversion in the gasifier shown in figure 1 is controlled by pyrolysis of raw coal and steam gasification reaction of char. As mentioned previously, it is well known that the rate of pyrolysis is rather faster than the steam gasification reaction. Therefore, carbon conversion in the gasifier can be considered to be controlled mainly by the steam gasification reaction of char. Factors which affect the steam gasification reaction of char are coal type, particle diameter, gasifier temperature, partial pressure of steam, residence time of coal particle and so on. One of the unknown factors which makes the analysis of fluidized bed gasifier much difficult is the residence time of coal particle inside the fluidized bed. As shown in figure 1, coal particle together with circulating fluidized bed material enter the gasifier from the one side, and move to the exit located at the another side of the reactor. Therefore, residence time in the fluidized bed is far from the plug flow reactor. The residence time is affected by the 'convection' flow of sand by the circulating bed material and also by the 'diffusion' of bed material caused by the fluidization. By these effects, residence time of coal inside the fluidized bed has a specific residence time distribution (RTD) affected by the flow rate of circulating bed material and also by the fluidization behavior. Together with the steam gasification reaction rate, the averaged carbon conversion in the gasifier can be expressed approximately by the following equation. ¥
h = ò E (t ) × X (t )dt 0
(1)
E(t) denotes the RTD of coal particle inside the fluidized bed and X(t) denotes the carbon conversion of coal particle at the certain time t, respectively. E(t) is the function of fluidization behavior, size of bed and coal particle, size of the gasifier and so on. Likewise, X(t) is the function of temperature, coal type, particle diameter, steam partial pressure around the particle and so on. Although E(t) and X(t) is not independent and can be affected by each other, it is useful if the gasifier performance can be predicted by the simple function shown in equation (1).
2.2 Residence time distribution (RTD) of coal in fluidized bed gasifier RTD in fluidized bed gasifier can be affected by convection and diffusion effect caused by the circulating flow of bed material and also by the mixing by the bubble inside the bed. Also, segregation of particle will have a significant effect on RTD when the characteristics of coal and bed particles differs much. The factors which affect the RTD in fluidized bed gasifier are shown in figure 2.
②diffusion: mixing by bubble movement
coal
① convection: flow of particles
bed material
③Segregation vertical distribution of coal
outlet
inlet
Fluidization gas Fig.2 Factors which affect the RTD of coal inside the gasifier From figure 2, distribution of coal concentration inside the bed can be approximately expressed by the following simple equation[4];
¶C ¶C ¶C ¶ ¶C ¶ ¶C +u + wS = (Dx ) + (Dz ) ¶t ¶x ¶z ¶x ¶x ¶z ¶z
(2)
Where C denotes the coal concentration inside the bed, u is the horizontal convection velocity and Dx is the horizontal diffusion coefficient of coal. Segregation of coal particle is also included in the
equation by the vertical convection velocity w multiplied by the segregation coefficient S. Dz is the vertical diffusion coefficient. Dx, Dz and S are function of particle characteristics, fluidization intensity, bed height and so on. It is difficult to derive general equations of these parameters, and so these values are obtained experimentally from the measurement of RTD using two dimensional fluidized bed with 1m width and 0.5 or 1m bed height. Figure 3 shows the example of RTD calculation result with different bed length according to equation (2). Residence time in figure 3 is shown by the dimensionless time which actual time is divided by the averaged residence time τ. Averaged residence time τ is calculated by dividing the mass of the bed by circulating flow rate of the bed material. There is a steep peak at the lower range of the residence time, and has a broad distribution at the higher range. If the reactor is similar to the plug flow reactor, residence time must have a steep peak at the averaged residence. However, by the effect of diffusion and segregation, the result shows that the residence time has a broad distribution, and the length of the fluidized bed strongly affects the RTD.
2
Cal.(L=1m) Cal.(L=3m) Cal.(L=9m) Cal.(L=18m)
E [-]
1.5 1
0.5 0 0
0.5 1 1.5 2 dimensionless time t/τ [-]
2.5
Fig.3 Prediction the effect of bed length on RTD [4]
2.3 Validation of the model of RTD In order to validate the model of RTD, measurement of RTD using bubbling fluidized bed is conducted. Figure 4 shows the apparatus for the experiment (L=3m). Bed particle is continuously fed from the one side (inlet) of the fluidized bed at a certain flow rate. Coal particle is instantaneously fed from the inlet. RTD of coal particle is calculated by measuring the coal concentration of the mixture of bed material and coal particle at the outlet of fluidized bed. Figure 5 shows the calculated RTD of the coal compared with the experimental data with different bed length (L=1,3m). The predicted and experimental RTD shows a good match, which
shows that residence time of coal particle in fluidized bed gasifier shown in figure 1 is far from a plug flow reactor, and RTD inside the fluidized bed can be approximately expressed by the combination of convection, diffusion and segregation effects shown in equation (2). It has to be mentioned that RTD strongly depends on the size of the reactor. Dx and Dz in equation (2) basically depends on the fluidization behavior and does not depends on the horizontal size of the fluidized bed. On the other hand, convection velocity u will be larger for larger horizontal size fluidized bed if the averaged residence time is set to be constant. Therefore, RTD becomes similar to the perfectly mixed reactor if the fluidized bed size is small, and becomes closer to that of plug flow reactor if the size becomes larger. 1.2 Exp.(L=1m) Cal.(L=1m)
E [-]
1 0.8 0.6
Sand, coal inlet
0.4 0.2 0 0
Over flow
0.5 1 1.5 2 dimensionless time t/τ [-]
1.2
Exp.(L=3m)
1
Cal.(L=3m)
E [-]
0.8
Fluidization air
2.5
0.6 0.4 0.2 0
Fig.4 Apparatus for RTD measurement (L=3m)
0
0.5 1 1.5 2 dimensionless time t/τ [-]
2.5
Fig.5 Validation of the RTD model [4]
3.Experimental for validation of carbon conversion modeling As shown in equation (1), carbon conversion is expected to be calculated from the combination of RTD calculation shown in equation (2) with the fundamental time history of steam gasification reaction of coal. In order to validate the model of carbon conversion, experiment using pilot-scale gasifier is conducted [5]. Figure 6 shows the schematic of pilot-scale experimental apparatus. The plant is configured with combustor, cyclone, gasifier, synthesis gas treatment system and flare stack. Combustor is cylindrical type, made by steel and refractory. Gasifier is square shaped, connected with steel pipe with cyclone standpipe and riser. Silica sand is installed inside the gasifier and the combustor as a bed material. Steam is produced by city gas fired boiler and fed to the gasifier as fluidization and gasification agents
after reheating by the electric heater. This is to maintain the flexibility of the test plant, and basically steam can be produced from the heat of synthesis gas and exhaust gas in the commercial plant. Coal is fed to the gasifier from the side wall by rotational feeder. Synthesis gas from the gasifier is cooled and cleaned by gas cooler and scrubber, and burned in flare stack after the measurement of gas composition and flow rate by gas chromatograph. Whole system is heated up by a city gas burner before the experiment. After the gasifier temperature reaches a set value, coal is fed into the gasifier to start the gasification. Synthesis gas is sampled from the duct after the gasifier cyclone in order to analyze the gas composition and tar concentration. Lignite and sub-bituminous coal is used as a fuel. Table 1 shows the results of proximate and ultimate analysis of coal used in the experiments.
Combustor
Gas cooler
Cyclone
Table 1 Physical and chemical analysis of coal Analysis \ Coal type Proximate(wt%)
Coal feeder
Gas cooler
Bag filter
Water spray Fan
Stack
A
B
Moisture Ash Volatile
31.8 2.2 33.4
16.4 0.8 40.8
Fixed Carbon
32.6
42.1
69.4 4.8 0.7 0.1 24.9 22,835
69.4 5.0 1.0 0.1 24.5 23,340
Elemental(wt%,daf)
Gasifier Fan
Secondary air Gas burner Steam Steam heater Primary air
Air heater
Flare stack
Carbon Hydrogen Nitrogen Sulfur Oxygen Heating Value (HHV) [kJ/kg]
Air
Fig.6 Pilot-scale gasifier [5]
4.Results 4.1 Gasifier performance Figure 7 shows the typical composition of synthesis gas from coal A. More than 50% of synthesis gas consists of hydrogen, which is the characteristic tendency of steam gasification. Calorific value of the synthesis gas is over 14,000kJ/m3N (dry,HHV), and contamination by nitrogen is less than 1%. Higher hydrogen synthesis gas cannot be achieved by the common air-blown gasifier (without water-gas-shift reaction), and is suitable for the synthesis of methanol or other chemical component. Composition of the syngas does not change dramatically by the gasification condition. Figure 8 shows the effect of gasification temperature on the gaisifer carbon conversion for
both coal A and B. Gasifier carbon conversion is calculated by the measurement of synthesis gas composition and also gas flow rate using argon as a tracer gas. Gasifer carbon conversion strongly increases with the gasifier temperature, which shows that gasifier temperature has a strong impact on the steam gasification reaction. Therefore, it is necessary to achieve higher gasifier temperature by increasing the circulating flow rate of bed material. Most important factor which affects the circulating flow rate is the superficial velocity of gas inside the riser combustor, and higher flow rate can be obtained by higher superficial velocity[5][6]. Carbon conversion of lignite coal is larger than that of sub-bituminous coal. It is well known that lignite coal has a larger amount of alkali and alkaline earth metal (AAEM) which work as a catalysis of gasification reaction. From the measurement of the amount of ion-exchanged Ca and Na, coal A has up to 0.4 mmol/g-char of exchangeable Ca+Na, which agrees well with the correlation between ion-exchanged Ca+Na and reactivity in the previous research [7]. It has to be mentioned that unburned carbon content in the ash collected from the riser is below 5%, and by including the carbon conversion in the riser combustor, total carbon conversion is
CH4 9%
CO2 19%
Other HC 3%
H2 51%
CO 18%
Fig.7 Syngas composition (coal A)
Gasifier carbon conversion [a.u.]
more than 99%. This is comparable to the existing high temperature gasifier.
750
Coal A
Coal B
800 850 Gasifier temperature [deg C]
900
Fig.8 Effect of gasifier temperature and coal type on carbon conversion
4.2 Evaluation of the carbon conversion model By introducing the RTD calculated from equation (2) and X(t) into equation (1), carbon conversion of the reactor can be predicted. X(t) in equation (1) is derived experimentally from the fundamental experiment using small scale fluidized bed gasifier made by cylindrical quartz glass with 50mm diameter and 650mm height [5]. Arrhenius type reaction rate equation including the effect of steam partial pressure is obtained from the time history of carbon conversion of lignite coal
by changing the bed temperature and steam partial pressure. Experimental data for the evaluation is obtained from the carbon conversion data using the apparatus shown in figure 6 by changing the gasifier temperature, gasifier size and steam flow rate. Figure 9 shows the comparison of the predicted and experimental value of carbon conversion in the gasifier. Although there is a certain difference, experimental and predicted value is approximately same even for the totally different experimental condition, which reveals that the simple model shown in equation (1) can be used to predict the basic performance of the fluidized bed gasifier The phenomena in the fluidized bed reactor is extremely difficult because of the influence of fluidization on mass and heat transfer and also on chemical reaction. The possible cause of the difference in figure 9 seems to be the change of the coal particle diameter during the reaction, temperature distribution in the fluidized bed, effect of the inhibition gas on the steam gasification reaction all of which is not considered in the model. Especially, it is well known that tar, hydrogen and hydrocarbon gas will strongly inhibit the steam gasification reaction of coal inside the fluidized bed [8]. However, from engineering point of view, it is useful to predict the gasifier performance simply, and this semi-empirical model will be of great use for the designing of the fluidized bed
Experimental [%]
gasifier.
Prediction [%]
Fig.9 Comparison of predicted and experimental value of gasifier carbon conversion
6.Conclusion The modeling for the prediction of carbon conversion in circulating fluidized bed gasifier with dual reactor (riser combustor and bubbling bed gasifier) is shown in the paper. The model is made by the combination of residence time distribution (RTD) of coal in bubbling fluidized bed with the fundamental rate equation of steam gasification. RTD of coal is calculated from the empirical model of particle movement in the fluidized bed including mixing and segregation effect. The model for
RTD prediction is validated by the fundamental experiment measuring RTD of coal in bubbling fluidized bed with different bed length. Rate equation of steam gasification is derived from the experiment using small scale fluidized bed steam gasifier. The gasification experiment using pilot scale DFBG is also performed, and the result of carbon conversion is compared with the prediction by the modeling work. From the result, although there is a certain difference between the experimental and predicted value, both value showed the same tendency for the effect of temperature, steam flow rate, size of the gasifier, which shows that the model basically can be used to predict the performance of the steam gasifier in DFBG for low rank coal.
7.Acknowledgment We express our gratitude to tekMIRA and Sojitz Corporation for their useful advice and cooperation work for the development of TIGARR gasifier.
8.References [1] L.Xiaojiang, H.Wu, J.Hayashi, C.Li,, Fuel 83 (2004) 1273-1279 [2] J.Corella, J.M.Toledo, G.Molina, Ind. Eng. Chem. Res. 46 (2007) 6831-6839 [3] G.Xu, T.Murakami, T.Suda, S.Kusama, Y.Matsuzawa, H.Tani, Proceedings of the 14th European Biomass Conf., Paris, France (2005) 670-673 [4] Z.Liu, M.Narukawa, M.Takafuji, T.Suda, Proceedings of the Fluidization XIII (2010) [5] T.Suda, M.Takafuji, Y.Matsuzawa, T.Fujimori, Proceedings of the Twenty Fourth Annual Int. Pittsburgh Coal Conf., Johannesburg, South Africa (2007) [6] H.T.Bi, J.R.Grace, J-X.Zhu, Int. J. Multiphase Flow 19(6) (1993) 1077-1092 [7] T.Takarada, Y.Tamai, A.Tomita, Fuel 64 (1985) 1438-1442 [8] B.Bayarsaikhan, N.Sonoyama, S.Hosokai, T.Shim, J.Hayahshi, C.Li, T.Chiba, Fuel 85 (2006) 340-349
Steam Gasification of Low Rank Coals with Ion-Exchanged Sodium Catalysts Prepared Using Natural Soda Ash
Yuu Hanaoka, Enkhsaruul Byambajav,a Takemitsu Kikuchi, Naoto Tsubouchi, and Yasuo Ohtsukab* Institute of Multidisciplinary Research for Advanced Materials, Tohoku University Katahira, Aoba-ku, Sendai 980-8577, JAPAN (Currently, aDepartment of Chemistry, National University of Mongolia Ulaanbaatar, MONGOLIA; bCooperative Research Center, Akita University, JAPAN) *Corresponding author (E-mail: [email protected])
ABSTRACT Ion exchange reactions of brown and sub-bituminous coals with natural soda ash, which is composed of > 99 % Na2CO3, have been studied at 20 – 40 °C without any pH-adjusting reagents, and the pyrolysis and subsequent steam gasification of the resulting Na-exchanged coals has been carried out with a fixed bed quartz reactor mainly at 700 °C. When the Na+ concentration and pH of an aqueous mixture of each coal and the soda ash are monitored in the ion exchange process, these factors decrease both at a larger rate with the brown coal that contains a higher content of COOH groups, showing that the ion exchange of Na+ with the H+ of the COOH takes place predominantly. About 65 % of the COOH can be exchanged with Na+ ions under the optimized conditions, irrespective of the coal type. The reactivity of these raw coals in steam at 700 °C is almost the same to be low, and char conversions are less than 20 mass % even after 2 h reaction. The exchanged Na promotes remarkably the gasification of both coals at this temperature, but the rate profile is different: The conversion for the brown coal increases linearly with increasing time and reaches almost 100 % at 1 h, whereas it needs approximately 2 h for the sub-bituminous coal to be gasified completely. The temperature dependency of the conversion with this coal reveals that the use of the Na catalyst can lower reaction temperature by about 120 °C, and the Arrhenius plots of the initial specific rate show that apparent activation energies are estimated to be 190 and 120 kJ/mol without and with the catalyst, respectively. The SEM-EPMA and XRD measurements of Na-bearing chars recovered after the pyrolysis and gasification suggest that the Na catalysts are finely dispersed at the initial stage of the reaction but that they may be deactivated by the formation of sodium silicates at high char conversions of more than 90 % even at a low temperature of 700 ºC.
1
INTRODUCTION Catalytic steam gasification of low-grade carbon resources, such as brown coal/lignite, sub-bituminous coal and petcoke, has recently attracted renewed interest to develop a novel SNG (Substituted Natural Gas) production process (Coal to SNG), advanced IGCC (Integrated Gasification Combined Cycle) and polygeneration technologies. In the Exxon Catalytic Coal Gasification process to produce CH4 from Illinois No. 6 coal, K2CO3 was used as a catalyst, and the PDU demonstration run at 900 kg/d was carried out in 1981 with a fluidized bed gasifier at 690 °C and 3.5 MPa [1]. On the other hand, it has been reported that Ni catalyst shows higher performance in the direct production of CH4 under the mild, bench-scale conditions of 600 °C and 1.9 MPa when an aqueous solution of Ni(NH3)6CO3, which can efficiently be recovered and recycled, is impregnated with an Australian brown coal [2], though the Ni is thought to be more expensive than K2CO3. In Japan, the STEP CCT (Strategic Technical Platform for Clean Coal Technology) project has very recently started, and one of the main targets is to develop an advanced IGCC process that can achieve very high generation efficiency of more than 65 % and features the gasification of low rank coals at less than 900 °C by using steam directly from HRSG after a gas turbine unit. One option to realize such the low temperature gasification is to utilize a catalyst. Although it has been shown that nanoscale particles of metallic Ni can exhibit extraordinary catalytic activity in the steam gasification of brown coals even at low temperatures of 450 – 500 °C [3-5], alkali and alkaline-earth metal carbonates, which are more readily available, may be rather suitable as the catalyst raw materials. In the present work, therefore, we focus on the utilization of inexpensive natural soda ash as a catalyst to achieve experimentally the complete steam gasification of low rank coals at less than 900 °C. The soda ash is from Wyoming, U.S.A. and composed mostly of Na2CO3, and the reserves are estimated at 100,000 Mt. Because it has been accepted that ion-exchanged Na catalysts are very effective for the steam gasification of brown coals [6], the ion exchange technique using an aqueous solution of the soda ash is selected and carried out to make clear the dynamic behavior. As well known, one of the major drawbacks to alkali metal carbonate catalysts is that they are deactivated by the reaction with inherent minerals, such as quartz, kaolinite and illite [7]. To minimize this possibility, relatively-low gasification temperatures of less than 750 °C are utilized in the present study.
EXPERIMENTAL Coal Samples An Australian brown coal and an Indonesian sub-bituminous coal, denoted as LY and AD
2
respectively, were used in this work. The as-received samples were air-dried at room temperature, crushed and sieved mainly to the particles with the size fraction of 250 – 500 μm. The proximate and ultimate analyses are summarized in Table 1, which reveals very low ash and sulfur contents of these coals. Table 1. Coal LY AD a b
Proximate and Ultimate Analyses of Coals
Proximate analysis (mass %, dry) Ash FCa,b VMa 1.7 52.3 46.0 1.3 48.8 49.9
C 66.2 71.8
Ultimate analysis (mass %, daf) H N S 5.1 0.38 0.26 5.1 0.52 0.11
Ob 28.1 22.5
VM, volatile matter; FC, fixed carbon. Determined by difference.
Catalyst Material and Ion Exchange Method Natural soda ash from Green River, Wyoming, U.S.A. was used as the precursor material of ion-exchanged Na catalyst. The major size fraction of the as-received sample was in the range of 150 – 650 μm, and it was employed without any pretreatment. The soda ash contains more than 99 mass % Na2CO3, and it is almost Cl-free. In the ion exchange method without any pH-adjusting reagents, an aqueous solution of the soda ash was first added to an aqueous mixture of coal particles and de-ionized water held at a constant temperature of 20 – 40 ºC, and the Na+ concentration and pH of the resulting mixture during stirring were then monitored on line with Na+-selective and pH electrodes, respectively. The initial concentration of Na+ ions was constant at 2500 mg Na/L, unless otherwise stated. The amount of actually-exchanged Na+ ions was calculated by the concentration change. After a predetermined time, the Na-exchanged coal was separated by filtration and water-washed repeatedly, and it was finally dried with a vacuum over held at 60 ºC. The dried sample was supplied to the following gasification experiment. Pyrolysis and Steam Gasification The pyrolysis and subsequent gasification run was carried out with a fixed bed quartz reactor at a constant temperature of mainly 700 ºC under ambient pressure. About 500 mg of the sample was first heated at 300 ºC/min up to 700 ºC in a stream of high purity He, then held for 10 min to remove volatile matter, and finally exposed to 50 vol. % H2O diluted with He to gasify in situ the resulting char. The pyrolysis run alone was also performed in the same manner as above to determine weight loss in this process. Char conversion in the steam gasification was estimated on a dry, ash-fee and catalyst-free basis from the change in the sample weight before and after reaction. Product gas after the removal of H2O was analyzed on line with two micro gas chromatographs (GC) with thermal conductivity detectors attached,
3
and H2, CO, CO2, and CH4 were mainly determined. Characterization The Na content of the dried exchanged coal was determined by the ICP-ES (inductively coupled plasma emission spectroscopy) method after the extraction of Na+ ions from the coal with an aqueous solution of ammonium acetate [8,9]. This determination method proves that the Na content is almost the same as the amount of exchanged Na+ ions estimated by the decreased Na concentration in the ion-exchange process. The TPD (temperature programmed desorption) run of the raw coal was carried out by heating it at 3 ºC/min in a stream of high purity He to analyze on line CO2 evolved by the GC and to determine total CO2. The content of free COOH groups was simply calculated based on the assumption that total CO2 originates from all COOH groups, which are unassociated (free) and associated with naturally-present Na, K, Mg and Ca cations. The N2 adsorption, XRD (X-ray diffraction) and SEM-EPMA (scanning electron microscope-electron probe micro analysis) measurements of selected char samples were also made.
RESULTS AND DISCUSSION Ion Exchange Reactions between Coal and Natural Soda Ash The concentration of Na+ ions in the aqueous mixture of the soda ash and LY or AD coal is plotted against stirring time in Figure 1, where the temperature is held at 20.0 or 40.0 ± 0.1 ºC. With LY coal, the concentration at 20 °C decreased rapidly at the initial stage to be almost constant after 10 min, and the rate of the ion exchange was larger at a higher temperature of 40.0 ºC, at which this reaction was complete within 5 min, but the concentration at a steady state was nearly equal to that at 20 °C. With AD coal, on the other hand, only the slight decrease in the Na+ concentration was observed at 20 °C, whereas it decreased gradually at 40 °C as the time increased. When the pH of the aqueous mixture was monitored, both coals provided the similar time changes as those in Figure 1, and the pH at 40 °C decreased from the initial 11.1 to 7.9 and 10.0 at 30 min with LY and AD, respectively. Because COOH groups can work as ion-exchangeable sites for Na+ ions at the final pH of approximately 9 [6], ion exchange reactions under the present conditions may take place predominantly according to the following equation: Na+ + -COOH → H+ + -COONa
(1)
Because the contents of free COOH groups determined by the TPD method were 2.0 and 1.3 meq./g-coal for LY and AD, respectively, the ratios of exchanged Na+/free COOH groups
4
3000
AD/40 ºC 2000
LY/20 ºC LY/40 ºC
1000
+
Na concentration, mg/L
AD/20 ºC
0
0
10 20 Time, min
30
Figure 1. Changes in Na+ concentration with stirring time in the process of ion exchange between LY or AD coal and natural soda ash at a constant temperature of 20 or 40 ˚C. Table 2.
Effect of Particle Size on the Proportion of Na+ Exchanged
Coal
LY LY LY AD AD AD a
Ion-exchange conditions Particle size Temperature (μm) (°C) 250 - 500 30.0 250 - 500 40.0 75 - 150 30.0 250 - 500 30.0 250 - 500 40.0 75 - 150 30.0
+
a
Na /COOH
0.62 0.62 0.64 0.33 0.42 0.63
+
Ratio of Na ions exchanged to free COOH groups.
could be estimated at 0.62 and 0.42 at 40 °C with LY and AD, respectively. Therefore, not only the rate of the ion exchange but also the amount of Na+ ions exchanged was higher with the former coal, as can be expected. It has been reported that ion exchange reactions between coal and alkali metal cations at ambient temperature are controlled by their diffusion within coal pores [10]. The size of coal particles was thus changed from the initial 250 – 500 μm to 75 – 150 μm. The results are summarized in Table 2. As expectable, the decreasing rate of the Na+ concentration at 30 °C was larger at a smaller size, irrespective of the coal type, and the effect was more considerable with AD. The Na+/COOH ratio at 30 °C was almost unchanged with the particle size for LY, whereas it was higher at 75 – 150 μm for AD and nearly equal to the value (0.64) for LY. This
5
agreement is interesting and suggests that the proportion of surface COOH groups accessible by Na+ ions is almost the same between the two coals. Steam Gasification of Na-Exchanged Coals
Figure 2 shows the catalytic effect of the exchanged Na on char conversion in the steam gasification of LY or AD at 700 ˚C. The reactivity of the raw coal without the catalyst was low, and the conversion at 2 h was about 20 % in both chars. On the other hand, the Na catalyst remarkably promoted the gasification of these chars. The reaction rate with LY char increased almost linearly with increasing time, as featured in the alkali metal-catalyzed gasification of coal chars, and this sample could be gasified almost completely within 1 h. In spite of the fact that Na loading was lower on AD char, the conversion at 0.5 h was close to that with LY char, meaning that the initial catalytic activity per g-Na was higher with AD char. As seen in Figure 2, however, the rate enhancement beyond 0.5 h was smaller with this sample, and it thus took about 2 h for the complete gasification. The different rate profiles observed with the two chars suggest that the Na catalyst on AD char may be deactivated more significantly at the latter stage of the gasification.
Char conversion, mass % (dacf)
100 2.9 % Na/LY
80 60
1.6 % Na/AD
40 Raw LY
20 Raw AD
0
0
1 2 Reaction time, h Figure 2. Changes in char conversion with reaction time in the steam gasification of the raw and Na+-exchanged coals at 700 ˚C.
When the gasification of AD coal was carried out in the range of 650 – 800 °C, char conversion at 1 h was 60 % at 800 °C for the raw sample, whereas it was almost 100 % at 750 °C for the Na-exchanged coal. The comparison of the temperature dependency between the two samples showed that the use of the catalyst could lower the gasification temperature by 120 °C. In addition, apparent activation energies without and with the Na were estimated at 190 and 120 kJ/mol from the Arrhenius plots using initial specific rates.
6
Yield of product gas, which was estimated based on the unit weight of char, increased in the order of CO < CO2 < H2, regardless of the coal type and the temperature, showing that the shift reaction of CO formed with H2O took place to a large extent. When all yields without and with the Na were plotted against char conversion, an almost linear relationship was observed with each of these products, which means that the catalyst has no significant effect on the product composition under the present H2O-excess and differential conditions. Catalyst States after Pyrolysis and Gasification
The SEM-EPMA measurements after the pyrolysis of Na-exchanged LY and AD coals at 700 ºC reveal the uniform distribution of Na species, independently of the coal type. The XRD profiles after the pyrolysis and gasification at 700 °C are shown in Figure 3, where char conversion data are also given. With the pyrolyzed samples, denoted as “0 conv.” in this figure, the XRD peaks of Na2CO3 could be detected with LY char, whereas no diffraction lines of Na species were detectable for AD char due probably to the low Na content and/or the high dispersion. When these pyrolyzed chars were gasified up to 50 – 55 % conversion, the Na2CO3 peaks were stronger with LY char, but contrarily the XRD line attributable to this species was very weak with AD char. The Na2CO3 content of this sample was estimated at 13 mass %, which was nearly equal to that of the pyrolyzed LY char, for which the XRD intensities of Na2CO3 were much larger as seen in Figure 3. These observations strongly suggest that the Na catalyst on the AD sample is more finely dispersed, which may account for the larger activity per g-Na observed at 0.5 h in Figure 2. By the analogy with reaction mechanisms proposed for the K2CO3-catalyzed gasification of coals and carbons [1,11-14], Na2CO3 may be transformed to sodium-carbon complexes, denoted as Na-char, as catalytically-active forms in the following manner: Na2CO3 + char → Na-char + CO2
(2)
The Na-char may be expressed chemically as -C-Na and -C-O-Na [11-15]. In the XRD measurements of LY and AD samples after the almost complete gasification at 700 °C observed in Figure 2, the diffraction peaks of Na compounds other than Na2CO3 newly appeared in both cases. These signals might be identified as Na2SiO3, which may be formed by the reaction of Na2CO3 with quartz (SiO2) in the latter stage of the gasification according to the following equation: Na2CO3 + SiO2 → Na2SiO3 + CO2
(3)
According to thermodynamic calculations, the Gibbs free energy changes of this reaction are < 0 at temperatures of > 400 °C, and it is more favorable at a higher temperature. It has been reported that the model reaction of K2CO3 and quartz takes place at more than 750 °C [7].
7
When the XRD intensity ratio of the new peak/Na2CO3 was compared between the gasified LY and AD chars, it was larger with the latter sample, suggesting that the Na catalyst may be deactivated more readily according to eq. 3. The decreased rate enhancement observed after 0.5 h in Figure 2 may be caused by the silicate formation. Although Na+ ions of Na2SiO3, if any, can readily be recovered and recycled because of the large solubility in water and the alkaline property of the aqueous solution, the Na-catalyzed gasification at lower temperatures of < 700 °C may be recommended to avoid the above-mentioned catalyst deactivation. B
Na2CO3
B BB B
LY char 55 % conv.
B B
BB B
B
B
Diffraction intensity
BB
B BB
LY char 0 % conv.
B B
B
AD char 50 % conv.
AD char 0 % conv.
10
20
30
40
50
60
70
80
2θ (Cu Kα), degree Figure 3. XRD profiles after the pyrolysis and gasification of Na-exchanged coals at 700 ˚C.
CONCLUSIONS
The results described here demonstrate that Na cations of natural soda ash, which is composed of > 99 % Na2CO3, can successfully be ion-exchanged predominantly with the protons of COOH groups in low rank coals in the absence of any pH-adjusting agents, and that the exchanged Na shows the high catalytic performance at low loading of 2 – 3 mass % in
8
the fixed bed steam gasification at 700 °C in particular for oxygen-rich brown coal.
ACKNOWLEGEMENT
The present work was supported in part by the Japan Coal Energy Center (JCOAL) commissioned by the Strategic Technical Platform for Clean Coal Technology (STEP CCT) of the New Energy and Industrial Development Organization (NEDO) with the subsidy of the Ministry of Economy, Trade and Industry of Japan. One of the present authors, Dr. Enkhsaruul Byambajav, is grateful for financial support from the Korea Gas Corporation and R&D Program of National University of Mongolia.
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]
Hirsch, R.L.; Gallagher, J.E., Jr.; Lessard, R.R.; Wesselhoft, R.D. Science 1982, 215, 121. Takarada, T.; Sasaki, J.; Ohtsuka, Y.; Tamai, Y.; Tomita, A. Ind. Eng. Chem. Res. 1987, 26, 627. Tomita, A; Tamai, Y. Fuel 1981, 60, 992. Tomita, A; Ohtsuka, Y.; Tamai, Y. Fuel 1983, 62, 150. Ohtsuka, Y.; Tomita, A.; Tamai, Y. Appl. Catal. 1986, 28, 105. Takarada, T.; Nabatame, T.; Ohtsuka, Y.; Tomita, A. Ind. Eng. Chem. Res. 1989, 28, 505. McKee, D.W.; Spiro, C.L.; Kosky, P.G.; Lamby, E.J. Chemtech 1983, 624. Morgan, M.E.; Jenkins, R.G.; Walker, P.L., Jr. Fuel 1981, 60, 189. Takarada, T.; Tamai, Y.; Tomita, A. Fuel 1985, 64, 1438. Bobman, M.H.; Golden, T.C.; Jenkins, R.G.; Solvent Extraction and Ion Exchange 1983, 1, 791. Freriks, I.L.C.; van Xechem, H.M.H.; Stuiver, J.C.M.; Bouwman, R. Fuel 1981, 60, 463. Mims, C.A.; Rose, K.D.; Melchior, M.T.; Pabst, J.K. J. Am. Chem. Soc. 1982, 104, 6886. Mims, C.A.; Pabst, J.K. Fuel 1983, 62, 176. Yuh, S.J.; Wolf, E.E. Fuel 1983, 62, 252. Yuh, S.J.; Wolf, E.E. Fuel 1984, 63, 1604.
9
Separation of Pyrolysis from Fluidized Bed Steam Gasification: Its Conception and Application Masahiro NARUKAWA*, Makoto TAKAFUJI, and Toshiyuki SUDA
Research Laboratory, IHI Corporation, Ltd., Isogo, Yokohama 235-8501, Japan * E-mail: [email protected] Abstract The pyrolysis separation dual fluidized bed gasification (P-DFBG) place a bubbling fluidized bed pyrolyzer above a bubbling fluidized bed gasifier involved in the normal dual fluidized bed gasification (N-DFBG) systems. The heat carrier particles circulated from the char combustor enter first the pyrolyzer to facilitate the pyrolysis reactions of fuel occurring therein, and the particles are in turn forwarded into the gasifier to provide endothermic heat for the steam gasification reactions of chars. By feeding fuels into the upper pyrolyzer, the pyrolyzer can separate pyrolysis gases from fuel chars so that the lower gasifier gasifies the resultant chars produced in the pyrolyzer. Therefore, the steam gasification reactions in the gasifier proceed without pyrolysis gases, which inhibit the steam gasification reactions of chars. Consequently, by using P-DFBG it is hopeful to increase gasification efficiency. This anticipation was verified through gasifying subbituminous coal in two 2.0 kg/h experimental setups configured according to the principles of P-DFBG and N-DFBG, respectively. Increases in carbon conversion and cold gas efficiency of P-DFBG compared with N-DFBG were about 3%, respectively. Keywords: Steam gasification, Pyrolysis separation, Dual fluidized bed gasification, Pyrolyzer, Subbituminous coal
1. Introduction Dual fluidized bed gasification (DFBG) with steam gasification has the advantage of producing middle-caloric gas free of serious N2 dilution, because it isolates combustion
reaction for endothermic heat required by pyrolysis/gasification reactions [1]. A combination of a bubbling fluidized bed gasifier and a pneumatic riser combustor is the superior technique for the DFBG. A bubbling fluidized bed gasifier in the DFBG with steam gasification has the advantage of producing gasification gases rich of H2 [2]. The steam gasification of chars occurring in the gasifier can be inhibited by pyrolysis gases such as H2, hydrocarbons, and so on [3]. Then, gasification efficiency is anticipated to improve by the separation of the pyrolysis gases from the steam gasification of chars. We recently developed the pyrolysis separation DFBG (P-DFBG) for fuel steam gasification. The P-DFBG place a bubbling fluidized bed pyrolyzer above the bubbling fluidized bed gasifier involved in the normal DFBG (N-DFBG) system. The pyrolyzer is fluidized by N2, while the gasifier is fluidized by steam. The circulated heat carrier particles (HCPs) enter first the pyrolyzer to provide endothermic heat for pyrolysis reactions of fuels and in turn move into the lower gasifier. The downcomers of the overall system are immersed into the particle beds of the bubbling fluidized beds. Fuels are fed into the freeboard of the pyrolyzer and then pyrolyzed by contacting the HCPs from the combustor. The pyrolyzer can separate pyrolysis gases from chars so that the lower gasifier works to gasify only the resultant chars produced in the upper pyrolyzer. Therefore, steam gasification reactions in the gasifier proceeds without pyrolysis gases, which inhibit steam gasification reaction of fuel chars. To demonstrate the superiority of P-DFBG and thereby to show its technical feasibility, in the present study experiments were conducted by feeding subbituminous coals (ca. 2.0 kg/h) into P-DFBG and N-DFBG systems, respectively.
2. Experiment The experimental apparatus and detection method about N-DFBG are available in the previous studies [2, 4, 5]. Modifying the N-DFBG setup according to the conception of pyrolysis separation results in the P-DFBG setup shown in Fig. 1. The setup consists of a pneumatic riser combustor, a fluidized bed pyrolyzer, a fluidized bed gasifier, and a loop-seal. The combustor has dimensions of 53.5 mm i.d. and 6200 mm height. The pyrolyzer is a rectangular bed with a total height of 1250 mm. Cross-sectional area and height of the pyrolyzer are 190 × 80 mm2 and 500 mm for its lower part and 190 × 180 mm2 and 550 mm for its upper part, respectively. The gasifier is also a rectangular bed with a total height of 1800 mm. The cross-sectional area and height of the
Combustion gas
Combustor Φ 53.5 × 6200 mm
Heat exchanger
Bag filter Gas analyzer
Gs measure Coal
Pyrolyzer 190 × 80 × 1250mm Pyrolysis and Gasification gases N2
Tar and Water trap Filter
IDF
Exhaust
FDF
Combustion tube
Heat exchanger
Bag filter Gas analyzer
Gas chromatograph
IDF
Exhaust
Gasifier 370 × 80 × 1800 mm Steam FDF
N2
Loop-seal 140 × 80 × 400 mm
Figure 1. Schematic diagram of the pyrolysis separation dual fluidized bed gasifier
(P-DFBG).
gasifier are 370 × 80 mm2 and 900 mm for its lower part and 370 × 180 mm2 and 700 mm for its upper part, respectively. The loop-seal is also a rectangular bed (140 × 80 × 400 mm). All ducts conveying HCPs have the same i.d. of 53.5 mm. The downcomers immerse into the particle beds of the pyrolyzer, gasifier and loop-seal, respectively. The adopted distributors for these beds are of multi-orifice nozzle type; the orifice diameter is 1.1 mm. Combustor and gasifier have independent cyclones, heat exchangers, bag filters, and induced draft fans (IDFs), thus enabling independent pressure control in both the reactor vessels. The pyrolysis gases generated in the pyrolyzer are mixed with the gasification gases in the freeboard of the gasifier, and then the mixed product gases are released from the gasifier. In the preset study, we discuss the amount of this product gases, namely overall gasification efficiency for the present P-DFBG. The IDF of the gasifier enables control of pressure in the pyrolyzer as well. The combustor, pyrolyzer, and gasifier were electrically heated and temperatures of up to about 1173 K are allowed. The mixed product gases are first burnt off in a combustion tube and in turn sent to the heat exchanger mounted in the exhausted line.
Table 1. Properties of coal used in the present study Proximate analysis [wt. % - as received] Moisture 21.1 Volatile matter
38.4
Fixed carbon
39.4
Ash Ultimate analysis [wt. % - dry basis] C H O N S HHV [J/g - as received] Size [mm]
1.1
69.4 4.9 23.8 0.4 0.1 22100 0.5 - 5.6
The involved experimental procedure and detection method are similar to those taken in the experiments with N-DFBG [2, 4, 5] except for the points mentioned here. By using a table feeder, fuels are carried with argon gas (tracer gas) into the freeboard of the pyrolyzer for P-DFBG, and into that of the gasifier for N-DFBG, respectively. In the present study, N2 was supplied into the pyrolyzer at about 873 K. Temperature of the steam supplied into the gasifier was at about 773 K. Nitrogen gas is also supplied into the loop-seal for the seal between the gasifier and combustor. For P-DFBG, only the mixed product gases released from the gasifier was sampled to measure its compositions and concentrations. The parameters used here are defined as:
] × 100
(1)
∑ [HHV × (Gas production rate)] × 100
(2)
Xc =
η=
∑ [M
cg
× (Gas production rate )
M cf × (Fuel feed rate) g
HHVf × (Fuel feed rate)
where Gas production rate =
Cg
C Ar
× (Volume flow rate of Ar)
(3)
In Eq. (1), Mcg indicates the moles of carbon per normal cubic meter of each constituent in product gas and Mcf indicates that per kilogram of coal feed. In Eq. (3), Cg indicates concentrations of each constituent in rude product gases and CAr indicates concentrations of
Table 2. Experimental conditions in the present study Coal Steam/Carbon Tp Tg Tc [kg/h] [mol/mol] [K] [K] [K] P-DFBG
1.9
1.5
1066
1085
1162
N-DFBG
2.1
1.4
1065
1083
1137
Up [m/s]
Ug [m/s]
4.2
Uc [m/s]
0.092 0.036
τp [s]
τg [s]
40
310
4.1
0.092 0.036
38
310
Ar in rude product gases. Normal condition denotes 273.15 K at 1 atm. The used fuel is subbituminous coal. Summarized in Table 1 are the major properties of the coal. The sizes of the used coal are 0.5 – 5.6 mm. Silica sand of 190 µm Sauter mean diameter is used as HCPs. Table 2 summarizes the experimental conditions in the present study. In both experiments the coal feeding rates were about 2.0 kg/h with similar reaction temperatures in combustor (about 1150 K) and gasifier (about 1084 K), respectively. The superficial gas velocity Ug in the gasifier and Up in the pyrolyzer were 0.092 m/s and 0.036 m/s at the operation bed temperatures Tg and Tp, respectively. The nominal residence time τp of particles (including coal) inside the pyrolyzer is defined as: τp =
Particle amount in the pyrolyzer Particle amount circulated in unit time
(4)
τg is defined by the similar equation for gasifier as well. The particle amount in the present system is about 30 kg. The particle amount circulated in unit time is measured experimentally at actual running temperature by using the method of Xu et al. [6]. 3. Results and Discussion Fig. 2 shows the time series of measured product gas concentrations (Fig. 2a) and its corresponding performance (Fig. 2b) for P-DFBG. The concentrations in Fig. 2a are normalized to the mentioned gas species, so that the summations of the concentrations shown in Fig. 2a are constantly 100 % for the all measurements. The time of coal feeding lasted for about 60 min; however, in the first 30 min no measurement of product gases was conducted, because the composition reached stability after about 30 min from the start of fuel feed. Fig. 2 shows that the experiment’s quasi-steady state could be surely maintained for the period of about 20 minutes measured here. The product gas constituent with the highest concentration was H2 (53 %), followed by CO2 (22%), CO (17 %), CH4 (6 %) and other hydrocarbons (1 %). The corresponding carbon conversion Xc and cold gas efficiency η remained consequently constant, namely about 48 % and 56 %, respectively. Fig. 3 compares the carbon conversion Xc, and cold gas efficiency η for the
60 (a) H2 CO2 CO CH4 Other HCs
Concentration [%]
50 40 30 20 10 0 0
5
10
15
20
25
Time [min]
Conversion and efficiency [%]
60 (b) 50 40 30 20
Xc η
10 0 0
5
10 15 Time [min]
20
25
Figure 2. Time series of (a) concentrations of product gases and (b) the corresponding carbon conversion and cold gas efficiency. The plotted data are the result of P-DFBG.
experiments conducted according to N-DFBG (open bars) and P-DFBG (solid bars) system. The values in Fig. 3 are the averages over the period in which the experiment was quasi-steady. Under the experimental conditions shown in Table 2, carbon conversion and cold gas efficiency of P-DFBG were 48 % and 56 %, respectively, whereas those of N-DFBG were 45 % and 53 %, respectively. Normalized concentrations of product gases for N-DFBG without N2 (fluidizing gas in pyrolyzer) and Ar (tracer) gases were comparable with P-DFBG (H2: 52 %, CO2: 23 %, CO: 17 %, CH4: 7 %, and other HCs: 2 %). Increases in carbon conversion and cold gas efficiency of P-DFBG compared with N-DFBG were about 3%,
80 Conversion and efficiency [%]
N-DFBG P-DFBG 60
40
20
0 Xc
η
Figure 3. Comparison of carbon conversion and cold gas efficiency for N-DFBG and P-DFBG.
respectively. This result suggests that the separation of pyrolysis gases from the steam gasification reactions decreased the concentration of inhibition gases in the gasifier and thus enhanced the steam gasification of chars. The carbon conversion efficiency as obtained by numerical simulation was higher than the results of P-DFBG experiment conducted in the present study, suggesting that gasification gases as well pyrolysis gases partly inhibit the steam gasification reactions in gasifier. Conclusions By using P-DFBG and N-DFBG, the separation of pyrolysis gases from the steam gasification of fuel chars was demonstrated to enhance gasification efficiency. Nomenclature CAr [%] Cg [%] HHVf [mol/kg] HHVg [J/Nm3] Mcf [mol/kg] Mcg [mol/Nm3] Tc [K] Tg [K]
: concentration of Ar in rude product gas : concentration of each constituent in rude product gas : higher heating value of 1 kg of coal : higher heating value of 1 Nm3 of each constituent in product gas : moles of C in 1 kg of coal : moles of C in normal 1 Nm3 of each constituent in product gas : characteristic temperature of riser combustor : characteristic temperature of fuel gasifier
Tp [K] Uc [m/s] Ug [m/s] Up [m/s] Xc [%] η [%] τg [s] τp [s]
: characteristic temperature of fuel pyrolyzer : superficial gas velocity in riser combustor at its operation temperature : superficial gas velocity in fuel gasifier at its operation temperature : superficial gas velocity in fuel pyrolyzer at its operation temperature : conversion of coal C into product gas (Eq. 1) : cold gas efficiency (Eq. 2) : nominal residence time of particles in fuel gasifier : nominal residence time of particles in fuel pyrolyzer
Acknowledgements This work was supported by New Energy and Industrial Technology Development Organization (NEDO), Japan. The authors gratefully acknowledge Mr. Yukio ODA for his help in the experiments. References [1] A. V. Bridgwater, The technical and economic feasibility of biomass gasification for power generation, Fuel 74 (1995) 631-653. [2] G. Xu, T. Murakami, T. Suda, Y. Matsuzawa, H. Tani, The superior technical choice for dual fluidized bed gasification, Ind. Eng. Chem. Res. 45 (2006) 2281-2286. [3] B. Bayarsaikhan, N. Sonoyama, S. Hosokai, T. Shimada, J. –i. Hayashi, C. –Z. Li, T. Chiba, Inhibition of steam gasification of char by volatiles in a fluidized bed under continuous feeding of a brown coal, Fuel 85 (2006) 340-349. [4] G. Xu, T. Murakami, T. Suda, Y. Matsuzawa, H. Tani, Gasification of coffee grounds in dual fluidized bed: performance evaluation and parametric investigation, Energy Fuels 20 (2006) 2695-2704. [5] G. Xu, T. Murakami, T. Suda, Y. Matsuzawa, H. Tani, Two-stage dual fluidized bed gasification: its conception and application to biomass, Fuel Process. Tech. 90 (2009) 137-144. [6] G. Xu, T. Murakami, T. Suda, Y. Matsuzawa, H. Tani, Particle circulation rate in high-temperature CFB: measurement and temperature influence, AIChE J. 52 (2006) 3626-3630.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: GASIFICATION EXACT TITLE OF PAPER: DYNAMIC SIMULATION OF THE GASIFICATION PROCESS – HOW TO SHORTEN COMMISSIONING TIMES, BUILD A OPERATORS TRAINING CENTER AND OPTIMIZE THE DESIGN Presenting Author: Friedemann Mehlhose, Dipl. Ing, Siemens AG Contact Information: Simens Fuel Gasificaiton Technology GmbH & Co.KG, Halsbrücker Straße 34, 09599 Freiberg, Germany Co-Author: Julia Kittel, Dipl.-Mat. Siemens AG Contact Information: Simens Fuel Gasificaiton Technology GmbH & Co.KG, Halsbrücker Straße 34, 09599 Freiberg, Germany
Abstract: The plant automation together with the design of the emergency shutdown system is one of the most important topics in the project execution of gasification plants. The Siemens Fuel Gasification Technology (SFGT) 500MW class gasifier is an improved design based on about 20 years operation experience with a 200MW gasification plant and the 5MW pilot plant at SFGT headquarter in Freiberg, Germany. The automation of the first SFGT 500MW gasification plant was developed with the Siemens Power Plant Automation System (SPPA) T3000 together with a dynamic simulation of the gasification island in “Dymola” and “SIMIT”. The knowledge and experiences from the pilot plant could be used perfectly as basis for the processes simulation and later on for the validation of simulation results. These validated simulation results have enabled SFGT to predict the behavior of the large scale gasification system properly and with only few uncertainties. The paper presents two different concepts of process simulation its goals and application. It shows how simulation can support the engineering and design as well as the automation of customer projects e.g. our 500MW gasification plant in the PR China. Dynamic testing of the control and emergency shutdown logics saves time and cost during plant commissioning. Customer training with a validated dynamic simulation and the same automation system used in real plants guarantee the best learning conditions. In the Research & Development program of SFGT dynamic process simulation helps very efficiently to make judgments on advanced design philosophies or cost saving measures. A real time data transfer from site to SFGT headquarter will support commissioning and operation as well as the validation of the dynamic model with a commercial size gasification plant. All this ideas were driven by the aim of cost reduction and an improved operation philosophy together with the highest demands of plant safety.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: GASIFICATION A DYNAMIC SIMULATOR OF A COMMERCIAL-SCALE IGCC PLANT Mi-Yeong Kim, Mrs., Korean Electrical Power Corporation [email protected] Yong-Jin Joo, Dr., Korean Electrical Power Corporation [email protected] In-Kyu Choi, Mr., Korean Electrical Power Corporation [email protected] Joong-Won Lee, Mr., Korean Electrical Power Corporation [email protected] Si-Moon Kim, Mr., Korean Electrical Power Corporation [email protected] Min-Churl Lee, Mr., Korean Electrical Power Corporation [email protected]
Abstract: In this study, the simulator for a commercial-scale plant of IGCC (Integrated Gasification Combined Cycle) is developed for dynamic tests and an education of plant operation. The simulator consists of a dynamic process model and a HMI (Human-Machine Interface). The dynamic process model of plant includes ASU, Gasifer, Gas treatment units, Combined Cycle using dynamic process simulation tool. Specially, a model of gasifier is analyzed in detail, flow which has gasification reaction in the gasifier is divided by 6 zones, and heat transfer and mass balance of each zone are calculated. The dynamic process model acts as a field and DCS (Distributed Control System) of plant. DCS in the simulator has master-control of plant like turbine lead mode, gasifier lead mode, coordination mode. The master-control means a plant selects to follow and control output of gas turbine or gasifier first as changing total power command and a plant will be operated under three kinds of master-control as following plant condition. The HMI of simulator works connection between controller of the dynamic model and operator, operator can control a whole plant. Operator can do start-up and shut-down of plant easily using HMI. In addition, Operator can monitor conditions of plant though graphs of real-time based data and history data of important process variables in the HMI. Developed simulator is possible to perform engineering studies of IGCC plant like dynamic test for process variables and changing feed stock and analysis for control logic and so on.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Manuscript PROGRAM TOPIC: GASIFICATION EXACT TITLE OF PAPER: ENTRAINED FLOW SLAGGING SLURRY GASIFICATION AND THE DEVELOPMENT OF COMPUTATIONAL FLUID DYNAMICS MODELS AT CanmetENERGY
Robin Hughes, Research Engineer, CanmetENERGY 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 996-0066 Dennis Lu, Research Scientist, CanmetENERGY 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 996-2760 Adrian Majeski, Research Engineer, CanmetENERGY 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 996-3735 Andrew Corber, Research Council Officer, National Research Council 1200 Montreal Rd., Ottawa, Ontario, CANADA, [email protected], (613)991-0642 Ben Anthony, FBC and Gasification Group Leader, CanmetENERGY 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 996-2868
The development of computational fluid dynamics (CFD) models representing entrained flow slagging slurry gasification has proven difficult due to limited information being available in the open literature regarding gasifier geometry, burner spray characterization, and local gas and char conditions. This paper describes efforts made by Canadian national laboratories CanmetENERGY and the National Research Council, to provide the data required for developing and validating gasification CFD models. The CanmetENERGY two tonne per day (slurry feed rate) pilot-scale gasifier has been modified to allow local gas and char conditions to be sampled from within the gasifier and from the syngas exiting the quench vessel during gasifier operation. A series of Canadian and U.S. solid feedstocks have been gasified and a subset of the results of these gasification tests is presented here. The National Research Council’s Institute for Aerospace Research spray characterization laboratory is determining droplet size and velocity characteristics for the CanmetENERGY slurry gasifier burner spray at elevated pressure and temperature. Gasifier geometry, burner spray characterization, local gas and char conditions, fuel characteristics, and slag viscosity measurements have been used in the development of CanmetENERGY CFD models representing the system. The data is being forwarded to our industrial, academic, and government research partners in Canada and the U.S.
I N T RO D U C T I O N
There is a general feeling that entrained flow slagging gasification reactor computational fluid dynamics (CFD) models require validation data to ensure that the models accurately represent the effects of rate of reaction, mass and heat transfer on system operation. It is very difficult to obtain the required reactor dimensions, fuel characteristics, and product properties from commercially operating gasification equipment, so CanmetENERGY has initiated a series of pilot scale gasification tests to provide the necessary data. This paper is the first in a series that will provide increasingly detailed results for the express purpose of validating CFD models. This paper provides dimensions of the current CanmetENERGY slurry injector and reaction chamber. Fuel characterization includes ultimate & proximate analysis, ash fusion temperatures, major oxides, surface area, pore size distribution, reactivity data via thermogravimetric analysis, slurry viscosity, coal particle size distribution, and particle density. In the gasification test reported here the coal was introduced into the gasifier as coal-water slurry. The atomization of coal-water slurries is quite difficult to predict by CFD, so a number of spray characterization methods have been employed to provide information regarding spray angles, droplet size, and droplet shape. The spray characterization was performed at atmospheric pressure here in part to provide the data for this publication, and in part to prepare for future testing to be performed at the operating pressure of the CanmetENERGY gasifier. E X P E R I M E N TA L M E T H O D
CHARACTERIZATION TECHNIQUES
STANDARD METHODS Where possible, standard test methods have been used to characterize both the coal for the gasification tests and the solids products from the tests. Table 1 provides a list of standard methods that have been used for the characterization of the fuel. Table 1: Standard Test Methods Used in Characterizing Solids Analysis Proximate Moisture TGA, Secondary Moisture, Ash TGA Volatile Fixed Carbon
Standard Test Method ASTM D5142 ISO 562 ASTM 3172
Ultimate C, H, N Total S
ASTM D5373 ASTM D5373
Calorific Analysis
ISO 1928
Ash Fusibility Major Oxides (X-Ray Fluoresence)
ASTM D1857 ASTM D4326
SURFACE PROPERTIES AND DENSITY
A Micrometrics TriStar II 3020 was used to determine the surface area, pore volume, and pore area of the fuels. The surface area was determined using the BET (Brunauer, Emmett, Teller) methodology. The BET method proceeds by the measurement device increasing the N2 gas pressure and allowing nitrogen gas to adsorb
onto the surface of the pores. The gas is assumed to form a continuous monolayer, which completely covers the solids’ surface. The number of molecules of gas that form the monolayer is then calculated and multiplied by the cross sectional area of one nitrogen molecule to get the total surface area of the coal particle. The nitrogen cannot reach micropores with diameters smaller than 5Å and therefore very small pores will not be accounted for using N2 porosymmetry. The pore volume and pore area were analyzed by the BJH (Barret, Joyner, Halenda) methodology. The TriStar II is capable of measuring the surface area, the pore volume, and the pore area of pores between 17 and 3000Å. Before analysis, 1-2g samples were degassed at 150°C under vacuum for 16 hours. All samples were degassed using a Micrometrics VacPrep 061. Sample vials were backfilled with nitrogen to avoid contamination while being loaded into the analyzer. Three samples of each coal were analyzed. The results provided here are averaged values of surface properties. Figure 5 provides an example of N2 adsorption and desorption isotherm data. The skeletal density was analyzed using helium pycnometry with a Micrometrics, AccuPyc II 1340. Each sample was loaded into a pre-weighed sample cup up to ¾ full. The sample mass was recorded. The sample was loaded into the gas pycnometer and the skeletal density was recorded after 10 minutes. Two separate coal samples were analyzed. The gas pycnometer calculates three skeletal densities for each analysis, therefore a total of six skeletal densities were used to find the average value. SLURRY VISCOSITY
Slurry viscosity was tested using a Fann Model 35A/SR6 Viscometer with Fann rotor #1, Fann bob #2 and Fann spring #1. Slurries were prepared in advance of gasification tests to determine viscosity versus shear rate at 45 wt% and 50 wt% solids. Viscosity was determined at shear rates of 37.7 s-1, 75.4 s-1, 113.1 s-1 and 226.2 s-1. Sample data are plotted in Figure 6. REACTIVITY
Samples were crushed using a mortar and pestle and sieved to ≤ 250µm. The samples were dried in accordance with ASTM method D 3173 - 87. Reactivity of the coal was tested using a Cahn 1000 thermogravimetric analyzer (TGA). Samples of between 30-50 mg were brought up to 400°C in a 40mL/min flow of nitrogen. The samples were held at these conditions for 5 minutes. Then, the gas was switched to a mixture of air (40mL/min) and nitrogen (30mL/min) and the temperature was ramped up from 400 to 900°C. The temperature was held at 900°C for 5 minutes to allow the final weight reading to stabilize. The weight of the sample was logged every 4 seconds for the duration of the run. The testing was performed at atmospheric pressure. An example of sample weight versus temperature data is plotted in Figure 7. These data were converted to a rate and plotted in Figure 8, to determine reactivity. PARTICLE SIZE DISTRIBUTION
Sieve analysis was used to characterize particle was used to determine the particle size distribution. The sieve analysis was performed using U.S. Standard sieves sizes 60, 100, 140, 200, 325 and 400 mesh. A maximum of 10 g of coal was sieved through the six screens for 20 minutes with constant vibration and a pulsing action every 5 seconds using a sonic sifter. The mass that was collected by each sieve was weighed and recorded. These masses were then converted into the percentage of particles that are smaller than a given size. GASIFICATION PILOT PLANT
The CanmetENERGY pilot scale gasifier (Figure 2) was operated in oxy-fired slurry feed mode at an operating pressure of ~1500 kPa(g) with a bituminous coal. Slurry was injected into the gasifier through a gasswirl atomizer with impinging plate and pintle to provide a hollow cone spray (Figure 1). The atomized slurry was injected downwards on the central axis of the refractory lined vessel. The gasification chamber is a flat-topped, open-bottom cylinder 254 mm ID with total length 2030 mm. The injector face was flush with the top of the
reaction chamber. There is no reduction in the reaction chamber diameter at the slag tap as is common in commercial down-fired gasifiers. The ‘Reactor Gas Sample’ probe (Figure 2) was located 1660 mm from the top head of the reaction chamber and withdrew the sample at a point 40 mm from the centerline of the reactor. There are no purge streams entering the gasifier. The sample probe is nitrogen cooled and is fitted with a shroud to avoid excessive slag flow into the sample probe which results in probe plugging. The heat loss from the gasifier has been estimated to be 17 kW at a similar pressure, temperature and gas composition when firing a pure liquid fuel. The reactant parameters are identified in Table 2. The reactant parameters were slightly different during the period when sampling from the quench vessel outlet (identified as ‘Quench Line’) versus the reactor sample probe, and so average parameters are given for both periods. All instrumentation calibration is NIST traceable. The syngas and slag exit the bottom of the gasifier and enter the quench vessel where they enter a metal walled down comer 330 mm ID. Within the down comer the slag and syngas are sprayed with water by four flat fan sprays at an of 15° below the horizontal plane. The sprays have an included angle of ~65 ° with a mean diameter estimated to be approximately 40 microns. The sprays are flush with the inner wall of the down comer and are located 300 mm below the gasification chamber outlet. The operating parameters of the quench vessel are given in Table 3.
Figure 1: Gas-swirl atomizer used for atomizing slurry with oxygen. Injector cooling assembly not shown. Drawing modified from Delavan Swirl-Air Nozzles brochure. Nozzle assembly 32740-13-
Table 2: Gasifier reactor parameters
Reactants Coal Slurry Flow Temperature Pressure Oxygen Flow Temperature Pressure
kg/hr ºC kPa (g)
kg/hr ºC kPa (g)
Quench Line Sampling Period
Reactor Sampling Period
78.9 30 1782
82.0 30 1709
40.0 21.2 1782
40.5 21.8 1709
Table 3: Quench parameters
Quench Vessel Operation Temperature Out Pressure Quench Water Flow Temperature Quench Jacket Out
Quench Line Sampling Period
Reactor Sampling Period
ºC kPa(g)
131 1527
131 1461
kg/hr ºC ºC
378 10 99
378 10 97
Note that the quench vessel has a water cooled jacket that will contribute to quench vessel heat loss - assume wall temperature of quench vessel same as temperature of water leaving quench jacket
Figure 2: CanmetENERGY gasifier and quench vessel indicating locations of reactant in and out flows and sample extraction. Thermocouple locations are indicated by “TE”. SLURRY INJECTOR SPRAY CHARACTERIZATION
Slurry spray characterization was performed using a different coal than was used in the gasification tests, however, the particle size distribution of the coals is quite similar. The spray characterization was performed in an open atmosphere bucket test with the spray tip injecting downwards approximately 1000 mm above a 45 gallon drum. The spray testing was performed to determine the performance of the analytical equipment prior to constructing a high pressure test device. The characterization is being supplied here to provide generic spray characteristics for the gasifier injector. LASER SHEET IMAGING
A StockerYale Lasiris Powerline 800mW, 668.5 nm continuous laser, outfitted with optics to produce a 1 mm thick light sheet with 30 degree fan angle was used to illuminate the spray. The light sheet was photographed using a Nikon D300 digital SLR, equipped with a Nikon AF Nikkor 85mm f/1.8D.
OIL SLIDE DROPLET MORPHOLOGY
An oil slide device (Figure 3) was used to capture droplets / particles from the spray in order to study the morphology of the droplets (shape, size, solid / liquid distribution) produced by the gasifier injector. The device consists of an outer tube with a slot (hole) machined into it a short distance from one end. A rod with an oil slide mounted onto it was inserted into the tube such that the oil slide was covered by the outer tube and the slot was sealed by the rod. The end of the rod was located at position 1 initially. The device was then inserted into the spray with the region of interest in the spray directed at the slot. The rod was then rapidly withdrawn from the tube with the end of the rod moving from position 1 to position 2. In withdrawing the tube the oil slide passed by the slot at which time the spray impinged on the oil slide. The particles and droplets become lodged in the oil. The slide then continues to position 2 where the spray can no longer impinge on the oil slide. The device is then removed from the spray. When the rod was withdrawn at an appropriate speed a reasonable number of distinct particles and droplets were collected. The oil droplet slides were photographed using a Nikon D300 digital SLR mounted to a Zeiss West Germany microscope. The images were processed using ImageJ freeware.
Figure 3: Device for collecting spray droplets / particles on an oil coated slide over a short time period, shown in cross-section. MALVERN PARTICLE / DROPLET SIZING
Droplet/particle size distribution in the spray was measured using a Malvern Spraytec particle size analyzer (Figure 4). The analyzer measures droplet size distributions using the technique of laser diffraction. In this technique the angular intensity of light scattered from a spray is measured as the spray passes through a laser beam. The recorded scattering pattern was analyzed using three separate optical models that are suitable for opaque or nearly opaque particles to calculate the particle size distribution; standard opaque, Fraunhofer, and optically dense. The laser beam travels through the entire spray and therefore allows this technique to provide a spatial and temporal average particle size in the measurement zone. The Malvern Spraytec specifications are shown in Table 4. The system was operated in continuous mode for the tests, which limits the speed of Malvern to 1 Hz. 30 sample points collected produce a spatial and temporal
average of the droplets in the spray. The spray was located roughly 300 mm from the receiving optics. At this distance we estimate the Malvern’s operating range is 1.5 µm to 600 µm.
Figure 4: Malvern Spraytec particle size analyzer; 1- HeNe laser, 2- Collimating optics, 3-Measurement zone, 4-Fourier lense, 5-Silicon diode detector array, 6-Data acquisition system.
Table 4: Malvern Spraytec specifications
Particle / Droplet Size Range Lens Range Working Ranges
Concentration Range
0.1 µm – 2000 µm 300 mm lens: 0.1 – 900 microns (Dv50 : 0.5 – 600 microns) 0.5 microns: 150mm 5 microns: > 1000 mm 10 microns: > 2000 mm Patented Multiple Scattering Correction Minimum acceptable transmission: 5% dependent on particle size range
Detection system
36 element log-spaced silicon diode detector array
Light source
632.8 nm, 2 mW helium-neon laser
R E S U LT S
FUEL CHARACTERIZATION
The fuel characteristics of the bituminous coal used in the gasification tests are provided in Tables 5 to 10 and Figures 5 to 8. Table 5: Coal proximate and ultimate analysis Ash Volatiles Fixed Carbon Inherent Moisture C H O N
wt% wt% wt% wt% wt% wt% wt% wt%
9.51 36.08 51.02 3.39 69.38 4.61 12 0.86
Table 6: Mineral oxides SiO2
wt%
51.79
Al2O3 Fe2O3
wt% wt%
17.8 7.24
TiO2
wt%
0.66
P2O5 CaO MgO
wt% wt% wt%
0.72 12.55 2.04
SO3
wt%
4.35
Na2O
wt%
0.3
K2O Ba Sr V Ni Mn Cr Cu Zn Loss on Fusion
wt% wt% wt% wt% wt% wt% wt% wt% wt% wt%
1.15 0.4669 0.0952 0.0151 0.0099 0.937 0.0201 0.0096 0.0107 0.66
Table 7: Ash fusion properties Reducing Initial Spherical Hemispherical Fluid
ºC ºC ºC ºC
1160 1227 1304 1379
Table 8: Particle size distribution by sieve Size 400+ 250 149 105 74 44 38 <38
wt% 0 10.79 9 9.06 9.78 8.95 9.39 43.03
Table 9: Density by He pycnometry Density
kg/m
3
1440
Table 10: Surface properties BET Surface
2
m /kg 3
930.3
Pore Volume
m /kg
3.5(10)
-6
Pore Area Average Pore Diameter
m /g
2
6.1(10)
-4
Å
121.015
2.5
V cm3/g (STP)
2
1.5
1
0.5
0 0
0.2
0.4
0.6
0.8
1
P/Po
Figure 5: N2 adsorption and desorption isotherms. 2000 1800
Viscosity (cP)
1600 1400 1200 1000 800 600 400 200 0 0
50
100
150
Shear Rate (1/sec)
200 45.00%
250 50.00%
Figure 6: Slurry viscosity at various shear rates. Gasification tests performed with 50 wt% solids loading.
1.2
Normalized Weight
1
0.8
0.6
0.4
0.2
0 0
100
200
300
400
500
600
700
800
900
1000
Temperature (deg C)
Figure 7: Loss in weight of coal in oxidizing atmosphere.
7
6
dw/dt (%/min)
5
4
3
2
1
0 0
100
200
300
400
500
600
700
800
900
Temperature (deg. C)
Figure 8: Change in weight of coal with time versus temperature.
1000
GASIFICATION
Figure 9 indicates reactant mass flow rates. Gasifier reactor operation (Table 11) and product gas analysis (Table 12 and Figure 10) are averaged over two time periods; ‘Quench’ and ‘Reactor’ as identified on Figure 10. The slurry flow rate appears to vary, but this is in fact due to the effect of slurry tank mixer vibration on the load cells which are used for determining flow. An alternate flow measurement device is required to avoid this apparent variation. Flow rate was determined by dividing the total slurry consumed by the elapsed time for each period of interest (about 20 minutes). Carbon conversion has been determined based on the sum of the carbon contents of the solids exiting the gasifier (bottom of quench vessel, liquid bag filters) divided by the carbon content of the coal entering the gasifier. After performing an ash balance across the reactor it was determined that not all of the ash entering the gasifier was collected. There are two reasonable explanations for where the balance of the ash is located. A portion of the ash may be built up on the wall of the reactor where it can be expected to have very low carbon content (<= carbon content of ash in the quench vessel). Another portion may be located in piping which can be expected to have similar carbon content to the ash in the bag filter. If we assume the former (ash in reactor) then carbon conversion was 99.7% and if we assume the latter then carbon conversion was 98.0%. No attempt has been made to determine carbon conversion on a gaseous species basis as the analysis is expected to have a greater degree of uncertainty. Table 11: Gasifier reactor operation Gasifier Pressure Temperatures TE-411 TE-412 TE-413 TE-414
kPa(g)
ºC ºC ºC ºC
1528
1461
Quench 1457 1389 1335 1237
Reactor 1478 1403 1346 1250
Quench 31.64 35.83 32.44 0.01 0.08 -99.99
Reactor 31.45 38.19 30.20 0.03 0.09 -99.97
Table 12: Product gas composition
H2 CO CO2 CH4 H2S COS
mol% mol% mol% mol% mol% mol%
100 95 Quench
Reactor
90 Slurry (kg/hr)
85 80 75 70 65 60 55 50 45
Oxygen (kg/hr)
40 35 30 25 20 15 10 5 0 11:31:12
12:00:00
12:28:48
12:57:36
13:26:24
13:55:12
14:24:00
14:52:48
15:21:36
15:50:24
Figure 9: Reactant mass flow rates (kg/hr), 120
0.1 Quench Line Sampling
Reactor Sampling
0.09
100 0.08
0.07 80 0.06
60
0.05
0.04
H2 N2 C0 C02 H2S CH4 C0S
40 0.03
0.02 20 0.01
0 10:48:00
11:16:48
11:45:36
12:14:24
12:43:12
13:12:00
13:40:48
14:09:36
14:38:24
15:07:12
15:36:00
0 16:04:48
Figure 10: Gas composition. Shaded regions indicate period for reported gas compositions in Table 12.
SLURRY INJECTOR SPRAY
Spray characterization tests were performed using a low rank coal slurry with a solids loading of 40% (~500 cP viscosity) at a slurry flow rate of 85 kg/hr. Atomization was assisted by nitrogen at 25 kg/hr in place of the oxygen that is used in gasification tests. The characterization was performed at atmospheric pressure in an unconfined space in the downwards vertical position. LASER SHEET VISUALIZATION
The laser sheet visualization (LSI) has provided an indication of the spray geometry through a plane at the centre line. Figure 11 shows an image generated by LSI. It can be seen that this is a hollow cone spray. The majority of the spray is composed of fine droplets, however, there are a number of larger droplets on the right hand side of the image. It can also be seen that, in addition to the image not being perpendicular to the centre line of the atomizer, that the spray is tending toward the right hand side of the image. This is likely a result of the nitrogen supply being connected perpendicularly to the ‘oxygen supply tube’ only a few tube diameters above the ‘oxygen jets’ during the spray characterization tests. This would result in a slightly higher pressure on one side of the ‘mixing chamber’ and the spray non-symmetry. This is not the case during gasifier operation in which the perpendicular portion of the supply tube is more than a meter from the ‘oxygen jets’. We can expect that the spray will be more symmetrical during gasifier operation. Figure 12 provides estimated spray dimensions derived from the LSI image in Figure 11. The units on the image are arbitrary, with the dimension of ‘126’ equaling 18.8 mm. The included spray angle is 62°.
Figure 11: Laser sheet image of slurry atomization.
Figure 12: Estimated spray dimensions from laser sheet imaging (LSI).
OIL SLIDES
It can be seen from oil slide images (Figure 13) that both water droplets, with included small coal particles, and nearly dry coal particles exist in the spray. In the images the jagged structures are the coal particles, while the spherical ones are water droplets. In some instances jagged structures with semi-spherical regions can be seen. These are coal particles with a layer of water on some parts of the coal particle. The variation in particle shape is an important finding since the optical measurement techniques assume the measured particles are spherical. As such, the data will have to be interpreted accordingly. When this technique is applied to transparent liquids, such as pure water droplets, the centre of the droplet is visible. This is in contrast to the current data, where virtually all the droplets sampled are opaque. We believe this is due to the presence of one or more small coal particles inside the water droplet which block the light. This is advantageous since the diagnostics can only be given one light scattering model per measurement. Figure 13a and 13b are taken from different locations within the spray, so we can see that there is variation in the quantity of water on the coal particles within the spray. During these tests it was not possible to definitively determine the location of the oil slide device in order to provide information on the degree of wetting at any location in the spray. However, in a standard photograph of slurry atomizer a dark fringe appears around the spray (Figure 14) which is not present when atomizing water. This may be a region where dry coal particles can be found. The oil slide images shown and others were processed to determine average particle diameter, perimeter, and circularity. The results show that the average particle diameter is approximately 70 micron. This is significantly larger than the Malvern results shown in the next section. There are several explanations for this including:
1.
When the droplet is sitting on the slide it is no longer round, but semi-circular. This will give the droplet a larger effective diameter due to the mass of the droplet spreading out.
2.
The Malvern can detect smaller droplets than the oil slide method. As such, the average diameter calculated by the Malvern should be lower than that of the oil slides.
3.
The diameters that are calculated from the oil slide images are done so by their confined areas. Many of the particles are not round and/or overlap with neighboring particles resulting in a larger average size. For example, if two 50 micron particles are touching, this will be interpreted as one very large particle rather than 2 small particles.
4.
The results from oil slide measurements can be biased significantly by which area of the slide is inspected. If an area with large droplets is selected the results will be very different than if an area with very small droplets is selected. The Malvern however, measures across the entire spray, offering a more accurate representation of the entire droplet population. a
b
Figure 13: Oil slide droplet images. (a) left – average diameter 82 micron, (b) right – average diameter 68 micron.
Dark region may be coal particles without water layer
Figure 14: Photograph of slurry atomizer during bucket testing. Slurry flow 82 kg/hr, N 2 flow 17 kg/hr.
MALVERN SIZE ANALYSIS
Given the results from the oil slides which indicated that all the particles are opaque, the Malvern data was collected using three different light scatter models that are best suited to opaque or nearly opaque particles. 1. Standard Opaque 2. Fraunhofer
3. Optically Dense Problems were initially encountered with spray accumulating on the Malvern optics, so an air purge system was connected to the Malvern transmitting and receiving optics to prevent particles from accumulating on the windows. The results were a significant improvement on the earlier data. Some particles still stuck to the optics; however, their effect was negligible. The obscuration stabilized after a few seconds of operation (Figure 15), along with the measured droplet size. This bodes well for future high pressure tests. If the windows to the pressure vessel can be kept clean, then data should be acceptable. At the flow rate and coal loading used for the tests the transmissivity was 10%. This is approaching the limit of the Malvern’s operating range. Given the high obscuration, the data could be subject to multiple scattering errors. The coal is opaque, so the low transmissivity could be the result of the laser light simply being blocked, rather than scattered. If this is the case, multiple scattering errors may not be an issue, and the data should be accurate.
Figure 15: Malvern software display of transmissivity (Trans) and droplet size (Dv(50)) with time.
A plot of the percentage of the total droplet/particle volume by droplet Sauter diameter is provided in Figure 16. Table 13 provides the average Sauter mean diameter for the three light scattering models employed. Table 13: Average droplet/particle sized by Marlvern Spraytec
Model SMD micron Standard Opaque 32.4 Fraunhofer 27.2 Optically Dense 30.1
7
6
% of Total Volume
5
4
3
2
1
0. 1 0. 1 0. 2 0. 3 0. 4 0. 5 0. 7 0. 9 1. 3 1. 7 2. 3 3. 2 4. 3 5. 9 8. 0 10 .8 14 .7 20 .0 27 .2 37 .0 50 .3 68 .3 92 . 12 9 6. 17 3 1 23 . 6 3. 31 3 7. 43 2 1 58 . 1 6. 79 1 6. 7
0
Average Doplet Size (um)
Figure 16: Malvern results: Droplet size distribution by volume.
CONCLUSIONS
The majority of data required to validate a computational fluid dynamics reactor model has been provided. There is further data that is being generated by CanmetENERGY to more completely define the system that will be published over the coming year. This data includes slag viscosity measurements and modeling, particle size distribution and composition of char and slag products, determination of fate of trace elements, and data relevant to heat transfer analysis. In addition, future tests will be performed with ‘de-tuned’ gasifier injectors that will reduce carbon conversion to gain data at intermediate stages of reaction as well as insertion of sample probes at locations where coal residence time is reduced. Also, results will be published on a variety of fuels including all ranks of coals, as well as residues from the oil and gas industry.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 GASIFICATION
Numerical simulation analyses of an entrained-bed gasification reactor Ming-Hong Chen, Graduate student, Department of Aeronautics and Astronautics, National Cheng Kung University, No.1, University Road, Tainan City 70101, Taiwan, ROC Tel.: +886 6-2757575x63641, E-mail address: [email protected] Tsung Leo Jiang, Professor, Department of Aeronautics and Astronautics, National Cheng Kung University, No.1, University Road, Tainan City 70101, Taiwan, ROC Tel: +886 6-2757575x63676, fax: +886 6-2389940, E-mail address: [email protected] Yau-Pin Chyou, Senior Specialist, Institute of Nuclear Energy Research Atomic Energy Council, No. 1000, Wenhua Rd., Jiaan Village, Longtan Township,Taoyuan County 32546, Taiwan (R.O.C.) Tel: +886 3-4711400-3422, E-mail address: [email protected] Chang-Bin Huang, Engineer, Institute of Nuclear Energy Research Atomic Energy Council, No. 1000, Wenhua Rd., Jiaan Village, Longtan Township,Taoyuan County 32546, Taiwan (R.O.C.) Tel: +886 3- 4711400-3429, E-mail address: [email protected]
Abstract: A three-dimensional numerical simulation model for the coal combustion and gasification of an entrained-bed gasification reactor has been developed by employing the computational-fluid-dynamic software FLUENT. The numerical simulation model is able to predict the flow and reaction characteristics of a gasification reactor fed by dry pulverized coal, wet pulverized coal, and/or biomass. It adopts several physical models, including the coal gasification model, the turbulence flow model, the turbulence reaction model, and the thermal radiation model, and can be applied to analyzing a gasification reactor with a multi-feeding fuel-injection system at varying operation conditions. The results obtained from the present study show that the generated syngas is primarily composed of carbon monoxide and hydrogen, and the predicted outlet gas composition is in good agreement with the experimental result. For coal gasification, a lower oxygen/carbon ratio is found to produce more carbon monoxide and hydrogen, and the outlet temperature is relatively lower. The produced carbon monoxide decreases with an increasing water/coal ratio. However, there exists an optimal water/coal ratio for the maximum hydrogen production. The generated syngas by using a general low-carbon high-oxygen biomass as the stock is predicted to have a composition much lower in carbon monoxide and higher in carbon dioxide than that by using the coal.
Introduction The development and management of energy has become a vital subject due to the first and second energy crisis occurred in 1973 and 1979 [13]. The shortage of oil and natural gas results in the soaring of the energy cost. The exhaustion of oil will be inevitable within 50 years under the current consumption rate. On the contrary, the coal, as 72% of the fossil fuel, will be the primary energy source. The critical problem of the traditional coalutilization by combustion is the release of the environmentally impacted products, such as CO2, NOx and SOx. Therefore, the development of the clean coal combustion technology has become a major task to meet the dual environmental and economical requirements. By minimizing the release of polluting molecules, the clean goal technique aims to make the coal a cheap and non-polluting source of energy. Typical clean coal combustion technologies include the pressurized fluidized-bed combustion (PFBC), the integrated gasification fuel cell (IGFC), and the integrated gasification combined cycle (IGCC). The IGFC system has the advantages of ultra high energy conversion efficiency and clean utilization, as shown in Fig. 1. With the commercialization of the high-efficiency turbine, the plant efficiency is able to reach 50%, and it can be up to 60% if an IGFC system is employed. The gasification process is a non-catalyzed, incomplete oxidation reaction [5]. The purpose of the gasification process is to convert solid fuel into gaseous species, such as CO and H2 for combustion or the chemical process. The traditional coal combustion process results in a large amount of CO2, H2O and thermal energy, but it is an endothermic reaction for the coal gasification process. Typical species within the gasification syngas include CO, H2 and CH4. Recently, most of the gasified syngas is processed into CO and H2 [6] for further applications. There are two primary reactions of the complicated gasification process, i.e., the
pyrolysis/devolatilization and the char oxidation reaction. The schematic diagram of the pyrolysis of a single coal particle is shown in Fig. 2. The volatile species undergo the gaseous oxidation reaction, producing CO2, H2O, and CO. After the devolatilization process, the remaining solid char reacts with CO2, H2O, and O2 as the consequential gasification process. With the purification process, the gasified syngas is applicable for the sequential utilization. The gasifier can be classified into three common types by the feeding velocity, i.e., the fixed-bed, the fluidized-bed, and the entrained-bed gasifier. The feeding velocity of the oxidizer of the entrained-bed gasifier is the fastest, and the size of employed the coal particle is the smallest among three types of gasifiers. The primary (for fuel) and the secondary (for oxidizer) feeding ports of the entrained-bed gasifier are at the same position, leading to the optimal mixing for the gasification process. The gasification rate of the entrained-bed gasifier is the fastest, and the amount of the fuel-processing per day is also the largest among the three types of gasifiers. Therefore, the entrained-bed gasifier is the main object in the development of the gasifier. There are some relevant studies on the numerical analysis of the entrained-bed gasifier. Syred et al. [8] showed that FLUENT 6.0 was able to analyze the particle combustion in a complicated flow field by the Lagrangian tracking method. Grabner et al. [9] developed a numerical model for the pressurized fluidized-bed gasifier. The trace of coal particle was described by the EulerianLagrangian model to predict the particle behavior, pyrolysis, and surface reactions. Wang et al. [10] found that the employed turbulent model is a significant factor on the evaluation of the gasification performance by using several turbulent models and devolatilization models for the entrained-bed gasifier. Ajilkumar et al. [11] presented a numerical model for the air-blown, labscale gasifier using the commercial software FLUENT 6.2. Wang et al. [12] examined four chemical reaction models on the prediction of the
performance of the entrained-bed gasification reactor. The coal gasification technology has the advantage of reducing the environmental impact, compared to the traditional coal combustion utilization. Therefore, it is essential to develop the capability of evaluating the gasification performance. The objective of the present study is thus to develop a three-dimensional, numerical model for the entrained-bed gasifier. The model is then applied to analyze the gasification process of an entrained-bed gasifier fed by dry pulverized coal, wet pulverized coal or biomass.
Numerical Model The numerical model for the entrained-bed gasifier is developed by using the commercial software FLUENT 6.3[13]. Numerically, the flow within the gasifier is assumed to be composed of the continuous phase for the gaseous flow and the discrete phase for the coal particles. Transport Equations of the Continuous Phase For the continuous phase, the governing equations include the mass, momentum, energy and species conservation equations. The mass conservation equation is: ui S p (1) xi where the source term on the RHS of the Eq. (1) is the mass source due to the pyrolysis of the coal particle. The momentum conservation equation is: ui u j x j (2) p ui u j u 'i u ' j SMk x j xi xi x j xi
with the Boussinesq eddy viscosity hypothesis, the Reynolds stresses in Eq. (2) can be represented as: u u j 2 u 'i u ' j k ij t i (3) x 3 x i j The transport equations for the turbulent kinetic energy and turbulent dissipation rate are:
t k xi ui t C1 P C2 xi xi xi k ui k P xi xi
(4) (5)
with empirical constants as [14]: C 0.09
C1 1.44 C2 1.92
(6)
k 1.0 1.3 The energy conservation equation is: (7) ui h K T S ph xi xi xi where the second term on the RHS is the source term to evaluate the radiation of the coal particles. The species conservation equation is: DkmYk S pYk Rk uiYk (8) xi xi The source term (second term on the RHS) is the species mass source/sink due to the pyrolysis or the oxidation reactions. The third term on the RHS is the species production or consumption due to the chemical reactions.
Transport Equations of the Discrete Phase For the transport equations of the discrete phase of the coal particles, the track of the coal particle is obtained by integrating the forces acting on the particle. The forces acting on the coal particle include the inertial of the particle and the drag force, and can be described as: du p dt
FD u u p
where
FD Re
(9)
18 C D Re p d p2 24
d p u p u
C D a1
a2 a 32 Re Re
(10)
The parameters in Eq. (9) are applicable to the spherical particle and a range of Re [15]. The track of the particle is evaluated by the instantaneous velocity, which is related to the mean velocity and the fluctuating velocity as: (11) u* u u ' where the fluctuating velocity in Eq. (11) is defined by the stochastic discrete random walk model. The initial particle distribution is described by the Rossin-Rammler distribution. The relation between the particle diameter and the particle mass fraction can be described as: n (12) Yd exp d p / d where d is the mean particle diameter, and n is the distribution parameter. The energy equation for the particle accounts for convection, radiation, devolatilization and surface reactions, and is given as: dT m p c p p hAp T TP p Ap R4 T p4 dt (13) dm p dm p h fg f h H reac dt dt where m p and Tp are the particle mass and particle temperature, respectively. temperature, defined as:
R is the radiating
R G / 4 1/ 4 ,
(14) and G is the incident radiation, which is related to the radiate intensity as: 4
G Id
(15) . The radiation in the gasifier is described by the P-1 model [16]. The convection coefficient between the particle and the gaseous flow field is evaluated by the empirical relation proposed by Ranz and Marshall [17] as: 0
Nu
hd p k
2.0 0 .6 Re 1d/ 2 Pr 1 / 3
.
(16)
Devolatilization The devolatilization reaction of the coal particle is described by the single rate reaction step model proposed by Badzioch et al.[18] as: k A1e E / RT (17)
As the volatile species diffuses out of the coal particle, the heterogeneous reactions on the particle surface occur. Those particle surface reactions are described by the implicit relations of Smith et al. [19] as: Rk ,r AprYk Rk ,r (18) Rk ,r
R Rkin ,r pn k ,r D0,r
Rkin,r ArT n,r e Er / RT
Nr
(19) (20)
The char and volatile species will undergo the oxidation reaction to produce thermal energy and gasification to generate CO and H2. These reactions are evaluated by the probability density function (PDF) tables as described below. Computational Domain and Boundary Conditions The employed coal is exploited by the China Shenhua Energy Company, and its properties are shown in Table 1. The LHV of the employed coal is 29.2 MJ/kg. The composition of the employed coal is shown in Table 2. The particle size distribution is shown in Table 3. In the present study, the gasification performance of coal will be compared with that of biomass. The properties of the employed biomass [20] are shown in Table 4. Three operating parameters, i.e., O/C ratio, feeding temperature, and water/coal ratio, are selected to investigate the effect on the gasification performance. The O/C ratio is defined as the mass ratio of the oxygen in the introduced oxidizer to the carbon of the employed coal. The generated syngas and the outlet temperature will be analyzed according to the employed operating conditions. The schematic diagram of the investigated entrained-bed gasifier [21] is shown in Fig. 3. The length of the entrained-bed gasifier is 2.72m and the diameter of the combustion chamber is 15cm. Three feeding ports are installed at the angles of 60°, 180° and 300° as shown in Fig. 3. Fuel is introduced from the inner feeding port via the discrete-phase injection while oxidizer is introduced from the outer ring with a fixed mass flow rate. The mass flow rate
of the oxidizer is 0.01116 Kg/s and the coal/gas ratio is 2.579. The feeding temperature is 400K, and the operating pressure is 13.7 atm. For simplification, it is assumed that the bottom of the reactor is closed and the wall of the reactor is adiabatic. Non-Premixed Combustion Model and Probability Density Function Under the assumption of the non-premixed combustion model, the instantaneous thermochemical state of the fluid is related to a conserved scalar quantity known as the mixture fraction, f. The mixture fraction can be written in terms of the atomic mass fraction as [22]
f
Z i Z i ,ox Z i , fuel Z i ,ox
(21) where Zi is the elemental mass fraction for element, i. The subscript ox denotes the value at the oxidizer stream inlet and the subscript fuel denotes the value at the fuel stream inlet. The species equations can be reduced to a single transport equation for the mixture fraction, f, under the assumption of equal diffusivities. The reaction source terms in the species equations cancel, and thus f is a conserved quantity. While the assumption of equal diffusivities is problematic for laminar flows, it is generally acceptable for turbulent flows where turbulent convection overwhelms molecular diffusion. The Favre mean (density averaged) mixture fraction equation is f f t f S m t t (22)
The source term Sm is due to the mass into the gas phase from reacting particles (i.e., coal). In addition to solving for the Favre mean mixture fraction, the conservation equation for the mixture fraction variance will also be solved [23], which is given as: f 2 f 2 t 2 (23) t f 2 C g t f Cd f 2 k t
where f f f . The values for the constants t ,
Cg , and Cd are 0.85, 2.86, and 2.0, respectively. The probability density function, written as p(f), can be thought of as the fraction of time that the fluid spends in the vicinity of the state f. Written mathematically: 1 p f f lim i T T i (24)
where T is the time scale and i is the amount of time that f spends in the Δf band. The shape of the function p(f) depends on the nature of the turbulent fluctuations in f. The shape of the assumed PDF, p(f), is described by the mathematical function, i.e., the -function. The -function PDF shape is given by the following function of f and f 2 :
p f where
f 1 1 f
1
f 1 1 f
1
df
(25)
f 1 f a f 1 f 2 (26) f 1 f (1 f ) 1 f 2 (27) The PDF shape of p(f) is a function of the mean mixture fraction and the mixture fraction variance only. Thus, solving the conservation equations for the mean mixture fraction and the mixture fraction variance at each point in the flow field, the assumed PDF shape can be computed and used as the weighting function to determine the mean values of species mass fractions, density, and temperature. Significant computational time can be saved by computing these integrals once, storing them in a look-up table, and retrieving them during the iteration.
Results and Discussion The probability density function (PDF) tables for the transport equations are generated for the non-premixed combustion model according to the
employed coal properties and operating conditions. The largest mean temperature is 3408 K, which means the largest possible temperature under the employed conditions, as shown in Fig. 4. To validate the proposed numerical model, the composition of the generated syngas is compared with the experimental results [21] as shown in Fig. 5. The generated syngas predicted is mainly composed of CO and H2, while the fractions of CO2 and H2O (product of the oxidation reaction) are relatively small. Therefore, the gasification process dominates within the gasifier rather than the oxidation reaction, and this trend is consistent with the feature of the investigated gasifier. In comparison with the experimental result, the discrepancy in the outlet gas composition is only about 5.8%, validating the proposed numerical model. The temperature distribution within the gasifier is shown in Fig. 6. The largest temperature within the gasifier is about 3018K, resulting from the combustion at the feeding region to provide the thermal energy for the gasification reaction. Since the combustion occurs near the feeding region only, the downstream temperature is relatively low. The outlet temperature is about 1600K. The oxygen distribution is shown in Fig. 7. Oxygen is almost exhausted near the feeding region, and, therefore, only the gasification reaction proceeds at the downstream region. The distribution of H2O is shown in Fig. 8. Corresponding to the consumption of oxygen, H2O is mainly produced near the feeding region. There is no significant variation for the H2O concentration at the downstream region due to the lack of oxygen. The distribution of CO is shown in Fig. 9. The fraction of CO is small near the feeding region, and it gradually increases along the reactor. The molar fraction of CO at the outlet of the gasifier is about 60%. The distribution of H2 is similar to that of CO except in the bottom region, as shown in Fig. 10, and its fraction is 18% at the outlet of the gasifier. The species distributions along the axial direction of the gasifier are shown in Figs. 11 and
12. The oxygen fraction increases at 0.2 of the dimensionless position, as shown in Fig. 11, and it is also the position of the feeding port introducing the oxidizer and fuel. The fraction of oxygen decreases afterward due to the oxidation reaction, and depletes at 0.22 of the dimensionless position. The gasification reaction proceeds after the oxidation reaction. The fraction of CO and H2 decreases at the region of the feeding port due to the oxidation, and gradually increases after the region of the oxidation reaction. It is noticeable that the fraction of the CO and H2 is relatively high within the region from 0 to 0.18 of the dimensionless position. This is because the bottom of the gasifier (0 of the dimensionless position) is closed, and the high operating pressure results in a slow velocity, leading to the relatively high fraction of the carbon, CO and H2. This trend was also observed in the study of Wang et al. [24]. The fraction of CO2 and H2O, as shown in Fig. 12, increases drastically at the region of the oxidation reaction (0.2 of the dimensionless position) and reduces to about 0.1 of the molar fraction afterward. The gasification performance using the biomass was compared with that using coal. The composition of the employed biomass is shown in Fig. 13. There is a significant composition difference between biomass and coal. The carbon of biomass is 30% less than that of coal while it is 30% higher in oxygen. The heating value of biomass is 62% lower than that of coal, as shown in Table 5, leading to 735K lower in the largest mean temperature, as shown in Table 5. The composition of the generated syngas using different fuels was compared at Fig. 14. The fraction of CO using biomass is 30% lower than that using coal. The fraction of H2 using biomass is 4% lower than that using coal. For the product of the oxidation reaction, the fraction of CO2 using biomass is 2.3 times larger than that using coal, and the fraction of H2O is also 3.9 times larger than that using coal. This is because the fraction of carbon in the employed biomass is relatively low, while the fraction of oxygen is relatively high, leading to less gasification reaction and more oxidation reaction. The outlet gas
temperatures are compatible in both cases (about 1600K), since both cases are under the same O/C ratio, as shown in Table 5. Overall, the gasification performance of coal, in terms of the production of CO and H2, is obviously better than that of biomass. The effects of the O/C ratio, the feeding temperature, and the water/coal ratio on the gasification performance are investigated. For the O/C ratio, two additional cases are examined besides the base case as shown in Table 6, and the results are compared in Fig. 15. Since part of the non-reactive carbon remains in the bottom of the reactor, the O/C ratio based on the actually consumed carbon is 0.63, 0.65, and 0.71, respectively. The predicted outlet concentrations of CO and H2 increase as the O/C ratio decreases while those of CO2 and H2O decrease as the O/C ratio decreases. The outlet temperature decreases with the decrease of the O/C ratio. This is because a lower oxygen fraction in the case of a lower O/C ratio leads to less oxidation reaction, more gasification reaction, and a lower outlet temperature. For the feeding temperature, two additional cases are examined to evaluate its effect on the gasification process, as shown in Fig. 16. The predicted outlet concentrations of CO and H2 increase while those of CO2 and H2O decrease as the feeding temperature increases. The outlet temperature increases about 10K as the feeding temperature increases from 300K to 500K. This is because part of the thermal energy from the higher feeding temperature is absorbed by the endothermic gasification reaction, resulting in only 10 K increase in the outlet temperature. To investigate the effects of the adjustment of water/coal ratio, the cases with 20%/80% and 40%/60% water/coal ratio are examined. Under the same introduced mass flow rate, the designated fraction of the fuel (20% and 40%) is replaced by water, and it is dry feeding for the base case. Detailed specifications are shown in Table 7, and gasification results are shown in Fig. 17. At a 20%/80% water/coal ratio, the predicted outlet
concentration of CO reduces by 8% while it significantly reduces by 28% at a 40%/60% water/coal ratio. For H2, its fraction increases about 4% at a 20%/80% water/coal ratio. At a 40%/60% water/coal ratio, however, the fraction of H2 reduces to the level of the dry-feeding base case. Therefore, there should be an optimal water/coal ratio for the maximum production of H2. The variation of CO2 is not significant at a 20%/80% water/coal ratio while it increases 8% at a 40%/60% water/coal ratio, which is apparently related to the reduction of CO. The fraction of H2O gradually increases with the increase of the water/coal ratio. The outlet temperature, as shown in Fig. 18, increases with the increasing water/coal ratio. This is because the O2/Coal ratio also increases in the adjustment of the water/coal ratio.
Conclusions A three-dimensional numerical simulation model for the coal combustion and gasification of an entrained-bed gasification reactor has been developed. The results obtained from the present study show that the generated syngas is primarily composed of carbon monoxide and hydrogen and, the predicted outlet gas composition is in good agreement with the experimental result. The gasification performance using biomass was investigated and compared with coal. The case using a general low-carbon high-oxygen biomass results in 30% lower for CO, 4% lower for H2, 2.3 times higher for CO2, and 3.9 times larger for H2O in the generated syngas. The outlet temperatures are comparable in both cases. Overall, the gasification performance of coal, in terms of the production of CO and H2, is better than that of biomass. In the parametric study, the results show that the outlet concentrations of CO and H2 increase and those of CO2 and H2O decrease as the O/C ratio decreases. The outlet temperature decreases with a decreasing O/C ratio. The results also show that the outlet concentrations of CO and H2 increase, while those of CO2 and H2O decrease as the feeding temperature increases. The produced CO decreases
with an increasing water/coal ratio. However, there exists an optimal water/coal ratio for the maximum H2 production.
[9]
Acknowledgement This work was sponsored by the Institute of Nuclear Energy Research Atomic Energy Council under contract No. 982001INER018. Part of the continuous follow-up activities is supported by National Science Council under the code No. NSC 98-3114-Y-042A-005. The authors would also like to thank Dr. Po-Chuang Chen and Mr. Yan-Tsan Luan for their assistance in revising the manuscript.
[10]
[11]
Reference [1] [2] [3]
[4]
[5]
[6] [7]
[8]
R. A. Hinrichs and M. Kleinbach, Energy: Its Use and the Environment, Sea Harbor Drive, Orlando, USA, 2002. R. A. Ristinen and J. J. Kraushaar, Energy and the Environment, Wiley, New York, 1999. L. A. Ruth, "Energy from Municipal Solid Waste: A Comparison with Coal Combustion Technology," Prog. Energy Combust. Sci., 24, pp.545-564, 1998. T. Kivisaari, P. Bjornbom, C. Sylwan, B. Jacquinot, D. Jansen and A. de Groot, "The feasibility of a coal gasifier combined with a high-temperature fuel cell," Chemical Engineering Journal, 100, pp.167-180, 2004. L. D. Smoot and D. T. Pratt, "Pulverized-Coal Combustion and Gasification Theory and Applications for Continuous Flow Processes," New York and London, 1979. L. D. Smoot and P. J. Smith, Coal Combustion and Gasification, Plenum Press, New York, 1985. D. R. Tree, D. L. Black, J. R. Rigby, M. Q. Mcquay and B. W. Webb, "Experimental Measurement in the BYU Controlled Profile Reactor," Prog. Energy Combsut. Sci., 24, pp. 355-383, 1998. N. Syred, K. Kurniawan, T. Griffiths, T. Gralton and R. Ray, "Development of fragmentation models for solid fuel combustion and gasification as subroutines for
[12]
[13] [14] [15]
[16] [17] [18]
[19]
inclusion in CFD codes," Fuel, 86, pp. 22212231, 2007. M. Grabner, S. Ogriseck and B. Meyer, "Numerical simulation of coal gasification at circulating fluidised bed conditions," Fuel Processing Technology, 88, pp. 948-958, 2007. T. Wang and A. Silaen, "Effects of Turbulence and Devolatilization Models on Gasification Simulation," 25th International Pittsburgh Coal Conference, Pittsburgh, USA, 2008. A. Ajilkumar, T. Sundararajan and U. S. P. Shet, "Numerical modeling of a steamassisted tubular coal gasifier," International Journal of Thermal Sciences, 48, pp. 308-321, 2009. T. Wang and A. Silaen, "Comparison of instantaneous, equilibrium, and finite-rate gasification models in an entrained-flow coal gasifier," 26th International Pittsburgh Coal Conference, Pittsburgh, USA, 2009. Fluent Incorporated, Fluent 6.3, User Guide, 2006. B. E. Launder and D. B. Spalding, Lectures in Mathematical Models of Turbulence, Academic Press, London, England, 1972. S. A. Morsi and A. J. Alexander, "An investigation of particle trajectories in twophase flow systems," J. Fluid Mech., 55, pp. 193-208, 1972. P. Cheng, "Two dimensional radiating gas flow by a moment method," AIAA, pp. 16621664, 1964. W. E. Ranz and W. R. Marshall, "Evaporation from drops. Part 1," Chem. Eng. Prog., 48, pp. 141-146, 1952. S. Badzioch and P. G. W. Hawksley, "Kinetics of Thermal Decomposition of Pulverized Coal Particles," Ind. Eng. Chem. Process Design and Development, 9, pp. 521530, 1970. I. W. Smith, "The combustion rates of coal chars: A review, in: Proceedings of the 19th International Symposium on Combustion,"
[20]
[21]
[22]
[23]
[24]
The Combustion Institute, pp. 1045-1065, 1982. P. Klimantos, N. Koukouzas, A. Katsiadakis and E. Kakaras, "Air-blown biomass gasification combined cycles (BGCC): System analysis and economic assessment," Energy, 34, pp. 708-714, 2009. J. H. Kuo, "Three dimension numerical simulation of pulverized coal gasification," Master thesis, National Cheng Kung Unversity, Taiwan, ROC, 2006. Y. R. Sivathanu and G. M. Faeth, "Generalized State Relationships for Scalar Properties in Non-Premixed Hydrocarbon/Air Flames," Combustion and Flame, 82, pp. 211230, 1990. W. P. Jones and J. H. Whitelaw, "Calculation Methods for Reacting Turbulent Flows: A Review," Combustion and Flame, 48, pp.1-26, 1982. T. Wang, A. Silaen, H. W. Hsu and M. C. Lo, "Part-Load Simulations and Experiments of a Small Coal Gasifier," 23rd International Pittsburgh Coal Conference, Pittsburgh, PA, USA, 2006. NOMENCLATURE
a1 a2 a3 A
Empirical constant for drag coefficient Empirical constant for drag coefficient Empirical constant for drag coefficient Surface area -1
-1
cp
Specific heat (J kg K )
C1 C2 Cd CD
Empirical constant for turbulence model
E f f FD G hi h fg
H reac Total enthalpy of reaction Radiation intensity I k Kinetic rate Fluid thermal conductivity (W m-1 K-1) kf ks L m m Mi n Nu
Solid thermal conductivity (W m-1 K-1) Cell length (mm) Mass (kg) Mass flow rate (kg s-1) Molecular weight (g mol-1) Distribution parameter Nusselt number
patm
Atmospheric pressure (atm)
Pn
SMk
Bulk partial pressure of the gas phase species (Pa) Prandtl number Reynolds number Chemical reaction rate (mol m-3 s-1) Gas constant (J mol-1 K-1) Mass into the gas phase from reacting particles Momentum sink/source term (N m-3)
Sp
Source term for the pyrolysis of the coal
S ph
particle Source term of the radiation of the coal
S pYk
particles Species mass source/sink due to the
T T u u u x
pyrolysis or the oxidation reactions Temperature (K) Time scale (s) Velocity (m s-1) Fluctuating velocity (m s-1) Instantaneous velocity (m s-1) X coordinate (mm)
Pr Re Ri R Sm
Empirical constant for turbulence model Empirical constant for mixture fraction
Cg
Drag coefficient Empirical constant for mixture fraction
C
Empirical constant for turbulence model
d Dkm
Diameter (mm) Mass diffusion coefficient for species in the mixture
Total energy (J) Mixture fraction (dimensionless) Fluctuating mixture fraction Drag force (N lbf ) Incident radiation Species enthalpy (J mol-1) Evaporative enthalpy (KJ kg-1)
X
Yd Yk Zi
x/L, dimensionless coordinate along the cell length (-) The initial particle distribution
Mass fraction of each species Elemental mass fraction for element, i
Greek letters Turbulent dissipation rate Effectiveness factor Dynamic viscosity (Pa s) Density (kg m-3) Stefan-Boltzmann constant
k
(5.67 ×108 W m-2 K-4) Empirical constant for turbulent kinetic energy
t R
Empirical constant for turbulent dissipation Empirical constant for the mixture fraction Velocity (m s-1) Delta function Radiating temperature (K) Solid angle (radians)
Subscripts Kinetic energy k Particle p Abbreviation ox Oxygen reac Reaction kin Kinetic
Table 1. Coal properties [13] Property Value/Unit Density 1300 kg/m3 Cp 1000 J/kg-K Thermal Conductivity 0.0454 W/m-K Latent Heat 0 Vaporization temp. 400 K Volatile fraction (%) 33.76 Binary Diffusivity 5E-4 m2/s Particle Emissivity 0.9 Particle Scattering Factor 0.6 Swelling Coefficient 2 Burnout Stoichiometric Ratio 2.67 Combustible Fraction (%) 60 Table 2. Coal composition Species Wt% C 0.80142 H 0.0423 O 0.14149 N 0.0118 S 0.003 Table 3. Particle size distribution Particle size distribution Minimum Maximum Average
Table 6. Investigated Oxygen/Coal ratio O/C O2/Coal(MAF) 0.33 0.27 0.45 0.36 0.57 0.46 Table 7. Oxygen/Coal ratio and coal slurry concentration for the adjustment of water/coal ratio Water/Coal O2/Coal(MAF) Coal(MAF)/Slurry 0%/100% 0.46 1.00 20%/80% 0.59 0.77 40%/60% 0.81 0.56
Fig. 1. Schematic diagram for the IGFC system [4]
(μm) 50 70 60
Table 4. Biomass composition Species Wt% C 0.497 H 0.062 O 0.440 N 0.010 S 0.000 Table 5. Comparison of the thermal properties for coal and biomass Property Coal Biomass LHV (J/kg) 2.92E+7 1.12E+7 Max. mean temp. (K) 3408 2673 Outlet gas temp. (K) 1591 1582
Fig. 2. Schematic diagram of the pyrolysis of a single coal particle [7]
Fig. 6. Temperature distribution of the gasifier along the axial direction Fig. 3. Gasifier geometry and size [21]
Fig. 7. Oxygen distribution along the axial direction of the gasifier Fig. 4. PDF table for the mean temperature
Fig. 8. Steam distribution along the axial direction of the gasifier
Fig. 5. Comparison of the outlet gas composition with experimental results [21]
Fig. 9. Carbon monoxide distribution along the axial direction of the gasifier Fig. 12. Distribution of CO2, H2O and O2 along the axial direction of the gasifier
Fig. 10. Hydrogen distribution along the axial direction of the gasifier
Fig. 13. Comparison of the composition of the employed coal and biomass
Fig. 11. Distribution of CO, H2, O2 and carbon along the axial direction of the gasifier Fig. 14. Comparison of the outlet gas composition using coal and biomass
Fig. 18. Outlet temperature under different water/coal ratios
Fig. 15. Outlet gas composition and temperature under different O/C ratios
Fig. 16. Outlet gas composition and temperature under different feeding temperatures
Fig. 17. Outlet gas composition under different water/coal ratios
Comparison of measured and calculated viscosities of German lignite based slags
Arne Bronsch* and Patrick J. Masset
Freiberg University of Mining and Technology Centre for Innovation Competence Virtuhcon – Group “Multiphase Systems” Fuchsmühlenweg 9, Reiche Zeche D-09599 Freiberg, Germany
Abstract The viscosity of German lignite based slags was investigated at temperatures up to 1700 °C with a Couette type viscometer under reducing conditions to simulate gasification conditions. The experimental values are compared with available models of the literature. For iron, sulphur and sodium rich slag and for silica poor slag some discrepancies have been observed. For the other investigated compositions a pretty good agreement was observed between experimental and calculated values.
*
Corresponding author. Tel./Fax: +49 3731 39 4809/4555
E-Mail: [email protected]
1
Introduction
The thermophysical properties of molten slags are of importance for several industrial processes: metallurgy, gasification/combustion, glass industry and so on. Depending on the industrial process the composition of the slags differs and it exhibits different viscous behaviour. The experimental conditions to operate the measurements are usually very delicate (corrosion issues). Then it becomes challenging to predict the viscosity of such media to reduce the experimental work. However, the prediction of viscous properties of molten real slags is not as obvious as it includes a high number of components. Since the middle of the last century prediction of the viscosity of molten slags was examined by several authors. Viscosity measurements on coal ash slags by [1] [2] [3] [4] [5], on artificial laboratory-prepared slags [4] [6] [7] and on natural slags such as magmatic liquids [8] have been reported. In addition, different models were developed to predict the viscosity of such mixtures by including the changes in microstructure of the slag as a function of the temperature and composition. For gasification processes, e.g. in entrained flow coal gasification, the properties of the liquid slag is important as it influences the reactor protection against high temperature corrosion [9]. In the field of gasification and for one coal, Vargas et al. [10] compared several viscosity models. Very often the developed models cover a limited composition range and cannot be used for a extended range as it may be useful in the case of coal based slags. The available models have been thoroughly reviewed and a summary of the equations, application conditions for each model is given in Appendix A. The objective of the present work is to compare the values of experimental viscosity measurements obtained with two German lignite coal slags with the available models in the literature.
2 2.1
Experimental Materials
In this study, two slags were prepared from two German lignites. In addition, an artificial twocomponent testing ash/slag (Ash/Slag T) was prepared to compare with literature data. For the ash preparation, an amount of around 5 kg (Sample A) and 1 kg (Sample B) of untreated coal was filled into crucibles and ashed automatically under air in a furnace for 48 hours. At high temperature it was hold for ca. 24 hours. To avoid any exchange of volatiles between the two coals, the samples were ashed separately. After cooling and checking of possible coal/coke odds (which would cause a second ashing) the ashes were weighted and filled into glass flasks. Amounts of ca. 72 g (Ash A) and 91 g (Ash B) were produced in one batch. The ash T was prepared to focus a composition with a low melting point at 1709 K [21] to measure in a wide range of temperature. SiO2 (Aldrich Chemistry 99.8 % purity) and 10.6 g of CaO (Aldrich Chemistry 99.9 % purity) were used to prepare approx. 30 g of the binary mixture. The slag sample was prepared by melting the ash inside a boron nitride crucible during the viscosity measurement without thermal pre-treatment.
2.2
Techniques
XRF-analysis was carried out with equipment by BRUKER AXS Tiger (wavelength dispersive). XRD-analysis was done with BRUKER AXS D8 Discover with cobalt tube, Goebel mirrors, XYZ table and Våntec-1 detector. Viscosity measurements were achieved with Ravenfield Model FG MkIV viscometer (Couette-type). A simplified overview is given in Fig. 1a. The system is mounted in a cabined. Heaters are made of graphite and a maximum temperature of 1700 °C is reachable. The furnace consist of two parts which are brought together to enclose the sample holder. Temperature measuring is done by Raytek precision pyrometer which receives radiation from
a graphite crucible considered as a black body. Inside the graphite crucible is a boron nitride crucible which contains the molten slag. For the torque transmission a bob made of molybdenum is connected to a shaft made of boron nitride. During experiments the housing was purged with nitrogen. The oxygen level was kept close to 0.004 % per volume. A standard rheometer type Physica MCR 301 by Anton Paar was used for viscosity measurements (sample B). The principle configuration is given in Fig. 1b. The rheometer is installed on a concrete table, which reduce vibrations. Inside the table is the high frequency furnace installed, respectively. An air bearing and a high precession drive enable high resolved torque measurements (minimum up to 0.1 μNm. The torque is directly captured by electrical parameters of the motor. Rheological properties are determined by torque and agitation. A bob (diameter 6 mm, height 20 mm) rotates inside a crucible (inner diameter 10 mm, inner height 60 mm). The gab width is around 4 mm.
(a)
(b)
Fig. 1: (a) Simplified assembly Ravenfield Model FG MkIV viscometer (Couette-type), (b) Schema of the viscosimeter Physica MCR 301 (Anton Paar).
3
Results and discussion
3.1
Characterization of the ash
The ash compositions and phases were determined by XRF (Table 1) and XRD (Table 2), respectively. The main characteristic of the ash A is the low silicon content (almost zero) and the high concentration of Na and Mg. Ash B exhibits a low Na content whereas it contents approx. 30 mass % eq. oxide of silica. The Mg concentration in ash B is low and close to 2 mass % (compared to 23 mass % in ash A). The Fe2O3 and CaO contents are similar in the investigated ashes. The amount of sulphur differs significantly from ash A to ash B, e.g. 50 % higher in the latter one. This is supported by phase analysis (Table 2), which shows that the amount of anhydrite in ash B is far higher than in ash A (42 mass % compared to 19 mass %). Sodium sulphate has been evidenced in ash A and represents the rest of sulphur based species in ash A. The main phase observed in ash A are: anhydrite, lime, periklas and Srebrodolskit + Braunmillerit with a concentration close to 20 mass %. Some remaining calcite was evidenced (~ 4 mass %) and may arise from incomplete ashing. In ash B, anhydride and quartz are the two main phases, representing 40 and 25 mass %, respectively. Calcium-silicates (wollastonite and larnite) were also detected (around 8 m/o in total).
Table 1: XRF analysis of prepared ashes at 815 °C in air (mass % equivalent oxide) Name
Al2O3 BaO CaO Fe2O3 K2O MgO
MnO
Na2O SiO2 SO3 TiO2 Total
Ash A
1.0
0.3
39.6
11.8
1.6
23.4
0.1
5
0.2
16.9
0.1
100
Ash B
2.1
0.1
28.4
12.7
0.1
2.4
0.3
-
29.6
24.2
0.1
100
Ash T
-
-
35.2
-
-
-
-
-
64.8
-
-
100
Table 2: XRD analysis of prepared ashes at 815 °C in air (in mass %) Phase
Formula
Akermanite
Ca2MgSi2O7
-
1.7
-
Anhydrite
CaSO4
19.4
40.2
-
Braunmillerite
Ca2FeAlO5
-
4.4
-
Calcite
CaCO3
4.2
-
-
Diopside
CaMgSi2O6
-
2.6
-
Gehlenite
Ca2Al2SiO7
-
0.2
-
Hematite
α-Fe2O3
-
0.7
-
Larnite
Ca2SiO4
-
3.9
-
Lime
CaO
22.7
0.4
35.2
Maghemite
γ-Fe2O3
-
2.4
-
Magnetite
Fe3O4
-
2.0
-
Mayenit
Ca12Al14O33
-
0.6
-
Periclase
MgO
22.7
2.0
-
Quartz
SiO2
-
25.9
64.8
Sodium sulfate
Na2SO4
9.7
-
-
Srebrodolskit
Ca2Fe2O5
-
9.3
-
21.3
-
-
-
3.7
-
100
100
100
Srebrodolskit + Braunmillerit Ca2Fe2O5+Ca2FeAlO5 Wollastonite Total
CaSiO3
Ash A Ash B Ash T
3.2 •
Viscosity measurements Sample T
The curves viscosity vs. temperature compared with literature data [8] [22] [23] is given in Fig. 2. The curve obtained with the sample T exhibits four distinct regions. Above 1720 K it was ascribed to the viscosity of the liquid phase. Between 1670 and 1720 K, it corresponds to the biphasic mixture liquid + CaSiO3. The third region in the 1500-1670 K temperature range may corresponds to the viscosity of the newly formed glass and it becomes a solid for temperatures lower than 1500 K.
Viscosity in Pa s
1.0E+03 1.0E+02 1.0E+01 1.0E+00 1.0E-01 1.0E-02 1.0E-03 1500
1600
1700
1800
1900
2000
T in K [8] 64.0 mass % SiO2 [23] 61.5 mass % SiO2
[22] 62.8 mass % SiO2 Slag T 64.7 mass % SiO2
Fig. 2: Comparing of own data and literature data for Slag T (system SiO2-CaO) [8] [22] [23]. •
Sample A
Silica-poor ash (sample A) was heated up and measurements were done in the range of 1570 K-1980 K. The results are plotted in Fig. 3. Viscosity exhibits an almost linear behaviour in the temperature range of 1570 K-1860 K. Around 1910 K a smooth decrease of the viscosity was observed and may be ascribed to a phase transition from rich liq. + sol. to pure liq. The measured viscosity is above 2 magnitudes (Streeter model) and 5 magnitudes
(Riboud model), respectively and therefore not useable for correct predictions in technical applications. Neither the alternating changes nor the value of viscosity were predicted correctly by all models. 1.0E+05
Viscosity in Pa s
1.0E+04 1.0E+03 1.0E+02 1.0E+01 1.0E+00 1.0E-01 1.0E-02 1.0E-03 1.0E-04 1.0E-05 1000
1200
1400
1600
1800
2000
T in K Bottinga-Weill
Urbain
Riboud
Streeter
Slag A
Fig. 3: Comparison of calculated (maximum/minimum) and measured viscosities for Slag A.
•
Sample B
Silica-rich Ash B was heated up above melting point and viscosity measurements were done in a temperature range of 1448 K-1823 K. The results are presented Fig. 4 and compared with calculated values from literature models (minimum and maximum values). Around 1450 K the viscosity decreases rapidly with increasing temperature. A linear behaviour was observed in the range of 1473 K-1823 K. For temperatures around 1648 K an increasing of viscosity was recorded which could occur by phase transitions. A graphical comparison shows measured values are in the limits of predictions. In the temperature range 1473 -1698 K, the Watt-Fereday’s model fits well the measured values. Above 1698 K Lakatos’ model agrees with the experimental pretty well. Also Lakatos shows the decrease of viscosity in the temperature range of 1450 K-1500 K even with a smaller slope.
1.0E+05 1.0E+04 Viscosity in Pa s
1.0E+03 1.0E+02 1.0E+01 1.0E+00 1.0E-01 1.0E-02 1.0E-03 1.0E-04 1.0E-05 1000
1200
1400
1600
1800
2000
T in K Kalmanovitch-Frank
Lakatos
Riboud
Watt-Fereday
Slag B
Fig. 4: Comparison of calculated (minimum/maximum results) and measured viscosities for Slag B.
3.3 •
Viscosity predictions Sample T
All used viscosity models decreasing viscosity with increasing temperature on the simple Slag T calculations (Fig. 5). The calculated viscosity for the composition of sample T varies from 103 Pa.s at 2000 K to 107 Pa.s at 1000 K. The lowest viscosity values were predicted by Kalmanovitch-Frank in the temperature range 1000 K-1600 K (8.7.103 Pa s-2.8 Pa s) and by Lakatos in the range 1700 K-2000 K (1.3 Pa s-1.4.10-1 Pa s). The highest viscosities are given by Watt-Fereday in the range 1000 K-2000 K (2.7.1010 Pa s-1.1.102 Pa s).
Viscosity in Pa s
1.0E+11 1.0E+09 1.0E+07 1.0E+05 1.0E+03 1.0E+01 1.0E-01 1.0E-03 1000 1200 Reid-Cohen Watt-Fereday Shaw Riboud Slag T - measurement
1400 1600 Sage-McIlroy T in K Bomkamp Lakatos Kalmanovitch-Frank
1800 2000 Modified Silica Ratio Bottinga-Weill Urbain Streeter
Fig. 5: Viscosity predictions of Slag T
When comparing measured and predicted values Fig. 6 is given for the best accurate model predictions. Measured viscosity is in the prediction range of Bomkamp (maximum) and (Kalmanovitch-Frank (minimum) at 1550-1550 K. The strong decrease of viscosity (ca. 2 magnitudes) during heating depends probably on phase transition and liquid-solid behaviour of partly molten slag occurring in the phase diagram CaO-SiO2 [21]. At 1550 K-1680 K recorded viscosity is similar to the model of Kalmanovitch-Frank. Deviations around 1 Pa s occur. Around 1680 K a viscosity decrease is recognized (ca. 0.6 Pa s-0.3 Pa s). Reason could be the beginning of complete melting process which is given around 1709 K in literature. Viscosity keeps almost constant in the temperature range of 1685 K-1725 K. It seems that melting process has finished at 1725 K, viscosity decreases up to the temperature limit of 1751 K (0.2 Pa s-8.5.10-3 Pa s).
1.0E+05
Viscosity in Pa s
1.0E+04 1.0E+03 1.0E+02 1.0E+01 1.0E+00 1.0E-01 1.0E-02 1.0E-03 1500
1600
1700
1800
T in K Bomkamp
Kalmanovitch-Frank
Slag T
Fig. 6: Comparing predicted (maximum/minimum values) and measured viscosity of Slag T
•
Sample A
Predicted viscosities characteristics in a temperature range of 1000 K–2000 K are given in Fig. 7. Slag A has a negligible amount of SiO2. It was observed that none of the existing models may predict correctly the viscosity of such type of slag in the temperature range of the investigations. The predicted values are somewhat two orders of magnitude lower than the measured values and the experimental data cannot be fitted. It seems that the viscosity of real slags exhibiting high concentrations of sulphur, iron and sodium can not be predicted with a high reliability with the present models. This point has been poorly discussed in the literature and need further assessment. The observed discrepancies may be ascribed to the change in the composition of the slag due to environmental conditions (oxygen partial pressure).
Viscosity in Pa s
1.0E+05 1.0E+04 1.0E+03 1.0E+02 1.0E+01 1.0E+00 1.0E-01 1.0E-02 1.0E-03 1.0E-04 1.0E-05 1000
1200
1400
1600
1800
Reid-Cohen
T in Ratio K Modified Silica
Bottinga-Weill
Urbain
Riboud
Streeter
Kalmanovitch-Frank
Slag A - measurement
2000
Fig. 7: Selected results of viscosity predictions for Slag A
•
Sample B
The viscosity predictions of Slag B show a pretty good agreement with the experimental in a wide temperature range of 1000-1900 K (Fig. 8). Deviations smaller than 2 magnitudes are estimated around 1500 K. The models of Lakatos [7] and Watt-Fereday [11] offer the best fitting of the experimental data. However, the present models do not integrate the phase changes and fail at lower temperatures below 1400 K.
Viscosity in Pa s
1.0E+05 1.0E+04 1.0E+03 1.0E+02 1.0E+01 1.0E+00 1.0E-01 1.0E-02 1.0E-03 1.0E-04 1.0E-05 1000 Reid-Cohen Watt-Fereday Shaw Riboud
1200
1400
1600
T in K Sage-McIlroy Bomkamp Lakatos Kalmanovitch-Frank
1800
2000
Modified Silica Ratio Bottinga-Weill Urbain Slag B
Fig. 8: Selected viscosity predictions for Slag B.
4
Conclusion
The viscosity of artificial and real coal slag has been measured. The measured values were compared with calculated values. It appeared that the literature model can partially predict the experimental data. For slag with low silica contents, none of the modes can fit the experimental data. The reasons of the discrepancies are that the models do not take into account the local chemistry and the changes of composition with temperature. This point need to be further investigated by including the thermodynamic data of the investigated mixtures.
Acknowledgement The financial support from BMBF (Federal Ministry of Education and Research) and SMWK (Saxon State Ministry for Higher Education, Research and the Fine Arts) is gratefully acknowledged. Acknowledgements are due to Prof. Scheller and Mr. Hötzel at Institute of
Iron and Steel Technology (University of Freiberg), for their help in measuring the viscosity of sample B. Also acknowledgement to the technical staff of the Institute of Energy Process Engineering and Chemical Engineering (University of Freiberg) for the chemical analysis.
References 1.
Reid, W.T. and P. Cohen, Transactions of the American Society of Mechanical
Engineers, 1944. 66: p. 83-97. 2.
Sage, W.L. and J.B. McIlroy, Relationship of coal ash viscosity to chemical
composition. Journal Name: Combustion; (United States); Journal Volume: 31:5, 1959: p. Medium: X; Size: Pages: 41-48. 3.
Hoy, H.R., A.G. Roberts, and D.M. Wilkins, IGE J, 1965(June): p. 444-469.
4.
Urbain, G., et al., Viscosity of silicate melts. Transactions & Journal of the British
Ceramic Society, 1981. 80: p. 139-141. 5.
Kalmanovitch, D.P. and M. Frank, Mineral matter and ash deposition from coal.
1988: p. 89-101. 6.
Bottinga, Y. and D.F. Weill, The viscosity of magmatic silicate liquids; a model
calculation. Am J Sci, 1972. 272(5): p. 438-475. 7.
Lakatos, T., L. Johansson, and B. Simmingsköld, Viscosity temperature relations in
the glass system SiO2-Al2O3-Na2O-K2O-CaO-MgO in the composition range of technical glasses. Glass Technology - European Journal of Glass Science and Technology, 1972. 13(3): p. 88-95. 8.
Urbain, G., Y. Bottinga, and P. Richet, Viscosity of liquid silica, silicates and alumino-
silicates. Geochimica et Cosmochimica Acta, 1982. 46(6): p. 11. 9.
Higman, C. and M. van der Burgt, Gasification Processes, in Gasification (Second
Edition). 2008, Gulf Professional Publishing: Burlington. p. 91-191.
10.
Vargas, S., F.J. Frandsen, and K. Dam-Johansen, Rheological properties of high-
temperature melts of coal ashes and other silicates. Progress in Energy and Combustion Science, 2001. 27(3): p. 237-429. 11.
Watt, J.D. and F. Fereday, FLOW PROPERTIES OF SLAGS FORMED FROM ASHES
OF BRITISH COALS-1. Inst Fuel-J, 1969. 42(338): p. 99-103. 12.
Quon, D.H.H., S.S.B. Wang, and T.T. Chen, Viscosity Measurements of Slags From
Western Canadian Coals. Journal of Engineering for Gas Turbines and Power, 1985. 107(3): p. 803-806. 13.
Shaw, H.R., Viscosities of magmatic silicate liquids; an empirical method of
prediction. Am J Sci, 1972. 272(9): p. 870-893. 14.
Vargas, S., F. Frandsen, and K. Dam-Johansen, ELSAM - Idemetsi Kosan cooperative
research project: performance of viscosity models for high-temperature coal ashes, in CHEC. 1997: Lyngby. 15.
Riboud, P.V., et al., Improvement of continuous casting powders. Fachberichte
Hüttenpraxis Metallweiterverarbeitung, 1981. 19(10): p. 859-869. 16.
Streeter, R.C., E.K. Diehl, and H.H. Schobert. MEASUREMENT AND PREDICTION
OF LOW-RANK COAL SLAG VISCOSITY. in ACS Symposium Series. 1984. Washington, DC, USA: ACS. 17.
Machin, J.S. and D.L. Hanna, VISCOSITY STUDIES OF SYSTEM CaO-MgO-A2O3-
SiO2: I, 40% SiO2. Journal of the American Ceramic Society, 1945. 28(11): p. 310-316. 18.
Machin, J.S. and T.B. Yee, VISCOSITY STUDIES OF SYSTEM CaO-MgO-Al2O3-
SiO2: II, CaO-A2O3-SiO2. Journal of the American Ceramic Society, 1948. 31(7): p. 200-204. 19.
Srinivasachar, S., et al., A fundamental approach to the prediction of coal ash deposit
formation in combustion systems., in PSI Technology Co. 1992.
20.
Harb, J.N., C.J. Zygarlicke, and G.H. Richards, The effect of particle composition and
temperature on the deposition of two Western US coals in a laminar drop-tube furnace. Vol. 66. 1993, London, ROYAUME-UNI: Institute of Energy. 21.
Verein Deutscher Eisenhüttenleute, V., et al., Slag Atlas. 2 ed. 1995, Düsseldorf:
Verlag Stahleisen GmbH. 616. 22.
Bockris, J.O.M., Lowe, D. C., Viscosity and the Structure of Molten Silicates.
Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 1954. 226(1167): p. 423-435. 23.
Neuville, D.R., Viscosity, structure and mixing in (Ca, Na) silicate melts. Chemical
Geology, 2006. 229(1-3): p. 28-41. 24.
Klein, L.C., B.V. Fasano, and J.M. Wu, Viscous flow behavior of four iron-containing
silicates with alumina, effects of composition and oxidation condition. Journal of Geophysical Research, 1983. 88(Supplement). 25.
Saito, N., et al., Viscosity of blast furnace type slags. Metallurgical and Materials
Transactions B, 2003. 34(5): p. 509-516. 26.
Saito, N., et al., Viscosity of Blast Furnace Type Slags. Metallurgical and Materials
Transactions B: Process Metallurgy and Materials Processing Science, 2003. 34(5): p. 509516. 27.
Greenberg, S., Viscosity of synthetic and natural coal ash slags, in Materials Science
and Technology Division. 1984, University of Chicago: Chicago. 28.
Mills, K.C., Mineral matter ans ash in coal. ACS Symposium, 1984. 301: p. 195-214.
29.
Mills, K.C., Viscosities of molten slags. 1992, National Physical Laboratory.
30.
Mills, K.C. and C.P. Broadbent, Evaluation of slag program for the prediction of
physical properties of coal gasification slags, in Impact ash deposition coal fired plants. 1993. p. 513-525.
31.
Senior, C.L. and S. Srinivasachar, Viscosity of ash particles in combustion systems for
prediction of particle sticking. Energy & Fuels, 1995. 9(2): p. 277-283.
Appendix A: Selected models for prediction of viscosity of molten slags
Drying Mechanism of Low Rank Coal with Different Reacting Conditions: Fixed Bed vs. Fluidized Bed Taejin Kang, Doman Jeon, Sihyun Lee*, Sangdo Kim* and Hyungtaek Kim Division of Energy Systems Research, Ajou University San 5 Woncheon-Dong, Youngtong-Gu, Suwon, Korea Tel: 82-31-219-2972, Fax: 82-31-219-2969 E-mail: [email protected] *Korea Institute of Energy Research
Abstract Low rank coal (lignite) can be successfully utilized in the thermal power plant through the pretreatment of coal with efficient way. The price of low rank coal is equivalent to one third of steaming coal. However, it is difficult to be utilized in thermal power plant mainly for two reasons: high moisture, and instability. In the present investigation, experiments are progressed with low rank coal to reducing moisture content. The experimental parameters of this study are drying temperature and time duration as well as particle size as divided three sections of 0.3~1mm, 1~1.18mm and 1.18~2.8mm. Temperature variation on the drying results that moisture content is not much changed below 80℃ and drying is saturated above 150℃. With different particle size investigated in this study, drying behavior is not much different with particle size. Furthermore, it also indicated that pore structure changed after dewatering which can be discerned the SEM microscopy such that Lignite progressively transforms pore structure of mesopore into micropore when it is dried. Consequently, it can be found that a total dimension of pore is reduced through the dewatering lignite. Investigation will be progressed with different reaction condition, especially in fluidized-bed. The resulting data will be used in designing the fluidized bed drying demonstration plant.
1. Introduction
Due to energy crisis, there has been an increasing interest in coal all over the world. Among energy resources, interest in Low Rank Coal is rising along with energy security. That is the coal which contains higher moisture content from 30% up to 60 %. In order to utilize those Low Rank Coal as fuel for generating electricity, it should be required to precede drying process prior to its utilization. When moisture gets dry, water content decreases below 10 % and then caloric value increases up to 6,000 ~ 6,500 kcal/kg. Although it seems feasible to achieve enhancement of Low Rank Coal, it is required to get method of stabilizing surface by means of interpretation of mechanism about reattachment because it can possibly cause reattachment of moisture. This study does the experiment on Low Rank Coal drying depending on different velocity, using fluidized-bed and fixed-bed characteristics, and putting drying parameter - temperature and diameter particle size in same condition. The reason why Indonesia lignite is selected as coal for experiment is that it is the High Rank Coal of all Low Rank Coal. Not only does it have great access in South Korea, but also has possibility of industrialization because of a large amount of reservoir.
2. Experimental
The substance used in this experiment is a lignite containing 30 % moisture contents, showing main characteristic in Table 1. Item
Unit
Result
%, Wt
33.02
Inherent Moisture
%, Wt
13.63
Ash
%, Wt
3.30
Volatilc Matter
%, Wt
44.44
Fixed Carbon
%, Wt
38.63
Gross Calorific Value(Air Dry Basis)
Kcal/Kg
5,370
Gross Calorific Value(Dry Basis)
Kcal/Kg
6,220
Total Moisture(As Received Basis) Proximate Analysis(Air Dry Basis)
Higher Heating Value
Table 1. Basis properties of the lignite
The whole process is illustrated as Fig.1. below. As it seems, it is used air as drying medium. It is carried out by taking advantage of heated gas. When it goes into reactor, the installation helps gas flow through Air Box spread out. For the purpose of steady flow velocity, Cock Valve is installed after regulator. minimum fluidization velocity on fluidized-bed experiment is equal to 1.5 times of higher, and done with a condition H/D=1. The rate of H/D in this experiment is equivalent of 40g coal. Also, diameter size of coal has three types: 0.3-1mm, 1.18-2.8mm, 2.8-4mm. The temperature of Inlet gas is followed by 100℃, 125℃, 150℃. The reactor is made of glass so that flow pattern can be overseen in person. In addition, it has insulation. In case of fixed-bed, it is conducted as 4.6cm/sec velocity
Fig.1. Sketch of Fluidized Bed Setup
3. Result
3.1 Fluidized bed Experiment depending on different diameter size was carried along with 1.5 times higher minimum fluidization velocity. Variation by temperature is shown as Fig.2, Fig.3, Fig.4. It is observed by putting different temperature condition in same diameter size.
Moisture Content(%)
35
0.3-1mm, 100℃ 0.3-1mm, 125℃
30
0.3-1mm, 150℃
25 20 15 10
5 0
0
2
4
6
8
10
Drying time(min)
Fig.2. Moisture content change of lignite by temperature at 0.3-1mm
Moisture 35 Content(%) 30
1.18-2.8mm, 100℃ 1.18-2.8mm, 125℃
25
1.18-2.8mm, 150℃
20 15 10 5 0 0
2
4
6
8
10
Drying time(min) Fig.3. Moisture content change of lignite by temperature at 1.18-2.8mm
Moisture 35 Content(%) 30
2.8-4mm, 100℃ 2.8-4mm, 125℃ 2.8-4mm, 150℃
25
20 15 10 5 0 0
2
4
6
8
10
Drying time(min) Fig.4. Moisture content change of lignite by temperature at 1.18-2.8mm
Except for the case of 0.3-1mm diameter size at 100℃, when the experiment is under 10 minute drying, it is recognizable that moisture content of coal is 5 % or so. The part shown as below 0% was omitted because of experimental error. It is found that different diameter size and temperature play crucial role in containing individual moisture content. From same diameter size, higher temperature indicates more active reaction, compared with at 100℃. In case of same temperature condition, it can be shown that diameter size does not have great impact on drying moisture contents when observing differences between 100℃ and 150℃. Although diameter size acts on time of drying, it seems to have same tendency in drying curve. In addition, although the differences between 0.3-1mm and 1.182.8, 2.8-4mm is identified at 100℃ and 125℃, in comparison between 1.18-2.8mm and 2.8-4mm, it is found there is little change.
3.2 Fixed bed
It is carried out with 4.6cm/sec, changing diameter and temperature. The influence depending upon diameter size is shown as Fig.5. Fig.6. Fig.7. Despite 50 minute drying, it turns out that moisture content is considered difficult to fall down below 15%. Additionally, it is found that moisture content reaches down 50 % of initial moisture content. As a result of fixed reactor, though it can be figured out that temperature and diameter size have an effect on drying, it is identified that influence of temperature is greater than diameter size. Aside from 0.3-1mm case, it is shown that there is little influence at 1.18-2.8mm, 2.8-4mm Moisture 35 Content(%) 30 25
20 15 10
0.3-1mm, 100℃ 0.3-1mm, 125℃ 0.3-1mm, 150℃
5 0 0
10
20 30 Drying time(min)
40
50
Fig.5. Moisture content change of lignite by temperature at 0.3-1mm
Moisture 35 Content(%) 30 25
20 15 10
1.18-2.8, 100℃ 1.18-2.8mm, 125℃ 1.18-2.8mm, 150℃
5 0 0
10
20
30
40
50
Drying time(min)
Fig.6. Moisture content change of lignite by temperature at 1.18-2.8mm
Moisture 35 Content(%) 30 25
20 15 10
2.8-4mm, 100℃ 2.8-4mm, 125℃ 2.8-4mm, 150℃
5 0 0
10
20
30
40
50
Drying time(min)
Fig.7. Moisture content change of lignite by temperature at 2.8-4mm
3.3 Comparison slope of Fixed Bed and Fluidized Bed.
As a result of experiment through fixed-bed reactor, fixed-bed reactor can be compared with fluidized-bed reactor. Thus it is figure that drying in the fluidized-bed reactor which had frequent contact between air and solid is more active. The difference between fixed bed and flow bed is shown as Fig.8. Fig.9. Fig.10. by drawing slope of conversion.
Fig.8. Conversion change of lignite by temperature 100℃
Fig.9. Conversion change of lignite by temperature 125℃
Fig.10. Conversion change of lignite by temperature 150℃
As a consequence of comparison between fixed bed and fluidized bed, except for experimental error, when presenting conversion of two reaction about time of drying, it is identified that two types show in a linear way. When it comes to slope, fluidized bed reaction has ten times greater slope than fixed bed. This is explained that fluidized bed reaction can remove moisture faster than fixed bed owing to difference of contact between gas and solid.
4. Conclusion
In order to use Low Rank Coal - lignite as electric-generating fuel, it is required to have drying processes. For this process performance, fixed-bed and fluidized-bed reaction are used in this experiment. Compared with fixed bed reaction, fluidized bed reaction takes less time to proceed drying process. It results from differences in contact between air and solid. The errors from fluidized bed reaction is caused by a decreased amount of moisture from heating for drying the coal. It is found that it can produce floating matter which is smaller than initial particle size. To be more clear mechanism, drying kinetics regarding fixed-bed and fluidized-bed should be added to this experiment along with variation of drying parameter.
ACKNOWLEDGMENT This work was supported by the Power Generation & Electricity Delivery of the Korea Institute of Energy Technology Evaluation and Planning(NP2008-0092-05) grant funded by the Korea government Ministry of Knowledge Economy
5. Reference 1) G. Couch, "Coal upgrading to reduce CO2 emission", IEA Clean Coal Center Report, p72, 2002 2) Edward K. Levy, Hugo S. Caram, Zheng Yao, Zhang Wei and Nenad Sarunac, KINETICS OF COAL DRYING IN BUBBLING FLUIDIZED BEDS, proceedings Fifth World Congress on Particle Technology, Orlando, Florida, April 2006. 3) Green Energy and Technology, Zero-Carbon Energy Kyoto 2009, DOI: 10.1007/978-4-431-99779-5_35 4) Wen, C.Y. and Yu, Y.H., "A generalized method for predicting the minimum fluidization velocity", AIChE J., 12, 610-612 (1966). 5) Duane G. Levine, Richard H. Schlosberg, Bernard G. Silbernagel, Understanding the chemistry and physics of coal structure(A Review), Proc. NatL Acad. Sci. USA Vol. 79, pp. 3365-3370, May 1982 6) Young-Cheol Bak, Hyun-Soo Yang and Jae-Ek Son, Change in Pore Structure of Chars Obtained under Different Pyrolysis Conditions, HWAHAK KONGHAK Vol. 28, No. 6, December, 1990, pp691-698 7) Christian Bergins, Kinetics and mechanism during mechanical/thermal dewatering of lignite, Fuel 82(2003) 355-364
THE NATURAL TECHNOLOGY FOR PRETREATMENT AND UTILIZATION OF THE ENERGETIC FLY ASH Maria KUSNIEROVA1, Maria PRASCAKOVA1, Peter FECKO2, Rudolf MATYSEK2 1
Institute of Geotechnics of Slovak Academy of Sciences Watsonova 45, 043 53 Kosice, Slovak Republic, [email protected] 2 VSB-Technical University of Ostrava, Faculty of Mining and Geology 17. listopadu Str. 15, 708 33 Ostrava-Poruba, Czech Republic
Abstract The progressive usage of primary raw materials force on to the higher wastes utilizations and processing in the case of the useful components containing. By the development of new technologies of this wastes processing can be imitate the processes that are running in the nature by the geo-sphere formation. Provided research confirmed that the fly ash can be used as the substituent for volcanic ashes by the synthetic zeolitees and mullite-corundum materials preparation, using the principles of natural genesis of raw materials. Key words: fly ash, volcanic glass, non-metallic materials Introduction Energetic fly ashes are by the course of the law categorized as wastes in the general. By its formation, composition and some facilities but also by the forecastable possibilities of its using are in many aspects equivalents to the some primary, commercially used non-metallic materials. The aim of the presented paper is comparison of the composition of examined blackcoal fly ashes with the selected primary non-metallic materials on the fundamentals basics and the definition of its industrial usage possibilities as the secondary crude. From the whole capacity of the energetic wastes from the combustion of the fossil fuels are the fly ashes of 75-85% volume [1]. Following the available data it can be anticipated, that yearly production of this waste is of 1 million ton volume in Slovakia [2]. Intensity of the fly ash as the secondary raw material utilization is very low and is related to the brown coal fly ashes and stabilisate that are certified and used mainly by the building materials production [3]. Volume of the black coal fly ashes and stabilisate utilization is inappreciable ant they are not used in principle and are only detonated at the dumps. The main barrier for the fly ash utilization is the high content of the unburned coal that is in some burning technologies of 20% high [1]. Nowadays dominate trends of the fly ashes utilization in the all volume without its qualitative pretreatment. However this way is the cheapest but on the one way it has its qualitative and quantitative limits and on the other way this can be not finding as rational because it don’t allow utilization of the useful components contained in the fly ash. As more rational can be considered approach to the fly ashes utilization after the technological treatment focused on the full or partial useful components separation and following complex utilization of the whole fly ash volume on the principle of the qualitative equivalence to the primary raw materials. The newest trends focused on the sophistic methods of the fly ash matter utilization as precursors for the composite materials preparation by equivalent principle and composition near to the natural materials and enabling to save the primary raw materials. In the tables 1 and 2 are shown comparisons of elementary composition and chosen elements occurrence forms in fly ashes and equivalent primary raw materials. Compared are also classical commercial usage of primary raw materials with prognosis for possibilities of usage of fly ashes and products of its treatment or modification.
PDF created with pdfFactory Pro trial version www.pdffactory.com
Table 1. The comparison of the appearance of the majority components forms obtained in the black coal fly ashes and in the primary non-metallic materials including their usage [1,8,9,10]. Classical commercial Primary occurrence Secondary equivalents Supposed usage Element usage forms in lithosphere contained in fly ash possibilities glass, abradant and Si in solution with Al, quartz, crisobalite, vitric phases production, Si Si clear Si production – not tridymite cenospheres composites economic Si solutions, Al Al production, abradant, corundum, bauxite, Hydrargillite, bohmite, Al production - not corundum ceramics bohmite, gibbsite accessories economic Heatproof composites, Si-Al adsorbents, ceramics, mullite, sillmanite, zeolites, clays, synthetic zeolites and crystalline heatproof materials, andalusite, analcim, mullite, sillmanite, geopolymeres and glass melt, abradant, phylipsite, mordenite, feldspars, volcanic production, additives to amorphous insulating materials auspidine glasses, the building bricks form production magnetite, maghemite, magnetite, hematite, accessory iron production, Fe production, steal and maghemite, hematite, Fe siderite-in pigments, iron filling Fe alloys, pigments siderite deulphurisation masses technologies building mat., gypsum gypsum and and plasterboards, gyps, anhydrite, gyps, anhydrite plasterboards decorating stones, limestone, dolomite, Ca majority obtained in production, additive to desulphurization calcite, marble stabilisate the building composite mediums, fertilizers materials rullite, illmenite, pigments production, rutil, illmenite, absolutely usage isn’t illmenorullite, Ti special alloys illmenorullite, anatas anticipated anatase energetic, organic chemistry
black and brown coal
C
unburned coal containing coke, graphite
energy production
Table 2. The comparison of the chosen elements in the energetic wastes on the basis of the black coal fly ashes and the primary non-metalic materials [1,8,9,10]. Energetic wastes on the basis Equivalent primary not-metallic materials of black coal fly ashes petrurgic black coal black coal ceramic clays black coal zeolites mullite basalt component fly ash stabilizate slag (%) (%) (%) (%) (%) (%) (%) SiO2 39.9-48.4 52.0 22.7 43.5-47.0 49.5-66-3 65.0-71.3 29.01 Al2O3 17.8-20.7 22.4 10.7 11.0-13.0 20.0-29.8 11.5-13.1 69.6 Fe2O3 5.8-10.2 5.4 4.26 4.0-7.0 2.2-3.8 0.7-1.9 0.5 CaO 2.0-3.7 4.1 40.8 10.0-12.0 0.4-1.2 2.7-5.2 MgO 0.9-1.2 1.8 8.0-11.0 0.6-0.95 0.6-1.2 TiO2 0.6-0.8 0.8 2.0-3.5 1.1-1.3 0.1-0.3 0.79 K2O 1.5-2.3 0.5 1.0-2.5 0.8-2.3 2.2-3.4 0.06 Na2O 0.5-0.8 2.5 2.5-3.5 0.1-0.2 0.2-1.3 0.18
Results and discussion In principle the main of the possible solutions of the fly ashes utilization is encrypted in some ways in the equivalent action of the primary non-metallic materials genesis creation. One of the good examples of the genesis equivalence could be mentioned comparison of the generation and alternation of the fly ash wit the natural volcanic ashes containing volcanic glasses, pumic and the other tufa rocks. Both of these materials are products of the thermal processes. By this processes occurs quick cooling of the Si-Al melt and that’s why aroused material is of the high content of not-crystalline vitric phases. This in natural conditions is the subjected to the devitrfication process under the zeolite and illite structures formation. These
PDF created with pdfFactory Pro trial version www.pdffactory.com
processes were spotted also in the case of the study of fly ashes dumped for the different time (see Tab. 3), where with the dumping time prolongation was loss of the amorphous vitric phase and creation of the illite and zeolite structures detected [5]. Table 3. Content of amorphous phases in fly ashes dumped for different time [5]. Energetic fly ash fresh 5 years dumped Amorphous phases 90.65 79.30 content (%)
20 years dumped 67.7
The zeolites are in the nature generated by the hydrothermal alternation of volcanic ashes. Achieved results focused on the modeling of the hydrothermal alternation condition of the energetic fly ashes mention possibilities of the amorphous phase of the fly ash structure transformation on the zeolite structure of the phylipsite type. In the figures 1-3 are shown the XRD patterns of fly ash, zeolite and zeolite prepared from fly ash.
Figure 1. XRD pattern of original fly ash
PDF created with pdfFactory Pro trial version www.pdffactory.com
Figure 2. XRD pattern of original zeolite
Figure 3. XRD pattern of zeolite prepared form fly ash
PDF created with pdfFactory Pro trial version www.pdffactory.com
It was experimentally confirmed that the zeolites from the fly ashes quality is influenced by its pretreatment what is documented in the table 4. Table 4. The comparison of the exchange capacities of natural zeolites and zeolitised fly ashes Sample Exchange capacity (mol. kg-1) High-class natural zeolite 1.88 Low-class natural zeolite 0.47 Fly ash 0.01 Zeolitized fly ash without tretment 1.24 Pretreated fly ash in Humpry spiral 0.1- 0.5 mm / light product 2.41 0.1- 0.5 mm / heavy product 0.60 0.071- 0.1 mm / light product 1.55 0.071- 0.1 mm / heavy product 0.70 0.04- 0.071 mm / light product 1.57 0.04 – 0.071 mm / heavy product 0.65
It was experimentally proved that zeolitized fly ashes applied in to the concretes can increase robustness of these concretes [7]. On the comparison base of the elementary content equivalence, occurrence of the fly ash components forms and the generation of the mullite in the nature it can be proposed that this material could be precursor for the mullite production under the extensional conditions. It is advisable to note that the mullite occurrence in the nature is very rare, but it is very interesting for its thermal properties for usage in the industry. All the volume of the mullite ceramics consumption is covered by the synthetic mullite produced from the natural precursors, as well as the sillmanite, cyanite, andalusite and topaz, or from the pure Al and Si oxides solution in appropriate stoichiometric ratio by the thermal synthesis. From the tables 2 and 3 is liquet that the black coal fly ashes don’t have the identical stoichiomethric Si and Al ratio, but contains forms of the natural precursor mineral and the abundant content of the mullite building elements. Experimentally was confirmed that by the stiochiomethric changing of the Si and Al ratio can be the more types of composite material consisting of the black coal fly ash and the Al additive produced. After the thermal synthesis are the mullite and corundum materials produced as it is shown in the figure 4 a and b [4]. (a)
(b)
Figure 4. The comparison of the stoichiomethric composition of the black coal fly ash before (a) and after (b) thermal synthesis.
Conclusion Study of the equivalence generation of the elementary composition and the elements occurrence forms in the fly ashes and the primary non-metallic materials offers many possibilities for the rationally usage of the black coal fly ashes as the worthful secondary raw material as well as its complex wasteless usage. The further equivalence study with the natural materials, its generation and usage gives many opportunities for the next research of the sophistic technologies and materials.
PDF created with pdfFactory Pro trial version www.pdffactory.com
Acknowledgement This research has been carried out within the project NFP 26220220051 Development of progressive technologies for utilization of selected waste materials in road construction engineering, supported by the European Union Structural funds and SGA project No. 2-008610. References [1] FEČKO Peter, KUŠNIEROVÁ Mária, RÁCLAVSKÁ Helena, ČABLÍK Vladimír, LYČKOVÁ Barbora: Fly ash. VŠB-Technical University of Ostrava, Faculty of Mining and Geology, 2005, ISBN 80-248-0836-6, 191 p. [2] MAAS Peter, MATYS Mirko: Stabilita odkalísk, PFUK, Katedra inžinierskej geológie, manuscript. [3] Personal consultation ENO Nováky. [4] KUŠNIEROVÁ Mária, LUPTÁKOVÁ Alena, ŠLESÁROVÁ Andrea: Bezodpadová technológia komplexného spracovania a využitia čiernouhoľných popolčekov, In. Recyklace odpadů XI./II., Košice 6-7. 12. 2007, ISBN 978-80-248-1676-0, s. 15-20. [5] KUŠNIEROVÁ Mária, ŠLESÁROVÁ Andrea, PRAŠČÁKOVÁ Mária: The signitificance of fly ash for their processing and utilization, In. Proc. Waste recycling-IX., Krakow, 17-19.11. 2005. [6] KUŠNIEROVÁ Mária, PRAŠČÁKOVÁ Mária, FEČKO Peter: Modification of ashes sorption properties by hydrothermal alternation. In: Proceedings of 22nd Annual International Pittsburgh Coal Conference Coal – Energy and the Environment, Pittsburgh, USA, 2005, CD. [7] JUNÁK Jozef, ŠTEVULOVÁ Nadežda, SINČÁKOVÁ A.: The tempemperature influence of final compressive strength of coal fly ash/ cement free samples, Czasopismo techniczne, chemia, 2-Ch/2008 zeszyt 16(105. Wydawnictwo politechniky Krakowskej im tadeusza Kosciuszki, ISSN 0011-4561, ISSN 1897-6298., s. 151-156. [8] ČABLÍK Vladimír: Získavání Ti a Al z popílku, Doktroská práce, VŠB TU Ostrava, 2001. [9] BERRI L., MASON B., DIETRICH R.: Mineralogia, Moskva, Mir, 1987. [10] BŰCHNER W., SCHLIEBS R., WINTER G., BŰCHEL K.H.: Průmyslová anorganická chemie, SNTL, Praha 1991.
PDF created with pdfFactory Pro trial version www.pdffactory.com
1 PHYSICAL STRUCTURE AND CHEMICAL PROPERTIES OF ORGANIC MATTER OF BROWN COALS FROM DIFFERENT FIELDS IN RELATION TO THE COMPOSITION OF MINERAL COMPONENTS P.N.Kuznetsov, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences 42 K.Marx str, 660049 Krasnoyarsk, Russia Fax: 7(391)2124720; E‐mail: [email protected] L.I.Kuznetsova Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences,
42 K.Marx str, 660049 Krasnoyarsk, Russia Fax: 7(391)2124720; E‐mail: [email protected]
Abstract—The cation-exchange forms of a considerable portion of metals that occur in brown coals from various deposits were identified. Based on swelling data, the interaction of the organic matter of coals with solvents was studied depending on the concentrations of mineral components. It was found that natural brown coals exhibit a densely crosslinked supramolecular structure with the predominance of molecular-size pores. In the course of decationization, the organic matter underwent partial depolymerization; the rate of diffusion and the accessibility of fragments to solvents with relatively bulky molecules dramatically increased.
Coals are naturally occurring concentrators for many chemical elements (according to various published data, for about 70 elements from the periodic table). The presence of mineral elements in coals is associated with many scientific and technical problems of considerable current interest because of the increasing use of fuel. The concentrations and speciation of mineral components in coals are of interest for solving a number of problems: the applicability of coals as power-plant fuel or a raw material for processing and the extraction of valuable mineral components for practical use and solving environmental problems. Depending on concentration, all mineral components are subdivided into macrocomponents, that is, the main ash-forming substances that occur in coal in considerable amounts, and trace components (or trace elements), which occur in low concentrations (≤0.1%) [1, 2]. In high-rank coals, a considerable portion of mineral components occurs as discrete minerals, which can be partially separated from the organic matter. In low-rank coals (especially, brown coals), only a small portion of the elements (such as silicon, aluminum, and iron) can be the constituents of discrete minerals; the majority of the elements is usually associated with difficult-to-separate clay materials or organic matter. Thus, alkaline earth metal compounds can occur in brown coal as the constituents of organic matter, primarily, as carboxylate derivatives [3–6]. In this state, they can enter into cation-exchange reactions; therefore, some brown coals can be used as inexpensive and efficient cation exchangers [7–9]. It was found that the compounds of alkaline earth metals, among which calcium is usually predominant, can exert a considerable effect of the occurrence of various chemical reactions in coal [10–21]. Thus, they decrease the yields of liquid products in liquefaction and pyrolysis [6, 13, 15, 16, 19]; on the contrary, they have a considerable promoting effect in gasification [11–13, 18, 20].
2
It is believed that mineral components can exert a catalytic effect [5, 12, 20] or affect the supramolecular structure of the organic matter of coal (OMC) [13, 21]. According to published data [15, 16, 21], polyvalent metal cations can serve as crosslinks in the organic matter to hinder its interaction with reagents. The concentrations of many trace components in coals are usually lower than Clarke numbers, that is, the average abundances in the earth’s crust. However, coals from some deposits can exhibit geochemical specialization with respect to particular elements; as a result of this, the concentrations of these elements can be higher than the Clarke numbers by factors of tens or hundreds [22,23]. High metal contents are characteristic of many deposits in the Lena Basin. Anomalously high concentrations of valuable rare earth metals were found in top and bottom coals at the Zhigansk coal deposit of Lena Basin [23]. Considerable concentrations of valuable rare and rare earth elements also occur in brown coals from the Kangalas deposit and northern regions of the Lena Basin [23]. Problems related to the rational use of coals with the simultaneous utilization of mineral components are of importance because of the large-scale development of metal-bearing coal deposits under strict requirements imposed on the environmental safety of coal power engineering and a deficiency of valuable metal reserves [1,24]. Brown coals from the Kansk-Achinsk Basin differ from coals from other large deposits in Russia in a comparatively low total ash content (80% of coals have ash contents lower than 8%, and a low concentration of trace elements, including toxic elements [25]. Large deposits of low-ash brown coals with the concentrations of ash-forming substances as low as 1–2% are concentrated in the Latrobe Valley in Australia; because of this, they are of considerable interest as high-quality raw materials for coal-chemical processes [26]. Studies oriented to the development of various processes for the production of ash-free coal have been in progress in the past few years. Thus, treatment with dilute acid solutions decreased the concentration of mineral matter to a level of 0.2% [27–30]. This coal can find use both as a reactive raw material for conversion into liquid fuels, valuable organic substances, or carbon materials and for efficient and environmentally safe power generation, in particular, at new gas turbine plants [30–35]. In this paper, we report the results of a comparative study of the supramolecular structure of the organic matter of brown coals from various deposits and the interactions with solvents depending on the composition of the mineral constituents. The samples of brown coals from the Borodino, Berezovo, and Aban deposits of the Kansk-Achinsk Basin; the Kangalas deposit of the Lena Basin; and the Yallourn deposit of Latrobe Valley in Australia were used. These brown coals were significantly different in the concentrations of mineral components. EXPERIMENTAL The representative samples of commercial coals from the above deposits were used. Samples were also taken at various sites of the Borodino deposit. To extract ion-exchange cations, particular samples were treated with dilute solutions of HCl at room temperature in an inert atmosphere. The composition of the organic matter of coals was determined using chemical elemental and functional
3
analysis and IR spectroscopy; the concentrations of metals were determined by atomic absorption spectrometry and X-ray fluorescence analysis; the porosities and specific surface areas were determined using mercury porosimetry and nitrogen thermodesorption, respectively. In terms of properties, coal can be classified with amorphous–crystalline polymers having inhomogeneous composition and irregular spatial structure [36, 37]. To characterize the chemical properties and supramolecular structure of coal in its interaction with solvents, we used a method of swelling in organic solvents. Swelling ratios were determined by a volumetric method; lowmolecular-weight aliphatic alcohols, tetrahydrofuran (THF), pyridine, pyridine derivatives, and tetralin were used as solvents. RESULTS AND DISCUSSION Coal Characterization Tables 1 and 2 summarize the compositions of the representative samples of commercial coals from various deposits. The organic matter exhibited similar ultimate analyses, except for Australian coal, which was characterized by a lower carbon content and increased oxygen content (28.8%), including carboxylic oxygen (6.6%). The concentrations of ash-forming substances varied from 1.0 to 7.8% (Table 2). The ash content was as high as 9.5% in particular regions of the Borodino deposit. Among all of the samples taken, Australian coal exhibited the lowest ash content (1.0%). The Australian coal was characterized by low and almost the same Ca and Mg contents, whereas Ca was predominant in Kansk-Achinsk and Kangalas coals (its concentration was higher than that of Mg by a factor of 4–7). Table 1. Composition of the initial samples of brown coals from various deposits Elemental composition, wt % (on a daf basis)
Coal
Borodino Berezovo Aban Kangalas Yallourn Borodino treated with 0.2 N HCl Berezovo treated with 0.2 N HCl Yallourn treated with 0.2 N HCl
Active oxygen, wt % (on daf basis)
C
H
N
S
O
COOH
OH
71.3 70.3 71.3 71.0 65.7 70.5
4.8 4.7 4.8 5.5 4.7 4.9
0.9 0.7 0.9 0.8 0.5 0.8
0.2 0.2 0.2 0.4 0.3 0.2
22.8 24.1 22.8 22.3 28.8 23.6
2.8 2.5 1.9 2.3 6.6 5.8
6.3 8.0 7.9 4.3 8.3 6.5
71.2
4.8
0.8
0.2
24.0
6.0
7.1
66.8
4.7
0.4
0.2
27.9
8.3
8.5
4
Table 2. Concentrations (wt % on a dry basis) of the main ash-forming elements in brown coals from various deposits Deposit
Ad
Al
Ca
Mg
Fe
Borodino Berezovo Aban Kangalas Yallourn
6.2 3.9 5.8 7.8 1.0
0.26 0.18 0.05 0.62 0.10
1.25 1.40 2.10 1.65 0.15
0.30 0.20 0.28 0.30 0.16
0.06 0.50 0.61 0.73 0.40
The concentrations of individual trace elements and the recoveries of these elements with a 0.2 N solution of HCl were determined for the samples taken from the Borodino deposit (Table 3). The concentrations of the alkaline earth metals Sr and Ba, as well as Mn, in the initial samples were somewhat higher than the Clarke numbers for coals (that is, the average abundances in the world’s coals). The concentrations of the other trace elements were much lower than the Clarke numbers for coal, and the concentrations of toxic elements were much lower than the maximum permissible concentrations (MPCs) (for example, the MPC of Mn is 1000 g/t). The above data fall within the range of values observed in coals from the Kansk-Achinsk Basin [25]. The treatment with a 0.2 N solution of HCl decreased the total ash content of coals by a factor of 2– 4. As a result of treatment under the specified conditions, minimum ash contents were 1.5 and 0.4% for Kansk-Achinsk and Yallourn coals, respectively. Note that the cations of alkaline earth metals (Ca, Mg, Sr, and Ba) were almost completely extracted in all cases. The recovery of Fe varied from 50 to 93% (Table 3). The cations of valuable trace elements such as Y and Yb readily passed into solution (by 95% or more). The recovery of potentially toxic elements (Mn, V, and Be) was also sufficiently high. The recovery of Ni was no higher than 12%; this suggests the occurrence of its non-ion-exchangeable form in native inorganic minerals. Table 3. Concentrations of trace elements (g/t) in the samples of coal from the Borodino deposit and recoveries after treatment with a 0.2 N solution of HCl Element Coal Clarke number* Average concentration in KanskAchinsk coals* In analytical samples Recovery of the element, %
Sr
Ba
Mn
V
Be
Ni
Ge
Ga
Sc
Y
Yb
130
120
100
23
2.4
8
1.5
7
–
7
0.9
162
124
130
4.1
0.4
3.5
0.7
1.0
–
1.9
0.3
44– 250 100
96– 204 >97
50– 0.9– <0.04– <0.4– <0.0 <0.2– <0.4 <0.4– <0.2– 500 3.1 0.2 1.2 2 1.5 0.5 0.3 73–97 55–73 >85 0–12 – 65–89 – >95 >95
5
The elemental composition of OMC and the concentration of hydroxyl groups changed only slightly as a result of treatment with an acid solution (Table 1). In this case, the concentrations of free carboxylic acids in Kansk-Achinsk and low-ash Australian coals increased, on the average, by factors of 2–2.5 and 1.3, respectively (Table 1). Figure 1 shows the IR spectra of coal before and after treatment with a dilute solution of HCl. It can be seen that, in decationized coal, the intensity of an absorption band due to free carboxylic acids at 1720 cm–1 increased and the intensities of the bands at 1560 and -1 1410 cm due to the ionic –COO groups of metal carboxylates decreased. This is consistent with conclusion [3-6] that cations extractable into an aqueous solution (among which Ca2+ cations are predominant) mainly occur as carboxylate salt fragments in brown coal. Effect of mineral components of coals on the properties of organic matter in the interaction with solvents Parent coals. The spatial structure of OMC is characterized by the presence of an amount of pores (open and closed); it includes amorphous (disordered) and crystalline (comparatively ordered) regions with different packing densities, fragment flexibilities, and fragment sizes. The specific surface areas of samples were as low as 1–3 m2/g, as measured using the thermal desorption of nitrogen. These values suggest a low open porosity with a pore size greater than 0.38 nm. According to mercury porosimetry data, the pore structure exhibited a bimodal pore-size distribution (Fig. 2). Peaks in a curve occurred in the regions of mesopores with radii of 10–20 nm and macropores with a broad maximum at 700–800 nm. Aban coal with an increased Ca content was Fig.2 Pore size edistributions for coal samples from prominent in terms of pore structure; it the (1) Aban, (2) Berezovo, (3) Borodino, and (4) Yallourn deposits [38]. contained a larger amount of well-
6
organized mesopores and a smaller amount of macropores. Australian coal with a low concentration of ashforming substances, including Ca, was characterized by a smaller porosity and a broader pore-size distribution. We experimentally found that all of the brown coals efficiently interacted with methanol and ethanol. Equilibration times were 15 and 30– 40 min in methanol and ethanol, respectively (Fig. 3). The equilibrium swelling ratios in ethanol for various coal samples (other than Australian one) fall within a comparatively narrow range with the average value of Qethanol = 1.60 ± 0.10. Australian Fig.3. Kinetics of swelling of brown coal in (1) ethanol, coal exhibited a higher swelling ratio (2) methanol, (3) n-propanol, and (4) isopropanol. (1.85); this was likely due to a higher oxygen content of this coal. The interaction with n-propanol, whose chemical properties (solubility parameters are 12.7 and 11.6 cal1/2/cm3/2, respectively; see Table 4) and a kinetic molecular diameter (0.49 nm) are similar to those of ethanol but the molar volume is greater, occurred much more slowly.
7
Isopropanol (a secondary alcohol) with almost the same chemical properties and molar volume as in n-propanol but with a greater kinetic diameter (0.56 nm) weakly interacted with coal. These results are consistent with published data on the swelling of other coals in alcohols [39, 40]. Thus, for subbituminous coal from Illinois, the initial rate of swelling in methanol was higher than that in n-butanol or isobutanol by a factor of 14 or 64, respectively [40]. As an example, Table 4 compares the swelling characteristics of a Borodino coal sample with ash and Ca contents of 7.2 and 1.8%, respectively, in various solvents. Along with swelling ratios, the numbers of absorbed molecules (N) calculated from these ratios are also given; these numbers were determined taking into account the molar volume V of a solvent (using the equation N = (Q – 1)/V). The value of N reflects the number of ruptured intermolecular bonds in OMC; this value characterizes the effects of solvents with different molecular sizes more adequately than the swelling ratio. It can be seen that almost the same numbers of moles were absorbed in methanol and ethanol media (11–12 mmol/(cm3 coal)). The amount of absorbed n-propanol was much smaller (N = 8 mmol/cm3). In the case of isopropanol, the value of N was as low as 3 mmol/cm3; the same numbers of more bulkier n-butanol and cyclohexanol molecules were absorbed. Of the solvents used, pyridine most efficiently interacted with coal. However, even a small methyl substituent in the pyridine ring dramatically decreased the ability to interact with coal. In benzylpyridine and tetralin, coal weakly swelled even as the temperature was increased to 250°C. Swelling in THF was much weaker than in pyridine, although their molecular sizes are similar. For coal samples from various deposits, swelling ratios in THF changed from 1.41 to 2.01 depending on Ca content. From a comparison between experimental data, it follows that the swelling of natural brown coals in the majority of solvents considerably depends on pore sizes in organic matter. Only small molecules of methanol and ethanol can readily penetrate into the bulk, interrupt molecular associations (mainly through hydrogen bonds), and cause swelling. Steric effects were pronounced even in a comparatively small linear molecule of n-propanol, whose critical diameter is 0.49 nm. From the swelling data, it follows that 70% bonds are accessible to this molecule, whereas only 25% bonds, which are cleaved by methanol and ethanol, are accessible to other alcohol molecules. Thus, the critical pore size in the starting brown coal is no greater than 0.5 nm. Decationized coals efficiently interacted with n-propanol, isopropanol, n-butanol, and cyclohexanol. The amount of absorbed alcohol molecules was almost the same (N = 11–12 mmol/cm3) regardless of molecular size; this fact suggests the absence of noticeable steric hindrances and the occurrence of specific interactions in the system. Swelling in pyridine also considerably increased, and swelling in 4-methylpyridine dramatically increased. This means that partial decation-ization leads to an increase in the size of transport pores in OMC to provide a greater accessibility of its fragments to solvent molecules. At room temperature, steric hindrances occur only in comparatively large bicyclic molecules of benzylpyridine and tetralin; that is, the critical pore size increased from 0.5 to 0.65 nm after decationization.
8
Fig.4. Kinetics of THF swelling of (1) the parent Kansk-Achinsk coal and of the treated with HCl solutions with different concentrations: (2) treated with water with no HCl; (3) 0.1 HCl; (4) 0.2 HCl; (5) 1.0 N HCl.
Fig.5. Change in the swelling ratio (1) of the parent coals and (2) of the decationized and Ca-exchanged coals in THF as a function of Ca concentration
Nevertheless, unlike the initial coal, which weakly swelled in tetralin even at 250°C, the coal after decationization began to swell at 150°C with the absorption of 5 mmol/cm3 (Table 4). This suggests that decationization labilized a supramolecular structure, which can be activated by a weak thermal action. Figure 4 shows the dynamics of swelling in THF for samples treated with HCl solutions with different concentrations. It can be seen that both the maximum swelling ratio and the rate of swelling increased as the acid concentration was increased (that is, as the residual calcium content of coal decreased). The equilibration time was 30–50 min for deeply decationized coal (1 N solution of HCl), whereas one to three days were required for the starting coal. Figure 5 shows the dependence of the equilibrium swelling ratios for various initial and decationized samples in THF on calcium concentrations in these samples. It can be seen that the swelling ratio nonlinearly increased from 1.4 to 2.25-2.65 as the Ca content decreased from 2.1 to 0.02-0.006%. In a regular series of samples obtained from the same parent coal, a strictly linear function of the following form was observed (Fig. 5, insert) with the correlation coefficient R = -0.99:
Analogous behaviors took place in the pyridine-coal system. The quantitative relationships found suggest that the swelling of brown coal in THF is con-
9
lg(Qt-1)/(Qeq-1)
trolled by metal-carboxylate groups, among which Ca carboxylates are predominant. In the OMC, they play a role of strong ionic crosslinks that facilitate the formation of a dense spatial structure; thus, they are responsible for its ability to interact with solvents. An analysis of the kinetics of swelling allowed us to obtain data on the mechanism of solvent molecule penetration into the bulk of OMC and on the mobility of its structural fragments. Usually, two the most commonly used models accepted for polymers are considered [41, 42]. Both of the models are based on the assumption that the stages of chemical interactions between solvent molecules and polymer fragments occur rapidly; kinetically, the process of swelling is controlled by transport of solvent molecules in the bulk of the polymer. In the first model, it is assumed that the rate of transport limits the ability of polymer chains to relaxation. That is, the rate of swelling depends on the relaxation constant of the polymer. The possibility of transport of this kind (hole or zone transfer) depends on the occurrence of a free volume in the polymer; this free volume depends on structure. In the second model, Fickian mode of solvent diffusion through polymer is a ratelimiting stage; that is, the process of swelling is determined by an apparent diffusion coefficient. The structure of OMC is more complicated than that of polymers. The OMC is characterized by a porosity; therefore, diffusion through open pores can occur in it as in any porous body. The above types of diffusion can be recognized based on the exponent n in the generalized kinetic expression [42, 43] (Qt –1)/(Qeq –1) = ktn, where Qt and Qeq are the current and equilibrium swelling ratios, respectively; k is the rate constant; and t is time. The exponent n is determined from the slope of an experimental straight line at the initial portion of the dependence of log[ (Qt – 1)/(Qeq – 1) on logt . Its numerical value characterizes the mechanism of transport: n = 0.85–1.0 if the transport is limited by relaxation chain vibrations, n = 0.45–0.50 in the case of Fickian mode transport with definite diffusion coefficient. In Figure 6, the kinetic dependances in coordinates of log[ (Qt – 1)/(Qeq – 1) upon logt are shown for THF swelling of different coals as the examples. The calculated values of n are given in Table 5. n=0,85 n=0,41 It was found that that the exponent n was 0.83 in the 2 3 -0,2 swelling of the parent brown coal in ethanol (small n=0,22 molecules); it somewhat increased after decationization -0,4 1 (to 0.95, Table 5). This fact suggests that the interaction -0,6 of ethanol with both parent and decationized brown coal depends on the ability of OMC fragments to undergo -0,8 relaxation vibrations. In n-propanol, the exponent n was 1,0 1,5 2,0 2,5 log 0.57 and 0.85 for the parent and decationized brown coal, Time, min respectively. Thus, in the former case (parent coal), Fig. 6. Kinetics of THF swelling of swelling results from both types of diffusion (that is, Fick the parent Borodino coals with and relaxation diffusion), whereas it results from only (1) 1.34%Ca, (2) 0.75 % Ca, and relaxation diffusion in the latter case (decationized coal). (3) its decationized sample.
10
Table 5. Exponent n in rate equations for the swelling of parent and decationized brown coals Solvent
Parent coal
Decationized coal
Ethanol n-Propanol n-Butanol THF Pyridine Methylpyridine
0.83 0.57 – 0.19–0.46 0.67 –
0.95 0.85 0.88 0.81–0.90 0.90 0.92
n
The change of the mechanism of diffusion in the parent coal on going from ethanol to npropanol with close chemical properties and critical molecular diameters is due to the fact that a large-volume cavity is required for the relaxation diffusion of a larger molecule. It is well known [44] that the probability of formation of a cavity with a diameter equal to or greater than a molecular volume is proportional to exp(–γVm/Vf), where Vm is the molecular volume, Vf is the free volume of the polymer, and γ is the hole overlap factor; that is, the probability of the event dramatically decreases with molecular volume. Kinetic regularities of this kind also occurred in the swelling of the parent coal in pyridine (n = 0.67) and methylpyridine. In decationized coal, organic matter is considerably depolymerized; because of this, it has fragments that are more flexible and a larger free volume. This increases the probability of formation of a large cavity near a diffusing molecule. As a result, molecular transport in decationized coal occurs by the relaxation mechanism. More complicated relations were observed in the coal–THF system, where the exponent n varied over a wide range from 0.19 to 1.0 depending on the concentration of Ca (Fig. 6 and 7). For initial coals with low concentrations of alkaline earth cations (0.41–0.75%), n = 0.41–0.46, which is close to values characteristic of the process limited by Fick diffusion. For samples with higher Ca 1,0 concentrations, anomalously low values of n = Decationized Yallourn 0.19–0.30 were obtained, which are inconsistent Kansk-Achinsk 0,8 Kangalas with the above models. A possible reason for the low rate of swelling 0,6 of the parent coal in THF and a deviation from the Fick kinetics can be associated with the densely 0,4 cross-linked amorphous–crystalline structure of OMC with slit-shaped pores of molecular sizes. The 0,2 diffusion of rigid cyclic THF molecules in this 0,0 0,5 1,0 1,5 system can occur in an amorphous part through Ca concentration, wt.% activated forms, in which the apparent diffusion Fig.7. Exponent n in the rate equation of coefficient strongly depends on the size and THF swelling of different coals from configuration of pores. During penetration coal various deposits as a function of Ca matter, diffusion coefficient can decrease because concentration. of narrowing of the pores and geometric resistance
11
due to the necessity of passing round crystalline regions and diffusing only through the amorphous part of OMC. In all of the decationized samples (regardless of the Ca content of the parent coal), the exponent n varied within the range of 0.80– 0.92; these values are characteristic of the relaxation transport model. This means that the fragments of organic matter in decationized coal can undergo conformational changes in the interaction with solvent molecules; OMC can exhibit elastic properties to some extent. CONCLUSIONS (1) A considerable portion of mineral components in brown coals from the deposits of the Kansk-Achinsk Basin, the Kangalas deposit, and the Yallourn deposit in Australia occurs as the carboxylate constituents of organic matter in a cation-exchangeable form. The presence of these structural units facilitates the formation of a densely crosslinked supramolecular structure of organic matter with molecular-size pores. Ionic crosslinks with the participation of alkaline earth cations, among which calcium cations are predominant, make a crucial contribution. Other polyvalent cations, including trace element cations, can also play the role of ionic crosslinks. A wide variety of these elements in brown coal can be responsible for a noticeable total contribution to the crosslinking and strengthening of the supramolecular structure, although the concentrations of particular elements are low. (2) The interaction of solvents with the fragments of the organic matter of natural brown coals depends on both the chemical properties of the solvents and, largely, the sizes of solvent molecules. Only small methanol and ethanol molecules can penetrate into the bulk of OMC and degrade intermolecular association (mainly by hydrogen bonds). The steric effects of the OMC structure are pronounced even with a comparatively small linear molecule of n-propanol. The interaction of OMC with larger molecules more strongly depends on the Ca content. At a low content, the transport of THF molecules can occur by Fick diffusion. In coals with a densely crosslinked structure (with a concentration of Ca higher than 0.75%), the slow transport of rigid cyclic THF molecules can occur through pseudo-fickian activated diffusion forms, in which an apparent diffusion coefficient depends on the size and configuration of pores, and can decrease because of geometric resistance due to the necessity of passing round crystalline regions and diffusing only through the amorphous part of OMC. (3) In the course of decationization, the supramolecular structure undergoes principal changes. The partial depolymerization of organic matter occurs, and this organic matter acquires to some extent elastic properties under the action of the majority of solvents. In decationized coal, the rate of diffusion and the accessibility of fragments to solvents with relatively bulky molecules dramatically increase. Solvent molecules diffuse through holes formed in the relaxation vibrations of comparatively flexible chains. Acknowledgements. This work was supported in part by the Siberian Branch of the Russian Academy of Sciences (Integration Program no. No 106).
12
REFERENCES 1. Shpirt, M.Ya., Kler, V.R., and Pertsikov, I.Z., Neorga-nicheskie komponenty tverdykh topliv (Inorganic Components of Solid Fuels), Moscow: Khimiya, 1990. 2. Yurovskii, A.Z., Mineral’nye komponenty tverdykh goryuchikh iskopaemykh (Mineral Components of Solid Fossil Fuels), Moscow: Nedra, 1968. 3. Loskutova, E.N., German, N.M., Bochkareva, K.I., and Shiryaeva, K.N., Piroliz burykh uglei (Pyrolysis of Brown Coals), Novosibirsk: Nauka, 1968. 4. Birgauz, R.Ya. and Kukharenko, T.A., in Khimiya i pere-rabotka topliva (Fuel Chemistry and Processing), Moscow: Izd. IGI, 1971, vol. 27, issue 1, p. 3. 5. Herman, R.G., Simmons, G.W., Cole, D.A., et al., Fuel, 1984, p. 673. 6. Kuznetsov, P.N., Khim. Tverd. Topl. (Moscow), 1998, no. 3, p. 53. 7. Golovin, G.S. and Maloletnev, A.S., Kompleksnaya pere-rabotka uglei i povyshenie effektivnosti ikh ispol’zovaniya (Integrated Processing of Coals and Increasing the Efficiency of Coal Utilization), Moscow: NTK Trek, 2007. 8. Golovin, G.S., Lesnikova, E.B., Artemova, N.I., and Dement’eva, O.A., Khim. Tverd. Topl. (Moscow), 1996, no. 5, p. 26. 9. Butler, C.J., Green, A.M., and Chaffee, A.L., Fuel, 2007, vol. 86, p. 1130. 10. Zekel’, L.A. and Shpirt, M.Ya., Khim. Tverd. Topl. (Moscow), 2005, no. 1, p. 91. 11. Miura, K., Hashimoto, K., and Silveston, P.L., Fuel, 1989, p. 1461. 12. Hashimoto, K., Miura, K., and Ueda, T., Fuel, 1986, vol. 65, p. 1516. 13. Barthe, P., Charcosset, H., and Guet, J.M., Fuel, 1986, vol. 65, p. 1330. 14. Wornat, M.J. and Nelson, P.F., Energy and Fuels, 1992, vol. 6, p. 136. 15. Kuznetsov, P.N., Kuznetsova, L.I., Bimer, J., et al., Fuel, 1997, vol. 76, p. 189. 16. Mochida, I., Sakata, R., and Sakanishi, K., Fuel, 1989, vol. 68, p. 3. 17. Aleksandrov, I.V. and Kamneva, A.I., Khim. Tverd. Topl. (Moscow), 1976, no. 2, p. 90. 18. Kuznetsov, P.N., Kuznetsova, L.I., and Kutikhina, E.A., Khim. Tverd. Topl. (Moscow), 2008, no. 3, p. 36. 19. Murakami, K., Shirato, H., Hanada, N., and Nish-iyama, Y., Energy and Fuels, 1998, vol. 12, p. 843. 20. Linares-Solano, A., Hippo, E.J., and Walker, Ph.L., Fuel, 1986, vol. 65, p. 776. 21. Wornat, M.J. and Sakurovs, R., Fuel, 1996, vol. 75, p. 867. 22. Nifontov, B.F., Potapov, V.P., and Mitina, N.V., Geokhimiya i otsenka resursov redkozemel’nykh ele-mentov v kuznetskikh uglyakh. Perspektivy pererabotki (Geochemistry and Rare Earth Element Resource Evaluation in Kuznetsk Coals: Prospects for Processing), Kemerovo: IUUKh SO RAN, 2003. 23. Kashirtsev, V.A., Zueva, I.N., Suknev, V.S., et al., Otechestvennaya Geologiya, 1999, no. 4, p. 65. 24. Pashkov, G.L., Soros. Obshcheobrazovat. Zh., vol. 7, no. 11, p. 67.
13
25. Gavrilin, K.V. and Ozerskii, A.Yu., Kansko-Achinskii ugol’nyi bassein (Kansk-Achinsk Coal Basin), Moscow: Nedra, 1996. 26. Li, C.-Z., Fuel, 2007, vol. 86, P. 1664. 27. Serferinoglu, M., Paul, M., Sandstrom, A., et al., Fuel, 2003, vol. 82, p. 1721. 28. Couch, G.R., Advanced Coal Cleaning Technology, London: IEA, Coal Research, 1991. 29. Steel, K.M. and Patrick, J.W., Fuel, 2001, vol. 80, p. 2019. 30. Li, Ch., Takanohashi, T., Yoshida, T., et al., Fuel, 2004, vol. 83, p. 727. 31. Rubiera, F., Arenillas, A., Arias, B., et al., Fuel, 2003, vol. 82, p. 2145. 32. Wang, J., Li, C., Sakanishi, K., Nakazato, T., et al., Fuel, 2005, vol. 84, p. 1. 33. Miura, K., Shimada, M., Mae, K., and Sock, H.Y., Fuel, 2001, vol. 80, p. 1573. 34. Yoshida, T., Takanohasi, T., Sakanishi, K., Saito, I., et al., Fuel, 2002, vol. 81, p. 1463. 35. Domazetis, G., Barila, P., James, B.D., and Glaisher, R., Fuel Proc. Technol., 2008, vol. 89, p. 68. 36. Krichko, A.A., Gagarin, S.G., and Makar’ev, S.S., Khim. Tverd. Topl. (Moscow), 1993, no. 6, p. 27. 37. Skripchenko, G.B., Ross. Khim. Zh., 1994, vol. 38, no. 5, p. 27. 38. Kageyama Y., Kaneko T, Tazawa K., Okuyama N., Saito I., Kuznetsov P.N. Chemistry for Sustainable Development, 1996, Vоl. 4, № 1, p. 37. 39. Pope, Ch.G., Fuel, 1984, vol. 63, p. 1681. 40. Aida, T., Fuku, K., Fujii, M., et al., Energy Fuels, 1991, vol. 5, p. 79. 41. Tager, A.A., Fizikokhimiya polimerov (Physical Chemistry of Polymers), Moscow: Khimiya, 1978. 42. Hall, P.J., Thomas, K.M., and Marsh, H., Fuel, 1992, vol. 71, p. 1271. 43. Otake, Y. and Suuberg, E.M., Fuel, 1989, vol. 68, p. 1609. 44. Denisov, E.T., Okislenie i destruktsiya karbotsepnykh polimerov (Oxidation and Degradation of Carbon-Chain Polymers), Leningrad, 1990.
A Study to Recover Coal from Turkish Lignite Fine Coal Tailings: Comparison of Falcon Concentrator and Multi Gravity Separator (MGS)
M. Fatih CAN, Selçuk ÖZGEN, Eyüp SABAH*, Department of Mining Engineering, Afyon Kocatepe University, 03200 Afyonkarahisar, Turkey
ABSTRACT Lignite coal is the primary domestic source of energy in Turkey, for this reason effective exploitation of the reserves of Turkey is very crucial. In Turkey the fine tailings of lignite coal processing plants are sent in most cases to the tailing ponds without any treatment. However, recovery of fine coals from coal preparation tailings and recycle of processing water are of both economic and environmental incentives, not only preserving natural resources but also reducing environmental consequences of discharging large volume of tailings. Recent developments in the use of various gravity equipments in fine-coal beneficiation have been discussed and their relative merits have been compared. In this study, the possibility of beneficiation of lignite tailings included quartz, kaolinite, siderite, mica/illite, dolomite, feldspar compounds in the Tunçbilek/Kütahya region was investigated by Multi Gravity Separator (MGS) and Falcon Concentrator and these two methods compared. The entire exercise revealed that the MGS could produce a clean coal with an ash content of 22.83%, 5696 kCal/kg calorific value and a recovery of 49.32% and that the Falcon could produce a clean coal with an ash content of 40.26%, 4224 kCal/kg calorific value and a recovery of 64.53% from a feed coal having an ash content of 66.21% and 1835 kCal/kg calorific value.
Key words: Fine lignite coal tailings, Waste processing, Hydrocyclone, Falcon Concentrator, Multi Gravity Separator (MGS)
*
Corresponding author: Prof. Dr. Eyüp Sabah
Afyon Kocatepe University Department of Mining Engineering 03200 AFYONKARAHİSAR e-mail: [email protected] Phone: +90-272-228-1423 / 1320 Fax: +90-272-228-1422
1. INTRODUCTION Coal is a fossil fuel which is the most abundant and geographically scattered all around the world. In world general 826 billion tones apparent reserve of coal is estimated. In other words, this means the world has coal reserves those can sustain more than 119 years [BP Statistical Rewiev, 2009]. However coal has this much long consume period, apparent oil and natural gas reserves consumption period is estimated between 40 to 60 years according to current production rates [Dağdelen, 2006]. For this reason coal will be one of the most important energy resources of the future.
As the demand on clean coal increase with developing environmental standarts, an increase of modern coal preparation plant numbers is observed in latest years. Modern coal preparation plants are consisting of coarse, medium and fine coal cleaning circuits. Fine coal has size distribution below 28 meshes (0.589 mm) and designated as slime in coal cleaning circuits. Generally, this size fractioned material must be processed in all coal preparation plants. In past year’s coals of below 28 meshes (0.589 mm) size discharge was economic, but with swiftly developing technologies fine coals are also utilized. Also today’s mining processes produces more fine materials as the mechanization applications have more frequently applied. Therefore it is a must to recycle the coal in fines. In this manner recovered fine coals are prevented to deploy in waste and avoid a series of problem in waste disposals.
Today 5% flotation and 95% gravimetric processes are used to prepare the produced coal all around the world. Gravimetric processes applied for coarse coal preparation have low cost, high capacity and selectivity. Gravimetric processes are also utilized for cleaning fine coal. However, some other problems arouse within the used separators. Only with developing technologies, some advanced gravity separation equipments as hydrocylones, Multi Gravity Separator (MGS), Falcon and Knelson concentrator, Kelsey Jig are used for ultra fine coal enrichment. There are many conducted researches with successful results on enrichment of coals [Honaker, 1995; Honaker et al., 1996; Honaker, 1998; Honaker, 2000; Parekh and Khalek, 2002; Majumder et al., 2007a; Majumder and Barnwal, 2008; Malkoç, 2008]. Yıldırım et. al. [1995], revealed clean coal with 75.8% efficiency and 7.01% ash content from 36.6% ash bearing run of mine Zonguldak coal by applying MGS equipment. Aslan et. al. [1999] examined enrichment of Yeniçubuk-Gemerek lignite coal with MGS and gained clean coal with 79.82% efficiency and 19.23% ash content from lignite coal of 35.75% ash content. In another research with MGS equipment for Manisa-Soma hard lignite coal which has 35.9% ash content was enriched and clean coal with 79.5% recovery efficiency and 18% ash content was obtained [Çiçek et al., 2002]. Abd-Elrahiem, [2003] utilize carrier column flotation to enrich very fine coal of 10.13% ash content, and as result obtained clean coal with 79.94% efficiency and 5.15% ash content. Li et. al. [2003] used cyclo-micro bubble column to clean 47.11% ash containing coal slime and finally obtained clean coal with 79.26% combustible recovery and 10.55% ash content. Majumder et. al. [2007b] was made to study the effects of different process variables on the performance of an MGS for the beneficiation of coal fines. The MGS is capable of producing clean coal having an ash content of 14.67% with a 71.23% yield (the float-sink yield was 72% at the same ash level) from a feed coal ash of 24.61%.
In this study, the aim is to recover -0.5 mm coal from Tunçbilek/Kütahya coal plant tailings and investigate the effect of MGS and Falcon concentrator parameters on coal enrichment. Ash content and coal recovery rates are designated as performance criteria.
2. MATERIALS AND METHODS 2.1 Materials Lignite coal tailings obtained from Tunçbilek Coal Preparation Plant of G.L.I of Turkish Coal Enterprises (Kütahya-Turkey) were used in this study. The samples were taken from slurry waste with standard of TS ISO 5667-10.
2.2 Methods 2.2.1. Characterization Tests A number of qualitative and quantitative analysis techniques were used to characterize the coal tailings. The chemical composition of the tailings was defined by X-ray fluorescence (XRF). The mineral composition of the tailings was determined by X-ray diffraction (XRD) method using a Rigaku-Giger Flex analyzer. The particle size distribution of the tailings was obtained using a Retsch AS200 Sieve Shaker and Fritsch-Analysette 22 Particle Size Analyzer. The specific gravity of the tailings was determined by Quantachrome Ultrapycnometer 1000. The ash and sulfur contents of the tailings were determined according to ISO 1171 and ISO 351, respectively. In addition, the calorific value was determined based on ISO 1928. Following the characterization tests, the beneficiation studies were immediately initiated before any physical and chemical decomposition of tailings.
2.2.2. Test Procedure Before MGS and Falcon Concentrator experiments, classification tests were performed using a hydrocyclone. The aim of this test was to separate the clay and/or carbonate minerals from the coal. Since large amounts of fine and ultra fine particles in the tailings cause a low cut point, a small diameter hydrocyclone (44 mm) was selected for the experiments. Underflow products were concentrated using a Multi-Gravity separator (MGS) or Falcon Concentrator. For each MGS and Falcon Concentrator tests, 2000 g of the dry coal sample was used. The effects of varying operating parameters such as drum speed, tilt angle, shaking amplitude, wash water rate, pulp feed rate and pulp solid ratio was investigated for MGS. The shaking frequency of the MGS was fixed at 4.9 cps for all the experiments. To obtain required solids ratio in feed, measured quantities of solids and water were mixed in the slurry tank. The MGS variables were adjusted at necessary levels. The feed slurry was pumped into the MGS drum at the required flow rate using the peristaltic pump while the MGS was in operation. Samples from the clean coal and tailing streams were collected at steady-state condition. The samples were filtered, dried and analyzed for ash content and combustible recovery. When changing a parameter in each test, the other parameters were kept constant and the optimization results were used in the other tests. However, a laboratory-scale Falcon SB-40 concentrator was used for the tests. Once the optimum hydrocyclone conditions were identified and applied, the concentrate was further processed in the Falcon concentrator. Ash rejection was achieved by optimizing variables such as gravity force, water pressure, pulp solids ratio and feed rate. The classification and concentration test results were analyzed with Minitab for MGS and Falcon Concentrator. The model included a set of equations relating the variables, ash content and combustible recovery in the clean coal.
3. Result and Discussion 3.1 Characteristics of Tailings The results of chemical analysis for the samples were presented in Table 1. As seen in Table 1, LOI of the samples are about 37% of Tunçbilek which means that the amount of organic material (coal) in the tailings was significant when compared to the amount of inorganic impurities. When Figure 1 was examined, - 80% of the Tunçbilek coal tailing was -28μm.
Table 1. Chemical analysis of coal tailings. Composition Weight (%)
Cumulative Distrubition, %
SiO2 Al2O3 Fe2O3 MgO CaO Na2O K2O TiO2 P2O5 MnO LOI
35.53 13.34 5.89 3.68 1.53 <0.11 1.23 0.50 0.10 0.13 37.60
1 10
2 20
4 30
Particle Size, µm
Figure 1. Particle size distribution of coal tailings.
Table 2 presents the ash of the coal samples with respect to particle size. Figure 1 shows the particle size distribution of tailings. The Tunçbilek tailings’ under 4 m size particles contained 37% clay minerals, for the sample containing particles between 4 and 63 m in size, 59% accounted for silt minerals, and finally for particles above 63 m 4% accounted for sand size (Wentworth, 1922).
Table 2. The ash content via different particle size of original coal wastes. Particle Size (µm) +425 -425+300 -300+106 -106+75 -75+38 -38 Total
Ash Content (%) 46.88 46.79 46.70 51.13 56.71 77.43 66.60
XRD analysis results of the plant tailings are indicated in Figure 2. As seen Figure 2, the main minerals of Tunçbilek are kaolinite, illite, mica and others. The qualitative and quantitative analysis results of tailings are presented in Table 3.
Figure 2. XRD diagram of Tunçbilek coal tailings.
Table 3. Main properties of tailings. Parameters
Properties
Specific gravity, gr/cm3 Solids content, % LOI, % Particle size, m Percentage of clay size particles Percentage of silt size particles Percentage of sand size particles 80% under size Total ash content, % Total sulphur, % Calorific value, kcal/kg
1.88 7.20 37.60 -500 37 59 4 28 m 66.40 1.38 1835
3.2 Studies on MGS with Tunçbilek Coal Tailings The variables and results regarding the obtained MGS final beneficiation tests with the Tunçbilek coal tailings were given respectively in Table 4. In these tests regression analysis was performed to determine the relationship between 6 variables and 2 response functions.
From the experimental results listed in Table 4 and the response functions representing the ash content of clean coal and the recovery of clean coal could be expressed as functions of the pulp solid ratio (s), drum speed (r), tilt angle (˚), shaking amplitude (sh), wash water rate (w), feed rate (f).
The relationship between response (ash content and recovery of clean coal) and variables were obtained with Minitab 15 for coded unit as follows:
Ash Content of the clean coal on MGS for Tunçbilek: (Ash)1 = 65.9-0.529*x1-0.233*x2+2.27*x3+0.824*x4+1.81*x5-1.53*x6
(1)
Recovery of the clean coal on MGS for Tunçbilek: (Recovery)1 = 181-2.89*x1-0,633*x2+2.30*x3+1.91*x4+8.04*x5-3.79*x6
(2)
Table 4. Level of variable and MGS test results for Tunçbilek coal tailings. Exp. No 1 2 3 4 5 6 7 8 9 10 11 12 13
x1 (s), % 15 15 15 15 15 15 15 10 20 15 15 15 15
x2 (r), rpm 229 229 229 229 229 170 201 201 201 201 201 201 201
Level of variables x3 (a), x4 (sh), x5 (w), ˚ mm lpm 2 10 1 2 10 3 2 10 5 0 10 3 4 10 3 2 10 3 2 10 3 2 10 3 2 10 3 2 15 3 2 20 3 2 10 3 2 10 3
x6 (f), lpm 1 1 1 1 1 1 1 1 1 1 1 2 3
Observe results Ash Content Recovery (%) (%) 19,21 23,98 20,86 41,72 26,44 56,12 16,75 29,12 25,81 38,31 37,68 78,57 19,72 44,38 32,48 70,28 27,19 41,34 31,47 63,22 36,18 74,18 22,83 49,32 26,37 47,79
3.3 Studies on Falcon Concentrator with Tunçbilek Coal Tailings Likewise, the variables and results regarding the obtained Falcon Concentrator final beneficiation tests with the Tunçbilek coal tailings were given respectively in Table 5. In these tests regression analysis was performed to determine the relationship between 4 variables and 2 response functions.
From the experimental results listed in Table 5 and the response functions representing the ash content of clean coal and the recovery of clean coal could be expressed as functions of the gravity force (x7), solids rate (x8), flow rate (x9), water pressure (x10).
The relationship between response (ash content and recovery of clean coal) and variables were obtained with Minitab 15 for coded unit as follows:
Ash Content of the clean coal on Falcon Concentrator for Tunçbilek: (Ash)2 = 38.8 – 0.00923*x7+ 0.047*x8 + 3.55*x9 + 1.97*x10
(3)
Recovery of the clean coal on Falcon Concentrator for Tunçbilek: (Recovery)2 = 81.0 – 0.0131*x7 + 0.117*x8 + 4.41*x9 + 5.86*x10
(4)
Table 5. Falcon concentrator parameters and test results. Level of variables Test No
X7 (gf), g
X8 (sr), %
X9 (fr), lpm
X10 (wp), psi
1 2 3 4 5 6 7 8 9 10
300 300 20 78 176 300 300 300 300 300
10 20 20 20 20 20 20 20 20 30
1 0.5 1 1 1 1 1 1.5 1 1
0.5 0.5 0.5 0.5 0.5 0.5 1 0.5 1.5 0.5
Observed results Ash Content, Recovery, % % 42.64 87.03 41.02 86.36 44.03 90.17 43.86 89.57 42.07 88.99 40.26 78.9 42.69 91.48 44.57 90.77 44.33 91.69 43.59 89.38
3.2 Modelling of MGS and Falcon Concentrator Eqs. (1-4) were derived from Tables 4 and 5 for the response factors, from which the response factors at any regime in the interval of our experiments can be calculated from these equations. The observed (experimental) results and the predicted values obtained using these model equations are given in Tables 6 and 7 and Figs. 3-4. The predicted values and the observed data points, indicating a very good fit (R2 value of 0.807 for ash content of the clean coal, R2 value of 0.944 for recovery of the clean coal for MGS) of the response equations. However, the match of predicted values with the actual data points indicates a poor fit (R2 value of 0.577 and 0.399 for ash and recovery values, respectively) of the equation for Falcon Concentrator.
In the final beneficiation tests that performed in MGS, optimum result is dependent test number 7 when took in consideration the ash contents and recovery of clean coal. The ash content of the clean coal is 19.72% and the coal recovery 44.38%. However, the coal that has the lowest ash content with 16.75% in test number 4; the highest recovery with 78.57% in test
number 6 were obtained. In the final beneficiation tests with the Tunçbilek coal tailings that performed in Falcon Concentrator, optimum result is dependent test number 6 when took in consideration the ash contents of clean coal and recovery of coal. The ash content of the clean coal is 40.26% and the coal recovery 35.04%. Namely at test 6 the lowest ash content of coal and the highest recovery with 91.69% in test number 9 were obtained.
Table 6. Observed and predicted values for ash content and combustible recovery of clean coal on MGS for Tunçbilek coal tailings Exp. No 1 2 3 4 5 6 7 8 9 10 11 12 13
Ash Content (%) Observed Predicted 19,21 17,67 20,86 21,29 26,44 24,91 16,75 16,75 25,81 25,83 37,68 35,04 19,72 27,81 32,48 30,46 27,19 25,17 31,47 31,93 36,18 36,05 22,83 26,28 26,37 24,75
Recovery (%) Observed Predicted 23,98 20,64 41,72 36,72 56,12 52,80 29,12 32,12 38,31 41,32 78,57 74,07 44,38 54,45 70,28 68,90 41,34 40,00 63,22 64,00 74,18 73,55 49,32 50,66 47,79 46,87
Table 7. Observed and predicted values for ash content and combustible recovery of clean coal on Falcon Concentrator for Tunçbilek coal tailings Exp. No 1 2 3 4 5 6 7 8 9 10
Ash Content (%) Observed Predicted 42.64 41.04 41.02 39.73 44.03 44.09 43.86 43.55 42.07 42.65 40.26 41.51 42.69 42.49 44.57 43.28 44.33 43.47 43.59 41.97
Recovery (%) Observed Predicted 87.03 85.58 86.36 84.54 90.17 90.42 89.57 89.66 88.99 88.37 78.90 86.75 91.48 89.68 90.77 88.95 91.69 92.61 89.38 87.92
Multi Gravity Separator Predicted Ash Content, %
Falcon Concentrator y = 0.7039x + 12.176 R² = 0.5771
y = 0.8069x + 5.1665 R² = 0.8072
Observed Ash Content, %
Figure 3. Predicted versus observed response of ash content on MGS and Falcon Concentrator
Predicted Recovery, %
for Tunçbilek Coal Tailings.
y = 0.3988x + 53.181 Multi Gravity Seperator y = 0.3988x + 53.181 R² = 0.3987 Falcon Concentrator R² = 0.3987
y = 0.9438x + 2.673 R² = 0.9438
Observed Recovery, %
Figure 4. Predicted versus observed response of combustible recovery on MGS and Falcon Concentrator Tunçbilek Coal Tailings.
4. CONCLUSIONS In this paper, it is aimed to gain fine coal particles from tailings by enhanced gravity methods after classification with hydrocyclone. Two different equipments (MGS and Falcon Concentrator) have been used and the beneficiation results have been compared.
After -0.5 mm in size the tailings of Tunçbilek coal preparation facility which their ash contents are 66.21% were performed pre-enrichment operating; they were enriched by using 6 various variables. The ranges of values of variables of hydrocyclone used in the design were; 44mm hydrocyclone diameter, apex diameter 3.2 mm, vortex diameter 14.3 mm, feed solid 5%, feed pressure 0.5 psi. The ranges of values of variables of MGS used in the design were shaking frequency (4.9 cps), feed solid (10-20), drum speed (170-229), tilt angle (0-4º), shaking amplitude (10-20 mm), wash water rate (1-5 lpm), feed rate (1-3 lpm). The ranges of values of variables of Falcon Concentrator used in the design were gravity force (20-300 g), solids rate (10-30%), flow rate (0.5-1.5 Lpm), water pressure (0.5-1.5 psi). The mathematical model equations were derived for both ash content and combustible recovery by using sets of experimental data and a mathematical software package.
Predicted values match the experimental values reasonably well, with R2 of 0.807 for ash content and R2 of 0.944 for combustible recovery of Tunçbilek coal for MGS. However, ratio solid in feed of 15%, drum speed of 229 rpm, tilt angle of 0º, shaking amplitude of 10, wash water rate of 3 lpm, feed rate of 1 lpm have been determined as optimum levels to achieve the minimum ash content clean coal of 16.75%, whereas it is 16.75% in the tests conducted or predicted, i.e. a 0% upgrading in the ash content could be obtained by the mathematical model. In the same way, gravity force of 300 g, solids rate of 20%, flow rate of 1 Lpm, water pressure of 0.5 psi have been determined as optimum levels to achieve the minimum ash
content clean coal of 40.26%, whereas it is 41.51% in the tests conducted or predicted, i.e. a 3.10% upgrading in the ash content could be obtained by the mathematical model.
Consequently, when the ash content of clean coal and the coal recovery efficiencies considered that provided both facilities tailing, 19.72% and 40.26% ash containing the clean coal with 44.38% and 78.90% recovery was provided from 66.2% ash containing the Tunçbilek coal tailing with MGS and Falcon Concentrator, respectively. Also 22.89% ash containing the clean coal with 60.01% recovery was provided from 52.65% ash containing the Soma coal tailing.
References Abd Elrahiem, F.H., 2003, Carrier flotation for desulfurization and deashing of difficult to float El-Maghara Coal, Ore Dressing, 10, 10-17. Aslan, N., Canbazoğlu, M., Ulusoy, U., 1999, An Investigation of washability characteristics of lignites from Yeniçubuk-Gemerek districts by MGS. In: 16th Mining Congress of Turkey, 321-326. BP Statistical Review, 2009, Statistical review of world energy, 46 p. Çiçek, T., Cöcen, İ., Engin, V.T., Demir, Ş., 2002, Recovery of fine coal from tailings of coal washing plants. In: IXth Int Min Pro Symposium, Turkey, 198-199. Dagdelen, R., 2006, General assessment of the hardcoal mining and restructuring programme of Turkish Hardcoal Enterprise, The 15th Coal Congress of Turkey, 413-426. Honaker, R.Q., 1995, Enhanced gravity separators: new alternatives for fine coal cleaning. In: Pro 12th Int coal pre con, Intertec Inc., Lexington, Kentucky, 282–92. Honaker, R.Q., Wang, D., Ho, K., 1996, Application of the Falcon concentrator for fine coal cleaning, Mineral Engineering, 9 (11), 1143-56. Honaker, R.Q., 1998, High capacity fine coal cleaning using an enhanced gravity concentrator, Mineral Engineering, 11 (12), 1191–1199. Honaker, R.Q., Singh, N., Govindarajan, B., 2000, Application of dense-mediumin an enhanced gravity separator for fine coal cleaning, Mineral Engineering, 13, 415–427. Li, B., Tao, D., Ou, Z., Liu, J., 2003, Cyclomicrobubble column flotation of fine coal, Separation Science of Technology, 38, 1125-1140. Majumder, A.K., Tiwari, V., Barnwal, J.P., 2007a, Separation characteristics of coal Fines in a knelson concentrator-a Hydrodynamic approach, Coal Preparation, 27, 126-137. Majumder, A.K., Bhoi, K.S., Barnwa,l, J.P.,2007b, Multi-gravity separator: an alternate gravity concentrator to process coal fines, Mineral Processing & Metallurgy Review, 24, 133-138. Majumder, A.K., Barnwal, J.P., 2008, New Possibilities in Fine Coal Beneficiation Techniques, Journal Institution Engineers-Pt MN, 89, 3-8. Malkoç, Ö., 2008, Clean Coal Gain from Wastes of Egean Lignite Enterprises Soma Washery, Undergraduate Thesis, Afyon Kocatepe University, Faculty of Engineering, Department of Mineral Engineering, Turkey, 103 p. Parekh, B.K., Khalek, M.A.A., 2002, Using Falcon concentrator, as a new technology, for removal of environmental pollutants of Egyptian coal, Journal Ore Dressing, 4 (7), 20-8. Turkish Standard, TS ISO 5667-10, 2002, Water quality-sampling-Part 10: Guidance on sampling of waste waters, Turkish Standards Institution, 8 p. Yıldırım, I., Ateşok, G., Çelik, M.S., 1995, Production of clean coal with low ash content for coal-water slurry by laboratory type Multi Gravity Separator, 14th Mining Congress of Turkey, Istanbul, 443–447. Wentworth, C.K., 1922, A scale of grade and class terms of clastic sediments, Journal Geology, 30, 377–392. World Coal Institute (WCI), 2005, The coal resource-A comprehensive overview of coal, London.
Manuscript Not AVAILABLE
EVALUATION OF DENSE MEDIUM SEPARATION PERFORMANCE OF İMBAT COAL PREPARATION PLANT
G.Özbayoğlu 1, Ü. Atalay 2, Ali İ.Arol 2, O.Sivrikaya 2 [email protected] 1.Engineering Faculty, Atilim University, Ankara-Turkey 2.Mining Engineering Department,Middle East Technical University, Ankara-Turkey
The aim of this study is determination of the washability characteristics of İmbat coal preparation plant feed and evaluation of performance of operating coal preparation plant for two different capacities. This plant was established to clean the r.o.m. coal with 2200 k.cal/kg average calorific value and 40-45 % ash content. The washability characteristics of plant feed were determined for the size fractions of + 50mm, -50 +18 mm and -18 mm fractions separately. A sink and float tests were carried out with representative test samples of each size fractions with the specific gravities of 1.30, 1.40,1.50, 1.60, 1.70ş 180 and 1.90 . Cumulative float ash curve, cumulative sink ash curve, Elementary ash curve, specific gravity curve and specific gravity distribution curve were drawn. It was possible to decrease ash content of coal down to 4.61%, 3.1% and 5.1% for the sizes of +50mm, -50 +18 mm and -18 mm fractions respectively at the lowest separation density of 1.30 g/cm3 with very low yield The operating coal preparation plant have a Drew boy and dense medium cyclones for cleaning of coarse and fine coals respectively. The first separation in the Drew boy is carried out at high density . The float fraction of first Drew boy is send to the second drew boy which is operating with low separation density. Here the float product for thermal power station is ptoduced. The fine fraction of plant feed is cleaned in single stage operation of dense medium cyclones. A detailed in-plant operation of the dense medium separation employed in İmbat coal preparation plant was also conducted to determine the relationship between the plant capacity and the plant performance. The separation performances achieved by the Drew boy and dense medium cyclone circuit under two different capacity conditions indicated that the increase in the capacity from 500 t/hour to 600 t/hour resulted a slight decrease in the performances of both equipments.
EFFECT OF SHAPE FACTOR ON COAL FLOTATION O. Güven, G. Bulut, K.T. Perek
Istanbul Technical University, Mining Faculty Mineral Processing Engineering Department 34468, Maslak, Istanbul, Turkey [email protected]
ABSTRACT
In this study, the flotation behavior of hard coal products having different particle shapes produced by different grinding conditions was investigated with and without reagents. As a reagent kerosene + isooctanol mixture was used in coal flotation. The flotation kinetics of coal particles with different shape was studied as a function of time. Shape characteristics of the particles were investigated by Leica Qwin program. The results showed that particles presenting lower fullness
ratio and roundness properties were recovered better during
flotation of the coal mineral studied. As a result, the shape of the particles produced by the ball mill changed upon to their roundness, their floatability was increased.
INTRODUCTION Flotation is one of the most complex mineral processing operations as it is affected by a very large number of variables. Many of these are beyond the control of the minerals engineer, and some can not be even measured quantitatively with the available instruments (Vapur et al., 2010). Numerous factors influence flotation. Although particle hydrophobicity and size are important and have received much attention in research, a few of investigations have reported the effect of particle shape on flotation recovery and kinetics (Hicyilmaz et al., 1995, Gorain et al., 1998 Koh et al., 2009, Mahmoud M. Ahmed, 2010, Melo and Laskowski, 2005). Grinding is an essential operation for determining the appropriate size for flotation. Some researchers have also investigated the difference of shape factor in terms of different grinding methods (Ulusoy et al., 2003). Grinding conditions play an important role on shape and morphological properties of particles. In general abrasion, impact and etc. were the main mechanisms forming during grinding operation and impact effect presents the main
characteristic of grinding in ball mills. Koh et al., 2009, investigated the effect of particle shape and hydrophobicity in flotation, they found that ground particles increased the flotation recoveries but in terms of contact angle measurements, hydrophobicity degree decreases in accordance with the increasing surface coverage ratio. In this paper the effects of particle shape, and kinetics on the flotation process were investigated. The flotation experiments were performed with hard coal samples with different grinding conditions in a laboratory type Denver cell. The flotation time was varied between 0.5 to 4.5 minutes for flotation tests with and without reagent. MATERIALS In this study, hard coal samples from Zonguldak, Turkey were used. The size distribution for samples from different grinding conditions was shown in Table 1. The chemical analysis showed that the constant carbon, volatile, ash and moisture contents were 58.2%, 27%, 14.1% and 3.5 % respectively. These results indicated that the sample was nearly dry and pure hard coal. Table 1. Size distribution of samples related with different grinding conditions
Size Range (mm) -0,5+0,3 -0,3+0,212 -0,212+0,150 -0,150+0,106 0,106+0,074 0,074+0,053 0,053+0,038 -0,038
Cumulative Undersize (%) Dry grinding Wet grinding conditions conditions 100,0 98,9 100,0 92,9 96,8 76,8 77,1 63,2 64,6 53,9 53,6 46,1 43,9 37,9 30,4
In experimental studies, the effects of grinding conditions on particle shape and also the effects of these materials on kinetics of the flotation were investigated. Grinding tests were carried out with a laboratory ball mill. Feed size of the coal samples were below 1 mm. Five minutes of dry grinding gave similar size distribution with wet grinding at
the same grinding time which provides a comparative study for their particle shape characteristics. Image analysis were carried out with Leica Qwin image analyze program (Leica QWin User Manual,1995) based on the particle projections obtained from the photographs. About 200 measurements were performed at -0,106+0,074 mm size fraction for determining the particle shape parameters. The average of lengths and breadths were taken as length (L) and breadth (B). From these results, shape factors called aspect ratio, roundness and fullness ratio can be defined as follows ;
Roundness =
4 * * A * 1,064 P2
Aspect Ratio=
Fullness ratio= sqrt(
L B
Area ) ConvexArea
In these formulas, P: perimeter and A: area of the particles. Convex area: the area of the polygon circumscribing the feature, formed by tangents to its boundary (Terence Allen, 2003). Flotation tests were performed in a 2.5 L Denver flotation cell. The flotation tests were carried out with and without reagent. The reagent used in flotation tests was a mixture of 80 % wt. kerosene and 20 % wt. isooctanol. These conditions were chosen to determine the kinetics of the
flotation. Flotation products were collected periodically between 0.5 to 4.5 minutes without reagent and 0.5 to 2.5 minutes with reagent. All flotation products were dried and then weighted to determine the mass recovery and also kinetics of flotation. The kinetics of the flotation was calculated by the equation,
ln(1 Rt / Rm ) where Rt: The
cumulative yield of each time interval and Rm: the maximum cumulative yield of flotation test (Arbiter and Harris, 1962). Results and Discussion The effects of grinding conditions on particle shape was investigated at particle size range of -0,106+0,074 mm. The results of image analysis was shown in Table 2.
Table 2. The image analysis results at different grinding conditions
Parameters 2
Area, µm Length, µm Breadth, µm Convex Area, µm2 Perimeter, µm Roundness Aspect Ratio Equivalent Diameter, µm Fullness Ratio
Dry grinding
Wet grinding
Total 18039,87 9759,83 6451,34 21824,06 4890,66 111,66
Mean Total 247,12 15161,82 133,7 8731,57 88,37 5530,64 298,96 18961,44 67 4314,2 0,73 1,51 105,17
Mean 233,26 134,33 85,09 291,71 66,37 0,71 1,58
1284,14
17,59
1110,87
17,09
-
0,91
-
0,89
The shape factor showed that the particles ground at dry grinding conditions has higher roudness value and fullness ratio in comparision with the particles ground at wet grinding conditions. The difference in particle shape of the ground samples at -0,106+0,074 mm size fraction for each dry (Figure 1a) and wet (Figure 1b) grinding conditions was shown in Figure 1a,b.
Figure 1a. Image analysis for dry ground Figure 1b. Image analysis for wet ground particles at -0,106+0,074 mm size particles at -0,106+0,074 mm fraction size fraction The results of the flotation tests were shown in Table 3. These results indicated that using reagents not only enhanced the flotation recovery but also a decrease on flotation time.
Table 3. The results of the flotation tests
Flotation Time (min) 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Flotation tests without reagent
Flotation tests with reagent
Cumulative yield (%)
Cumulative yield (%)
Dry grinding conditions 6.1
Wet grinding conditions 6.7
Dry grinding conditions 55.3
Wet grinding conditions 59.0
9.8 13.2 15.9 18.3 20.3 22.0 23.7 25.8
12.4 17.1 21.1 24.8 28.5 31.2 33.6 36.2
71.8 80.6 84.1 85.8 -
81.8 88.3 90.2 91.9 -
The kinetics of the flotation was calculated by the equation described above. The flotation rate constants (k- values) were plotted versus flotation time (Figure 2) in order to determine the effect of grinding conditions and reagent on the kinetics of the flotation.
Figure 2. The kinetics of flotation versus time
As this figure reveals that similar values were obtained for each test group carried out with and without reagent (Table 4). Therefore it was found that grinding conditions did not lead to a considerable difference for kinetics of flotation. Table 4. Flotation rate constant (k), min-1 values for different grinding conditions Conditions With reagent Without reagent
Flotation rate constant (k), min-1 Dry grinding conditions Wet grinding conditions 1,9434 1,9941 0,6135 0,6515
Conclusions These results showed that grinding conditions play an important role on shape and morphological properties of particles. It was found that different grinding conditions resulted by particles with different shape factor. The particles ground in dry ball milling has higher roundness which decreases the flotation recovery. The kinetics of flotation showed that using reagent increases the flotation recovery to the maximum point at nearly half time of tests without reagent. References Ahmed, M. M., 2010. Effect of communition on particle shape and surface roughness and their relation to flotation process. Int. J. Miner. Process 94, 180–191 Allen, T., 2003. Powder sampling and particle size determination. Elsevier, USA Arbiter, N., Harris, C.C., 1962. Flotation kinetics in froth flotation. 50th Anniversary Volume, pp. 215-246, Editor: Fuerstenau D.W., AIME, Newyork. Gorain, B.K., Harris, M.C., Franzidis, J.P., Manlapig, E.V., 1998. The effect of froth residence time on the kinetics of flotation. Mineral Engineering 11 (7), 627-638 Hicyilmaz, C., Bilgen, S., Atalay, U., Ulusoy, U.,
1995. Effect of
shape
characteristics of particles on flotation. Fizykochemiczne Problemy Mineralurgii, 29, 31-38 Koh, P.T.L., Hao, F.P., Smith, L.K., Chau, T.T.,Bruckard, W.J., 2009. The effect of particle shape and hydrophobicity in flotation. Int. J. Miner. Process 93, 128–134 Leica QWin User Manual, 1995. Leica Q500MC QWin User Manual
Melo, F., Laskowski, J.S., 2005. Fundamental properties of flotation frothers and their effect on flotation. Minerals Engineering, 19, 766-773 Ulusoy, U., Hicyilmaz, C., Yekeler, M., 2003. Role of shape properties of calcite and barite particles on apparent hydrophobicity. Chemical Engineering and Processing 43, 10471053 Vapur, H., Bayat, O., Ucurum, M., 2010. Coal flotation optimization using modified flotation parameters and combustible recovery in a Jameson cell. Energy conversion and management 51, 1891-1897
REFORMING OF LOW RANK COAL BY SOLVENT TREATMENT AT AROUND 350oC Xian LI, Ryuichi ASHIDA, Hiroyasu FUJITSUKA, Kouichi MIURA Department of Chemical Engineering, Kyoto University, Kyotodaigaku Katsura, Nishikyo-ku, Kyoto 615-8510, Japan
Abstract: Recently, we proposed degradative solvent extraction at around 350C as an efficient upgrading method of low rank coals. In this study, eight low rank coals including lignites and sub-bituminous coals were thermally treated and fractionated by using 1-methylnaphthalene as solvent at 350C. The coals were partly decomposed and were separated into three fractions having different molecular weights. The yields of the three fractions respectively ranged from 17wt% to 27wt% (soluble), 4 wt% to 17 wt% (deposit), and 47 wt% to 64 wt% (residue; upgraded coal) on d.a.f. coal basis. The total higher heating value (HHV) of the fractions (sum of HHV of soluble, deposit and residue on raw coal basis) was not lost. The carbon contents of the three fractions were much larger than those of their parent coals, suggesting that significant amount of oxygen was effectively removed from the coals during the treatment. The interesting findings were that the solubles and deopsits obtained from the eight coals were respectively very close to each other in elemental composition, chemical structure, molecular weight distribution, pyrolysis behavior, and thermal plastic behavior. Thus, the proposed degradative solvent extraction method was found to be effective in converting low rank coals into upgraded coal and compounds having very similar chemical and physical properties. Key words: Low rank coal; Fractionation; Upgrading; Solvent extraction; Degradative extraction.
1. Introduction Low rank coals are currently used just as fuels near coal mines, because their utilization is limited due to the high moisture content and the high spontaneous combustibility
after
dewatered.
Therefore,
both
dewatering and upgrading are necessary to utilize low rank coal more effectively even as just fuels. In 1996 Song and Schobert stressed the importance of preparing chemicals and materials from coal 1). If we also think of the low rank coal as the feedstock of chemicals and materials, it is necessary to recover precursors of chemicals and/or materials from the low rank coals in addition to dewatering and upgrading. The authors have proposed a method to upgrade and fractionate coal by using non-polar solvent at below 350C 2, 3). In this work the degradative extraction of low rank coals is proposed as a method to recover several fractions having similar chemical and physical properties from wide range of low rank coals. 2. Experimental 2.1 Coal sample and solvent Eight low rank coals including lignites from Thailand (MM) and Philippine (PH), brown coals from Indonesia (WR), Australia (LY), and Malaysia (BB and MB), and two sub-bituminous coals from Indonesia (AD and TH) were employed in this work. And the properties of the coals were shown in Table 1. 1-methylnaphthalene (1-MN) was used as non hydrogen donor solvent for the
degradative extraction. Table 1 Properties of the eight low rank coals Coal MM LY WA BB MB PH AD TH
Ultimate analysis (wt%, d.a.f.) C H N O+S(diff.) 66.4 3.9 1.9 27.8 66.7 4.7 0.9 27.7 67.1 5.1 1.0 26.9 71.0 4.9 1.3 22.8 71.7 4.8 1.7 21.9 72.2 4.6 0.9 22.3 72.9 5.1 1.0 21.0 80.7 5.0 2.0 8.2
Proximate Analysis (wt%, d.b.) VM FC Ash 50.2 24.0 25.8 51.5 47.0 1.5 50.5 47.9 1.5 43.4 52.5 4.1 42.8 52.6 4.6 51.5 36.7 11.8 51.7 46.5 1.8 41.6 50.2 8.2
Moisture wt%, a.r. 12.2 56.3 37.0 21.9 24.1 25.8 21.2 10.3
2.2 Experimental procedure Since the treatment in organic solvent at around 350°C was found to be effective in upgrading low rank coal in our previous works 4, 5), the thermal treatments of the eight low rank coals were performed in a batch reactor (autoclave) at 350°C. The schematic diagram of the apparatus is shown in Fig. 1. A stainless steel autoclave (350 cm3, 55 mm I.D.) was charged with as-received coal (13 g on dry basis) and 300 cm3 of 1-MN. A stainless filter (65 mm O.D. and 0.5 µm opening) was equipped at the bottom of the autoclave. After sufficiently purging the autoclave with N2, the sealed autoclave was heated up to 350°C, where it was kept for 60 min under autogenous pressure. The extracts and the residue (upgraded coal; UC) were separated by opening the valve connected below the filter at the extraction temperature. The extracts with the solvent were collected in a stainless steel vessel equipped under the valve which was cooled by water. The solvent containing the extracts was filtrated using a PTFE membrane filter (0.5 μm opening) to separate the extracts into the extract that precipitated at room temperature (deposit) and the extract soluble in solvent even at room
90
Raw coal
UC
Deposit
Soluble
85 80 75 70 65 60 9 8 7 6 5 4 3 2 30 25 20 15 10 5 0
MM
LY
WA
BB
MB
PH
AD
TH
Coal
Fig. 3 Carbon, hydrogen and oxygen contents of the three fractions and raw coals
CO2
Other gas
C content (wt%, d.a.f.)
3. Results and discussion 3.1 Yields of products First, the water inherent in the coals was completely removed as liquid water while the coals were heated up to 350°C. Fig. 2 shows the yields of soluble, deposit, UC, H2O and gaseous products. The yields of the same fractions obtained from the eight coals were different, depending on the coal type. The soluble and deposit yields respectively ranged from 17 wt% to 27 wt% and from 4 wt% to 17 wt% on d.a.f. coal basis. This means that 22 wt% – 41 wt% of the coals were extracted by the 1-MN. The gaseous products consisted almost solely of CO2. H2O yields reached 10 wt%, especially for the lower rank coals. This result suggests that the degradation of oxygen functional groups occurred during the treatment.
H content (wt%, d.a.f.)
Fig. 1 Schematic diagram of the apparatus used for low rank coal extraction
contents of soluble and deposit were respectively rather close to each other. The oxygen contents of the soluble and deposit were lower than those of their parent coals. The carbon contents of soluble were surprisingly from 81.8 wt% to 84.0 wt%, and the oxygen contents were as small as from 6.5 wt% to 10.2wt% on d.a.f. sample basis. So, the elemental compositions of the solubles were rather close to that of bituminous coal. In addition, the carbon and oxygen contents of the UC were respectively increased and decreased as compared to those of their parent coals. Fig. 4 shows the elemental compositions of extracts and raw coals on H/C vs. O/C diagram. For the soluble, all of the data points converged on the values of H/C 1.1 and O/C 0.07 – 0.09. On the other hand, all the data points converged on the values of H/C 0.77 – 0.91 and O/C 0.12 – 0.18 for the deposit. These results show that the proposed degradative extraction can convert the wide range of low rank coals into soluble and deposit that respectively have similar elemental compositions.
O content (wt%, d.a.f.)
temperature (soluble). The solvent dissolving the soluble was evaporated at around 140°C under reduced pressure to recover the soluble as solid. The yields of UC, deposit, and soluble were determined by weight. The gaseous products were analyzed by gas chromatograph, and the yield of H2O was determined by the oxygen balance.
Soluble
80
H2O
1.2
60
UC
40 20 0
MM
LY
WA
BB
MB
PH
AD
Raw Soluble Deposit UC
-CO2
Deposit
Atomic ratio, H/C [-]
Yield (wt%, d.a.f.)
100
-CH4
1.0
0.8 -H2O
TH 0.6
Fig. 2 Yields of products. 0.0
3.2 Properties of extracts and UC Fig. 3 shows the elemental compositions of the three fractions and raw coals. The carbon and hydrogen
0.1
0.2
0.3
0.4
Atomic ratio, O/C [-]
Fig. 4 Elemental compositions of fractions and raw coals on H/C vs. O/C diagram.
30
other as compared to those of the raw coals. This was
Deposit Soluble
deposits and solubles were respectively very close to each
25 20 15
also the case with the aliphatic C-H stretching vibration (2980 – 2845 cm-1). Furthermore, the absorption bands of the whole region from 1800 cm-1 to 700 cm-1 were also
Residue
Raw coal
HHV [MJ/kg-raw coal, daf]
35
10
rather similar to each other for deposits or soluble. Fig. 7
5
shows the hydrogen distributions measured by 1H-NMR
0
MM
LY
WA
BB
MB
PH
AD
for solubles and deposits. The hydrogen distributions of
TH
the solubles and deposits were respectively rather close to
Fig. 5 HHVs of raw coals and fractions.
each other. Thus, the results of FT-IR and hydrogen Fig. 5 shows the higher heating value (HHV) of raw coals and fractions estimated by the Motto-Spooner equation 6). The total HHV of fractions (sum of HHV of soluble, deposit and UC on raw coal basis) was larger than that of their parent coal, except WA and PH. This indicates that the treatment is an endothermic process for the low rank coals. Thus, the proposed degradative extraction does not cause the heating value loss for low rank coal.
solubles and deposits are respectively rather close to each other. Fig. 8 shows the molecular weight distributions (MWDs) measured by LD-TOFMS for solubles and deposits. The results show that the MWDs of solubles were rather close to each other and the molecular weights were around 280. This was also the case with the deposits, and the molecular weights of the deposits ranged from 400 to 800.
80
60
distribution analyses show that the chemical structure of
Raw coal
Soluble
Depodit
MM
MM
LY
LY
WA
WA
40
Absorbance (a.u.)
0 100
Deposit
80 60
Intensity (a.u.)
Intensity (a.u.)
20
BB MB PH
BB MB PH
40
AD
AD
20 0 140 120
TH
Soluble
MM WA MB AD
100 80
1000
LY BB PH TH
2000
TH
3000
1000
Mass/Charge
2000
3000
Mass/Charge
Fig. 8 MWDs of solubles and deposits.
60 40 20
Weight loss (wt%, d.a.f.)
3000 2000 1000 -1 Wave number (cm )
Fig. 6 FT-IR spectra of extracts and raw coals. 1.0 H
0.8
Hydrogen distribution (-)
Hydrogen distribution (-)
1.0 H
0.6 0.4 0.2 0.0
H
Har
MM
LY
WA
BB
MB
Soluble
PH
AD
H
0.8
80 Deposit 60 40 Soluble 20 0 0
H
Normalized displacement (-)
100
0
4000
0.0 -0.2 -0.4
Deposit Soluble
-0.6 -0.8
(a) (b) 300
600
900
-1.0 0
o
200
400
600
o
Temperature ( C)
Temperature ( C)
0.6 0.4 0.2
Fig. 9 TG (a) and TMA (b) curves of solubles and
H
deposits.
Har
0.0 TH MM
LY
WA
BB
MB
PH
AD
TH
Deposit
Fig. 7 Hydrogen distributions of solubles and deposits. Fig. 6 shows the FT-IR spectra of extracts and raw coals obtained by the commonly-used KBr pellet method. The strength and shape of the peak attributed to stretching vibration of hydroxyl group (3600 – 3200 cm-1) of
Fig. 9 shows the thermogravimetric (TG) and the thermomechanical analysis (TMA) curves of the solubles and deposits during heating up to 900°C (TG) / 600°C (TMA) at same heating rate of 10 K/min. Both TG and TMA curves of solubles and deposits are respectively rather similar to each other. Additionally, the TMA results show that all of the solubles and deposits completely melted at lower than 100°C and 250°C, respectively.
Thus, the solubles and deposits respectively have rather similar pyrolysis and thermal plastic behaviors, regardless of the coal type. 4. Conclusion Eight low rank coals including lignites and sub-bituminous coals were thermally treated and fractionated by 1-MN at 350°C to obtain extracts (soluble and deposit) and UC. The total HHV of the fractions was not lost in the treatment process. The carbon and oxygen contents of the three fractions were respectively much higher and much lower than those of their parent coals. The solubles and deposits obtained from the eight coals were respectively very close to each other in elemental composition, chemical structure, MWDs, pyrolysis and thermal plastic behavior. Thus, the proposed degradative extraction method was found to be effective in converting low rank coals into UC and compounds having similar chemical and physical properties without heating value loss. Acknowledgement The assistance of Kobe Steel Co. for construction of the extraction apparatus is gratefully acknowledged. Reference 1) Song, C.S. and Schobert, H.H., Non-fuel uses of coals and syn hesis of chemicals and materials. Fuel 1996;75:724-736 2) Ashida R, Morimoto M, Makino Y, Umemoto S, Nakagawa H, Miura K, Saito K, Kato K. Fractionation of brown coal by sequential high temperature solvent extraction. Fuel 2009; 88: 1485-1490 3) Ashida R, Nakagawa K, Oga M, Nakagawa H, Miura K. Fractionation of coal by use of high temperature solvent extraction technique and characterization of the fractions. Fuel 2008; 87: 576–82 4) Miura K, Ashida R, Umemoto S, Sakajo A, Saito K, Kato K. Extraction and upgrading of brown coal through treatment in solvent at high temperature, Proceedings - Annual International Pittsburgh Coal Conference 2008, 25th, 464/1–464/12 5) Ashida R, Umemoto S, Hasegawa Y, Miura K, Kato K, Saito K, Nomura S, Upgrading of low rank coal through mild solven treatment at temperatures below 350°C Proceedings - Annual International Pittsburgh Coal Conference 2009, 26th, 51/1–51/12. 6) Mott R A, Spooner C E. The Calorific Value of Carbon in coal: The Dulong Relationship. Fuel in Science and Practice 1940; 19: 226-231, 242-251.
t
An Experimental Investigation of Factors related to Coke Strength Degradation in Coke Milli-Structure Tetsuya KANAI*, Yoshiaki YAMAZAKI*, Kenichi HIRAKI*, Xiaoqing ZHANG*, Masakazu SHOJI*, Hideyuki AOKI* and Takatoshi MIURA* *Department of Chemical Engineering, Graduate School of Engineering, Tohoku University, 6-6-07 Aoba, Aramaki, Aoba-ku, Sendai, 980-8579, Japan
Abstract In order to clarify the factors related to coke strength degradation, relationship between coke strength and pore structure is quantitatively investigated. The tensile strength of coke is measured by diametral-compression test and analyzed by Weibull plot. Pore structure is analyzed with wide range(10 mm×10 mm) and high resolution(2.43 μm/pixel) photographs which are acquired by combining approximately 20 photographs. Absolute maximum pore length, pore area ratio and pore roundness are measured by image analysis of the photographs. By comparison of strength and absolute maximum pore length in coke, strength is degraded with an increase in area ratio of pores over 1000 μm in absolute maximum length. From image analysis, it is found that the pore roundness decreases with an increase in absolute maximum length. The length of 1000 μm corresponds with the critical crack length calculated by Griffith equation in scale.
1 Introduction For stable operation of blast furnace, sustaining the flow passes of liquid metal and high temperature reducing gas in the blast furnace is necessary. Since coke has those roles, strength is regarded as the most important quality for coke. Recently, demand for steel in Asia such as China and India is increasing and that causes the rapid rise in coal price. The steel industry has no other choice to use a lot of slightly caking coal with low price. Now, in order to make design guides of strong coke despite high-volume use of slightly caking coal, the factors related to the coke strength degradation should be identified. To evaluate degradation of the coke strength, many factors are needed to be considered because coke has a lot of cracks and pores. Though drum index provides appropriate value for furnace operation, it is not appropriate for investigating the effects of micro structure such as pore and crack on coke strength because fracture morphology in drum is complicated. Therefore, quantitative evaluation from a point of view of material mechanics is necessary. Quantitative analysis of mechanical properties in coke milli-structure is essence for making designs of strong coke which has complex hierarchical structure.1) Coke milli-structure is defined as coke lump size of which is about 10 mm.1) However there are few reports about quantitative evaluation of relationship between coke structure and strength considering coke milli-structure. Recently it has been recognized that coke strength is affected by mainly defect structures such as pore and crack rather than coke matrix property.2-3) In this study, the effects of pore structure on strength in coke milli-structure is quantitatively investigated by pore structure analysis and tensile strength measurement for coke which is made by varying coal blending ratio and coal particle size.
2 Experimental 2. 1 Specimen making Coke samples for the test specimen are provided by Nippon Steel Corporation. Manufacturing condition of cokes is shown in Table 1.The cokes are made by blending of strongly caking coal Goonyella and noncaking coal NCBC. Procedure of making coke samples is shown in Figure 1. The coke taken out from coke oven are bored and cut to the disc with the diameter of about 19 mm and the thickness of about 5 mm.
Table 1 Manufacturing condition of cokes
Goonyella [ mass %]
NCBC [ mass %]
Coal size [mm] (Mass fraction)
100 50 50 0
0 50 50 100
< 3 (85 %) < 1 (100 %) < 3 (85 %) < 3 (85 %)
A B C D
Coal charging density
Coke lump Fig. 1 Procedure of making coke samples
Y X
¥ Fig. 2 A schematic diagram of diametral-compression test
3
[kg/m ] 830 850 850 700
2. 2 Diametral-compression test Measurement principle of Diametral-compression test is shown in Figure 2. Tensile stress σ is generated at x direction by applying load P at coke from y direction. Tensile stress σ is calculated from following Eq. (1) : ,
(1)
where d and l are diameter and thickness of disk specimen, respectively. The tensile stress calculated by substituting maximum load P into Eq. (1) is regarded as tensile strength σt. Applying and measuring of load P is conducted by use of Auto Graph (AG-150 kN, Shimadzu Co.). Loading rate is 2 mm/min. In order to evaluate the tensile strength statistically, tensile strength is evaluated by Weibull analysis.
2. 3 Pore Structure analysis of coke surface Acquiring the image of coke surface Coke specimens are impregnated with epoxide-based resin by vacuum impregnation. After that the specimen is grinded and buffed. Digital microphotograph of coke grinded surface is acquired by use of optical microscope Nikon LV-100-POL. Size of photograph is 3.11 mm ×2.33 mm, resolution is 2.43 μm/pixel. About 20 photographs acquired in one coke surface are combined to one photograph whose size and resolution are 10.0 mm×10.0 mm and 2.43μm/pixel, respectively. The photograph is shown in Figure 3. White colored part is coke matrix and black or gray colored part is pore. Two combined photographs are acquired every condition of coke. Acquired photograph is binarized by use of image analysis software (Win roof ver. 6.1). Binarization is carried out with discriminant analysis method which is one of the auto binarization methods. A binarized image is shown in Figure 4. Green colored part is pore and black colored part is coke matrix.
Pore
Pore (Green)
Coke matrix
Coke matrix (Black)
1 mm
1 mm
Fig. 3 Conbined photograph of coke surface
Fig. 4 Binary image of Fig . 3
Pore structure analysis Absolute maximum pore length, pore area and pore roundness are measured from the binarized images by use of image analysis software (Win roof ver. 6.1). Where absolute maximum length is defined as maximum length of two point on pore contour. Roundness is represented as Eq. (2) :
(2)
,
where S and L are pore area and boundary length of pore, respectively. Roundness takes a value between 0 and 1. The larger roundness means that the pore structure is like a circle.
3 Results and discussion 3.1 Diametral-compression test Weibull plot of tensile strength σt acquired from diametral-compression test is shown in Figure 5. where cumulative breaking probability F(σt) is calculated by median rank method. The cumulative breaking probability with median rank method is represented as Eq. (3) :
(3)
,
where i is the strength order in magnitude and m is the number of total specimen. Scale parameter calculated from Weibull plot is shown in Table 2. Scale parameter denotes the tensile strength when F(σt) reaches 63%. The strongest coke is B and the weakest coke is D. Strength of coke A and C is more or less the same.
2.0 A
[-]
1.0
Coke sample B C D
0.0 -1.0 -2.0 -3.0 -4.0 0.0
1.0
2.0
lnlnσσt [-][-]
Fig. 5 Weibull plot of each coke sample
3.0
Table 2 Scale parameter of each coke sample
Scale parameter [MPa]
A
B
C
D
5.37
7.67
5.20
3.28
3. 2 Pore structure analysis of coke Absolute maximum pore length Area ratio of pores in each absolute maximum length to total area of photograph is shown in Figure 6. Absolute maximum pore length is ranged 0-100 μm, 100-500 μm, 500-1000 μm and 1000- μm. In coke B, area ratio of pores under 500 μm is larger than other cokes. Area ratio of pores over 1000 μm is smaller than other cokes. In coke D, area ratio of pores under 500 μm is smaller than other cokes. Area ratio of pore over 1000 μm is larger than other cokes. These observations imply that the weaker coke the more large absolute maximum length pore has.
0.8
Absolute Maximum of pore[-] [-] Area ratio of pores inlength photograph
0.6 0.5
A
B
C
A
0.7
D
B
C
D
0.6 0.4
0.5 0.4
0.3
0.3
0.2
0.2 0.1
0.1 0
0 0-100
100-500 500-1000
1000-
Absolute maximum pore length [μm] Fig. 6 Area ration of pores in each absolute maximum pore lengrh
0-100
100-500 500-1000
1000-
Absolute maximum pore length [μm] Fig. 7 Average pore roundness in each absolute maximum pore length
Pore roundness Average pore roundness in each absolute maximum length is shown in Figure 7.In all cokes, pore roundness is decreases with an increase in absolute maximum length. This indicates that pore shape becomes distorted with an increase in pore size regardless of types of coke. Pore roundness in coke A, absolute maximum length of which is larger than 500 μm is smaller than other cokes. It is considered that pore shape becomes complicated as a result of linking pore. Coke A which is consist entirely of strongly caking coal has a tendency to expand larger bubbles during softening.
3. 3 Factors related to coke strength degradation Relationship between coke strength and pore structure Factors related to coke strength degradation are investigated from relationship between coke strength and pore structure. Relationship between area ratio of pore and scale parameter is shown in Figure 8. Strength is degraded with an increase in area ratio of pores over 1000 μm in absolute maximum length. On the other hand, strength is not degraded with an increase in area ratio of pores under 1000 μm in absolute maximum length. This indicates that pores over 1000 μm in absolute maximum length affect the strength degradation. The effect of coke strength degradation with
Scale parameter [MPa]
pores under 1000 μm in absolute maximum length is small.
9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0
-100 [μm] 100-500 [μm] 500-1000 [μm] 1000- [μm]
A : Black C : Orange 0 0
0.1 0.1
B : Blue D : Green
0.2 0.3 0.4 0.2 0.3 0.4 Area ratio of pore [-]
0.5 0.5
0.6 0.6
Fig. 8 Relationship between area ratio of pore and scale parameter The effect of coke strength degradation with the pores over 1000 μm in absolute maximum length is considered. It has been recognized that the pore roundness of which is about 0.6-1.0 decrease in homogenized elastic modulus which represents macro stiffness. 4) On the other hand, the pore roundness of which is small decrease homogenized elastic modulus 4). It is indicated that the low-roundness pores over 1000 μm in absolute maximum length decrease macro stiffness of coke and coke strength. Strength of coke A is as large as that of coke C despite area ratio of pores over 1000 μm is smaller than coke C. This is because pore roundness of A is smaller than that of coke C as shown in Figure 7.
Griffith crack Size of crack length which affects coke break is investigated with Griffith equation as : ,
(4)
where σ and c are tensile stress and half length of crack, respectively. Fracture occurs when this stress intensity factor KI exceeds fracture toughness KIc which is one of material properties. Critical crack length of tensile failure of coke is calculated as 2c = 1.59 mm by substituting parameters of tensile failure of coke KIc = 0.25 MPa 5) and tensile stress σ = 5 MPa 5) into Eq. (4). It is indicated that the crack in 1 mm scale is origin of the tensile fracture of coke from critical crack length. This corroborates the result that the factor related to coke strength degradation is the pore over 1000 μm in absolute maximum length.
4 Conclusion The effects of pore structure on strength in coke milli-structure is quantitatively investigated by pore structure analysis and tensile strength measurement for coke which is composed of varying coal blend ratio and coal particle size. The conclusions are summarized as follows : The factor related to coke strength degradation is the pore over 1000 μm in absolute maximum length. Pore roundness is decreases with an increase in absolute maximum length. The length of 1000μm corresponds with the critical crack length calculated by Griffith equation in scale.
Acknowledgment Coke sample used in this study is provided by Nippon Steel Corporation. We would like to express our gratitude to Nippon Steel Corporation.
Nomenclature c
:
half length of crack
[m]
d
:
diameter of specimen
[m]
F
:
cumulative breaking probability
[-]
i
:
strength order of coke sumple
[-]
KI
:
stress intensity factor
[Pa・m1/2]
KIc
:
fracture toughness
[Pa・m1/2]
l
:
thickness of coke sample
[m]
L
:
boundary length of pore
[m]
n
:
number of coke sample
[-]
P
:
load
[N]
S
:
area of pore
[m2]
:
tensile stress
[MPa]
:
tensile strength
[MPa]
t
References 1)
M. Sakai, R. Nishimura, M. Nishimura and K. Fukuda : Tetsu-to-Hagané, 92 (2006), 164.
2)
T. Ogata, K. Ueoka, Y. Morozumi, H. Aoki, T. Miura, K. Uebo, K. Fukuda and K. Matsudaira : Tetsu-to-Hagané, 92 (2006), 171.
3)
K. Ueoka, T. Ogata, Y. Morozumi, H. Aoki, T. Miura, K. Uebo and K. Fukuda : Tetsu-to-Hagané, 92 (2006), 184.
4)
K. Ueoka, T. Ogata, Y. Matsushita, H. Aoki, T. Miura, K. Fukuda and K. Matsudaira : Tetsu-to-Hagané, 93 (2007), 728.
5)
T. Arima: Tetsu-to-Hagané, 87 (2001), 274.
An Experimental Study on the Effect of Metallic Iron Particles on Strength Factors of Coke after CO2 Gasification Reaction Yoshiaki Yamazaki1, Kenichi Hiraki1, Tetsuya Kanai1, Xiaoqing Zhang1, Masakazu Shoji1, Hideyuki Aoki1 and Takatoshi Miura1 1
Department of Chemical Engineering, Graduate School of Engineering, TOHOKU University, 6-6-07, Aoba, Aramaki, Aoba-ku, Sendai 980-8579, Japan
Abstract In order to prevent the pulverization and the fracture of coke lump in blast furnace, control of degradation part (reaction mechanism) in coke lump and that of the embrittlement behavior are significant. Addition of catalyst particle is an effective and a simple method for advancement of the CO2 gasification reactivity and control method for reaction mechanism. In this study, the effect of iron particles on coke-matrix state after gasification reaction is investigated experimentally. Coke-matrix vanishing is evaluated by spatial distribution of lump porosity and microscopic observation. Elastic modulus of coke-matrix is evaluated by nano-indentation method. Coke lumps with and without iron-particles (ferrous coke and formed coke, respectively) were used. These coke lumps were gasified by CO2-containing gas. Reaction temperature was set at 1173 K. Reaction gas compositions were set at 100/0 and 50/50 in ratio of CO2/CO. In each reaction gas composition, in ferrous coke, a decrease in the elastic modulus of coke-matrix with progress of gasification is smaller than that in formed coke and coke-matrix vanishing occurred. It is suggested that the iron particle promotes gasification reaction of coke-matrix selectively around itself (and coke-matrix in that part is rapidly vanished). It is also suggested that this reaction mechanism maintains elastic modulus of coke-matrix because of the local rapid gasification reaction around the iron particle. Key words: ferrous coke; coke strength; porosity; matrix stiffness; coke gasification 1. Introduction The permeability through the blast furnace is reserved by the void of metallurgical coke bed. If the void is infilled by the lump-size degradation or coke breeze generation, the permeability is blocked and stable operation of blast furnace is conspicuously impeded. It is concerned that the CO2 gasification reaction which vanishes coke-matrix promotes pulverization and lump-strength degradation. In general, after gasification reaction, changes of coke microstructure in mm-scale are observed as follows: (i) Coke-matrix (solid) is vanished [1, 2]. (ii) Elastic modulus of coke-matrix decreases [3]. In practical knowledge, on the other hand, the pulverization resistance of highlyreactive coke (Coke Reactivity Index, CRI: gasification reactivity of 5 mm lump-coke at 1373 K in temperature) is higher than that of low CRI one while conversion (gasification ratio of whole lump) is constant. Nishi et al. [4, 5] have proposed the reaction mechanism which progresses via unreacted-core model as the coke has high pulverization resistance (Coke Strength after Reaction, CSR: Drum test of coke after CRI test) by gasification reaction. Watakabe et al. [1] have reported that spatial gradient of gasification ratio based on porosity of high CRI coke is larger than that of low CRI one. In addition, CSR of high CRI coke is higher than that of low CRI coke. It is suggested that the pulverization in outer region of the coke lump is selectively progressed since coke in outer region of the lump embrittles with progress of gasification reaction (coke-matrix vanishing and elastic modulus degradation) and unreacted-core strength is reserved. Thus, it is supposed that the lump-size degradation with the fracture is slightly progressed in high CRI coke and CSR increases. In this way, in order to prevent the pulverization and the
fracture, control of degradation part (reaction mechanism) in coke lump and that of the embrittlement behavior is significant. Addition of catalyst particle is promising as an effective and a simple method for advancement of the CO2 gasification reactivity (including CRI). Many kinds of alkali metal, alkali earth metal and transition metal have been known as the catalyst [6]. However, the catalysts are narrowed down to calcium [7] and iron [8] from a practical standpoint. Each of the catalyst activity is nearly equal. Anyway, gasification reaction is promoted on interface of the catalyst particle and coke-matrix. Lump embrittlement of coke which has 50 vol. % in pore is affected by cokematrix state as described previously (degree of vanishing and elastic modulus). However, the effect of the catalyst particle on coke-matrix after gasification reaction has not been investigated. In this study, iron-particles is used as the catalyst and the effect of iron particles on coke-matrix state after gasification reaction is investigated experimentally. Cokematrix vanishing is evaluated by spatial distribution of lump porosity and microscopic observation and elastic modulus of coke-matrix is evaluated by nano-indentation method. 2. Experiment 2. 1 Samples Coke lump with and without iron-particles was made. Both cokes are called ferrous coke and formed coke, respectively. Slightly-caking coal and non-caking coal were used. Blending ratio of slightly-caking and noncaking coals whose diameter was under 3 μm was 70/30 based on mass and both coals were mixed well. In making
Fig. 1 A photograph of a sample Table 1 Experimental conditions of CO2 gasification reaction MBR Mixing ratio [mass%] Reaction temperature [K] Reaction gas composition [vol%] Gas flow rate [NL/min] Conversion (Carbon basis) [mass%]
0 and 30 1173 CO2/CO = 100/0 and 50/50 5
Fig. 2 Schematic diagram of the experimental apparatus
20
of ferrous coke, mixed coals were also mixed with 30 mass % of iron ore particles whose diameter is under 250 μm. Mixed materials were pressed into 6 mL of briquette at 296 MPa and were carbonized at 1273 K for 6h. A photograph of a sample is shown in Fig. 1. 2. 2 CO2 gasification Schematic diagram of experimental apparatus is shown in Fig. 2. A coke sample was hanged from the weighing scale to alumina reaction tube filled with alumina ball for heat transfer to gas. Reaction tube was heated by electric furnace. Reaction gases were led into the reaction tube from the bottom. The gases after gasification were cooled by water-cooling tube, and were then ejected outside. Weight loss with gasification reaction was measured by weighing scale. Experimental conditions are shown in Table 1. Reaction temperature was set at 1173 K which is Thermal Reserve Zone (the area where CO2 gasification reaction of coke principally develops) temperature in case of using highly reactive coke. Reaction gas compositions were set at 100/0 and 50/50 in ratio of CO2/CO. Final conversion was 20 mass% based on carbon in coke. 2. 3 Spatial distribution of porosity Cross-sectional digital images were taken by optical microscope (LV-100-POL, Nikon) and the spatial distribution of porosity after gasification reaction was measured by image analysis (Winroof 5.01, Mitani Corporation). Conceptual diagram of the taking procedure of digital image is shown in Fig. 3. Coke samples were resin-mounted and were cut and polished. From end to end of coke samples, digital images were three times taken in each sample. Taking area of digital images (3.14 mm x
2.35 mm, 2.45 μm/pixel) was slid aside in half length of image size. 2. 4 Elastic modulus of coke-matrix 2. 4. 1 Measurement Elastic modulus of coke-matrix in mm-scale was measured by nano-indentation method. Resin-mounted specimens which are the same as mentioned in 2.3 were used again. Measurement part was outer region (vicinity of surface) and inner region (vicinity of center) of test specimen. Measurement conditions are shown in Table 2. The number of measuring points was 50 by each sample and gas composition. 2. 4. 2 Principle of nano-indentation method Load cycle indentation using sub-micron indentation instruments is now a means of determining the deformation properties such as hardness and elastic modulus. Fig. 4 shows schematic representation of a section through an indentation. A diamond tipped indenter with a precise geometry is pressed into a specimen with an increasing load up to a predetermined limit, and is then removed. The deformation properties can be determined using the load and displacement data obtained during the loading–unloading sequence. In this study, calculating method of elastic modulus was based on the method proposed by Oliver et al. When Berkovich triangular indenter which has 115-degree in angle is used, elastic modulus Eeff can be calculated by following formula: P hs = 0.75 max , (1) S (2) hc = hmax − hs , A ≈ 23.97hc2 ,
(3)
Fig. 4 Schematic representation of a section through an indentation
Fig. 3 Conceptual diagram of the taking procedure of digital image
Table 2 Measurement condition of nano-indentation method
Indenter Loading/unloading velocity Maximum load Holding time in maximum load The number of measurements
Berkovich triangular pyramid 3 100 2 50
[mN/s] [mN] [s]
100 Coarse pore 80
100 Coarse pores
-0.5
P oros ity [% ]
P o ro s ity [% ]
80 60
Coarse inert 40
-0.25
60 40
20
20
0
0
0 r/D [-]
0.25
0.5
(a) formed coke
-0.5
-0.25
End of lump after gasification
0 r/D [-]
0.25
0.5
(b) ferrous coke
Fig. 5 Spatial distributions of porosity after gasification (CO2/CO = 100/0). (a) formed coke, (b) ferrous coke
dP 2 S= Eeff A , = dh π 1 1 −ν 2 1 −ν i2 . = + Eeff E Ei
(4) (5)
3. Results and discussion 3. 1 In case of gas composition in CO2/CO = 100/0 3. 1. 1 Spatial distribution of porosity after gasification
Fig. 5 shows spatial distributions of porosity after gasification. In formed coke, porosity was distributed uniformly along the radial direction. In ferrous coke, meanwhile, porosity in outer region of lump was significantly larger than that in inner region. Relationships between porosity and conversion based on carbon mass of coke lump are shown in Fig. 6. In outer region, porosity increased with an increase in progress of gasification. On the other hand, in inner region, porosity hardly changed. It is suggested that chemical reactivity of gasification is
Fig. 6 Relationships between porosity and conversion based on carbon mass of coke lump In ferrous coke lump at reaction gas composition CO2/CO = 100/0
(a) formed coke
(b) ferrous coke
Fig. 7 Relationships between reaction time and conversion of (a)formed coke and (b) ferrous coke lump in CO2/CO = 100/0 advanced by the presence of iron-particles, and the gasification outer the coke lump is selectively progressed. Reaction controlling mechanism of ferrous coke lump and formed coke lump are estimated. Fig. 7 shows relationship between reaction time and conversion. In formed coke, the behavior of gasification progress is not that of homogeneous reaction model despite uniform porosity distribution. In ferrous coke, conversion apparently decreased from gasification initiation because of the oxidation of iron-particles by CO2. Then conversion increased. After a decrease in conversion, the progress behavior was similar to formed coke. It is suggested that reaction mechanism of both cokes is similar. Fig. 8 shows results of the reaction controlling mechanism estimation. The largest R2 is focused since the true mechanism shows linear plot, diffusion in boundary film-controlling is the true mechanism both ferrous coke and formed coke. Despite not the same spatial distribution of porosity in each
coke, each reaction controlling mechanism is the same. It is supposed that reaction gas diffuses into inner region of the lump through the pore inner coke after passing through the boundary film because of 50 vol. % in pore. In ferrous coke, it is supposed that the gas reacts rapidly soon after diffusing into the lump due to iron catalysis, and gasification in outer region of lump significantly progresses. In formed coke, meanwhile, it seems that the gas diffuses into inner the lump and reacts slowly compared with ferrous coke. 3. 1. 2 Elastic modulus of coke-matrix Elastic moduli of formed coke-matrix before and after gasification are shown in Table 3 and that of ferrous cokematrix are shown in Table 4. In formed coke, elastic modulus after gasification in inner region decreased with the progress of gasification. This result suggests gasification reaction progresses ununiformly and
0.20
0.18
0.18
0.16
0.16
0.14 R = 0.9915
0.12 0.10 2
R = 0.9879
0.08
0.12 f(xB) [-]
f(xB) [-]
2
R = 0.9991
2
0.14
0.10 0.08 2
R = 0.9979
0.06
0.06 0.04
0.04
2
R = 0.8878
0.02 0.00
0.02
2
R = 0.9079
0.00 0
4000
8000 Reaction time [s]
12000
16000
6000
(a) formed coke
8000
10000 Reaction time [s]
12000
14000
(b) ferrous coke
Fig. 8 Results of the reaction controlling mechanism estimation for (a) formed coke and (b) ferrous coke lump in CO2/CO = 100/0
Fig. 9 Microstructure of ferrous coke before and after gasification in outer region of coke lump in CO2/CO = 100/0 corresponds with the results as described previously. It has been known volume of nm-scale micro pore inside cokematrix increases with progress of gasification. In addition, a decrease of elastic modulus of coke-matrix is caused by an increase in the pore volume. The results in this study correspond with these results. On the other hand, in ferrous coke, the elastic modulus hardly changed with progress of gasification. Microstructure of ferrous coke before and after gasification in outer region of coke lump is observed as shown in Fig. 9. In inner region, iron-particles are completely surrounded by coke-matrix. In outer region, where is greatly gasified part, coke matrix surrounding iron-particles does not exist. Therefore, coke-matrix surrounding iron-particles is vanished with iron-particles. It is suggested that weight loss of whole ferrous coke lump is caused not by an increase in nm-scale pore but by the coke-matrix vanishing. Considering that a decrease in elastic modulus of coke-matrix does not occur, it is suggested that the vanishing rate of coke-matrix which contact iron-particles is greater than increasing rate of the
Table 3 Elastic modulus of coke matrix (formed coke) Outer region [GPa]
Inner region [GPa]
Before reaction
29.8 ± 4.0
31.1 ± 2.8
After reaction (CO/CO 2=0/100)
25.4 ± 4.6
29.2 ± 5.1
After reaction (CO/CO 2=50/50)
23.2 ± 3.2
31.9 ± 3.6
Table 4 Elastic modulus of coke matrix (ferrous coke) Outer region [GPa]
Inner region [GPa]
Before reaction
28.2 ± 3.6
30.9 ± 2.7
After reaction (CO/CO 2=0/100)
29.3 ± 4.4
29.2 ± 3.7
After reaction (CO/CO 2=50/50)
24.4 ± 4.6
31.2 ± 5.3
100
80
80
po ro sity [% ]
Coarse pore
Porosity [%]
100
60 Coarse inert
40
60 40 20
20 0 -0.5
-0.25
0
0
0.25
0.5
r/D [-] (a) formed coke
-0.5
-0.25
0
0.25
0.5
r/D [-]
(b) ferrous coke
Fig. 10 Spatial distributions of porosity after gasification (CO2/CO = 100/0)
(a) formed coke
(b) ferrous coke
Fig. 11 Results of the reaction controlling mechanism estimation and relationship between reaction time and conversion based on carbon mass. (a) formed coke, (b) ferrous coke. nm-scale pores. Probably, gasification reaction on the surface between iron-particles and coke-matrix is more rapid than the gas diffusion into the nm-scale pores. 3. 2 In case of gas composition in CO2/CO = 50/50 3. 2. 1 Spatial distribution of porosity after gasification Fig. 10 shows spatial distributions of porosity after gasification. Both ferrous coke and formed coke, porosity was distributed uniformly along the radial direction. Porosity of ferrous coke is slightly higher than that of formed coke. Reaction controlling mechanism of ferrous coke lump and formed coke lump are estimated. Fig. 11 shows results of the reaction controlling mechanism estimation and relationship between reaction time and conversion based on carbon mass of each coke lump. Both the behaviors of gasification progress are similar. In addition, in both ferrous coke and formed coke, the lump reaction mechanism is described by homogeneous reaction model. When CO is contained with reaction gas, CO
adsorbs competitively with CO2 to active site on cokematrix. Then CO2 gasification reaction is inhibited. Thus, the chemical reactivity in CO2/CO = 50/50 is smaller than that in CO2/CO = 100/0. So it is supposed that unreactedcore is not formed since reaction gas can diffuse enough into inner region of lump, even if iron-particles exist in the lump. 3. 2. 2 Elastic modulus of coke-matrix Elastic moduli of formed coke-matrix before and after gasification are shown in Table 3 and that of ferrous cokematrix are shown in Table 4. A decrease in elastic modulus of ferrous coke-matrix is smaller than that of formed coke, although elastic moduli of ferrous coke and formed coke decreased together with progress of gasification. Iron-particles inhibit a decrease in elastic modulus of coke-matrix not only in CO2/CO = 100/0 but also in CO2/CO = 50/50. Microstructure of ferrous coke before and after gasification is observed and is shown in
Iron particle
Iron particle
1 mm Before gasification reaction
1 mm After weight loss of 20 %
Fig. 12 Microstructure of ferrous coke before and after gasification in outer region of the lump in CO2/CO = 50/50 Fig. 12. Before gasification, iron-particles are completely surrounded by coke-matrix. After gasification, coke matrix surrounding iron-particles does not exist. It is clear that iron-particles vanish coke-matrix not only in CO2/CO = 100/0 but also in CO2/CO = 50/50. Considering these results, It is suggested that the iron particle promotes gasification reaction of coke-matrix selectively around itself (and coke-matrix in that part is rapidly vanished). It is also suggested that this reaction mechanism maintains elastic modulus of coke-matrix because of the local rapid gasification reaction around the iron particle. 4. Conclusion The effect of iron particles on coke-matrix state after gasification reaction is investigated experimentally. The results are concluded as follows: (i) CO2/CO = 100/0: In formed coke, spatial distribution of porosity was uniform. Elastic modulus of coke-matrix in outer region of the lump decreases with progress of gasification reaction. In ferrous coke, porosity in outer region of the lump was significantly larger than that in inner region which hardly changed with progress of gasification reaction. It is supposed that the gas reacts rapidly soon after diffusing into the lump due to iron catalysis, and gasification in outer region of lump significantly progress. In addition, elastic modulus of cokematrix hardly decreases and coke-matrix vanishing only occurred. (ii) CO2/CO = 50/50: Both ferrous coke and formed coke, porosity was distributed uniformly along the radial direction. The chemical reactivity in CO2/CO = 50/50 is smaller than that in CO2/CO = 100/0. So it is supposed that unreacted-core is not formed. Porosity of ferrous coke is higher than that of formed coke. In addition, in ferrous coke, a decrease in the elastic modulus of coke-matrix with progress of gasification is smaller than that in formed coke and coke-matrix vanishing occurred not only in CO2/CO = 100/0 but also in CO2/CO = 50/50. (iii) It is suggested that the iron particle promotes gasification reaction of coke-matrix selectively around itself (and coke-matrix in that part is rapidly vanished). It
is also suggested that this reaction mechanism maintains elastic modulus of coke-matrix because of the local rapid gasification reaction around the iron particle. Typical condition of gasification reaction is set in this study. The effect of iron-particles in actual blast furnace condition will be investigated in the future. Nomenclature A : Contact area E : Elastic modulus of sample Eeff : Effective elastic modulus of indenter and sample Ei : Elastic modulus of indenter hc : Indentation depth hmax : Maximum displacement hs : Surface displacement of contact boundary Pmax : Maximum load S : Tangent stiffness at initiation of unload Acknowledgements This study was carried as a part of the research activities "Fundamental Studies on Next Innovative Iron Making Process" programmed for the project "Strategic Development of Energy Conservation Technology Project". The financial support from New Energy and Industrial Technology Development Organization (NEDO) is gratefully acknowledged. References [1] S. Watakabe and K. Takeda: Tetsu-to-Hagané, 87 (2001), 467 [2] T. Ogata, K. Ueoka, Y. Morozumi, H. Aoki, T. Miura, K. Uebo and K. Fukuda: Tetsu-to-Hagané, 92 (2006), 171 [3] H. Hayashizaki, K. Ueoka, T. Ogata, Y. Yamazaki, Y. Matsushita, H. Aoki, T. Miura, K. Fukuda and K. Matsudaira: Tetsu-to-Hagané, 95 (2009), 460 [4] T. Nishi, H. Haraguchi and Y. Miura: Tetsu-to-Hagané, 70 (1984), 43 [5] T. Nishi, H. Haraguchi and T. Okuhara: Tetsu-toHagané, 73 (1987), 1869
[6] J. Lahaye and P. Ehriburger: Fundamental Issues in Control of Carbon Gasification Reactivity, Kluwer Academic Publishers, 1991 [7] S. Nomura, H. Ayukawa, H. Kitaguchi, T. Tahara, S. Matsuzaki, S. Koizumi, Y. Ogata, T. Nakayama and T. Abe: ISIJ Int., 45 (2005), 316 [8] H. Fujimoto and M. Sato, J. Jpn. Inst. Energy, 89 (2010), 21
Spectral Emissivities of Ni and Fe based Boiler Tube Materials with Varying Chromium Content at High Temperature Atmospheres Miki Shimogori*, Fabian Greffrath**, Viktor Scherer** Alfred Gwosdz*** and Christian Bergins*** * Babcock-Hitachi K.K.Kure Research Laboratory ** Ruhr University of Bochum *** Hitachi Power Europe GmbH
ABSTRACT To evaluate radiative heat transfer characteristics of boiler tube materials, spectral emissivities of 6 types of Fe-based and Ni-based alloys with varying chromium contents of 2, 9, 18, 23 and 25% have been investigated under high temperature oxidizing conditions. Metal samples were heated in an electric furnace and spectral emissivities were determined by using a monochromator system at wavelengths from 1 to 17µm in the temperature range from 773 to 1073K. After the measurements, the sample surfaces were analyzed by means of SEM (scanning electron microscope). The experimental values of the spectral and total emissivities were discussed in relationship to chromium content in metals. The obtained results are as follows: (1) Spectral emissivities of metals containing more than 9% of chromium have the typical wavelength and temperature dependencies of metal emissivities; they decrease with increasing wavelength and increase with increasing temperature. In contrast, spectral emissivities of the metal containing 2% of chromium have weak dependencies both on wavelength and temperature; they were high under most measurement conditions and changed periodically with the wavelength, (2) Spectral emissivity levels tend to decrease with increasing chromium content in metals. The SEM analysis showed that the thickness of the oxidation layer formed on the sample surface was small for the sample containing higher amounts of chromium, and (3) Total emissivity also tends to decrease with increasing chromium content in metals. SEM observation indicates that the growth of the oxidation layer increases metal emissivity. Based on SEM and emissivity measurement results, it appears that the difference in spectral emissivities among metals with varying chromium content is due to the thickness of the oxidation layer.
1.INTRODUCTION In the fields of industrial boilers, where radiative heat transfer is dominant, knowledge of the emissivities of boiler tube materials is crucial for the design. Emissivities of real surfaces are very sensitive to their surface states; they depend on material, chemical components, roughness etc 1)-3). In boilers, metal surfaces are exposed to high temperature oxidizing atmospheres. Changes in operating conditions and the usage of different tube materials make estimating emissivity of actual heat transfer surfaces a serious challenge.
1
In early studies, a considerable number of attempts have been made to obtain values of spectral emissivity of metals. Previous studies have mainly been conducted for pure metals such as platinum, copper, alumina and tungsten in reducing atmosphere4)-6). These studies mainly focus on space technology. Recently, spectral emissivity characteristics of commercial alloys have been discussed in relationship to steel grades, surface treatment processes and atmospheres
7)8)
. They showed that these factors severely affect spectral
emissivities. Unlike studies made extensively on specific engineering fields, emissivity data of alloy tube materials to evaluate heat transfer performances in boilers is still scarce. The purpose of this paper is to evaluate radiative heat transfer characteristics of boiler tube materials in high temperature oxidizing atmospheres. For this purpose, spectral emissivities of Fe-based and Ni-based alloys with varying chromium contents of 2, 9, 18, 23 and 25% have been determined by using a monochromator system at wavelengths from 1 to17µm in the temperature range between 773 and 1073K.
2. Nomenclature
:Stefan-Boltzmann constant (5.67×10-8 W/m2K4) Ss :radiant intensity of sample [W/m3] :radiant intensity of black body [W/m3] SR λ :wavelength [m] T :temperature [K] ε(λ,T) :spectral emissivity at temperature T [-] ε(T) :total emissivity at temperature T [-] Eb(λ,T) :monochromatic emissive power of black body [W/mK] σ
3.EXPERIMENTAL STUDY 3.1 Experimental set up and procedure All experiments described in this paper have been carried out at a radiation test rig of the department of energy plant technology (LEAT) at the Ruhr-University Bochum, Germany 9). This test rig, as depicted in Fig. 1, consists of two main parts: The first part comprises the heating unit holding both the sample in the sample holder and the blackbody reference radiator, whereas the second part consists of the spectral radiometer. The heating element is a vertical, hollow cylindrical electric furnace, which is used to heat both the sample and the blackbody reference radiator up to the measuring temperatures. To avoid heat loss it is covered with insulating slabs at the top and bottom. The blackbody is laid out as a cavity radiator and made of a heat resistant nickel-based alloy. The cylindrical samples have a diameter of 22mm and 4mm height and sit in a sample holder, which in turn lies on top of the blackbody. The blackbody sits on a ceramic tube, which can be shifted up and down inside the heating element by means of a Bowden cable. For measurement of the sample radiation the sample holder is shifted up to the top of the heating element directly beneath the insulation slab. The radiation emitted vertically by the sample surface leaves the
2
heating element through an optical access and is detected by the spectral radiometer. Temperatures are determined by type K thermocouples in the sample and the sample holder as well as type N thermocouples in the bottom of the blackbody reference radiator.
(2) Signal detecting parts (h) Detectors
(g) Monocrometor
(b) Sample
(a) Electric furnace
(c) Sample holder (f) Insulation (d) Black body
(e) Thermocouples (1) Sample heating parts
Figure 1 Schematic drawing of the set up 9)
(2) Signal detecting parts
(3) Data collecting parts
(h) (n)
(m) Grating (k) (j)
(o) Personal computer (l)
(g) Monochromator (i)
Figure 2 Operation mode of the spectral radiometer 9) Radiation measurement is carried out by a computer-aided spectral radiometer system that is based on a Czerny-Turner grating monochromator, as depicted in Fig. 2. Thermal radiation emitted by the sample (or the blackbody respectively) passes through the optical access and hits a gold-plated off-axis mirror that reflects the radiation signal onto the entrance slit of the monochromator system. In front of this slit the signal is periodically intercepted by a rotating aperture plate, a so-called chopper, which by this means
3
imposes a modulation of known phase and frequency on the signal. This modulation is synchronized with a lock-in amplifier that in turn is able to gain actual sample signals and provide high signal-to-noise ratios. Inside the monochromator radiation is collimated by a curved mirror and diffracted off of three gratings, each with a different grating constant, which are mounted onto a rotatable turret. The dispersed radiation is then reflected by another mirror and refocused on the exit slit, where a lN2-cooled MCT-detector (Mercury-Cadmium-Tellur) is installed. To avoid superimposition of diverse diffraction maxima of different wavelengths a filter wheel with six edge filters is inserted into the optical path. The stepping motors that select the appropriate optical filter as well as the grating and diffraction angle according to the target wavelength are controlled by computer software. For the precise alignment of the spectrometer with the sample a laser pointer is used prior to the measurements by transferring its beam back from the detector position through the whole monochromator system onto the sample surface. Determination of spectral emissivities is based on a static radiation comparison method, in which the ratios of the detected radiance signal from the sample and the reference radiator at the same temperature and wavelength are calculated (see equation 1):
ε (λ , T ) =
S S (λ , T ) S R (λ , T )
(1)
The spectral emissivities are determined in the wavelength range from 0,98µm to 17,02µm using a step size of 40nm, resulting in a total of 402 records per measurements. The measuring temperatures can be adjusted in a range between about 773K and 1273K by means of a thyristor-operated furnace control. The lower temperature limitation results from the operational range of the heating element, whereas the upper limit is restricted by the thermal stress of the blackbody material. The wavelength interval is limited by the sensitivity of the used detector. The step size of 40nm has been chosen as a good compromise between spectral resolution and acceptable measuring expenditure. 3.2 Preparation of metal samples Table 1 shows the classification of metals used for the current series of emissivity measurements and Table 2 lists their chemical components. Flat, disk-shaped test samples with a diameter of 22mm and a thickness of 4mm were prepared from plates or tubes of the corresponding metals. To measure the sample temperature, a Type K thermocouple with 1mm diameter was inserted into a hole at the side of the sample and located 2mm beneath the center of the disk surface. To investigate the effect of surface oxidation, additional pre-oxidized samples were prepared for each metal by exposing the samples to a temperature of 973K for 24hours in an electric furnace.
4
Table 1 Classification of metal samples used for a series of tests Classification
Materials 7CrMoVTiB10-10 T92 SUS304H HR6W HR3C Alloy617
Nominal components 2Cr 9Cr 18Cr-8Ni 23Cr-43NiWNbTi 25Cr-20Ni Ni-based
Fe based Fe-Ni based Fe based Ni based
Table 2 Chemical components of metal samples used for a series of tests (unit; wt%) Materials
C
Si
Mn
P
S
Ni
Cr
W
Ti
Nb
Fe
7CrMoVTiB10-10 T92 SUS304H HR6W HR3C Alloy617
0.08 0.09 0.08 0.09 0.05 0.08
0.30 0.22 0.22 0.29 0.41 <0.30
0.50 0.45 0.76 0.92 1.22 <0.30
<0.02 0.02 0.03 0.01 0.02 <0.01
<0.1 <0.01 <0.01 <0.01 <0.01 <0.01
0.20 9.33 44.64 19.79 Balc.
2.40 8.81 18.73 22.83 24.82 23.00
1.64 6.87 -
0.08 0.09 0.40
0.06 0.52 0.18 0.45 -
Bal. Bal. Bal. Bal. Bal. <1.5
4. RESULTS AND DISCUSSTION 4.1 Spectral emissivity of metal samples Spectral emissivities of metal samples without pre-oxidation are depicted in Figures 3 to 8. All specimens were tested at 773, 873, 973, and 1073K. The pictures embedded in each figure present the sample surface after measurement. As seen in Figure 8, the damage on the surface of 7CrMoVTiB10-10 was severest among samples tested. Because of its weak oxidation resistance due to low chromium content, an oxidation layer was formed on the surface of 7CrMoVTiB10-10 and partially flaked. In contrast, no remarkable damage was observed on the surfaces of other samples. As shown in Figures 3 to 7, spectral emissivities of metals containing more than 9% of chromium were higher at the shorter wavelength and lower at the longer wavelength. Furthermore, the spectral emissivities increased with an increase in measurement temperature. The results of 7CrMoVTiB10-10 shown in Figure 8 differed from other results, as its spectral emissivity showed periodic variation in the wavelength range from 8 to 16µm.
5
Without pre oxidation (after measurements)
Spectral emissivity [ - ]
Cr 9% %
836K 927K 1019K 1107K
Wavelength[nm] Figure 3 Spectral emissivities of T92 without pre-oxidation as a function of wavelength in oxidizing atmosphere Without pre oxidation (after measurements)
Spectral emissivity [ - ]
Cr 19% %
835K 927K 1023K 1116K
Wavelength[nm] Figure 4 Spectral emissivities of SUS304 without pre-oxidation as a function of wavelength in oxidizing atmosphere
6
Without pre oxidation (after measurements)
Spectral emissivity [ - ]
Cr 23% %
843K 932K 1018K 1110K
Wavelength[nm] Figure 5 Spectral emissivities of HR6W without pre-oxidation as a function of wavelength in oxidizing atmosphere Without pre oxidation (after measurements)
Spectral emissivity [ - ]
Cr 25% %
847K 935K 1025K 1115K
Wavelength[nm] Figure 6 Spectral emissivities of HR3C
without pre-oxidation as a function of wavelength in
oxidizing atmosphere
7
Without pre oxidation (after measurements)
Spectral emissivity [ - ]
Cr 23% %
848K 933K 1017K 1098K
Wavelength[nm] Figure 7 Spectral emissivities of Alloy617 without pre-oxidation as a function of wavelength in oxidizing atmosphere
Spectral emissivity [ - ]
Cr 2% %
Without pre oxidation (after measurements)
838K 928K 1014K 1103K
Wavelength[nm] Figure 8 Spectral emissivities of 7CrMoVTiB10-10
without pre-oxidation as a function of
wavelength in oxidizing atmosphere
8
4.2 Effect of surface oxidation Considering that the oxidation layers grow during the measurements, temperature dependence was assumed to be dominated by the layer growth. Hence it is assumed that if the sample surface was oxidized enough before measurement, temperature dependencies of spectral emissivity will decrease. To confirm this, pre-oxidized samples were tested. As expected, the temperature dependences of all samples with pre-oxidation were smaller than those without pre-oxidation. For example, in case of HR6W (in Fig. 9), the increase in spectral emissivity with increasing temperature was less than 0.05 across all measured wavelengths ranges. As discussed later in Fig.12, the difference in emissivity spectra among each metal also decreased by pre-oxidation. In the case of 7CrMoVTiB10-10 containing 2% of chromium, there is another point worth mentioning. The emissivity spectra of the pre-oxidized specimens (in Fig. 10) showed pronounced tendencies of periodic variation in comparison with those without pre-oxidation (in Fig. 8). More specifically, in pre-oxidized cases, the cycle of variation in spectral emissivity became shorter and the variation width decreased. This periodic change diminished as measurement temperature increased (see the results at 1014 and 1105K). Similar periodic tendencies have been observed in some early studies2), 7), 10) conducted for oxidized metal samples. Our results obtained for 7CrMoVTiB10-10 are consistent with them. Ane et al.10) have studied the applicability of the stratified media theory to the calculation of the emissivities of oxidized stainless steel. They concluded that the theory gives a good interpretation of experimental data, however, it is difficult to fit them. With pre oxidation 24h at 973K
(after measurements)
Spectral emissivity [ - ]
Cr 23% %
843K 932K 1018K 1110K
Wavelength[nm] Figure 9 Spectral emissivities of HR6W with pre-oxidation as a function of wavelength in oxidizing atmosphere
9
Spectral emissivity [ - ]
Cr 2% %
With pre oxidation 24h at 973K (after measurements)
833K 921K 1014K 1105K
Wavelength[nm] Figure 10 Spectral emissivities of 7CrMoVTiB10-10 with pre-oxidation as a function of wavelength in oxidizing atmosphere 4.3 Effect of chromium contents in metals To clarify the effect of chromium content on spectral emissivity, spectral emissivities of all metals containing different chromium contents were compared at 923K. A temperature of 923K is a typical design temperature of boiler super heater tubes. Figures 11 and 12 present the results of samples without and with pre-oxidation, respectively. In Fig.11, a general tendency can be detected. The spectral emissivity levels throughout the range of measurements decrease in the following order: 7CrMoVTiB10-10 > T92 > SUS304 > Alloy617 > HR6W > HR3C. Considering that chromium is an oxidation resistant material, this result clearly indicates that spectral emissivity is lower for metals with higher resistance of oxidation. As shown in Fig.12, for samples with pre-oxidation, the effect of chromium content on spectral emissivity became smaller. More specifically, the difference in spectra among each metal decreased and all curves of spectra have a common "arc" shape, with the exception of 7CrMoVTiB10-10. Notably, the emissivity of Alloy 617 increased dramatically at wavelengths shorter than 6µm.
10
1.0
Spectral emissivity(-)
0.8
0.6
0.4
7CrMoVTiB10-10(2Cr) T92 (9Cr) SUS304H (18Cr) HR6W (23Cr) HR3C (25Cr) Alloy617 (23Cr)
0.2
0.0 0
2000
4000
6000 8000 10000 12000 14000 16000 Wavelength(nm)
Figure 11 Spectral emissivities of alloys with varying Cr contents (without pre-oxidation, measurement temperature 923K) 1.0
Spectral emissivity(-)
0.8
0.6
0.4 7CrMoVTiB10-10(2Cr) T92 (9Cr) SUS304H (18Cr) HR6W (23Cr) HR3C (25Cr) Alloy617(23Cr)
0.2
0.0 0
2000
4000
6000 8000 10000 12000 14000 16000 Wavelength(nm)
Figure 12 Spectral emissivities of alloys with varying Cr contents (with pre-oxidation, measurement temperature 923K)
4.4 Surface analysis results To analyze the oxidation conditions of each sample surface, SEM analyses of metal samples have been conducted after measurements. Fig.13 gives SEM images of tested samples taken at the cross-section near their surfaces. SEM observation shows that the thickness of the scale of 7CrMoVTiB10-10 was about 10µm and the intergranular oxidation layer was formed under the scale. This indicates that the sample was heavily oxidized during measurement and corresponds to a visual analysis showing that a flaked oxidation layer was formed on its surface. In the cross-section image, a flaked scale with a micron-order thickness
11
was also observed for the case of T92. For other samples, SUS304, HR6W and HR3C, no flaked scale was observed. In terms of Alloy617, it should be noted that internal oxide caused a specific morphology on its
Internal oxide
surface.
Internal oxide
Figure 13 Photographs of sample surfaces and SEM images of cross-section of sample surfaces The SEM analysis shows that the thickness of the oxidation layer was small for the samples containing higher amount of chromium. This indicates that the growth of the oxidation layer increases metal emissivity. From a series of studies, it appears that the difference in spectral emissivities among metals with varying chromium content is due to the different thickness of the oxidation layer.
12
One may remember that the emissivity of Alloy 617 increased dramatically at the shorter wavelength after pre-oxidation. At this moment, there is no further information available to discuss the relations between specific surface morphology of Alloy617 and its emissivity. Recent studies
11), 12)
have suggested
that the surface asperity will control spectral emissivity characteristics, which is called as "micro-cavity effect". This will be examined in future work. 4.5 Total emissivity of each alloy Total emissivities were calculated using spectral emissivity data of tested metals (see equation2):
∫ ε(λ ,T )Eb (λ ,T )dλ ε(T ) = λ1 λ 2 ∫λ1 Eb (λ ,T )dλ λ2
(2)
Fig.14 shows temperature dependent total emissivities of the 6 tested metals with and without pre-oxidation. In the case of samples without pre-oxidation (in the left figure), total emissivities increased with an increase in temperature and emissivity levels clearly decreased with increasing Cr content, such as: 7CrMoVTiB10-10(Cr:2%)>T92(Cr:9%)>SUS304H(Cr:19%) ≒ Alloy617(Cr:21-23%)>HR6W(Cr:23%) >HR3C(Cr:25%). In the case of samples with pre-oxidation, except for 7CrMoVTiB10-10, total emissivities also increased with an increase in temperature. Total emissivity of 7CrMovTiB10-10 is almost constant at 0.8 throughout the range of tested wavelengths. The total emissivities are ranked in the following order: 7CrMoVTiB10-10 (Cr:2%)> Alloy617(Cr:21-23%)> SUS304H(Cr:19%)>T92(Cr:9%)≒HR3C(Cr:25%) >HR6W(Cr:23%). The change in emissivity levels of each sample with and without pre-oxidation depends on metal material. For example, the emissivity levels of HR6W and SUS304H did not change remarkably whether pre-oxidation treatment was done or not, however, that of T92 decreased and those of HR3C and Alloy617 increased. Keeping in mind that actual boiler tubes have oxidized surfaces, the emissivity data of pre-oxidized ones should be applied to evaluate heat transfer characteristics.
13
1.0
1.0
with pre-oxidation
without pre-oxidation 0.8
Total emissivity[ - ]
Total emissivity[ - ]
0.8
0.6
0.4 7CrMoVTiB10-10 T92 SUS304H HR6W HR3C Alloy 617
0.2
0.6
0.4
7CrMoVTiB10-10 T92 SUS304H HR6W HR3C Alloy 617
0.2
0.0
0.0 773
873 973 1073 Sample temperature[K]
1173
773
873 973 1073 Sample temperature[K]
1173
Figure14 Temperature dependence of total emissivities of 6 types of alloys with and without pre-oxidation
5. CONCLUSIONS To evaluate radiative heat transfer characteristics of boiler tube materials, spectral emissivities of 6 types of Fe-based and Ni-based alloys with varying chromium contents of 2, 9, 18, 23 and 25% have been investigated at wavelengths from 1 to 17µm in high temperature oxidizing atmospheres ranging from 773 to 1073K. The main conclusions are summarized as follows: (1) Spectral emissivities of metals containing more than 9% of chromium have typical wavelength and temperature dependence of metals; they decrease with increasing wavelength and increase with increasing temperature. In contrast, spectral emissivity of the metal containing 2% of chromium has weak dependences both on wavelength and temperature; it is high at almost all measurement conditions and changes periodically along with wavelength. (2) Spectral emissivity levels tend to decrease with increasing chromium content in metals. The SEM analysis showed that the thickness of the oxidation layer formed on sample surface was small for the sample containing higher amount of chromium. (3) Total emissivity also tends to decrease with increasing chromium content in metals. (4) SEM observation indicates that the growth of the oxidation layer increases metal emissivity. Based on SEM and emissivity measurement results, it appears that the difference in spectral emissivities among metals with varying chromium content is due to the thickness of the oxidation layer.
14
REFERENCES 1) R. Siegel and J. R. Howell, Thermal radiation heat transfer, 4th edition, Taylor & Francis, New York, 2001. 2) M. F. Modest, Radiative heat transfer, McGRAW-HILL, New York, 1993. 3) A. G. Emslie, ”A review of some problem areas in the theory of thermal radiation of solids”, Symposium on thermal radiation of solids, pp.3-9, San Francisco, March 1964. 4) H. E. Bennett, “Influence of Surface Roughness, Surface Damage, and Oxide Films on Emittance”, Proceedings of Symposium on Thermal Radiation of Solids, pp.145-152, NASA SP-55, 1965. 5) H. H. Blau, “Surface Properties of Metals”, Proceedings of Symposium on Thermal Radiation of Solids, pp.159-164, NASA SP-55, 1965. 6) H. E. Clark, D. G. Moore, “Method and Equipment for Measuring Thermal Emittance of Ceramic Oxides from 1200 to 1800K”, Proceedings of Symposium on Thermal Radiation of Solids, pp.241-258, NASA SP-55, 1965. 7) W. Bauer, W. Gräfen, and M. Rink, “Spectral Emissivities of Heat-treated Steel Surfaces”, AIP Conference Proceedings, Vol.7, pp.807-812, 2003. 8) T. Kruczek, Z. Rudnicki and A. Sachajdak, “Measurements of the Spectral Emissivity of Steel in a gaseous Reduction Atmosphere”, Proceedings of the 3rd International Symposium on Radiation Transfer, “Radiation Transfer–III”, Antalya, Turkey, 2001. 9) S. Bohnes, V. Scherer and S. Wirtz,“Methodik zur experimentellen Untersuchung der Strahlungs- und Wärmeleitungschaften von Aschen und Ablagerungen aus Kraftwerkskesseln ” , VGB PowerTech, vol.89/2009 , pp.71-81, April 2009. 10) A. M. Ane and M. Huetz-Aubert, "Stratified Media Theory Interpretation of Measurements of the Spectral Polarized Directional Emissivity of Some Oxidized Metals", Int. Journal of Thermophysics, Vol.7, No.6 (1986), pp.1191-1208. 11) H. Sai, Y. Kanamori, K. Hane, and H. Yugami, "Numerical Study on Spectral Properties of Tungsten One-dimensional Surface-relief Gratings for Spectrally Selective Devices", J.Opt.Soc.Am.A,22 (2005), pp.1805-1813. 12) K. Hanamura, and Y. Kameya, "Spectral Control of Thermal Radiation Using Rectangular Micro-cavities on Emitter-surfaces for Thermophotovoltaic Generation of Electricity", Journal of Thermal Science and Technology, Special issue, 3 (2008), pp.1-12.
15
Twenty-Seventh Annual International PITTSBURGH COAL CONFERENCE Istanbul, TURKEY, October 11-14, 2010
Effect of MgO Additive on the Reduction of Ash Deposition of Upgraded Brown Coal Katsuya Akiyamaa, Haeyang Paka, Yasuaki Uekib, Ryo Yoshiiec, Ichiro Narusec E-mail address: [email protected] (K.Akiyama) Mechanical Engineering Research Laboratory, KOBE STEEL, LTD., 1-5-5 Takatsukadai, Nishi-ku, Kobe, 651-2271, JAPAN b Energy Science Division, EcoTopia Science Institute, NAGOYA UNIVERSITY, Furo-cho, Chikusa-ku, Nagoya, 464-8603, JAPAN c Department of Mechanical Science & Engineering, NAGOYA UNIVERSITY, Furo-cho, Chikusa-ku, Nagoya, 464-8603, JAPAN a
Abstract Ash with low melting temperatures causes slagging and fouling problems in pulverized coal combustion boilers. Ash deposition on the heat exchange tubes affects the decrease in the overall heat transfer coefficient due to the low thermal conductivity of the ash as well as several other operation problems. Therefore, the operational conditions of the boilers directly relate to the ash deposition behavior. The objectives of this study are to evaluate the effect of MgO addition with the coal on the reduction of ash deposition during upgraded brown coal (UBC) combustion and to understand the reduction mechanisms of ash deposition. The melting temperature of the UBC ash is 1494 K, which is relatively lower than that of bituminous coal ash. Before the actual ash deposition experiments, the molten slag fraction in the UBC ash was estimated by means of chemical equilibrium calculations while varying the mixing mass ratio of MgO to coal ash. The results of a simulation indicate that the MgO addition played a role in decreasing in molten slag fraction. It was confirmed that Mg formed solid composites with Si, Fe, Al, Ca, and Mn and played a role in decreasing the molten slag fraction in ash on the tube. As a next step, ash deposition tests were conducted using a pilotscale pulverized coal combustion furnace equipped with a refractory wall. The thermal load in the furnace was fixed at 149 kW, and the maximum gas temperature exceeded 1750 K. A water-cooled stainless-steel tube was inserted at 1573 K in the furnace to measure the quantity of the ash deposits. As a result, the MgO addition contributed to the decreasing rate of ash deposition even for the UBC. These calculations and experimental results suggested that the MgO addition contributed to the reducing UBC ash deposition on the tube. Key Words: Reduction of ash deposition, MgO additive, Upgraded brown coal (UBC)
1. Introduction Bituminous coals, which are categorized as high-rank coals, are most commonly consumed in industrial sectors. In contrast, demand for a low-rank coals, such as sub-
bituminous coal, lignite, and brown coal, is limited because of its lower calorific value and/or higher moisture content. Therefore, low-rank coal is usually utilized as fuel only in some specific local regions around its mines. Recently, there have been demands for a technology to utilize the low-rank coals as a common fuel because of the limited availability of reserves of the high-rank coals. As some Indonesian low-rank coals have high moisture contents, their heating values are relatively low. However, the contents of the ash, nitrogen, and sulfur in these coals are smaller. If the effective drying or dewatering technologies for these coals could be developed, it would be possible to use them as new fossil fuels. Kobe Steel has developed an UBC process based on the slurry dewatering process [1-3]. Akiyama et al. also evaluated the combustion characteristics of UBC [4]. However, the ash melting points of Indonesian low-rank coals are relatively low. Therefore, slagging and fouling problems may occur in pulverized coal combustion boilers. The ash deposits on the heat exchanger tubes reduce the overall heat transfer coefficient because of its low thermal conductivity. Consequently, it is difficult to use such low-rank coals alone in the pulverized coal combustion boiler. Ash deposition phenomena are influenced by coal type (ash compositions, melting temperature, and distribution of mineral matter), reaction atmosphere, particle temperature, the surface temperature of heat exchanger tubes and tube materials, flow dynamics, and so forth. Several reviews relating to the ash deposition characteristics have already been reported [5]. For instance, Raask elucidated the deposit initiation [6], and Walsh et al. [7] and Baxter [8] studied the deposition characteristics and growth. Beer et al. [9] attempted to develop theories of ash behavior. Benson et al. [10] summarized the behavior of ash formation and deposition during coal combustion. Naruse et al. [11] evaluated the ash deposition characteristics under high-temperature conditions. Vuthaluru et al. [12] evaluated the ash formation of brown coal. Li et al. [13] investigated coal char-slag transition under oxidation conditions. Bai et al. [14] studied the characterization of low-temperature coal ash at high temperatures under a reducing atmosphere. Abbott et al. [15] investigated the adhesion force between slag and oxidized steel. Harb et al. [16] predicted ash behavior using a chemical equilibrium calculation. Hansen et al. [17] quantified ash fusibility using differential scanning calorimetry, and Ichikawa et al. [18] measured the liquid phase ratio of ash using differential thermal analysis. Song et al. [19] investigated the effect of coal ash composition on ash fusion temperatures. Even for those references, however, precise and quantitative knowledge of the deposition of coal ash with a low-melting temperature during coal combustion has been insufficient. Our previous studies [20] have proven that the molten slag fraction in ash obtained by chemical equilibrium calculations is one useful index with which to predict the coal blending method to reduce the deposition fraction of ash. The objectives of this study are to evaluate the effect of MgO addition into the coal on the reduction of ash deposition for UBC and to understand the reduction mechanisms of ash deposition. The melting temperature of the UBC ash is 1494 K, which is relatively lower than that of bituminous coal ash. Before the actual ash deposition experiments, the molten slag fraction in the UBC ash was estimated by means of chemical equilibrium calculations while changing the mixing mass ratio of MgO to coal ash. As the next step, ash deposition tests were also conducted using a pilot-scale pulverized coal combustion furnace equipped with a refractory wall.
2. UBC Process Figure 1 shows a schematic diagram of the UBC process. The UBC process developed by Kobe Steel can increase the quality of low-rank coal by means of a slurry dewatering technique at a low temperature and pressure without chemical reactions. Kobe steel started
developing this UBC process for Australian brown coal in 1993. The 3 t/d pilot plant for Indonesian UBC has been operated successfully since 2001. Recently, the 600 t/d demonstration plant has been operated in Satui, South Kalimantan. In the UBC process, the low-rank coal is first crushed and then is mixed with oil and a small amount of asphalt to produce the slurry. This coal–oil slurry is heated in order to be dewatered. The UBC produced adsorbs very little moisture since the asphalt can deactivate the active sites effectively. After the coal is separated from the oil, it is dried and pressed to make transportable briquettes. As mentioned above, the UBC process can effectively remove the moisture in low-rank coal and can increase its heating value without changing its other properties. Therefore, the UBC produced from Indonesian low-rank coal has useful properties, such as low nitrogen, sulfur, and ash contents. However, this process cannot change the melting point of UBC ash any more. 1 6 5
M
2
1. 2. 3. 4. 5.
M
8
9
10
4 7 11 3
6. 7. 8. 9. 10. 11.
Raw Coal Additive (Asphalt) Recycle Oil Slurry Making Unit Slurry Dewatering Unit (Evaporator) Steam Compressor Waste Water Centrifugal Machine Dryer Roll Type Briquette Machine UBC Briquette
Figure 1. Schematic diagram of UBC process.
3. Chemical Equilibrium Calculations We used the Indonesian UBC and three types of bituminous coal with different melting temperatures and ash compositions, as shown in Table 1, for the calculations. As seen from the table, the initial deformation temperature of Coal A under the oxidizing condition is the highest among them. Prominent features of Coal B and Coal C are large CaO and Na2O contents in the ash, respectively. For the ash in the UBC, Fe2O3 content was the largest. The ash particles tend to adhere more on the tube as the amount of molten particles increases. In this calculation, therefore, we calculated the molten fraction of each ash by chemical equilibrium theory, using Fact Sage Ver. 6.1 software. Table 2 shows the conditions of the chemical equilibrium calculations. The gaseous compositions given were assumed to be near a burner in the pulverized coal combustion boiler. In the calculations, the partial pressures of CO, CO2, and H2 gas were fixed. We performed these calculations in 50 K increments to determine the mass percentage of molten slag in the ash in the temperature range 1273–2073 K.
4. Experimental Section Figure 2 shows a schematic diagram of the pilot-scale pulverized coal combustion refractory furnace used in this study. The furnace was 3.65 m long and 0.4 m in inner diameter. It was insulated by a refractory material to reduce the heat loss. The pulverized coal weighed by the volumetric feeder was transported to the burner with nitrogen gas. The primary air at ambient temperature and the secondary air heated at 573 K were supplied separately. City gas was supplied to the furnace in order to heat it to 1750 K. In the
experiments, we inserted a thermocouple and a gas sampling probe into the furnace to measure the gas temperature and gaseous concentrations, respectively. The coal was pulverized with 85% in mass fraction less than 75 μm. Table 1. Properties of coals tested. Coal sample Heating value [M J/kg] Proximate analysis [wt%, dry]
Fuel ratio [-] Ultimate analysis [wt%, d.a.f]
Ash fusion temperature [K]
Ash compositions [wt%]
Ash Volatile matter Fixed carbon Carbon Hydrogen Nitrogen Sulfur Oxygen (B alance ) Oxyidizing Initial deformation Hemisperical Fluid Reducing Initial deformation Hemisperical Fluid SiO2 Al2 O3 CaO Fe 2 O3 M gO Na 2 O K2 O SO3 P2 O5 TiO2 V2 O 5 M nO
Coal A 29.02 12.18 32.83 54.99 1.67 81.39 5.26 1.83 0.50 11.02 1753 >1823 >1823 1607 1821 >1823 65.90 22.90 1.08 4.58 0.79 0.47 1.37 0.62 0.31 1.37 0.07 0.03
Coal B 27.00 6.46 34.60 58.94 1.70 80.94 4.75 1.03 0.17 13.11 1433 1473 1493 1373 1453 1473 28.85 17.53 24.97 8.79 1.22 0.59 0.22 12.38 0.17 0.56 0.01 0.25
Coal C 29.94 8.53 38.63 52.84 1.37 78.61 5.27 1.42 0.16 14.54 1393 1423 1453 1393 1423 1453 53.04 10.03 14.54 5.25 2.49 2.15 0.78 6.96 0.11 0.56 <0.01 0.02
UBC 25.27 6.27 49.09 44.64 0.91 72.91 5.65 1.05 0.34 20.05 1440 1494 1651 1433 1453 1473 41.10 12.50 10.30 16.30 8.36 0.13 1.50 8.34 0.04 0.64 0.03 0.26
Table 2. Conditions of the chemical equilibrium calculations. Temperature (K) Gas composition (%)
Ash composition (wt%)
1273 ~ 2073 8.3E-8 O2 CO 2 12.3 CO 8.2 1.5 H2 N2 70.6 7.4 H2 O SiO 2 , Al2 O 3 , CaO, Fe2 O 3 , MgO, Na2 O, K 2 O, SO 3 P 2 O 5 , TiO 2 , V2 O 5 , MnO
We used MgO with a mean particle diameter of 0.2 μm and purity of 99.9% as an additive to the UBC. The pulverized UBC and MgO additive were mixed using a mixer before the ash deposition test. Table 3 shows the experimental conditions. The heat load was fixed at a given value of 149 kW for all the tests. The oxygen concentration of the flue gas at the furnace outlet was fixed at a given value of 1.0% as a dry basis. The oxygen concentration at the inserted ash deposition tube was approximately 1.5% as a dry basis. We inserted the deposition probe, made of stainless steel (SUS304), 1.9 m from the burner tile where the
temperature was about 1573 K. The temperature around the ash deposition tube is similar to that in the slagging area near the burner in the pulverized coal combustion boiler. The length and outer diameter of the ash deposition tube were 0.2 and 0.0318 m, respectively. The deposition probe mainly consisted of a test piece for deposition and a water-cooled probe.
Figure 2. Schematic diagram of a pilot-scale pulverized coal combustion refractory furnace. Table 3. Experimental conditions, Coal sample Heating Value [kW] Combustion stoichiometric ratio [-] Oxygeon concentration at furnace outlet [%] Oxygeon concentration around ash deposition tube [%] Gas and mean particle temperature around ash deposition tube [K] Gas velocity [m/s] Surface temperature of ash deposition tube [K] Ash deposition tube diameter [mm] Ash deposition tube length [mm] Exposure time for ash deposition [min]
UBC 149 1.05 1.0 1.5 1573 ∼ 2.6 773 31.8 200 100
Two thermocouples were installed between the test piece and the water-cooled probe. We assumed the temperatures measured by those thermocouples to be the surface temperature of the test piece of deposition. The surface temperature was controlled at 773 K by adjusting the flow rate of cooling water inside the tube. We observed the cross-sectional structures and compositions of ash particles before adhering to the tube and the deposition layer on the tube using scanning electron microscopy (SEM) and electron probe micro analysis (EPMA). The ash particles just before adhering to the tube were also collected in the filter, using a water cooled sampling probe and a suction pump.
5. Results and discussion 5.1. Fraction of the molten slag with and without MgO addition obtained by chemical equilibrium calculations
Molten Slag Fraction in ash [wt%]
Figure 3 shows the calculated results of the molten slag fraction in ash for four types of coal ash. The molten slag fraction is defined as the content of molten slag in the total ash. For Coal B ash, Coal C ash, and UBC ash, the molten slag formed at relatively low temperatures. For Coal A ash, however, the molten slag did not form at temperature less than 1273 K. 100 Coal C_ash
80 60 Coal B_ash
UBC_ash Coal A_ash
40 20 0 1073
1273 1473 1673 Temperature [K]
1873
Figure 3. Calculated results of the molten slag fraction in ash for four types of coal ash.
Molten slag fraction in ash with MgO [wt%]
Figure 4 shows the relationship between the additive fraction of MgO and the molten slag fraction in the ash with MgO at 1573 K. In the calculations, we simply added the MgO to the each coal ash sample. From the figure, the molten slag fraction in Coal B ash, Coal C ash, and UBC ash indicated quite large values over 90%. The molten slag fraction in Coal A ash indicated a relatively low value around 40%. However, the molten slag fraction in ash with MgO decreased as the additive fraction of MgO in three types of the coal ash increased. As for UBC ash, the molten slag fraction in ash with MgO became almost the same as that of Coal A ash when the additive fraction of MgO was 25%. Therefore, the MgO addition to coal ash played a role in the decrease of the molten slag fraction in the ash. T = 1573 K
100 80
Coal C_ash
60 UBC_ash 40 20
Coal A_ash without MgO Additive
Coal B_ash 0
0 10 20 30 40 50 Additive fraction of MgO in coal ash [wt%]
Figure 4. Relationship between additive fraction of MgO in coal ash and the molten slag fraction in the ash with MgO. Figure 5 shows the relationship between the additive fraction of MgO in coal ash and the increase-decrease mass ratio of the molten slag. The increase-decrease mass ratio of molten
Increase-decrease mass ratio of molten slag [-]
slag is defined as the mass ratio of the molten slag with MgO additive to the molten slag without MgO additive. Therefore, a ratio of less than 1 means that production of the molten slag actually decreased even if MgO was added to the ash. This figure shows that the MgO addition was effective for Coal B ash and the UBC ash. Under a condition of 25 wt% addition of MgO, production of molten slag in UBC ash almost halved. T = 1573 K
1.5
1.0
Coal C_ash UBC_ash
0.5 Coal B_ash 0.0
0 10 20 30 40 50 Additive fraction of MgO in coal ash [wt%]
Figure 5. Relationship between additive fraction of MgO in coal ash and the increase– decrease mass ratio of molten slag. Figure 6 shows the calculated results of compositions in both the molten slag and solid phases for the UBC ash with and without MgO at 1573 K. The additive fraction of MgO is 25 wt% in the ash. The minor components with less than 0.1 wt% in both phases are negligible in this figure. For the UBC ash without MgO addition, SiO2 slag was significant in the slag phase formed. When MgO was added, the fraction of the molten slag phase deceased, and some alumino-silicates compounds were produced. As a result, a large amount of molten slag, including SiO2 and FeO, decreased by adding MgO to the UBC ash. UBC_ash with MgO UBC_ash
MnO MnO MgO MgO FeO FeO Fe2O3 Fe2O3 CaO CaO
Al2O3 Al2O3 SiO22 SiO
0
5
10
15
20
25
30
35
Mass Persentage [%]
(a) Compositions in the molten slag phase.
40
UBC_ash with MgO UBC_ash
MnMgSiO MnMgSiO44 AlMg2O4 AlMg2O4 CaFeSiO CaFeSiO44 FeAl2O4 FeAl2O4 Fe2SiO4 Fe2SiO4 Al3O4 Al3O4 CaMgSiO CaMgSiO44 MgAl2O4 MgAl2O4 FeMgSiO FeMgSiO44 Mg2SiO4 Mg2SiO4 KAlSi2O6 KAlSi O 2
6
0
5
10
15
20
25
30
35
40
Mass Persentage [%]
(b) Compositions in the solid phase. Figure 6. Calculated results of the compositions in the molten slag and solid phases for UBC ash with and without MgO at 1573 K.
5.2. Ash deposition behavior during coal combustion Figures 7 and 8 show profiles of the temperatures and both gaseous concentrations and unburnt fraction Uc* along the central axis of the furnace during UBC combustion, respectively. The temperature at the location of the ash deposition tube was approximately 1573 K. The unburnt fraction is defined as: U c = 100 − η = *
Uc C Ash × × 100 1 − U c 100 − C Ash
(1)
where Uc* [%] is the unburnt fraction, η [%] is the combustion efficiency, Uc [%] is the unburnt carbon concentration in the fly ash, and CAsh [%] is the ash content in coal. At the probe insertion position, the unburnt carbon in the reacting particles was almost zero, and CO remained only a little. 1773
Temperature [K]
1723 1673 1623 1573 1523 1473
Position of tube inserted
1423 1373 0.0
0.5 1.0 1.5 2.0 2.5 Distance from burner tile [m]
3.0
Figure 7. Temperature profiles along the central axis of the furnace.
O2, CO [Vol.%]
2.5 2.0
O2
Position of tube inserted
0.8 0.6
1.5 0.4
1.0 0.5 0.0 0.0
Unburnt CO 0.5 1.0 1.5 2.0 2.5 Distance from burner tile [m]
0.2 0.0 3.0
Unburnt fraction Uc* [%]
1.0
3.0
Figure 8. Profiles of gaseous concentrations and the unburnt fraction along the central axis of the furnace. Figure 9 shows photos of ash deposition on the tube for the UBC with and without MgO after 100 min of exposure time. For the UBC without MgO, most of the deposited ash was in a molten state. The deposition layer on the tube was very thin. For the UBC with MgO, on the other hand, the ash deposition layer was mainly composed of fine brown particles.
Figure 9. Photos of ash deposition on the tube after 100 min. Figure 10 shows the relationship between the calculated molten slag fraction in the ash and the deposition fraction of ash (φash) obtained experimentally. Here, φash is defined as: φ ash =
Dash Fash − probe ⋅ t
Fash − probe =
Ap Af
Fash − furnace
(2)
(3)
In Eqs. (2) and (3), Dash denotes the deposition mass of ash for a certain exposure time (t). Fash-probe and Fash-furnace are the mass flow rates of ash in the cross-sectional areas of the probe (Ap) and the furnace (Af), respectively. The term φash_t=100 is the experimental result after 100 min. As seen in the figure, the deposition fraction of ash also decreased with a decrease of the calculated molten slag fraction. This result suggested that MgO addition to the UBC could contribute to reducing the ash deposition to the same level as that of Coal A.
Deposition fraction of ash (φash_t=100) [-]
T = 1573 K
0.15
UBC_ash 0.10
Coal A_ash 0.05
0.00
UBC_ash+MgO(25%) UBC_ash+MgO(50%)
0
20 40 60 80 100 Molten slag fraction in ash [wt%]
Figure 10. Relationship between molten slag fraction in the ash and the deposition fraction of the ash.
5.3. Mechanisms for the reduction of ash deposition Figure 11 shows the cross-sectional structures and compositions of the ash particles just before adhering to the ash deposition tube. The additive fraction of MgO in the UBC ash was 25 wt%, and the mean particle diameter of added MgO was 0.2 μm. Comparing the UBC ash with and without MgO, the fine particles with rich Mg existed uniformly in surroundings of the ash particles for the UBC ash with MgO. This observation indicated that the MgO particles did not always react with the UBC ash particles directly. UBC
10μm
Fe
Mg
Si
Al
UBC with MgO
Figure 11. Cross-sectional structures and compositions of ash particles just before adhering to the ash deposition tube. Figure 12 shows the cross-sectional structures and compositions of the deposit on the tube. We collected the samples of ash deposits on the tube after 10, 30, and 100 min of the exposure time. Comparing the results after 100 min for the UBC ash with and without MgO, a thick layer of Fe was produced for the UBC ash without MgO. This thick layer was mainly composed of Fe, Si, and Al. For the UBC ash with MgO, on the other hand, the layer was thinner, and Si and Al did not enrich the layer. As MgO was added to the UBC, the MgO concentration in the deposit increased. Based on these observations, Figure 13 summarizes the mechanisms for the reduction of the ash deposition for the UBC ash with MgO. For the UBC ash without MgO, the UBC ash directly reacted with the tube to produce an oxidized film that contained Fe, Si, and Al. For the UBC ash with MgO, on the contrary, the existence of MgO inhibited direct reactions between the ash particles and the tube surface. According to the results of the chemical equilibrium calculations mentioned above, MgO addition to the UBC could reduce the production of molten slag. In other words, the surface state of the deposit will always be dry so that the UBC ash with MgO did not easily adhere to the tube.
(a) UBC ash (b) UBC ash with MgO additive Figure 12. Cross-sectional structures and compositions of the deposit on the tube.
MgO Particle Coal Ash Particle
Rebound Stick
Stick
Tube
Tube
(a) UBC
(b) UBC with MgO
Oxidized Iron
Figure 13. Mechanisms for the reduction of ash deposition on the tube by MgO addition to the UBC. 6. Conclusions We conducted chemical equilibrium calculations and ash deposition tests in order to evaluate the effects of MgO addition to the coal on the reduction of ash deposition for UBC and understand the reduction mechanisms of the ash deposition for the MgO addition. The MgO addition played a role in the decrease of the molten slag fraction. The amount of ash deposition for the UBC with MgO was relatively depressed. The MgO inhibited production of the molten slag on the tube. One of the reduction mechanisms was due to the production of solid phase compositions of alumino-silicates due to MgO addition.
Acknowledgments This study was partly supported by JCOAL, which provided our UBC sample.
References [1] Sugita S, Deguchi T, Shigehisa T, Katsushima S. Demonstration of UBC Process in Indonesia, ICCS&T, Okinawa, Japan, Oct. 2005. [2] Sugita S, Deguchi T, Shigehisa, T. Demonstration of a UBC (Upgrading of Brown Coal) Process in Indonesia. Kobe Steel Eng. Rep. 2006;56(2):23–26. [3] Yamamoto S, Sugita S, Mito Y, Deguchi T, Shigehisa T, Otaka Y. Development of Upgraded Brown Coal Process, The Institute for Briquetting and Agglomeration (IBA), Savannah, GA, Oct. 2007. [4] Akiyama, K, Tada, T. Combustion characteristics of upgraded brown coal (UBC), Proc. 6th Int. Symp. Coal Combust., China, 2007, 527–532. [5] Frandsen, FJ. Ash Research from Palm Coast, Florida to Banff, Canada: Entry of Biomass in Modern Power Boilers. Energy Fuels 2009;23(7):3347–3378. [6] Raask E. Mineral impurities in coal combustion, Hemisphere Publishing Corporation, Washington, 1985, 169–189. [7] Walsh PM, Sayre AN, Loehden DO, Monroe LS, Beér JM, Sarofim AF. Deposition of bituminous coal ash on an isolated heat exchanger tube: Effects of coal properties on deposit growth. Prog. Energy Combust. Sci. 1990;16(4):327–345. [8] Baxter LL. Influence of ash deposit chemistry and structure on physical and transport properties. Fuel Process. Technol. 1998;56 (1-2):81–88. [9] Beer JM, Sarofim AF, Barta LE. Inorganic transformations and ash deposition during combustion. In Engineering Foundation; Benson, S. A., Ed, ASME: New York; 1992.
[10] Benson SA, Jones ML, Harb JN. Ash formation and deposition Chap.4. In Fundamentals of coal combustion for Clean and Efficient Use, Smoot, L. D., Ed, Elsevier Science, New York, 1993, p. 299. [11] Naruse I, Kamihashira D, Khairil; Miyauchi Y, Kato Y, Yamashita T, Tominaga H. Fundamental ash deposition characteristics in pulverized coal reaction under high temperature conditions. Fuel 2005;84(4):405–410. [12] Vuthaluru HB, Wall TF. Ash formation and deposition from a Victorian brown coal modelling and prevention. Fuel Process. Technol. 1998;53(3):215-233. [13] Li S, Whitty K. J. Effect of Temperature and Residual Carbon. Energy Fuels 2009;23(4):1998–2005. [14] Bai J, Li W, Li B. Characterization of low-temperature coal ash behaviors at high temperatures under reducing atmosphere. Fuel 2008;87(4-5):583–591. [15] Abbott MF, Conn RE, Austin LG. Studies on slag deposit formation in pulverized-coal combustors: 5. Effect of flame temperature, thermal cycling of the steel substrate and time on the adhesion of slag drops to oxidized boiler steels. Fuel 1985;64(6):827–831. [16] Harb JN, Munson CL, Richards,GH. Use of equilibrium calculations to predict the behavior of coal ash in combustion systems. Energy Fuels 1993;7(2):208–214. [17] Hansen LA, Frandsen, FJ, Dam-Johansen K. Ash fusion quantification by means of thermal analysis, An Engineering Foundation Conference on Impact of Mineral Impurities in Solid Fuel Combustion, Hawaii, Nov., 1997. [18] Ichikawa K, Oki Y, Inumaru J, Ashizawa M. Study on the Predicting Technique of Ash Deposition Tendency in the Entrained-Flow Coal Gasifier. Trans. Japan Soc. Mech. Eng., Ser. B 2001;67:144–150. [19] Song, W. J, Tang, L. H, Zhu, X. D, Wu, Y. Q, Zhu, Z. B, Koyama, S. Effect of Coal Ash Composition on Ash Fusion Temperatures. Energy Fuels 2009;24(1):182–189. [20] Akiyama K, Pak H, Tada T, Yoshiie R, Naruse I. Characteristics of Ash Deposition Behavior of Upgraded Brown Coal (UBC) and Bituminous Coal. Proc. of the 6th Mediterranean Combust. Symp., France, China; June, 2009.
MODELING AND OPTIMIZATION OF NOX EMISSION AND PULVERIZED COAL FLAME IN UTILITY SCALE FURNACES Srdjan Belosevic, Miroslav Sijercic, Branislav Stankovic, Nenad Crnomarkovic, Slobodan Djekic* Institute of Nuclear Sciences Vinca, Laboratory for Thermal Engineering and Energy, University of Belgrade, P.O. Box 522, 11001, Belgrade, Serbia, tel.: +381-11 3408-764, e-mail: [email protected] * Electric Power Industry of Serbia, Vojvode Stepe 412, 11000 Belgrade, Serbia, +381-11 3972-997, e-mail: [email protected] Abstract: The emission of NOx is of great concern to designers and operators of most industrial furnaces and boilers. The pulverized coal flame in utility scale boilers is also of great importance, affecting the levels and distribution of temperature and heat flux. Numerical studies of combustion and heat transfer processes in energy conversion systems can describe how the fuel chemical energy is converted into thermal energy with high efficiency and acceptable emission. Although there is much technology now available to compute complex flows in energy systems, development of submodels describing individual processes, as well as comprehensive CFD codes are increasing worldwide. A comprehensive 3D differential mathematical model and software were previously developed in-house and validated against experimental data. A practical motivation was to solve operation problems in tangentially-fired furnaces of the power plant Kostolac-B 350 MWe boiler units. The software is aimed for prediction of processes and operation situations in utility boiler pulverized coal-fired furnaces and it is adapted to be used by engineering staff dealing with the process analysis in boiler units. Characteristics of the model are Eulerian-Lagrangian approach to multiphase flow, k-ε turbulence model, particles-to-turbulence interactions modeled by PSI Cell method, diffusion model of particle dispersion, six-flux method for radiation modeling, heterogeneous reactions in kinetic-diffusion regime on the basis of experimentally obtained case-study coal kinetic parameters, within a “shrinking core” concept and with respect to the model of char oxidation, as well as homogeneous reactions controlled by chemical kinetics or turbulent mixing. In addition, submodel describing formation and destruction of thermal and fuel NOx has been developed and validated against available data obtained by monitoring of NOx emission from boiler units. The main motivation for this study was to achieve optimal position of flame with acceptable levels of NOx emission. The flame position depends on many influencing parameters. Selected predictions of pulverized coal flame geometry and position are given in the case-study furnace under different operating conditions, like fuel and air distribution. Even when both the fuel nitrogen content and the combustion temperature are not very high, the emission of NOx may still surpass environmental limits if the combustion process is not managed correctly. It is therefore essential to understand the NOx formation process so that the NOx emission can be controlled. Although post-combustion clean-up is viable, modifying combustion process often controls NOx most economically. In air staging method, e.g., the portion of combustion air is introduced downstream, through special, over-fire-air ports. In this work, the numerical study has been performed to achieve both NOx emission reduction and favorable position of flame in the case-study furnace, by investigating the impact of pulverized coal distribution over the burner tiers, without need for construction changes. The contributions of fuel and thermal NOx are reported as well. The results of the model can help in increasing combustion efficiency, lowering emission of pollution, fuel savings and corresponding economy and enviromental benefits during the facility exploitation.
Keywords: modeling, optimization, NOx emission, flame, pulverized coal, utility boiler
1. INTRODUCTION There is an ever increasing necessity for efficient and low-emission utilization of lower and variable quality fuels in a wide range of boiler load changes during the exploitation of pulverized coal-fired power plants. Moreover, there is a need to improve the performance of existing energy conversion systems, much of which has many remaining years of useful life. A priority tasks are both the increase of combustion and heat transfer efficiency and the emission reduction of nitrogen oxides and other pollutants. In attaining these goals, numerical methods can help a great deal, complementing the experimental investigations and providing detailed and complete information on processes in the furnaces under different operating conditions [1-20]. Such numerical studies encompass a great variety of problems in combustion, as well as exploitation and control of pulverized coal-fired power plants. A number of the studies are focused on modeling NOx formation/destruction and predicting NOx emission [1, 3, 5, 6, 14, 19, 20]. The commercial CFD codes for flow, combustion and heat transfer simulations, like [7-9, 12, 14, 20], the models developed in-house, e.g. [1-6, 10, 11, 15-19] and the combination of the former and the (sub)models incorporated as a user subroutines, like [13], are all applied for the purposes.
The optimization of the pulverized coal turbulent diffusion flame geometry and position is of special practical interest, because of the significant impact on heat transfer within the furnace and the thermal load of the water walls. The flame position, that depends on different parameters, strongly affects the furnace exit gas temperature (FEGT), thus influencing the boiler unit efficiency. Higher efficiencies of combustion and heat transfer lead to lower consumption of fuel for required power output, which implies emission reduction of pollutants, like greenhouse gases. A too high flame position may cause the thermal overload of the super heaters in convective pass, especially in tower-type boilers. Due to the need for intensive heat transfer during operation, care is also taken to ensure that the flame does not descend too low into the hopper, or approach too close to the furnace walls, to prevent the heating surfaces from an excessive fouling and slag deposition. Velocity, temperature and concentration fields and heat fluxes, determined by numerical predictions, offer the information important for the flame optimization in coal-fired power plants. An overview of the comprehensive combustion models, applied for the pulverized coal flame modeling and simulation of processes in large-scale furnaces is given in [10, 16, 18]. Especially useful for the purpose may be to combine the numerical experiments with the up-to-date experimental techniques, like visualization of 3D temperature distributions in utility boilers [21]. The model can suggest solutions for the problems diagnosed by the flame visualization and the solutions then can be assessed by means of this experimental method. To predict the operation situations in pulverized coal-fired boiler furnaces, a comprehensive differential 3D mathematical model of two-phase turbulent reactive flow with heat and mass transfer and a finite volume code were developed and validated against experimental data [10, 16-18]. The comprehensive model offers such a composition of modeling approaches so as to balance sophistication with computational practicality. The software helps in optimization of the combustion, heat transfer and flame and supports efficient and environmental friendly operation of the facility. It is to be used primarily by engineering staff dealing with the process analysis in boiler units and can be coupled with another in-house developed software, aimed for calculation of air-coal dust mixture preparation [22]. One of the most important tasks in thermal plant engineering is flue-gas emission reduction. Nitrogen oxides consist of nitric oxide (NO), nitrogen dioxide (NO2) and nitrous oxide (N2O), as well as some other oxides, considerably less influential. Although N2O contributes to the greenhouse effect in the troposphere and participates in ozone depletion in the stratosphere, N2O emissions are not typically significant within pulverized coal combustion systems. NO and NO2 are collectively referred to as NOx. Oxides of nitrogen have been identified as direct precursors of photochemical smog and they contribute to acid rain. Total amount of nitrogen oxides, emitted from all thermal power plants of the Electric Power Industry of Serbia, was 58030 tons per year (data valid for 2008 [23]). European Union legislation regarding the maximum levels of NO emissions in coal-fired power plants requires maximum limit of 500 mg/Nm3 of NO at 6% O2 for the plants with power output of 500 MWth or more. Emission of NOx in domestic thermal power plants highly exceeds new European emission limits of 200 mg/Nm3 (valid from 2016 on) [23]. The application of primary measures for NOx control (combustion modifications within the furnace, with optimization of process parameters) is a simple and cost effective means of NOx emission reduction (up to 60%), whereas secondary measures, based on the flue gas post-combustion clean-up, are much more expensive. High costs have forced as to look into less expensive alternatives, such as better tuning of the combustion/aerodynamics parameters. Even when the fuel nitrogen content and the combustion temperature are not very high, the emission of NOx may still surpass environmental limits if the combustion process is not managed correctly. It is therefore essential to understand the NOx formation process so that the NOx emission in boilers can be controlled. As a result, there is considerable demand for predictive procedures for pollutant emissions. Even a relatively moderate reduction of polution can provide extremely important economy and enviromental benefits during the facility exploitation. During the past several decades, investigations of NOx formation and dectruction proceses comprised measurements-experimental examinations of reaction kinetics parameters [24-35] and modeling of complex processes of different fuels combustion [36-42], as well as the demonstration of applicability of alternative technologies for NOx control [43-46], such as low-NOx burners. A great deal of these activities consists of numerical simulations of chemical reactions and the processes related to NOx emission from the furnaces [1, 3, 5, 6, 14, 19, 20, 40]. Computer modeling of NOx formation and dectruction processes in combustion systems offers a tool that can be used for investigation and better understanding of these complex systems. A comprehensive overview of modeling approaches 2
for prediction of nitrogen oxides formation and destruction in combustion systems and involving chemical reactions are given in [47, 48]. Although one day it will no doubt be possible to employ the full, or nearly full, NO chemistry in a reliable form for engineering calculation, present-day constraints of practicality dictate the use of simplified chemical models, but in conjunction with the detailed CFD calculations. This approach is referred to as comprehensive modeling [48]. In addition, the predictions are usually the outcome of a mathematical model that comprises a central code for the aerodynamics, heat-transfer and two-phase combustion, coupled to an NO post-processor. The former has undergone continuous development within this research group over many years and it has been the subject of previous publications, like [10, 16-18]. The post-processor (in-house developed model of NOx formation and destruction) is described in brief in the present paper, along with an assessment of its performance. NO from combustion systems results from three main processes: thermal NO, fuel NO and prompt NO. The nitrogen present in fossil fuels such as coal and fuel oil is typically the most significant source of NO formed during combustion. Referred to as fuel NO, it typically accounts for 75–95% of the total NO in coal combustors. Contribution of thermal NO (formed from oxidation of atmospheric nitrogen) does not become significant until temperatures in coal flames are greater than 1650 K [47]. At the temperatures below 1600 K–1800 K, thermal NO formation is significantly reduced [15, 19, 47, 48]. Though, for the reasons of the model completeness, we have chosen not to ignore the existence of thermal NO in the present model. Prompt NO is defined as the NO compound formed from the attack by hydrocarbon fragments on molecular nitrogen in the flame zone. It is significant only in very fuelrich flames and its contribution to the total NO formed is likely to be small in combustors operating mostly in fuel-lean or close to stoichiometric ratios; therefore it is neglected in this study. The NOx submodel, developed to describe formation and destruction of thermal and fuel NOx in the furnace, has been preliminary validated by comparison with available experimental data from measurements of NOx emission from boiler units in operation. The paper presents selected results of the thermal-flow conditions in pulverized coal tangentiallyfired dry bottom furnace of power plant Kostolac-B 350 MWe boiler unit, simulated by means of the inhouse developed software. The flame geometry and position within the pulverized coal furnace and NOx emission are investigated in dependence on the operation parameters, such as distribution of aircoal dust mixture and preheated air over individual burners and burner tiers and the cold air ingress. We have tried to achieve NOx emission reduction by investigating numerically the effect of fuel and air distribution, not requiring any construction changes in the boiler unit. Since different fuel and air distributions strongly affect the flame characteristics, a numerical study has been performed to minimize NOx content in the flue gases of the case-study furnace, by means of an optimal distribution of pulverized coal over the burner tiers, providing in the same time a favorable position of the flame within the furnace. 2. MATHEMATICAL MODEL AND NUMERICAL CODE For prediction of processes in two-phase turbulent reactive flow in large-scale pulverized coal-fired boiler furnaces at stationary conditions, a comprehensive 3D differential mathematical model and computer code were developed in-house. During the comprehensive model development, we investigated numerically and optimized a number of calculation parameters and constants in the models of individual processes. Some numerical approaches were compared with the alternative ones in order to justify the selection done. The model has been already described in details [10, 16-18], while general features and characteristic elements are presented here as well. The two-phase flow is treated by Eulerian-Lagrangian approach. Gaseous phase is described by timeaveraged Eulerian partial differential equations for mass, momentum, energy, gas mixture components concentrations, turbulence kinetic energy and its rate of dissipation. In general index-notation, for general variable Φ :
⎛ ∂ (ρU jΦ ) = ∂ ⎜⎜ ΓΦ ∂Φ ∂x j ∂x j ⎝ ∂x j
⎞ ⎟ + SΦ + S Φp ⎟ ⎠
(1)
with additional sources due to particles S Φp , while ρ, Uj, ΓΦ and SΦ denote gas-phase density (kg/m3), velocity components (m/s), transport coefficient and source term for Φ, respectively. To close the gas 3
phase conservation equations, k-ε turbulence model is used. The dispersed phase is described by differential equations of motion, energy and mass change in Lagrangian field. Particle velocity vector is a sum of convective and diffusion velocity. The convective velocity is obtained from the equation of motion, by particle tracking along the trajectories with constant particle number density. Particle dispersion by turbulence is modeled by diffusion model (introducing particle diffusion velocity, depending on particle turbulent diffusion) [10]. Influence of particles on gas phase is accounted for by PSI Cell method. Particle-to-wall collisions are considered with a given degree of elasticity. As opposed to a number of comprehensive combustion models, turbulence modulation is also accounted for. This is done in conservation equations for turbulence kinetic energy k (m2/s2) and its dissipation rate ε (m2/s3) by additional source terms: 2kS pm + S pk , 2εS pm + S pε , S pε = Cε3
ε
S pk , where k 2 kS pm and 2εS pm are source terms due to the mass exchange between phases S pm , while S pk is due to the fluctuations of drag forces, given as a function of particle response time, turbulence time scale and particle concentration [16]. Although the effect of turbulence modulation is usually neglected in the comprehensive combustion models, preliminary numerical investigations have suggested that the particles can affect the turbulence energy in the range of approximately 20-30% [16], which in turn may influence the turbulent combustion. The particle concentration cannot be directly obtained by solving the particle motion using a Lagrangian approach. However, two calculation procedures can be used for the purpose: approximation of the particle number density (particle concentration) by means of the Eulerian approach and determination of the particle concentration field on the basis of the calculation results in the Lagrangian field. In the first procedure, used in the paper, the continuity equation for the particle concentration (particle number density) Np (1/m3) is solved. The equation is obtained in the form of Eq. (1), with the zero source terms and the coefficient of particle turbulent diffusion Γp (m2/s): Γ p = ν tp / σ p ,
ν tp = ν t ⋅ (1 + τ p / τ t ) , where νt and νtp are the fluid turbulent diffusivity (m2/s) and the particles −1
turbulent diffusivity (m2/s), τp and τt are the particle response time (s) and the turbulence time scale (s), while σp=1.3-1.6 is the Prandtl-Schmidt number (-). With respect to the fact that there is a small Reynolds number of the particle-to-gas relative motion in most cases, it is acceptable to use the gas velocity at the point considered in the Eulerian field instead of the particle velocity. This approximation is used only in the equation for particle number density (for the rest of calculations the results of the integration along the particle trajectories are applied). This procedure gives reasonable results for the particle concentration when there are two, non-premixed streams, like here, in the case of the pulverized coal diffusion flame. A convection-radiation heat transfer is considered and the six-flux method used to model thermal radiation (diffusion type-equations are solved for total radiation fluxes), with isotropy or anisotropy of radiation scattering. Source terms due to radiation are added in energy equations for gas and particles. The gas is considered to exchange thermal radiation over the entire space, while the particle over its diffusive surface. Absorption and scattering properties of particles are obtained by numerical optimization. Source terms due to radiation are added in the gas phase and dispersed phase energy equations. Modeling of heterogeneous oxidation has provided sufficient accuracy in prediction of general effects of combustion in utility scale furnace, while for simulation of local domains under substoichiometic conditions heterogeneous gasification and homogeneous reactions of combustion of released gases are also essential. The homogeneous reactions are controlled by slower of two processes: chemical kinetics (given by Arrhenius expression) or turbulent mixing (modeled by the eddy-break-up model) [17]. There were necessary experimental data only for particle kinetics of Serbian lignites considering the entire coal particle, which influenced the modeling approach. Individual processes in complex combustion process are treated together on the basis of the particle kinetics and experimentally obtained coal kinetic parameters. The heterogeneous reactions are based on these experimentally obtained case-study coal kinetic parameters. The methodology is general and could be adapted without difficulties to include additional reactions. Since the char combustion is by far slower process 4
compared with the devolatilization and the volatiles combustion, it controls the overall combustion process. So, the coal particle combustion is modeled with respect to the char combustion model. The “shrinking core” concept is used. The chemical kinetics is known to be more important in the domain of lower temperatures; in general, kinetics controls the coal combustion for temperatures up to 1500 K. Relative importance of chemical kinetics and the particle surface diffusion was numerically investigated as well. For temperatures lower than 1200 K, kr, Arrhenius reaction rate in kinetic regime (m/s) was found to be lower than kd, a diffusion parameter of mass transfer (m/s), for two orders of magnitude, while in the range of 1200 K-1500 K, kr was lower for one order of magnite, which was the most often case during the examinations. For temperatures higher than 1500 K the two quantities approached to each other, so that around 1800 K (approximately maximal local temperatures of gas) kr becomed higher to some extent. These findings held for both examined particle diameters: 150 μm and 50 μm. Although the examinations suggested that the kr was mostly the slower, i.e. controling process, the overall numerical results still did not give enough arguments to justify neglecting the diffusion. So, the optimal choice was combined kinetic-diffusion regime, like it was done in the combustion model. Change of the coal particle mass mp (kg), equal to the reaction rate ℜ p (m/s), is given in combined kinetic-diffusion regime: −
dm p dτ
=ℜp =
ox E − Ap M p χ mol , k r = Ae RT , k d = S ⋅ D h 1 1 + kr kd
d
p
, D = 9.8 ⋅10 −10 ⋅ T 1.75
(2)
Molecular diffusivity D (m2/s) is given by empirical expression for high-temperature combustion products [49]. In addition, τ, Ap, Mp, χoxmol, A, E, R, Sh, dp, T, are: time (s), particle cross section area (m2), molar mass (kg/mol), oxidant molar concentrations (kmol/m3), pre-exponential factor in Arrhenius expression (m/s), activation energy of the coal in Arrhenius expression (J/kmol), universal gas constant [=8.314 kJ/(kmol⋅K)], Sherwood number (-), particle diameter (m) and gas temperature (K), respectively. Total particle mass change due to reactions is the sum of changes due to the individual processes. The reactions of complete oxidation of carbon and hydrogen are considered by corresponding reaction rates, while sulfur is taken into account through equivalent carbon content. Mass and heat addition due to combustion is considered by additional sources due to particles in the conservation equations, Eq. 1. Prandtl-Schmidt numbers have been optimized to 0.7 for energy and 0.86 for concentrations equations. Determination of kinetic parameters for six Serbian lignites was done on the basis of the coal combustion experiments in vertical cylindrical 15 kW laboratory furnace (with combustion chamber internal diameter 71 mm) [50]. All tests were carried out under steady operating conditions. The pulverized coal moisture content was 8 mass % and percentage residue on sieve R90=50%, the fuel flow rate was up to 4 kg/h, air-coal dust mixture inlet temperature was 438 K and air excess was 1.25. Variations of percentage residue on sieve R90 were only ±1%. Reproducibility of the experiments was satisfactory. The coal flow rate differences (occurring only occasionally) were never in excess of ±10%. Differences for primary and secondary air flow rates were about ±3%. Ash content was determined in solid samples taken by isokinetic suction probe from several inspection ports along the furnace height. The gas temperature change along the furnace height was obtained by means of thermocouples. The concentration of oxygen in combustion products was also measured. It was assumed that the reaction rate constant k (m/s), determined on the basis of the measurements and −
E
corresponding expressions [50], followed a well known Arrhenius expression: k = A ⋅ e R ⋅T (m/s) , ln k = ln A - ( E RT ) where A is pre-exponential factor (m/s), E is the coal activation energy (kJ/kmol), R is universal gas constant [=8.314 kJ/(kmol⋅K)] and T is gas temperature (K). Values of the reaction rate constant calculated for several vertical levels of the combustion chamber were presented in semilogarithm diagram, with lnK at ordinate plotted against 1/T at abscise. Intersection point between the interpolating curve (in this case straight line) and ordinate gave value of pre-exponential factor A, while the coal activation energy E was obtained from the interpolating line inclination angle. The kinetic parameters obtained for the case-study coal (lignite Kostolac-Drmno) and used in the mathematical model were: A=5.5⋅103 m/s, E=9.95⋅104 kJ/kmol [16, 50]. Experimental error, related to
5
a dispersion of the coal reaction rate values around the interpolating line, could be estimated to be less than 5%. Initial and boundary conditions usual for elliptical partial differential equations are applied. Conditions near the walls are described by the “wall functions”. For energy equation, it is possible to set either the wall temperature or the wall heat flux. Discretization of partial differential equations is performed by the control volume method and hybrid-differencing scheme. To ensure the right choice of the discretization scheme, the numerical results obtained by the hybrid scheme were compared to the predictions determined by means of the power-law scheme [16], which is to some extent less efficient, but provides an extremely good representation of exponential behavior, considered to be the closest to the exact solutions. Discretized equations are solved by the SIPSOL method (a derivative of the well known SIP algorithm, here extended for 3D flows). Coupling of the continuity equation and the momentum equations is done by using SIMPLE algorithm, while the stabilization of iteration procedure is provided by under-relaxation method. In order to provide accurate results and reliability when used in different operating conditions, the code was at first carefully verified by the grid-independence study and assessment of numerical error, while the model was validated against experimental data and applied for parametric calculations [10, 16]. A 3D staggered, structured numerical grids were used. To evaluate the influence of numerical grid and particle tracking on accuracy, three meshes were employed (with 200000, 549250 and 731250 grid nodes) in conjunction with three different numbers of particle trajectories (50, 200 and 800 trajectories per each vertical tier of each burner). The numerical results suggested the mesh with 549250 nodes and 800 trajectories per burner (6400 trajectories in total-all the burners in operation) as a proper choice, providing computational efficiency, converged solutions and accuracy in the same time. The diffusion model of particle dispersion was applied, providing statistically independent results in conjunction with less particle trajectories, when compared to the stochastic models. The model was validated against the available results of large-scale measurements in the case-study boiler furnace for three operation regimes and the accuracy of the computations was assessed as satisfactory [16]. Local gas temperatures were measured by water-cooled suction pyrometer with thermocouple protected by a shield from influence of radiation, while the flame temperatures were measured by optical pyrometer. Already validated model was used for the computations in the same case-study furnace at different operating conditions. Submodel of NOx formation and destruction processes has been developed within the comprehensive modeling approach, requiring a joint solution of detailed CFD equations for turbulent flow, combined with reduced chemical reaction mechanisms. The submodel is executed after the flame structure has been predicted, in a common “post-processor” way. The basis for this approach is that the formation of pollutant species does not affect considerably the flame structure, which is governed by fast fueloxidizer reactions. Important advantage of the approach is computational efficiency: pollutant submodels typically converge in a fraction (10%) of the time required to converge the combustion case. Thus, NOx submodel parameters and pollutant formation mechanisms can be more easily investigated by solving the NOx submodel using restart files for a pre-calculated flame structure. The homogeneous reactions of NO formation and destruction processes are described by the submodel in dependence on the kinetics parameters. The kinetics parameters have been numerically examined and optimized. In general, the submodel allows considering also the heterogeneous reactions of NO formation and reduction with char, the latter being more significant. For prediction of thermal NO, simplified Zeldovich expression is used, obtained by assuming initial concentrations of NO and OH are low so that only the forward rates of the Zeldovich mechanism are significant [48]. Thus, overall rate of thermal NO formation and destruction reactions can be expressed as: d [ NO] dt
= 2k1 ⋅ [O] ⋅ ⎡⎣ N 2 ⎤⎦ ,
⎡⎣mol ⋅ m -3 ⋅ s -1 ⎤⎦ ,
8
k1 = 1.8 ⋅ 10 ⋅ e
−
38 370 T ,
3
-1 -1
[m ·mol ·s ]
(3)
Concentrations of [O], [N2] and [NO] are given in (mol/m3], while k1 is reaction rate constant, taken from relevant references. Rate equation (3) is coupled to the fuel oxidation process through competition for the oxygen atom, whose local concentration [O] must be estimated when a comprehensive kinetic scheme is not used to compute the fuel oxidation chemistry. Accurate estimation of the nonequilibrium oxygen concentration is critical for the correct prediction of thermal NO reactions. In fuel6
lean combustion zones, oxygen atoms have been assumed to be in equilibrium with O2. The concentration of the oxygen atom can be estimated from the partial equilibrium of oxygen dissociation: 0.5 ⋅ O 2 ⇔ O , [6], as done here. The NOx submodel comprises also the reactions of fuel NO formation and depletion, in the first place through hydrogen cyanide (HCN), as an intermediate compound from volatilization, taking into account significant impact of local oxygen content. The experimental evidence suggests that the majority of the fuel NO emissions from coal combustion results from gas-phase reactions involving devolatilized material. The primary nitrogen compound, which here is the atomic nitrogen content of the volatiles, evolves as an intermediate nitrogenous compound, which can be hydrogen cyanide, ammonia or any other complex nitrogenous compound. Solomon & Colket [32], suggested that almost all coal nitrogen is contained in the aromatic rings; hence HCN is favored as a main intermediate compound. The intermediate compound is then either oxidized to form NO or reduced with a further nitrogen-containing species to yield molecular nitrogen. De Soete favours a yet simpler proposal where the HCN reacts directly either with oxygen to form NO or with nitric oxide to produce N2. We have adopted De Soete’s recommendations [30] (this model may not necessarily to be appropriate for strongly fuel-rich conditions such as may be encountered in heavily staged combustion). The chosen model is in conjunction with De Soete’s global-reaction kinetics for the gas-phase NO formation [30, 34]. The fuel NO formation reaction rate is given as: dX NO dt
10
= r1 = 3.5 ⋅ 10
⋅ X HCN ⋅
α XO 2
⋅e
−
33 732.5 T ,
[1/s]
(4)
The equation is used for fuel-lean conditions, that prevail in the pulverized-coal-fired furnaces and, as proposed by Lockwood and Romo-Millares [40], the constant of Eq. (4), i.e. the pre-exponential factor, is increased by 3.5 compared with original value (De Soete [30], more proper for fuel-rich conditions). As explained in [40], because the effects of the temperature fluctuations were not considered, an adjustment of one of the model parameters (pre-exponential factor in equation for the reaction between HCN and O2) was required. A sensitivity analysis was made, to determine the best value, while the values presented in the literature (De Soete [30] and Smith, Hill & Smoot [38]) were found not to give the correct magnitude of NO concentrations. The best fit of this coefficient for the data presented in [40], based only on experimental NO emission values, was found to be 3.5 x 1010 1/s. In forthcoming years, the value has been extensively used for NOx predictions in utility scale boilers, e.g. [1, 3]. However, calculation of the reaction rate could be performed by means of Eq. (4) with different values of pre-exponential factor in dependence on the local fuel concentration (fuel-rich, or fuel-lean conditions). Coefficient „α“, used in Eq. (4), depends on the local concentration of oxygen, according to De Soete [30]: α=1 XO2 ≤ 4.1·10-3 4.1·10-3 ≤ XO2 ≤ 1.11·10-2
α=-3.95-0.9·ln XO2
1.11·10-2 ≤ XO2 ≤ 0.03
α=-0.35-0.1·lnXO2
(5)
XO2 ≥ 0.03 α=0 For the NO depletion rate the following expression has been selected according to the literature: dX NO dt
12
= r = −3.0 ⋅ 10 2
⋅ X HCN ⋅ X
NO
⋅e
−
30 208.2 T
,
[1/s]
(6)
XHCN, XO2 , XNO are corresponding “mole fractions” (mol/mol). It is reasonable to presume that the release rate of the fuel nitrogen into the gas phase is proportional to the devolatilization rate, as done in Lockwood & Romo-Millares [40]. The release of the main intermediate compound (HCN) by devolatilization of the pulverized coal particle is taken into account implicitely; it is proportional to the nitrogen content in the fuel and numerically described in accordance with the used model of the coal particle combustion, which presumes that all individual processes (including devolatilization) in complex combustion process are taking place simultaneously [10, 16-18]. Source of HCN due to the release from the pulverized coal is calculated within the subroutine for Lagrangian particle tracking, as a sum of sources from all particle trajectories passing through the considered controle volume. The model can be adapted to take into account more detailed data on the composition of volatiles and the 7
fuel devolatilization kinetics, when there are such. For the cases in this study the effect of the NO reduction reaction by the char was found to be small, like in Lockwood & Romo-Millares [40]. Transport partial diferential equations for NO and intermediate compounds from volatiles, like HCN, are solved in Eulerian field. The source of NO in corresponding transport equation is obtained in dependence on the total net formation/destruction rate of NO (i.e. both fuel and thermal NO), while the source of HCN comprises both the HCN release by devolatilization and HCN depletion in the gaseous phase. The NOx submodel has been incorporated into the finite volume numerical code. The code has been verified with the grid independence study. For numerical predictions of NOx the numerical grid with 130x65x65 grid nodes has been selected. Other calculation parameters and operating conditions are given in details within the description of corresponding numerical experiments, for individual testcases, Tables 1-2.
Figure 1. The main form for data input.
The comprehensive model provides local values and corresponding fields of velocity components, temperature, component concentrations of the gas mixture and radiation heat fluxes, as well as the particles concentration, trajectories, velocities and temperatures during combustion. By means of developed NOx post-processor, concentration fields of nitrogen compounds, like NO, HCN, etc., can be obtained as well. To improve further the software applicability, special user-oriented forms have been introduced for data input, as shown in Figure 1. The new software solutions facilitate variation of operation parameters, turning on/off individual burners, selection of the data input mode (with possibility of coupling with the module for air-coal dust mixture preparation), restart from previous iterations (if needed) and so on [18]. 3. OPERATING CONDITIONS Kostolac Power Plant steam boiler units B-1 and B-2 (nominal steam capacity 1000 t/h and power output 350 MWe at full load, each) are of tower-type with natural circulation. The units differ to each other by the preparation of pulverized coal: centrifugal separators and louver separators (characterized by simplicity of construction) on the air-coal dust mixture ducts are used in the B-1 and the B-2 unit, correspondingly. The water-wall dry-bottom furnaces (dimensions: 15.1 m x 15.1 m x 43.0 m), with after combustion device-grate, are identical in both units. The furnace, burning pulverized coal-Serbian
8
lignite Drmno, is tangentially fired by eight jet-burners (with one fan-mill per burner) having four vertical tiers, each: two lower-stage burners (often called the “main burners”, used for combustion of larger particle size classes) and two upper-stage burners (supplying smaller particle size classes). For standard operating conditions at full load, when seven burners are in operation, total coal feed rate is 424.3 t/h and total air flow rate is 1050⋅103 Nm3/h. Pulverized coal mass flow rate is distributed approximately 70% per lower-stage and 30% per upper-stage burners. Air-coal dust mixture (temperature: 200 OC, pulverized coal and transport fluid flow rate per burner: 10.384 kg/s and 43.726 kg/s). Inlet air (temperature: 288 OC; secondary air flow rate per burner: 42.2 kg/s; air flow rate through after combustion device: 15.18 kg/s). The case-study coal, Serbian lignite Kostolac-Drmno: proximate analysis (%), as received (moisture 43.93, ash 22.25, volatile 21.39, fixed carbon 12.43); lower heating value (kJ/kg), as received: 7326.9; ultimate analysis (%), as received (carbon 22.46, hydrogen 2.12, oxygen 7.7, nitrogen 0.9, sulphur-combustible 0.64); kinetic parameters (A=5.5⋅103 m/s, E=9.95⋅104 kJ/kmol). The pulverized coal moisture content: 8.83%. Coal particle density is 1300 kg/m3. Results of the coal mills investigations, as well as the thermo-technical measurements in the boilers are presented as well, for the B-1 boiler unit [51, 52] and the B-2 boiler unit (with fixed louverswithout steering [52] and with steering, but when movable elements of louvers were closed [53]). According to experimental investigations in existing operating conditions, 30-47% of pulverized coal and 51-63% of gaseous phase in air-coal dust mixture are distributed over the upper-stage burners in the B-1 boiler unit, while in the B-2 boiler unit only 16-25% of pulverized coal and 42-49% of the gaseous phase [52]. 4. RESULTS AND DISCUSSION By using the developed software, a numerical study has been performed on the effect of different operation parameters on the position and geometry of the pulverized coal flame and the NOx emission from the case-study boiler furnace. Within a number of test-cases, Tables 1, several influencing parameters are considered, such as distribution of coal and air flow rates over the burner tiers and over the seven individual burners in operation, as well as the cold air ingress.
Table 1. Distribution of the fuel and the combustion air mass flow rate over the burner tiers Testcases
*
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17* 18** 19 20 21
Fuel distribution over the burner tiers (%) Lower-stage Upper-stage burners burners Lower Upper Lower Upper 45.5 52.4 30.4 0.0 34.9 35.2 43.2 28.6 36.8 27.6 40.0 35.0 24.5 28.0 27.0 26.0 45.5 35.2 34.9 50.0 50.0
24.5 0.0 25.8 31.6 49.1 44.8 43.2 52.7 38.4 43.6 50.0 35.0 45.5 48.0 43.0 38.0 24.5 44.8 49.1 50.0 50.0
19.5 32.7 30.7 32.0 6.2 8.0 7.9 8.3 9.4 13.9 4.5 15.0 10.5 11.0 13.0 15.0 19.5 8.0 6.2 0.0 0.0
10.5 14.9 13.1 36.4 9.8 12.0 5.7 10.4 15.4 14.9 5.5 15.0 19.5 13.0 17.0 21.0 10.5 12.0 9.8 0.0 0.0
Air-coal dust mixture gas phase through the lower-stage burners (%) 57.0 37.5 48.9 24.4 58.5 60.3 57.4 56.0 51.1 53.2 64.0 57.0 57.0 55.1 51.3 47.8 57.0 60.3 58.5 100.0 100.0
Secondary FEGT NOx emission air through (oC) (mg/Nm3) the lowerstage burners (%) 67.8 1021 1037 68.6 1004 998 74.8 1096 1148 67.8 1072 74.2 1020 854 67.8 1033 841 67.8 982 828 74.2 1026 914 74.2 1044 940 74.2 1068 1118 74.2 997 841 67.8 1058 1079 67.8 1064 1154 71.4 1047 931 71.4 1075 1118 71.4 1100 1244 67.8 995 948 67.8 904 838 60.0 991 808 90.0 1027 992 70.0 988 858
Seven burners operating unevenly The share of cold air ingress in the total amount of air 30% (for the rest of the test-cases: 7.5%)
**
9
Test-cases 1-21 consider the same values of coal and air flow rates, in total and per burner (with uniform operation of seven coal mills, except in the test-case 17, in which burners operate unevenly), as for the standard operating conditions at full load. The cases differ to each other by the parameters given in Tables 1-2. Distributions of the fuel, the air-coal dust mixture gaseous phase and the secondary air (preheated air supporting combustion), over the burner tiers, shown in Table 1, are not selected at free will, because they are related to each other. The distributions in the test-cases 1 and 17 are exactly as for the standard operating conditions at full load, while for the test-cases 2-7 and 8-10 the distributions are given according to the investigations of the coal-mills [52] and [53], correspondingly, although the reactants flow rates are equal to the ones in the standard conditions. As indicated, the test-case 17 considers the uneven distribution of fuel and air over the seven burners which are turned-on, as presented in Table 2, but again for the overall conditions at full load. Additional numerical analysis has been done with respect to different distribution of the fuel and the preheated air over the burner tiers (test-cases 11-16 and 19-21), while the test-case 18 investigates the effect of the increase of the cold air ingress. For all of the test-cases, the representative mean particle diameter of mono-dispersed pulverized coal is selected, with respect to the sieve analysis (grinding fineness of R90= 55.0%, R1000= 2.0%, where R90 denotes percentage residue on sieve with grid spacing 90 μm), Rosin-Rammler-Sperling distribution of particle size classes and a set of numerical experiments. Repeated numerical experiments have suggested that initial coal particle size of dp= 150 μm provides a proper representation of the coal particles size distribution. Table 2. Different operation regimes of individual burners, at the full load of the boiler unit Burner Test-case 1: seven burners operating evenly Flow rate of the pulverized coal/ gas phase of the air-coal dust mixture (kg/s) Flow rate of the secondary air* (kg/s) Test-case 17: seven burners operating unevenly Flow rate of the pulverized coal/ gas phase of the air-coal dust mixture (kg/s) Flow rate of the secondary air* (kg/s)
1
2
3
4
5
6
7
8
10.4/ 43.7 38.2
10.4/ 43.7 38.2
10.4/ 43.7 38.2
10.4/ 43.7 38.2
10.4/ 43.7 38.2
10.4/ 43.7 38.2
10.4/ 43.7 38.2
0.0
11.4/ 47.8 41.8
10.4/ 43.7 38.2
10.4/ 43.7 38.2
8.4/ 35.5 31.0
10.4/ 43.7 38.2
10.4/ 43.7 38.2
11.4/ 47.8 41.8
0.0
9.8
9.8
*
Including the tertiary air (core air)-portion of preheated air supplied through the middle of the air-coal dust mixture ducts
4.1. Numerical predictions of the flame position and geometry The gas temperature field in characteristic sections of the furnace, shown in Figure 2, depicts the influence of the fuel and air distribution over the burner tiers on the position and geometry of the turbulent flame, while Figure 3 presents the effect of the air ingress. Both vertical and lateral positions of the flame should be noticed, as well as the occupation of the furnace by the flame. As an additional indicator of the situation in the furnace, FEGT is also presented in Table 1 (obtained by averaging the predicted gas temperatures at the furnace exit), in most cases being lower for the lower-positioned flame. In this study, the test-case 1 could be considered as the reference one, for it is related to the standard operating conditions at full load: approximately 70% of coal and secondary air are supplied to the lower-stage burners, there is a relatively uniform thermal load of majority of the furnace volume, while the flame core descends to the region of the lower-stage burners. In the test-case 4, Figure 2, due to the pulverized coal deposition within the air-coal dust mixture ducts of the lowest-stage burners in the as-found boiler condition [52], 31.6% of fuel is supplied to the higher level of the main burners, while the rest of the pulverized coal (68.4% in total) is supplied to the upper-stage burners. This distribution of fuel over the tiers of burners gives an apparently uneven thermal load of the furnace and extremely high position of the flame, identifying the operation regime as the critical one, with respect to the fact that the flame is very likely to enter the region of heat exchangers in the convective pass. A successful attempt to optimize the flame position numerically is given in the test-case 5. Compared to the previous situation, the conditions of uniform combustion seem to be attained with the flame brought down and FEGT approaching the value for the standard operating 10
Figure 2. Influence of the fuel and air distribution over the burner tiers on the flame geometry and position in the test-cases 4, 5 and 7.
conditions at full load. Here, 84% of the pulverized coal is supplied to the lower-stage burners, as given in Table 1. The test-case 7 explores the possibility to increase the share of the coal introduced through the lower-stage burner levels over 85% (86.4% exactly, distributed evenly over the two levels). The flame descends into the furnace hopper to a considerable extent. However, this does not have to be an insoluble problem, provided there are no overheating of the hopper wall-tubes at the side closer to the flame and that there is a proper air/fuel ratio for the near-complete combustion in the hopper region (due to the tertiary air intake through the after combustion device) to avoid an excessive unburned combustibles in bottom ash. In spite of the high-temperature and wide core of the flame throughout the longitudinal section, the cross-sections in Figure 2 suggest that the flame does not approach to close to the water-walls to cause the thermal overload of the pipes or the problems with slagging. In the same time, the value of FEGT seems to be lowered just as to suggest the high efficiency of the heat transfer in the furnace. In contrast to the rest of the test-cases, where the share of the cold air ingress (entering the furnace around the burners), in the whole amount of air is 7.5%, the test-case 18 explores the effect of a significant air ingress increase to 30%, with the cold air also entering around the burners. The coal and the fluid are distributed just as in the test-case 6. The increase of the cold environmental air suction into the slightly under-pressurized furnace chamber leads to an excessive decrease of FEGT (904 ºC, compared to 1033 ºC, Table 1) due to the cooling of the flue gases. It seems that the overall flame geometry, although different, is not changed essentially, Figure 3, due to the cold air ingress near the place at which the combustion air is introduced as well. In a number of the as-found boiler conditions 11
Figure 3. Influence of the cold air ingress on the flame in the test-cases 6 and 18.
the cases with even higher ingress of cold air were recognized as well, due to the bad sealing of the furnace walls [51, 52]. With respect to numerical predictions of the flow and temperature field in different operation regimes of individual burners, Table 2, for the rest of conditions as in the standard case at full load of the boiler unit, the optimization of the central vortex and the pulverized coal flame position was performed, Figure 4. The standard working regime is achieved by 7 burners (and 7 coal mills) operating evenly, by which the symmetry of the flow and temperature field is disturbed, giving displacement of the vortex and the flame to the water-wall at which the burner is turned-off. When all of the burners are operating uniformly, there is aerodynamic and temperature symmetry in the furnace. This offers the most favorable conditions, but they are never achieved in exploitation because one burner is always kept as the back-up. As known and confirmed also by simulations [10, 16], the symmetry, being favorable regarding the water-walls thermal stresses as well as diminishing the conditions for slagging, can be maintained when two burners on opposite walls are turned-off. Here, the solution was sought for the problem by experimented numerically with uneven distributions of fuel and air over the seven burners in operation (some amount of air was also supplied to the turned-off burner, number 8, to prevent it from being overheated), Table 2. In the test-case 17, by uneven operation of the mills, i.e. the burners, the vortex was properly centered. Though, the flame is still relatively close to the wall around the turned-off burner and there is also an increase of the space occupied by the flame in the cross-section through the lowest-stage burners, Figure 4. Further careful numerical testing and optimization of the flame geometry and position would be desirable, taking into account the complete flame geometry.
Figure 4. Numerical optimization of the vortex and flame position in the cross-section through the lowest-stage burners at x=12.1 m, in the test-cases 1 and 17.
12
4.2. Numerical prediction of NOx concentration field and emission
4.2.1. Validation of the calculations and the mathematical model of NOx formation and destruction In order to perform preliminary validation of the developed NOx submodel and the NOx predictions, numerical results were compared with the available experimental data obtained by periodical measuremets of NOx emission from the power plant Kostolac B-1 and B-2 utility boiler units [23]. Operating conditions for steam boiler units (in the case of nominal steam capacity and power output of 350 MWe at full load, each) were already described, section 3. The operating conditions and the characteristics of the fuel-Drmno lignite and the pulverized coal are provided also in [16, 18]. In the test-cases considered for comparisons there were the same flow rates of the air-coal dust mixture and the combustion air, both in total and per individual burner as for the reference test-case 1, while they differed to each other with respect to the distribution of air-coal dust mixture over the burner tiers, Table 1. The fuel and air distributions in the test-cases corresponded to the ones that could have been expected in the existing operating conditions at the real boiler units. Table 3 presents both the comparison between predictions and measurements and the NOx emission limits. To evaluate the predictions, it was necessary to correct-recalculate the predicted average NOx mass fractions (kg/kg) at the furnace exit to the conditions required by the corresponding standards for NOx emission limits (like the European Union Directive 2001/80/EC): p=1 atm, t=0 OC, emission limit values for NOx (measured as NO2, dry basis, corrected to 6% O2 in flue gases, for solid fuels and the boiler facility power output > 500 MWth), as well as to express the emission in corresponding unit: [mg/Nm3]. A reasonable degree of prediction ability was demonstrated. However, it should always bear in mind that it is not simple to trace the origins of eventual predictive difficulties, because the defects in the NOx post-processor cannot be separated easily from those of the parent code for the flow and combustion. Table 3. Comparison between predicted and measured NOx emission Numerical predictions of NOx emission from the utility boiler furnace [mg/Nm3] Power plant Kostolac B-1 Test-case 1 1037
Test-case 3 1148
NOx emission [mg/Nm3] Power Plant Kostolac – * Limit values periodical measurements
Power plant Kostolac B-2
B-1
B-2
Serbian regulations
European Union Directive 2001/80/EC
Test-case 14 931
1051
893
450
500 (**200)
*
Emission limit values for NOx (measured as NO2, dry basis, corrected to 6% O2 in flue gases, for solid fuels and the boiler facility power output > 500 MWth) [23] ** From 2016 on
On the other hand, it is clear that the NOx emissions detected at the boiler units are fairly above the emission limits, especialy those which will become valid in the near future (from 2016 on). Thus it is necessary to reduce the emission from the analyzed boiler unit in the forthcoming period, by means of different measures, in the first place primary ones. In doing this, numerical simulations can help a great deal in examination, evaluation and selection of alternative solutions, with respect to their applicability and efficiency. 4.2.2. Numerical results for the reference test-case Numerical predictions given in Figure 5, for the test-case 1, show the dependence of NO content in the furnace of the Kostolac B boiler units on the flue gases temperature and the concentration of reactants, in this case HCN and oxygen, that produce fuel NO by homogeneous reactions, Eq. (4). Reactions of NO reduction with HCN, Eq. (6), are also taking place. Since fuel NO is by far more important than thermal NO (in the temperature domain for the furnace considered), the dependence of fuel NO actually determines the character of the total NO concentration field. Figure 5 suggests significant influence of the concentration of HCN (as intermediate compound released by devolatilization) and thus also of the nitrogen content in the coal which is the source of HCN, on the NO content. A narow 13
zone of higher NO concentrations, placed in the region in which the major part of the fuel is introduced through the lowest-stage burners, corresponds to the maximum of the HCN content. Wider zone of high NO concentrations can be seen in downstream (upward) direction and is related to the region of intensive chemical reactions of HCN depletion and NO formation. In contrast to thermal NO, the content of fuel NO (and consequently also the total NO) is less affected by temperature. In addition to the nitrogen content in the fuel, it is strongly influenced by air to fuel ratio [46] (air excess), i.e. oxygen concentration. This can be clearly seen from the comparison between the concentrations, given in Figure 5. So, the NO concentration field does not follow only the temperature and the HCN concentration field, but even more the O2 concentration field. In spite of high temperatures, there is no high NO concentration in the central region of the furnace, because of the oxygen depletion by intensive reactions of the fuel combustion. In the model, this strong dependence of NO content on oxygen is described by the Eq. (5), that gives dependence of the coefficient α on the local oxygen concentration, where α is exponent of the mole fraction XO2 in Eq. (4) for fuel NO formation reaction rate.
Figure 5. Gas temperature field and O2 , HCN and NO concentration fields in the Kostolac B utility boiler unit furnace, test-case 1
Analysis of the results showed that the contribution of thermal NO was of the order of several percents of the total NO. The region in which thermal NO might have been imortant was restricted to the narrow zone in which it was produced, corresponding to the maximal local gas temperatures in the 14
furnace (T=1650 K-1800 K), where the presence of thermal NO could have been only expected, as already explained. Since in the cases considered, as clarified by more detailed analysis of numerical data, the local gas temperatures rarely exceeded this temperature limit, the contribution of thermal NO to the total NO was not so significant. 4.2.3. The effect of selected operating conditions on NOx emission: a numerical analysis 4.2.3.1. The effect of the air-coal dust mixture distribution over the burner tiers on NOx emission Developed mathematical model of NOx formation/destruction was applied to predict and analyze several test-cases, with respect to NOx emission from furnaces of the Kostolac B utility boiler units. In the test-cases, there were the same flow rates of the air-coal dust mixture and the combustion air, both in total and per individual burner, as for the reference test-case 1. They differed to each other with respect to the working parameters given in Tables 1-2. The main difference was the pulverized coal distribution over the burner tiers. Although both FEGT and NOx emission are given in Table 1 for a number of test-cases, only a few situations are presented in Figures 6-11, in order to provide a concise analysis. Figures 6-11 present numerical results for the field of local concentrations, i.e. mass fractions of NO in characteristic sections of the boiler furnace. In the test-cases 1 and 3 the centrifugal separators are used, while in the test-cases 11 and 14-16 the louver separators are applied. While the air-coal dust mixture distibutions for the test-cases 1 and 3 (Kostolac B-1) and 14 (Kostolac B-2) were obtained during the examinations of the coal mills in existing operating conditions, the test-case 11 was performed as an atempt to optimize this distribution numerically, with respect to the reduction of NOx content in the furnace flue gases. Figures 10-11 show numerical simulations for two cases with opened movable elements of the louvers (test-cases 15 and 16), in comparison with the situation when they are closed, test-case 14, Figure 9. As given in Table 1 and summarized in the titles of Figures 6-8 for the test-cases considered, increase of the fraction of the pulverized coal flow rate through the lower-stage burners provides a decrease of both FEGT and NOx emission. Similar can be concluded for the test-cases given in Figures 9-11. Figures 6-11 also suggest that higher FEGT and NOx emission correspond, in these cases, to the higher position of the flame (i.e. to the wider flame at the furnace exit) and vice versa. However, one should be very carefule in trying to draw some general conclusions. Working situations in furnaces are very complex and require comprehensive analysis. For example, fraction of the pulverized coal flow rate through the lower-stage burners (so called main burners) is the same for the test-cases 1 and 15, i.e. 70%, but FEGT and NOx emission are lower in the test-case 1, with considerably lower position of the flame, Figure 6, compared to Figure 10. This is probably due to the fact that in the test-case 1 considerably more fuel is supplied through the lower tier of both the main burners and the higher-stage burners, as given in Table 1. Moreover, there is lower NOx emission, slightly lower FEGT and lower position of the flame in the test-case 3 than in the test-case 16, Figures 7 and 11, although there is to some extent lower fraction of the pulverized coal flow rate through the lower-stage burners in the testcase 3. The explanation is most likely similar to the one given for the previous two cases. As demonstrated, the analysis has to account for the fuel distribution over each of the burner tiers: two lower-stage and two upper-stage burners. Anyhow, the numerical study is to be performed for each case individually, because there are a lot of mutually dependent parameters influencing the situation in the boiler furnace. This kind of comparative analysis of the gas temperature and the NOx concentration field, in conjunction with the values of FEGT and NOx emission from the furnace, offers the possibility of optimizing both the flame and the emission. Higher position of the flame causes the higher values of FEGT, which suggests also the lower efficiency coefficient of the boiler. For the test-cases presented in Figures 6-11, the lowest emission achieved by numerical optimization (in the test-case 11) was approximately 32% lower that the highest one (in the test-case 16). This is not to be neglected, especially because in this example the emission reduction is attained simply by using an appropriate combustion organization within the furnace considered, with no need for construction changes in the furnace and the boiler.
15
Figure 6. NO content in the furnace and gas temperature field for the test-case 1 (fraction of the pulverized coal flow rate through the lower-stage burners 70%, FEGT=1021 OC, NOx emission=1037 mg/Nm3)
Figure 7. NOx content in the furnace and gas temperatur field for the test-case 3 (fraction of the pulverized coal flow rate through the lower-stage burners 56.2%, FEGT=1096 OC, NOx emission= 1148 mg/Nm3)
Figure 8. NOx content in the furnace and gas temperature field for the test-case 11 (fraction of the pulverized coal flow rate through the lower-stage burners 90%, FEGT=997 OC, NOx emission=841 mg/Nm3)
16
Figure 9. NOx content in the furnace and gas temperature field for the test-case 14 (fraction of the pulverized coal flow rate through the lower-stage burners 76%, FEGT=1047 OC, NOx emission=931 mg/Nm3)
Figure 10. NOx content in the furnace and gas temperature field for the test-case 15 (fraction of the pulverized coal flow rate through the lower-stage burners 70%, FEGT=1075 OC, NOx emission= 1118 mg/Nm3)
Figure 11. NOx content in the furnace and gas temperature field for the test-case 16 (fraction of the pulverized coal flow rate through the lower-stage burners 64%, FEGT=1100 OC, NOx emission= 1244 mg/Nm3)
17
4.2.3.2. The effects of the fuel and air distribution over the burners, the cold air ingress and the combustion air distribution over the burner tiers on NOx emission The test-case 17 examines what happens with the emission when there is an uneven distribution of the fuel and the combustion air-secondary air over the individual burners, used in order to centralize the flame within the furnace cross section, Figure 4. It is important that this, aerodynamically and thermally favorable-central postition of the flame, does not cause the emission increase (but the opposite, see Table 1), in the case considered. In the test-case 18, compared with the test-case 6, with the rest of the operating conditions being the same, an excesive ingress of the cold air slightly reduces the emission (due to the additional cooling of the gases), but also causes an excesive reduction of FEGT. The other interesting situation is given in the test-case 19, which differs from the test-case 5 only by 20% decrease of the fraction of secondary air injected through the lower-stage burners, namely 60% instead of 74.2%, Table 1. Compared to the test-case 5, with already achieved favorable conditions (values of FEGT and the emission, as well as the flame positioned mainly in the central-burners region of the furnace), further decrease of both the FEGT and the emission are provided in the test-case 19. With these modifications of the combustion air distribution, one of the lowest values of the emission is achieved, Table 1, with the flame additionally descended, Figure 12.
Figure 12. NO content in the furnace and gas temperature field for the test-cases 5 and 19 (fractions of the secondary air flow rate through the lower-stage burners: 74.2% and 60%, FEGT=1020 OC and 991 OC, NOx emission= 854 mg/Nm3 and 808 mg/Nm3)
Finally, the test-cases 20 and 21 demonstrate also the influence of different distribution of secondary air over the burner tiers, for the same distribution of coal. Here, however, all air-coal dust mixture is introduced through the lower-stage burners, while the upper-stage burners are used only for injecting the preheated air. Injection of 20% less secondary air through the lower-stage burners decreases FEGT and considerably decreases NOx emission. Although the contribution of thermal NO (strongly dependent on the gas temperature) is not so significant in the cases considered, it may still be supposed 18
that the part of the explanation for the obtained trend may be that less combustion air in the lower-stage burners region will delay, to some extent, the NO formation reactions to the upper regions of lower temperatures. Nevertheless, it seems that the emission reduction obtained by the corresponding modifications of the combustion air distribution over the burner tiers in the test-cases 19-21 could be explained mainly by a proper tuning of the local air excess in the burners region. Performed numerical analyses demonstrate the possibility of applying the developed predictive tool for NOx emission in finding the solutions and selecting between alternative combustion modifications and other primary measures that may be appled for reduction of the emission from pulverized coalfired boiler furnaces. This kind of analysis enables an efficient optimization of the operating conditions and the case-study furnace operation with respect to both the emission reduction and the possibility of keeping, or even increasing the energy efficiency of the facility. CONCLUSIONS A comprehensive numerical study was performed on the effects of different operation parameters on the pulverized coal flame position and NOx emission from tangentially-fired furnaces of 350 MWe utility boiler units. The predictions of flow, combustion and heat transfer were performed by means of previously developed CFD model and computer code. NOx emission predictions were done by recently developed NOx formation and destruction submodel and the post-processor, using simplified chemical models in conjunction with detailed CFD calculations. The coal and the combustion air distribution over the burner tiers, the cold air ingress and different operation regimes of individual burners were investigated and found to affect the flame and the emission considerably. It was possible to achieve favorable vertical position of the flame within the furnace by means of careful numerical optimization of the pulverized coal and the combustion air distribution over the burner tiers. Non-uniform distribution of coal and air over the individual burners offered an option to attain a central position of the flame within the furnace cross section, which might have affected the thermal load of the water-walls and the conditions for fouling and slag deposition. For prediction of thermal NO simplified Zeldovich expression was used, while the reactions of fuel NO formation and destruction were modeled as well. Fuel NO formation was considered through oxidation of HCN as the main intermediate compound from volatiles, taking into account strong influence of local oxygen content. For fuel NO formation reaction rate, correction of pre-exponential factor was adopted, according to relevant references, for predominantly fuel-lean conditions in pulverized coal-fired furnaces. The developed NOx post-processor was applied, along with a comprehensive model of pulverized-coal combustion, to predict NO emissions from the utility boiler furnaces. NOx submodel was preliminary validated by comparison between the predictions and the available results of measurements of the emission from power plant Kostolac B-1 and B-2 boiler units. The results suggested that the present NO model was capable of reproducing the different trends in NO formation/destruction, associated with the dissimilar operating conditions. Numerical results for the gas mixture temperature field and the O2, HCN and NO concentration fields are given as well, for the reference test-case. The numerical study was performed to achieve both NOx emission reduction and favorable position of the flame in the case-study furnace, by investigating the corresponding combustion modifications, without need for construction changes in the facility. The study showed that it was possible to achieve this task by careful optimization of the coal and the combustion air distribution over the burner tiers. In general, both FEGT and NOx emission were decreased and the flame was descended when introduced larger fractions of pulverized coal flow rate through the lower-stage burners and/or the lower stages of both the main burners and the upper-stage burners. The emission reduction could be obtained also by a proper tuning of the local air excess in the burners region. However, the complex situation in the boiler furnace requires accounting for the fuel and air distribution over each of the burners and the burner tiers and performing the numerical study for each case individually. Numerical predictions enable impact-analysis of many operation parameters, optimization of combustion, heat transfer, flame and emission, thus solving critical operation problems, like excessively high or low flame position, or high emission. This kind of numerical analysis enables optimization of the operating conditions in the boiler furnaces both with respect to the emission reduction and the energy efficiency improvements.
19
ACKNOWLEDGMENTS This work has been supported by the Republic of Serbia Ministry of Science and Technological Development (project: TR-18007) and the Electric Power Industry of Serbia. REFERENCES [1] L. M. R. Coelho, J. L. T. Azevedo, M. G. Carvalho, Application of a global NOx formation model to a pulverized coal fired boiler with gas reburning, Proceedings, Fourth International Conference on Technologies and Combustion for a Clean Environment, Lisbon, 7-10 July 1997, Paper 9.4, pp. 1/8-8/8. [2] J. Fan, P. Sun, X. Zha, K. Cen, Modeling of combustion process in 600 MW utility boiler using comprehensive models and its experimental validation, Energy & Fuels 13 (1999), pp. 1051-1057. [3] M. Xu, J. L. T. Azevedo, M. G. Carvalho, Modelling of the combustion process and NOx emission in a utility boiler, Fuel 79 (2000), pp. 1611-1619. [4] Ch. Zheng, Zh. Liu, X. Duan, J. Mi, Numerical and experimental investigation on the performance of a 300 MW pulverized coal furnace, Proceedings of the Combustion Institute 29 (2002), pp. 811-818. [5] R. He, T. Suda, M. Takafuji, T. Hirata, J. Sato, Analysis of low NO emission in high temperature air combustion for pulverized coal, Fuel 83 (2004), pp. 1133–1141. [6] K. Li, S. Thompson, J. Peng, Modelling and prediction of NOx emission in a coal-fired power generation plant, Control Engineering Practice 12 (2004), pp. 707–723. [7] J. Pallares, I. Arauzo, L. I. Diez, Numerical prediction of unburned carbon levels in large pulverized coal utility boilers, Fuel 84 (2005), pp. 2364-2371. [8] R. Vuthaluru, H. B. Vuthaluru, Modelling of a wall fired furnace for different operating conditions using FLUENT, Fuel Processing Technology 87 (2006), pp. 633-639. [9] A. Bosoaga, N. Panoiu, L. Mihaescu, L. I. Backreedy, L. Ma, M. Pourkashanian, A. Williams, The combustion of pulverized low grade lignite, Fuel 85 (2006), pp. 1591-1598. [10] S. Belosevic, M. Sijercic, D. Tucakovic, Three-dimensional modeling of utility boiler pulverized coal tangentially fired furnace, International Journal.of Heat and Mass Transfer 49 (2006), pp. 3371-3378. [11] E. I. Karpenko, V. E. Messerle, A. B. Ustimenko, Plasma-aided solid fuel combustion, Proceedings of the Combustion Institute 31 (2007), pp. 3353–3360. [12] M. Kumar, S. G. Sahu, Study on the effect of the operating condition on a pulverized coal-fired furnace using computational fluid dynamics commercial code, Energy & Fuels 21 (2007), pp. 3189-3193. [13] N. Hashimoto, R. Kurose, H. Tsuji, H. Shirai, A numerical analysis of pulverized coal combustion in a multiburner furnace, Energy & Fuels 21 (2007), pp. 1950-1958. [14] L. I. Dıez, C. Cortes, J. Pallares, Numerical investigation of NOx emissions from a tangentially-fired utility boiler under conventional and overfire air operation, Fuel 87 (2008), pp. 1259–1269. [15] J. Makovička, Mathematical model of pulverized coal combustion, Dissertation thesis, Czech Technical University in Prague, Faculty of Nuclear Sciences and Physical Engineering, Prague, Czech Republic, 2008. [16] S. Belosevic, M. Sijercic, D. Tucakovic, N. Crnomarkovic, A numerical study of a utility boiler tangentially-fired furnace under different operating conditions, Fuel 87 (2008), pp. 3331-3338. [17] S. Belosevic, M. Sijercic, P. Stefanovic, A numerical study of pulverized coal ignition by means of plasma torches in air-coal dust mixture ducts of utility boiler furnaces, International Journal.of Heat and Mass Transfer 51 (2008), pp. 1970-1978. [18] S. Belosevic, M. Sijercic, N. Crnomarkovic, B. Stankovic, D. Tucakovic, Numerical prediction of pulverized coal flame in utility boiler furnaces, Energy & Fuels 23 (2009), pp. 5401-5412. [19] R. Straka, Mathematical model of pulverized coal fired boiler, Dissertation thesis, Czech Technical University in Prague, Faculty of Nuclear Sciences and Physical Engineering, Prague, Czech Republic, 2009 (through private communication). [20] E. H. Chui, H. Gao, Estimation of NOx emissions from coal-fired utility boilers, Fuel 89 (2010), pp. 29772984. [21] H-Ch. Zhou, Ch. Lou, Q. Cheng, Zh. Jiang, J. He, B. Huang, Zh. Pei, Ch. Lu, Experimental investigations
20
on visualization of three-dimensional temperature distributions in a large-scale pulverized-coal-fired boiler furnace, Proceedings of the Combustion Institute 30 (2005), pp. 1699-1706. [22] D. Tucakovic, T. Zivanovic, V. Stevanovic, S. Belosevic, R. Galic, A computer code for the prediction of mill gases and hot air distribution between burners’ sections at the utility boiler, Applied Thermal Engineering 28 (2008), pp. 2178-2186. [23] M. Gavrić, A. Vlajčić, B. Čeperković, Green Book of the Electric Power Industry of Serbia, Belgrade, June 2009, Published by: Electric Power Industry of Serbia, Sector for Public Relations, Belgrade, http://www.eps.rs/publikacije/ZK/EPS – Zelena knjiga.pdf [Accessed 17.02.2010] (in Serbian). [24] C. P. Fenimore, Formation of nitric oxide in premixed hydrocarbon flames, Thirteenth Symposium (International) on Combustion, The Combustion Institute, 1971, pp. 373-380. [25] D. Iverach, K. S. Basden, N. Y. Kirov, Formation of nitric oxide in fuel-lean and fuel-rich flames, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 767-775. [26] V. Quan, F. E. Marble, J. R. Kliegel, Nitric oxide formation in turbulent diffusion flames, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 851-860. [27] F. V. Bracco, Nitric oxide in droplet diffusion flames, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 831-842. [28] C. T. Bowman, Kinetics of nitric oxide formation in combustion processes, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 729-738. [29] C. T. Bowman, Kinetics of pollutant formation and destruction in combustion, Progress in Energy and Combustion Science 1 (1975), pp. 33-45. [30] G. G. De Soete, Overall reaction rates of NO and N2 formation from fuel nitrogen, Fifteenth Symposium (International) on Combustion, The Combustion Institute, 1975, pp. 1093-1102. [31] J. M. Levy, L. K. Chan, A. F. Sarofim, J. M. Beer, NO/char reactions at pulverized coal flame conditions, Eighteenth Symposium (International) on Combustion, The Combustion Institute, 1981, pp. 111-120. [32] P. R. Solomon, M. B. Colket, Evolution of fuel nitrogen in coal devolatilisation, Fuel 57 (1978), pp. 749755. [33] J. A. Miller, C. T. Bowman, Mechanism and modelling of nitrogen chemistry in combustion, Progress in Energy and Combustion Science 15 (1989), pp. 287-338. [34] G. G. De Soete, Heterogeneous N2O and NO formation from bound nitrogen atoms during coal char combustion, Twenty-Third Symposium (International) on Combustion, The Combustion Institute, 1990, pp. 1257-1264. [35] I. Aarna, E. M. Suuberg, A review of the kinetics of the nitric oxide-carbon reaction, Fuel 76 (1997), pp. 475-491. [36] J. W. Mitchell, J. M. Tarbell, A kinetic model of nitric oxide formation during pulverized coal combustion, AIChE Journal 28 (1982), pp. 302-311. [37] E. E. Khalil, Modelling of Furnaces and Combustors, Abacus Press, UK, 1982. [38] P. J. Smith, S. C. Hill, L.D. Smoot, Theory for NO formation in turbulent coal flames, Nineteenth Symposium (International) on Combustion, The Combustion Institute, 1982, pp. 1263-1270. [39] R. D Boardman, L. D. Smoot, Prediction of nitric oxide in advanced combustion systems, AIChE Journal 34 (1988), pp. 1573-1576. [40] F. C. Lockwood, C. A. Romo-Millares, Mathematical modeling of fuel NO emissions from PF burners, Journal of the Institute of Energy 65 (1992), pp. 144-152. [41] J. M. Jones, P. M. Patterson, M. Pourkashanian, A. Williams, Approaches to modelling heterogeneous char NO formation/reduction during pulverised coal combustion, Carbon 37 (1999), pp. 1545-1552. [42] M. Xu, Y. Fan, J. Yuan, C. Sheng, H. Yao, A simplified fuel-NOx model based on regression analysis, International Journal of Energy Research 23 (1999), pp. 157-168. [43] L. J. Muzio, G. C. Quartucy, Implementing NO control: research to application, Progress in Energy and Combustion Science 23 (1997), pp. 233-266. [44] S. Li, T. Xu, Sh. Hui, X. Wei, NOx emission and thermal efficiency of a 300 MWe utility boiler retrofitted by air staging, Applied Energy 86 (2009), pp. 1797–1803.
21
[45] Sh. Su, J. Xiang, L. Sun, S. Hu, Zh. Zhang, J. Zhu, Application of gaseous fuel reburning for controlling nitric oxide emissions in boilers, Fuel Processing Technology 90 (2009), pp. 396–402. [46] B. G. Miller, Sh. F. Miller, J. L. Morrison, A. W. Scaroni, Cofiring coal-water slurry fuel with pulverized coal as a NOx reduction strategy, http://www.energy.psu.edu/factssheets/55_Cofiring_CWSF.pdf [Accessed: 25.02.2010]. [47] A. M. Eaton, L. D. Smoot, S. C. Hill, C. N. Eatough, Components, formulations, solutions evaluation, and application of comprehensive combustion models, Progress in Energy and Combustion Science 25 (1999), pp. 387-436. [48] S. C. Hill, L. D. Smoot, Modeling of nitrogen oxides formation and destruction in combustion systems, Progress in Energy and Combustion Science 26 (2000), pp. 417-458. [49] T. V. Vilenskij, D. M. Hemaljan, Dynamics of Pulverized Fuel Combustion, Energija, Moscow, 1978 (in Russian). [50] A. V. Saljnikov, B. S. Repic, P. T. Radulovic, L. L. Jovanovic, Combustion kinetics of coal dust, Journal of Engineering Physics and Thermophysics 68 (1995), pp. 225-229. [51] B. Perković, D. Adamović, V. Joksimović, M. Erić, Adjustment and optimization of the operation of power plant Kostolac-B boiler unit 1, after overhaul and reconstructions done in 2002/2003, Technical Report NIVITE 297, Institute Vinca and Power Plant Nikola Tesla, 2005 (in Serbian). [52] P. Radovanović, B. Perković, D. Adamović, Thermotechnical investigations and operation analysis of the boiler units B1 and B2 in power plant Kostolac before rapairs done in 2007 and 2008, Technical Report NIVLTE 373, Institute Vinca and Power Plant Nikola Tesla, 2008 (in Serbian). [53] P. Radovanović, B. Perković, D. Adamović, Thermotechnical investigations of the boiler units B1 and B2 in power plant Kostolac after rapairs done in 2008, Technical Report NIV-LTE 409, Institute Vinca and Power Plant Nikola Tesla, 2009 (in Serbian).
22
OBSERVATION OF HEAT RELEASE REGION AS FUNCTIONS OF COAL PROPERTIES IN TURBULENT JET PULVERIZED COAL FLAMES Yon-Mo Sung, Cheor-Eon Moon, Seong-Yool Ahn, Jae-Woo An School of Mechanical Engineering, Pusan National University, 30 Jangjeon-dong, Gumjeong-ku, Busan, 609-735, Republic of Korea Gyung-Min Choi*, Duck-Jool Kim Pusan Clean Center, Pusan National University, 30 Jangjeon-dong, Gumjeong-ku, Busan, 609-735, Republic of Korea *
Corresponding Author: [email protected] Telephone: +82-51-510-2476 FAX: +82-51-512-5236
Presented at The 27th Annual International Pittsburgh Coal Conference Istanbul, Turkey October 11-14, 2010 ABSTRACT One issue of interest is to develop diagnostic methods for the monitoring and control of the pulverized coal flames in power plants. The purpose of this study is to establish visualization and diagnostic methods in the pulverized coal combustion fields. An advanced instrumentation and research methodology was employed to observe the structure of pulverized coal flame in a laboratory scale burner. The effects of pulverized coal properties, volatile matter, particle size and moisture content, on the heat release region in turbulent jet pulverized coal flames were investigated experimentally. To understand the accuracy of line of sight measurement in the two-dimensional (2-D) visualization, point measurements of chemiluminescence intensity by Cassegrain optics were also conducted. The heat release region for the structure of pulverized coal flame was observed through visualization by CH* chemiluminescence image with an intensified high-speed camera, and by CH* chemiluminescence intensity for local point measurements. The streamwise length of the heat release region based on 2-D visualizations was about 11.4% longer than that of point measurements and increased proportionally to the volatile matter content. The temperature rise for 35~45 µm coal particles was faster than that for 75~90 µm particles, which resulted in a shift of reaction region toward upstream direction. The -1-
coal moisture content less than 15%, however, had little effect on the structure of the pulverized coal flame. The obtained results give us useful information for evaluating practical pulverized coal flames.
Keywords Pulverized coal combustion, Heat release region, Chemiluminescence intensity, Twodimensional visualization, Point measurement, Volatile matter, Particle size, Water contents
Introduction The combustion of pulverized coal is a complex process involving conductive, convective, and radiative heat transfer; turbulent fluid motion; coal particle devolatilization; volatile reaction; char reaction; particle dispersion; ash formation; and other processes, interactively and simultaneously [1]. Thus, the fundamental mechanism of pulverized coal combustion and the structure of pulverized coal flame have not fully understood. To obtain optimal conditions of pulverized coal combustion, the major processes such as ignition, drying, devolatilization, and volatile and char reactions of pulverized coal particles must be understood. All these processes depend on burning condition and coal properties [2]. Coal properties and combusting condition include volatile matter content, particle size, moisture content, coal rank, gas temperature, system pressure and residence time [3-9]. Many studies have been conducted to understand the detail characteristics of pulverized coal combustion [10] such as the structure of pulverized coal flame, pulverized coal particle ignition [11-13], devolatilization [14,15], and volatile and char combustion [16]. The yield of volatile is likely to be fairly high in actual pulverized coal flame, and the volatile reaction is also of great importance for the ignition of coal particles [17]. The interactions of volatile flame and char combustion of coal particles were investigated experimentally and numerically [17]. A surrounding mantle of volatile products was observed during the early stages of combustion when bituminous coal particle are more sensitive to the volatile yield than to the kinetics of devolatilization, and relatively insensitive to the kinetics of char combustion for mediumvolatile coal flame [18]. The phenomena of coal combustion can be described by considering three major regions: a particle heatup region, a devolatilization region, and a soot growth region [19]. There is an extensive review for diagnostic techniques for the monitoring and control of practical flames [20]. Burning coal may give luminous (carbon) flames, and at times the base of such flames may show C2 and CH emission [21]. The OH* radical seems more suitable for lean -2-
flames, which CH* and C2* have a more monotonic behavior and stronger dynamics for richer flames [22]. Thus, luminosity flame structure is observed through CH * chemiluminescence intensity and is considered to be a good representative of the heat release rate [23]. CH * chemiluminescence signals also provide a better indicator for the onset of pulverized coal particle ignition and the volatile reaction region than a direct photograph of the pulverized coal flame [24]. The measurement of CH* chemiluminescence from ensemble-averaging of ICCD image data can reveal the underlying trends in particle ignition and devolatilization [25]. Because the line of sight method by high speed camera might cause measurement error due to superposition, the verification of diagnostic methods is needed to obtain reliable information from the pulverized coal combustion. In this study, to estimate the availability of 2-D visualization method using an intensified high speed camera, pulverized coal jet flames were visualized and the results were compared with those of point measurements with a multi-color integrated cassegrain receiving optics (MICRO) system [26] of chemiluminescence intensity. In addition, the relationship between pulverized coal properties and the structure of pulverized coal flame was investigated. To investigate the effects of pulverized coal properties on the heat release region in turbulent jet pulverized coal flame, various pulverized coal properties including five different types of pulverized coal, two particle sizes of bituminous coal, and three values of moisture contents for sub-bituminous coal were taken into consideration.
Fig. 1. Schematic of pulverized coal burner and supply system.
Experimental Pulverized coal combustion burner Figure 1 shows a schematic diagram of a pulverized coal combustion burner and supplying system. The burner has a coaxial dual piping structure with a main burner port (inner diameter: 6 mm) and an annular slit burner (width: 0.5 mm) installed outside of the main burner port. In -3-
this study, the feed rate of pulverized coal particles were consistently regulated and maintained by a screw-type coal feeder (FEEDCON-µM, Nisshin engineering) and coal particles were carried by the main air flow. Methane was supplied from the slit burner port outside the main burner port in the dual pipe structure to stabilize the two-phase jet flame. The air and methane flow rates were controlled by a mass flow controller (KOFLOC-3660, Kojima instruments). Table 1 shows the proximate analysis, ultimate analysis, and heating values for five different types of coal (NCA, ECM, Vitol, PCC, and MHU). The pulverized coal feed rate, air flow rate and CH4 flow rate was 1.04 × 10-4 kg/s, 5.00 × 10-4 m3/s and 5.00 × 10-5 m3/s, respectively. Table 1 Properties of five coals used in this study
Coals a
M
Proximate analysis
Ultimate analysis
(wt. %, air-dry)
(wt. %, dry)
b
VM
c
e
CV
d
FR
FC
Ash
C
H
O
N
S
Ash
(kcal/kg, air-dry)
NCA
2.88
24.77
57
15.35
72.62
3.99
5.69
1.55
0.35
15.8
2.30
6907
ECM
2.58
32.51
49.86
15.05
68.9
4.54
8.6
1.65
0.87
15.44
1.53
6977
Vitol
3.76
32.87
50.29
13.08
70.68
4.49
9.22
1.46
0.56
13.59
1.53
6954
PCC
15.42
37.83
41.83
4.92
70.41
5.38
16.02
1.39
0.99
5.81
1.11
6918
MHU
14.5
40
40.1
5.4
71.4
4.97
14.78
1.47
1.06
6.32
1.00
6930
a
M: Moisture; bVM: Volatile Matter; cFC: Fixed Carbon; dFR: Fuel Ratio; eCV: Calorific Value
Fig. 2. Schematic of simultaneous optical measurement system. Simultaneous optical measurement system Figure 2 shows a schematic diagram of the simultaneous optical measurement system. To visualize the pulverized coal flame and to measure mean particle temperature, we used a high-4-
speed camera (FASTCAM-1024PCI, Photron) coupled with an image intensifier (UVi Camera Intensifier, Invisible Vision) and two-color radiation pyrometer (METIS-MQ11, Sensortherm), respectively. An image intensifier and a 431 nm filter (CH* chemiluminescence) were used in conjunction with a high-speed camera to obtain high magnification images of the pulverized coal flame.
In order to estimate the availability of 2-D visualization method, we measured
chemiluminescence intensity in pulverized coal flame along streamwise axis using a MICRO system [27,28]. Because this cassegrain optics has quite high spatial resolution, superposition due to line of sight can be eliminated effectively. The collected emission was sent to the photomultiplier spectroscopic unit which consists of an optical band-pass filter (center wavelength, FWHM, transmitting efficiency) and a photomultiplier (R106UH, Hamamatsu) for detecting the CH* (431.4 nm, 1.5 nm, 49.2%) emission intensity. The detailed optical system for local point measurement is described in [26-29]. The analog current signals from the photomultipliers were converted into a voltage signal by an I/V converter, and digitized by an A/D converter (NI-DAQ, National Instruments) at a sampling rate of 5 kHz. The two-color radiation pyrometer [30], which measure temperature from the ratio of radiation signals of two adjacent wavelengths (0.7 to 1.1 µm), was used to measure the temperature of the pulverized coal particles. The particle temperature was measured at the center of the flame in the axial direction in interval of 5 to 10 mm. The analog current signal on the basis of the ratio of the radiation intensity of two adjacent wavelengths were converted into a voltage signal by an I/V converter, and digitized by an analog-to-digital converter at a sampling rate of 5 kHz. The methane flow rate was the minimum amount required to form a stable pulverized coal flame.
Results and Discussion Effect of coal type (volatile matter) The pulverized coal combustion phenomena proceed in the following order. First, pulverized coal is preheated from the burner port and the moisture is dried. Then its volatile matter is emitted and ignited simultaneously. Both volatile matter and fixed carbon burn after reaching the maximum temperature of the particles. Finally, the temperature is rapidly increased because of radiative and conductive heat transfer and then the coal transforms to ash through complete combustion [31]. To investigate the effects of pulverized coal properties on the heat release region in a turbulent jet pulverized coal flame, five different types of coal were selected as shown in Table 1. In this experiment, because turbulent pulverized coal flame was opened to atmosphere, char reaction could not maintain in the downstream. Visualized emission region, therefore, indicates from ignition region to end of volatile matter reaction including partial char reaction. -5-
Figure 3 shows CH* chemiluminescence intensity by the MICRO system and that by intensified high-speed camera measured at the center of the flame in the axial direction. The emission intensity distribution of CH* chemiluminescence can be represented by a ninth-order polynomial equation. From the intersection of the straight line (dotted thick line) with the baseline, the onset of ignition (assumed to correspond to the onset of devolatilization) and the end of volatile reaction can be estimated [24]. The onset of heat release region, defined as the onset of the CH* signal, occurs around 20 mm downstream from the burner rim. The length of heat release region by intensified high-speed camera was about 30 mm (13%) longer than that by MICRO system. This overestimation by the high-speed camera is ascribed to a superposition of emission along the line-of-sight. MICRO system was lifted to keep horizontal balance, but camera was kept at same position.
Fig. 3. Averaged CH* chemiluminescence image and its intensity profile including polynomial fitting based on 2-D measurements by intensified high-speed camera, and point measurements by MICRO system for PCC coal flame. Top (no filtered) and bottom (431 nm filtered) flame images obtained from high-speed (average of 500 images) camera. Figure 4 shows history of heat release region and instantaneous images obtained for PCC coal flame. The onset position and the end of heat release region fluctuated temporally due to complexity of coal ignition process and turbulence.
The averaged onset position of heat
release region, the length of heat release region and the end of heat release region were 38 mm, 234 mm and 272 mm, respectively, though the width of heat release region fluctuated spatially. Because a methane pilot flame was adopted for flame stabilizing, the fluctuation of onset of heat release region was relatively smaller than that of end of heat release region. Because the heat release region of pulverized coal flame fluctuated significantly even for constant mass flow rate, -6-
a multi-dimensional time-series measurement, such as high speed camera, can give us useful information.
Fig. 4. History of heat release region and instantaneous photographs for PCC coal flame. Figure 5 shows estimated lengths of heat release region by an intensified high-speed camera and the MICRO system for five different types of coal in Table 1. The length of heat release region measured by MICRO system (point measurements) increased slightly with increasing volatile matter. The results from 2-D measurements (intensified high-speed camera) showed similar tendency with those from point measurements though averaged discrepancy between two measurement methods was about 11.4%. The discrepancy between point measurement and 2-D measurement increased in proportional to the length of heat release region. Note that some compensation is needed in order to estimate the heat release region using high speed camera.
Fig. 5. Estimated length of heat release region according to volatile matter content (%). -7-
Effect of coal particle size To clarify the effects of coal particle size on the heat release region, experiments with coal particle sizes of 35~45 µm and 75~90 µm were conducted. The coal (Vitol in Table 1) samples from the pulverizer mills in thermal power stations were size classified using standard sieves (325 to 450 mesh and 170 to 200 mesh). Figure 6 shows direct photographs of pulverized coal flames and the effects of coal particle size on the heat release region based on the MICRO system. The length of heat release region of averaged 500 images for 35~45 µm and 75~90 µm were 220 mm and 270 mm, respectively. For the smaller particle size, because the total particle reaction surface area increased and chemical reaction and heat transfer rates also increased, the flame propagation velocity increased definitely. The length of the heat release region for the smaller particle size was reduced 17 mm (8%) compared to the larger one. CH* chemiluminescence intensity and flame luminosity of smaller particle coal flame were stronger at the upstream of the flame compared with those of larger particle one. This strong intensity of flame emission resulted from enhanced chemical reaction rate.
Fig. 6. Examples of CH chemiluminescence intensity and polynomial fit for Vitol coal flame 75~90 µm particle diameter, 35~45 µm particle diameter. Top (75~90 µm) and bottom (35~45 µm) flame images (no filtered) obtained from high-speed (average of 500 images) camera. Figure 7 shows the mean particle temperature with different particle sizes on the central axis of the pulverized coal flame. The temperature rise of the 35~45 µm particle size was faster than the 75~90 µm particle size, and the peak temperature was 35 K (3%) higher. These results indicate that mixing of the volatiles with air and volatile reactions at the particle surfaces are more active in the case of the smaller particle size.
-8-
Fig. 7. Size effects on mean particle temperature profiles for Vitol coal flame. Effect of coal moisture content The coal drying process can improve combustion performance such as an increased flame temperature and low unburned carbon compared to that of raw coals [5,32,33]. To investigate the effect of coal moisture content on the heat release region, experiments for three cases (1.52, 9.49 and 15.42%) of PCC coal were conducted. The PCC coal was characterized as subbituminous with an inherent moisture content of 15.42% in Table 1. The distribution of CH* chemiluminescence intensity measured by the MICRO system at the center of the axial direction of the flame with coal moisture content is presented in Figure 8. The length of heat release region for the three cases was almost identical regardless of moisture content of coal. The discrepancy between the three cases was only 3 mm (1.1%).
Fig. 8. Moisture effects on heat release region for PCC coal flame. Figure 9 shows profiles of mean particle temperature at the center in the axial direction of the flame as a function of coal moisture content. The averaged particle temperatures for each case -9-
were 1401 K, 1422 K, and 1444 K, respectively. The moisture content had little effect on the length of the heat release region and the mean particle temperature compared to volatile matter and coal size when it is less than 15%. That is, coal drying process may be effective up to 15%.
Fig. 9. Moisture effects on mean particle temperature for PCC coal flame.
Conclusions The effects of pulverized coal properties on the heat release region were investigated in turbulent jet pulverized coal flame. Five types of pulverized coal (two particle sizes for a bituminous coal and three moisture contents for a sub-bituminous coal) were considered. The main results of this study are summarized as follows: 1. The length of the heat release region by 2-D measurements was about 11.4% longer than by point measurements, and increased proportionally to the volatile matter content. 2. The temperature rise of the 35~45 µm coal particles was faster than the 75~90 µm particles, and the peak temperature was 3% higher. The length of the heat release region was reduced by about 10%. 3.
Moisture content had little influence on pulverized coal flame when the moisture content was below 15%.
Acknowledgement This work is supported by research [2007-C-CD12-01-3-010] of the Korea Energy Management Corporation and project of the Pusan Clean Coal Center Research Grant.
References [1] L. D. Smoot, P. J. Smith, Coal combustion and gasification, Plenum press, New York, 1985. [2] A. Williams, M. Pourkashanian, J. M. Jones, Prog. Energy Combust. Sci. 27 (2001) 587-610. [3] H. Katalambula, J. Hayashi, K. Kitano, Journal of Chemical Engineering of Japan 33 (2000) - 10 -
49-56. [4] H. Katalambula, J. Hayashi, K. Kitano, T. Chiba, Energy and Fuels 11 (1997) 1033-1039. [5] R. Kurose, H. Tsuji, H. Makino, Fuel 80 (2001) 1457-1465. [6] Y. Zhao, H. Y. Kim, S. S. Yoon, Fuel 86 (2007) 1102-1111. [7] K. Annamalai, W. Ryan, S. Dhanapalan, Prog. Energy Combust. Sci. 20 (1994) 487-618. [8] S. C. Saxena, Prog. Energy Combust. Sci. 16 (1990) 55-94. [9] T. Suda, K. Masuko, J. Sato, A.Yamamoto, K. Okazaki, Fuel 86 (2007) 2008-2015. [10] J. O. L. Wendt, Prog. Energy Combust. Sci. 6 (1980) 201-222. [11] M. Zhu, H. Zhang, G. Tang, Q. Liu, J. Lu, G. Yue, S. Wang, S. Wan, Proc. Combust. Inst. 32 (2009) 2029-2035. [12] L. D. Timothy, A. F. Sarofim, J. M. Beer, Symposium (international) on Combustion 19 (1982) 1123-1130. [13] M. Gieras, R. Klemens, P. Wolanski, S. Wojcicki, Symposium (international) on Combustion 21 (1986) 315-323. [14] W. J. Mclean, D. R. Hardesty, J. H. Pohl, Symposium (international) on Combustion 18 (1981) 1239-1248. [15] H. Kobayashi, J. B. Howard, A. F. Sarofim, Symposium (international) on Combustion 16 (1977) 411-425. [16] J. B. Howard, R. H. Essenhigh, Symposium (international) on Combustion 11 (1967) 399408. [17] J. Yu, M. Zhang and J. Zhang, Proc. Combust. Inst. 32 (2009) 2037-2042. [18] S. M. Choi, KSME International Journal. 4 (1990) 71-77. [19] W. R. Seeker, G. S. Samuelsen, M. P. Heap, J. D. Trolinger, Symposium (international) on Combustion 18 (1981) 1213-1226. [20] J. Ballester, T. García-Armingol, Prog. Energy Combust. Sci. 36 (2010) 375-411. [21] A. G. Gaydon, The Spectroscopy of Flames, 2nd ed., Wiley, New York, 1974. [22] N. Docquier, S. Belhalfaoui, F. Lacas, N. Darabiha, Proc. Combust. Inst., 28 (2000) 17651774. [23] F. Lacas, B. Leroux, N. Darabiha, Proc. Combust. Inst. 30 (2005) 2037-2045. [24] A. Molina, C. R. Shaddix, Proc. Combust. Inst. 31 (2007) 1905-1912. [25] C. R. Shaddix, A. Molina, Proc. Combust. Inst. 32 (2009) 2091-2098. [26] F. Akamatsu, T. Wakabayashi, S. Tsushima, M. Katsuki, Y. Mizutani, Y. Ikeda, N. Kawahara, T. Nakajima, Meas. Sci. Technol. 10 (1999) 1240-1246. [27] S. M. Hwang, R. Kurose, F. Akamatsu, H. Tsuji, H. Makio, M. Katauki, JSME International Journal 49 (2006) 1316-1327. [28] S. M. Hwang, R. Kurose, F. Akamatsu, H. Tsuji, H. Makio, M. Katauki, Energy and Fuels - 11 -
19 (2005) 382-392. [29] J. R. Kim, F. Akamatsu, G. M. Choi, D. J. Kim, Thermochim. Acta 491 (2009) 109-115. [30] S. M. Godoy, F. C. Lockwood, Fuel 77 (1998) 995-999. [31] H. Tsuji, A. K. Gupta, T. Hasegawa, M. Katsuki, K. Kishimoto, M. Morita, High Temperature Air Combustion, CRC Press LLC, 2003. [32] A. Bosoaga, N. Panoiu, L. Mihaescu, R. I. Backreedy, L. Ma, M. Pourkashnian, A. Williams, Fuel 85 (2006) 1591-1598. [33] B. W. Asay, L. D. Smoot, P. O. Hedman, Combust. Sci. Technol. 35 (1983) 15-31.
- 12 -
Mathematical Model of the Low-Temperature Oxidation of Coal in Coal Stockpiles and Dumps Ing. Marian BOJKO, Ph.D, Department of Hydromechanics and Hydraulic Equipment, Faculty of mechanical Engineering, VŠB-Technical University of Ostrava, tř. 17. listopadu 15, 708 33 Ostrava-Poruba, tel. (+420) 597 324 385, e-mail [email protected]
Doc. RNDr. Milada KOZUBKOVÁ, CSc.,
Department of Hydromechanics and Hydraulic Equipment, Faculty of mechanical Engineering, VŠB-Technical University of Ostrava, tř. 17. listopadu 15, 708 33 Ostrava-Poruba, tel. (+420) 597 323 342, e-mail [email protected]
Ing. Zdeněk MICHALEC,
Institute of Geonics AS CR, v. v. i., Studentská 1768, 708 00 OstravaPoruba, tel. (+420) 596 979 228, e-mail [email protected]
Abstract Article defines mathematical model of the low-temperature oxidation of bituminous coal. The mathematical model defines single phase mathematical model with porous zone as coal where consumption of oxygen, production of smoke exhaust and heat are solved as source terms in transport equations. The rate constant defines by Arrhenius expression. Parameters of Arrhenius equation (activation energy and pre-exponential factor) are determined from experimental measuration. For numerical calculation method of finite volume (software ANSYS FLUENT 12) was used. 1
INTRODUCTION
Problem of low-temperature oxidation of coal stockpiles and dumps is presented as dangerous risk with respect to environment, when influence of atmospheric conditions causes self ignition of coal stockpile. The spontaneous ignition of coal matter in coal stockpile and dump is strongly dependent on wind velocity. Flow in coal stockpile or dump is laminar and flow in boundary layer is fully turbulent. Then the simulation includes two diametrically different regimes of flow. This problem is characterized by changes of coal temperature, humidity and oxygen consumption [1], [2], [3], [4]. Low-temperature oxidation of coal is heterogeneous chemical reaction producing carbon dioxide ( CO2 ), carbon monoxide ( CO ), water vapor ( H 2 O ) and oxy-coal. Solution of this problem can be solved by two methods, i. e. solution of two-phase mathematical model with heterogenous reaction or simplified by single phase model. The simplified single phase approach is based on experimental data and kinetics of reaction and will be tested. 2
THERMO-KINETICS DATA OF COAL FROM LABORATORY EXPERIMENT
Low-temperature oxidation of bituminous coal proceeds in two regimes. In the first one the gradual accumulation of heat is observed (so called incubation process) which can keep on several weeks. Subsequently coal oxidation is noticeably accelerated and in case of optimal conditions the temperature accumulation passes into spontaneous combustion. Turning point is characterized by critical temperature which is changed in interval 70°C ÷ 100°C. For these two regimes there are two different chemical reaction mechanisms defined using reaction rate determined by Arrhenius equation. Coal oxidation reaction is a function of temperature and oxygen concentration (O2). The basic parameters of Arrhenius equation are thermo-kinetic data (activation energy and pre-exponential factor) defined often by 1
experimental measurement. From literature sources it is evident the great variance of thermokinetic data see [5], [6], in particular in case of coal types. In this paper the thermo-kinetic data from laboratory experiment using method of adiabatic oxidation [7] for numerical simulation are applied. Activation energy E and pre-exponential factor A are input parameters in Arrhenius equation and are used for definition of rate constant k as follows:
E
k A e RT where R 8.3145J / molK is universal gas constant.
(1)
Critical temperature from experimental measuration is defined Tkr=350.24 K. For first and second oxidation regime the activation energy and pre-exponential factor are evaluated in table 1. Tab. 1 - Activation energy E and pre-exponential factor A Temperature T≤Tkr T≥Tkr E [kJ.mol-1] 9.98 55.4 -1 A [s ] 0.00345 20600 The above evaluated values were used for numerical simulation. 3
MATHEMATICAL MODEL OF SINGLE PHASE LOW-TEMPERATURE OXIDATION
The region is supposed as porous region consisting of coal particles in case of laminar flow with superficial velocity (our problem) using porosity and inverse values of permeability, which is accounted for in continuity, momentum and all scalar transport equations: continuity equation u j 0 (2) t xj where [kg.m-3] is density of mixture and u j [m.s-1] is the component of velocity vector. momentum equation ui ui u j ui p f i ui t xj xi x j xj
(3)
where p [Pa] is static pressure, [Pa.s] is molecular viscosity, g [m.s-2] is gravity acceleration vector, [m2] is permability of porous zone. Local mass fraction of specie Yi Error! Reference source not found. is defined in the form m V Yi i i i i i (4) m V where mi [kg] is mass of specie i , m [kg] is total mass of mixture, i [1] is volume fraction of specie i in mixture. Another variable which is using in theory of species is molar concentration Ci [kmol.m-3]. equation of species transport in conservative form is in form [8]
2
Yi u jYi J j ,i Ri Si t x j xi
where J j ,i Di,m
(5)
Yi [kgm-2s-1] is diffusion flux, Di,m [m2s-1] is diffusion coefficient, Ri x j
[kg.m-3.s-1] is the net rate of production of species i by chemical reaction, S i is the rate of creation by addition from the dispersed phase plus any user-defined sources [8]. An equation of this form will be solved for N 1 species where N is the total number of fluid phase chemical species present in the system. The mass fraction of the species must sum to unity. energy equation T g c pg (1 ) s c ps c pgu jT jl u j t x j xl (6) T c pTJ j ,i S h x j x j x j where T [K] is temperature, is porosity of the medium, s , g [kg.m-3] is solid resp. fluid medium density, c ps , c pg [J.kg-1.K-1] is solid resp. fluid medium specific heat, s , g [W.m.K-1] is solid resp. fluid medium thermal conductivity, [W.m-1.K-1] is effective thermal conductivity g 1 s and S h is source term. To respect the influence reaction mechanism between porous zone and fluid it is essential to add the source terms of reactant and product and energy including dependence on kinetic energy of reaction rate. The source terms are related to reaction rate of coal oxidation expressed by Arrhenius formula. Relation reaction rate of coal oxidation on temperature must be derived from experiment [7]. SO YO A exp E RT 1
2
2
where A is pre-exponential factor, E is activation energy for the reaction, R is universal gas constant. 4
SIMULATION OF LOW TEMPERATURE OXYDATION IN REAL TERRAIN WITH COAL STOCKPILE
CFD simulation perform in 4,3 x 4,3km and high 400m domain. This domain includes real Earth's surface, mine buildings and coal stock, Fig.1. The mine and coal stack is situated in Orlová Lazy near city Karviná.
Fig. 1 – Coal stock and mine buildings
3
4.1 Grid, physical properties, boundary conditions
Domain included entering surfaces, which were used to unification of air inlet and outlet altitude and flow area respectively. This modification make possible the definition of turbulent quantities. The mesh size ratio and cell size near the wall must be observed, because turbulent quantities equilibrium is strongly dependent on quality of mesh [8]. The mesh size is relatively extensive (ca 1mil. cells) if the mesh criteria are within accepted limits, see Fg. 2. The number of cells must be smaller with respect to implementation of coal oxidation and spontaneous ignition. Mine buildings can be substituted by simply porous domain, which has pressure drop equal to origin domain with buildings see Fig.2. Complicated geometry of the buildings was replaced by simply porosity volume. Resistant coefficient of the porous domain was specified by means of thin domain near the buildings. Pressure resistance coefficient is calculated by means of pressure drop and change of reference velocity of wind for real geometry with buildings. N E
W
Inlet 4
Inlet 3
S
Fig. 2 Grid of terrain with coal stock and detail of mine buildings Physical properties Physical properties of coal (density, specific heat, thermal conductivity) are defined in Tab 2. Values of permeability and porosity of coal are gained from literature [9] and moreover there exit dependence between permeability and porosity for spherical particles in form: d 2 3 (7) 2 1501 In this table the constants used in source terms in mathematical model. Tab. 2 – Physical properties of coal, porosity and permeability of coal stock Unity Coal -3 Density kg.m 1300 -1 -1 Specific heat J.kg .K 1090 Thermal conductivity W.m-1.K-1 0.275 Porosity [-] 0.15
4
[m-2]
Permeability Boundary conditions
7.785.10-12
CFD simulation is calculated with mesh see Fig 2. Only two dominant directions of the wind are simulated. Two dominant directions of the wind are South and East. Velocity profile in boundary layer is described by power function, and temperature profile is described by linear function. Atmosphere is stable, so the temperature of air degreasing with altitude. The air compressibility influence is needed for boundary condition definition, because the domain high is inconsiderable and density of air is function of altitude. Velocity profile is defined by power function (8) with reference velocity vref 5 m/s. Reference velocity is defined in zref 10 m altitude. Power coefficient for stable state of atmosphere is p 0.14 . v vref
z 10
p
(8)
where: z – altitude m ,
m v ref – reference velocity in 10 m altitude , s p – power coefficient - , Temperature profile is defined by linear function, which describes degreasing of temperature as function of altitude. Temperature of Earth's surface is chosen t 0 19 °C and T0 292 K respectively. Turbulent kinetic energy and altitude
k
and dissipation are defined by frictional velocity v 0.2258
3
v 0.4 z
(9)
2
k
v 0.3
(10)
where:
m v* – frictional velocity , s Previous formulas is used for formulation of boundary profiles on inlets. Inlet of domain is defined by means of mass flow rate and direction vector. 4.2 Results of numerical calculation
Results of numerical calculation are evaluated by velocity vectors in two cross sections from East direction of wind, see Fig. 2. From results is evidently that maximum of velocity is 1.29 m/s. Then from result you can see detail of velocity field around of coal stock.
5
Fig. 3 – Velocity vectors from East direction of wind with detail of coal stock Next result is the temperature field through coal stock from East direction of wind, see Fig. 4. The range of temperature is <293K ÷ 306K>. Maximum temperature is near Earth surface and surface of coal stock.
Fig. 4 – Temperature field through coal stock from East direction of wind
Fig. 5 – Temperature field through coal stock from South direction of wind 6
Fig. 5 shows the temperature field through coal stock from South direction of wind. You can see that maximum of temperature is 297K and position of maximum is below the surface of coal stock. Comparison between simulations of various direction of wind is show in Fig. 6. 410
Temperature [K]
390
370
South
North
East
West
350
330
310
290 0
50
100
150
200
250
300
Time [days]
Fig. 6 – Maximum of temperature histories inside of coal stock for all directions of wind 5
CONCLUSIONS
This paper occupies with numerical solution of low-temperature oxidation of coal. In the first part are using experimental results of spontaneous combustion kinetics and rate in two temperature intervals, where the parameters of Arrhenius rate (activation energy and preexponential factor) were derived. Then the singe phase mathematical model of lowtemperature oxidation was defined. This model used the source term of oxygen consumption (O2) and production of carbon dioxide (CO2), carbon monoxide (CO), water (H2O) and heat. These terms were formed by user´s defined functions. Subsequently this model was used in real terrain with coal stockpile. Numerical calculation was realized for four direction of wind (North, South, East, West). From results of maximum of temperature histories inside of coal is evidently that direction of wind is important for spontaneous heating. Work is financed by project Czech Science Foundation - GA 105/08/1414 Mathematical modelling of coal spontaneous in coal stockpiles and dumps REFERENCES [1]
[2] [3] [4] [5]
KRISHNASWAMY, S., GUNN, R. D., AGARWAL, K. P. Low-temperature oxidation of coal 2. An eprimental and modelling investigation using a fixed-bed isothermal flow reactor, Fuel, 1996, Vol. 75, No. 3, pp. 344-352. ISSN 0016-2361. SUJANTI, W., ZHANG, D. A laboratory study of spontaneous combustion of coal: the influence of inorganic matter and reactor size. Fuel, 1999, Vol. 78, pp. 549-556. ISSN 0016-2361. YUAN, L., SMITH, A. Numerical study on effects of coal properties on spontaneous heating in longwall gob areas, Fuel, 2008, Vol. 87, pp. 3409-3419. ISSN 0016-2361. ARISOY, A. AKGUN, F. Modelling of spontaneous combustion of coal with moisture content included, Fuel, 1994, Vol. 73, pp. 281-286. ISSN 0016-2361. KRAJČIOVÁ, M. JELEMENSKÝ, L., KIŠA, M., MARKOŠ, J. Model prediction on self-heating and prevention of stockpiled coals. Journal of Loss Prevention in the Process Industries, 2004, (17), pp. 205-216. ISSN 0950-4230.
7
[6] [7]
[8] [9]
ROSEMA, A., GUAN, H., VELD, H. Simulation of spontaneous combustion, to study the cause of coal fires in the Rujigou Basin. Fuel, 2001, Vol. 80, pp. 7-16. ISSN 0016-2361. CYGANKIEWICZ, J. Badanie skłonności polskich węgli do samozapalenia metodą testu adiabatycznego. In Větrání a bezpečnost dolů, mezinárodní kongres, Ostrava, VŠB-TU Ostrava, 2000, str, 22-57. FLUENT: Fluent 12 - User’s guide, Fluent Inc. 2007. VŠB-TU Ostrava. URL:http:// http://spc.vsb.cz/portal/cz/documentation/manual/index.php. LIMING, Y., SMITH, A. C. CFD modeling of spontaneous heating in a large-scale coal chamber. Journal of Loss Prevention in the Process Industries. 22, 2009, p. 426-433. ISSN: 0950-4230.
8
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission PROGRAM TOPIC : 2. GASIFICATION – SYNGAS CLEANUP SLIPSTREAM TESTS OF PALLADIUM SORBENTS FOR HIGH TEMPERATURE CAPTURE OF MERCURY, ARSENIC AND SELENIUM FROM FUEL GAS Hugh G.C. Hamilton (Dr.) Johnson Matthey Technology Centre e-mail: [email protected] Tony Wu : [email protected] Subhash Datta : [email protected] Robert C. Lambrecht : [email protected] John Wheeldon : [email protected] National Carbon Capture Center Southern Company, Wilsonville, Al 35186 U.S.A. Tel : 00 1 205 670 5875 Fax : 00 1 205-670-5843 Evan J. Granite : [email protected] Henry W. Pennline : [email protected] National Energy Technology Laboratory, United States Department of Energy P.O. Box 10940, Pittsburgh, PA 15236-040 USA Tel : 00 1 412 386 4607 Fax : 00 1 412 386 6004 Hugh Hamilton : [email protected] Liz Rowsell : [email protected] Stephen Poulston : [email protected] Andrew Smith : [email protected] Johnson Matthey Technology Centre, Blounts Court, Sonning Common, Reading, BerkshireRG4 9NH UK Tel : 00 44 (0)118 924 2000 Fax : 00 44 (0)118 924 2254
Keywords: National Carbon Capture Center (NCCC), mercury, arsenic, selenium, gasification, sorbent Abstract: In gasification for power generation, the removal of mercury and other trace elements such as arsenic, selenium and phosphorus by sorbents at elevated temperatures preserves the high thermal efficiency of the integrated gasification combined cycle system. Most commercial sorbents display poor capacity for elemental mercury at elevated temperatures. Palladium is an attractive sorbent candidate for the removal of mercury and the trace elements from fuel gases at elevated temperatures. The National Carbon Capture Center at the Power Systems Development Facility (PSDF) in Wilsonville, Alabama, is a large-scale flexible test facility established to develop and demonstrate a wide range of advanced power generation technologies that are critical to developing highly efficient power plants that capture carbon dioxide. The palladium-based sorbents have been tested for
extended periods of time in slipstreams of fuel gas at the NCCC. These results will be described, and possible future testing will be discussed.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: GASIFICATION EXACT TITLE OF PAPER: MERCURY MEASUREMENT AND REOMOVAL FROM AN ENTRAINED FLOW SLAGGING COAL GASIFIER
Dennis Lu, Research Scientist, CanmetENERGY/Natural Resources Canada 1 Haanel Drive, Ottawa, Ontario, CANADA K1A 1M1, [email protected], (613) 996-2760 Robin Hughes, Research Engineer, CanmetENERGY/Natural Resources Canada 1 Haanel Drive, Ottawa, Ontario, CANADA K1A 1M1, [email protected], (613) 996-0066 Karl Abraham, Senior Program Engineer, Environment Canada 1 351 St-Joseph Blvd., 9th Floor, Gatineau, QC K1A 0H3 CANADA, [email protected], (819) 953-2079 Ben Anthony, FBC and Gasification Group Leader, CanmetENERGY/Natural Resources Canada 1 Haanel Drive, Ottawa, Ontario, CANADA K1A 1M1, [email protected], (613) 996-2868
Abstract: In typical synthesis gas (syngas) from entrained flow slagging coal gasification thermodynamic calculations predict that only the elemental form of mercury (Hg0) is stable rather than the bivalent oxidized form (Hg ++), such as is present in HgCl2. Therefore, Hg0 is expected to be dominant in such a reducing environment. However, the chemical and physical processes governing the interactions of mercury forms with syngas components are poorly understood, particularly the results of heterogeneous reactions involved in gasification syngas are lacking. Data on Hg emissions from gasification systems have not been sufficiently reliable and the mass balance closures have high associated error ranges because of problems with sampling and analysis, which make understanding mercury characteristics under gasification conditions difficult. This paper presents studies on mercury measurement specifically applicable to an entrained flow slagging gasifier at the CanmetENERGY 0.6MW pilot scale gasification plant. Mercury speciation has been successfully measured directly from a high-P and high-T gasifier vessel and as well quenched downstream syngas containing nitrous and sulfurous species. A bench-scale fixed-bed system was also used to investigate the Hg removal performance of sorbents, including commercial activated carbon, sulphur- and alkali-doped activated carbons, limestone and dolomite. The fixed bed was operated above the dew point temperature of the synthesis gas for the activated carbon sorbents, and for CaO-based sorbents at a higher temperature in the range of 500-700°C, which has been chosen to match the operating conditions of the CaO-sorbent looping process for CO2 capture.
PERFORMANCE IMPROVEMENT OF A DESULFURIZATION SORBENT FOR WARM SYNTHESIS GAS CLEANUP Jeom-In Baek, Jungho Ryu, Tae Hyoung Eom, Joong Beom Lee, Yong-Ho Wi, Chong Kul Ryu* KEPCO Research Institute 305-380 Munji-ro 65, Yuseong-gu, Daejeon, Republic of Korea Tel: 82-42-865-5250, Fax: 82-42-865-5202, e-mail: [email protected]
Abstract KEPCO Research Institute has improved the performance of a solid regenerable desulfurization sorbent prepared by spray-drying method. Here, we present a newly developed desulfurization sorbents which showed improved physical properties and reactivity compared to our previous desulfurization sorbents. The attrition resistance of the new desulfurization sorbents was much higher than the previously developed sorbents. Other physical properties such as average particle size, tap density, and shape were suitable for the fluidized-bed applications. Sulfur sorption capacity of the new sorbent, which was measured by thermogravimetric analyzer using a simulated synthesis gas containing 1% H2S, was around 10 wt% at the reaction temperatures of 500 and 650 oC for absorption and regeneration, respectively. In the future works, an in-depth sorbent analysis and reactivity test according to the reaction temperature change will be carried out to improve the performance of the spray-dried desulfurization sorbent and a fluidized-bed desulfurization process. Introduction Integrated Gasification Combined Cycle (IGCC) is considered as a clean power generation technology with a high thermal efficiency and low emission. The higher cost of electricity (COE) of IGCC power plant than that of pulverized coal (PC) power plant has been barrier to the widespread use of IGCC plant. However, when considered CO2 capture together the COE of coal IGCC with CO2 capture is evaluated less than that of a PC power plant with CO2 capture (DOE/NETL, 2007; IEA, 2008). The international cooperation to reduce CO2 emission for the mitigation of climate change is demanding more cost-efficient power plants with CO2 capture process. Therefore, it is expected that the world IGCC capacity will be rapidly increased in the near future. It was suggested that the efficiency of IGCC without CO2 capture would decrease around 69% and the COE would increase 3240% when the currently available commercial CO2 capture process, Selexol, was adopted (DOE/NETL, 2007). Minimizing the COE increase and efficiency loss is required to accelerate the deployment of IGCC with CO2 capture. New technologies such as
advanced air separation unit, warm gas clean-up and CO2 capture at warm temperatures should be developed to reduce the COE increase of IGCC caused by CO2 capture. A raw synthesis gas from a coal gasifier contains hydrogen sulfide (H2S). It is necessary to remove H2S to avoid the poison of the catalyst used in downstream and to protect the plants from corrosion. Current commercial-technologies to remove H2S are using wetscrubbing method which uses physical solvents or chemical amines. The temperature of gas stream in this capture process should be cooled down to around 40 oC, which results in a efficiency loss. Warm synthesis gas clean-up is attractive because it can be operated at an elevated temperature, reducing the efficiency loss. Therefore, warm synthesis gas cleanup is considered as a key technology to develop an advanced IGCC system with high thermal efficiency. Solid regenerable sorbents are used for H2S removal in warm synthesis gas clean-up. A lot of work has been done to develop solid sorbents for desulfurization at warm temperatures (Bu et al., 2008; Gupta and Gangwal, 1992; Gupta et al., 1997; Lee et al., 2008; Sanchez-Hervas et al, 2005). RTI/Eastman has been developing desulfurization process using ZnO-based sorbents. In the tests at a pilot plant installed in Eastman’s Kingsport coal gasification plant, sulfur removal more than 99.9% was proved during operation over 3000 h (Gupta et al., 2008). KEPCO Research Institute also has been developing solid regenerable desulfurization sorbents for warm synthesis gas clean-up. ZAC-32N prepared by spray drying method showed >99% sulfur removal from coalderived synthesis gas during continuous steady state operation for 720 h in a bench scale circulating fluidized bed process, which was developed by Korea Institute Energy Research (KIER). To improve the performance of ZAC-32N, several new desulfurization sorbents including PS-2 were designed and tested (Baek et al, 2009). Although some properties such as surface area and sulfur sorption capacity of PS-2 were improved, the mechanical strength of PS-2 was less than ZAC-32N. In this study, a new zinc-based sorbent was designed to improve the performances of our previously developed desulfurization sorbents. Physical properties and H2S sorption capacities of newly spray-dried desulfurization sorbent were investigated and compared with previous ones. Experimental Section Sorbent Preparation: New desulfurization sorbent, PS-8, was designed based on the PS-2 composition to improve the mechanical strength. Raw materials for the preparation of PS-8 were composed of 50 wt% ZnO as an active material, 50 wt% matrices containing multi binders, and a Ni-based promoter by 7.5 wt% to improve the attrition resistance and regenerability of the sorbents. Commercially available powder type raw materials were well mixed with pure water. The mixed slurry was comminuted with ball mill to control particle size and make homogeneous colloidal slurry. The homogenized slurry was spray-dried to form free-flowing solid sorbents with suitable shape, particle size, bulk density suitable for fluidized-bed applications in a 10 kg batch scale. The spray-dried green body was calcined at 650 C in a muffle oven for 5 h after pre-drying at 120 C overnight to enhance mechanical strength with burning out the organic additives added during slurry preparation.
Physical Characterization: The shape of calcined desulfurization sorbents was obtained using scanning electron microscope (SEM). Average particle size of the sample was measured using a MEINZER II sieve shaker. The measuring procedure for particle size and size distribution was followed by the American Society for Testing and Materials (ASTM) E-11. Packing density of each sorbent was determined using the Autotap instrument (Quantachrome) proposed in ASTM D 4164-88 to measure the tapped density of formed catalyst. The packing density were obtained by dividing the sorbent mass by its tapped volume. Brunauer-Emmett-Teller (BET) surface area of the sample was determined by N2 physisorption. Pore volume and porosity was obtained with Hg porosimetry. The Hg intrusion volume between particles was subtracted from the total Hg intrusion volume to calculate porosity contributed by the pore volume only within the particle. Attrition resistance of the calcined oxygen carrier was measured with a modified three-hole air-jet attrition tester based on the ASTM D 5757-95. The attrition resistance was determined at 10 standard liters per minute (slpm) over 5 h as described in the ASTM method. The attrition index (AI) is the percent fines generated over 5 h. The fines are particles collected at the thimble after 5 h from the start, which was attached to the gas outlet. AI = [total fine collected for 5 h/amount of initial sample (50 g)] x 100 %
(4)
A lower AI indicates a better attrition resistance of the bulk particles. The AIs of fresh Akzo fluid catalytic cracking (FCC) catalyst, which was used as a reference for comparison, was 22.5%, under the same measurement condition. TGA Chemical Reactivity: Chemical reactivity of the desulfurization sorbents was measured with a thermal gravimetric analyzer (Thermo Cahn, TherMax500) which was modified to supply water vapor into the reactor tube. The overall schematic diagram of TGA system was shown in Figure 1. An alumina crucible was used to hold the desulfurization sorbents. The gas flow rate was controlled by mass flow controller. Reacting gas mixture was introduced into the lower part of the reactor tube and flowed upward. An inert gas, N2, was fed into the balance chamber and flowed downward to the outlet of TGA furnace to protect the balance from the diffusion of reactant gas and reduce the buoyancy effect. The weight and furnace temperature data were continuously recorded by data acquisition unit. H2S absorption and sorbent regeneration were conducted under atmospheric pressure. A simulated synthesis gas composed of 1% H2S, 3% CO2, 30% H2, 64% CO and 2% H2O was used for absorption. Air was used as a regeneration gas. Before the sulfidation step, clean synthesis gas (without H2S) was supplied for 30 min to obtain weight baseline. After absorption, the reactor was purged with N2 and heated to the regeneration temperature. The amount of samples used for reactivity tests were around 20 mg scale. A total flow rate was 0.15 slpm in all steps. To prevent water vapor condensation, electric line heaters were installed along the gas line and gas stream was heated above saturation point. The sulfur sorption capacity was determined from the weight increase during H2S absorption.
Data acquisition
Microbalance E/B
Sytem outlet Vent
Air for pressure control
Back pressure regulator
Balance purge gas
CO2 TGA furnace H2O
N2
N2
CO H2 H2S/CO
Vent
Furnace gas Thermocouple
H2O vapor N2 Air
Figure 1. Schematic diagram of TGA system for desulfurization tests. Results and Discussion Physical Properties: The physical properties of the spray-dried desulfurization sorbents were summarized in Table 1. All three sorbents had spherical shape. The average particle size (115 m) of newly developed sorbent, PS-8, was similar to that of previous sorbents, ZAC-32N and PS-2. Particle size of PS-8 ranged from 58 to 196 m for more than 99 wt% of the sorbents. PS-8 had much narrow particle size range compared to the previously developed sorbents, which will improve the operating stability of a circulating fluidizedbed system. The bulk density of PS-8 was also increased to 0.92 g/cm3, which is more preferable value for fluidized-bed applications. The BET specific surface area of PS-8 (52.9 m2/g) was similar to that of PS-2 (51.3 m2/g) and higher than that of ZAC-32N (39.3 m2/g). The porosity and pore volume of PS-8 were similar to those of ZAC 32N and PS-2. The attrition resistance is an important physical property which decides the applicability of the sorbent for fluidized-bed process. Sorbents with higher AI above 60% are not suitable for bubbling bed reactor and those with AI above 30% are not suitable for transport reactor. The mechanical strength of PS-8 was much higher than that of ZAC 32N and PS-2. And
Figure 2. SEM images of spray-dried desulfurization sorbents.
Table 1. Physical properties of the desulfurization sorbents calcined at 650 °C ZAC 32N PS-2 PS-8 Shape Sphere Sphere Sphere Avg. particle size/µm 111 124 115 Size distribution/µm 38 – 250 57 – 303 58 – 196 Bulk density / (g/cm3) 0.77 0.86 0.92 2 BET / (m /g) 39.3 51.3 52.9 Hg porosity/% 35.6 32.1 32.8 3 Hg pore volume / (cm /g) 0.48 0.44 0.40 Attrition index (AI) / % 31.4 36.0 11.2 it was also higher than that of commercial FCC catalyst (22.5%). The high attrition resistance of PS-8 could be obtained by changing the multi-binder matrices. The increased attrition resistance will contribute to the reduction of sorbent make-up requirement.
Sulfur sorption capacity / wt%
TGA Chemical Reactivity: The sulfur sorption capacity of PS-8 at an absorption temperature of 500 C was presented in Figure 3, along with sorption capacities of ZAC32N and PS-2. The sulfur sorption capacities of sorbents were stabilized from the second cycle. PS-8 showed the sulfur sorption capacity of around 10 wt% after second cycle, which is similar to that of PS-2 and higher than that of ZAC-32N. The higher sorption capacity of PS-8 can be ascribed to the increased surface area resulted from the support matrix change because more active sites can be exposed to the surface. There was significant reduction of sorption capacity at the second cycle, compared with the first cycle. Several explanations can be given for this reduction although they are not clear at this moment. The assumption that zinc migrated to the surface and evaporated during cyclic can be one of the explanations. Another explanation is that sulfur was irreversibly absorbed on the support of sorbents. Detailed analysis on the structure of the fresh and used sorbents should be carried out to find out the accurate cause of the sulfur sorption capacity reduction at the second cycle. 16
ZAC-32N PS-2 PS-8
14
12
10
8 1
2
3
Cycles
Figure 3. Sulfur sorption capacities of the spray-dried desulfurization sorbents
Summary A dry regenerable zinc-based desulfurization sorbent for H2S removal from warm synthesis gas was prepared by spray-drying method. The physical properties of newly developed sorbent, PS-8, were characterized in terms of average particle size (115 µm), size distribution (58196 µm), bulk density (0.92 g/cm3), attrition resistance (AI, 11.2%), surface area (53 m2/g), and porosity (33%). The physical properties of PS-8 were suitable for fluidized bed-applications. Its attrition resistance was greatly improved compared with our previously developed sorbents, ZAC-32N and PS-8, and the sulfur sorption capacity of PS-8 was also improved compared to that of ZAC-32N. In our future works, the sulfur sorption capacity according to the reaction temperature change and the reason for the sorption capacity reduction at the second cycle will be investigated to improve the performance of PS-8 and to provide better operating condition of a warm synthesis cleanup process. Acknowledgement This work was supported by Energy Efficiency and Resources R&D program (2008CCD11P040000) under the Ministry of Knowledge Economy, Republic of Korea. The authors also would like to thank the Korea Electric Power Corporation (KEPCO) and the Korea Western Power Company, Ltd. for supporting the program. References Baek, J.-I.; Eom, T. H.; Lee, J. B.; Wi, W.-H.; Ryu, C. K., “Solid regenerable sorbents for H2S removal from syngas at a warm temperature”, 26th Annual International Pittsburgh Coal Conference, September 2023, 2009. Bu, X.; Ying, Y.; Zhang, C.; Peng, W., “Research improvement in Zn-based sorbent for hot gas desulfurization”, Power Technology 2008, vol. 180, 253258. DOE/NETL, “Cost and performance baseline for fossil fuel plant”, DOE/NETL-2007/1281, USA, 2007. IEA , “CO2 capture and storage”, OECD/IEA, 2008. Gupta, P. R.; Gangwal, S. K., “Enhanced durability of desulfurization sorbents for fluidized-bed applications”, DE93000247, 1992 Gupta, R. K.; Turk, B. S.; Vierheilig, A. A., “Desulfurization sorbents for transport-bed applications”, In proceedings of advanced Coal-based power and environmental systems ’97 conference, Pittsburgh, PA, July 2224, 2997. Gupta, R.; Turk, B.; Lesemann, M., Schlather, J.; Denton D., “Status of RTI/Eastman warm gas clean-up technology and commercialization plans”, Gasification Technologies Conference, October 8, 2008 Lee, J. B.; Baek, J.-I., Ryu, C. K.; Yi, C. K.; Jo, S. H.; Kim, S. H., “ Highly attritionresistant zinc-based sorbents for H2S removal by spray-drying technique”, Ind. Eng. Chem. Res. 2008, vol. 47, 44554464. Sanchez-Hervas, J. M.; Otero, J.; Ruiz, E., “A study on sulphidation and regeneration of Z-
sorb III sorbents for H2S removal from simulated ELCOGAS IGCC syngas”, Chem. Eng. Sci. 2005, vol. 60, 29772989.
Proceedings of the 27th International Pittsburgh Coal Conference, Istanbul, Turkey, October 11-14, 2010
INVESTIGATION OF THE COAL GASIFICATION PROCESS UNDER VARIOUS OPERATING CONDITIONS INSIDE A TWO-STAGE ENTRAINED FLOW GASIFIER Armin Silaen and Ting Wang * [email protected], [email protected] Energy Conversion & Conservation Center University of New Orleans New Orleans, Louisiana, USA through an incomplete combustion. Feedstock is partially combusted with oxygen and steam at high temperature and pressure with only less than 30% of the required oxygen for complete combustion being provided. The syngas produced can be used as a fuel, usually for boilers or gas turbines to generate electricity, or can be used to make a synthetic natural gas, hydrogen gas or other chemical products. The gasification technology is applicable to any type of carbonbased feedstock, such as coal, heavy refinery residues, petroleum coke, biomass, and municipal wastes.
ABSTRACT Numerical simulations of the coal gasification process inside a generic 2-stage entrained-flow gasifier fed with Indonesian coal at approximately 2000 metric tone/day are carried out. The 3-D Navier-Stokes equations and eight species transport equations are solved with three heterogeneous global reactions, three homogeneous reactions, and two-step thermal cracking equation of volatiles. The Chemical Percolation Devolatilization (CPD) model is used for the devolatilization process. Finite rates are used for the heterogeneous solid-to-gas reactions. Both finite rate and eddy-breakup combustion models are calculated for each homogeneous gas-to-gas reaction, and the smaller of the two rates is used. The water-shift reaction rate is adjusted to match available syngas composition from existing operational data without catalyst. This study is conducted to investigate the effects of different operation parameters on the gasification process including coal mixture (dry vs. slurry), oxidant (oxygen-blown vs. air-blown), and different coal distribution between two stages. In the two-stage coal-slurry feed operation, the dominant reactions are intense char combustion in the first stage and enhanced gasification reactions in the second stage. The gas temperature in the first stage for the dry-fed case is about 800 K higher than the slurry-fed case. This calls for attention of additional refractory maintenance in the dry-fed case. One-stage operation yields higher H2, CO and CH4 combined than if a two-stage operation is used, but with a lower syngas heating value. High heating value (HHV) of syngas for the one-stage operation is 7.68 MJ/kg, compared to 8.24 MJ/kg for two-stage operation with 72%-25% fuel distribution and 9.03 MJ/kg for two-stage operation with 50%-50% fuel distribution. Carbon conversion efficiency of the air-blown case is 77.3%, which is much lower than that of the oxygen-blown case (99.4%). The syngas heating value for the air-blown case is 4.40 MJ/kg, which is almost half of the heating value of the oxygen-blown case (8.24 MJ/kg).
1.1 Global Gasification Chemical Reactions This study only deals with global chemical reactions of coal gasification (Smoot and Smith, 1985) that can be generalized in reactions (R1.1) through (R1.9) below: Heterogeneous (solid and gas) phase C(s) + ½ O2 → CO,
ΔH°R = -110.5 MJ/kmol
(R1.1)
C(s) + CO2 → 2CO, ΔH°R = +172.0 MJ/kmol (R1.2) (Gasification, Boudouard reaction) C(s) + H2O(g) → CO + H2, ΔH°R= +131.4 MJ/kmol (R1.3) (Gasification) C + 2H2 → CH4,
ΔHoR = -87.4 MJ/kmol (Direct methanation)
(R1.4)
ΔH°R = -283.1 MJ/kmol
(R1.5)
Homogenous gas phase CO + ½ O2 → CO2,
1.0 INTRODUCTION
CO + H2O(g) → CO2 + H2 , ΔH°R = -41.0 MJ/kmol (Water-shift)
(R1.6)
CO + 3H2 → CH4 + H2O, ΔHoR = -205.7 MJ/kmol (Methanation)
(R1.7)
CH2.121O0.5855 → 0.5855CO + 0.2315H2 + 0.4145CH4 (R1.8) (Volatiles cracking) CH4 + ½ O2 → CO + 2H2 (R1.9) (Volatiles gasification via CH4)
Gasification is the process of converting various carbonbased feedstocks to clean synthetic gas (syngas), which is primarily a mixture of hydrogen (H2) and carbon-monoxide (CO) with minor methane (CH4) and inert nitrogen gas,
1
In this study, the methanation reactions are not considered. Reactions (R1.8) and (R1.9) involve volatiles. The volatiles are modeled to go through a two-step thermal cracking (R1.8) and gasification processes (R1.9) via CH4 as the intermediate product. The coal used in the study is subbituminous from Indonesia, whose compositions are given in Table 1a. It has a moisture content of 8.25%. Its moisturefree (MF) proximate and ultimate analyses compositions are listed in Table 1b. The compositions of volatiles are derived from the values of coal heating value, proximate analysis, and ultimate analysis.
configurations: single-stage down fired system and two-stage with multiple feed inlets. The model was constructed using GLACIER, an REI in-house comprehensive coal combustion and gasification tool. The basic combustion flow field was established by employing full equilibrium chemistry. Gas properties were determined through local mixing calculations and are assumed to fluctuate randomly according to a statistical probability density function (PDF), which is characteristic of the turbulence. Gas-phase reactions were assumed to be limited by mixing rates for major species as opposed to chemical kinetic rates. Gaseous reactions were calculated assuming local instantaneous equilibrium. The particle reaction processes include coal devolatization, char oxidation, particle energy, particle liquid vaporization and gasparticle interchange. The model also includes a flowing slag sub-model.
Table 1a Compositions of Indonesian sub-bituminous coal.
Volatile H2 O ash C H N S O Total, wt % HHV, kcal/kg
Weight % 38.31% 8.25% 3.90% 37.95% 2.68% 0.69% 0.31% 7.91% 100.00% 5690
The U.S. Department of Energy/National Energy Technology Laboratory (NETL) developed a 3D CFD model of two commercial-sized coal gasifiers [Guenther and Zitney (2005)]. The commercial FLUENT CFD software is used to model the first gasifier, which is a two-stage entrained-flow coal slurry-fed gasifier. The Eulerian-Lagrangian approach is applied. The second gasifier is a scaled-up design of transport gasifier. The NETL open source MFIX (Multiphase Flow Interphase eXchanges) Eulerian-Eulerian model is used for this dense multiphase transport gasifier. NETL also developed an Advanced Process Engineering Co-Simulator (APECS) that combines CFD models and plant-wide simulation. APECS enables NETL to couple its CFD models with steady-state process simulator Aspen Plus.
Table 1b Moisture-free (MF) compositions of Indonesian sub-bituminous coal.
Proximate Analysis (MF), wt% Volatile 51.29 Fixed Carbon (FC) 47.54 Ash 1.17 100.00
Ultimate Analysis (MF), wt% C 73.32 H 4.56 O 20.12 N 0.72 S 0.11 Ash 1.17 100.00
Silaen and Wang (2006) carried out a study that focused on the effect of flow injection directions on the gasification performance using the same generic two-stage entrained flow gasifier as studied by Chen et al. and Bockelie et al. Horizontal injection direction was compared to downward and upward direction. The results revealed that the horizontally tangential injection direction gave the best gasifier performance. Changing the direction of the first-stage injectors downward resulted in a carbon fuel conversion reduction, but produced more H2. Changing the direction of the second-stage injectors, however, did little to affect the overall flow patterns due to the smaller-quantity of coal injection (25%); therefore the gasifier performance was essentially insignificantly affected.
1.2 Recent Research Chen et al. (2000) developed a comprehensive threedimensional simulation model for entrained coal gasifiers which applied an extend coal gas mixture fraction model with the Multi Solids Progress Variables (MSPV) method to simulate the gasification reaction and reactant mixing process. The model employed four mixture fractions and separately track the variable coal off-gas from the coal devolatilization, char-O2, char-CO2, and char-H2O reactions. Chen et al. performed a series of numerical simulations for a 200 ton per day (tpd) two-stage air blown entrained flow gasifier developed for an IGCC process under various operation conditions (heterogeneous reaction rate, coal type, particle size, and air/coal partitioning to the two stages).
Silaen and Wang (2010) conducted a study that investigates the effects of different parameters on gasification performance including five turbulence models, four devolatilization models and three coal particle sizes. The Eulerian-Lagrangian approach with finite global reaction rates was applied. A two-step decomposition model was applied to volatiles cracking and gasification via benzene as the intermediate product. The results revealed that the standard kε and the Reynolds Stress Model (RSM) models gave consistent results. Smaller particles have a higher surface/volume ratio, react faster than larger particles, and produce syngas with higher heating value than larger particles. High inertia possessed by larger coal particles propel the
Bockelie et al. (2002) of Reaction Engineering International (REI) developed a CFD modeling capable of entrained flow gasifiers that focuses on two gasifier
2
solid species mass fraction on the surface, and particle surface area. The reaction rates are all global net rates, i.e., the backward reaction, calculated by equilibrium constants, are included in the global rate. Therefore, the finite rate employed in this study implicitly applies to the local equilibrium approach. Reaction rate constants used in this study are summarized in Table 2.
particles cross the gas streamlines and increase particle-gas mixing which result in enhanced reaction rate, but they take longer to complete reaction of the entire particle. The single rate devolatilization model and the chemical percolation model produced moderate and consistent devolatilization rate. This study is the continuous work of Silaen and Wang (2006, 2010) and focuses on investigating the effects of different operation parameters on the gasification process including coal mixture (dry vs. slurry), oxidant (oxygen-blown vs. air-blown), and different coal distribution between two stages.
The reaction rate of the water-shift, adopted from Jones and Lindstedt (1988), is found to be too fast in this study because the rate is obtained with the presence of catalyst. Considering no catalyst is added in a typical gasifier, the water shift reaction rate is purposely slowed down to make the syngas composition consistent with that in the actual production of a commercial entrained-flow gasifier with coalslurry feed from bottom.
2.0 COMPUTATIONAL MODEL The computational model and submodels (devolatilization, reactions, particle dynamics, gasification) used in the study are the same as developed by Silaen and Wang (2010), so all equations and detailed modeling intricacies are not repeated here, but they are briefly summarized below. The time-averaged steady-state NavierStokes equations as well as the mass and energy conservation equations are solved. Species transport equations are solved for all gas species involved. The standard k-ε turbulence model is used to provide closure. Silaen and Wang (2010) applied five turbulence models (standard k-ε, k-ω, RSM, k-ω SST, and k-ε RNG) and reported that the standard k-ε turbulence model yields reasonable results without requiring very much computational time when compared to other turbulence models. Enhanced wall function and variable material property are used. The P1 model is used as the radiation model.
For liquid droplets, water evaporates from the particle’s surface when the temperature is higher than the saturation temperature (based on local water vapor concentration). The evaporation is controlled by the water vapor partial pressure until 100% relative humidity is achieved. When the boiling temperature (determined by the air-water mixture pressure) is reached, water continues to evaporate even though the relative humidity reaches 100%. After the moisture is evaporated, due to either high temperature or low moisture partial pressure, the vapor diffuses into the main flow and is transported away. Please refer to Silaen and Wang (2010) for details. Table 2 Summary of reaction rate constants used in this study Reaction Rate Constant Solid-gas heterogeneous reactions: n k = AT exp(-E/RT) C(s) + ½O2 → CO
C(s) + CO2 → 2CO
n
k = AT exp(-E/RT)
(Gasification, Boudouard reaction) C(s) + H2O(g) → CO + H2
n=0 2
(Combustion)
The flow (continuous phase) is solved in Eulerian form as a continuum while the particles (dispersed phase) are solved in Lagrangian form as a discrete phase. Stochastic tracking scheme is employed to model the effects of turbulence on the particles. The continuous phase and discrete phase are communicated through drag forces, lift forces, heat transfer, mass transfer, and species transfer. The finite-rate combustion model is used for the heterogeneous reactions, but both the finite-rate and eddy-dissipation models are used for the homogeneous reactions, and the smaller of the two is used as the reaction rate. The finite-rate model calculates the reaction rates based on the kinetics, while the eddy-dissipation model calculates based on the turbulent mixing rate of the flow. Gasification or combustion of coal particles undergoes the following global processes: (i) evaporation of moisture, (ii) devolatilization, (iii) gasification to CO and (iv) combustion of volatiles, CO, and char. The Chemical Percolation Devolatilization (CPD) model [Fletcher and Kerstein (1992), Fletcher et. al (1990), and Grant et. al (1989)] is chosen as the devolatilization model based on the finding by Silaen and Wang (2010) that the Kobayashi two-competing rates devolatilization model [Kobayashi et. al. (1976)] is very slow, while the CPD model gives a reasonable result. For solid particles, the rate of depletion of the solid, due to a surface reaction, is expressed as a function of kinetic rate,
Parameters
n
k = AT exp(-E/RT)
(Gasification)
-0.5
A = 0.052 kg/m .Pa 7 E = 6.1x10 J/kmol n=0 2
-0.5
2
-0.5
A = 0.0732 kg/m .Pa 8 E = 1.125x10 J/kmol n=0 A = 0.0782 kg/m .Pa 8 E = 1.15x10 J/kmol
Gas phase homogeneous reactions: n k = AT exp(-E/RT) CO + ½ O2 → CO2
n=0 12
CO + H2O(g) → CO2 + H2 (Watershift)
n
k = AT exp(-E/RT)
A = 2.2x10 E = 1.67x108 J/kmol n=0 2
A = 2.75x10 E = 8.38x107 J/kmol Eddy-dissipation only CH2.121O0.5855 → 0.5855CO + 0.2315H2 + 0.4145CH4 Eddy-dissipation only CH4+ ½O2 → CO + 2H2
The computation is carried out using the finite-volumebased commercial CFD software FLUENT 12.0 from ANSYS, Inc. The simulation is steady-state and uses the pressurebased solver, which employs an implicit pressure-correction scheme and decouples the momentum and energy equations. SIMPLE algorithm is used to couple the pressure and velocity. Second order upwind scheme is selected for spatial discretization of the convective terms. For the finite rate model where the Eulerian-Lagrangian approach is used, the iterations are conducted alternatively between the continuous
3
and the dispersed phases. Initially, two iterations in the continuous phase are conducted followed by one iteration in the discrete phase to avoid the flame from dying out. Once the flame is stably established, twenty iterations are performed in the continuous phase followed by one iteration in the dispersed phase. The drag, particle surface reaction, and mass transfer between the dispersed and the continuous phases are calculated. Based on the dispersed phase calculation results, the continuous phase is updated in the next iteration, and the process is repeated. Converged results are obtained when the residuals satisfy mass residual of 10-3, energy residual of 10-5, and momentum and turbulence kinetic energy residuals of 104 . These residuals are the summation of the imbalance in each cell, scaled by a representative for the flow rate. The computation was carried out in parallel processing on two dual-core Pentium clusters with 12 nodes each. 2.1 Physical Assumptions
Characteristics
of
the
Model
and
This paper studies a two-stage entrained flow coal gasifier as shown in Fig. 1. Meshed computational domain is shown in Fig. 2. The grid consists of 1,106,588 unstructured tetrahedral cells. In the simulations, the buoyancy force is considered, varying fluid properties are calculated for each species and the gas mixture, and the walls are assumed impermeable and adiabatic. The flow is steady and no-slip condition (zero velocity) is imposed on the wall surfaces.
Fig. 2 Meshed computational domain of the two-stage entrained-flow gasifier.
3.0 BOUNDARY AND INLET CONDITIONS Raw Syngas
Indonesian sub-bituminous coal is used as feedstock in this study; its composition is given in Table 1. Boundary conditions for the baseline case are shown in Fig. 3. The summary of the studied cases are listed in Table 3. In the baseline (Case 1) of this study, coal-slurry-fed and two-stage configuration is used with fuel distribution of 75%-25% between the first and the second stages. Total mass flow rates of the coal slurry and the oxidant are 21.9 kg/s and 9.92 kg/s, respectively. The total mass flow rate of the dry coal powder case (Case 2) is 12.90 kg/s. The difference in fuel mass flow rates is caused by water added for slurry. The moisture in the coal is included in both slurry and dry feed cases. The coal/water weight ratio of the coal slurry is 60%-40%. Oxidant/coal slurry feed rate gives O2/coal equivalence ratio of 0.4. The equivalence ratio is defined as the percentage of oxidant provided over the stoichiometric amount for complete combustion.
Top view of second stag injectors 9m
2nd Stage
Coal Slurry
1st Stage
2.25 m
Top view of first stage injectors
The oxidant is considered as a continuous flow and coal slurry is considered as a discrete flow. The discrete phase only includes the fixed carbon and water from the moisture content of coal (8.25% wt) and water added to make the slurry. The slurry coal is treated as particles containing both coal and liquid water. Other components of the coal, such as N, H, S, O, and ash, are injected as gas, together with the oxidant in the continuous flow. N is treated as N2, H as H2, and O as O2. S and ash are not modeled and their masses are lumped into N2. The coal slurry size is uniformly given as 50
Coal Slurry & Oxygen 0.75 m 1.5 m 0.75 m
Fig. 1 Schematic of the two-stage entrained-flow gasifier.
4
μm for the purpose of conveniently tracking the particle size reducing rate. Investigation of effects of coal particle size on gasification performance has been performed by Silaen and Wang (2009 and 2010) and is not repeated here.
4.0 RESULTS AND DISCUSSIONS The following five cases are studied: Case 1: Baseline case, oxygen-blown, coal slurry, 75%-25% distribution in 2 stages Case 2: Oxygen-blown, dry coal, 75%-25% distribution in 2 stages. Case 3: Oxygen-blown, coal slurry, 50%-50% distribution in 2 stages. Case 4: Oxygen-blown, coal slurry, 100% distribution in the 1st stage. Case 5: Air-blown, coal slurry, 75%-25% distribution in 2 stages
The walls are assigned as adiabatic with internal emissivity of 0.8. The boundary condition of the discrete phase at walls is assigned as “reflect”, which means the discrete phase elastically rebound off once reaching the wall. At the outlet, the discrete phase simply escapes/exits the computational domain. The gasifier is operating at 24 atm. P = 24 atm
4.1 Baseline Case (Case 1) • • • •
The baseline case (Case 1) is the two-stage oxygen-blown operation with coal slurry distribution of 75%-25% between the first and the second stages. Gas temperature and species mole fraction distributions on the horizontal and center vertical planes in the gasifier are shown in Fig. 4. The gas temperature is seen higher in the region between the first stage and second stage injection locations than in the region above the second stage injection location. Maximum gas temperature in the first stage reaches 2400K (3860°F). The dominant reaction in the first stage is the intense char combustion (C + ½ O2 → CO and CO + ½ O2 → CO2) in the first stage and gasification reactions (mainly char-CO2 gasification, C + ½ CO2 → CO) in the second stage. Oxygen is completely depleted through the char combustion in the first stage. Char gasification is enhanced in the second stage with the injection of the remaining coal without oxygen. Char is gasified with CO2 produced in the first stage through reaction C + CO2 → CO and with H2O through reaction C + H2O → CO + H2.
Pressure: 24 atm No slip condition at wall Adiabatic walls Inlet turbulence intensity 10%
2nd stage Coal: 5.47 kg/s, 300 K
1st stage Oxidant: 9.92 kg/s, 425 K Coal: 16.43 kg/s, 300K
Fig. 3 Boundary conditions of the baseline case of the two-stage entrained-flow gasifier. Table 3 Parameter and operating conditions of the studied cases. The changed parameters are highlighted.
Parameters Type Fuel Oxidant Stage Distribution Fuel Oxidant Mass flow rate Fuel (kg/s) Oxidant (kg/s)
1
Case 1
Case 2
Case 3
Case 4
Case 5
Slurry Oxygen
Dry Oxygen
Slurry Oxygen
Slurry Oxygen
Slurry Air
2
75% 100%
25% 0%
16.43 9.92
5.47 0.00
Total
1 75% 100%
21.9 9.92
9.47 9.92
2
Total
1 50% 100%
25% 0%
2
Total
50% 0%
3.16 12.90 10.95 10.95 21.90 0.00 9.92 9.92 0.00 9.92
* Oxidant for Case 5 is air (78% N2, 22% O2).
5
1
2
Total
1
100% 100%
0% 0%
21.90 9.92
0.00 21.90 8.12 0.00 9.92 21.00
75% 100%
2
Total
25% 0% 2.70 10.82 0.00 21.00
Fig. 4 Gas temperature and species mole fraction distributions for Case 1 (2-stage, 75%-25%, coal slurry, oxygen-blown).
Mass-weighted averages of gas temperature and species mole fractions along the gasifier height for Case 1 are shown in Fig. 5. The dips in the graphs occur at the injector elevations at heights of 0.75 m for the first stage and 3 m for the second stage. The CO2 mole fraction and the gas temperature decrease from roughly 27% to roughly 19% as the gas flows from the first stage to the second stage. On the other hand, CO mole fraction increases from 12% to 20%, due to the endothermic char-CO2 (R1.2) gasification mentioned above. Meanwhile, the increase in the average mole fraction of H2 in the second stage is negligible. This may indicate that char-CO2 gasification is more dominant than char-H2O gasification in the second stage. 0.4
1st stage fuel injection 2nd stage fuel injection
significant change in the syngas temperature and compositions. The significant temperature drop from roughly 1900K (2960°F) to 1500K (2240°F) across the second stage clearly indicates the advantage of injecting only coal at the second stage to protect the refractory liner and reduce the maintenance cost. Fig. 6a shows helical flow pathlines inside the gasifier; the helical pattern lengthens the flow residence time to allow more time for the reactions to take place. Velocity vectors on vertical midplane and horizontal injection levels are presented in Fig. 7. Due to the vortex generated by the tangential fuel injections in the first stage, strong upward flow occurs near the wall, and weak downward flow occurs in the center. The central core near the second stage exhibits an almost stagnant region due to the opposing fuel injections at the second stage. The flow below the first stage injection level is weak, which could result in some gas being trapped. The momentum of each jet in the second stage is not strong enough to reach the center, and hence the jets are swept upward by the strong main flow from the first stage. Figure 6b shows the coal particle distribution. The particles injected in the first stage are depleted fairly quickly, while those injected in the second stage are depleted at a slow rate.
2000
Gas mole fraction
0.3 H2O
CO
1800
1600
0.2 CO2 1400
Gas temperature (K)
H2
0.1 T O2
Exit syngas temperature and mole fraction compositions are listed in Table 4. Carbon conversion efficiency is 99.4%, which is the comparison of the total mass of carbon injected into the gasifier to the total mass of carbon at the gasifier exit. The high heating value (HHV) of the exit syngas is 8.24 MJ/kg.
1200
CH4
0
1000 0
1
2
3
4
5
6
7
8
9
Gasifier height (m)
Fig. 5 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Case 1 (2-stage, 75%-25%, coal slurry, oxygen-blown).
At the gasifier height of 8.5 m, the graphs for the average gas temperature and gas mole fractions flatten out. This indicates that the rates of reactions are slowing down. Making the gasifier longer or higher would probably not make
6
The distribution of gas temperature presented in Fig. 8 shows that the local highest temperature in the first stage is approximately 3200 K (5300°F), which is 800 K (1440°F) higher than the coal slurry case (Case 1). Unlike the coal slurry case, the dry coal case does not have a lot of H2O to absorb the heat released by the char combustion, nor does much water react with char through the char-H2O gasification. H2O presented in Case 2 comes from the moisture contained in the coal, while H2O in Case 1 comes from both the moisture contained in the coal and water added to the coal to make the slurry. This higher gas temperature means that the fuel injectors and refractory liner in the first stage will experience more severe thermal loading and maintenance issues than in the coal slurry operation. As seen in Fig. 9, the average CO mole fraction in the first stage is slightly higher than in the coal slurry case (Case 1), approximately 19% versus 12%. The same is observed for the average CO2 and H2 mole fractions, 30% for CO2 and 34% for H2 in the dry coal case compared to 27% for CO2 and 31% for H2 in the coal slurry case.
(a)
Similar to the coal slurry operation in Case 1, char gasification is enhanced in the second stage by injecting the remaining fresh coal. But because the coal injected is dry coal, char gasification that occurs is mainly char-CO2 gasification.
(b)
Fig. 6 (a) Flow pathline colored by the residence time temperature and (b) particle distribution for Case 1. Table 4 Exit syngas temperatures and compositions. Case 1 Fuel distribution Oxidant Fuel type Exit syngas: T (K) T (°F)
Case 2
Case 3
Case 4
Case 5
2-stage 2-stage 2-stage (75%-25%) (75%-25%) (50%-50%)
1-stage
2-stage (75%-25%)
oxygen slurry
oxygen slurry
air slurry
oxygen dry
oxygen slurry
1310 1898
1882 2928
1250 1790
1407 2073
1143 1598
31.7% 20.2% 18.9% 1.2% 26.7% 1.3% 0.0%
33.8% 31.4% 19.0% 1.7% 0.8% 13.3% 0.0%
31.1% 19.7% 19.2% 1.3% 27.4% 1.3% 0.0%
32.2% 21.5% 18.0% 0.7% 26.3% 1.3% 0.0%
19.0% 7.6% 12.5% 0.4% 16.4% 44.1% 0.0%
99.4% 8.24
100.0% 9.45
98.3% 9.03
94.8% 7.68
77.3% 4.40
Mole fraction: H2 CO CO2 CH4 H2O N2 O2 Carbon conversion efficiency HHV at 25°C (MJ/kg)
4.2 Effects of Coal Mixture (Slurry vs. Dry) Case 2 is conducted to investigate the effects of using dry coal as the fuel. Coal and oxidant feed rates are kept the same as for Case 1. Nitrogen is used as the transport gas for the coal powder. The amount of N2 transport gas used is 0.3 times the mass of coal powder. The same fuel and oxidant distributions as in Case 1 are used, which is two-stage operation with 75%25% fuel distribution between the first and second stages and 100% oxidant injected into the first stage with no oxidant injection at the second stage.
Fig. 7 Velocity vectors and temperature field on the center vertical plane and injection planes for Case 1.
7
Fig. 8 Gas temperature and species mole fraction distributions for Case 2 (2-stage, 75%-25%, dry coal, oxygen-blown).
respectively. The syngas HHV of the dry coal case is also higher than the coal slurry case, 9.45 MJ/kg versus 8.24 MJ/kg. Of course, a higher heating value is better. However, in addition to potential increased maintenance issue related to fuel injectors and refractory liner, the higher syngas temperature of the dry coal case means that thermal efficiency will reduce when the syngas temperature is cooled down to the acceptable level for operating the downstream gas clean-up system. Although syngas cooler can transfer the thermal energy of the high raw syngas temperature to high-pressure steam, degradation of the energy quality will inevitably affect the overall plant thermal efficiency.
3000 1st stage fuel injection 2nd stage fuel injection
0.4
0.3
CO
2500
0.2 CO2 2000 0.1
Gas temperature (K)
Gas mole fraction
H2
T O2
4.3 Effects of Fuel Distribution
CH4
H2O 0
1500 0
1
2
3
4
5
6
7
8
9
In the baseline case (Case 1), fuel is distributed by 75%25% between the first and the second stages. Cases 3 & 4 are conducted to study the effects of different fuel distributions. In Case 3, the fuel is evenly distributed between the first and the second stages, i.e. 50%-50%. In Case 4, all (100%) of the fuel is injected in the first stage. In other words, Case 4 simulates the one-stage operation of the gasifier. The same total feed rate of coal slurry and oxidant in Case 1 is used in Cases 3 & 4. As in Case 1, all of the oxidant is injected in the first stage.
Gasifier height (m)
Fig. 9 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Case 2 (2-stage, 75%-25%, dry coal, oxygen-blown).
Both Figs. 8 and 9 show a significant increase in CO (from approximately 19% to 31%) and decrease in CO2 (from approximately 30% to 19%) in the second stage, due to the charCO2 gasification. Minor char-H2O reaction also occurs in the second stage. The small decrease in H2 in the second stage is due to dilution by the additional coal.
Fig. 10 presents the comparison of average gas temperature and species mole fractions for Cases 1, 3, and 4. Higher massweighted average gas temperature 2500 K (4040°F) occurs in the first stage for Case 3 (50%-50%) compared to 1900 K (2960°F) of Case 1 (75%-25%) and is due to the higher O2/char ratio in the first stage for Case 3. Higher O2/char causes more char to burn, resulting in a higher average gas temperature. However, counter-intuitively, lower O2/char ratio in Case 4 (100%-0%) in the 1st stage also produces higher average gas temperature than Case 1. A plausible explanation would be that the higher temperature in Case 3 is not actually caused by rich combustion as first thought, but it is caused by less water
The average temperature of the exit syngas listed in Table 4 is 1882 K (2928°F), which is 572 K (1030°F) higher than the syngas for the coal slurry case (Case 1), due to lack of steam in the dry coal operation. Compared to the coal slurry case, there is less H2O to absorb the heat from the char combustion and less H2O to react with C through the endothermic char-H2O reaction. H2 and CO2 contents of the syngas are higher than those of the coal slurry case, 33.8% and 31.4% versus 31.7% and 20.2%,
8
turbine combustor without going through the gas cleanup. In reality, the sensible heat will be used to produce steam to produce power through the steam turbine because the syngas temperature needs to be reduced for cleaning and desulfurization.
presence, and hence, less heat capacity to absorb heat generated by combustion. This explanation can be supported by the high oxygen and CO2 concentrations but low CO and H2 concentration in the first stage of Case 3 shown in Fig. 10. This means combustion in the 1st stage in Case 3 is complete (i.e. high CO2) but the gasification process is less productive (i.e. low CO and H2). On the other hand, in Case 4 when 100% coal is injected in the 1st stage, oxygen is quickly consumed (i.e. low O2) to produce CO with high temperature. The relatively lower average gas temperature in the injector area for Case 1 (75%25%) has the benefit of helping prolong the life of fuel injectors and refractory liners.
Based on the syngas temperature and composition, the 50%-50% fuel distribution (Case 3) gives the best result. It has the highest syngas HHV (9.03 MJ/kg) even though its carbon conversion efficiency (98.3%) is slightly lower than that of the 75%-25% case (Case 1 with carbon conversion efficiency of 99.4%). Besides the highest syngas HHV, Case 3 has the lowest syngas temperature (1250 K, 1790°F). This lowest syngas temperature compared to the other cases means that there will be less energy loss during the syngas clean-up process. However, its mass-weighted average of gas temperature (2500 K, 4040°F) in the first stage is highest compared to those of the other cases, 1900 K (2960°F) for Case 1 and 1500 K (2240°F) for Case 4. This high gas temperature will put the gasifier's fuel injectors and wall refractory bricks in a higher thermal loading; consequently, they will be more prone to failure and require more maintenance.
The graph of O2 mole fraction for Case 3 shows that a little amount of O2 still exists when the gas reaches the second stage injection level. This means that char has a good opportunity to react with the abundant O2 at the first stage. Meanwhile, for Case 1 (75%-25%) and Case 4 (100%-0%), O2 is quickly completely depleted in the first stage. The comparison of CO and CO2 mole fractions for all three cases confirms that char combustion is more intense in Case 3. Case 3 has the highest CO2 mole fraction and the lowest CO mole fraction in the first stage. It implies that a large amount of char in the first stage goes through complete combustion. Case 4 (100%-0%), which has the lowest O2/char ratio in the first stage, has the lowest CO2 mole fraction and the highest CO mole fraction.
Velocity vectors on vertical midplane and horizontal injection levels for Case 3 are presented in Fig. 10. With 50% of the fuel injected in the second stage, the fuel jests are stronger than in Case 1 (Fig. 7) and are able to penetrate deeper to the center crossing the upcoming flow from the first stage.
The exit syngas composition listed in Table 4 indicates that among the three cases, Case 4 (100%-0%) yields the highest H2 production – 32.2% compared to 31.7% for Case 1 (75%-25%) and 31.1% for Case 3 (50%-50%). Case 4 also has the highest CO production – 21.5% compared to 20.2% for Case 1 and 19.7% for Case 3. However, Case 4 has the highest exit syngas temperature at 1407 K. Syngas temperature for Cases 1 and 3 are 1310 K and 1250 K, respectively.
4.4 Effects of Oxidant (Oxygen-Blown vs. Air-Blown) Case 5 simulates the air-blown two-stage operation of the gasifier. Air with composition of 22% O2 and 78% N2 by weight is used as the oxidant. The O2/C mole ratio is maintained the same as in Case 1 (oxygen-blown) which is 0.4. Total feed rate of coal and oxidant combined is the same as for Case 1. Similar to Case 1, the fuel is distributed 75% and 25% between the first and the second stages.
Even though Case 4 has the highest H2, CO and CH4 combined, its syngas high heating value is the lowest among three cases. Case 4’s HHV is 7.68 MJ/kg, compared to 8.24 MJ/kg for Case 1 and 9.03 MJ/kg for Case 3. This is due to the lower carbon conversion efficiency of Case 4 (94.8%) compared to the other two cases (99.4% for Case 1 and 98.3% for Case 3). The exit syngas of Case 4 contains the most unreacted char. Thus, combined with its high temperature, it has the lowest HHV. Note that when the syngas exit temperature is high more chemical energy has been converted to the sensible heat of the syngas and less chemical energy is reserved in the syngas. This sensible heat could be effectively used in the gas turbine combustor if the syngas could be fed directly into the gas
As expected, the mass-weighted average of gas temperature in the first stage shown in Fig. 12 is lower than in Case 1 (oxygen-blown) due to the abundance of N2 as a diluent in the air-blown case. The maximum cross-sectional mass weighted average gas temperature is approximately 1450 K (2150°F), while the maximum average gas temperature in the oxygenblown case is 2000 K (3140°F).
9
0.4
1st stage fuel injection
2900
1st stage fuel injection 2nd stage fuel injection
2400
H2 mole fraction
Gas temperature (K)
2nd stage fuel injection
1900
1400 Case 3 (50%-50%)
900
0.3
Case 3 (50%-50%)
0.2
Case 1 (75%-25%)
Case 1 (75%-25%)
Case 4 (100%-0%)
Case 4 (100%-0%)
400
0.1
0
1
2
3
4
5
6
7
8
9
0
1
2
Gasifier height (m) 0.3
3
4
0.4
1st stage fuel injection
7
8
9
CO2 mole fraction
2nd stage fuel injection
0.2
0.1 Case 3 (50%-50%) Case 1 (75%-25%)
0.3
0.2
Case 3 (50%-50%)
0.1
Case 1 (75%-25%)
Case 4 (100%-0%)
Case 4 (100%-0%)
0
0
0
1
2
3
4
5
6
7
8
9
0
1
2
0.15
1st stage fuel injection
O2 mole fraction
0.3
Case 3 (50%-50%)
0.2
4
5
6
7
8
9
1st stage fuel injection 2nd stage fuel injection
2nd stage fuel injection
0.4
3
Gasifier height (m)
Gasifier height (m)
H2O mole fraction
6
1st stage fuel injection
2nd stage fuel injection
CO mole fraction
5
Gasifier height (m)
0.1
0.05
Case 3 (50%-50%) Case 1 (75%-25%)
Case 1 (75%-25%) Case 4 (100%-0%)
Case 4 (100%-0%)
0
0.1 0
1
2
3
4
5
6
7
8
0
9
1
2
3
4
5
6
7
8
9
Gasifier height (m)
Gasifier height (m)
Fig. 10 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Cases 1, 3 and 4. 0.6
1500
1st stage fuel injection 2nd stage fuel injection
0.5
Gas mole fraction
0.4 1300 0.3 1200
T
H2O
0.2
H2
1100
CO2
0.1
Gas temperature (K)
1400
N2
CO
O2
CH4
0
1000 0
1
2
3
4
5
6
7
8
9
Gasifier height (m)
Fig. 12 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Case 5 (2-stage, 75%-25%, coal slurry, air-blown). Fig. 11 Velocity vectors and temperature field on the center vertical plane and injection planes for Case 3 (50%-50%).
10
through the char-H2O gasification. This higher gas temperature means that the fuel injectors and refractory walls in the first stage will experience higher thermal loading than in the coal slurry operation. The syngas HHV of the dry coal case is also higher than the coal slurry case -- 9.45 MJ/kg vs. 8.24 MJ/kg. However, the higher syngas temperature of the dry coal case would result in a lower plant thermal efficiency because it needs to be cooled before it goes through the gas clean-up system downstream of the gasifier. Consequently, a lot of energy will be downgraded (i.e. loss of exergy) via waste heat exchanger even though part of the energy can be recovered to produce superheated steam to generate electricity through the steam turbine.
Table 5 Comparison of exit syngas temperature and composition between Cases 1 and 5 after N2 is removed from the syngas.
Fuel distribution Oxidant Fuel type Exit syngas: T (K)
Case 1
Case 5
2-stage (75%-25%) oxygen slurry
2-stage (75%-25%) air slurry
1310
1143
H2 CO CO2 CH4 H2O
32.1% 20.5% 19.1% 1.2% 27.1%
34.0% 13.6% 22.4% 0.7% 29.3%
O2 Carbon conversion efficiency HHV at 25°C (MJ/kg)
0.0%
0.0%
99.4% 8.25
77.3% 7.26
Mole fraction:
Effects of Fuel Distribution between Two Stages Due to less water to absorb heat, reducing the fuel feed in the first stage does result in higher gas temperatures in the first stage. One-stage operation yields higher H2, CO and CH4 combined than if a two-stage operation is used but with a lower syngas heating value. The 50%-50% fuel distribution case yields the highest syngas HHV and lowest syngas exit temperature among the studied cases. The exit syngas of onestage operation contains the most unreacted char, combined with its high exit temperature, results in the lowest heating value.
The syngas composition listed in Table 4 shows that the mole fraction ratio of CO/H2 is 0.4 for the air-blown case (Case 5), which is much lower than those of the oxygen-blown case (Case 1). The syngas HHV for Case 5 is approximately only half of Case 1, 4.40 MJ/kg vs. 8.24 MJ/kg. The syngas of Case 5 is diluted with N2, which causes this low heating value. However, its low carbon conversion efficiency at 77.3% also contributes to this low syngas heating value. Low carbon conversion efficiency is due to the lower overall gas temperature inside the gasifier, where less energy is available to drive the endothermic gasification reactions.
Effects of Oxidant (Oxygen-Blown vs. Air-Blown) Gas temperature inside the gasifier for the air-blown case is lower than in the oxygen-blown gasifier due to the abundant presence of N2. Lower than the oxygen-blown case (99.4%), the carbon conversion efficiency of the air-blown case is 77.3%. The syngas heating value for the air-blown case is 4.40 MJ/kg, which is almost half of the heating value of the oxygen-blown case (8.24 MJ/kg). Even when N2 is removed for comparison, the HHV of the air-blown case is still about 1 MJ/kg less than the oxygen-blown case.
To give a fair comparison between the syngas in Cases 1 and 5, syngas compositions and heating values for both cases are recalculated after the N2 contained in the syngas are removed. The recalculated compositions are compared in Table 5. The mole fraction of H2 (34.0%) for the air-blown case (Case 5) becomes slightly higher than the oxygen-blown (Case 1, 32.1%), but the CO mole fraction for the air-blown (13.6%) is 6.5 percentage points lower than the oxygen-blown case. As expected, the heating value of the syngas increases from 4.40 MJ/kg to 7.26 MJ/kg after N2 is removed. Nonetheless, this recalculated syngas heating value is still lower by roughly 1 MJ/kg than that of the oxygen-blown case (8.25 MJ/kg) even after N2 is removed.
6.0 ACKNOWLEDGEMENTS This study was supported by the Department of Energy contract No. DE-FC26-08NT01922 and the Louisiana Governor's Energy Initiative via the Clean Power and Energy Research Consortium (CPERC), administered by the Louisiana Board of Regents. 7.0 REFERENCES
5.0 CONCLUSIONS
Bockelie, M.J., Denison, K.K., Chen, Z., Linjewile, T., Senior, C.L., and Sarofim, A.F., CFD Modeling For Entrained Flow Gasifiers in Vision 21 Systems, Proceedings of the 19th Annual International Pittsburgh Coal Conference, Pittsburgh, PA. September 24-26, 2002.
Five cases of different operating conditions are simulated and the results show: Effects of Coal Mixture (Slurry vs. Dry)
Chen, C., Horio, M., and Kojima, T., Numerical Simulation of Entrained Flow Coal Gasifiers, Chemical Engineering Science, 55, 3861-3833, 2000.
The temperature in the first stage for the dry-fed case is approximately 2800 K (4580°F), which is 400 K (720°F) higher than the slurry-fed case. Unlike the slurry-fed case, the dry-fed case does not have a lot of H2O to absorb the heat released by the char combustion, nor does much steam react with char
Fletcher, T.H., and Kerstein, A.R., Pugmire, R.J., Grant, D.M., Chemical Percolation Model for Devolatilization: 2.
11
Temperature and Heating Rate Effects on Product Yields, Energy and Fuels, 4, 54, 1990. Fletcher, T.H., and Kerstein, A.R., Chemical Percolation Model for Devolatilization: 3. Direct Use of 13C NMR Date to Predict Effects of Coal Type, Energy and Fuels, 6, 414, 1992. Grant, D.M., Pugmire, R.J., Fletcher, T.H., and Kerstein, A.R., Chemical Percolation of Coal Devolatilization Using Percolation Lattice Statistics, Energy and Fuels, 3, 175, 1989. Guenther, C., and Zitney, S.E., Gasification CFD Modeling for Advanced Power Plant Simulation, Proceedings of the 22th International Pittsburgh Coal Conference, Pittsburgh, Pennsylvania, September 12-15, 2005. Jones, W.P., and Lindstedt, R.P., Global Reaction Schemes for Hydrocarbon Combustion, Combustion and Flame,73,233,1998. Kobayashi, H., Howard, J. B., and Sarofim, A. F., Coal Devolatilization at High Temperatures, 16th Symp. (Int'l.) on Combustion, The Combustion Institute, 1976. Silaen, A. and Wang, T., Effects of Fuel Injection Angles on Performance of A Two-Stage Coal Gasifier, Proceedings of the 23rd Pittsburgh Coal Conference, Pittsburgh, Pennsylvania, Sept. 25-28, 2006. Silaen, A. and Wang, T., Comparison of Instantaneous, Equilibrium and Finite Rate Gasification Models in an Entrained Flow Coal Gasifier, Paper 14-3, Proceedings of the 26th International Pittsburgh Coal Conference, Pittsburgh, Pennsylvania, Sept. 21-24, 2009. Silaen, A. and Wang, T., Effect of Turbulence and Devolatilization Models on Gasification Simulation, International Journal of Heat and Mass Transfer, 53, 2074-2091, 2010. Smoot, D.L., and Smith, P.J., Coal Combustion and Gasification, Plenum Press, 1985.
12
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: GASIFICATION: GASIFICATION SCIENCE AND MODELING EXACT TITLE OF PAPER: START-UP BEHAVIOR OF A FIXED BED GASIFIER: ONE DIMENSIONAL MODELING Giampaolo Mura, Prof., University of Cagliari - Department of Chemical Engineering and Material Science [email protected]] Mariarosa Brundu, PhD, University of Cagliari - Department of Chemical Engineering and Material Science [email protected]
Abstract: This work copes with the development of a mathematical model for the investigation of the transient behavior of a countercurrent fixed bed gasifier. The phenomenological model is based on heat and mass continuity equations. Heterogeneousness is somehow considered by the insertion of two separated heat balances, one for the gas and one for the solid phase. All the main phenomena involved in the gasification process are inserted: drying, pyrolysis, gasification and combustion reactions of the solid phase and homogeneous gas phase reactions including secondary pyrolysis reactions. The system is described with a pseudo homogeneous approach. Moisture loss calculation is carried out by the introduction of a first order kinetic on the moisture content of the bed; a competitive reaction model is used for primary pyrolysis; heterogeneousness of the system is considered for gas phase reactions by the introduction of a shrinking core reaction model where the external diffusion and the kinetic resistance are considered. The solid phase is constituted by four pseudo components: coal, ash, char and moisture. The ash behavior is described by the introduction of a shell progressive model with variable particle diameter. Gas species considered by the model are: CO, CO2, H2, H2O, CH4 and tar. Input for the model are flowrates, temperature and composition profiles at the initial state for both: the gas and the solid phase. The model was used to study the start up of an air blown atmospheric gasifier in the case of a Pittsburg n°8 coal seam feedstock. The initial conditions chosen for the dynamic simulation are in accordance with the start up procedure of an existing gasification pilot plant. Output for the model are the variation with time of temperature, composition and spatial velocity profiles of the system. In particular the dynamics is analyzed with reference to the bed behaviour in a long time investigation. Scarce information about this topic was before present in the literature. The influence of steam injection also revealed the presence of multiple steady states for this system.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11–14, 2010 PROGRAM TOPIC: GASIFICATION
Entrained Flow Coal Gasification: Modeling, Simulation & Experimental Uncertainty Quantification For a Laboratory Reactor Institute for Clean and Secure Energy, The University of Utah Philip J. Smith, professor & director ([email protected]) Charles Reid, graduate student ([email protected]) Julen Pedel, graduate student ([email protected]) Jeremy Thornock, asst. research professor ([email protected])
Abstract: Modeling and simulation on petascale computing platforms offers unprecedented opportunities to explore new transformative technology options in entrained-flow coal gasification. However, before these advanced simulation tools can be used with confidence, formal validation and uncertainty quantification is required. In this paper we explore the uncertainty in predictions from large eddy simulations (LES) of the Brigham Young University (BYU) pilot-scale entrained flow reactor1 . The specific focus of this paper is to produce predictive capability for gasification with quantified uncertainty bounds through a formal validation and uncertainty quantification (V/UQ) analysis. We have employed the Data Collaboration methods of Michael Frenklach and coworkers at the University of California-Berkeley in our V/UQ analysis 2,3. Data Collaboration requires consistency between simulation results and experimental data. Having developed the ARCHES code, that combines LES with the Direct Quadrature Method of Moments (DQMOM) for the gasification simulation, we now perform V/UQ employing the BYU experimental data. The simulation produces temporally and spatially resolved data of the reacting, multiphase flow field, including the moment description and evolution of the full particle number density function. Specifically, we have studied three key particle behaviors; particle size segregation (Stokes number effects), particle clustering, and particle pyrolysis. We perform the V/UQ analysis with the spatially resolved compositional data measured in the BYU gasifier. In this study we have extracted experimental error for each of the measurements taken. Used prior ARCHES validation information to produce prior uncertainty bonds on the most sensitive simulation parameters. We have included uncertainties in numerical parameters, models, and scenario parameters. The resulting V/UQ analysis produced posterior uncertainty bounds on both the quantities of interest and the uncertain parameter space studied.
Introduction: Utilization of domestic sources of fuel such as coal is growing increasingly important, but under increasing scrutiny. There is a critical need to retrofit existing industrial and applied-scale facilities that use coal to meet regulatory requirements, and to explore new technologies and techniques for carbon emission reduction. With the continually increasing power of computational resources, predictive simulation tools are playing an increasingly important role in this process. As computing moves from peta- to exa-scale, simulation codes and models that can utilize these resources are needed. The large-eddy simulation tool Arches is one such simulation tool. In the current work, a dispersed-phase model is implemented in Arches, and preliminary results are shown. A procedure for implementing the direct quadrature method of moments in a large-eddy simulation code, and applying the method to a complex, multiphysics problem (coal gasification), has been developed. Simulations of the gasifier of Brown et al1 are performed. The gasifier is , and simulations are performed using a grid, with a resolution of . A two-step Kobayashi-Sarofim devolatilization model4 is used, with rate constants given by them. To model the particle-laden gasifier, the direct quadrature method of moments is coupled with large eddy simulation. Recent developments in the direct quadrature method of moments have resolved issues encountered with DQMOM when incorporating complex gasification physics. The DQMOM formulation is explained, the technique used to address the numerical issues in DQMOM is explained, and simulation results are compared to results from Brown et al.
Large Eddy Simulations: Simulation of turbulent flows, particularly large-scale and high Reynolds number flows, is a difficult and expensive problem. The cost of computations in which all length and time scales are resolved scales as Re11/4. In addition, that cost increases for complex geometry. Resolving all scales of the turbulent flow in a direct numerical simulation (DNS) has not yet been attempted for these multiphysics problems. In order to move computations to the scale of applied and industrial systems, it is necessary to model some or all scales of the flow. Most coal combustion simulations have resorted to a Reynolds-Averaged Navier Stokes (RANS) analysis to circumvent the massive computational burden. In contrast to RANS, the large eddy simulation (LES) technique resolves the largest scales of the flow up to a cutoff length ∆, but models the smallest scales of the flow. This is done by applying a low-pass filter to the velocity field to reduce high-frequency processes, allowing for a balance between computational cost and increased accuracy. The smallest (filtered) scales must then be modeled. However, the modeling is simplified over RANS turbulence models because, as stated by the Kolmogorov Hypothesis, at high enough wavenumbers, the turbulence becomes statistically homogeneous, meaning the turbulence at these scales can be modeled without introducing significant modeling error. Almost 99% of the computational cost of DNS is spent on resolving the smallest 1% of flow scales, so LES can substantially reduce computational costs by eliminating the cost of resolving turbulence at small scales and using a simplified model instead. For complex multiphysics problems like coal gasification, there are a wide range of physical processes happening. Each of these have a characteristic length (l) or wavenumber (k). If the wavenumber associated with a particular physical process is greater than the Nyquist limit, kN = 1/∆, then it falls in the resolved spectra. Thus, it is essentially DNS, for that process, on the LES grid. In the case of chemical reactions, if the Damköhler number is small enough, the reactions are sufficiently slow to be completely resolved. Likewise, particle motion can be completely resolved if the particles are ballistic, and the Stokes number of the particles is large enough, because small (unresolved) fluctuations in velocity will have no affect on particle momentum. Our analysis shows that for the entrained flow gasification cases we have studied, the Stokes Number of the essentially all of the pertinent coal particles and the Damköhler number of the coal particle devolatilization and char oxidation processes are all above the Nyquist limit for the simulations that we can perform on reasonable computer clusters. Thus, the coal process are fully resolved (ie. essentially DNS). Of course, there are still gas phase turbulent mixing and reaction processes that are not resolved and must be modeled in traditional LES
fashion. For this reason, large eddy simulation holds great promise for pulverized coal combustion and entrained-flow gasification applications with multiple, coupled physical processes.
Coal Particle Approach: � � jet,� � Nξ Nξ 3 Number Density Function. 3 In the distribution of particles can∂ be�described � � � ∂f ∂ ∂ a particle-laden ∂f ∂ ∂f using a number + density((NDF), f )− denoted D Gj,f >and f )+ +h(ξ, x, t)The (4) f(x,t), a function(< of space time and hasΓunits of [#/volume]. xi which=is− ξj ∂t i=1NDF ∂xiis a function of a i=1 ∂x ∂x ∂ξ ∂ξ ∂ξ i i j j j set of independent variables j=1for particles. This vector j=1of internal coordinates, denoted ξ;
may include temperature, composition, or size of particles. In this way the NDF can describe the distribution of
whichthe represents the spatial and with temporal evolution of ξ.the NDF,density including all particle number density of particles particular values of Thethe number is written as f(ξ; x,phenomena t): drag force, particle size change, breakage, etc.) Nξ � �∂f (ξ;� �� Nξ Nξ 3 x, t) ∂ ∂� � � ∂ ∂f ∂+ ∂ ∂f (�u (ξ; x, t))+ Γξ (�G (ξ;(4) x, t)) = h (ξ; x, t) (1) i |ξ�f j |ξ�f Dxi = −∂t (< f )+ +h(ξ, x, t) j,f > j ∂ξ Quadrature Approximation: Direct Quadrature Method of Moments ∂xG i j ∂x ∂x ∂ξ ∂ξ ∂ξ i i j j j j=1 =1 j=1 j=1
n the Direct Quadrature Method NDF is approximated by � � the � � a sum � of delta functions, as dx N Niξ ξ 3 3 of Moments, � � � is the ξ-space convection term j = dξ j /dt. Each is∂term conditioned on the ,value ξ because the values of j∂f ial and Gtemporal the the including all particle phenomena ∂evolution ∂ GNDF, ∂f ∂ andofG ∂f where ui>isfof)− the x-space convection ui >=fof thex, ξ-space j is +model (< u D = − (< G )+ Γ +h(ξ, existing t) (4) Lagrangian hown onthese Figure 1. The integral of the NDF becomes a summation a set of N weighted abscissas: i,fchanges for different xi j,f values ξj dt terms particles and different of ξ. Because many ∂t etc.) ∂xi ∂xi ∂x ∂ξj ∂ξj ∂ξj ge, breakage, i=1 j=1 j=1 dξji derivatives, modelsi=1 give expressions for these temporal existing Lagrangian models can be “plugged in” to the convection term Gjˆ = . Each � is N conditioned on the value of ξ because NDFwhich transport equation framework, anddttherefore can Eulerian well. represents the spatial and temporal evolution of be theextended the NDF, to including allmodels particleas phenomena w(x)g(x)dx wα < >α (5) the values of these model terms ≈changes forg different particles and different (drag force, particle size change, breakage, etc.) ximation: Direct Quadrature Method of Moments Direct Quadrature Method of Moments. In the Direct Quadrature Method of Moments, the NDF is approximated α=1 values of ξ. Because many existing Lagrangian models give expressions for these by a sum of delta functions, as shown on temporal derivatives, existing models can beas“plugged in” to the thod Moments, NDF iswith approximated byLagrangian a sum of delta functions, For of aFigure multivaritae NDF, i.e. multiple interQuadrature Approximation: Direct Quadrature Method of Moments 1. The the integral of the NDF becomes a NDF transport equation and therefore can be extended to Eulerian ralcoordinates, of the NDFthe becomes aof summation ofas: aframework, set of N weighted abscissas: summation ofNDF a set N be weighted abscissas: al can written In the Direct Quadrature Method of Moments, the NDF is approximated by a sum of delta functions, as models as well. shown on Figure 1. TheNintegral of the NDF becomes a summation of a set of N weighted abscissas: ˆ � w(x)g(x)dx ≈ w < g > (5) N ˆ ξ N α α N � � 3.2 Eulerian � � � vs. Lagrangian wα (x, t) α=1δ ξj − �ξj (x,w(x)g(x)dx wα < g >α (5) f (ξ; x, t) ≈ t)� ≈ α
α=1
j=1 e. with multiple The inter-most common and widely-used models for the dispersed phase are LaFor For a multivaritae NDF, i.e.i.e. with multiple intera multivaritae NDF, with multiple inter(6)track a large number of characteristic particles models. These models be written as: grangian the NDF can be written as: nal coordinates, the NDF can be written as: wherenal Ncoordinates, is the number of internal coordinates ξ as they move through the domain and interact with the fluid; an ordinary difnd wα is called a ferential weight and represents the form number in the of equation N N associated particles per volume with the abscissa � � � � � � x,t)� t) ≈ wα (x, t) δ ξj − �ξj (x, t)�α .δ ξj − �ξf j(ξ; (x, α α=1 dx DQMOM solves transportj=1equations for the = Sx (2) (6) dt where Nξ is number of internal coordinates eights and abscissas of the quadrature approxima(6) where Nξthe is the number of internal coordinates and w is a weight and represents the number α on. These transport equations are simply scalar dξ = Sξ andcoordinates wα is called a weight and represents the number of internal (3) ofofequations, particles per volume volume associated with the dt particles per associated with the abscissa ansport and are easily implemented represents the number abscissa α. code.α. The transport equations for the to a with CFD is solved for each particle, and information from all particles is then collapsed ated the abscissa DQMOM solves transport equations for the eights w weighted abscissas ςj,α = w FigureThese 1: DQMOM approximation NDF with 3 DQMOM solves transport for the onto the Eulerian grid to yield the methods are describedofinthe greater α and and α �ξj � α NDF. weights abscissas of the equations quadrature approximaweights and abscissas of the quadrature an be expressed as: quadrature nodes rt equations fordetail the in [11, 10]. are simply scalar tion. These transport equations approximation. Theseand transport equations are transport equations, are easily implemented uadrature approximaAlternatively, Eulerian methods are based on solving scalar transport equa∂ ∂ simply scalarcode. transport equations, and areforeasily into a CFD The transport equations the ons are simply scalar for (wwαtions ) +intoweighted (ustatistical ) = ainformation (7)(moments) about the solid phase. Moments of i,f w α implemented the LES code. αςj,α = wα �ξj �α Figure 1: DQMOM approximation of the NDF with 3 weights abscissas α and ∂t ∂x i e easilycan implemented an NDF be expressed as: are defined as quadrature nodes port equations for the ∂ ∂ ´ ∞ ∂ ∂ (u(7) , jthe = t) 1NDF .dξ . . Nwith (8) i,f ςjα ) = ξ scissas ςj,α = wα �ξ �αα ) +Figure ξ kbjα f of (ξ; x, 1:w(ς DQMOM 3 (ui,f )jα =)a+ j(w α α ∂x approximation −∞ ∂t i ∂t ∂xi (4) quadrature nodes mk = ´ ∞ f (ξ; x, t) dξ where ui,f is the exact velocity of∂ the weight w or weighted abscissa ς . The source terms a −∞ α α α and bj,α ∂ (ς ) + (u ς ) = b , j = 1 . . . N (8) jα i,f jα jα ξ re obtained by solving a linear system For more details on the full DQMOM derivation, see ∂t of equations. ∂x (unless (7) k = 0, in which casei there is no denominator) and denote the expected f wα ) = aα ]. where uvalue velocity of the weight abscissa ςαand . The source terms aα and bj,α of exact ξ over the entire NDF.wThe most common least expensive Eulerian i,f is the α or weighted The LES simulations of particle-laden coaxial jet presented in this work were done using four internal are obtained by solving a linear system of equations. For more details on the full DQMOM derivation, see moment method is the multi-fluid model [5, 14], which treats the solid phase ∂ ∂ [3]. oordinates: the particle d, pjand the three velocity components the particle velocity: upx , upy , upz ., as(u ai,fseparate fluid. method solves transportofequations for the number (ςjα )The + LES ςdiameter = 1This ...N (8) jα ) =ofbjα ξ simulations particle-laden coaxial jet presented in this work were done using four internal ∂xi distribution, nd a∂tmondisperse i.e. one quadrature node. th st α=1
ξ
nto a CFD code. The transport∂equations ∂for the (ςjα ) + α �ξ(u (8) i,f ςjα ) = bjα , j = 1 . . . Nξ weights wα and weighted abscissas ∂t ςj,α = w ∂xi j �α Figure 1: DQMOM approximation of the NDF with 3 can be expressed as: quadrature nodes where ui,f is the exact velocity of the weight wα or weighted abscissa ςα . The source terms aα and bj,α re obtained∂by solving∂a linear system of equations. For more details on the full DQMOM derivation, see (wα ) equations + (ufor = aα wα and(7) i,f w α )weights The transport the weighted abscissas ςj,α = wα (ξj)α can be expressed as: 3]. ∂t ∂xi The LES simulations of particle-laden coaxial jet presented in this work were done using four internal oordinates: the particle diameter∂dp(ςand the∂three of the particle velocity: upx , upy , u(8) pz ., (ui,fvelocity ςjα ) = bcomponents jα ) + jα , j = 1 . . . Nξ ∂t ∂x i nd a mondisperse distribution, i.e. one quadrature node. where ui,fui,fisisthe velocity of the weight ςα . The source α or weighted where the exact exact velocity of the weight wα or w weighted abscissaabscissa ςα . The source terms aα andterms bj,α areaα and bj,α are obtained bybysolving a linear linearsystem system of equations. For details more on details the fullderivation, DQMOM obtained solving a of equations. For more the fullon DQMOM seederivation, see Particle Velocity 5 3]. Marchisio and Fox . The LES simulations of particle-laden coaxial jet presented in this work were done using four internal Particle flow, Velocity. a gas-solid flow, theisparticle motion is affected by the drag force, which can be described n a gas-solid theIn particle motion affected by the drag force, which can be described the coordinates: particle diameter dp and the three velocity of the mass particle velocity: uby upy ,Stokes upz ., px , the by the the Stokes drag law. For a meso-scale size particle, whencomponents the other additional forces are omitted, drag law. For a meso-scale size particle, when the other additional mass forces are omitted, the momentum and a mondisperse distribution, i.e. one node. momentum equation for the particle canquadrature be expressed by an ordinary differential equation. quation for the particle can be expressed by an ordinary differential equation. � � Particle Velocity dui,p = � fdrag (ui,g − ui,p ) + gi (ρp − ρg ) + Fi,v (9) dt τ ρ m p p p th here i denotes the i direction, g iis the gravity force acting on the particle, Fv are the other body force n a gas-solid flow, the particle motion is affected by the drag force, which can be described by the Stokes cting on thethe particle, u is the gparticle velocity, and facting is the coefficient ofother dragforces force, which drag edrag i denotes ith direction, isg isthe gravity force the Fthe are other bodyhas for where the pith direction, thewhen gravitythe force acting on theon particle, Fforces bodythe acting v are the law. Fori denotes a meso-scale size particle, other additional massparticle, are the momentum vomitted, ose relationship with particle Reynolds number [6]: on the particle, u is the particle velocity, and f is the coefficient of the drag force, which has a close p drag gequation on the particle, iscan thebe velocity, and fondrag is the coefficient of the drag force, which ha for thetheparticle expressed byforce an acting ordinary differential where irelationship denotes ithup direction, gparticle is the number gravity the particle, Fequation. v are the other body forces 6: with particle Reynolds relationship with particle Reynolds [6]: is the coefficient of the drag force, �number � which has a acting on the particle, up is the particle velocity, and fdrag � 1 Re < 1 du f g (ρ − ρ ) F p i,p drag i p g i,v close relationship with particle Reynolds = number + (9) [6]: (ui,g − u0.687 i,p ) + = dt fdrag ρ m 1 < Re < 1000 i 1τ1p + 0.15Rep p p p Rep < 1 1 0.0183ReRep < 1 Re > 1000 0.687 p p 0.687 f =
1 + 0.15Re < Rep < 1000 fdrag = 1 + 0.15Re drag 1p< Rep <11000 p 0.0183Re ρpRe |u1000 p 0.0183Re Reupg |> 1000 p dpp > p− Rep =
.
ρp dp |up − ug | µg Rep = . p − ug | ρp dp |u Re =µg .
n Equation (9), τp is the particle relaxation ptime
µg
In Equation (9), τp is the particle relaxation time
uation (9), τp is the particle relaxation time ρp d2p τp =
τp . =
ρp d2p . 18µg
(10)
18µg 2 ρp dmomentum where τp is the particle relaxation time The conservation law of the gas phase is, p The momentum momentum conservation law of theof gasthe phase The conservation law gasis, phase is, τp = . ∂ρu
18µg
∂ρu ∂t
+ ∇ · ρu = ∇ · D∇u − ∇p + Sdrag he momentum conservation law phase ∂t of the + ∇gas · ρu = ∇ ·is, D∇u − ∇p + Sdrag
(11)
(10
(
(11
where Sdrag is the momentum exchange between the particles and the gas phase.
∂ρu here Sdrag is the momentum exchange particles and+the gas phase. + ∇between · ρu = the ∇ · D∇u − ∇p Sdrag ∂t is the momentum exchange between the particles and the gas phase.
(
e Modeling Sdrag is the momentum exchange between the particles and the gas phase. where Sdrag
Modeling Arches: Experimental setup
The simulation tool used in this work is ARCHES. ARCHES is a three-dimensional, Large Eddy Simulation In this study, applied to by thepersonnel particle-laden coaxial jet thatfor wasClean used by [7] for his experiments. (LES) LES codeisdeveloped within the Institute andBudilarto Secure Energy within the University of A description of the nozzle is low-Mach given in Figure Particles are conveyed with air in to thesimulate primaryheat nozzle Utah. ARCHES uses a number2a. (M < 0.3), variable density formulation and mass whereas the secondary nozzle is fed with air. Air and particles are injected at room temperature in a 18x18 ninch thischamber. study, LES is applied to theon particle-laden coaxial jeton that usedatbythe Budilarto [7]field for his experiment The experiments focus the effects of velocity ratio thewas flow-field near region region of a coaxial velocity is defined ratio of the to central description of jet. theThe nozzle is ratio, givenwhich in Figure 2a.as the Particles areannular conveyed withinlet airvelocity, in the primary nozz U /U = V R, was varied at three different values: 0, 1.0, 1.5 by varying the volumetric flow rate of the a 0 hereas the secondary nozzle is fed with air. Air and particles are injected at room temperature in a 18x1 annular jet. The Reynolds number based on the average central inlet velocity is 8,400. Table 1 lists the flow s study, LES The is applied to the particle-laden coaxial jet that ratio was used byflow-field Budilartoat[7] fornear his experime nch chamber. experiments focus on the effects of velocity on the the region fiel conditions for the experiments.
Experimental odeling setup
perimental setup
transfer in reacting flows 7. The LES algorithm is composed of the numerical differencing scheme and a solution method for solving the filtered, density-weighted, time-dependent coupled conservation equations for mass, momentum, mixture fraction, coal particle moment equation(s), and enthalpy in a Cartesian coordinate system. This set of filtered equations is discretized in space and time and solved on a staggered, finite volume mesh. The staggering scheme consists of four offset grids, one for storing scalar quantities and three for each component of the velocity vector. An explicit time-stepping scheme is used to advance the simulation forward in time. For the spatial discretization of the LES scalar equations, flux limiting and upwind schemes for the convec- tion operator are used to ensure that scalar values remain bounded. For the momentum equation, a central differencing scheme for the convection operator is used. All diffusion terms are computed with a second-order approximation of the gradient. Overall, ARCHES is second-order accurate in space and time. The ARCHES code is massively parallel and highly scalable due to the integration of ARCHES into the Uintah Computational Framework (UCF)8.
Verification, Validation and Uncertainty Quantification: The process of V&V/UQ is to quantify the uncertainty of a predictive simulation tool by requiring a consistent data set as defined by Feely et al.9 for the intended use of the simulation tool. Here, a data set consists of a set of experimentally observed data and simulation data. Within this data set, uncertainty is defined as ym(x)−ye ≤u where ym is the simulation prediction as a function of input parameters x, ye are the experimentally observed data and u is the total uncertainty. The uncertainty is derived from two sources; uncertainty arising from imprecise physical measurements and uncertainty arising from inputs (x) to the model. Experimental uncertainty is composed of bias errors, reproducibility errors, and instrument error. Each experimental observation, ye,i, has lower (le) and upper (ue) uncertainty bounds. The model inputs, x, are classified as one of three categories; 1) model parameters 2) scenario parameters or 3) numerical parameters. Model parameters are generally physical constants that are determined from a priori knowledge (ie, a reaction rate constant.) Scenario parameters consist of initial and boundary conditions (eg, inlet flow rate.) Numerical parameters relate to the discrete representation of the continuous model (eg, spatial discretization scheme.) Each component of x, like the experimental observations, is also associated with an estimated level of uncertainty consisting of upper (β) and lower bounds (α). Given x with its associated uncertainty, results from the simulation tool maps out a region Rs. Any value within Rs is feasible if it satisfies the criteria le ≤ ym(xi) − ye ≤ ue. Values of xi, such that αi ≤ xi ≤ βi satisfying this criteria are feasible parameters because they lie within a region that has been observed experimentally. In this manner, the experimental observations inform the simulation tool and potentially effect the bounds of x. By the same token, we can also examine when individual observations within ye are consistent with each other through the model. For example, consider two experimental observations, ye,1 and ye,2, contained within ye. The experiments ye,1 and ye,2 are said to be consistent with each other if they can both be described with ym given a set of feasible parameters, x. In this manner, the simulation tool can inform the experiment by explaining two independent experimental observations, thus describing consistency between experiments through the model. The process of discovering the consistent data set with the feasible parameter space has been termed data collaboration10 and reduces to a constrained optimization problem. As with most optimization problems, one must perform function evaluations to determine the optimal parameters. This, however, presents a unique challenge for the size and scope of the coal gasification problem that we wish to address here. That is, each function evaluation of our simulation tool ARCHES requires a substantial amount of computational resources. Consequently, we utilize methods to minimize the total function evaluations by choosing them in optimum locations within x and using presumed function descriptions (surrogates) to describe ym(x) . In this manner we attempt to perform as few function evaluations as necessary.
Results: Single Phase Coaxial Jet. In this study, LES is applied to the non-reacting particle-laden coaxial jet that was used by Budilarto11 for his experiments. Simulations were performed for the three velocity ratios cases without particle. The two main objectives are first to validate our modeling approach by ensuring that mean velocities are predicted correctly and secondly to have a reference case to then understand the effect of particles on the gas dynamics. Figure 2(a) (b) (c) are instantaneous profiles of the x-component of the gas velocity. Figure 2(d) compares the simulation results to the experimental data for the mean gas velocity along the centerline.
(a) Gas x-velocity profile (m/s) - VR = 0
(b) Gas x-velocity profile (m/s) - VR = 1.0
1
Ug/U0
0.8
0.6
0.4
VR = 0, LES VR = 0, Exp. VR = 1, LES VR = 1, Exp. VR = 1.5, LES VR = 1.5, Exp.
0.2
0
0
2
4
6
8
10
12
14
16
x/d
(c) Gas x-velocity profile (m/s) - VR = 1.5
(d) Comparison between LES and exprimental data of gas averaged x-velocity along the centerline for different velocity ratios / No particles
Figure 3: Simulation results of gas x-velocity without particles
Velocity profiles are predicted with a maximum error less than 10%. Results were considered accurate enough to validate the modeling approach. Possible ways to improve predictions are specifying better inlet boundary conditions and increasing mesh resolution. Two-Phase Coaxial Jet. Again, using the Budilarto experiments, simulation results for a coaxial jet for different velocity ratios and particle size are presented in Figures 3 through 7. Figure 3 shows instantaneous profiles for VR = 0. The gas velocity profiles for 25-micron and 70-micron particles look similar [(a), (b) ]. The particle velocity profiles [(c),(d)] are however very different from the gas phase velocity. First, it should be noticed that, for numerical reason, particle velocity and number density are set to zero when the particle concentration is six orders of magnitude smaller than the inlet concentration. Then, it can be observed that the large particles have more inertia; their speed decreases more slowly along the centerline and they are relatively unaffected by eddies in the gas phase. Small particles have a velocity profile which is closer to the gas velocity and are clearly affected by eddies. This is also confirmed by profiles of the number density (i.e. number of particles per volume) [(e),(f)]; the small particles spread more than the large ones and also cluster faster. Figure 4 and 5 presents profiles respectively for VR = 1.0 and VR = 1.5. The correlation between the gas phase and the particle phase for small particles is explicit. Effects of the annular jet can be observed on the small particles velocity at the edges of particle jet. Large particles velocity is less influenced by the gas phase and decreases slowly along the centerline. Figure 6 presents comparisons of the mean gas and particles velocity along the centerline between the simulation predictions and experimental data. For VR = 0 [(a),(b)], small particles velocity is almost the same as the gas velocity. Initially, their velocity is slightly slower than the gas velocity, but around x/d =4, the two lines cross and because of inertia the particle velocity becomes larger than the gas velocity. Large particles have a velocity profile very different from the gas velocity. Initially, their speed is lower and when x/d <6 they slow down the gas phase while their own speed increases. Then the gas velocity drops below the particle velocity and particles are now a positive source of momentum to the gas phase. For VR = 1.0 [(c),(d)], it can be noticed that simulations predicts slightly higher velocities than experiments, which was expected given that the gas velocity was also over predicted in the case with no particles. Results for VR = 1.5 [(e),(f)] are in good agreement with experimental data. Figure 7 provides plots of the mean concentration of particles along the radial direction at x = 15d. Values are normalized by the centerline value of the weight. For every velocity ratio, it can be observed that the fine particles have a wider distribution whereas the coarse particle jet remains more compact. Simulations give a good prediction for VR = 0 even though the spreading rate is under predicted for VR = 1.0 and VR = 1.5. Simulations predict a higher spreading for VR = 1.5 than for VR = 1.0, which is also observed experimentally. A lower spreading rate for VR = 1.0 than for VR = 0 is predicted, which is probably a consequence of the lower shear stress; this behavior is however not confirmed by the experimental data. Possible reasons for the discrepancy are inaccurate inlet conditions and mesh resolution. The particle velocity at the inlet is indeed specified with no fluctuations, which creates a slower mixing and spreading rate. A finer mesh resolution could also improve the mixing in the gas phase and have an impact on the solid phase. Overall, simulation results show that the modeling approach is able to capture particle segregation.
(a) Gas x-velocity (m/s)
(b) Gas x-velocity (m/s)
(c) Particle x-velocity (m/s)
(d) Particle x-velocity (m/s)
(e) Number density
(f) Number density
Figure 4: Simulation results (VR = 0) for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)]
Figure 3: Simulation results (VR = 0) for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)]
(a) Gas x-velocity (m/s)
(b) Gas x-velocity (m/s)
(c) Particle x-velocity (m/s)
(d) Particle x-velocity (m/s)
(e) Number density
(f) Number density
Figure 5: Simulation results (VR = 1) for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)]
Figure 4: Simulation results (VR = 1) for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)]
(a) Gas x-velocity (m/s)
(b) Gas x-velocity (m/s)
(c) Partcile x-velocity (m/s)
(d) Particle x-velocity (m/s)
(e) Number density
(f) Number density
Figure 6: Simulation results (VR = 0) for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)]
Figure 5: Simulation results (VR = 0) for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)]
1
1
0.9
0.8
0.8
U/U0
U/U0
0.9
0.7
0.7
0.6
0.6
Ug−25um−LES Ug−25um−Exp. Up−25um−LES Up−25um−Exp.
0.5
0.4
0
2
4
Ug−70um−LES Ug−70um−Exp. Up−70um−LES Up−70um−Exp.
0.5
6
8 x/d
10
12
14
0.4
16
0
2
4
(a) VR = 0
8 x/d
10
12
14
16
10
12
14
16
12
14
16
(b) VR = 0
1
1
0.9
0.8
0.8
U/U0
U/U0
0.9
0.7
0.7
0.6
0.6
Ug−25um−LES Ug−25um−Exp. Up−25um−LES Up−25um−Exp.
0.5
0.4
6
0
2
4
Ug−70um−LES Ug−70um−Exp. Up−70um−LES Up−70um−Exp.
0.5
6
8 x/d
10
12
14
0.4
16
0
2
4
(c) VR = 1.0
6
8 x/d
(d) VR = 1.0
1 1
0.9 0.9
0.8
U/U0
U/U0
0.8
0.7 0.7
0.6
0.6
Ug−25um−LES Ug−25um−Exp. Up−25um−LES Up−25um−Exp
0.5
0.4
0.4 0
2
4
Ug−70um−LES Ug−70um−Exp. Up−70um−LES Up−70um−Exp.
0.5
6
8 x/d
10
(e) VR = 1.5
12
14
16
0
2
4
6
8 x/d
10
(f) VR = 1.5
Figure 7: Comparison between simulations and experimental data of time-averaged gas and particles velocities Figure 6: Comparison between simulations and experimental data of time-averaged gas and particles along the centerline for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)] velocities along the centerline for 25 um particles [(a) (c) (e)] and 70 um particles [(b) (d) (f)]
Weight−25um−LES Weight−70um−LES Weight−25um−Exp. Weight−70um−Exp.
1
0.8
DR/DRmax
DR/DRmax
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0
Weight−25um−LES Weight−70um−LES Weight−25um−Exp. Weight−70um−Exp.
1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
r/d
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
r/d
(a) VR = 0
(b) VR = 1.0
Weight−25um−LES Weight−70um−LES Weight−25um−Exp. Weight−70um−Exp.
1
DR/DRmax
0.8
0.6
0.4
0.2
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
r/d
(c) VR = 1.5
8: Radial profiles of particles number density taken at x = 15d Figure 7: Radial profiles ofFigure particles number density taken at x = 15d
Gasification: A first LES validation study of the BYU gasifier was performed using the Arches LES code. Experimental data were analyzed and experimental uncertainty numbers were produced from this analysis. These uncertainty numbers were then used to perform a two-parameter gasification validation study. Several numerical issues impeding the progress of simulations were resolved, and full gasification simulations using a coupled Direct Quadrature Method of Moments (DQMOM)-Large Eddy Simulation (LES) algorithm were performed. Color plots of gas-phase mass fractions at a particular instant in time for CO and CO2 are shown in Figure 8. The affect of large-scale structures on the carbon conversion is clearly shown. Figure 9 shows a plot of the zeroth moment, which is the number density function for the coal particles. From this figure, the dispersion of particles is clearly visible.
(a)
(b)
Figure 8: Gas-phase mass fractions (a) CO (0.0 ωCOCO ≤ 0.28) (0.0 ≤ and ωCO % (b) ≤ 0.31). Figure 2: Gas-phase massfor fractions for ≤(a) (0.0and ≤ ω(b) ≤20.28) CO2 COCO 2
(0.0 ≤ ωCO2 ≤ 0.31).
8
Figure 3:of Plot of m moment m000 ,density the number density of particles in000the flow Figure 9: Plot moment 000, the number of particles in the flow (1.5e04 ≤ m ≤ 1.0e10). (1.5e04 ≤ m000 ≤ 1.0e10).
In addition to the gasification simulations performed, a preliminary validation case matrix was also run for the 4 ofResults Discussion purpose applying theand Data Collaboratory approach to validating the simulation results. toward this end, a 4case validation matrix was run, and two parameters were identified as key variables.
Color plots of gas-phase mass fractions at a particular instant in time for CO and • Solids flowrate m ̇ solid (variation from a baseline value) CO shown in Figure 2. The affectofof10% large-scale structures on the carbon 2 aremass conversion is clearly shown. Figure 3 shows a plot of the zeroth moment, which • Kobayashi-Sarofim devolatilization model activation energy E2 (variation of 10% from a baseline value) is the number density function for the coal particles. From this figure, the One parameter,ofthe solids mass flowrate, represented a “scenario parameter,” a parameter that is related to dispersion particles is clearly visible. the physical conditions of the gasifier. The other represented a model parameter, that is, Figure 4 shows plots of radial concentration profiles at various heights in a numerical value that affects behavior an important The devolatilization model is of high importance in coal the the gasifier. Theofresults do notsub-model. compare favorably, and are particularly undergasification, and the second Kobayashi-Sarofim reaction dominates at higher temperatures, so the value E2 predicted at the centerline midway through the reactor. There are several pohas the strongest impact on the devolatilization rate in practical coal gasification systems. tential sources of error, including the particle velocity model, a poor gas-phase model, or under-prediction of rawa coal reaction rates. Incorrect predicThemixing validation results, shown in figure showed limited sensitivity to both parameters. However, before drawing conclusions, more favorable gasification would be needed. tion of any of these quantities could lead toresults under-prediction of combustion and gasification reaction products near the centerline. Additionally, no models have been implemented for slagging, fouling, or soot evolving from the coal, all of which will affect the species concentrations. References: 1 B.W. The poor comparison results is P.O. not of great concern; the results presentedof entrained-flow Brown, L.D. Smoot, P.J.ofSmith, and Hedman, "Measurement and prediction gasification AIChE Journal, vol. 34, 1988, pp. 435-446. representprocesses," a first-pass gasification calculation. Little effort to date has been focused on improving the gasification physics, validating gas-phase and particle2 Frenklach, M., Packard, A., Seiler, P., and Feeley, R., “Collaborative Date Processing in Developing phase mixing, or implementing alternatives to the particle velocity model in Predictive Models of Complex Reaction Systems,” International Journal of Chemical Kinetics, 36, 57–66 the Arches LES code. Efforts have primarily focused on achieving numerical (2004). stability in the dispersed-phase model, DQMOM, in order to get to the point 3 Feeley, R., Seiler, P., Packard, A., and Frenklach, M., “Consistency of a Reaction Dataset,” Journal of where significant effort can now be focused on improving the coal combustion Physical Chemistry A,physics. 108, 9573–9583 (2004). and gasification 4
H. Kobayashi, J.B. Howard, and Adel F. Sarofim. Coal devolatization at high temperatures. Sixteenth Symposium (International) on Combustion, pages 411–425, 1976.
9
5
Daniele Marchisio and Rodney O. Fox. Solution of population balance equations using the direct quadra- ture method of moments. Journal of Aerosol Science, 36:43–73, 2005. 6
L. Douglas Smoot and Philip J. Smith. Coal Combustion and Gasification. Plenum Press, 1985.
7
J. Spinti, J. Thornock, E. Eddings, P. Smith, and A. Sarofim. Heat transfer to objects in pool fires. In WIT Press, editor, Transport Phenomena in Fires, 2008. 8
R. Rawat, S.G. Parker, P. Smith, and C. Johnson. Parallelization and integration of fire simulations in the uintah pse. In 10th SIAM Conference on Parallel Processing for Scientific Computing, 2001. 9
R.Feely, P. Seiler, A. Packard, and M. Frenklach. Consistency of a reaction dataset. J. Phys.Chem., 108:9573–9583, 2004. 10
M. Frenklach, A. Packard, P. Seiler, and R. Feely. Collaborative data pro- cessing in developing predictive models of complex reaction systems. Int. J. Chem. Kinet., 36:57–66, 2004. 11
Stephanus Budilarto, Chemical Engineering, PhD, Purdue University, “Experimental Study of Velocity Ratio, Particle Size and Size Distribution Effects in Particle-Laden Jet Flow”, 2003
NUMERICAL SIMULATION OF THE HYDRODYNAMICS OF A FLUIDIZED BED COMBINED WITH AN ENTRAINED BED GASIFIER Jiantao Zhao Institute of Coal Chemistry, Chinese Academy of Sciences Taiyuan, China, 030001, [email protected], Tel: +86-351-4031362 Jiejie Huang Institute of Coal Chemistry, Chinese Academy of Sciences Taiyuan, China, 030001, [email protected], Tel: +86-351-20211371 Yitian Fang Institute of Coal Chemistry, Chinese Academy of Sciences Taiyuan, China, 030001, [email protected], Tel: +86-351-20211371 Yang Wang Institute of Coal Chemistry, Chinese Academy of Sciences Taiyuan, China, 030001, [email protected], Tel: +86-351-4048313
Abstract: The ash agglomerating fluidized bed (AFB) coal gasification process, developed by the Institute of Coal Chemistry, Chinese Academy of Sciences, has the advantages of moderate operation temperature, lower oxygen consumption, lower operation cost, higher coal adaptability, and especially the fitness to coals of high ash content and high ash fusion temperature. The commercial pressured AFB gasifier (300t/d, 0.6MPa) have been put into operation for the ammonia and methanol synthesis. However, the carbon conversion (~90%) still has the potential to be highly improved by increasing the gasification efficiency of fly ash with high carbon content and low reactivity. Therefore, a new concept for coal gasification was evolved for coal gasification, combining the fluidized bed gasifier and the entrained flow gasifier into one unit. The combined gasifier can gasify feedstock with higher reactivity in the fluidized bed and further gasify fly ash with lower reactivity in the entrained flow bed at higher temperature (~1200 oC). The preliminary experiments have been carried out to confirm its feasibility. The present work aims to establish a rational mathematic model to simulate the fluidization dynamics to assist with the design, operation and scale-up of the gasifier. The CFD model is established base on the results from the simulation on the entrained flow bed and the fluidized bed. The advanced hybrid grid technology was used to build the numerical mesh. A series of simulation works were carried out using the CFD model established for the pilot scale coal gasifier. The gas-solid flow field, particle distribution, the mutual influence of the gas and particle flow between the entrained flow bed and the fluidized bed were discussed in detailed. Keywords: coal gasification; fluidized bed; entrained flow bed; numerical simulation 1. Introduction Gasification is a key step for high-efficiency and less-polluting coal utilization. At present, two types of gasification reactor—the fluidized bed and the entrained flow bed are considered as the promising technologies for large-scale coal gasification and have been commercially available. The fluidized bed gasifier usually operates at moderate temperature (~1000oC) with the advantages of lower oxygen consumption and lower operation cost. However, the total carbon conversion is not very high (~90%) because of lower reactivity of fly-ash [1] and lower reaction ability of the free board region [2]. The entrained flow gasifier has the advantages of high carbon conversion (>98%) and high effective gas
(CO+H2) content (>90%) due to its operation at higher temperature (1200~1500oC). However, the entrained flow gasifier has the higher oxygen consumption, lower thermal efficiency and higher cost. Based on many years’ R&D work on the Ash-agglomerating Fluidized Bed gasification process, an integrated coal gasifier has been developed by the Institute of Coal Chemistry, Chinese Academy of Sciences (ICC, CAS). It is a kind of new gasifier which combines the advantages of high temperature entrained flow gasification and moderate temperature fluidized bed gasification. The relatively lower temperature semi-coke in the dense phase of the fluidized bed and the gas produced by fluidized bed part have the function to cool down the high-temperature product gas and slag coming from the entrained flow part so that the slag particles can be mixed into the dense phase and then discharged together with fluidized bed’s coarse ash in dry mode. This kind of integrated gasifier possesses has many advantages like simpler structure, lower cost, higher carbon conversion, lower oxygen consumption and wider coal adaptability. At present, the conception confirmation of this integrated coal gasifier has been accomplished [3], and the R&D on process optimization and scale-up is now carrying out. It is well know that the traditional scale-up process in chemical industry usually needs a long time and is very costly. In recent years, the rapid development of Computational Fluid Dynamics (CFD) and the computer technology has made it possible to study gas-solid two phase flow more accurately and economically. Moreover, the reliable simulation software has been presented for the description of fluidized bed and entrained flow reactor. Therefore, CFD method is used to study the gas-solid flow of the coupling process of the fluid bed and the entrained flow bed. The objective of this work hopes to provide some instructive information for the fundamental research and the technical development of this new process. 2. Mathematical model 2.1 Physical model The pilot-scale integrated gasifier attached an entrained flow bed on the side of the fluidized bed, as shown in Fig.1. Both the fluidized bed and the entrained flow bed have the inner diameter of 300mm. The lower part of the fluidized bed is height of 1300mm and the length of the entrained bed is 2100mm, respectively. The free board has the inner diameter of 500mm and its height is 1000mm in the simulation. The properties of the gas and particulates are listed in Table 1.
Fig.1. Structure and mesh scheme of the integrated gasifier
Table 1. Properties of particulate phases and gas phase Values Coarse particle Fine particle
Parameters Particle density /(kg/m3)
1200
900
Particle mean diameter /mm
1.1
0.06
Minimum fluidization velocity /(m/s) Voidage at minimum fluidization Gas dynamical viscosity /(Pa·s) Gas density /(kg/m3)
0.4 0.42 1.79×10-5 1.225
---
2.2 CFD model equations Numerical schemes based on the Eulerian-Eulerian and the Eulerian-Lagrangian approachs have been used for CFD simulation of the gas-solid multiphase flow. The Eulerian-Eulerian approach, namely two-fluid hydrodynamic model, considers the particulate phase to be a continuous fluid interpenetrating and interacting with the fluid phase, which is suitable to deal with the dense solid flow such as the fluidized bed process. The Eulerian-lagrangian approach considers the particulates as the discrete phase, which is fit for the simulation of the dilute solid phase. The integrated gasifier is composed of the fluidized bed with dense particulate phase and the entrained bed with dilute particulate phase. However, it is difficult to deal with the different gas-solid flow with the two approaches in a single unit. Therefore, the Eulerian-Eulerian model was use for its simulation [4, 5]. This approach results in mass, momentum, and energy balance equations for both gas
and solid phases, as given below: Continuity equations: q q q q vq 0 ; t Momentum equations: For gas phase:
p or q s, g , but p q
( g g v g ) ( g g v g v g ) g p g g g g t g g ( Fg Flift , g Fvm , g ) K sg (v s v g ) msg v sg
(1)
(2)
For solid phase: In the integrated gasifier, there are two different solid phases. One is the coarse particle phase with larger size in the fluidized bed; the other is the fine particle phase with smaller size in the entrained bed. For the coarse particle phase, the interaction of particles should be considered in the dense gas-solid flow. ( s1 s1v s1 ) ( s1 s1v s1 v s1 ) s1p p s1 s1 s1 s1 g t s1 s1 ( Fs1 Flift ,s1 Fvm ,s1 ) K s1, g (v s1 v g ) ms1g v s1g
For Fine particle phase, the interaction of particles can be neglected.
(3)
( s 2 s 2 v s 2 ) ( s 2 s 2 v s 2 v s 2 ) s 2 p s 2 s 2 s 2 g s 2 s 2 ( Fs 2 Flift ,s 2 Fvm,s 2 ) t K s 2 g (v s 2 v g ) m s 2 g v s 2 g
(4)
The fluid-solid momentum exchange coefficient between gas and solid phase can be expressed as follows: 1 2 (5) K s 2, g p d s 2 C D v f v p ; 8 The drag force coefficient can be determined as follows:
C D max 24(1 0.15 Re 0.687 ),0.44
(6)
Turbulence model: The RNG k-ε model [6] is used for the turbulent flow of the gas phase. ( U ) t
C 1 Pk C 2 k
(7)
( Uk ) t k
k Pk R
(8)
R
C 3 1 0 2 1
eff t ;
3
k
;
t C
S k ; S 2Sij Sij k2
;
Sij
1 vi v j 2 x j xi
,
; Pk t U (U U T )
C 0.085 ; k 0.7179 ; 0.7179 ; C 1 1.4 ; C 2 1.68 ; 0 4.38 ; 0.012
Table 2. Simulating conditions of the integrated gasifier Case Parameters Values T /K 293 P /MPa 0.1 H0 /m 0.65 0-3 Uf /(m/s) 0.85 0 Ue/Uf 0 1 Ue/Uf 0.25 2 Ue/Uf 0.5 3 Ue/Uf 1.0 H0 /m 0.7 Uf / (m/s) 0.85 4 Ue/Uf 0.5 H0−static bed height, m; P−pressure, MPa; T−temperature, K; Ue−gas velocity in the entrained bed, m/s; Uf−gas velocity in the fluidized bed, m/s
The calculated domain was meshed by an advanced hybrid grid technology to build rational numerical cells. The total number of the elements was 20410. A commercial software package, Fulent 6.1, was employed in this study. The simulation was conducted in the transient mode. The time step and total time was selected to 0.0005s and100s respectively. The operating conditions used in the simulations are listed in Table 2. Five cases were conducted in simulation. Case 0 was only the fluidized bed. From case 1 to case 3, the entrained bed gas velocity gradually increased to investigate the solid movement and the interacting between the fluidized bed and entrained bed. Case 4 was to measure the influence of the dense phase bed height of the fluidized bed on the mixture of coarse and fine particles. 3. Simulation results and discussion 3.1 Gas-solid flow pattern The gas-solid hydrodynamics in the single fluidized bed and the entrained bed have been extensively studied respectively [7-9]. The hydrodynamic characteristics of the combined gasifier are the interactive behaviors of gas and solids between the two reactors. The effects of the gas velocity, bed height (hold-up) in the fluidized bed were investigated by the numerical simulation in the research works. Figure 2 shows the gas and particle velocity distributions for cases 1-3. From case 1 to case 3, the gas velocity is increased gradually. In the entrained bed, it can be seen that the flow patterns of gas and fine particle are similar due to the strong interaction in each operating cases. Fine particles entrained by high velocity gas from nozzle injects into the vessel. Swirling eddies appear near the wall along the axis direction. The rising gas velocity can decrease the swirling flow and make gas and particle flows tend toward the plug flow in the body of entrained bed although it may promote the swirling near the nozzle because of the high velocity jet causing the vacuum suction. In the connecting area of the entrained bed and fluidized bed, gas and fine particles form a strong up-flow. Meanwhile, the violent swirling flow along axial direction is produced due to the intensive momentum exchange among gases and particles from different beds. Higher gas velocity from the entrained bed may enhance the swirling flow. The flow pattern means that the inner circulation of gas and solid may occur in the joint area to promote the mixture efficiency, which is beneficial to the heat and mass transfer in the gasifier.
(1) Gas
(2) Fine particle (a) Case 1
(1) Gas
(2) Fine particle (b) Case 2
(1) Gas
(2) Fine particle (c) Case 3
Fig. 2. Speed vectors of gas and fine particle for case1-3
3.2 Fine Particle distribution The volume fractions of fine particle for case 1-3 are shown in Fig. 3. In the entrained bed, the fine particle distribution is not uniform because the injecting gas from nozzle leads the intensive swirling flow. Higher particle concentration appears near the nozzle area and the underside of the wall. In order to prevent coal ash from agglomerating and adhering to the wall, particle deposition should be as low as possible. From the comparison of different cases, it also can be seen that the higher gas velocity can reduce the particle deposit near the wall; however, it may promote the suction effect to increase the particle concentration near the nozzle. The higher particle concentration suggests a high temperature area formed near the nozzle area, which is harmful for the long-term operation of the nozzle. Therefore, the appropriate gas velocity is the key to the operation of the novel gasifier. Figure 4 and 5 shows the cross-sectional fine particle distribution and average volume fraction at the joint of the fluidized bed and the entrained flow bed gasifier, respectively. It can be seen that the fine particle concentration is higher on the side of the joint. Furthermore, lower gas velocity in the entrained bed leads the decrease of gas-solid drag force so that more fine particles enter the fluidized bed. In case 1, the fine particle volume fraction in the joint was much larger than that in the case 2 and case 3, as shown in Figure 5. More fine particles entered into the fluidized bed suggest that high temperature slag particles formed in the entrained bed can be mixed into the fluidized bed so that they are cooled and discharged as dry mode to improve the heat efficiency in the gasification process.
(a) Case 1
Case 2
Case 3
Fig. 3. Fine particle volume fraction distributions
(a) Case1
(b) Case2
(c) Case3
Fig. 4. Fine particle volume fraction distributions at the cross-section of the joint
Volume fraction of fine particle
0.0035 0.0030 0.0025 0.0020 0.0015 0.0010 0.0005 0.0000 Case 1
Case 2
Case 3
Fig.5. Fine particle volume fraction at the cross-section of the joint 3.3 Effects of the fluidized bed height The gas and solid interaction between the entrained bed and the fluidized bed is greatly influenced by the dense phase height in the fluidized bed. If the bed height is above cross-section of the joint zone, the gas and solid from different reactors can be extensively mixed, whereas the mixture may be lessened. Figure 6 and 7 show the fine particle and coarse particle volume fraction distributions and the cross-sectional fine particle volume at the joint for case 2 and case 4. The coarse particle volume fraction suggests the dense phase height in the fluidized bed. It can be seen that the fluidized bed height is below the joint zone in case 2, whereas that in case 4. The fluidized bed height is mainly influenced by the static bed height and the fluidizing gas velocity and less influenced by the gas velocity of the entrained bed. The simulation results show that the dense phase of the fluidized bed higher than the joint area is beneficial for
the mixture of fine particles and coarse particles. In case 4, more fine particles are entered into the fluidized bed because of the entrainment of fluidizing coarse particles.
Case 2 (a) Coarse particle
Case 2
Case 4
Case 4
(b) Fine particle Fig.6. Fine particle and coarse particle volume fraction distributions for case2 and case4 4. Conclusions A 3-D CFD simulation model by combining TFM (two-fluid model) and KTGF (kinetic theory of granular flow) based on commercial CFD package has been developed for the novel fluidized-entrained bed combined gasifier. A series of numerical simulations were performed under several operating conditions. The gas-solid mutual interaction between the entrained flow bed and the fluidized bed including the gas-solid flow fields and particle concentration distributions were predicted in detail using the model. It was concluded that the appropriate gas velocity in the entrained bed and the dense phase height in the fluidized bed are critical for the gasification process. Although only cold gas-solid flow was
Volume fraction of fine particle
0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000 Case 2
Case 4
Fig.6. Cross-sectional fine particle volume at the joint for case2, 4 considered and no more comparisons were conducted with experimental data, this work indeed will provide a base for the future comprehensive gasification model developing and give a better understanding to the combining gasifier. Acknowledgements This work was supported by the Special Funds for Major State Basic Research Projects Program (the 973 Program) (Grants 2005CB221200) and Knowledge Innivation Program of the Chinese Academy of Sciences (Grants KGCX2-YW-320). References 1. Y. T. Fang, The Study on Gasification Kinetics of Char Fines from Free-board of Fluidized Bed Gasifier, Master thesis, ICC, CAS, 1994. 2. Z. Tang, Simulation of Fluidized Bed Coal Gasification Process, PhD. thesis, ICC, CAS, 1997. 3. Jinhu Wu, Yitian Fang. A new integrated approach of coal gasification: the concept and preliminary experimental results, Fuel Processing Technology, 2004, 86: 261-266. 4. Matteo Chiesa, Vidar Mathiesen, Numerical simulation of particulate flow by the Eulirian-Lagrangian and the Eulerian-Eulerian approach with application to a fluidized bed, Computers&Chemical Engineering, 2005, 29: 291-304. 5. William Vicent, Salvador ochoa, An Eulerian model for the simulation of an entrained flow coal gasifier, Applied Thermal Engineering, 2003, 23: 1993-2008. 6. C. G. Speciale, S.Thangam, Analysis of an RNG based turbulence model for separated flows[J]. Int. J. Eng. Sci., 1992, 30(10): 1379-1388. 7. Jiantao Cao, Zhonghu Chen, Yitian Fang, et al, Simulation and experimental studies on fluidization properties, Powder Technology, 2008, 183: 127-132. 8. Kun Gao, Jinhu Wu, Yang Wang, et al, Bubble dynamics anf its effect on the performance of a jet fluidized bed gasifier simulated using CFD, Fuel, 2006,85:1221-1231. 9. H. Tominaga, T. Yamashita, Simulator development of entrained flow gasifier at high temperature and high pressure atmosphere, 9th Japan Australia Joint Meeting, Melbounre, Vic, Australia, 1999
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
NUMERICAL SIMULATION OF GASIFICATION PROCESS IN A CROSS-TYPE TWO-STAGE GASIFIER Yau-Pin Chyou, Chang-Bin Huang, Yan-Tsan Luan Nuclear Fuels and Materials Division Institute of Nuclear Energy Research Atomic Energy Council Longtan, Taoyuan, Taiwan (R.O.C.)
ABSTRACT Numerical simulation of the oxygen-blown coal gasification process inside a cross-type two-stage (E-Gas like) gasifier is studied with the commercial CFD solver ANSYS FLUENT. The purpose of this study is to use CFD simulation to improve understanding of the gasification processes in the E-Gas like gasifier. In this paper, chemical reaction time is assumed to be faster than the time scale of the turbulence eddies. All the species are assumed to mix in the intermolecular level. The 3-D Navier-Stokes equations and species transport equations are solved with the eddy-breakup reaction model (instantaneous gasification). The influences of coal slurry concentration and O2/coal ratio on the gasification process are investigated. Under the condition of feeding carbon being almost completely converted, low slurry concentration is preferred over high concentration if more H2 is wanted with lower syngas temperature; while higher slurry concentration is preferable for producing more CO with higher syngas temperature. The case of higher O2/Coal ratio results in more combustion and leads to lower syngas heating values and higher temperatures. Meanwhile, lower O2/Coal ratio involves more gasification reactions and results in higher CO concentration and lower temperature. The flow behavior in the gasifier, especially the single fuel injection design on the second stage, is examined and discussed. In summary, the trends of simulated results of coal combustion and gasification processes in the cross-type two-stage gasifier are reasonable and the developed simulation model in this study can be used as a tool for preliminary examination of the global effect of thermal-flow and turbulence on gasification process. Key words: clean coal technology, gasification modeling, integrated gasification combined cycle (IGCC). INTRODUCTION Fuel combustion is usually deemed as the main source of greenhouse gas emission inventories, and it typically contributes over 90% of CO2 emissions and 75% of total greenhouse gas emissions in developed countries (Garg et al., 2006). In Taiwan, 53.35% of electricity was generated from coal-fired power plants, 20.35% came from LNG-fired (Liquid Natural Gas), 18.10% came from Nuclear power, * Corresponding Author’s E-mail: [email protected]
Ting Wang Energy Conversion & Conservation Center University of New Orleans New Orleans, Louisiana, USA 3.31% came from oil-fired power plants, while renewable energy, including conventional and pumped storage hydro, generated only 4.9% of electricity in 2009 (BOE, 2010). Coalfired power plants generate the most electricity, but it also discharges the most CO2 emissions as contrasting to other electricity utilities. More than 45% of CO2 emissions were from coal-fired power plants. Therefore, to reduce CO2 emission, it is important to improve the efficiency of coalfired power plants or to capture CO2 from them. There are many ways to reduce carbon emissions and the associated carbon footprints. Technologies for employing renewable energy such as solar, wind, ocean, hydro, and biomass have been developed and are growing at a fast pace. However, it is undeniable that the most polluted fossil fuel, coal, is still the primary energy source throughout the world. Even in Taiwan, a country without domestic coal production, there is still 34% of energy consumption that comes from coal, in which, 77.62% is utilized for power generation, 1.29% for blast furnace, 8.52% for coke production, and 12.58% for industrial and others (BOE, 2010). The supply of coal is abundant, cheap, stable, and will last for 122 years (BP, 2009). The demand and consumption of natural gas and gasoline have increased during the past few decades. Due to the limited amount of petroleum reserves on the earth, oil prices will expectedly keep growing within the next forty to fifty years. Therefore, the importance of using coal fuel has been emphasized because coal fuels have higher stability and wider variety of sources. However, coal is not considered a clean fuel; thus, future efforts should focus on developing the usages of clean and affordable coal fuel. Gasification is the process of converting various carbonbased feedstocks to clean synthetic gas (syngas), which is primarily a mixture of hydrogen (H2) and carbon-monoxide (CO). This conversion is achieved through the reaction of the feedstock with oxygen and steam at a high temperature and pressure with less than 30% of the required oxygen for complete combustion being provided. The syngas produced can be used as a fuel, usually as a fuel for boilers or gas turbines to generate electricity, or it can be used as a source for manufacturing ammonia or hydrogenation applications in refineries to make methanol, hydrogen, or other chemical products. The gasification technology is applicable to any type of carbon-based feedstock, such as coal, natural gas,
burned to produce carbon dioxide, water, NOx, SOx and other trace elements. However, when air or oxygen is insufficient, incomplete or partial combustion occurs, and coal is gasified. In general, coal gasification is categorized for three main steps: (1) devolatilization, (2) volatile combustion and thermal cracking, (3) char combustion and gasification (or char oxidation). The char oxidation is the most important reaction mechanism to produce syngas, which contains complex chemical reactions interacting with the turbulent flow. Because coal combustion has a long history of being applied in industry, the combustion model study is relatively mature. Smoot (1984) and Eaton et al. (1999) comprehensively reviewed the development of coal combustion modeling in fixed, fluidized, and entrained bed. Considering the turbulence, particle dispersion and chemical reaction, Smoot and Smith (1985) and Hill and Smoot (1993) developed a two-dimensional (PGCG-2) and threedimensional (PGCG-3) pulverized coal combustion model at Brigham Young University. The model solved mass, momentum, and energy conservation equations to simulate coal gasification and combustion. Coal particles under high temperature stream, the heat transfer between dispersed and fluid phases, and the resulting chemical reaction and radiation are illustrated in detail by Smoot & Smith (1985). The previous studies of coal devolatilization and volatile combustion models developed in the coal combustion models are applied in coal gasification. Comprehensive research has been carried out to understand velocity and temperature distribution as well as syngas composition in the gasification chamber (Chen et al. 2000, Choi et al. 2001, Vicente et al. 2003, and Watanabe & Otaka 2006). Chen et al. (2000a) implemented a three-dimensional model to simulate a 200-ton two-stage air blown entrained type gasifier, developed for an IGCC process. The simulation showed that turbulent fluctuations affected the volatile and char-oxygen reaction and significantly influenced the temperature and gas composition. Chen et al. (2000b) investigated the coal gasification under different parameters such as air ratio and coal particle size. They found that carbon conversion is independent of the devolatilization rate and less sensible to coal particle sizes, but it is sensible to the heterogeneous char-oxygen, char-CO2 and char-steam reaction kinetics. Besides, air ratio had a significant effect on syngas composition. The research team at the Energy Conversion & Conservation Center (ECCC) at University of New Orleans has made a significant effort to develop a gasification numerical simulation through the commercial software Ansys/Fluent. Silaen & Wang (2005, 2006, 2009, and 2010) used the geometry and the operating conditions based on the information in Bockelie et al. (2002) and Chen et al. (2000a). They concluded that coal slurry feedstock produced more H2 than coal powder feedstock, while coal powder feedstock generated more CO. An instantaneous gasification model can provide an overall approach on gasifier performances in terms of temperature, heating value, and gasification efficiency but not adequately catch the local gasification process inside the gasifier. In 2010, they investigated the effects of different turbulence and coal particle size on coal gasification in an entrained-flow gasifier. They concluded that turbulence model significantly influences the gasification results, and only
heavy refinery residues, petroleum coke, biomass, and municipal wastes. Syngas can be employed in the Integrated Gasification Combined Cycle (IGCC) to produce electricity. Compared to regular or supercritical pulverized coal (SCPC) combustion power plants, IGCC plants can achieve higher efficiency and lower emissions. For power generation, gasification integrated with IGCC is considered a clean and efficient alternative to coal combustion.. The high-pressure and high-temperature syngas from the gasifier can take advantage of the new generation of advanced turbine systems (ATS) to a potential efficiency of more than 50% (LHV). Furthermore, the syngas stream can also be tapped to produce methanol and hydrogen. The gasification technology becomes more important when integrated with carbon capture and sequestration (CCS) technology. CCS is the technology that captures CO2 by a physical or chemical process and stores it. In a typical case, a fossil-fired power plant implemented with CCS technology can reduce 90% of CO2 emissions. Because IGCC+CCS capture CO2 before the syngas combustion, it is easier and cheaper to capture CO2 compared to PC power plants. In the next decade, coal-fired power generation will continue to be the main source of electric power in Taiwan; therefore, it is essential to improve the coal combustion efficiency or to develop new technologies such as coal gasification that can extend the usability of coal more cleanly. Before building an experimental gasification facility, this research team is interested in utilizing Computational Fluid Dynamics (CFD) to help understand the thermal-flow and gasification process in the gasifier and guide the design of an experimental gasifier. Hence, the objective of this study is to establish a preliminary coal gasification model to improve the understanding of the gasification processes in a cross-type (EGas like) two-stage gasifier as well as to investigate the effects of operating parameters such as slurry concentration and O2/coal ratio on gasification performance. LITERATURE REVIEW In an entrained type gasifier, coal particles are injected into the gasification chamber and mixed with an oxidant stream at a high speed. Under the assumptions of fast chemical reactions and local chemical equilibrium, the reaction and particle movement are dominated by turbulence. Therefore, turbulence models play an important role in coal combustion and gasification simulation, and it affects not only the flow field inside the gasifier but also the coal conversion rate of gasification. Launder & Spalding developed the k-ε model in 1972, and it has since been widely used due to its simple physics and robust feature. Over the past decades a considerable number of models, k-ε model, Reynolds stress model (RSM), large eddy simulation (LES), and direct numerical simulation (DNS), have been developed to study and simulate the turbulent flows.. Hunt & Savill (2004) gave an introduction of the turbulence models and the appropriate conditions to use these models. The chemical reactions between coal combustion and gasification are similar. Under high temperature, coal is first decomposed through pyrolysis to vaporized volatile substances. With abundant air or oxygen, coal is completely 2
standard k-ε model and RSM models gave the consistent result in their study. The Kobayashi devolatilization model produces a slower devolatilization rate than the other models. The single rate model and the chemical percolation model produces moderate and consistent devolatilization rate.
Mass of dry coal (mositure free) Mass of slurry (1) Mass of oxygen O 2 /Coal = Mass of coal (MAF, mositure and ash free) (2)
Slurry concentrat ion =
MODEL DESCRIPTION E-Gas (Destec) type Gasifier The E-Gas™ gasifier consists of two stages, a slagging first stage and an entrained-flow, non-slagging second stage, as shown in Fig. 1. The first stage is a horizontal, refractorylined vessel in which carbonaceous fuel is partially combusted with oxygen at an elevated temperature and pressure, 2500°F/420 psia (1400°C/29bar). Approximately 80% of preheated slurry and oxygen are fed to each of two opposing mixing nozzles, one on each end of the horizontal section of the gasifier. The geometry is designed to provide a means for thoroughly mixing the reactants and to disperse this mixture, so a high carbon conversion is realized. E-Gas™ has developed its own proprietary design for these slurry mixers. The oxygen feed rate to the mixers is carefully controlled to maintain the gasification temperature above the ash fusion point of 2400 – 2600 °F (1589 – 1700K) (NETL, 2000), ensuring good slag removal and high carbon conversion. The raw fuel gas flows upward into the second (upper) stage of the gasifier while the molten slag flows down the walls of the gasifier and passes into a slag quench bath. In this upper vertical cylindrical stage, the remaining coal slurry is fed to have additional gasification occur and cool down the syngas temperature. The fuel is almost totally gasified in this environment to form syngas consisting principally of hydrogen, carbon monoxide, carbon dioxide and water. Sulfur in the fuel is converted to primarily hydrogen sulfide (H2S) with a small portion converted to carbonyl sulfide (COS). With appropriate processing downstream, over 98-99% of the total sulfur can be removed from the feedstock prior to combustion in the combustion turbine (NETL, 2000; NETL, 2002). Descriptions of Baseline Case Since the exact dimension of the E-Gas gasifier is not known to the authors, a geometry of a cross-type, two-stage gasifier has been built and simulated based on Fig. 1 and the E-gas (Destec) gasification process with feedstock information published in the open literature (NETL, 2000 and 2002). The Fuel properties of the received Illinois #6 Coal are shown in Table 1 and Table 2. The conditions of coal/water slurry feed, prepared using Illinois #6 coal of the baseline case are shown in Table 3. Approximately 80% of the slurry is gasified / combusted with oxygen in the first (lower) stage using two burners positioned on the opposing ends of this horizontal cylindrical section. In this upper vertical cylindrical stage, only the remaining coal slurry is fed without oxygen. The slurry concentration is 0.67 and O2/coal ratio is 0.91 for the baseline case. The definition of slurry concentration and O2/coal ratio in this paper are as follows:
Figure 1 E-Gas ( Previous Destec ) two-stage entrained flow gasifier. (Source: Destec, 1996)
Table 1 Ultimate analysis of feedstock coal
Ultimate Analysis Moisture Carbon Hydrogen Nitrogen Chlorine Sulfur Ash Oxygen Total
3
Wt. % 11.12 63.75 4.50 1.25 0.29 2.51 9.70 6.88 100
Wt. %, dry 71.72 5.06 1.41 0.33 2.82 10.91 7.75 100
including the enthalpy formation from the chemical reaction of the species. The energy E is defined as p v2 E = h− + (7) ρ 2
Table 2 Proximate analysis of feedstock coal
Proximate Analysis Moisture Fixed Carbon Ash Volatiles Total
Wt. % 11.12 44.19 9.70 34.99 100
Wt. %, dry 49.72 10.91 39.37 100
where h is the sensible enthalpy and for incompressible flow and is given as p h= Y jh j + (8)
∑
ρ
j
Yj is the mass fraction of species j and
Table 3 The feedstock conditions of base case
T
h=
Feedstock conditions of base case Flow rate (tons/day) - Coal (Wet base): - Slurry Concentration: - ASU O2/Coal (MAF): 4627
Second Stage Flowrates (tons/day) - Coal (Dry base, MF): - H2O: - O2 - N2
711
Tref
Turbulence Models Standard k-ε Model – The standard k-ε model, based on the Boussinesq hypothesis, relates the Reynolds stresses to the mean velocity as ⎛ ∂u ∂u j ⎞ 2 ⎟ − ρkδ ij (10) − ρ ui′u′j = μt ⎜ i + ⎜ ∂x j ∂xi ⎟ 3 ⎝ ⎠
1793 896 1841 97
where k is the turbulent kinetic energy, and μt is the turbulent viscosity given by μ t = ρC μ k 2 / ε (11)
474 237 0 0
where Cμ is a constant and ε is the dissipation rate. The equations for the turbulent kinetic energy (k) and the dissipation rate (ε) are: ⎡⎛ ⎤ ⎞ ∂ (12) (ρui k ) = ∂ ⎢⎜⎜ μ + μt ⎟⎟ ∂k ⎥ + Gk − ρε ∂xi ∂xi ⎢⎣⎝ σ k ⎠ ∂xi ⎥⎦ 2 ⎡⎛ ⎤ ⎞ ∂ (13) (ρuiε ) = ∂ ⎢⎜⎜ μ + μt ⎟⎟ ∂ε ⎥ + C1ε Gk ε − C2ε ρ ε σ ε ⎠ ∂xi ⎦⎥ ∂xi ∂xi ⎣⎢⎝ k k
METHODOLOGY Governing Equations The equations for conservation of mass, conservation of momentum, and energy equations are given as: v ∇ ⋅ (ρv ) = S m (3) v vv v ∇ ⋅ (ρv v ) = −∇p + ∇ ⋅ ⎛⎜τ ⎞⎟ + ρg + F (4) ⎝ ⎠ ⎛ v v v ⎞ ∇ ⋅ (v (ρE + p )) = ∇ ⋅ ⎜ λ eff ∇T − h j J j + ⎛⎜τ eff ⋅ v ⎞⎟ ⎟ + S h ⎜ ⎝ ⎠⎟ j ⎝ ⎠ (5) where λeff is the effective conductivity (λ+λt, where λt is the turbulence conductivity) and Jj is the diffusion of species j.
The term Gk is the generation of turbulence kinetic energy due to the mean velocity gradients. The turbulent heat flux and mass flux can be modeled with the turbulent heat conductivity (λt) and the turbulent diffusion coefficient (Dt), respectively. μ ∂T ∂T ρc p ui′T ′ = −λt (14) = −c p t ∂xi Prt ∂xi
∑
The stress tensor τ is given by 2 v ⎤ v ⎡ v τ = μ ⎢ ∇v + ∇v T − ∇ ⋅ v I ⎥ 3 ⎣ ⎦
(
)
(9)
p , j dT
where Tref is 298.15 K.
2550 0.67 0.91
First Stage Flowrates (tons/day) - Coal (Dry base, MF): - H2O: - O2 - N2
∫c
ρ ui′C ′ = − ρDt
μ ∂C ∂C =− t ∂xi Sct ∂xi
(15)
The constants C1ε, C2ε, Cμ, σk, and σε used are: C1ε = 1.44, C2ε = 1.92, Cμ = 0.09, σk = 1.0, σε=1.3. The turbulence Prandtl number, Prt , is set to 0.85, and the turbulence Schmidt number, Sct , is set to 0.7.
(6)
where μ is the molecular dynamic viscosity, I is the unit tensor, and the second term on the right-hand side is the effect of volume dilatation. The first three terms on the right-hand side of equation (5) represent heat transfer due to conduction, species diffusion, and viscous dissipation. Sh is a source term
Reaction Model Coal gasification reactions occur when coal is heated with limited oxygen and steam in a gasification reaction
4
the solid-gas reaction process can be modeled as homogeneous combustion reactions. This approach is based on the locally-homogeneous flow (LHF) model proposed by Faeth (1987), implying infinitely-fast interphase transport rates. The instantaneous gasification model can effectively reveal the overall combustion process and results without dealing with the details of the otherwise complicated heterogeneous particle surface reactions, heat transfer, species transport, and particle tracking in turbulent reacting flow. The eddy-dissipation model is used to model the chemical reactions. The eddy-dissipation model assumes the chemical reactions are faster than the turbulence eddy transport, so the reaction rate is controlled by the flow motions.
chamber. The main global reactions in a gasification process are as follows: Heterogeneous (solid and gas) phase C(s) + 1 2 O 2 → CO, ΔH o R = −110.5 MJ/kmol (R1) C(s) + CO 2 → 2CO, ΔH o R = +172.0 MJ/kmol (Gasification, Boudouard reaction)
(R2) C(s) + H 2 O(g) → CO + H 2 , ΔH o R = +131.4 MJ/kmol
(Gasification) (R3) Homogenous gas phase CO + 1 2 O 2 → CO 2 , ΔH o R = −283.1 MJ/kmol
Solution Methodology The CFD solver used in this study is the commercial CFD code ANSYS FLUENT V.12.0. FLUENT is a finitevolume-based CFD solver written in C language and has the ability to solve fluid flow, heat transfer and chemical reactions in complex geometries, and supports both structured and unstructured meshes. Buoyancy induced flow is calculated. The density in the buoyancy term in the momentum equation is calculated using the ideal gas law. The segregated solution method is used. Segregated solution solves the governing equations of continuity, momentum, energy, and species transport sequentially (segregated from one another). The non-liner governing equations are linearized implicitly with respect to the dependent variables. The second-order discretization scheme is applied for the momentum, the turbulence kinetic energy, the turbulence kinetic dissipation, the energy, and all the species. The SIMPLE algorithm is used in this study as the algorithm for pressure-velocity coupling. The built-in standard k-ε turbulence model is used.
(R4) CO + H 2 O(g) → CO 2 + H 2 , ΔH o R = −41.0 MJ/kmol (Watershift)
(R5) CH 2.760 O 0.262 → 0.262CO + 1.011H 2 + 0.123C 6 H 6 (Volatiles cracking)
(R6) C 6 H 6 + 3O 2 → 6CO + 3H 2 (Volatiles gasificati on via C 6 H 6 )
(R7) Reactions given in R1 and R4 are two exothermic reactions that provide the complete energy for the gasification. Based on these global reactions, approximately 22% of the stoichiometric oxygen is required to provide sufficient energy for gasification reactions. In real applications, 25~30% of the stoichiometric oxygen is provided to ensure high-efficient carbon conversion. Partial combustion occurs when the coal mixes with oxygen (R1). The energy released from (R1) also heats up any coal that has not burned. When the coal is heated without oxygen, it undergoes pyrolysis during which phenols and hydrocarbon gases are released. At the same time, char gasification (R2) takes place and releases CO. If a significant amount of steam exists, gasification (R3) and water shift reaction (R5) occur and release H2. Reactions (R6) and (R7) involve volatiles. The volatiles are modeled to go through a thermal cracking (R6) and gasification processes (R7) via C6H6. The enthalpy of volatiles is calculated from the coal heating value and fully combustion of the carbon from the proximate analysis. The global reaction mechanism is modeled to involve the following chemical species: C, O2, N2, CO, CO2, H2O, H2 and volatiles (see reactions R1 through R7). All of the species are assumed to mix in the molecular level. In this study, the instantaneous gasification and equilibrium models will be established. The instantaneous model assumes that coal vaporizes very fast into gas without going through heterogeneous finite-rate reaction process. The equilibrium model will follow the conventional equilibrium concept by incorporating the equilibrium constants into the CFD simulation. The interphase exchange rates of mass, momentum and energy are assumed to be infinitely fast. Carbon particles are made to gasify instantaneously; therefore
Boundary Conditions The schematic diagram of investigated gasifier is shown in Fig. 2. The height of the whole gasifier is 12m, and the length of the horizontal section (1st stage) is 8m. The diameter of the horizontal section is 2m, and the one of the vertical section that includes a convergent-divergent section is 1.6m. There are two opposing inlets in the horizontal section with a height of 1m, while there is only one inlet in the vertical section with a height of 3.6m. The feedstock of the fuel is about 2550 ton/day with a 0.67 coal slurry concentration and a 0.91 O2/coal ratio. The inlet temperature of both coal slurry and oxidant is 425K, and the operating pressure is 28atm. The wall is assumed as no slip condition and adiabatic wall condition. The gridindependent studies have been tested in this study. However, even the mesh quantities grow to 2 million, the gridindependent is still not achieved. In order to save the computer resource, there are 1.05 million unstructured meshes used in the computational domain.
5
RESULTS AND DISCUSS
Figure 2
Baseline Case Figure 4 shows the velocity contours and vectors along xy plane at z=0 in the gasifier of the baseline case. The velocity vectors in each circle show the velocity field in a cuttingplane of the gasifier. The results show that the flow is injected strongly in the core of the 1st stage; however, there is backflow in the outer rim region of the bottom vessel. The four cutting-plane vectors of the bottom vessel show the backflow forms an annular recirculation region surrounding the jet core flow in the bottom cylinder on both ends. The slow-moving recirculation zones occupy a portion of the gasifier like blockages and reduce the effective area for the core flow to pass through. This leads to a longer residence time for the flow trapped in the recirculation zones but accelerates the core flow due to reduced effective throughflow area in the 1st stage. When the raw fuel gas produced in the 1st stage flows upward into the 2nd stage of the gasifier, it is accelerated through the converging throat region. Since there is an inlet feeding with coal slurry in this throat region, the flow field is deflected by the fuel jet and two separated-flow recirculation zones immediately form downstream of the second stage fuel jet locations. Each of these two recirculation zones contains a pair of counter-rotating vortices as shown in the third crosssectional velocity vector plot from top of Fig. 4. Due to the large include-angle of the divergent section immediately downstream of (or above) the throat, the recirculation zones sustain about three throat-diameters above the second fuel injection location and the flow reattaches to the wall at around 7 meter high of the gasifier. These recirculation zones increase the residence time of the trapped flow and provide well mixing between the unreacted coal slurry and the hot gases from the 1st stage. The lengthened residence time seems beneficial for allowing more thorough reactions, but the yield is low in this region because the produced syngas wastes time recirculating inside this region and does not effectively contribute to the syngas production rate at exit of the gasifier. Therefore, these recirculations actually exert a non-ideal condition because they lengthen only a small part of mass flow in these dead-flow regions and force the main-body flow move faster with less residence time by reducing the effective cross-section area of the flow passage. The accelerated main flow results in requiring a longer gasifier chamber to achieve satisfactory residence time. This uneven distribution of flow fields and residence times leads to the consideration of using multiple tangential jets in the second stage, which can provide a uniformly distributed spiraling flow field with evenly lengthened residence time. In this manner, the length of the gasifier chamber could be shortened and the cost reduced. Another option to reduce the recirculation zones volumes is to reduce the divergent section included angle by increasing the throat area.
Computational grid and boundary conditions
Grid Sensitivity study A grid sensitivity study is conducted by using four different grids: 120,000 meshes, 300,000 meshes, 680,000 meshes, and 105,000,000 meshes. Fig. 3 shows the temperature distribution along the gasifier height of the four different grids. In most part the temperature increases as the mesh number increases. The result of 680k grid almost coincides with that of 1.05M grid except in the region near the second fuel injection (height 4~6 m), where the largest temperature difference is about 40K or less than 3%. Considering the difference in most of the area is less than 1%, the 1.05M grid is then used in this study. 2300
Temperature (K)
2100 1900 1700 1500 1300 1100 0
2
4
6
8
10
12
Gasifier hight (m) 120k
300k
680k
1.05M
Figure 3 Grid sensitivity of vertical temperature distribution.
6
Velocity Field (m/s)
Figure 4 The velocity contours and vectors along x-y plane at z=0 in the gasifier of the baseline case.
7
Figure 5 shows the flow pathline in the gasifier. As the pathline diagram indicates, the gas injected into the gasifier from the opposing inlets in the horizontal section causes a strong upward flow in the middle of the 1st stage. Then most of the fluid turns an angle of 90 degrees and flows upward to the gasifier outlet. The residence time of the fluid from inlet to outlet ranges from 1.15 to 26.6 seconds. However, it can be seen that in the computational domain, the majority of the pathline colors are blue and turquoise ranging between 2 and 6 seconds, and few pathline showing longer residence time in green, yellow, and red. These are the fluid particles that have hung around recirculated for a couple of times before flowing to the outlet. Generally speaking, the residence time of an entrain-bed type gasifier is usually in the scale of seconds, and the result of the 2 – 6 seconds residence time seems to be reasonable. Figure 6 shows the 3-D temperature contours of the gasifier. Figure 7 shows the contours of temperature and mole fractions of CO2, CO and H2 along x-y plane in the gasifier. Combustion occurs when coal slurry and oxidant are injected into the first stage. The carbon and the oxygen react immediately and generate CO via R1 reaction. Then CO reacts with oxygen again to produce CO2 via R4 reaction. The oxidation and combustion reactions generate significant heat and the temperature in the middle of first stage is raised to over 2300K. Due to the feed of oxygen being insufficient for full combustion, some of the coal is consumed and converts into CO and H2 via gasification process by reacting with CO2 and H2O via R2 and R3 reactions. Because these two equations are endothermic reactions, the temperature in the exit of bottom vessel is lowered to around 2000K. When the hot raw syngas transports into the upper stage, it reacts with the coal slurry, which is fed from the upper jet. There is no additional oxidant being fed into the second stage, so the main reaction in this region is gasification. More CO and H2 are produced in this section. When syngas exits the gasifier, the temperature is cooled down to 1591K, due to endothermic gasification process and the mole fractions of H2 and CO are raised to 0.417 and 0.372, respectively. The exit temperature of E-gas gasifier is 1311K according to open literature [NETL, 2000]; however, the predicted value in the base case is 1591K. The difference between the simulation and the actual operation of E-gas gasifier is due mainly to the following causes: (1) The wall of gasifier is set to be adiabatic, which implies that there is no heat loss through the wall boundary; (2) The devolatilization process has not been implemented in this study; thus the energy needed to drive out the volatiles are not consumed. Due to the fast reaction in the eddy-dissipation model, there is no H2O remained in the exit syngas. The mole fraction of H2 in the simulation results is also higher than the counterpart documented in the NETL report (2000).
Pathline (Colored by time, sec)
Figure 5 Flow pathlines of the gasifier.
Temperature (K)
Figure 6
8
3-D temperature contours of the gasifier.
(a) Temperature (K)
(b) Mole fraction of CO2
(c) Mole fraction of CO
(d) Mole fraction of H2
Figure 7 The contours of temperature and mole fractions along x-y plane in the gasifier of baseline case.
9
Effects of Slurry Concentration Cases A1 to A5 in Table 4 are the cases showing the effects of slurry concentration of gasification process. In the coal slurry concentration series cases, the weight of feeding coal and oxidant are fixed; higher value of coal slurry concentration indicate less water in the slurry.
Table 5 The input conditions and simulation results showing effects of O2/coal ratio on gasification process.
Case A2
CaseA3
O2/Coal
0.91
0.91
Coal/Slurry
0.55
Temperature (K) Mole Fraction (%)
Case A1
Case A4
Case A5
0.91
0.91
0.91
0.6
0.67
0.75
0.8
1424
1494
1591
1675
1722
CO
25.0
30.7
37.2
45.4
50.1
H2
47.7
45.0
41.7
37.8
35.5
CO2
22.1
18.7
14.9
10.1
7.4
Baseline Case
Case B2
CaseA3 Baseline Case
Case B4
Case B5
O2/Coal
0.85
0.88
0.91
0.95
0.98
Coal/Slurry
0.67
0.67
0.67
0.67
0.67
Temperature 1448 (K)
1521
1591
1684
1750
CO
38.6
37.9
37.2
36.3
35.6
H2
42.1
41.9
41.7
41.5
41.3
CO2
13.2
14.1
14.9
16.0
16.9
Mole Fraction (%)
Table 4 The input conditions and simulation results showing effects of slurry concentration on gasification.
Case B1
Equations R1, R4 and R7 are the main exothermic reactions in the gasification process. When O2/coal ratio increases, there is more oxygen for R1, R4 and R7 reactions, and the temperature of the exit syngas becomes higher. The purpose of the gasification process is to reduce the combustion process and generate more syngas, i.e., CO and H2. When the input oxygen concentration is higher, there is more combustion reaction with less syngas production. However, the energy for devolatilizing and thermally cracking coal is from the combustion process. If there isn’t enough oxygen for the combustion process, there will be insufficient energy for gasification process, and the syngas production will be less. More oxygen is helpful for the combustion process and leads to higher temperatures in the gasifier. The exit temperature increases with elevated O2/coal ratio. In Case B1, due to the lack of oxygen input into the gasifier, the temperature is lower than in other cases. The temperature of exit syngas is 1448K in case B1 and 1750K in Case B5. The mole fractions of CO, H2 and CO2 are shown in Table 5. When there is more oxygen in the gasifier, there is also more CO2 generated from the combustion process via R4 reaction; meanwhile, the mole fraction of CO is less. The change of H2 from Case B1 to B5 is not distinct. The source of H2 is mainly from Char-H2O gasification. In the O2/Coal series cases, the feedstock of coal and water are fixed. The only changes in the series cases are oxygen inputs. So the change of H2 is not obvious in the O2/coal series cases.
The results in Table 4 show that when the value of coal slurry increases (i.e., less water), the mole fraction of CO increases, and on the contrary, the mole fraction of H2 becomes lower due to less water in the feedstock. However, when the value of coal slurry decreases (i.e., more water), the mole fraction of CO2 becomes higher, but the temperature decreases. When there is more water in the feedstock, more carbon reacts with H2O as well, which generates CO and H2 via R3 reaction. Then CO reacts with H2O and generates CO2 and H2 via R5 reaction. In the results, there are more R2, R3 and R5 reactions in lower slurry concentration cases comparing to higher concentration cases, especially more R3 and R5 reactions. Although carbon will react with CO2 to generate CO via R2 reaction, the concentration of carbon is less in low slurry case, there is also less carbon to react with CO2 via R2 reaction comparing with R3 and R5 reactions. This leads to more CO2, more H2 and less CO produced in the low slurry case. The results show that the temperature decreases with reduced slurry concentration. The temperature of exit syngas in Case A1 is 1424K and in Case A5 is 1722K. The equation R5 is an exothermic reaction, but there are more endothermic reactions, R2 and R3, which deplete more energy. This is the reason why the exit temperature of Case A1 is lower than that in other high slurry concentration cases. In summary, Table 4 shows lower slurry (or more water) produces more CO2 and provides the syngas with higher concentration of H2 and lower concentration of CO.
Comparison with the EPA data Table 6 shows that the operating parameters used in this paper include the operating conditions published in NETL (2000) and EPA (Nexant, 2006). The slurry concentration varies from 0.55 to 0.80, and the O2/Coal ratio varies from 0.85 to 0.98. It can be seen that the simulating parameters used in this paper are consistent with an actual working gasifiers.
Effects of Oxygen / Coal ratio Cases B1 to B5 in Table 5 show the results of the effects of O2/coal ratio. In the O2/coal series cases, the weight of feeding coal and water are fixed; higher value of O2/coal ratio indicates more oxidant in the feedstock.
10
higher slurry concentration is more preferable for producing more CO with higher syngas temperature and less CO2. Higher O2/coal ratio also leads to higher temperature. When the higher O2/coal ratio, there is more combustion reaction and generates more CO2. The effect of O2/coal ratio on H2 production is not obvious, especially for a high O2/coal ratio. The single jet injection in the second stage induces large recirculation regions, which introduces inefficiency and reduced syngas production by trapping a portion of the flow, reducing the effective flow passage area, speeding up the main flow, and decreasing the residence time of main flow. An alternate design of using multiple tangential jets injections in the second stage or enlarging the throat diameter can be considered. Although the simplified instantaneous gasification model is used in this study, the trends of simulated results of the coal combustion and gasification process in the cross-type, two-stage gasifier are reasonable. Future activities need to be considered in the finite-rate reaction mode, especially the water-shift reaction rate without catalysts, and investigate the effects of turbulent model on the gasification process.
Table 6 Operating parameters in the simulated cases
Coal(MF)/Slurry
0.85
0.88
0.55 0.60 0.64 0.67
x
0.75 0.80
x
O2/Coal(MAF) 0.91 0.95 x x x x(NETL)
x
0.98
x(EPA) x
x x
Table 7 shows the comparison of outlet species mole fraction with EPA results, which are computed by Aspen Plus. The major difference is the H2O mole fraction. The outlet species concentrations of H2 and CO2 in this study are higher than EPA results, while the values of CO and H2O in this study are lower than EPA results. And according to the R5 reaction, CO reacts with H2O and produces CO2 and H2. It can be judged that the major difference is caused by the water shift reaction rate of R5. It is clear that the water-shift reaction rate is too fast in the current model using the eddy-dissipation model. Implementation of a correct finite-rate model for water-shift reaction is essential for achieving a more accurate prediction. This will be left as a future task for this research.
ACKNOWLEDGEMENTS The authors are grateful to National Science Council of TAIWAN ROC for the financial support of the code number NSC 98-3114-Y-042A-005.
Table 7 The comparisons of simulated results with EPA's results.
Mole fraction (%) CO H2 CO2 H2O
Simulated results 32.2 42.8 19.1 0.0
REFERENCES BOE (Bureau of Energy), (2010). Energy Statistical Hand Book 2009. Bureau of Energy, R.O.C. (in Chinese). Bockelie, M.J., Denison, K.K., Chen, Z., Linjewile, T., Senior, C.L., Sarofim, A.F., (2002). CFD modeling for entrained flow gasifiers in vision 21 systems, Ninteenth International Pittsburgh Conference, Pittsgurgh, PA, USA. BP, (2009). BP Statistical Review of World Energy. BP. Chen, C., Horio, M. & Kojima, T., (2000a). Numerical simulation of entrained flow coal gasifiers. Part I: modeling of coal gasification in an entrained flow gasifier. Chem. Eng. Sci., 55, 3861-3874. Chen, C., Horio, M. & Kojima, T., (2000b). Numerical simulation of entrained flow coal gasifiers. Part II: effects of operating conditions on gasifier performance. Chem. Eng. Sci., 55, 3875-3883. Choi, Y. C., Li, X. Y., Park, T. J., Kim J. H. & Lee, J. H., (2001). Numerical study on the coal gasification characteristics in an entrained flow coal gasifier. Fuel, 80, 2193-2201. Destec Energy, Inc., (1996). The Wabash River Coal Gasification Repowering Project. TOPICAL REPORT NUMBER 7, NOVEMBER 1996. Eaton, A. M., Smoot, L. D., Hill S. C. & Eatough C. N., (1999). Components, formulations, solutions, evaluation and application of comprehensive combustion models. Prog. Energy Combust. Sci., 25, 387-436.
EPA’s results 42.3 30.7 9.6 14.9
CONCLUSION A three-dimensional computational model has been established to simulate the coal combustion and gasification in a cross-type two-stage gasifier. The instantaneous gasification model is adapted in this study to investigate the gasification processes. The instantaneous gasification model significantly reduces the computational time but can only provide a qualitative trend of gasification process for preliminary investigation. Although the instantaneous gasification model is simplified, it can adequately capture the global feature of the effect of the thermal-fluid field (including turbulence structure) on chemical reactions. The results show that when the value of slurry concentration (carbon to water ratio) increases, the temperature of exit syngas is also higher. When the slurry concentration is lower, there are more R2, R3 and R5 reactions and leads to a higher concentration of H2 and CO2; meanwhile, the concentration of CO and temperature is lower. Under the condition of feeding carbon being nearly completely converted, low slurry concentration is preferred over high concentration if more H2 is wanted with lower syngas temperature, while 11
Faeth, G.M., (1987). Mixing, Transport and Combustion in Sprays, Progress in Energy Combustion Science, Vol. 13, pp. 293-345. Garg, A., Kazunari, K., and Pulles, T., (2006). 2006 IPCC Guidelines for National Greenhouse gas Inventories, Chapter 1: Introduction. The Intergovernmental Panel on Climate Change (IPCC). Hill, S.C. and Smoot, L.D., (1993). A Comprehensive Three-Dimensional Model for Simulation of Combustion Systems: PCGC-3, Energy & Fuels, 7, 874-883. Hunt, J. C. R. & Savill, A. M., (2004). Guidelines and criteria for the use of turbulence models in complex flows. In Prediction of Turbulence Flows, eds. G. F. Hewitt and J. C. Vassilicos, Cambridge University Press, Cambridge Jones, W.P., and Lindstedt, R.P., (1988). Global Reaction Schemes for Hydrocarbon Combustion, Combustion and Flame, 73, 233-249. Launder, B. E. & Spalding, D. B., (1972). Lectures in mathematical models of turbulence. New York, Academic Press NETL, (2000). Destec Gasifier IGCC Base Cases, PEDIGCC-98-003. NETL, (2002). The Wabash River Coal Gasification Repowering Project: A DOE Assessment, DOE/NETL-2002/1164. Nexant, Inc., (2006). Environmental Footprints and Costs of Coal-Based Integrated Gasification Combined Cycle and Pulverized Coal Technologies, United States Environmental Protection Agency (EPA), EPA Contract No. 68-W-03-33, Work Assignment 2-02. Silaen, A. & Wang, T., (2005). Simulation of coal gasification inside a two-stage gasifier. TwentySecond International Pittsburgh Conference, Pittsgurgh, PA, USA. Silaen, A. & Wang, T., (2006). Effects of Fuel Injection Angles on Performance of A Two-Stage Coal Gasifier. Twenty-Third International Pittsburgh Conference, Pittsgurgh, PA, USA. Silaen, A. & Wang, T., (2009). Comparison of Instantaneous, Equilibrium, and Finite-rate Gasification Models in an Entrained-flow Coal Gasifier. Twenty-Sixth International Pittsburgh Conference, Pittsgurgh, PA, USA. Silaen, A. & Wang, T., (2010). Effect of turbulence and devolatilization models on coal gasification simulation in an entrained-flow gasifier. International Journal of Heat and Mass Transfer 53, 2074–2091. Smoot, L. D., (1984). Modeling of coal-combustion processes. Prog. Energy Combust. Sci., 10, 229-272 Smoot, L. D. & Smith, P. J., (1985). Coal combustion and gasification. Plenum Press, New York Vicente, W., Ochoa, S., Aguillon, J. & Barrios, E., (2003). An Eulerian model for the simulation of an entrained flow coal gasifier. Appl. Therm. Eng. 23, 1993-2008 Watanabe, H. & Otaka, M., (2006). Numerical simulation of coal gasification in entrained flow coal gasifier. Fuel, 85, 1935-1943 12
URANIUM AND SOME OTHER TRACE METAL ELEMENT CONCENTRATION OF SOME TURKISH COAL ASHES Isik Ozpeker, Dr., Fikret Suner, Dr., Mehmet Maral, MSc, Tahsin Aykan Kepekli, MSc. Istanbul Technical University, Dept. of Geological Engineering, Istanbul – Turkey [email protected], [email protected], [email protected], [email protected] ABSTRACT This study focuses on uranium and some other trace element concentrations and distributions of some coal occurrences that had been formed in Trachea and Anatolia, Turkey. Coal occurrences are in different age and rank. Most of them are young and their ranks are low. Ashes of coal samples have been picked up from different parts of Turkey and were analyzed and evaluated in terms of uranium and some trace element contents. Chemical investigations were performed on the coal ashes via fluorometric method for analyzing uranium concentration, some trace and major element concentrations were analyzed by AAS (Atomic Absorption Spectrometry) and FP (Flame Photometry) methods. The analyses results show that the uranium content in coal ashes change between 0 – 178 ppm, while the average of Turkish coals is 10 – 33 ppm. Ni, Co, Cu, Zn, Pb, Ag, Fe, Ca, Na and K concentrations were also detected. The mentioned trace element concentrations are over the world averages in most of coal ash samples. Uranium was enriched in the western Anatolia, especially in Mugla – Yatagan, Aydın – Soke, Kutahya – Gediz and Acemkiri coal fields. Also, an asphaltite sample from Sirnak includes noticeable amount of uranium concentration. Uranium accumulation of the coal samples probably depended on surrounding units as the source rocks. Keywords: low rank coal, ash, trace element, Anatolia
INTRODUCTION Coal is clearly important and Turkish lignite reserves are estimated to be over 8 billion metric tons beside 1.1 billion tons of hard coal reserves [1], [2]. Up to 15% of the total coal reserves in Turkey are of sub bituminous rank or higher [1]. Turkey has provided over half of electric generation from burning coal and natural gas [2]. Tons of coal ashes are produced every year. The trace elements' occurrence in coal is to be very important in terms of environmental, economic, by-product, and technological behavior of coals and their
1
effects on human health at present. Although, there are associated with the organic matter of the coal, usually their concentration in coal is related minerals, which show varieties both chemical composition and physical properties. While there are forming ash, their associated trace elements will be fallow [3]. The previous investigations have studied on trace elements in Turkish coals [4]. Furthermore, trace elements analyses were performed using instrumental neutron activation analysis [5] - [7] and recently, there has been more interest in trace elements in Turkish coal [8] - [14]. Our analysis and previous researches’ results were reevaluated and interpreted together. Therefore, samples were obtained from as many working mines as possible. The coals in Turkey are generally low rank (lignite or sub bituminous) formed in several different depositional environments at different geologic times and have differing chemical properties (Fig. 1). Eocene coals are limited to northern Turkey; Oligocene coals, found in the Thrace Basins of northwestern Turkey, are intercalated with marine sediments; Miocene coals are generally located in Western Turkey. The coal deposits, which have limnic characteristics, have relatively abundant reserves. Pliocene– Pleistocene coals are found in the eastern part of Turkey. Most of these coals have low calorific values, high moisture, and high ash contents [4].
Fig.1. Location map of the Turkish coal occurrences [15]
2
MATERIALS AND METHODS Different major coal occurrences were sampled and analyzed. Also previous researchers’ results were re-evaluated for trace elements. Chemical investigations were performed on the coal ashes via fluorometric method for analyzing uranium concentration, some trace and major element concentrations were analyzed by AAS (Atomic Absorption Spectrometry) and FP (Flame Photometry) methods.
RESULTS AND DISCUSSION Trace element contents of coals are related to their mineralogy and major elements [11], [16]. Element associations with coal ash have been widely reported, and information obtained on the degree of organic – inorganic association [8], [11], [16] - [20]. Significant positive correlations with ash content for As, Cd, Cr, Ni, Pb, Sb and U were recorded for the A1 lignite bed Paleocene, Calvert Bluff Formation of Wilcox Group, USA [11], [21]. Some trace element U, Ni, Pb, Zn, Cu and Co concentrations in ashes of the major coal fields are present in various concentrations (Table 1) in comparison with the range for most world coals and ashes [22], [23]. Table 1. Some trace elements analysis of the coal ash samples (*[23], **[24], +[25], ++ [26]) +
Sample No
Amasya-Suluova 17145 Ankara-Ayaş 16039 Ankara-Beypazarı-Nallıhan Ankara-Beypazarı Ankara-Beypazarı 7TKİ Aydın-Söke-Dedeman BüyükD Aydın-Söke-Dedeman KesmeliD Balıkesir-Gönen-Şaroluk 17617 Balıkesir-Gönen-Akas 12TKİ Bingöl-Karlıova 17110 Bingöl-Karlıova 17135 Bolu-Reşadiye Bolu-Kamil Bilgehan Bolu-Adasal No:1 Bursa-Orhaneli Bursa-Orhangazi Bursa-Keles Çanakkale-Çan 1
++
++
++
++
++
U ppm 9* 2 10 38 2
Ni ppm 76* 172 225 N/A 171
Co ppm 37* 56 11 N/A 17
Cu ppm 68* 102 172 N/A 17
Zn ppm 117* 216 101 N/A 30
Pb ppm 45* 68 29 N/A 5
10 136 15 14 N/A N/A N/A 0 40 4 3 N/A 17 N/A
310 165 127 90 52 49 117 101 N/A 223 N/A 102 N/A 68
6 38 23 33 19 6 21 2 N/A 30 N/A 16 N/A 56
64 80 72 2276 44 56 44 82 N/A 65 N/A 72 N/A 269
180 130 110 140 103 66 247 64 N/A 82 N/A 77 N/A 322
140 64 51 180 70 18 58 5 N/A 57 N/A 82 N/A 269
3
Çanakkale-Katar Ticaret Çanakkale-Örencik 6TKİ Çankırı-Ilgaz 17707 Çankırı-Orta 8TKİ Edirne-Keşan-Mandadere Edirne-Hacıömer 15940 Edirne-Hacıömer 15941 Erzurum-Oltu-Balkaya Isparta-Türk Civa TKİ276 İstanbul-Kilyos-Kısırkaya İstanbul-Ağaçlı İstanbul-Demirciköy İstanbul-Demirciköy Kastamonu-Azdavay Kastamonu-Azdavay Kastamonu-Azdavay (Artık) Kütahya-Gediz Büyük D. Kütahya-Acemkırı D. Manisa-Akhisar 16195 Manisa-Soma 16745 Manisa-Soma TKİ10 Manisa-Soma TKİ167 Maraş-Elbistan-Afşin 16195 Maraş-Elbistan-Afşin (gitya) Maraş-Elbistan-Afşin TKİ180 Muğla-Milas-Sekköy Muğla-Milas-Sekköy 17983 Muğla-Milas-Sekköy 17931 Muğla-Milas-Sekköy17999 Muğla-Milas-Sekköy17878 Muğla-Karakuyu-Yatağan Muğla-Karakuyu-Yatağan 17935 Muğla 17919 Muğla 17937 Muğla 17938 Muğla-Yatağan 17939 Sivas-Kangal 19808 Yozgat-Sorgun TKİ276 Zonguldak-İncirharmanı Zonguldak (Lavvar artık) Şırnak Asfaltit TKİ261 Yatagan** Soma** Seyitomer** Yenikoy** Catalagzi** Afsin-Elbistan** Tuncbilek**
5 1 2 6 0 N/A N/A 4 N/A 5 N/A 1 2 11 2 N/A 178 149 53 8 28 14 4 1 5 1 20 20 42 7 47 98 N/A N/A N/A 17 63 3 4 N/A 83 N/A N/A N/A N/A N/A N/A N/A
N/A 250 440 178 261 158 120 N/A 125 590 244 194 N/A N/A N/A 98 1302 N/A 105 89 103 N/A 200 N/A 170 N/A N/A 173 N/A 161 N/A N/A 151 189 151 113 144 77 125 69 2698 134.7 79.7 1975.9 79.6 118.5 119.4 1986.1
N/A 49 44 28 21 7 32 N/A 2 63 50 60 N/A N/A N/A 38 48 N/A 22 399 25 N/A 46 N/A 13 N/A N/A 26 N/A 23 N/A N/A 29 22 34 32 11 61 35 33 8 N/A N/A N/A N/A N/A N/A N/A
N/A 74 101 84 87 92 71 N/A 60 212 3360 368 N/A N/A N/A 198 247 N/A 52 79 26 N/A 87 N/A 40 N/A N/A 14 N/A 14 N/A N/A 50 86 80 58 42 40 1782 160 208 136.7 59.8 98.8 39.8 98.8 39.8 99.3
N/A 138 222 101 138 178 173 N/A 70 184 246 221 N/A N/A N/A 104 200 N/A 26 153 250 N/A 122 N/A 40 N/A N/A 42 N/A 72 N/A N/A 28 114 150 156 66 70 134 148 2800 404.1 195.2 112.6 71.6 138.3 79.6 137
N/A 60 86 12 60 66 52 N/A 80 138 180 250 N/A N/A N/A 88 60 N/A 2 124 32 N/A 78 N/A 14 N/A N/A 5 N/A 2 N/A N/A 2 52 30 20 5 82 176 100 36 57.7 79.7 79 39.8 118.5 79.6 99.3
The average U concentrations generally increased north western and south western of Anatolia. The U concentrations were insignificantly different from each other and low contents in Marmara and the Black Sea regions while moderate U concentrations were noticeable in Aegean and Central Anatolia regions. The concentration of Ni in the coal ash samples was high and changed between 49 – 1302 ppm. The average Ni values are
4
high in the South Marmara and western Anatolia. Analysis suggests that the central Anatolia has a significantly higher than does the eastern Anatolia. The coal ash samples generally have low and moderate Co contents. The maximum value of Co detected is 620 ppm. The average concentration of Co in Turkish coal ash is 27 ppm. The higher value of Cu detected was 3360 ppm. The average concentration of Cu in coal ash was 190 ppm. The average Cu values are high in the north western Anatolia. The average Zn concentrations changed from 30 to 322 ppm and, had insignificant changes in Anatolia and Trachea. Also Pb values showed similarities like Zn and changed between 2 – 269 ppm.
CONCLUSION The trace element concentration of coal ashes was higher, especially Ni and U values. Furthermore, Co was significant in Agean. Cu, Pb and Zn content of ash exhibited similarities. The mentioned elements were intense western Anatolia.
REFERENCES [1] Tuncali, E., Ciftci, B., Yavuz, N., Toprak, S., Koker, A., Acik, H., Gencer, Z., Tahin, N., 2002. Chemical and technological properties of Turkish Tertiary coals. General Directorate of Mineral Research & Exploration, Ankara, Turkey. [2] Lynch, R., 2003. An energy overview of the Republic of Turkey. U.S. Department of Energy. http://www.fe.doe.gov/international/turkover.html. [3] Davidson, R. M., 1996. Trace elements in coal. Energia 7, No. 3 [4] Palmer, C.A., Tuncali, E., Dennen, K.O., Coburn, T.C., Finkelman, R.B., 2004. Characterization of Turkish coals: a nationwide perspective. International Journal of Coal Geology 60, 85–115. [5] Ayanoglu, S.F., Gunduz, G. 1978a. Neutron activation analysis of Turkish coals; I. Elemental contents. Journal of Radioanalytical Chemistry 43, 155–157. [6] Ayanoglu, S.F., Gunduz, G. 1978b. Neutron activation analysis of Turkish coals; II. Analysis of ashes and the effects of burning condition on percent transference. Journal of Radioanalytical Chemistry 43, 159–164. [7] Ayanoglu, S.F., Gunduz, G. 1978c. Neutron activation analysis of Turkish coals; III. Relation between composition of coal and local earth crust. Journal of Radioanalytical Chemistry 43, 165–167.
5
[8] Karayigit, A. I., Akgun, F., Gayer, R. A., Temel, A. 1999. Quality, palynology, and paleoenvironmental interpretation of the Ilgin Lignite, Turkey. International Journal of Coal Geology 38, 219–236. [9] Karayigit, A. I., Gayer, R. A., Querol, X., Onocak, T. 2000a. Contents of major and trace elements in feed coals from Turkish coal-fired power plants. International Journal of Coal Geology 44, 169–184 [10] Karayigit, A. I., Spears, D. A., Booth, C. A. 2000b. Antimony and arsenic anomalies in the coal seams from the Gokler Coalfield, Gediz, Turkey. International Journal of Coal Geology 44, 1–17. [11] Karayigit, A. I., Spears, D. A., Booth, C. A. 2000c. Distribution of environmental sensitive trace elements in the Eocene Sorgun coals, Gediz, Turkey. International Journal of Coal Geology 42, 297–314. [12] Querol, X., Whateley, M. K. G., Fernandez-Turiel, J. L., Tuncali, E. 1997. Geochemical controls on the mineralogy and geochemistry of the Beypazari lignite, central Anatolia, Turkey. International Journal of Coal Geology 33, 225–271. [13] Querol, X., Alastuey, A., Plana, F. et al. 1999. Coal geology and coal quality of the Miocene Mugla basin, southwestern Anatolia, Turkey. International Journal of Coal Geology 41, 311–332. [14] Gurdal, G. 2008. Geochemistry of trace elements in Çan coal (Miocene), Çanakkale, Turkey. International Journal of Coal Geology 74, 28–40. [15] http://www.mta.gov.tr/v1.0/images/daire_baskanliklari/enerji/siteharitalar/4.jpg [16] Spears, D.A., Zheng, Y., 1999. Geochemistry and origin of elements in some UK coals. Int. J. Coal Geol. 38, 161–179. [17] Nicholls, G.D., 1968. The geochemistry of coal-bearing strata. In: Murchison, D.G., Westoll, T.S. _Eds.., Coal and Coal Bearing Strata. Oliver and Boyd, Edinburgh, pp. 269–307. [18] Goodarzi, F., 1988. Elemental distribution in coal seams at the Fording coal mine, British Columbia, Canada. Chemical Geology 68, 129–154. [19] Finkelman, R.B., 1994. Modes of occurrence of potentially hazardous elements in coal: levels of confidence. Fuel Processing Technology 39, 21–34. [20] Mukhopadhyay, P.K., Goodarzi, F., Crandlemire, A.L., Gillis, K.S., MacNeil, D.J., Smith, W.D., 1998. Comparison of coal composition and elemental distribution in selected seams of the Sydney and Stellarton Basins, Nova Scotia, Eastern Canada. Int. J. Coal Geol. 37, 113–141. [21] Crowley, S.S., Warwick, P.D., Ruppert, L.F., Pontolillo, J., 1997. The origin and distribution of HAPs elements in relation to maceral composition of the A1 lignite bed _Paleocene, Calvert Bluff Formation, Wilcox Group., Calvert mine area, east–central Texas. Int. J. Coal Geol. 34, 327–343. [22] Swaine, D.J., 1990. Trace Elements in Coal. Butterworths, London, 278 pp.
6
[23] Meij, R., 1995. The distribution of trace elements during the combustion of coal, Env. Aspects of Trace elements in coal, Ed: Swaine and Goodarzi. [24] Bayat, O., 1998. Characterisation of Turkish fly ashes. Fuel Vol. 77, pp. 10591066. [25] Ozpeker, I., Maytalman, D., Kural, O., Uranium concentration of some coal ashes. [26] Ozpeker, I., Maytalman, D., Trace element distribution in some coal ashes.
7
TRANSFORMATIONS OF KARAMAN -ERMENEK LIGNITES OF TURKEY UNDER ACCELERATED ELECTRONS IMPACT Islam Mustafayev* , Fethullah Chichek* , Guven Onal**
*Institute of Radiation Problems, Azerbaijan National Academy of Sciences. 31a H.Javid ave, Baku-Azerbaijan, [email protected] **Istanbul Technical University. Mining faculty. 80260 Ayazaga, Istanbul-Turkey Abstract. The regularities of transformation of lignites from Karaman-Ermenek deposits of Turkey under accelerated electron impact were studied. The absorbed doze in lignites changed within the limits of 1170-3120 kGy. As basic indexes of process rate of gas formation, decreasing of initial mass of lignite, the contents of sulfur in the solid have been defined. The gaseous products Н2, CO and СН4 were identified. The specific features of radiation-chemical decomposition of organic mass of lignite under accelerated electrons impact are discussed . Introduction. The study of radiation-chemical transformation of coal has great importance for creation of scientific basis of radiation-chemical technology of fossil fuels, establishment of the role of radiation in genesis and metamorphism of fossil fuels as well as determination of radiation stability of carbonaceous materials used in nuclear technology. Researches on radiation-chemical processes of transformation of fossil fuels have been spent for last 25 years in USA, Japan, Italy, Russia and other countries within the framework of the Programs on atomic-hydrogen energy and new methods of processing fossil fuels /1-3/. Radiation-stimulated processes in low-grade lignites of Turkey are not regularly investigated. The separate researches on radiation-chemical desulphurization and co- copyrolysis of lignites of Yatagan, Soma and Yenikoy has been conducted in the Istanbul Technical University and Dokuz Eylul University /3-5/. In this work the regularities of transformation of lignites from Karaman-Ermenek deposits of Turkey under accelerated electrons impact were studied.
EXPERIMENTAL The basic characteristics of the investigated samples are: ashes-19.26 %, moisture-15.6 %, sulphur-2.01 %, calorific value - 3775 kcal/kg. Processes were spent under the impact of accelerated electrons on linear electron accelerator (ELU-6) with power of electron beam P= 12 20 W. The rate of the absorbed doze was determined by combined method-cylinder of Faraday
and Calorimeter and equaled to 4680- 7800 kGy/h. The absorbed dose in lignites changed within the range of 1170-3120 kGy. Irradiation was carried out in semiflowing installation, gas products have been collected in gasometer and liquid- in receiver. As basic parameters of process rate of gas formation, losses of initial mass of lignite, the contents of sulfur in solid have been defined. The gases of Н2, CO and СН4 were identified by chromatography. Decrease of lignite organic weight was determined by gravimetric method. Definition of the contents of sulphur is based on spectroscopic determination of sulphur dioxide-products of oxidation of sulfur.
RESULTS AND DISCUSSION In figure 1.kinetics of formation of gases Н2, CO and СН4 at the radiation-chemical
x 1018 molec/gram
Concentration of gases,
decomposition of lignite sample was illustated.
16 14 12 10 8 6 4 2 0
H2 CO CH4
0
10
20
30
40
50
T, min
Fig 1.Kinetics of formation of gas Н2, СО, СН4 at decomposition of lignite under accelerated electron impacts The formation rate (W, 1015 molec/g sec) and radiation–chemical yield (G, molec/100 eV) of gases was determined. In our experiments these parameters are: W (H2) =3.6, W (CO) = 5.2, W (CH4) =0.98, radiation-chemical yield of these gases are: G (H2) =0.172, G (CO) =0.25 and G (CH4) =0.047. It should be mentioned that these values exceeds radiation-chemical yield of gases in case of γ-radiolysis. On table 1. the values of radiation-chemical yield of gases Н2, CO and СН4 from lignites of various deposit of Turkey at the γ-radiolysis were presented.
Table 1. Radiation-chemical yields (Gx103 molec/100 eV) of gases at the γ-
of
Turkish lignites
Lignite type
Н2
СО
СН4
С2Н6
С2H4
C3
C4
C5
Karaman –
24
-
0.7
-
-
-
-
-
22
-
1.7
-
-
-
-
-
Nevshehir, raw 24
-
4.9
2
-
-
-
-
21.7
0.7
0.8
1.0
0.6
0.3
-
-
-
-
Ermenek, cleaned Karaman – Ermenek, raw
Silopi, raw
62
Trakya, mixed
19
4
1.0
The radiation-chemical yield of these gases at the γ-radiolysis are changed: G (H2) =0.019-0.062, G (CO) =0.004 and G (CH4) =0.001-0.021 molec/100 eV. Comparison shows that in case of irradiation by accelerated electrons the radiation-chemical yield of gases 3-9 times for hydrogen, 62 times for CO, 2-47 times for СН4 exceeds the values in case of γ-radiolysis. The observable difference in the yield of gases at the impact of γ- radiation and accelerated electrons can be connected by the influence of high value of dose rate which can lead to qualitatively new processes, or change of physical condition (temperature, pressure, etc) of reaction zone. In our early researches /3-6 / it has been shown that hydrogen and methane from lignites are formed due to recombination reactions of radiolytic radicals of Н, СН3: M ® CH3, H, CO, M +, M* H+H ® H2 CH3+H ® CH4 and etc Carbon monoxide is formed at radiation splitting of oxygen-containing functional groups. Change of recombination mechanism to reaction of separation in the gas formation demands temperature more 200С which in our case is not observed. Small increase in gases yield with activation energy Е = 10-15 кJ/mole is possible due to increase of diffusion of particles in reactionary volume. It means that active particles formed at radiolysis can be stabilized in defective structure of lignite mass. Rise in temperature even above 500С can lead to simplification of their diffusion and by that increase in recombination rate of radical products with formation H2 and CH4. At braking accelerated electrons in the sample volume local rise in
temperature and increase in the yield of products may be observed. Such rise in temperature does not happen under the action of γ- radiation and yields of gases considerably decrease under the influence of accelerated electrons. The course of kinetic curves at rather high doses (increase of inclination) also testifies local heating of lignite mass. The dependence of content of volatile products in structure of lignite at absorbed doze at irradiation in the range of doze D=1170-3120 kGy was illustrated in fig. 2. The lignite mass during irradiation decreases up to 76.5 %. During irradiation volatile products decrease correspondingly from 35 % to 11.5 %, so approximately 23.5 % of volatile products are allocated in the form of gases and liquid. In the same figure laws of change of sulphur relative content in solid product, concerning initial value were shown too.
120
Content, %
100 80 Organic mass, %
60
Sulphur, %
40 20 0 0
10
20
30
40
50
T, min
Fig.2. Dependence of relative changes of lignite mass and sulphur content at the irradiation by electron beam It should be noted that rather deep decomposition of lignites at thermal decomposition below 400С temperatures is impossible. The observable phenomenon is connected with destructive effect of ionizing radiation. Material balance of this process it is resulted in the table 2.
Table 2. Material balance of process of lignite radiation-thermal decomposition (in mass %)
Initial mass
Solid product
Liquid product
Gases
100
74-78
20-24
1-2
It is evident that in semiflowing installation the products of reactions basically are in liquid phase. The changes of absolute content of sulphur in lignite sample at absorbed dose of accelerated electrons were adduced in fig.3. The contents of sulphur in solid decreases from 2.01 % up to 1.62 % in the course of irradiation of 40 minute in W=12 W .
S, abso lute, %
2,5 2 1,5 1 0,5 0 0
10
20
30
40
50
T, min
Fig.3. Influence of irradiation (absorbed doze) on absolute content of sulphur in solid product of radiation-chemical decomposition of lignite
If we consider that there is also decomposition of organic mass of lignite a degree of relative reduction the contents of sulfur reaches up to 57.1 % (fig 2-3). It shows that in spite of decays up to 23.5 % of organic weight, and thus leaves up to 43 % of sulphur. It testifies about selective desulphurization of lignite mass under electron beam impact. Last circumstance is connected with selective cleaning of sulphur from organic mass of lignites, which is connected with highest electronic density of atoms of sulphur than atoms of carbon and hydrogen. Radiation action on chemical bounds with more density of electrons is more effectively. The integrated influence of radiation and temperature can lead to increase of selectivity of radiation desulphurization due to stimulation of separation processes with participation of sulphur-content groups. Thus, at radiation-chemical decomposition of lignites under the impact of accelerated electrons a number of the specific features differing influence of γ-radiation was shown: 1)
At braking electron there is local heating of lignite mass, that stimulates diffusion processes in gas formation
2)
At pulse impact of accelerated electrons with frequency 50 Hz and duration of the impulse 2.5 sec happens high-speed (104-105 degree/s) heating of lignite mass. It leads to deep decomposition of organic mass with formation of liquid products.
These features of influence of accelerated electrons on lignites can be used by development of processes of radiation-chemical technology of processing fossil fuels.
REFERENCES 1. D. D’Anjou, R. Litman. Radiochemical and Radioanalytical letters, 50, No1, (1983), p.37 2. Yermakov A., Popov V.N., Dzantiyev V. Radiation-chemical influence of irradiation on thermoradiation processes of coal gasification/ Chemistry of High Energy, 1988, v.22, p.132136. 3. I.Mustafayev, H.Hajiyev, K.Yagubov, B.Dzantiyev. Hydrogen and Hydrogen-containing gas production from fossil fuels by thermoradiation. Proc. of the 7-th World Hydrogen Energy Conference. Moscow, USSR, 25 - 29 September, 1988. v - 2, p. 955 - 965 4.I.Mustafayev, A.Yamik, H.Mahmudov. Coal Desulphurization by Electron Beam. 5-th Inter. Mineral Processing Symposium Capadociya, Turkey, 6 - 8 September, 1994, Progress in Mineral Processing Technology Ed. H.Demirel , p. 343 - 346 5. I.Mustafayev, M.Kemal, Y.Tolgonay, A.Yamik. Radiation-stimulated Processes of Destruction and Polycondencation in Coals. J. Radioanalitical and Nuclear chemistry letter, 1994, 187 (5), 355 – 365 6. Mustafayev, F.Chichek. Gas Formation at Radiation-Chemical Decomposition of Turkish Lignites. XXIII International Mineral Processing Congress,.3-8 September, 2006, Proceedings V. 3, p.2554-2558, Promedadvertsing-2006.
Desulfurization and Kinetics of Removal of Sulfur from High Sulfur Coal under Hydrogen Atmosphere Guojie Zhang, Yongfa Zhang* , Fengbo Guo, Bingmo Zhang Key Laboratory of Coal Science and Technology of Shanxi Province and the Ministry of Education ,Taiyuan University of Technology , Taiyuan 030024 , China Abstract: The reaction between hydrogen and sulfur in high sulfur coal at high temperature was studied in this paper. Crashed and sieved high sulfur coal sample (with particle size of 0.6mm) was placed in batches in 23 mm I·D· differential reactor. The release of hydrogen sulfide at run temperature and under different hydroatmospheres was followed by a hydrogen sulfide detector. The desulfurization yield was obtained from elemental analysis of residual char. The hydrogen can greatly promote the effect of desulfurization and more than 65% sulfur in the coal can be removed. The releasing curves of H2S in hydropyrolysis process obviously showed two peaks. The desulfurization process in hydropyrolysis of high sulfur coal can be regarded as in two stages according to the evolution profiles of H2S. The first peak at 250~450 ℃ was from the desulfurization of aliphatic sulfide and the second peak at 450~650℃could be from both the sulfur in pyrite and aromatic thiophenic structure. Results show that the desulfurization of high sulfur coal could be described much better with the grain reaction model than with the random pore model. The random pore model is only adapted to the initial stage of sulfur removal from high sulfur coal under hydrogen atmosphere while the grain reaction model is adequate the whole stage. Key words:high sulfur coal, hydropyrolysis desulfurization, kinetics, grain reaction model, random pore model
1. Introduction China is the world's largest coal producer and consumer. Coal accounts for 75% of primary energy in China, high sulfur coal reserve is about 30% of total coal reserves,and the proportion of high sulfur coal mining has also been increasing year by year
[1]
. Along with the rapid
development of economy in China, a large amount of SO2 and NOx emissions owing to coal consumption in primary energy, have resulted in severe environmental pollution. It not only will result in the formation of photochemistry smoke, acid rain and the human body breath endanger, but also will destroy the ecological environment, as well as bring the serious influence on the mankind's life and production. According to the resources and the economic base of our country, in next 50 years, the primary energy sources are coal-based in china [1,2]. So, developing a cheap, easy operation coal desulfurization technology will have far-reaching economic and environmental significance. Many desulfurization for coal before combustion methods have been proposed [1]. But so far, an economical and effective method of desulfurization has not been found. Except liquefied *Corresponding author. Tel. : +86 0351 6018676 E-mail address: [email protected]
pyrolysis and gasification conversion, hydropyrolysis (Hypy) is the third ways conversion method, Hypy has got a great concern as a safe and clean alternative [3]. In this paper, hydrogen pretreatment on the impact of coal pyrolysis desulfurization and kinetics of removal of sulfur from high sulfur coal under hydrogen atmosphere at high temperature was studied.
2. Desulfurization reaction model Sulfur compounds in coal are divided into two categories: inorganic sulfur and organic sulfur. Inorganic sulfur is mainly pyrite sulfur and sulfate sulfur; organic sulfur includes aliphatic sulfur, aromatic and thiophene sulfur in coal. The reaction of sulfur compounds is very complexity in coal hydropyrolysis, The results were obtained by using the first order reaction model to treat the various parts of experimental data with different conversion levels [4]. (Ai-S)s(solid) + H2(gas) = (Ai) s(solid) + H2S(gas)
(1)
Taking into account structure characteristics of high-sulfur coal particles, the grain reaction model and the random pore model are used to define the mechanism of high sulfur coal desulfurization. [5-7] According to grain reaction model, the solid particles are made of a large number of tiny crystals. The diameter of tiny crystal is about nm, the particle size of 10-6 ~ 10-7. These crystals may be spherical, cylindrical or sheet of non-porous material. Anywhere these materials contacting with the gas reactants, the reaction cannot be completed immediately. In addition, the particles can also be seen as a pile of ceramic aggregate, closely together. The gas diffusion in small aggregate diameter is belonged to Knudsen`s type, which is much slower than the gas diffusion in the main hole. Although they are not absolute non-porous material, but have similar reaction behavior of non-porous material. The reaction step of the grain model is including the gas reactant diffusion from outer to inner, particle surface reaction, the diffusion of product gas from inside to outer. After block of coal grinded, the pore structure of coal particles destroyed, the coal particles make up a large number of small spherical dense granules. The reaction of each particle is subject to unreacted core model. The continuity equation of solid coal expressed by unreacted core model of a single particle is followed as:
∂rgc ∂t
Deg rgc ks C A
=
Deg rgc + ks CSgo (1-
rgc rgo
(2)
)
The boundary conditions can be expressed as rgc
t=0
(3)
=r g
The conversion scores of a single particle can be expressed as rgc 3 x =1-( ) , g rgo
Simultaneously
υS (or υA )=-
∂Cs ∂t
1
r =r (1-x ) gc go g
3
(4)
C Sg = C So (1 − xg ) =-Cso
∂ (1-x g ) ∂t
=Cso
∂x g ∂t
where 2
k s C A g C S O (1 -x g ) 3 3 υS = • rg o 1 + δ 2 [1 -(1 -x ) 1 3 ](1 -x ) 1 3 g g g ∂x g ∂t
2
=
k s C A g (1 -x g ) 3 3 • rg o 1 + δ 2 [1 -(1 -x ) 1 3 ](1 -x ) 1 3 g g g
where δ g2 =
k s rgo Deg
(5)
,which can be seen as the modulus of particles reaction.
Assumption coal particle used size uniformity in experiments, and material layer is very thin ( δ = 0 ). Simultaneously, the size of coal particle is fine, internal mass transfer resistance of 2 g
single-particle can be negligible. The reaction rate of experiment coal samples can be expressed as: ∂x g ∂t
=
3k sC Ago r
go
2 (1-x ) 3 g
(6)
Let be β = t
3ksCAgo
(7)
rgo
where 1 1-(1-x) 3 =β t t s
(8)
Random pore model assumes that the fine grinding of coal particles have a porous structure, the desulfurization reaction meets the assumptions of the random pore model. The desulfurization rate of a single particle can be expressed as: E − dx R rT n =k o e g(PH )f(x)(1-x) dt 2
(9)
For this experiment, g(PH2),f(x) can be considered constant, let be n=1, the integral equation: (10)
ln(1-x)=-β t 2
where β =k e 2 0
E R rT
g(P )f(x) H2
(11)
Because coal particles used in experiment are in the same conditions, similarly, the desulfurization reaction meets the assumptions of the grain model. The desulfurization conversion can be expressed by equation (10).
3. Experimental The apparatus is shown in Figure 1. The reactor was a stainless steel pipe, mounted in a vertical tube furnace which was controlled on the basis of bed temperature. The product gases were analyzed by gas Chromatography (GC)-950 with a FPD detector.. The gas (in coke oven gas) intake can be measured using mass flow meter. Furthermore, the output of product gas flow can be measured using soap liquid meter.
The reactor was brought up to temperature with nitrogen purge. Hydrogen was introduced into the system at the start of the run. Adjustments were then made to bring the flow to the desired levels. The gas After the run, the reactor was disassembled; the product char was weighed, and analyzed by means of KZDL--3C sulfur determinator and VG Scientific ESCALab220i-XL Electron Spectrometer for residual sulfur. All experiments with larger errors in repeated experiment were rejected. The proximate and ultimate analysis of the coal sample is presented in Table 1. Table.1 The proximate and ultimate analysis of the coal sample Proximate analysis/ %(mass) ,ar
Ultimate analysis/ %(mass) ,daf
Mad
Aad
Vad
Cad
Had
Nad
Sad
Oad*
0.92
22.48
29.84
70.02
4.38
0.95
2.92
21.73
*by difference
The formula of desulfurization rate (SR) of coal was as followed:
SR =
( M • S0 − m • s) × 100% M • S0
Fig.1
High sulfur coal desulfurization system
4. Results and discussion 4.1 Effects of flow on desulphurization rate
Fig .2 Effects of flow on desulphurization rate(Heating rate:15℃/min, T: 550℃) The results are presented in Figure 2 as per cent of the desulfurization rate after treatment vs. hydrogen flow. Inspection reveals that desulfurization rate increases as the hydrogen flow increasing. The 67.6 % of the sulfur is removed at 650ml/min of hydrogen flow. When the
hydrogen flow is continued to increase, desulfurization rate would decrease. At the beginning, desulfurization rate increasing, which are mainly due to hydrogen flows increasing to external diffusion power increase, coal macromolecular hydrogenation reaction probability increased. At the same time, the gas diffusion rate of volatile is also accelerated from coal particles interior to the surface, the secondary reaction of Sulfide gas and residue is reduced. Then low level of desulfurization rate may be a result of "air holes" caused by large hydrogen flow (>650ml/min) [8]. As the reaction temperature increased, the coal began to soften to form molten state. The flow of feed gas increases to a certain value, which is through the molten resin layer, the deflection is produced to form many small "air holes". Then the reaction gas is directly outflow from the "air holes", as a result, the sulfur in coal cannot fully contact reaction with hydrogen, some of the sulfur radicals further polymerize to form stable and difficult removal of thiophene compounds.
Relative volume,cm3/g
Figures 3 shows pore size distribution of Hypy char sample at 550℃. 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 -0.01 0
300ml/min 600ml/min 900ml/min
20
40 60 80 Raduis,pore rangs,nm
100
Fig.3 porous ranges of char sample after conversion Can be seen from figure 3, the porous amount of char residue increased with gas flow increased, decreased after gas flow exceeding a certain number of values. The results show that as the gas flow rate is less than a certain value, the reaction of hydrogen and coal samples is more fully in the Hypy process, the pore inducing effect is better. When the hydrogen flow is more than the certain value, hydrogen and coal samples cannot have full access to reaction caused by the deflection forming "air holes". Then the reaction gas is directly outflow from the "air holes", as a result, the sulfur in coal cannot fully contact reaction with hydrogen, the micropore volume of residues is significantly reduced.
4.2 H2S concentration in the gas under Hydropyrolysis
Fig.4 H2S concentration in the gas under Hydropyrolysis ( Q: 60ml/min T: 750℃)
As can be seen from Figure 4, the escape curve of hydrogen sulfide showed two main peaks: 250~450℃low-temperature peak; 450~650℃ high-temperature peak. Hydrogen sulfide formation reaction has three types in temperature range 250 to 450℃: (1) Pyrite hydrogenation reaction; (2) Aliphatic sulfur side chain fracture reaction; (3) Thiophene heterocycles hydrocracking reaction. Pyrite hydrogenation reaction was complete under the temperature of less than 450℃. In these three reactions, aliphatic sulfur side chain fracture was most complex, however, reaction temperature is low, aliphatic sulfur side chain fracture was completely reduced or decomposed under the temperature of less than 400℃. The contribution of thiophene heterocycles hydrocracking reaction was least compared to pyrite hydrogenation and aliphatic sulfur side chain fracture reaction. Hydrogen sulfide formation reaction was mainly from thiophene sulfur and FeS sulfur desulfurization in temperature range 450 to 650℃. The reduction temperature of thiophene sulfur was high, more than 450℃. As a result, increase of reaction temperature on hydrogenation pyrolysis desulfurization reaction was favorable. However, the reaction of organic sulfur cannot be completely in temperature range 450 to 650℃. FeS hydropyrolysis reaction is very difficult under the temperature of less than 800℃. The contribution of high peak is mainly thiophene sulfur reduced. a
b
c
e
d
f
(a)450℃Hypy char; (b)550℃Hypy char;(c)650℃Hypy char; (d)750℃Hypy char (e) Pyrolysis char; (f) coal sample Fig.5 XPS spectra of sulfur element Characteristic peak of sulfide is shown in Table 2. From the XPS spectra of the sulfur element (Fig. 5), it can be seen that desulfurization rate in hydrogen atmosphere was higher than direct pyrolysis the desulfurization. With the temperature increasing, the desulfurization rate
increased, then declined again, which is due to the presence of mineral matter in coal []. Comparison table 2 and Figure 5 can be seen, coal sample contains large amounts of inorganic sulfur (Fig. 5 (f)). Compared Pyrolysis char (Fig. 5 (f)) to coal sample (Fig. 5 (e)), Sulfur content in pyrolysis char residue obviously decreased. Under 450℃, there exists still a small amount of inorganic sulfur and organic sulfur in Hypy char. However, compared with the pyrolysis char, the sulfur content in char is lowered (Fig. 5 (a)). When the temperature increased to about 550℃ ~650℃,
The inorganic sulfur of residual char is almost removed completely, the amount of
organic sulfur is also removed (Fig. 5 (b,c)). However, the temperature is higher than 650 ℃ with sufficient hydrogen, the inorganic sulfur in residues increases caused by the minerals in coal [7]. Table 2 Characteristic peak of sulfide [9,10] Characteristic peak, eV
Sulfide
162.7~163.0
Sulfide sulfur, Sulfide, mercaptans or benzene disulfide and series Organic sulfur compounds(Non-aromatic sulfur,including alkyl -, aryl - sulfide and polysulfides) Thiophene (aromatic sulfur) Sulfoxide, Sulfone Sulfate sulfur Inorganic sulfur
163.1~163.5 164.1~164.5 169.8~166.2 167.8~166.2 169.1~169.5 > 170
4.3 Reaction model validation
Fig.6 Effect of temperature on sulfur removal in coal In the hydrogen conditions, desulfurization experimental data was shown in figure 6, under the temperature rang 300~700℃. Can be seen from Figure 2, in the early stages of reaction, increasing temperature, desulfurization rate was significantly increased, sulfur in the coal seemed easy to remove. However, sulfur content of residual char of was used to analyze, it was found that the desulfurization ratio in the 500 ℃ temperature was better than 700℃. This was mainly because that sulfur of coal easily was converted to a stable heterocyclic compounds under high temperature, such as thiophene forming by organic sulfur being very difficult to decompose [11], which is an endothermic reaction. Therefore, increasing reaction temperature was not only enhancing desulfurization rate, but also promoted organic sulfur conversion between different forms, made desulfurization rate decline. Figure 6 data was treated in accordance to the equation (8) and (10), the results shown in
Figure 7. Can see from the figures, the random pore model could be better desulfurization Kinetics than the grain reaction model. Early in the desulfurization, the random pore model data was fairly consistent with experimental data. While the declination of the particle reaction model and experimental data was relatively large. In reaction latter, the two models and were better consistent with experimental data. The random pore model was always consistent with the experimental data in total reaction process. 1.2
0
1.2
0
随机孔模型 0.05
1
粒子模型
1
0.002
粒子模型
0.2
0.4
0.25
0.2
0.3
0
0.8
0.004
0.6
0.006
0.4
0.008
0.2
0.01
0.35 50
100
150
200
0
250
0.012
0
t/min
50
100 150 t/min
(a) 0.01
0.008 0.007
0
7
0.1
6
随机孔模型
5
粒子模型
0.005
0.3
0.004 0.4
0.003 0.002
随机孔模型
0.001
粒子模型
0
100
150 200 t/min
250
Ln(1‐x)
Ln(1‐x)
0.006
1‐(1‐x)(1|3)
0.2
50
0 0.1 0.2 0.3 0.4
4
0.5 3
0.6 0.7
2
0.8
0.5
1
0.6
0
0.9 1 0
300
30
60
90
120
150
t/min
(b)
(d)
8
0
随机孔模型
7
0.2
粒子模型
6
0.4
5
Ln(1‐x)
250
(b)
0.009
0
200
1‐(1‐x)(1|3)
0
1‐(1‐x)(1|3)
0.15
0.6
4 0.8
3
1‐(1‐x)(1|3)
0.6
随机孔模型
Ln(1‐x)
Ln(1‐x)
0.8
1‐(1‐x)(1|3)
0.1
1
2
1.2
1 0
1.4 0
30
60
90
t/min
120
150
(e) Fig.7 grain reaction and random pore model (a) 300℃; (b) 4000℃; (c) 500℃; (d) 500℃(e) 600℃; (f) 300~600℃
5、Kinetics In order to investigate the desulfurization activation energy of high-sulfur coal, the followed methods were used to obtain activation energy in this paper, the advantage of this method is not considering the reaction mechanism change.
dx = kF ( x) dt
(12)
Let integral tx
∫0
dt =
dx
x
∫0 kF(x)
k = A exp −
E RT
Bexp
So t x =
A
Where: B=
∫
x
0
E R T ,as Lnt = E + b x
RT
dx B ,b=Ln A F ( x)
9
9.5
8.5
9
8
8.5
Lntx
Lnt x
7.5 7
8
7.5
6.5 7
6
x=0.1, Ea=34.16KJ.moll‐1
5.5
6
5 0.0007
x=0.3, Ea=28.58KJ.moll‐1
6.5
0.0012
0.0017
0.001
0.0022
0.0012
0.0014
10.5
9.5
10
9
9.5
Lnt x
Lntx
10
8.5
9
8
x=0.5, Ea=23.31KJ.moll‐1
x=0.7, Ea=20.75KJ.moll‐1
7.5
7 0.001
0.0018
8.5
8 7.5
0.0016
T‐1/K‐1
T‐1 /K‐1
0.0012
0.0014
T‐1 /K‐1
0.0016
0.0018
0.001
0.0012
0.0014
0.0016
0.0018
T‐1 /K‐1
Fig.8 Calculation of activation energy at designated reaction fraction Figure 8 was shown that the desulfurization activation energy of coal was low at hydrogen atmosphere. Overall, the sulphur of Goudi high-sulfur coal was easy to removal. Yergey 与文献 18 has reported the same results. With increasing of the reaction fraction(x), the activation energy was lower. Increasing reaction temperature was beneficial to desulfurization of coal hydropyrolysis. At the same time, as the reaction temperature increased, the desulfurization process changed: the inorganic sulfur removal was gradually changed into organic sulfur removal. The desulfurization activation energy of organic sulfur is relatively small [7].
6. Conclusion
1)The desulfurization rate of high sulfur coal is related with temperature. Because there is the conversion of organic sulfur forms in coal under high temperature, the higher reaction temperature is not more conducive to the removal of sulfur in coal. Suitable desulfurization temperature is the range from 500 to 600℃. When desulfurization reaction is controlled by chemical reaction, the desulfurization activation energy is in the range 20 ~ 40 kJ.mol – 1. 2) The hydrogen sulfide releasing curves of in Hypy process show two peaks: The first peak at 250~450℃, which is mainly due to pyrite hydrogenation and aliphatic sulfur side chain fracture reaction; and the second peak at 450~650℃, which is mainly due to thiophene sulfur reduced 3) Compared the grain reaction model with the random pore model, the random pore model is more consistent with experimental data: Early, the random pore model data was fairly consistent with experimental data, and the grain reaction model was relatively deviation large with experimental data; Latter, the two models and were better consistent with experimental data. The random pore model was always consistent with the experimental data in the total reaction process.
Acknowledgment Project supported by the National Basic Research Prograin of China (Grant No.2005CB221202), Innovation Team Funds Scheme, the Natural Science Foundation of China (Grant No. 21006066) and the youth found of School Funds of Taiyuan University of Technology (TYUT).
References [1] Tian Zhengshan, WangQuankun, Bai Suzhen. Present situation and future progress of combusting desulfurization technology in high sulfur coal. Chemical Industry Times, 2009, 23(7): 53-56. [2] Xu qiang, Qian jin. Investigation on development and application of clean coal technologies in China. Coal Ash China, 2003, 4: 35-39. [3] Fu Zhixin, Guo Zhancheng, Wang Shen-xiang. Desulfurization of high sulfur coal by recycling additional gases during coking process. The Chinese Journal of Process Engineering, 2007, 7(5): 910-915. [4] Chen Haohan, Li Baoqing, Yang Jili, et al. Study of Desulphurization and Its Kinetics in Hydropyolysis of Yanzhou Bituminous Coal. Journal of Fuel Chemistry and Technology.1997, 25(2):114-117 [5] Guo Hanxian. Applied Chemical Engineering Kinetics.Beijing: Chemical Industry Press, 2003 [6] Levenspiel O. Chemical Reaction Engineering.New York: John Willey & Sons Inc.1999.,570-577 [7] LIU Junli , TANG Huiqing, GUO Zhancheng, Kinetica of removal of organicsulfur from semicoke under hydrogen atmosphere. Journal of chemical Industry and Engineering (China), 2004, 155(8): 1335-1440. [8] Zhang Guojie, Zhang Yongfa, Xie Kechang. Study on desulphurization of high sulfur coal in hydropyrolysis. Chemical Engineering (China), 2006, l 34(4): 55-58. [9] Chen Peng. Application of XPS in study forms of organic sulfur in macerals of Yanzhou coal. Journal of Fuel Chemistry and technology, 1997, 25(3): 238-241. [10] Zhu Zhenping, Cui Hong, Li Yunmei, et al. investigation of sulfur and nitrogen forms in two high sulfur coals. Environment Chemistry, 1995, 14(6): 483-488. [11] Li Bin, CaoYan, Zhang Jianmin, et al. The progress of release from coal pyrolysis and gasification progress. Coal Conversion, 2001, 24(3): 6-11. [12] Yergey A L,Lampe F W,Vestal M L et al. Nonisothermal kinetics Studies of the Hydrodesulfurization of Coal. Ind Eng Chem Process Des Develop,1974,13(3):233-240 [13] Cypres R,Mingels W,Lardlnois J P,et al. Feasibility Study of the Hydropyrolysis of Coal,UER 14110[R]. Luxembourg. Commission of the European Communites,1992.
2010 International Pittsburgh Coal Conference Istanbul, Turkey, October 11-14, 2010 PREPARATION OF ALTERNATIVE FUEL FROM COMPOST AND COAL SLURRIES ZAJONC Ondrej, RACLAVSKY Konstantin, SKROBANKOVA Hana, JUCHELKOVA Dagmar, RACLAVSKA Helena VSB – Technical University Ostrava, Czech Republic, [email protected]
Introduction New strategies in municipal solid waste (MSW) management, i.e., a separate collection of the organic fraction (EU Directive 1999/31/EC, EU Directive 2008/98/EC) and a reduction of the biodegradable MSW fraction allocated in landfills (EU Directive 2003/33/EC), have favored the development of composting as a useful biotechnology in transforming organic wastes into suitable agricultural products. In order to meet the requirements of the Directive in the Czech Republic in by the year 2013 it will be allowed only 697 tons for deposition on landfills and it will be necessary to find other utilization for 2477 tons of waste. The produced compost which does not meet the requirements for quality (concentrations of trace metals, C/N ratio and content of PHs) can be utilized for energy generation. It is supported by Directive 2009/28/EC of the European Parliament and the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC. One of the most important steps in the production of pellets is an increase of pellet density and at the same time preservation of their sufficient mechanical durability and low consumption of energy. The aim of this paper is to describe the preparation of suitable fuel mixture which has required mechanical properties using waste products.
Material The pellets were prepared from compost produced in the facility of Kompostarna Bruzovice, operated by Frydecka skladka Ltd., located in Moravskoslezske Beskydy Mts., Czech Republic. In this plant, the compost is produced by classical method, in a land-farming hall, in aerated and drained heaps frequently turned over in order to achieve good aerobic conditions. The compost is prepared from following components: saw dust 5%, grass 60%, horse manure 20%, waste from processing of potatoes 10% and vegetable waste 5%. The gross calorific value of compost is varying around 13.78 MJ/kg d.m. (dry matter). The direct energetic utilization of compost is complicated by its high moisture which has value around 50 % even at the end of composting process, depending at the storage conditions. The decrease of moisture required six months when it reached 13.3 % (Vergnoux et al. 2009). This long time is difficult from the point of view of storage capacities. The preliminary drying for production of pellets resulted in moisture ranging from 20 to 30 %. Optimization of compost moisture can be achieved by means of suitable additives (waste paper – cardboard). The data in Table 2 show that addition of 20 % of dry cardboard decreased moisture approximately by 20 %. The graph of the thermal degradation of compost in oxidizing atmosphere (air) up to 1000 oC is in Fig.1. The weight losses were determined from thermogravimetric (TG) curve. The total loss of weight was 61.5 %. Three temperature maxima were read on DTG curves: up to 200 oC – loss of water (19.7 wt %), 200 – 400 oC – oxidation of organic matter (loss 28.3 wt. %), and 400-600 oC – oxidation of cellulose, hemicelulose and lignine (loss 20.5 wt.%). During the process of composting humic acids are formed by transformation of carbohydrates. From analysis of major components it results that compost contained 31wt.% of lignine, 8wt.% of hemicelulose, 24 wt.% of cellulose and 7wt.% of
humic acids. From the point of view of pelletization technology, the particle size of input components is very important. Grain size analysis of compost is listed in Table 1. The two grain-size classes (0.5 – 1 mm and 0.25 – 0.5 mm) have the highest percentages – both approximately 26 % of particles. Table 1 Grain size analysis of compost – content of particles in individual classes (%) before milling and after milling. Grain size (mm)
> 2,5
2,0-2,5
1,6-2,0
1,0-1,6
0,5-1,0
0,25-0,5
0,16 -0,25
0,063-0,16
0,045- 0,063
<0,045
Before After
11,25 1,0
6,89 2,01
9,9 3,1
9,89 15,1
18,85 25,8
18,5 26,1
6,45 6,90
9,44 11,2
3,8 3,8
5,0 5,0
Table 2 Ultimate and proximate analysis of used materials and mixtures Cardboard
Compost
Coal slurry
Mixture 1
Mixture 2
% % kJ/kg d.m. kJ/kg d.m. kJ/kg in original state wt.% wt.% wt.% wt.% wt.%
Moisture 7,34 62,79 11,24 22,18 17,54 Ash content 2,40 20,53 21,04 20,81 17,21 NCV 15704 14954 23313 20454 21071 GCV 14202 13786 22122 19063 19880 GCV 12990 3686 19377 14881 15990 C 40,43 37,90 44,12 H 7,29 5,61 5,78 O 50,41 34,69 26,79 N 1,85 1,20 2,27 < 0.01 < 0.01 S 0,74 Explanations: NCV – net calorofic value, GCV – gross calorific value, Mixture 1: 40 % coal slurry + 60 % kompost. Mixture 2: 40 % coal + 40 % compost + 20 % cardboard
Fig.1 DTA/TG analysis of compost Waste cardboard is another raw material which was used as an additive in order to decrease moisture. In Europe, 5.4 million tonnes of cardboard is sold, and 62 % from this amount is used as a packing material for foodstuffs. In Europe the 41.5% of used cardboard is recycled, the rest is used for energy recovery and composting. In the Czech Republic, salvage and sorting of scrap cardboard reach up to 71%, while approximately 28 % is recycled. The further utilization of scrap cardboard remains problematic. The cardboard contains cellulose, inorganic fillers (SiO2, kaolin, aluminium oxide, gypsum, calcium carbonate, titanium dioxide, zinc oxide, perlite, vermiculite) and starch adhesive
bonding agent which can be gelatinized (Mihara T. Et al. 1981). In our case, cardboard was used as an additive which decreases the moisture of mixture for production of pellets. Alternatively, saturated mixture of water and cardboard was used for increasing of durability of pellets. In water leachate, there were identified following organic compounds by HPLC: cyclohexanone, 2-Cyclohexen-1-one, 3,5,5-trimethyl, Ethanol, 2-(2-butoxyethoxy), o-hydroxybiphenyl, dibutyl phthalate, diethyl phthalate, phenylethyl alcohol, phosphoric acid tributyl ester, pentacosane, hexacosane, heptacosane, iocosane. Saturated water leachate contained 2.5 % of starch. Lignosulfonate was not identified, although it is formed during production of cellulose and it is often used for gluing of cardboard (Kim et al.). Energy parameters of cardboard are listed in Table 2. Slurries from bituminous coal dressing represent third raw material for production of pellets. Material from slurry pond Novy York in Ostrava-Karvina Coal District, Czech Republic was used in this study. Coal slurries are no more produced in Ostrava – Karvina Coal District due to changes in technology of coal beneficiation, but they are still deposited in dried slurry ponds. At present, these slurries are used as a fuel by medium-sized or large energy sources. Coal slurries have higher concentrations of sulphur, chlorides, trace elements and high content of ash in comparison with coal.
Method DTA/TG analysis was performed at the instrument NETZSCH STA 409 EP. The sample (approx.10 mg) was heated up to 1000C with heating rate 10C min-1 in crucibles from -Al2O3 in dynamic oxidizing atmosphere of air with flow 100 cm3 min-1. Analyses of samples were performed according following standardized procedures: CEN/TS 15414-3:2006 Solid recovered fuels – Determination of moisture content using the oven dry method – Part3: Moisture in general analysis sample. CEN/TS 15403:2006 Solid recovered fuels – Methods for determination of ash content. CEN/TS 15415:2006 Solid recovered fuels – Determination of particle size distribution by screen method. CEN/TS 15400: 2006 Solid recovered fuels – Method for the determination of calorific value. CEN/TS 15407:2006 Solid recovered fuels – Method for the determination of carbon (C), hydrogen (H) and nitrogen (N) content. CEN/TS 15401: 2006 Solid recover fuels – Methods for determination of bulk density CEN/TS 15639 Solid recovered fuels - Methods for the determination of mechanical durability of pellets. Pellets were prepared from material which was crushed in the hammer crusher for biomass 9FQ 50 with throughput of 800 kg/hour (manufacturer Pest Control Corporation Vlcnov, Czech Republic). Production of pellets was performed using the briquetting press JGE 150 (manufacturer Pest Control Corporation Vlcnov, Czech Republic).
Results Forces, that cause pellet damage (i.e. fragmentation and abrasion of pellet) during handling, transportation and storage may be divided into three general classes: compression, impact and shear. Compression forces result in a crushing action, impact forces result in shattering both on the surface and of the pellet and along any natural cleavage planes of the pellet and shearing forces result in abrasion of pellet edges and surfaces (Kaliyan N., Morey R.V., 2009). The strength and durability of the densified products depend on the physical forces that bond the particle together. The binding forces that act between the individual particles in densified products have been categorized into five major groups They are: solid bridges, attraction forces between solid particles, mechanical interlocking bonds, adhesion and cohesion forces and interfacial forces and capillary pressure (Kaliyan N., Morey R.V., 2009). The effectiveness of the inter-particle bonds created during the densification
process has been measured by durability. Durability or abrasive resistance test simulates either mechanical or pneumatic handling. Holmen tester have been used for measuring durability of the densified products - pellet durability index (PDI). The PDI is calculated as the ratio of weight after tumbling over the weight before tumbling, multiplied by 100. A sample is circulated in the Holmen tester for 30 – 120 s. The remaining pellets are collected, sieved with sieve size of about 80 % of pellet diameter, weighted, and the PDI is calculated. The value of PDI can be influenced by moisture when it is higher than 10 % (Arshadi et al. 2008). The PDI values were determined for all samples at same moisture (4-5%). Mechanical properties of pellets depend on a shape and grain size of individual particles. Using material with variable particle size is advantageous because it increases mechanical durability of pellets. The mixture of particles will make interparticle bonding with nearly no inter-particle spaces (Arshadi et al., 2008, Kaliyan N., Morey R.V., 2009, Sultana et al. 2010). Other authors (Turner R. 1995, Franke) recommended particle size for good pellet quality 0.6-0.8 mm. Large particles are fissure points that cause cracks and fractures in pellets. Only Payne J.D. (2006) recommended feed particle size distribution to produce high quality pellets around 3 mm. For the preparation of pellets, fine grained coal slurries were sorted into for grain-size classes. Compost was used without sorting and cardboard was crushed to the form of !foam”. The influence of grain size distribution on the pellet durability index (PDI) is illustrated in Figs.2,3 and 4. The highest value of PDI was obtained for the grain size class with smallest particles (< 0.25mm for coal) and also in the grain size class 2-4 mm, which is in accordance with conclusion of Payne J.D. (2006). The selection of binding agent represents an important factor which has influence on PDI. Lignine, proteins, starch and soluble carbohydrates are used as “natural binders” for biomass. The bonding strength of these binders is influenced by binding forces between lignine and fat, hydrogen bridges between highly polarized components like lignine, cellulose, starch and proteins, and further by gelatinization of starch, dissolution and partial recrystallization of water-soluble carbohydrates (Kaliyan N., Morey R.V., 2009) According US Patent Application 20070006526 molten hydrocarbons, wood tar, starch have been used to bind coal fines together for forming briquettes and pellets. A ratio of about 5% to about 25% biomass with the balance being coal plus binder is acceptable for briquetting a pelleting. Pellets produced only from cardboard had PDI value 99% (30s) and pellets were Compaq. PDI for pellets produced from compost reached value of 95% (30s) and 90.6% (60s). The measured value of PID was very low in case of pellets produced from carbon and coal (Fig.4), pellets had disintegrating character. Neither cardboard nor its saturated leachate were not able to create the sufficient bonding force between particles. Starch is not effective binder for ultra-fine clean coal, below 0.072 mm (Mehta R.H. and Parekh B.K. 1995). Coal slurries had 65% of material that passed the sieve 0.25 mm. Cardboard facilitated only the decrease of moisture of compost. Pellets with compost had relative higher values of PDI. In this case, compost plays role of the natural binder where mostly higher content of lignine is utilized (up to 31 wt.%). The influence of grain size of coal particles on the PDI value is not very pronounced in pellets produced from 40wt.% of coal + 40wt.% of compost + 20% of cardboard. The pellets produced from coal and compost had the highest value of PDI for grain size under 0.25 mm and grain size 2 – 4 mm. The mean value of PDI calculated from all grain size classes is 91.8 for pellets produced from coal and compost and 91.6 for pellets produced from coal, compost and cardboard. It is apparent from these results that admixture of approximately 20 wt.% of compost is sufficient for achieving of optimal value of PDI.
94
Coal+compost+cardboard
PDI (%)
92
90
92.6
91
91.2
90 89
89
88
Coal+compost+cardboard 90.5
91
90.1
90
89
91.5 91.1
91
Coal+compost
91
93.6
PDI (%)
93
92
Coal+compost 93,6
89.6
88
88.7
87
87.1
86
86.3
85
87
84
86
83 < 0,25
0,25 – 1
1-2
2-4
< 0,25
Grain size of coal particles (mm)
0,25 – 1
1-2
2-4
Grain size of coal particles (mm)
Fig.2 Pellet durability index in Holmen tester,30 s Fig.3 Pellet durability index in Holmen tester, 60 s
100 80
PDI (%)
70
Coal 70 % + cardboard 30% 89.6 78.9
30s
60s
60 50
52.8
40
41.4
30
37.6
20 10
14.2
0
8.3
11
< 0,25 0,25 – 1 1-2 2-4 Grain size of coal particles (mm)
Fig.4 Pellet durability index in Holmen tester
Conductivity (µS/cm)
90
800 700 600 500 400 300 200 100 0
y = 285.89x + 115.6 R² = 0.9912
0
1
2
3
Sample g/20 ml od deionized water
Fig.5 Leaching of cardboard
Conclusions Meeting the requirements for mechanical quality of pellets produced from slurry can be achieved by utilization of compost as a binder. Concentration of 20 wt. % is sufficient. Cardboard as a binder in pellets produced from bituminous coal slurries proved not to be successful, but it is important additive for decreasing of moisture of pellets (up to 5 wt.%), it decreases also content of ash and it increases moderately gross calorific value. The application of compost as a binder in production of pellets from coal slurries can be important contribution or use of charges which are not suitable for direct applications in agriculture or land reclamation.
Acknowledgement This work was supported by research projects of Ministry of Education, Youth and Sport of the Czech Republic: DeCOx, No. MSM 619890019, and Research of factors influencing the mechanical durability of pellets from biomass, No. SP201027.
References Arshadi M., Gref R., Geladi P., Dahlqvist S.A., Lestander T. (2008): The influence of raw material characteristics on the industrial pelletizing process and pellet quality. Fuel Processing Technology, V.89, 1442-1447. Kaliyan N., Morey R.V. (2009): Factors affecting strenght and durability of densified biomass products. Biomass and Bioenergy, 33, 337-359. Kaliyan N., Morey R.V. (2010): Natural binders and solid bridge type binding mechanisms in briquettes and pellets made from corn stover and switchgrass. Bioresources Technology, V.101, 10821090. Mehta R.H., Parekh B.K. (1995): Pelletization studies of ultra-fine clean coal. In: High Efficiency Coal Preparation: An International Symposium, Chapter 37 - Dewatering, Thickening, and Agglomeration, 40 – 49.
Mihara T. et al. (1981): Inorganic fiberboard for heat insulation. Abstract Bull. of Inst. of Paper Chem., V. 51, No. 11, 1235-1237. Payne J.D. (2006): Troubleshooting the pelleting process.In: Feed Technology Technical Report Series. Singapore. American Soybean Association International Marketing Southest Asia, 17-23. Sultana A., Kumar A., Harfield D. (2010): Development of agri-pellet production cost and optimum site. Biosource Technology (in print). Turner R. (1995): Bottomline in feed processing: achieving optimum pellet quality. Feed Management, 46, 30-33. US Patent Application 20070006526 - Fuel pellet briquettes from biomass and recovered coal slurries. Vergnoux A., Guiliano M., Le Dréau Y., Dupuy N., Doumenq P. (2009): Monitoring of the evolution of some compost properties by NIR spectroscopy. Science of the Total Environment, 407, 2390-2403.
Manuscript Not AVAILABLE
Investigation of Effect of Reagents on the Coal Recovery from Coal Washing Plant Tailings by Floatation Oktay Bayata, Metin Ucurumb and Huseyin Vapura [email protected] a b
Mining Engineering Department, Cukurova University, Adana, Turkey Mining Engineering Departmen,t Nigde University, Nigde, Turkey
ABSTRACT: In this study, effects of collector and frother types on the coal recovery from Tunçbilek Ömerler coal washing plant tailings by froth floatation were investigated applying statistical analysis. For this purpose, collectors used on the flotation experiments (kerosene, diesel oil and commercial fatty acids as ionic collectors) were compared by ANOVA, and frothers (pine oil and MIBC) were compared by t-analysis. As a result of statistical analysis, it was found that there was not a significant difference among collector types, but there was a significant difference among frothers types. Despite their lower floatation yield and higher consumptions, ionic collectors can be used successfully in coal floatation where cleaner concentrates are required from high ash coal fines. The results showed that the decrease of 53.44% in the ash content of the coal has been obtained in the floatation experiments whereas, the combustible recovery was 87,61% using commercial fatty acids as ionic collectors.
Manuscript Not AVAILABLE
Column Flotation of Fine and Coarse Coal using a Novel approach D.P. Patil, B.K. Parekh University of Kentucky Center for Applied Energy Research, Lexington, KY 40511, USA Edgar B. Klunder National Energy Technology Laboratory, Pittsburgh, PA 15236, USA
ABSTRACT: Froth flotation, applied to the separation of solid particulates, has been practiced commercially for a long time in the coal and mineral industries. The potential benefits of establishing a deep froth, especially in column flotation have been shown by a number of researchers, and that includes demonstrating that the froth phase is much more efficient at mineral upgrading than is the pulp phase. This approach could be useful in where the particles have difficulty in reporting from pulp phase to froth phase. Hence, it is expected that introduction of particles into the froth phase will significantly improve the grade and recovery of particles. In this paper, a novel way of operating a flotation column was implemented, and the results were compared to those when operating the same column in the conventional fashion. Tests were conducted with both fine and coarse coals. Feeding into the froth zone enhances bubble-particle contact as observed by a higher product yield of 79.4% compared to a conventional column flotation yield of 73.4%, both at about 9% product ash. It was also observed that the novel froth feed improved the recovery of coarse (+ 1 mm) coal particles from 0.9 weight percent to 2.2 weight percent compared to the conventional way of feeding slurry to pulp. Similarly, recovery of 1x0.6 mm particles improved from 4.6 weight percent to 8.3 weight percent at the same ash level. Positive results were also obtained by external reflux of a portion of the concentrate back into the top of the column. The potential to simultaneously achieve improvement in both recovery and grade can be explained by application of conventional mass transfer concepts, analogous to developments in twophase foam fractionation.
STUDY OF THE LIGNITE QUALITATIVE PARAMETERS MODIFICATION, DURING ITS STORAGE Prof. dr. Sanda Krausz – University of Petrosani Eng. Daniel Burlan – National Society of Lignite, Oltenia Dr. Mihail Dafinoiu – National Society of Lignite, Oltenia Dr. Ion Bacalu – National Society of Lignite, Oltenia Assoc.prof.dr. Nicolae Cristea – University of Petrosani ROMANIA Abstract: The research goes out upon the quest of establishing the qualitative parameters variation based on the lignite storage conditions, from different open pits of the biggest coal basin from Romania. We start from the premise that the physical-chemical processes which inevitably occur during the lignite storage have negative effects on their quality. The lignite oxidation can cause the calorific power decrease, the burning capacity diminution and the modification of the lignite grains proprieties. Experimental approach was carried out in two stages: in the laboratory and on field. The lignite qualitative parameters modification depending on storage duration and coal moisture was studied and the implications of the storage duration are discussed. The effect of the climatic factors on the deposited lignite quality has also been analyzed: the temperature, the air moisture, the atmospheric pressure, the wind action.
1. Introduction The Romanian power necessary is mostly covered, about 98.6%, from internal production: 56.2% from thermo-electric power stations, 26.4% from hydro-electric power stations and 17.4% from nuclear and wind power stations. Lignite makes up the main raw material in thermo-electric power stations and its production is provided from Oltenia area, where 8 quarries are running with a reached production of about 32.5 mil. tonnes / year, the calorific value being between 1100 and 2400 kcal/kg. In the latest two years, the electric production in thermo-electric power stations decreased on an average yearly rate of 3.6%, concomitantly with a yearly increase rate of 7.5% in hydro-electric power stations and of 61% in nuclear power stations (I.N.S. 2009) In National Strategy for Sustainable Development, Romania reconsider its economical policy in electrical and thermal energy production and despite the problems with coal usage, this remains, for at least 50 years, the most important power resource. Currently, however, as a result of lower and especially fluctuating lignite demand in power plants, the gap between the two phases increased exploitation and delivery and implicitly, its storage period in mining deposits that frequently exceed three months. As a result, physical and chemical processes caused mainly by coal oxidation – an inherent and permanent phenomenon during exploitation until combustion – have become more intense, causing deterioration and in time, a loss of quality. This led to the initiation of this research which results will establish specifically, to what extent, the prolonged storage of lignite, because of oxidation, can worsen its important quality features, mainly the calorific value. The ultimate goal of this research was to establish whether a more accurate foreseeing of production is needed in relation to fluctuating demand and the design of a
conducted storage to ensure that the quality parameters remain unaffected within a time limit. 2. Generalities on coal oxidation during storage Research results on coal oxidation during the storage period, made to different types of coal over the years are not yet conclusive, probably due to the complexity of involved factors. Also, the possibility to adjust the storage duration made the interest in determining the effects of this process to be greatly diminished. As shown in the literature, investigations on the dependence of coal reactivity on its structure were conducted in laboratory conditions. Their results were particularly important from a theoretical viewpoint, covering knowledge of the physicochemical phenomena taking place and little of them have been applied in storage practice. Coal has in its natural composition oxygen in various proportions, depending on their degree of carbonization. The higher the carbon content, the smaller the oxygen content, so that the carbon content is in the position to be a way of assessing the capacity of oxidation of coal. Coal oxidation can be started even by depositing a fresh extracted coal quantity on an already stocked coal, the latter inducing heat transfer phenomena, with the formation of certain auto-heating nuclei. The phenomenon can be easily observed during the cold season with higher humidity, when the vapor formation is visible, when vapors are released into the air. What can’t be seen is the vapor adsorption on the surface of coal grains, but it was established that an increase of 1% in coal moisture due to vapor condensation is accompanied by an increase in temperature with 17oC, which increases the oxidation rate. (Parks, 1963). It was found that after 20 years of oxidation, coal still produces carbon oxides. CO2 release was estimated to achieve a consumption rate of 88% of adsorbed oxygen (Huggins, 1989). Outdoor storage practice causes visible coal degradation by losing its mechanical strength and smoothness, by deposition of iron hydroxides, sulfates, iron and calcium carbonate on cracks (Davidson, 1990). The main features of coals which have a strong tendency of auto oxidation and implicitly of ignition are: high speed of oxidation, characteristic in particular for the lower rank coal containing a relatively high content of moisture, oxygen, volatile matter, high friability and the presence of finely divided pyrite in coal mass. During a prolonged contact with air, coal changes its physical and chemical properties, which undergo a change more or less deep. Deterioration starts by adsorbing oxygen at the surface and along cracks in the coal, causing a surface oxidation of organic matter and some minerals. In the same time, on cracks are deposited iron hydroxides, iron and calcium sulfates and carbonates (Ghethner, 1987). Among petrographic constituents, clarain and vitrain undergo the easiest alteration, durain has higher resistance and fusain is the most difficult to be altered. It was established that low temperature oxidation of coal is followed by autoheating, a process that has a relatively long duration (Bacalu, Krausz, 2009). When critical temperature (400o ÷ 600oC) is reached, the properly firing is released. The oxidation force that can release the coal auto-firing depends on many factors: the oxygen content existing in coal or adsorbed on its surface, the coal surface state, the volatile content, the heat from the environment, the content of pyrite. Also, production of mixtures of CH4 and CO2 from chemical interaction caused by oxidation processes, contribute greatly to the auto-firing initiation. 2
Moreover, the coal storage manner is important: in big stacks, the environmental heat losses are small and the auto-firing risk increases compared with small stacks in which the potential for ignition is lower (Huidu E., 1996). The initial higher decrease in oxygen consumption when the freshly excavated coal is exposed to air, suggests that the coal – oxygen complex formed can be highly resistant to oxidation. It is probably a peroxide type formed during the hydrocarbon and other organic material oxidation. The complete low-temperature oxidation of fine coal yields an almost completely soluble product in caustic solutions. In fact, the amount of soluble material formed by low temperature oxidation is a good measure of the coal oxidation state. These soluble materials can be ullmine, ullmic acid, humine and humic acids. After precipitation from solutions, using acids, these solid suspensions consume atmospheric oxygen particulates at low temperature and have an indefinite composition. When oxidized coal is heated above its decomposition temperature or oxygen is continuously added in reaction, the carbon-oxygen complex decomposes to form carbon dioxide, carbon monoxide and water, their weight being variable. It was found that, at a temperature of 200oC, there is decomposition in carbon dioxide, carbon monoxide and water, and the characteristic rate of oxidation of coal is restored to levels close to those determined for fresh coal. The complex properties may vary depending on the formation circumstances and the type of coal. Studies on the initial appearance of carbon dioxide when fine coal is rapidly heated in oxygen showed that for a peroxidated coal, carbon dioxide appears first at about 125oC, while at the fresh excavated coal and stored in stacks, it appears at about 93oC. Minerals presented in coal structure can have considerable influence in the process of oxidation. An ash quantitative spectroscopic analysis showed that bassanit (CaSO4 · 1/2H2O) can be used as an indicator of the oxidation process (Pearson, 1979). The presence of iron in different forms (pyrite, illit, jarosite, goethite, lepidocrocit etc.) can be another cause of increasing coal oxidation process (Huggins, 1980). Also, oxidation is hastened because of the presence of soluble salts of iron from coal. As we mentioned, moisture is one of the basic parameters in the oxidation process, with major influence not only on pyrite and other minerals oxidation, but also of the organic part in coal. Water plays an important role in coal oxidation thus in peroxide forming, initiating oxidative reactions in organic macerales. Accelerating effect of moisture on the oxidation – degradation processes, shows that water enters into interaction with oxygenated groups on coal surface. High humidity tends to facilitate the formation of CO2 and prevents carbon monoxide formation (Bowman, 1980). High humidity during storage of coal seems to "preserve" the coal reactivity to subsequent oxidation. 3. Experimental research 3.1 Lignite characteristics As it’s known, the coal is composed of three components: organic mass (substances with complex structure consisting of carbon, oxygen, sulfur and nitrogen), inorganic mass (mineral substances that turn into ash as a result of coal calcination) and water, and from chemically standpoint, they contain: carbon, hydrogen, oxygen, nitrogen, sulfur, mineral mass (ash after calcination) and moisture. 3
To establish the exact composition of lignite from Oltenia, determinations were performed on three sets of representative samples collected from the 26 layers of the eight quarries in operation. The analysis results are considered like average values and are calculated in several different bases: AR (as-receive – coal sample containing total humidity, noncombustible mineral mass and organic mass), for the sample for analysis (ground below 0.20 mm) and for DAF – dry, ash free state, (coal sample without moisture and without non-combustible mineral mass, but containing sulfur). Results are shown in Table 1. Table 1
Ultimate analysis results
Chemical element Carbon Hydrogen Nitrogen Sulfur Oxygen
AR 26.93 2.24 0.6 1.45 9.725
Weight percent, wt % analysis 44.25 3.67 2.4 -
DAF 66.86 5.59 1.5 3.56 -
Ash analysis As a result of the determinations performed on each sample, there were established content ranges of different oxides in coal ash: SiO2 between 44.89 and 52.31%, Al2O3 between 21 and 26.48%, Fe2O3 between 8.14 and 11.76%, CaO between 6.77 and 13.62%, SO3 between 4.2 and 11.03%, MgO between 2.43 and 351%, TiO2 between 0.85 and 1.72, K2O between 1.4 and 2.18%, and Na2O between 0.33 and 0.72%. In addition to the macro-elements, lignite concentrates a whole series of microelements in quantities between 1ppm and 1%. The tendency of accumulation of microelements in coal mass decreases in the series: Ge > W > Ga > Be > Nb > Mo > Sc > Y > La > Zn > Pb. Elements accumulated in coal are: Li, Na, K, Rb, Mg, Ca, Sr, Ba, V, Cr, Ni, No, Pb, Ag, B and Ge. Average content of ash on the 3 series of samples was 23.91%, 22.92% and 25.43% respectively. Lignite hygroscopic moisture has a constant value on the three samples, around 5%. Calorific value of lignite It was determined on each sample the gross (heat power) calorific value and lower (net) calorific value was calculated. The results - the average values - were: - heat power value: 2466, 2500 and 2470 kcal/kg, respectively - lower calorific value: 2107.4, 2135.3 and 2107.7 kcal/kg respectively. Limits of variation of lower calorific values for the analyzed samples are between 1700 and 2600 kcal/kg.
4
3.2 Laboratory experimental research on auto-heating and auto-oxidation of coal in quarries 3.2.1 Determination of oxidation capacity – auto-heating of lignite in quarries Since the lignite exploited from the eight quarries has different characteristics, a prior research was designed to establish quarries with the most liable to qualitative depreciation through oxidation (Krausz and coworkers, 2003). With this purpose, to establish of oxidative and auto-heating capacity, we used a method based on recording the changes in temperature of the mixture of coal samples under 0.2 mm size, and a hydrogen peroxide solution (hydrogen peroxide) at a concentration of 20% (Ionescu and coworkers, 1993). It was measured the initial temperature of the mixture and then, changes were recorded every minute, until the temperature of 50oC is reached (first stage of heatingoxidation). After reaching this temperature, the time intervals when the temperature increases by 10 degrees have been pursued and noted until the maximum temperature is reached (phase II). Shortly after reaching maximum temperature, the reaction begins to decrease in intensity, a continued cooling of hydrogen peroxide solution – coal mixture being observed. The times when temperature decreased with 10 degrees are recorded, until temperature value reach 50oC. The results showed that the higher the sample’s calorific value, the heating duration is lower, and vice versa; the lower coal quality, the auto-oxidation duration is greater, so establishing which are the two quarries (among the eight taken in the analysis) that are more important to have in view for the investigation of qualitative degradation phenomenon during lignite storage, because these quarries contain the highest availability for oxidation – ignition coals. These careers were Lupoaia and Roşiuţa. Measurements for the two quarries (and not a single) were done for a greater certainty of obtained data. 3.2.2. Determination of lignite characteristic alteration depending on storage duration, in laboratory conditions Three identical samples, in terms of quantity (5 kg) and sizes (2.5 - 5 mm), but with different moisture content were taken from each Lupoaia and Roşiuţa quarries. The first sample is in anhydrous state (after drying) and the other two with larger moisture, are obtained by spraying the sample with water. Thus the Lupoaia samples have: average carbon content of 58.58%, the ash content of 29.7%, hygroscopic moisture of 4.91%, and moistures for samples II and III of 17.50% and 28.10%. Samples from Roşiuţa have: carbon content of 57.15%, ash content of 36.4%, hygroscopic moisture of 3.81% and moistures for samples II and III of 16.30% and 29.80%. The 2x3 samples were stored under identical conditions, in cylinders made of cardboard and fitted with 10 mm diameter holes to provide a natural draft of air through the lignite granules. Temporal variation of qualitative parameters was monitored and the results are presented in Table 2.
5
Table 2 Sample
I Wi=0 II Wi=17,5% III Wi=28,1%
Qualitative parameters variation during storage Storage Parameters DAF DAF N H VDAF CDAF duration, % % % % days LUPOAIA QUARRY initial after 5 days after 10 days after 20 days after 30 days initial after 5 days after 10 days after 20 days after 30 days initial after 5 days after 10 days after 20 days after 30 days
0,70 0,72 0,76 0,77 0,80 0,70 0,67 0,66 0,61 0,65 0,70 0,65 0,68 0,70 0,71
3,52 3,36 3,38 3,36 3,40 3,52 3,63 3,77 3,89 3,92 3,52 3,71 3,79 3,93 3,99
Qsup Kcal/kg
41,60 41,45 41,40 41,00 41,00 41,60 41,38 41,27 40,75 40,15 41,6 41,45 41,4 41,00 39,50
58,58 58,45 58,33 58,00 57,83 58,58 58,33 58,21 57,93 57,60 58,58 57,90 57,88 57,69 57,45
3814 3801 3785 3751 3740 3814 3768 3744 3742 3739 3814 3765 3744 3739 3698
38,62 38,48 38,52 38,45 38,45 38,62 38,59 38,75 38,81 38,76 38,62 39,43 39,89 40,09 39,21
57,15 57,21 57,18 57,11 57,10 57,15 57,83 56,91 56,78 56,75 57,15 56,73 56,68 56,65 56,61
3371 3364 3342 3328 3328 3371 3345 3338 3337 3324 3371 3321 3315 3308 3303
ROŞIUŢA QUARRY I Wi=0 II Wi=16,3% III Wi=29,8%
initial after 5 days after 10 days after 20 days after 30 days initial after 5 days after10 days after 20 days after 30 days initial after 5 days after 10 days after 20 days after 30 days DAF
0,59 0,63 0,64 0,62 0,64 0,59 0,58 0,61 0,57 0,58 0,59 0,61 0,63 0,58 0,62
DAF
3,25 3,12 3,15 3,10 3,15 3,25 3,41 3,28 3,35 3,38 3,25 3,61 3,52 3,59 3,51
DAF
DAF
Notation in table: N – nitrogen content; H - hydrogen content ; V - volatile matter content; C i h carbon content; W – initial moisture; W – hygroscopic moisture; Qsup – superior (heat) caloric value
Findings: nitrogen content has a slightly variation, even after the 30 days storage period; hydrogen content, has a slightly decreasing trend for dried coal samples and increase in wetted samples, for both quarries; volatile matter content slightly decreases over time for dried samples and a slightly increase is observed with the moisture increase, recording the same type of variation for both quarries, carbon content decreases over time, especially since the initial moisture content of coal is higher, ebbing is of 1.28% for sample I, 1.67% for sample II and 1.93% for sample III in the case of Lupoaia lignite and much weaker (from 0.1 to 0.8%) in the case of Roşiuţa.
6
Changes in heat power calorific value are presented graphically in figures 1 and 2, showing the qualitative depreciation that suits to carbon’s content reduction. This depreciation is more important as the sample has higher moisture content. The decrease in calorific value is 1.95% respectively 1.28% for dried lignite, 1.97% respectively 1.39% for the samples II and 3.05% respectively 2.02% for the samples III. These results confirm that oxidation process is more intense on higher coal quality.
calorific value(kcal/kg)
The caloric value variation on storage time for Lupoaia lignite
3820 3800 3780 3760 3740 3720 3700 3680 3660 3640
Initial
after 5 days
after 10 days
after 20 days
after 30 days
Wi=0
3814
3801
3785
3751
3740
Wi=17,5%
3814
3768
3744
3742
3739
Wi=28,1%
3814
3765
3744
3739
3698
Period (days)
Figure1 - The calorific value variation depending on storage time and moisture content for Lupoaia lignite
calorific value(kcal/kg)
The caloric value variation on storage time for Roșiuța lignite
3380
3360
3340
3320
3300
3280
3260
Initial
after 5 days
after 10 days
after 20 days
after 30 days
Wi=0
3371
3364
3342
3328
3328
Wi=16,3%
3371
3345
3338
3337
3324
Wi=29,8%
3371
3321
3315
3308
3303
Period (days)
7
Figure 2 - The calorific value variation depending on storage time and moisture content for Roşiuţa lignite 3.2.3 Research to determine the influence of industrial storage time on qualitative parameters of lignite from Lupoaia quarry Research in quarry was conducted over a period of six months, covering the months from January to June, to follow the progress of phenomenon depending on climate and weather changes within three seasons. The study was conducted for both quarries, based on monitoring of production and supply of coal stored at each loader and establishing lower calorific values as the main quality parameter. Table 3
Quarry
Lignite characteristics Lignite characteristics Layer no.
Total moisture W t, %
Ash content anh A ,%
Lower calorific value, Qlow,
V VI VII VIII X V-VI VII-VIII X-XII
40,2 41,9 40,5 40,7 40,0 42,0 42,3 42,5
28,9 28,7 30,1 28,1 40,21 27,9 28,7 35,8
2264 2192 2218 2282 1825 2210 2183 1923
Lupoaia
Roşiuţa
kcal/kg
Characteristics of directly sampled lignite from exploited layers are shown in Table 3 and constitute the basic parameters for daily reporting of the production features. Based on recorded data, were calculated the average monthly values by weighting, based on daily deliveries and the regression equations were obtained. The decreasing trend of lower calorific value of lignite during storage is clearly shown by these equations and is also presented graphically in figures 3 and 4. Delivered coal quality variation depending on the storage time on the careers LUPOAIA
2200
calorific value [Kcal/Kg]
2100
y = -13,702x + 2094,9 R 2 = 0,8721
2000 y = -13,96x + 2086,3 2 R = 0,7
1900
y = -11,835x + 2099,9 R2 = 0,7522
1800
1700
y = -16,563x + 2116,2 2 R = 0,9015
y = -10.397x + 2117 R² = 0.5822
1600
y = -13,255x + 2083,1 2
R = 0,8874 1500 0
10 January Liniară (January)
20 February Liniară (Marth)
30 Marth Liniară (April)
40 April Liniară (May)
50 60 storage time [days] May Liniară (June)
June Liniară (February)
8
Calorific value [Kcal/Kg]
Figure 3 Monthly variation of lower caloric value depending on storage time for Lupoaia quarry lignite Delivered coal quality variation depending on the storage time on the career ROSIUTA 2300 2200 2100 2000 1900 1800
y = -11,984x + 2075,2 R2 = 0,8631
y= -16,799x + 2108,9 R2 = 0,8388
1700
y = -16,009x + 2117,4 R2 = 0,7997
1600
y = -14,628x + 2105,5 2 R = 0,8744
y = -13,415x + 2093,8 R2 = 0,894 y = -13,523x + 2095,9 R2 = 0,9356
1500 0
5 January May Liniară (Marth)
10
15 February June Liniară (April)
20
25 Marth Liniară (January) Liniară (May)
30
35
April Liniară (February) Liniară (June)
40
storage time [days]
Figure 4 Monthly variation of lower caloric value depending on storage time for Roşiuţa quarry lignite Stationary period of coal deposits is variable from 5 to 50 days. The start of depreciation process occurs when coal is stocked in the deposit (when the calorific value is of approximately 2100 kcal/kg), and after an average of 30 days stocking, quality loss is over 10%. As the storage time of coal increases, the monthly regression straight line position changes, achieving a considerable difference between them. During probation period, the climatic factor changes were pursued: air and soil temperature, atmospheric relative humidity, atmospheric pressure, wind speed and specific weather phenomenon such as fog, rain, snow depth (Bacalu, Krausz, 2009). Variations in these parameters during the analysis are presented in Table 4. Table 4
Month January February March April May
Climatic factor values from January to June Air Soil Air relative temperature, temperature, humidity, o o C C % -1,74 -2,07 94,35 -3,97 -3,65 83,43 4,26 4,97 71,35 10,01 11,19 70,30 19,59 22,34 59,03
Atmospheric pressure, mBar 992,23 995,58 997,04 1289,34 991,54
Wind speed, m/s 0,14 0,41 0,56 0,73 5,43 9
June 22,79 27,12 66,13 991,02 0,55 Average 8,49 9,98 74,10 1042,79 1,30 The most influential factors on the storage behavior of lignite are air and soil temperature, relative humidity and wind speed. Air and soil temperature retains the same increasing course in January-June period, from -10 to 30oC, with negative values in winter months (January and February). The effect favoring the oxidation process occurs in the period with higher temperatures, during the summer months. This variation is the inverse of relative air humidity levels, which have high values in winter and low values in summer. Corroborating the evolution of the three parameters with their effect on the oxidation process, we have the explanation for the fact that lignite depreciation is not according to month and season. Thus, in January, when the negative temperature are recorded, relative humidity has a maximum value (94.35%) favoring oxidation; in June, the maximum temperature (23oC for air and 27oC for soil) favoring oxidation, even if relative humidity drops to 66.13%. The wind speed records increases almost constant on average 0.3 m/s every month, from January to April and higher in May (5.43 m/s). The correlation between calorific value decrease and wind speed increase is positive and accountable through its effect on increasing the oxidation process. Based on the obtained data it was determined a depreciation coefficient also due to these factors, representing the percentage reduction in calorific value from the initial value to that which is at delivery, values that are presented in Table 5 for the two quarries. Table 5 Quarry
Depreciation coefficient values Specification
Caloric initial value, delivery Lupoaia kcal/kg Storage time, days Depreciation coefficient, % Caloric initial value, delivery Roşiuţa kcal/kg Storage time, days Depreciation coefficient, %
Month January February March April 2155 2114 2134 2153
May 2151
June 2139
1834
1972
1918
1888
1974
1943
27
11
13
15
8
10
14.90 2176
6.72 2168
10.12 2210
12.31 2200
8.23 2210
9.16 2210
1880
1922
1950
1951
2034
2004
15
13
11
11
5
6
13.60
11.35
11.76
11.32
7.96
9.32
Depreciation coefficient value is increasing with the storage time of lignite, having a maximum value of 14.9% for 27 days of storage for Lupoaia lignite and 13.6% after 15 days storage lignite from Roşiuţa. Although not all climatic factors have an influence as strong, cumulative effect of their action is evident. 4. Conclusions 1. Investigation conducted under laboratory conditions revealed a high oxidation capacity of lignite. Also it was established a different behavior for the components of the 10
studied samples, distinguishing in all situations that there is a low carbon content decrease over 1% and accordingly a calorific value decrease of more than 2%. 2. The results for the two quarries of Oltenia basin are similar, validating the accuracy of determinations. Because they contain lignite with the highest oxidative capacity, the results from their study are even more justified for lignite from the other quarries, where oxidative capacity of lignite is lower. 3. Climatic factors influence the process of lignite depreciation, but their simultaneous influence determined a mutual compensation of these influences. 4. The long-time storage of lignite determines its unquestionable quality depreciation, which begins to appear even after five days; keeping stocks for a period exceeding 45 days leads to a sharp decrease in calorific value due to coal oxidation process. The start of the oxidation process occurs from excavation and dumping in storage place, the phenomenon is increasing day by day and leads to formation of nuclei of fire in different areas of the store, reaching finally to coal auto-firing. 5. Research results show the need for planning excavation process based on deliveries, in order to obtain duration of the material storage below the 5 days, after registering a qualitative depreciation of lignite. With this purpose, it is required a more careful planning in time of the production or consideration of special measures for the disposal of lignite, in conditions which decrease its oxidation.
References Bacalu I., Krausz S. (coord.), 2009 – Study of the dependence between the qualitative parameters of lignite and its exploitation and storage conditions, Doctoral thesis, University of Petrosani. Bowman R., 1980 – Low temperature oxidation of bituminous coal. Infrared spectroscopic study of samples from a coal pile, Fuel, 59. Davidson R., 1990 – Natural oxidation of coal, IEA Coal Research. Ghethner J.S., 1987 – The mechanism of the low temperature oxidation of coal by O2: observation and separation of simultaneous reactions in situ FT-IR difference spectroscopy, Applied Spectroscopy. Huggins, F.E., 1980 – Mossbauer detection of goethite in coal and its potential as indicator of coal oxidation, International Journal of Coal Geology. Huggins, F.E., 1989 – Chemistry of coal weathering, Elsevier Science Publishers, Amsterdam. Huidu E., 1996 – Coal storage systems, Technical Publishing House, Bucharest. Ionescu C., Matei A., Traistă E., 1993 – General Chemistry, University of Petrosani P.H. Krausz S., Trotea T., Bacalu I., 2003 – Behavior of the lignite from Oltenia National Company at oxidation and self heating, Mining Revue nr.12. Matei A., Toth I., Traistă E., Bacalu I., 2001 – Coal yard self combustion monitoring by gaseous emission control, Proceeding of Third International Symposium Mining and Environmental Protection, Vrdnik, Serbia. Parks, B.C.,1963 – Chemistry of coal utilization, New York – London, Lowry P.H. 11
Pearson D.E., 1979 – Mineral matter as a measure of oxidation of a coking coal, Fuel, 44. *** - 2009, Data of Statistic National Institute, Bucharest.
12
NATURAL-GAS-LEVEL EMISSIONS WHEN BURNING NAPHTHA (WITHOUT WATER INJECTION) IN A COMMERCIAL GAS TURBINE USING THE LPP TECHNOLOGY, CREATING A “CLEAN POWER” ALTERNATIVE FOR AN INTEGRATED GASIFICATION COMBINED CYCLE (IGCC) POLYGEN PLANT
Leo D. Eskin, Richard J. Roby, Michael S. Klassen, Richard G. Joklik and Maclain M. Holton LPP Combustion, LLC 8940 Old Annapolis Road, Suite K Columbia, MD 21045
Abstract Emissions test results from operation of a commercial dry, low-emissions Capstone gas turbine demonstrate the commercial feasibility of using the Lean, Premixed Prevaporized (LPP) Technology to burn a range of light liquid fuels in a power generation gas turbine, without water injection, while simultaneously achieving ultra-low, natural-gas-level emissions for NOx, CO and particulates. The presented test results have significant implications for future Integrated Gasification Combined Cycle (IGCC) plants. The Integrated Gasification Combined Cycle technology, as currently defined, couples a complex coal gasification process plant with a custom, coal syngasfired, combustion turbine combined cycle power plant. The IGCC process is a two-stage combustion design with gas cleanup between the stages. The first stage employs a gasifier where partial oxidation of the coal occurs. The second stage utilizes the gas turbine combustor to complete the combustion with the gas turbine/combined cycle (GT/CC) technology. Due to the impracticality of storing significant quantities of the coal syngas, it is necessary to ensure that the combustion turbine remains operational whenever the gasification plant is running. The shutdown of the combustion turbine requires immediate shutdown of the gasification plant. In addition, it is difficult to operate the gasification plant at part load; hence it is preferable to run the combustion turbine in a base load configuration. These are significant operating limitations. The current test results demonstrate the commercial viability of combining the gasto-liquid technology (GTL), e.g. Fischer-Tropsch synthesis or similar, and the Lean Premixed Prevaporized combustion technology to create a much more robust power generation system. The GTL process is a method whereby coal syngas is transformed into one or more forms of liquid fuel. These coal liquids can include diesel fuel, kerosene and naphtha (among others). The conversion of coal syngas to liquids is a well-known process and has been utilized for many years. The LPP process transforms a 27th Intl. Pittsburgh Coal Conference
1
October 2010 - Istanbul, Turkey
wide variety of liquid fuels into a substitute natural gas (or LPP gas) which the current results clearly show can be burned in conventional natural gas dry low emissions combustion hardware, precluding the use of water or steam to achieve low criteria pollutant (NOx, CO and PM) emissions levels. By combining the LPP combustion technology with the GTL process, IGCC operation is made much more flexible, dependable, and the overall economics are improved. Three alternative IGCC design scenarios are presented which would allow the IGCC plant to use a standard, natural gas fired combustion turbine together with an LPP skid to provide increased operating flexibility for the plant and to reduce the plant capital equipment cost. Emissions test results for a Capstone gas turbine are presented for both naphtha and bioethanol as a liquid fuel source for the LPP skid. The tests were conducted at the LPP Combustion facility located in Columbia, MD (just north of Washington, DC) and the generated power was sold to Baltimore Gas and Electric as part of a net metering agreement. Further gas turbine testing is in progress using additional liquid fuels such as diesel and biodiesel. Introduction The Integrated Gasification Combined Cycle (IGCC) technology has been recognized as an efficient and clean method of converting coal (or other carbonbased feedstock) into electricity, process heat, high-value liquid fuels and other chemicals. Several IGCC plants are now operating commercially both within the United States and internationally. While a number of IGCC power and polygeneration plants are currently under construction (notably in China), more widespread adoption of coal gasification and the IGCC technology has been hindered by overall plant economics, and by the inherent complexity associated with building and operating such facilities. The gasifier in a commercial IGCC plant is a very large device which generates syngas at high temperature and pressure (approx. 400 psia and 2600 F), and the gasifier has a large thermal mass. Hence, it is desirable to operate the plant at a stable base-load condition over a long period of time. Unfortunately, reduced demand for electrical power (at night and on weekends) leads to inefficient and less economical plant operation during these low-load time periods, due to decreased part-load performance of both the gasification block and the gas turbine power block. A typical coal-based IGCC plant is shown diagrammatically in Figure 1 below. Coal is partially oxidized by oxygen (or air) at high pressure and temperature, and the resulting syngas stream is cooled, cleaned of impurities (and potentially CO2) and burned in a combustion gas turbine. As shown in Figure 1, gas turbine modifications are normally required to accommodate the increased volumetric flow rate for the syngas, which occurs because of the reduced heating value of the syngas (125-250 BTU/scf LHV) as compared to the heating value of natural gas (800-1,000 BTU/scf LHV). These
27th Intl. Pittsburgh Coal Conference
2
October 2010 - Istanbul, Turkey
custom hardware changes increase the gas turbine capital and maintenance costs and restrict the turbine operation. As part of an overall effort to improve reliability, operability, and plant economics, a broad range of IGCC design and optimization studies have been conducted by the Electric Power Research Institute1,2 (EPRI) and by the United States Department of Energy3 (DOE), under the auspices of the Vision 21 program4,5,,among others. Common to all of the above studies, and common to the existing commercial IGCC facilities, is the requirement that the combustion turbine be fueled by the syngas produced in the gasification block. This design constraint couples the operation of the gasification block and the gas turbine power block, due to the impracticality of storing large quantities of syngas, and decreases the overall plant availability and reliability. Practical considerations of a large power generation facility, such as load following and part load operation, become much more difficult with the added complexity and operational requirements of the gasification plant. IGCC Plant Performance Enhancement Using the LPP Technology The process described herein proposes to combine the gas-to-liquid technology (GTL), e.g. Fischer-Tropsch synthesis or similar, and the Lean, Premixed Prevaporized (LPP) combustion technology with the coal gasification technology to create a much more robust power generation and co-production system. This combination would burn the coal-derived liquids using the same sophisticated combustor hardware used today. This technology has dramatically lowered the pollutant emission levels from natural gas power plants over the last 20 years. The emissions using the coal-derived fuels would be similar to those of natural gas. The GTL process is a method whereby syngas is transformed into one or more forms of liquid fuel. These liquid fuels, known generically as coal liquids or F-T liquids, may include diesel fuel, kerosene and naphtha, among others. The conversion of syngas to liquids is a proven technology that has been in
1
Evaluation of Alternative IGCC Plant Designs for High Availability and Near Zero Emissions: RAM Analysis and Impact of SCR. EPRI, Palo Alto, CA: 2005. 1010461. Integrated Gasification Combined Cycle (IGCC) Design Considerations for High Availability— Volume 1: Lessons from Existing Operations. EPRI, Palo Alto, CA: 2007 1012226. 2
3
http://www.fossil.energy.gov/programs/powersystems/vision21/
4
“Topical Report – Task 1 Topical Report, IGCC Plant Cost Optimization,” Gasification Plant Cost and Performance Optimization, United States Department of Energy, National Energy Technology Laboratory, Contract No. DE-AC26-99FT40342, May 2002. 5
“Topical Report – Task 2 Topical Report, IGCC Plant Cost Optimization,” Gasification Plant Cost and Performance Optimization, United States Department of Energy, National Energy Technology Laboratory, Contract No. DE-AC26-99FT40342, September 2003.
27th Intl. Pittsburgh Coal Conference
3
October 2010 - Istanbul, Turkey
commercial use around the world for many years6,7,8. The recently invented LPP process transforms a wide variety of liquid fuels into a vaporized fuel stream (or LPP Gas™) which may be burned in conventional, natural gas, dry low emissions combustion hardware, precluding the use of water or steam to achieve low criteria pollutant (NOx, CO and PM) emissions levels. By combining the LPP combustion technology with the GTL process, IGCC operation is made much more flexible, dependable, and the overall economics is improved.
Air or O2 Coal
Water/Steam from Syngas Cooler
Clean Gasification Syngas Slag
Sulfur, CO2
Ambient Air
Hot Exhaust Gas
Modified Gas Turbine
Cold Exhaust Gas to Stack
Heat Recovery Steam Generator
High Pressure Steam
Electrical Power
Electrical Power
Steam Turbine Low Pressure Steam to Condenser
Figure 1: Simplified Typical Coal-based IGCC Plant Process Flow Diagram LPP Technology and Liquid Fuel Combustion Overview Traditionally, spray diffusion combustors have been employed in gas turbines that operate on liquid fuels such as fuel oil #1 and fuel oil #2. However, this diffusion mode of operation tends to produce unacceptable levels of NOx emissions. The current technology for burning liquid fuels in gas turbines is to use water and/or 6
“Topical Report – Volume I, Process Design – Illinois No. 6 Coal Case with Conventional Refining”, Baseline Design/Economics for Advanced Fischer-Tropsch Technology, U.S. Department of Energy, Contract Number DE-AC22-91PC90027, October, 1994. 7
“Topical Report – Volume IV, Process Flowsheet (PFS) Models”, Baseline Design/Economics for Advanced Fischer-Tropsch Technology, U.S. Department of Energy, Contract Number DE-AC2291PC90027, October, 1994. 8
“Topical Report VI – Natural Gas Fischer-Tropsch Case, Volume II, Plant Design and ASPEN Process Simulation Model”, Baseline Design/Economics for Advanced Fischer-Tropsch Technology, U.S. Department of Energy, Contract Number DE-AC22-91PC90027, August, 1996.
27th Intl. Pittsburgh Coal Conference
4
October 2010 - Istanbul, Turkey
steam injection with conventional diffusion burners. Emissions levels for a typical “state of the art” gas turbine, such as a GE 7FA burning fuel oil #2 in diffusion mode with water/steam injection, are 42 ppm NOx and 20 ppm CO9. Water/steam injection has a dilution and cooling effect, lowering the combustion temperature and thus lowering NOx emissions. But at the same time, water/steam injection is likely to increase CO emissions as a result of local quenching effects. Thus, the “wet” diffusion type of combustion system for liquid fuels must trade off NOx emissions for CO emissions. In recent years, stringent emissions standards have made lean, premixed combustion more desirable in power generation and industrial applications than ever before, since this combustion mode provides low NOx and low CO emissions without water addition. Lean, premixed combustion of natural gas avoids the problems associated with diffusion combustion and water addition. Thus, lean, premixed combustion is the foundation for modern Dry Low Emissions (DLE) gas turbine combustion systems. When operated on natural gas, DLE combustion systems provide NOx and CO emissions of 25 ppm or less with no water addition. However, these systems cannot currently operate in premixed mode on liquid fuels because of autoignition and flashback within the premixing section. Plee and Mellor10 characterized autoignition of the fuel/air mixture in the premixer as an important factor that causes flashback in practical combustion devices. Autoignition of the fuel/air mixture occurs before the main combustion zone, when the ignition delay time of the fuel/air mixture is shorter than the mean residence time of the fuel in the premixer. Autoignition especially occurs with the higherorder hydrocarbon fuels, such as fuel oils, which have shorter ignition delay times compared to natural gas11. The short ignition delay times of vaporized higher hydrocarbons have proven difficult to overcome when burning in lean, premixed mode. Nevertheless, in order to overcome high NOx levels produced by spray combustion, gas turbine designers still desire to use lean, premixed, prevaporized (LPP) combustion. Several approaches have been reported in the 12,13,14,15,16,17,18,19 to overcome flashback and autoignition in the premixers literature
9
Davis, L. B. and Black, S. H., 2000, “Dry Low NOx Combustion Systems for GE Heavy-Duty Gas Turbines”, GER 3568G. 10
Plee, S. L. and Mellor, A. M., 1978, “Review of Flashback Term Reported in Prevaporizing/Premixing Combustors”, Combust. Flame, 32. pp. 193-203. 11
Oumejjoud, K., Stuttaford, P., Jennings, S., Rizkalla, H., Henriquez, J., Chen, Y., 2005, “Emission, LBO and Combustion Characterization for Several Alternative Fuels”, Proc. ASME Turbo Expo 2005, Paper# GT2005-68561. 12
Maier, G. and Wiitig, S., 1999, “Fuel Preparation and Emission Characteristics of a Pressure Loaded LPP Combustor”, 30th AMA Fluid Dynamics Conference, AIAA- 99-3774.
13
Imamura, A., Yoshida, M., Kawano, M., Aruga, N., Nagata, Y. and Kawagishi, M., 2001, “Research and Development of a LPP Combustor with Swirling Flow for Low NOx”, 37th Joint Propulsion Conference & Exhibit, AIAA-2001-3311.
27th Intl. Pittsburgh Coal Conference
5
October 2010 - Istanbul, Turkey
of LPP combustors. These approaches attempt to achieve low NOx emissions by designing premixers and combustors that permit rapid mixing and combustion before spontaneous ignition of the fuel can occur. In most of the work reported on LPP combustion systems in the literature, the fuel is sprayed directly into the premixer so that the liquid fuel droplets vaporize and mix with air at lean conditions. Typically, swirlers with multi-port liquid fuel injection systems are employed for better fuel/air mixing20. However, unlike these attempts to alter hardware, there has been no reported work on altering fuel combustion characteristics in order to delay the onset of ignition in lean, premixed combustion systems. In the present solution, vaporization of the liquid fuel in an inert environment has been shown to be a technically viable approach for LPP combustion. As described by Roby et al.21, a fuel vaporization and conditioning process has been developed and tested22 to achieve low emissions (NOx and CO) comparable to those of natural gas while operating on liquid fuels, without water or steam addition. In this approach, liquid fuel is vaporized in an inert environment to create a fuel vapor/inert gas mixture, LPP Gas™, with combustion properties similar to those of natural gas. Premature autoignition of the LPP Gas™ was controlled by the level of inert gas in the vaporization process. Tests conducted in both atmospheric and high pressure test rigs utilizing typical swirl-stabilized burners (designed for natural gas) found operation similar to that achieved when burning 14
Ikezaki, T., Hosoi, J. and Hidemi, T., 2001, “The Performance of the Low NOx Aero Gas Turbine Combustor Under High Pressure”, ASME paper 2001 -GT-0084.
15
Lin, Y., Peng, Y. and Liu, G., 2004, “Investigation on NOx of a Low Emission Combustor Design with Multihole Premixer-Prevaporizer”, Proc. ASME Turbo Expo 2004, Paper# GT2004- 53203. 16
Lee, C., Chun, K. S., Locke and R. J., 1995, “Fuel-Air Mixing Effect on NO X Emissions for a Lean Premixed-Prevaporized Combustion System”, 33rd AIAA Aerospace Sciences Meeting and Exhibit, Paper# AIAA-95-0729. 17
Michou, Y., Chauveau, C., Gijkalp, I. and Carvalho, I. S., 1999, “Experimental Study of Lean Premixed and Prevaporized Turbulent Spray Combustion”, 37th AIAA Aerospace Sciences Meeting and Exhibit, AIAA 99–0332. 18
Hoffmann, S., Judith, H. and Holm, C., 1998, “Further Development of the Siemens LPP Hybrid Burner”, ASME International Gas Turbine & Aeroengine Congress & Exhibition, ASME 98-GT-552. 19
Mansour, A., Benjamin, M., Straub, D. L. and Richards, G. A., 2001, “Application of Macrolamination Technology to Lean, Premixed Combustion”, ASME J. of Eng. Gas Turb. Power, 123, pp. 796–802. 20
Lin, Y., Peng, Y. and Liu, G., 2004, “Investigation on NOx of a Low Emission Combustor Design with Multihole Premixer-Prevaporizer”, Proc. ASME Turbo Expo 2004, Paper# GT2004- 53203. 21
Roby, R. J., Klassen, M. S. and C. F. Schemel, 2006, “System for Vaporization of Liquid Fuels for Combustion and Method of Use”, U.S. Patent, #7,089,745 B2. 22
P. Gokulakrishnan, M. J. Ramotowski, G. Gaines, C. Fuller, R. Joklik, L. D. Eskin, M. S. Klassen and R. J. Roby, 2008, "Experimental Study of NOx Formation in Lean, Premixed, Prevaporized Combustion of Fuel Oils at Elevated Pressures", Journal of Engineering for Gas Turbines and Power, Vol. 130.
27th Intl. Pittsburgh Coal Conference
6
October 2010 - Istanbul, Turkey
natural gas. Emissions levels were similar for both the LPP Gas™ fuels (fuel oil #1 and #2, Biodiesel and F-T synthetic JP-8) and natural gas, with any differences in NOx emissions ascribed to fuel-bound nitrogen present in the liquid fuel. Also, tests showed that the LPP combustion system helps to reduce the NOx emissions by facilitating stable combustion even at very lean conditions when using liquid fuels. Extended lean operation was observed for the liquid fuels due to the wider lean flammability range for these fuels compared with natural gas. An added advantage of the fuel vaporization and conditioning process is the ability to switch between LPP Gas™ and natural gas ‘on-the-fly’ in the same combustor, without significantly affecting the flame stability. Elevated pressure tests were conducted on a full temperature, full pressure combustor test stand capable of supplying combustor air at typical compressor discharge temperatures and pressures. During these high pressure gas turbine burner tests, the liquid fuel was supplied in gaseous form from the LPP liquid fuel vaporizer skid shown in Figure 2. The testing involved a study of emissions and combustion characteristics, such as flame stability and lean blow-out limits. The tests were performed at typical compressor discharge temperatures. For the high pressure tests, typical compressor discharge pressures were also used. The same fuel nozzle used for natural gas testing was also used for liquid fuel testing on LPP Gas™ without any modifications.
Figure 2: LPP liquid fuel vaporizer skid used for gas turbine burner testing at elevated pressures. Figure 3 shows NOx and CO emissions at full load conditions for both natural gas and fuel oil #2. During the testing, emissions and dynamics data were taken over a range of lean equivalence ratios from approximately 0.75 to the lean blow-off (LBO) limit. However, the emissions data is plotted against measured exhaust gas temperature in order to provide a common temperature reference. The lowest 27th Intl. Pittsburgh Coal Conference
7
October 2010 - Istanbul, Turkey
temperature data points shown in Figure 3 reflect the experimentally observed LBO limit. Figure 3 shows that fuel oil #2 LPP Gas™ has an extended LBO limit compared to natural gas and thus can achieve NOx emissions nearly as low as natural gas despite the fuel-bound nitrogen. Figure 3 also shows that the crossover point between NOx and CO emissions extends to lower temperatures (and therefore lower equivalence ratios) for fuel oil #2 LPP Gas™ as compared to natural gas. As can be seen from the figure, fuel oil #2 LPP Gas™ showed increased flame stability and an extended LBO limit at lower temperatures (equivalence ratio) compared to natural gas. Figures 4 and 5 show similar results (using the same burner at atmospheric-pressure) for a variety of fuels, including a F-T derived jet fuel (S-8).
NOx & CO - ppmvd (at 15% O2)
45 40
NOx (Fuel Oil #2) CO (Fuel Oil #2)
35
NOx (Natural Gas ) CO (Natural Gas )
30 25 20 15 10 5 0 1100
1200
1300
1400
Exhaust Temperature (K) Figure 3: Comparison of NOx & CO emissions measurements for fuel oil #2 and natural gas as a function of measured exhaust gas temperature for a single fuel nozzle at Solar Turbines Taurus 60 full load conditions (100%). Combustion air temperature was 648 K, combustor pressure was 12.6 atm, and fuel dilution was 5:1 (molar basis).
27th Intl. Pittsburgh Coal Conference
8
October 2010 - Istanbul, Turkey
CENTAUR 50 DATA (1 ATM) 50 NOx - ppmvd (at 15% O2)
Fuel Oil #2
40
Fuel Oil #1 Na tura l Ga s Biodie se l B100 (SME)
30
Etha nol (ASTM D-4806) Na phtha (Pe troleum) S-8 (FT-GTL)
20
JP-8 (4-12-07)
10 0 1500
1600
1700
1800
1900
2000
Exhaust Temperature (F)
Figure 4: Comparison of NOx emissions measurements for fuel oil #2, fuel oil #1, soy methyl ester biodiesel, ethanol, naphtha, synthetic JP-8 (S-8), JP-8 and natural gas as a function of measured exhaust gas temperature for a single fuel nozzle at Centaur 50 full load conditions (100%). Combustion air temperature was 627 K, combustor pressure was 1 atm CENTAUR 50 DATA (1 ATM) 25 CO - ppmvd (at 15% O2)
Fue l Oil #2
20
Fue l Oil #1 Na tura l Ga s Biodie se l B100 (SME)
15
Etha nol (ASTM D-4806) Na phtha (Pe trole um ) S-8 (FT-GTL)
10
JP-8
5 0 1500
1600
1700 1800 Exhaust Temperature (F)
1900
2000
Figure 5: Comparison of CO emissions measurements for fuel oil #2, fuel oil #1, soy methyl ester biodiesel, ethanol, naphtha, synthetic JP-8 (S-8), JP-8 and natural gas as a function of measured exhaust gas temperature for a single fuel nozzle at Centaur 50 full load conditions (100%). Combustion air temperature was 627 K, combustor pressure was 1 atm. 27th Intl. Pittsburgh Coal Conference
9
October 2010 - Istanbul, Turkey
Gas Turbine LPP Fuel Testing Testing of the LPP Combustion System was also performed using both naphtha and bioethanol to fuel a Capstone C30 gas turbine. The Capstone C30 microturbine is a 30 kW power generation system designed for use with natural gas. The C30 initially ignites and operates in diffusion mode until approximately 22 kW, when it switches to premixed mode. In premixed combustion mode, it produces nominally 9 ppm NOx and 30 ppm of CO. The combustion system operates at approximately 4 atm. An LPP system was developed that interfaces with the natural gas fuel feed, allowing for operation on LPP fuel, natural gas or a mixture of these fuels. Figure 6 shows the test facility developed for the C30 testing.
Figure 6: Capstone C30 Test Facility used to evaluate emissions for various fuels using the LPP Combustion Technology. NOx and CO emissions were measured for naphtha, pure ethanol and ethanol/water blends and compared to baseline operation on methane. NOx emissions when running on either naphtha or ethanol are comparable to or less than the emissions on methane (see Figure 7), and show a trend of decreasing emissions as the water content in the ethanol increases. At a water content greater than 35% in the ethanol, stable combustion could not be maintained. CO emissions with naphtha are equal to or lower than those with methane for both diffusion mode operation and for premixed mode operation (see Figure 8). CO emissions with ethanol are equal to or higher than those with methane for diffusion mode operation, and show a trend of lower emissions as the ethanol water content increases. CO emissions with ethanol are approximately equal to or lower than methane for premixed mode operation (see Figure 8). 27th Intl. Pittsburgh Coal Conference
10
October 2010 - Istanbul, Turkey
80 Baseline CH4
70
100% EtOH 95% EtOH
NOx [ppm] @15% O2
60
90% EtOH
50
65% EtOH Naphtha
40 30 20 10 0 0
5
10
15
20
25
30
Load [KW]
Figure 7: C30 NOx emissions for naphtha, pure bioethanol, blends of 95/5%, 90/10% and 65/35% bioethanol with water, and methane as a function of gas turbine load. The turbine operates in diffusion mode for loads below 25 KW, and operates in premixed mode at 25 KW and above. 250 Baseline CH4 100% EtOH
200
95% EtOH CO [ppm] @15% O2
90% EtOH
150
65% EtOH Naphtha
100
50
0 0
5
10
15 Load [KW]
20
25
30
Figure 8: C30 CO emissions for naphtha, pure bioethanol, blends of 95/5%, 90/10% and 65/35% bioethanol with water, and methane as a function of gas turbine load. The turbine operates in diffusion mode for loads below 25 KW, and operates in premixed mode at 25 KW and above. 27th Intl. Pittsburgh Coal Conference
11
October 2010 - Istanbul, Turkey
These tests demonstrate that the LPP Combustion System is capable of burning naphtha (a fuel readily available as a 25%-40% product stream from the FischerTropsch process) and ethanol (an important renewable liquid fuel), in a gas turbine combustor with NOx and CO emissions similar to those obtained from operation on natural gas. These results were obtained using a commercial DLE gas turbine combustion system designed for use with natural gas with no modifications to the combustion hardware. The pollutant emission levels achieved are much lower than can be obtained when using these fuels in conventional spray diffusion mode, which is how liquid fuels are burned today in gas turbines or reciprocating engines. The LPP Combustion System has demonstrated natural gas level emissions from naphtha without the need for expensive post-combustion cleanup such as selective catalytic reduction (SCR). An added advantage of the LPP fuel vaporization and conditioning process is the ability to achieve fuelinterchangeability of a natural gas-fired combustor with liquid fuels. Improved Plant Heat Rate While beneficially reducing NOx emissions, water injection used for NO control in traditional spray flame (diffusion) combustors incurs a substantial negative impact on the efficiency and maintenance of a liquid-fueled gas turbine. Water addition reduces the NOx emissions by reducing the flame temperature, but this in turn also reduces the Brayton Cycle thermodynamic efficiency of the gas turbine. In addition, the energy required to vaporize the injected water further reduces the gas turbine efficiency. Lastly, the reduced firing temperature reduces the exhaust temperature of the gas turbine for liquid-fueled operation, which in turn reduces the amount of steam production for a combined cycle plant. All of these effects serve to increase the net plant heat rate (and cost of operation) for a traditional liquid fueled gas turbine combined cycle power plant. In contrast, a combined cycle plant which is fueled by prevaporized fuel is not subject to the above losses. A gas turbine fueled in this way may be operated at the full natural gas firing temperature, thereby avoiding the above reductions in efficiency. Since there is no water injection with the LPP System, that vaporization loss is also avoided. Combined cycle performance calculations have shown that for a typical single pressure level heat recovery steam generator (HRSG) and GE Frame 7EA class gas turbine, one can expect to achieve at least a two percent (2%) improvement in the overall combined cycle plant heat rate when burning liquid fuel in LPP mode as compared against burning the same liquid fuel in traditional spray-flame diffusion combustors. These calculations were made using the GateCycle™ 23 power plant modeling software, and the losses associated with a fully integrated LPP system were included in the calculation of the net plant heat rate. This level of heat rate improvement is quite substantial, and represents an annual fuel savings of over five million dollars for base load operation of a GE Frame 7EA combined cycle plant (126 MW).
23
www.gepower.com/prod_serv/products/oc/en/opt_diagsw/gatecycle.htm
27th Intl. Pittsburgh Coal Conference
12
October 2010 - Istanbul, Turkey
LPP/IGCC Integration Scenarios Integration of the LPP and IGCC technologies can take place through several potential scenarios, as described below. In Scenario #1 (shown in Figure 9), it is proposed to enhance the coal gasification portion of the IGCC process plant with a GTL process plant to transform the coal syngas into coal liquids. These coal liquids could then be readily stored in tanks, for use as necessary. In this scenario, it is further proposed to replace the syngasfired combustion turbine used in a typical IGCC design with a conventional naturalgas fired combustion turbine, combined with an LPP skid to transform the coal liquid fuels into LPP Gas™ which will be burned by the conventional dry low emissions combustion turbine. By creating coal liquids, the gasification block would no longer require continuous operation of the combustion turbine, and the gasification block could maintain full base load operation, regardless of the power generated by the power block. If the combustion turbine load is reduced or removed altogether, the excess coal liquids produced would be stored as necessary in nearby tanks, or would be distributed via pipeline, truck or train. This configuration provides the widest load-following capability for the power block, with no minimum power generation requirement. Truck, Rail or Pipeline
Air or O2
Tanks
Clean Gasification Syngas
Coal
Slag
Gas to Liquid Coal Conversion Liquid
Sulfur CO2 LPP Gas
Hot Exhaust Gas Water/Steam from Syngas Cooler
Heat Recovery Steam Generator
Standard Gas Turbine
Cold Exhaust Gas to Stack
LPP Skid
Ambient Air Electrical Power
High Pressure Steam
Steam Turbine
Electrical Power
Low Pressure Steam Figure 9: Scenario #1: Integrated LPP Gas™ IGCC Plant
27th Intl. Pittsburgh Coal Conference
13
October 2010 - Istanbul, Turkey
In Scenario #2 (shown in Figure 10), the coal gasification portion of the IGCC process plant is enhanced with a GTL process plant to transform a portion (but not all) of the coal syngas into coal liquids. The syngas stream that was not converted to liquid would be burned in one or more syngas-fired combustion turbines, sized for base load operation during low plant load conditions (at night, etc.). The coal liquids would be consumed by standard combustion turbine hardware burning LPP Gas™, which would operate in “peaking mode” as necessary, and would allow the overall plant to respond to electrical load changes without having to change the rate of production of the syngas. In Scenario #3 (shown in Figure 11), it is proposed to completely decouple the gasification/GTL plant and the power plant. The coal liquids would be produced at the gasification/GTL plant and would be shipped to stand-alone combustion turbines that are equipped with the LPP technology. This would provide the added benefit of allowing the gasification/GTL plant to be sited at any location, including a location in close proximity to the coal (or other feedstock) source. A site within close proximity to the coal source would reduce the transportation cost for the coal, and would facilitate disposal of the slag waste product resulting from the gasification plant. Air or O2 Coal
Gasification
Clean Syngas
Truck, Rail or Pipeline
Coal Liquid Tanks
Gas to Liquid Conversion
LPP Skid
Slag Sulfur CO2 Ambient Air Ambient Air
Modified Gas Turbine
Hot Exhaust Gas
Standard Gas Turbine Water
Cold Exhaust Gas to Stack
Heat Recovery Steam Generator Water Heat Recovery Steam Generator Electrical Power
Steam Turbine
Cold Exhaust to Steam
High Pressure Steam
Electrical Power
Low Pressure Steam
Steam Turbine Low Pressure Steam
Figure 10: Scenario #2: Integrated LPP Gas™/Syngas IGCC Plant
27th Intl. Pittsburgh Coal Conference
14
October 2010 - Istanbul, Turkey
Air or O2 Coal
Gasification
Gasification/CTL Plant
Clean Syngas
Tanks
Gas to Liquid Coal Liquid Conversion
Slag Sulfur CO2
Power Generation Plant
Truck, Rail or Pipeline
LPP Skid
Heat Recovery Steam Generator
LPP Gas
Ambient Air
Standard Gas Turbine
Cold Exhaust Gas to Stack
Water
HP Steam Hot Exhaust Gas
Electrical Power
Electrical Power
Steam Turbine LP Steam
Figure 11: Scenario #3: Non-Integrated LPP Gas™ IGCC Plant Additional Benefits and Synergies of the LPP/IGCC Integration The combination of the coal gasification process with the LPP process also offers the following advantages, in addition to those described above: 1. Beneficial use of waste Nitrogen (N2) produced by the air separation unit portion of the coal gasification plant: Most coal gasifiers use high purity oxygen (O2) to partially oxidize the coal and create the coal syngas. This oxygen is produced by an air separation unit (ASU) that separates the oxygen and nitrogen from ambient air. The nitrogen produced by the ASU is considered to be a waste product stream, and is injected into the clean syngas burned by the combustion turbine in an attempt to reduce NOx emissions by the combustion turbine. However, nitrogen gas can be used by the LPP process to create the LPP Gas™. Hence, by using the waste nitrogen already available from the ASU, the energy requirements of the LPP process are substantially reduced. Note that the low NOx combustion hardware present in natural gas combustion turbines does not require the use of supplemental nitrogen as is the case for syngas combustion hardware. 2. Significant improvement in the IGCC value proposition may be realized by using lower value FT liquids (such as naphtha) as a feedstock for the LPP process, in place of the higher value coal liquid product streams such as kerosene or diesel fuel: Naphtha comprises a sizable portion (30-45%) of the total output stream of a Fischer27th Intl. Pittsburgh Coal Conference
15
October 2010 - Istanbul, Turkey
Tropsch based coal liquids plant. Unsuitable for use as a transportation fuel, the monetary value of naphtha is expected to significantly decline as commercial production of coal liquids increases in the future. Since naphtha represents a very good feedstock for the LPP process, the LPP technology will be able to convert the low value liquid stream into a high value LPP Gas™ stream, improving overall plant economics. 3. It may be possible to substantially reduce the plant capital cost if a spare gasifier is not needed for the coal gasification plant: As noted, the gasifier hardware portion of a coal gasification plant operates at a very high temperature and pressure. It has been found that the reliability of the gasifier hardware is such that traditional IGCC plant economics may require that a spare gasifier be built as a “hot standby” in case the primary gasifier fails or requires maintenance. The standby gasifier is needed because 1) there is a long lead time required to repair a gasifier, and 2) the syngas produced cannot be stored for use while the gasifier is being repaired. The gasifier hardware can cost tens or hundreds of millions of dollars in a typical IGCC plant. In the event of a gasifier failure, the integrated LPP/IGCC plant could continue to produce electricity by converting the stored coal liquid to LPP Gas™ and burning it in the gas turbine at the plant. 4. Ownership and operation of the CTL and power blocks may be separated: One of the operational issues of concern to IGCC plant owners is the fact that the coal gasification process is a complex chemical process for which the power industry does not have extensive experience. Scenarios #1 and #3 decouple the coal gasification/CTL plant from the power generation plant. This would allow, for example, a process plant company to own and operate the gasification/CTL plant, while a utility or independent power producer would operate a standard combustion turbine plant, along with the LPP skid. Summary and Conclusions The LPP Combustion technology presented in this paper allows the cleanest combustion of liquid fuels, achieving natural gas levels of criteria pollutants (NOx, CO, SOx & PM). In addition, no “net” carbon emissions are obtained when used with biofuels. This technology provides the capability for tremendous fuel flexibility and low emissions not previously attainable in modern DLE gas turbines with liquid fuels, and provides substantial operating cost savings for liquid fueled combined cycle power plants. The flexibility and economics of an IGCC plant can be considerably improved with the introduction the LPP technology, by cleanly burning coal-derived liquids as part of the plant configuration. An enhanced IGCC plant configuration has been 27th Intl. Pittsburgh Coal Conference
16
October 2010 - Istanbul, Turkey
proposed, which integrates the LPP technology with the traditional IGCC design. The combined LPP/IGCC concept overcomes several notable challenges associated with the operation and profitability of existing Integrated Gasification Combined Cycle facilities. By operating in a combined IGCC and CTL (polygen) configuration, the resulting plant has much more flexibility to handle turn-down due to reduced power demand. The resulting coal liquids are more readily stored than syngas, allowing the coal gasification block to continue to operate at design capacity even when the power production needs are reduced. Use of the LPP technology decouples the gasification and power blocks, allowing the gasification to take place at locations that may be more advantageous from an economic or environmental point of view. Other benefits include use of waste nitrogen for the LPP process, use of sidestream products such as naphtha as a feedstock for the LPP process and elimination of the need for a spare gasifier. It is anticipated that these enhanced configurations will significantly advance the acceptance and implementation of the IGCC concept in the future.
27th Intl. Pittsburgh Coal Conference
17
October 2010 - Istanbul, Turkey
Laminar Flame Speed Study of Syngas Mixtures (H2-CO) with Straight and Nozzle Burners Nicolas Bouvet, Christian Chauveau and Iskender Gökalp CNRS-Institut de Combustion, Aérothermique, Réactivité et Environnement, 1C Av. de la Recherche Scientifique,45071 Orléans Cedex 2, France Seong-Young Lee and Robert J. Santoro Propulsion Engineering Research Center and Department of Mechanical and Nuclear Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A. Laminar flame speeds of undiluted syngas (H2/CO) mixtures have been studied at atmospheric conditions using chemiluminescence and schlieren techniques for straight cylindrical and nozzle burner apparatus. A wide range of mixture composition, from pure H2 to 1/99 % H2/CO, has been investigated for lean premixed syngas flames. To achieve a better flame stabilization and reduce flame flashback propensity, two nozzle burners of different sizes have been designed and fabricated and were further used to compare the flame cone angle and the total surface area of the flame techniques. Results are compared to predictions using recent H2/CO mechanisms developed for syngas combustion. I. Introduction Sythetic gas, better known as “syngas”, has been widely studied because of its numerous industrial applications: chemicals and gases synthesis, methanol production [1] alternative fuel for spark-ignition engine [2] etc. One of the most promising and active research areas at the moment is certainly the power generation with the development and optimization of Integrated Gasification Combined Cycle (IGCC) plants. Processes of gasification allow a wide range of solid combustibles, including coal, biomass and municipal solid wastes (MSW) to be converted into syngas mixtures that will be used in a gas turbine engine to generate electricity. In this paper, syngas is considered as an alternative fuel, as its production and combustion involve a considerably clean conversion of solid fuels into energy: gas clean up, sulphur removal, CO2 capture along with lean premixed combustion in modern gas turbine combustors are the key answers for demanding environmental standards [3,4]. One of the particular concerns, however, is the variability of syngas composition recorded in the industry. Some IGCC project shows H2/CO ratios between 0.33 and 2.36 by volume [5]. Studies on syngas production from coal have shown that parameters such as coal quality and origins, reaction temperatures in gasifiers, oxygen/coal ratio or steam/coal ratio have an important impact on the composition of the synthesised syngas [6]. Clearly, a strong flexibility is expected while designing the new generation of IGCC power plants, particularly for the hardware related to syngas combustion. Modern combustors, for example, will have to deal with various syngas and hydrogen-enriched fuels, which requires a better understanding of their combustion complexities [7] Phenomena, such as autoignition, flame flashback and combustion instabilities still have to be better understood in order to generalize the use of syngas by designing safer facilities [8] The knowledge of flame speed of such mixtures is therefore highly required.
This study proposes a systematic investigation of laminar flame speeds of pure H2/CO syngas mixtures at atmospheric conditions. Few studies related to laminar flame speeds of H2/CO mixtures are available in the literature. Furthermore, most of them have investigated a very limited number of syngas compositions for equivalence ratios not always representative of lean premixed combustion in a gas turbine engine. Studies of flame speed were performed by Scholte and Vaags [9,10] for rich H2/CO mixtures on a nozzle burner using the schlieren method. An identical diagnostic was used by Günther and Janisch [11] on a straight burner to study syngas mixtures at stoichiometric conditions. Yumlu [12] measured adiabatic burning velocities of various hydrogen/carbon monoxide mixtures for an equivalence ratio of 0.6, using a flat flame burner with direct cooling. Flame speeds were obtained by extrapolation to zero heat loss of burning velocities obtained for different gas flow rates. Vagelopoulos and Egolfopoulos [13] investigated laminar flame speed and extinction strain rates of hydrogen and carbon monoxide mixtures using the counterflow, twin-flame technique along with laser Doppler velocimetry. McLean et al. [14] and Brown et al. [15] reported flame speeds obtained with the constant-pressure expanding spherical flame method, followed very recently by Sun et al. [16] for pressures up to 40 atm. Laminar flame speed and stretch effects were studied by Hassan et al. for pressures between 0.5 and 4 atm [17]. Huang et al. [18] performed flame speed measurements of a pure reformer gas (28% H2, 25% CO and 47% N2 by volume) at atmospheric conditions using digital particle image velocimetry and the counterflow twin flame apparatus. Natarajan et al. [19] investigated the influence of CO2 addition and reactant preheating on laminar flame speeds of syngas mixtures obtained by the cone surface method based on broadband chemiluminescence. Flame speeds near extinction have been studied very recently by Love et al. [20] and Subramanya and Choudhuri [21] on a water-cooled nitrogenstabilized flat flame and a twin flame counterflow burner, respectively. Detailed chemical models for H2/CO combustion are also available in the recent literature. Three of them have been selected to perform flame speed computations for the present study. Sivaramakrishan et al. [22], proposed a H2/CO mechanism based on the works of Davis et al. [23], with updated parameters for the reaction OH + HO2 = H2O + O2 for combustion at elevated pressures. Sun et al. [16] presented a H2/CO mechanism to model their laminar flame speeds obtained by the constant-pressure spherical flame technique. A limited number of other experimental data has been used to validate their mechanism and a new reaction rate constant for the reaction CO + HO2 → CO2 + OH was determined using ab initio calculations. A mechanism for CO, CH2O and CH3OH combustion has recently been proposed by Li et al. [24] with new rate constant recommendations for the reactions CO + OH = CO2 + H and HCO + M = H + CO + M. These chemical models have been selected for their recent updates of reaction rates and thermodynamics data as far as their validation against various experimental results are related to syngas combustion. II. Experimental Apparatus A schematic of the experimental apparatus used is shown in Figure 1. Each gas of the H2/CO/Air mixture is initially stored in separated tanks. Degrees of purity are respectively 99.95% for the hydrogen and 99.995% for the carbon monoxide. Air used is of breathing quality (99.95%). Each gas flow rate is carefully controlled by calibrated mass flow meter with an accuracy of about 1% full scale. A mixing section allows a rapid mixing of the reactants prior to injection into the burner.
Figure 1. Schematic of the experimental apparatus.
Figure 2. Diagnostics set-up. Straight Burners A series of six stainless steel burners have been designed and fabricated. Tested burners diameters are 3, 4, 6, 8, 12, and 16 mm. A nominal burner length of 900 mm was chosen based on the 50 x diameter criteria, therefore ensuring a fully developed laminar flow at the burner rim. Particular care was taken to fabricate sharp-edged burner rims so as to avoid any irregularity that could affect the flame stabilization. The apparatus also integrates a pilot flame ring with a premixed mixture of methane and air as reactants. Tests were conducted to determine if the pilot flame could enhance the syngas flame stability at lower equivalence ratios. As flames obtained on the 3 mm diameter displayed ridges characteristic of cellular instabilities and the use of the 16 mm tube was severely restricted by the available flow rates, results presented in this study were exclusively obtained for the 4, 6, 8 and 12 mm diameter tubes. Diagnostics Set-up The Z-type two-mirror schlieren system is arranged as shown in Figure 2: the light beam emitted by a tungsten lamp (30W) goes first through a condenser lens. The diverging beam issued from an aperture located at the focal point of the condenser lens is used as a point source for the first spherical mirror (fl=1m). The top of the burner is placed in the middle of the test region defined by the parallel beam formed between two spherical mirrors. The
resulting image is formed on a digital camera with 5 Mega pixels. A vertical knife edge located at the focal point of the second mirror is used to block the white beam and therefore to form the schlieren images. OH* chemiluminescence images were recorded with an intensified CCD (ICCD) camera (PIMAX: 512´512 pixels) equipped with an UV lens (f#/4.5) and filters (WG305+UG11). OH* chemiluminescence images were captured by accumulation, for each case, of 50 instantaneous images. During the experiment, the ICCD camera was moved to achieve the best resolution possible for each series of measurements. Similarly to Natarajan et al. [19] an horizontal knife edge was used to reduce the intensity from the base of the flame and make its tip more visible. This method substantially improves the trace of the flame reaction zone boundaries and thus the laminar flame speed computations. Schlieren and OH* chemiluminescence images were simultaneously taken for every flame studied. III. Flame Speed Calculations Methods The syngas laminar flame speed (SL) calculations were performed using two different approaches. First, an averaging method has been used to determine SL. Assuming that the burning velocity is the same over the total surface area of the flame (A), the flame speed can be calculated as follows: SL = A / Q where Q is the total volumetric flow rate of the unburned mixture. This method requires knowledge of the total area of the flame, deduced by analysis of the chemiluminescence images. As shown in Figure 3, a FORTRAN program performs a three-point Abel inversion [25] of the ICCD images to deduce the 2-D boundaries of the flame, based on the maximum emission of OH*. Using the flame axisymmetry hypothesis, the flame speed is calculated from each half of the computed flame.
Figure 3. OH* chemiluminescence image processing: (Left) OH* chemiluminescence of the syngas flame and (Right) image after Abel inversion and maximum intensity trace. Two main concerns need to be considered when using this method: first of all, the flame speed is not roughly constant over the entire surface of the flame as shown by Lewis and Von Elbe [26]. Indeed, slower velocities due to heat sink effects have to be expected at the burner
rim and higher velocities due to intense burning rate at the flame tip. Furthermore, the flame area should be defined just upstream of the preheat zone where the properties of the unburned gas are not modified. Calculations based on the OH* chemiluminescence will consequently lead to higher calculated flame surfaces, implying an underestimation of the calculated flame speeds. Still this method yields good results compared to other methods found in the literature. Figure 4 displays a comparison of different data set points for the laminar flame speed of pure H2 mixtures. The data are taken over the 4 mm burner and compared to various published data. The model results by Li et al. [24] are also superimposed in the figure. Flame tips tend to be closed above equivalence ratios of ~0.8, while below equivalence ratios of ~0.8, the flame tip appears open. A more detailed description for the opened flame tips will be presented in a later section. Although laminar flame speeds from this study are, as expected, lower than predictions, present results show a very good agreement with recent data points obtained in spherical bomb devices by Lamoureux et al. [27]. The second method employed is the cone angle method. This method is particularly suitable for flames displaying a straight cone and therefore, for experiments using aerodynamically contoured nozzles. In this case, the velocity (Uo) of the unburned mixture at the nozzle exit can be considered as constant and the expression of the laminar flame speed (SL) is given by SL = Uo sin α, where Uo is the bulk velocity of the unburned mixture and α the half cone angle of the flame. A FORTRAN program was used to perform edge detection on acquired schlieren images and therefore determine the angle α.
Figure 4. Laminar flame speeds of pure H2/air mixtures. Model: Li et al. [24] Flame Tip Opening In the attempt to study ultra lean mixtures of syngas, the problem of flame tip opening arises. This phenomenon, mainly observed for rich heavy-hydrocarbon/air mixtures and lean hydrogen/air mixtures has been studied by various investigators, including experiments of Law et al. [32] for methane, propane and butane/air mixtures and also mathematical description of the phenomenon by Buckmaster [33]. Figure 5 shows chemiluminescence images of syngas flames with different H2/CO compositions for an equivalence ratio close to 0.6. It can be seen that for the first case of pure hydrogen flame, the tip is obviously open. This trend is due to the highly diffusive nature of hydrogen, characterized by a Lewis number (ratio of thermal diffusivity to mass diffusivity) lower than unity, combined with flame curvature effects. As the deficient species, hydrogen, approaches the flame tip, its concentration is further reduced due to the defocusing effect of the tip curvature. The rate of thermal diffusion being much lower than mass diffusion in this case, the burning intensity is considerably weakened at the tip and local extinction is consequently observed. By further adding CO, the tip opening is reduced (cases (b) and (c)) and a closed tip is finally obtained for the 40/60% H2/CO composition (d).
a b c d Figure 5. Open flame tip phenomenon – 4 mm burner – (a) pure H2, (b) 80/20% H2/CO, (c) 60/40% H2/CO, and (d) 40/60% H2/CO. The flame tip opening phenomenon described previously was mainly observed for mixtures with higher hydrogen content (pure hydrogen to 80/20% H2/CO) for which it was impossible to obtain closed tips for equivalence ratios below 0.8. As the influence of a possible mixture loss through the open tip on the laminar flame speed seems difficult to evaluate and would require complex modelling of the syngas flame, results herein presented only consider the case of closed laminar premixed cones. IV. Results and Discussion A wide range of syngas mixture compositions have been tested on four different straight burners with diameters varying from 4 to 12 mm. The PREMIX code [34] was used to systematically calculate laminar flame speed for equivalence ratios and mixtures compositions and of interest. In the PREMIX code, Soret effects and multicomponent diffusion were considered while performing the calculations. The continuation mode of the PREMIX code was used to gradually enlarge the computation domain and refine the resolution grid so as to improve the convergence and precision of the calculations. Mechanisms were computed with the thermodynamics and transport data selected by the different investigators. Figure 6 displays the results for a 50/50 % H2/CO mixture. Experimental results show a very good agreement with the works of Natarajan et al. [19] using Bunsen burner flames, and Mclean et al. [14] with constant-pressure expanding spherical flames. Model predictions are also closely describing the increase of laminar flame speed from 42 cm/s at an equivalence ratio of 0.62 to 107 cm/s at an equivalence ratio of 1. The overall agreement of the models with the available experimental data is not surprising as the tested mechanisms have been validated using Mclean et al. [14] data points for 50/50 and 5/95% H2/CO mixtures.
Figure 6. Laminar flame speeds of 50/50% H2/CO/Air mixture. Symbols represent experimental data and lines represent computations.
Figure 7. Laminar flame speeds of various syngas mixtures. Symbols represent experimental data (6 mm burner diameter), and dotted lines represent computations performed with the mechanism of Sun et al. [16] However, it is interesting to mention that mechanisms of Sun et al. [16] and Li et al. [24] presenting similar predictions over the range of investigated equivalence ratios, are closer to the experimental data points, while mechanisms from Sivaramakrishnan et al. [22] and Davis et al. [23] seem to predict lower flame speeds. Sun et al.16 mechanism has been selected to further compare experimental results and numerical predictions obtained for compositions varying from 10/90% to 70/30% H2/CO (see Figure 7). As expected, the laminar flame speed increases with increasing hydrogen content in the mixture. It is also worth mentioning that for a given mixture, the flame speed has a linear evolution for equivalence ratios between 0.6 and 1. In general, experimental results are slightly lower than calculated flame speeds. The choice of OH* chemiluminescence to determine the flame surface area and heat losses at the burner rim are seen to be the main reasons. Still, over the range of targeted compositions, agreement between experimental results and computed flame speeds is fully satisfactory. Influence of small hydrogen addition on the laminar flame speed of CO/air mixtures is shown in Figure 8. Results are here plotted against various experimental data points of different investigators along with different syngas mechanism predictions. Flame speed measurements of the present study show an excellent agreement with the numerical computations presented. It is interesting to notice that the before-mentioned offset between predictions of Sivaramakrishnan et al. [22] and predictions of other investigators is gradually reduced while the hydrogen content is further decreased. A more detailed analysis of the mechanisms will be performed to identify key reaction parameters responsible for the observed differences. It is also worth mentioning that the laminar flame speeds of CO/air mixtures are very sensitive to small hydrogen additions: at an equivalence ratio of 0.6, the 1/99% H2/CO mixture flame speed is approximately 10 cm/s, while 10/90% H2/CO mixtures display a laminar flame speed higher than 20 cm/s. This nonlinear increase in the flame speed for the first added 10% of hydrogen has been previously observed experimentally by Scholte and Vaags [10] and further explained by Vagelopoulos and Egolfopoulos [13], identifying the main CO oxidation reaction (CO + OH = CO2 + H) as a key reaction for such mixtures. On the other hand, flame speeds obtained for mixtures with high hydrogen contents show less sensitivity for small CO additions (Not shown here).
Figure 8. Laminar flame speeds of 10/90, 5/95 and 3/97 and 1/99% H2/CO mixtures, lines represent computations performed with the mechanism of Sun et al.[16], Sivaramakrishnan et al. [22] and Li et al. [24] solid symbols are experimental data from the present study, open symbols experimental data by various authors.
In order to study the laminar flame speed of ultra lean syngas mixtures, a premixed methane/air pilot flame was used to achieve a better flame stabilization. Results for three different mixture compositions are presented in Figure 9. Experimentally, the stability of the studied flames was greatly improved, reaching equivalence ratios close to 0.3 for mixtures with higher contents of CO. Cellular instabilities previously observed at an equivalence ratio of 0.6 close to flame blow-off conditions were in this case observed at much leaner conditions. However, it can be noticed that flame speeds obtained with the pilot flame seem to be systematically higher than the values suggested by the trend without a pilot flame. For instance, duplicated measurements for the 50/50% H2/CO mixture for an equivalence ratio of 0.62 show that the determined flame speed with the pilot flame is approximately 7 cm/s higher than the case without pilot flame. This indicates that heat transfer from the pilot flame to the unburned mixture is not negligible and consequently, cooling of the burner would be required.
Figure 9. Laminar flame speeds of 10/90, 50/50 and 70/30% H2/CO mixtures with (Solid symbols) and without (Open symbols) pilot flame. Lines represent computations performed with the mechanism of Sun et al. [16]
Nozzle Burner Development and Technique Comparison To achieve better flame stabilization for lean combustion and reduce the flame flashback propensity as far as reducing heat transfers from the pilot flame to the unburned mixture, two stainless steel nozzle burners, integrated with a cooling head, have been designed and fabricated. Both contractions have an inlet diameter of 50 mm and outlet diameter of 4 and 8 mm respectively. Their design is based on the works of Cohen et al. [35]. The cross sectional velocity at each burner exit has been measured using a hot-wire anemometer to ensure that the contraction nozzle allows a flow velocity nearly constant at the burner exit. Cold mixture velocity profiles for the 4 and 8 mm nozzle burners are displayed in figures 10 and 11. Both contractions show flat velocity profiles over about 63% and 71% of the nozzle diameter respectively. By elimination of the steep velocity gradient at the exit boundary, the fabricated nozzles were expected to enhance the flame stability and allow the formation of straight-sided flame cones required to accurately apply the flame cone angle method.
Figure 10. Velocity profile comparison for the 4 mm nozzle burner (3mm above burner tip - Total flow 1-D: 3184 sccm, straight burner: 3162 sccm and nozzle burner: 3153 sccm)
Figure 11. Velocity profile comparison for the 8 mm nozzle burner (3mm above burner tip - Total flow 1D: 3290 sccm, straight burner: 3222 sccm and nozzle burner: 3249 sccm) A considerable discrepancy for flame speeds extracted from two measurement techniques is shown in Figure 12 in case of 50/50% H2/CO mixture. Flame speeds determined from
chemiluminescence images are lower up to 30 % in comparison to flame speeds calculated from schlieren images. Results for the 40/60 H2/CO% mixture for the 8 mm nozzle burner appear to be more comprehensive. A maximum 15% decay between both methods can be noticed, and results obtained from the total area method for the nozzle burner closely reproduce those established for straight burners of different diameter sizes. In general, the cone angle method gives higher flame speed results than the flame surface area. Differences observed between the two set of measurements can be further explained by examining the schlieren and chemiluminescence images shown figure 12 and 13. While flames formed on the 8 mm nozzle burner display very straight-sided cones, flames stabilized on the 4 mm burner seem less affected by the presence of the nozzle and flanks of the flame are more curved. This would seem that the 4mm diameter nozzle was inappropriate to measure the flame speed for the range of flow rates considered. These observations agree with Andrews et al.36 review on the determination of burning velocities mentioning that the flame cone angle is strongly dependant on gas flow rates and burner design. In general, large burner diameters are selected for the cone angle method to minimize the influence of the curvature at the flame base and tip. In practice, this requirement is restrictive for the range of investigated composition as syngas mixtures with higher hydrogen contents cannot be properly stabilized on large burner diameters. Influence of the nozzle cooling, pilot flame and stabilisation of ultralean syngas flames on the nozzle burners will be considered in the next study.
Figure 12. Comparisons for the determination of the laminar flame speed of 50/50% H2/CO + Air mixture. NB = Nozzle Burner, SB = Straight Burner, computations performed with the mechanism of Sun et al. [16].
Figure 13. Comparisons for the determination of the laminar flame speed of 40/60% H2/CO + Air mixture. NB = Nozzle Burner, SB = Straight Burner, computations performed with the mechanism of Sun et al. [16] V. Summary and Future Work An extensive series of laminar flame speeds with various syngas mixtures have been performed using straight cylindrical burners. Results from the chemiluminescence technique show a good agreement with the experimental works of Natarajan19 et al. and McLean14 et al. and also with predictions calculated with recent H2/CO mechanisms. In the attempt to study mixtures in the ultra lean domain, flame flashback and flame tip opening phenomena strongly limited the equivalence ratio range of investigation, particularly for mixtures with higher hydrogen contents. Further measurements will be conducted using the designed contractions believed to improve the flame stability at leaner conditions. A parallel study of syngas flame speeds has been engaged at the CNRS Orléans using the counterflow twin flame apparatus enclosed in a high pressure chamber along with Particle Imaging Velocimetry. This configuration will allow a more accurate determination of laminar flame speeds by minimizing heat losses, flame stretch and curvature effects encountered during the present study.
Acknowledgments This work is supported by the CNRS and The Pennsylvania State University Propulsion Engineering Research Center. The authors thank Mr. Larry Horner for his help in fabrication and assembly of the different burner setups used for the present study. References 1Wender,
I., “Reactions of Synthesis Gas”, Fuel Processing Technology, Vol. 48, 1996, pp. 189-297. N.N., Miraglia, Y.C., Raine, R.R., Bansal, P.K., Elder, S.T., “Spark Ignition Engine Performance with ‘Powergas’ Fuel (Mixture of CO/H2): A Comparison with Gasoline and Natural Gas”, Fuel, Vol. 85, 2006, pp. 1605-1612. 3Moliere, M., “Benefiting from the Wide Fuel Capability of Gas Turbines: A Review of Application Opportunities”, 2Mustafi,
Proceedings of the ASME Turbo Expo, GT-2002-30017, Vol. 1, ASME, New York, 2002, pp.227-238. 4Richards, G.A., McMillian, M.M., Gemmen, R.S., Rogers, W.A., Cully, S.R., “Issues for Low-Emission, Fuel-Flexible Power Systems”, Progress in Energy and Combustion Science, Vol. 27, 2001, pp 141-169. 5Brdar, R.D., Jones, R.M., “GE IGCC Technology and Experience with Advanced Gas Turbines”, GE Power Systems, GER4207, 2000. 6Lee, J.G., Kim, J.H., Lee, H.J., Park, T.J., Kim, S.D., “Characteristics of Entrained Flow Coal Gasification in a Drop Tube Reactor”, Fuel, Vol. 75, 1996, pp. 1035-1042. 7Lieuwen, T., Richards, G., “Burning Questions”, Mechanical Engineering, Vol. 128, 2006, pp. 40-42. 8De Biasi, V., “Targeting Turbine Research Needed to Burn Syngas and Hydrogen Fuels”, Gas Turbine World, Vol. 35 2005, pp. 28-30. 9Scholte, T.G., Vaags, P.B., ”The Influence of Small Quantities of Hydrogen and Hydrogen Compounds on the Burning Velocity of Carbon Monoxide”, Combustion and Flame, Vol. 3, 1959, pp. 503-510. 10T.G. Scholte, P.B. Vaags, “Burning Velocities of Mixtures of Hydrogen, Carbon Monoxide and Methane with Air”, Combustion and Flame, Vol. 3, 1959, pp. 511-524. 11Günther, R., Janish, G., “Messwerte der Flammengeschwindigkeit von Gasen und Gasgemischen“, Chemie Ingenieur Technik, Vol. 43, N° 17, 1971, pp. 975-978. 12Yumlu, V.S., “Prediction of Burning Velocities of Carbon Monoxide-Hydrogen-Air Flames”, Combustion and Flame, Vol. 11, 1967, pp. 190-194. 13Vagelopoulos, C.M., Egolfopoulos, F.N., “Laminar Flame Speeds and Extinction Strain Rates of Mixtures of Carbon Monoxide with Hydrogen, Methane and Air”, Proceedings of the Combustion Institute, Vol. 25, 1994, pp. 1317-1323. 14McLean, I.C., Smith, D.B., Taylor, S.C., “The Use of Carbon Monoxide/Hydrogen Burning Velocities to Examine the Rate of the CO + OH Reaction”, Proceedings of the Combustion Institute, Vol. 25, 1994, pp. 749-757. 15Brown, M.J., McLean, I.C., Smith, D.B., Taylor, S.C., “Markstein Length of CO/H2/Air Flames, using Expanding Spherical Flames, Proceedings of the Combustion Institute, Vol. 26, 1996, pp.875-881. 16Sun, H., Yang, S.I., Jomaas, G., Law, C.K., “High Pressure Laminar Flame Speeds and Kinetics Modeling of Carbon Monoxide/Hydrogen Combustion”, Proceedings of the Combustion Institute, Vol. 31, 2007, pp. 439-446. 17Hassan, M.I., Aung, K.T., Faeth, G.M., “Properties of Laminar Premixed Flames at Various Pressures”, Journal of Propulsion and Power, Vol. 13, N° 2, 1997, pp. 239-245 18Huang, Y., Sung, C.J., Eng, J.A., “Laminar Flame Speed of Primary Reference Fuels and Reformer Gas Mixtures”, Combustion and Flame, Vol. 139, 2004, pp. 239-251. 19Natarajan, J., Nabdula, S., Lieuwen, T., Seitzman, J., “Laminar Flame Speeds of Synthetic Gas Fuel Mixtures”, Proceedings of GT2005, ASME Turbo Expo 2005, GT2005-68917, June 2005. 20Love, N.D., Periasamy, C., Gollahalli, S.R., Choudhuri, A.R., “Laminar Burning Velocity of Synthetic Gas Premixed Flames Near Extinction Conditions”, 4th International Energy Conversion Engineering Conference and Exhibit, AIAA 20064122, June 2006. 21Subramanya, M., Choudhuri, A., “Investigation on the Flame Extinction Limit of Fuel Blends”, 41st AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA 2005-3586, July 2005. 22Sivaramakrishnan, R., Comandini, A., Tranter, R.S., Brezinsky, K., Davis, S.G., Wang, H., “Combustion of CO/H2 Mixtures at Elevated Pressures”, Proceedings of the Combustion Institute, Vol. 31, 2007, pp. 429-437. 23Davis, S.G., Joshi, A.V., Wang, H., Egolfopoulos, F., “An Optimized Kinetic Model of H2/CO Combustion”, Proceedings of the Combustion Institute, Vol. 30, 2005, pp. 1283-1292. 24Li, J., Zhao, Z., Kazakov, A., Chaos, M., Dryer, F.L., Scire, JR., J.J., International Journal of Chemical Kinetics, Vol. 39, 2007, pp. 109-136. 25Dash, C.J., “One-Dimensional Tomography: A Comparison of Abel, Onion-peeling, and Filtered Backprojection Methods”, Applied Optics, Vol. 31, N°8, 1992, pp. 1146-1152. 26Lewis, B., Von Elbe, G., Combustion, Flames and Explosions of Gases, third ed., Academic Press, Orlando, 1987, p. 283. 27Lamoureux, N., Djebaïli-Chaumeix, N., Paillard, C.-E., “Laminar Flame Determination for H2-Air-He-CO2 Mixtures Using the Spherical Bomb Method”, Experimental Thermal and Fluid Science, Vol. 27, 2003, pp. 385-393. 28Tse, S.D., Zhu, D.L., Law, C.K., “Morphology and Burning Rates of Expanding Spherical Flames in H2/O2/Inert Mixtures up to 60 Atmospheres, Proceedings of the Combustion Institute, Vol. 28, 2000, pp. 1793-1800. 29Kumar, K.S., “Laminar Burning Velocities of Lean Hydrogen-Air Mixtures”, Explosion Dynamics Laboratory Report, FM97-15, 1998. 30Dowdy, D.R.D., Smith, D.B., Taylor, S.C., “The Use of Expanding Spherical Flames to Determine Burning Velocities and Stretch Effects in Hydrogen/Air Mixtures, Proceedings of the Combustion Institute, Vol. 23, 1990, pp. 325-332. 31Egolfopoulos, F.N., Law, C.K., “An Experimental and Computational Study of the Burning Rates of Ultra-Lean to Moderately-Rich H2/O2/N2 Laminar Flames with Pressure Variations”, Proceedings of the Combustion Institute, Vol. 23, 1990, pp. 333-340. 32Law, C.K., Ishizuka, S., Cho, P., “On the Opening of Premixed Bunsen Flame Tips”, Combustion Science and Technology, Vol. 28, 1982, pp. 89-96. 33Buckmaster, J., “A Mathematical Description of Open and Closed Flame Tips”, Combustion Science and Technology, Vol. 20, 1979, pp. 33-40. 34Kee, R.J., Grcar, J.F., Smooke, M. D., Miller, J.A., Technical Report SAND85-8240, Sandia National Laboratories, Albuquerque, NM, 1987. 35Cohen, M.J., Ritchie N.J.B., “Low-Speed Three-Dimensional Contraction Design”, Journal of the Royal Aeronautical Society, Vol. 66, 1962, pp. 231-236. 36Andrews, G.E., Bradley, D., “Determination of Burning Velocity: A Critical Review”, Combustion and Flame, Vol. 18, 1972, pp. 133-153.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC:
Coal-Derived Products
- Coal Utilization By Products (Ash, Fertilizers, Etc,)
Structural Changes in Bituminous Coal Fly Ash Due to Treatments with Aqueous Solutions Roy Nir Lieberman Department of Biological Chemistry, Ariel University Center of Samaria, Israel Contact Information: Ariel, 40700, Israel, Phone: 00972-547-699701, fax: 00972-9-7417429, email: [email protected] Haim Cohen1,2, Ariel Goldman3, Roy Nitzsche1,4 1. Department of Biological Chemistry, Ariel University Center at Samaria, Ariel, 40700 Israel, phone: 00972-52- 4306878, fax: 00972-8-9200749, email: [email protected] 2. Chemistry Department, Ben-Gurion University of the Negev, Beer Sheva, Israel 3. Department of Civil Engineering, Ariel University Center at Samaria, Ariel, 40700, Israel 4. TU Bergakademie Freiberg, Fakultät 4, Institut für Energieverfahrenstechnik und Chemieingenieurwesen 09599 Freiberg, Germany.
Abstract: Coal fly ash is produced in Israel via the combustion of class F bituminous coals. The bulk of coal fly ashes produced in Israel stems from South African and Colombian coals and therefore these ashes were the subject of the present study. It has been shown that during treatment of the flyash with aqueous solutions appreciable Structural Changes in the Matrix do occur. The flyash can be used as a scrubber and fixation reagent for acidic wastes. Recently it was reported that the scrubbing product can serve as a partial substitute to sand and cement in concrete. The bricks have proved to be strong enough according to the concrete standards. In order to have a better understanding of the fixation mechanism we have decided to treat the flyashes with water or acidic solution (0.1M HCl) thus changing the surface of the flyash particles. Surface analysis of the treated and untreated fly ashes have demonstrated that the treated flyashes have changed appreciably its' interactions with transition metal ions. Three possible modes of interactions were observed: cation exchange, chemical bonding and electrostatic adsorption of very fine precipitate at the flyash surface.
Influence Factors on Density and Specific Surface Area (Blaine Value) of Fly Ash from Pulverized Coal Combustion Hiromi Shirai, Michitaka Ikeda and Kenji Tanno Central Research Institute of Electric Power Industry Address: 2-6-1 Nagasaka, Yokosuka-shi, kanagawa-ken 240-0196, Japan Phone: +81 46 856 2121 Fax: +81 46 856 3346 Email: [email protected] Abstract In the Japanese electric power industry, it is desirable to reduce the cost of the treatment and expand the range of use of fly ash, such as in concrete admixtures. Therefore, it is necessary to form high-quality fly ash. To obtain high-quality fly ash, it is important to clarify the factors affecting its properties. In this study, the factors affecting the density and specific surface area (Blaine value) of coal fly ash were investigated on the basis of experimental results obtained using our combustion test facility and the ash data from the boiler of an actual electric power plant. The density was affected by the ash particle size, the true density of the component materials and aluminum content which is closely related to the fusibility. The specific surface area was affected by the particle size distribution and particle shape. The shape was affected not only by the ash particle size but also by the unburned carbon concentration, the ash fusibility and the ash content in the coal. It was also found that the specific surface area of the ash generated from our combustion test facility is higher than that of ash from actual boiler for ash with the same particle size. This result indicates that the shape of particles is affected by their heating and formation histories in the boiler. On the basis of the above findings, correlation equations were obtained for the density and Blaine value.
2. Experimental Section 2.1 Coal sample In the study using our combustion test facility, 18 types of
bituminous coal were used. The range of properties of each coal type is shown in Table 1. Coals with different combustibilities and fusibilities were selected to cover the range of properties of coal burned in Japanese pulverized coal-fired power plants, from which fly ash is produced.
Table 1 Coal properties Property
Our test furnace
Moisture [wt%]*
2 - 13
1-9
Ash [wt%]**
7 - 17
5 - 15
Volatile matter [wt%]**
31 - 47
28 - 45
Fixed carbon [wt%]**
41 - 63
42 - 58
Fuel ratio [-]
0.8 - 2.3
1.0 - 2.1
HHV [MJ/kg]** Ash composition [wt%]
1. Introduction The amount of fly ash discharged from existing pulverized coal-fired power plants is increasing and has reached over 10 million tons per year in Japan. The ratio of fly ash in Japan that is effectively used is more than 90%. However, over 60% of fly ash is treated so that it can be used as a material in cement. In the Japanese electric power industry, it is important to reduce the cost of the treatment and expand the range of the use of fly ash, such as in concrete admixtures. Therefore, it is necessary to form high-quality ash. The Japanese Industrial Standard (JIS) for the use of fly ash in cement classifies fly ash into three grades. The grade is determined by the properties of the fly ash (particle size, unburned carbon concentration, density, specific surface area obtained by the Blaine method) and the properties of the mortar (fluidity, solidity) to which the fly ash is added. To manage the ash quality, it is important to develop a method of predicting the ash properties. Various methods (1),(2),(3) of predicting particle size and unburned carbon concentration have been proposed. However, the study (4) of the prediction of ash density and specific surface area has been insufficient. The purpose of this study is to clarify influence factors on the density and specific surface area (Blaine value) from experimental results obtained using our combustion test facility and the ash data from the boiler of an actual electric power plant.
Actual boiler***
27 - 32
27 - 31
SiO2
45 - 73
47 - 75
Al2O3
18 - 41
15 - 30
Fe2O3
1 - 13
2 - 11
CaO
0.5 - 10
0.8 - 13
MgO TiO2
0.2 - 1.8
0.2 - 2.0
0.5 - 3.5
0.6 - 1.6
P2O5
0.05 - 2.1
0.07 - 1.3
Na2O
0.05- 1.7
0.1 - 1.7
K2O Unburned carbon
0.7 - 3.3
0.4 - 3.3
1.6 - 14.5
0.9 - 7.7
* At equilibrium humidity, ** Dry basis, ***Blended coal properties are included.
Temp. controller (Simulated GGH)
Stack
ESP
Fuel Supply Equipment
Lime-limestone
Pulverized Coal Refuse Biomass Furnace
Temp. controller (Simulated GGH)
AH
De-SOx Orimulsion Heavy oil
De-NOx
Water-cooled Gas Cooler Bag Filter
Scrubber with NaOH Soulution
Figure 1 Schematic diagram of the pulverized coal combustion facility – 50µm. Mixed ash collected from all parts of the facility, except the bottom of the furnace, was used as a sample. On the other hand, the two-stage combustion is being introduced in Japan, t. However, the two-stage combustion ratio at an actual boiler is unknown. The size of pulverized coal particles in two-stage combustion is similar to that in our furnace.
Air injection port for the two-stage combustion
900
400
7,991
1,900
No.3 Burner No.2 Burner No.1 Burner
1,000 1,000
1,600
Sampling port
Burner
Front view
Side view
Figure 2 Details of the furnace 2.2 Pulverized coal combustion facility A schematic diagram of the pulverized coal combustion facility is shown in Figure 1. The facility consists of a pulverizer, a furnace with three burners and a gas treatment system. The coal feed rate is determined by the thermal input of coal. The thermal input is 760 kW×3 burners. The coal feed rate is approximately 100 kg/h×3 when high-quality bituminous coal is burned. Details of the furnace are shown in Figure 2. The furnace has a width of 0.9 m and a depth of 1.9 m, a height of 9.5 m. The mean resident time of the combustion gas is approximately 4 s at 1200 oC. The wall has a water wall structure whose inert surface is covered with a refractory material. In this study, the outlet O2 concentration was set to 4%, which is equivalent to an excess air ratio of 1.24. The standard two-stage combustion ratio is 30%, and the median diameter of the pulverized coal particle size is 35
2.3 method of measuring ash properties The ash particle size is measured using a laser diffraction particle size analyzer. The particle density was determined by the displacement of ethanol in a Le Chatelier flask. The specific surface area was determined by the Blaine method, in which the specific surface area is derived from the resistance to the flow of air through a layer of ash particles using a known standard material. 3. Results and Discussion 3.1 Ash particle density The ash particle density ρash [g/cm3] depends on the true density without closed void ρt [g/cm3] and the ratio of closed void volume to a particle volume (the void ratio: ε [−] ). The particle density is given by
ρ ash = ρ t (1 − ε )
(1)
The factors affecting the true density and the void ratio are investigated. True density The true density is the average density of the composed materials. However, it is difficult to analyze the composition of materials because a large part of the materials is amorphous. Therefore, the true density was estimated by the chemical analysis (JIS M 8815) of oxides in the ash. The selected oxides were SiO2(density ρi: 2.2 g/cm3), 3Al2O3・2SiO2(3.1), Fe2O3 (5.2), CaO(3.4), MgO(3.7), P3O4(2.4), TiO2(4.2), Na2O (2.3) and K2O (2.4). In this study, all Al2O3 was assumed to be mullite (3Al2O3・2SiO2) because more than 50% of Al2O3
3.0
0.30
Data from our test facility 2.9
Original ash (Dv-ash:22-25µm)
2.8
Pulverized ash (Dv-ash:4.0-5.5µm)
‐2%
Dv-ash:20-25μm
Temperature at liquid phase ratio of 90 wt%
0.20
2.6
ε [-]
Measured density [g/cm3]
Melting point (JIS M 8801)
0.25
2.7
2.5
0.15
0.10
2.4 0.05
2.3
0.00 1,200
2.2
1,300
1,400
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
Figure 3 Comparison between estimated true density and the density after removing closed voids by grinding ash
1,800
Newlands two-stage combustion ratio:30% D v-coal:16μm Newlands two-stage combustion ratio:30% D v-coal:49μm Newlands two-stage combustion ratio:30% D v-coal:91μm Newlands two-stage combustion ratio:0% Dv-coal:49μm Newlands two-stage combustion ratio:15% D v-coal:49μm Newlands two-stage combustion ratio:40% D v-coal:49μm Wambo two-stage combustion ratio:30% Dv-coal:31μm Wambo two-stage combustion ratio:30% Dv-coal:48μm Wambo two-stage combustion ratio:30% Dv-coal:69μm
Data from our test facility Dv-ash:20-25µm
0.25
0.20
ε [-]
0.35
0.20
1,700
0.30
Estimated true density ρa [g/cm3]
0.25
1,600
Figure 5 Influence of fusibility of ash on void ratio
2.0
0.30
1,500
Temperatur [oC]
2.1
ε [-]
Data from our test facility
Fluid point (JIS M 8801)
+2%
0.15
0.10
0.05
0.00
0.15
0
5
10
15
20
25
30
35
40
45
50
55
60
WAl [wt%]
0.10
Figure 6 Influence of Al2O3 content on void ratio
0.05 Data from our test facility 0.00 10
15
20
25
30
Dv-ash [µm]
Figure 4 Influence of ash particle size on void ratio exists as mullite. The content ratio Wi [wt%] of each oxide was recalculated so that the total content ratio of the oxides was 100 wt%. The estimated true ash particle density ρa [g/cm3] was estimated using the following equation.
ρa =
∑
100 (Wi / ρ i )
(2)
The ash was also ground to a volume mean particle diameter DV-ash [μm] of less than 6 μm to remove as many of the closed voids as possible. DV-ash is defined as
DV − ash =
∑ ( XiDi ) ∑ Xi
(3)
where Di [μm] is the diameter and Xi [-] is the weight ratio of the particles. The comparison between ρa and the density of pulverized ash particle is shown in Figure 3. It was found that ρa is almost equal to the pulverized ash density. This result clarified that our method can approximately estimate the true density.
Closed voids The void ratio ε [−] was calculated from the measured ash particle density and ρa. The factors affecting the formation of closed voids were clarified using the void ratio. The relationship between ε and DV-ash is shown in Figure 4. ε was strongly affected by the ash particle size; ε increases with decreasing DV-ash. Small ash particles are formed from the small coal particles, which burn rapidly and whose temperature also increases rapidly during combustion. Additionally, the number of coalescent mineral particles is less than that of large coal particles. Small coal particles are advantageous for the coalescence and melting of mineral particles. Therefore, ε is lower in smaller ash particles. On the other hand, ε for Newlands coal was different from that of Wambo coal at the same particle size range. This indicates that there are other factors strongly affecting ε. Other possible factors were investigated. ε did not correlate with the fuel ratio, which is the index of combustibility. Moreover, ε did not correlate with the melting point of the ash. However, it was found that ε correlates with the temperature at which the liquid phase fraction is 90%, which is the weight ratio of the liquid phase of ash to the total of the solid and liquid phases of ash as shown in Figure 5. These phases were estimated by FactSage 5.5 which is a thermochemical software and database package. Moreover, it
12,000
2.8 Our test furnace (R:0.63) Boiler A (R:0.92) Boiler B (R:0.74) Boiler C (R:0.78) Boiler D (R:0.94)
2.6 2.5
+3% -3%
R: Correlation coefficient
ρash-est [g/cm3]
Our test furnace Boiler A Boiler B Boiler C Boiler D
10,000
SAb [cm2/g]
2.7
2.4
8,000
6,000
2.3
4,000
2.2
2,000
2.1 0
2.0
0
5
10
15
20
25
Ds-ash [μm]
1.9
Figure 8 Influence of ash particle size on specific surface area (Blaine value)
1.8 1.8
1.9
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8 4.0
ρash-act[g/cm3]
Figure 7 Correlation between measured density and estimated density
3.5 3.0 2.5
SAb/SAd [-]
was found that the formation of mullite affects the above temperature. Figure 6 shows the correlation between ε and the content ratio of Al2O3 WAl [wt%] which comprise in the fly ash in the form of mullite. This correlation shows that the content of Al2O3 is an important factor affecting ε.
2.0 1.5 Our test furnace Boiler A Boiler B Boiler C Boiler D
1.0 0.5
Correlation equation for ash particle density The correlation equation for the ash particle density was set up as follows:
ρ ash − est = ρ a (1 − ( kd1 + kd 2 × (W Al / 100 ) + kd 3 × (W Al / 100 ) 2 + kd 4 × D p ))
(4)
where kdi is the coefficient calculated by the least-squares method. The correlation between the actual density ρash-act [g/cm3] and the estimated density ρash-est [g/cm3] is shown in Figure 7. It was clarified that the factors affecting density revealed in this study is correct, because the difference between two densities is within 5% of the actual density. 3.2 Specific surface area In JIS, the specific surface area of the ash, SAb [cm2/g-ash], is measured by the Blaine method. However, SAb is not the true specific surface area of the ash, though it is an important value for evaluating the outer surface area excluding the pore surface area. The specific surface area of a powder is affected by the particle size distribution and particle shape. Particle size The specific surface area is strongly affected by the surface area of fine particles. Therefore the influence of the ash particle size on SAb is investigated using the surface mean diameter DS-ash [µm], which is defined as D S − ash =
∑ Xi
∑ ( Xi Di )
(5)
0.0 0
5
10
15
20
25
30
35
Ds-ash [μm]
Figure 9 Influence of ash particle size on shape of ash particles The relationship between DS-ash [µm] and SAb is shown in Figure 8. If the shapes of the particles are the same, SAb increases as DS-ash decreases. However, no clear dependence of DS-ash on SAb was observed. This indicates that the influence of the particle size on SAb is weaker than that of the particle shape. Particle shape The specific surface area SAd [cm2/g-ash] when ash particles were assumed to be spherical was calculated from the ash particle size distribution and ash density. The relationship between DS-ash and the ratio of SAb to SAd, which is the reciprocal number of Carman’s shape factor, is shown in Figure 9. The ratio increases with the ash particle diameter. This indicates that the sphericity of the ash particles decreases as the diameter increases. It is estimated that the number of coalesced particles of minerals affects the particle shape similarly to its influence on density. Next, SAb/SAd in our furnace was compared with that in an actual boiler. For the same particle size range, SAb/SAd in our furnace was observed to be higher than that in the boiler. This result indicates that the sphericity of particles in our facility is lower than that in the boiler because the resident time of ash particles at the high-temperature zone of our furnace is shorter than that of the boiler.
5.0
4.0
Ds-ash:9-10µm
Data from our test furnace 3.5
Ds-ash:11-13µm
4.0
3.0
3.5
2.5
SAb/SAd [-]
SAb/SAd [-]
4.5
Data from our test furnace
3.0 2.5
1.5
2.0
1.0
1.5
0.5
1.0 0
2
4
6
8
10
12
0.0 1,200
14
Uc [wt%]
1,300
1,400
1,500
1,600
1,700
Melting point [oC]
Figure 12 Influence of fusibility of ash on shape of ash particles
Figure 10 Influence of unburned carbon concentration on shape of ash particles
1800
4.0 Uc:1-3wt%, Ds-ash:10-13µm
Data from our test furnace
Data from our test furnace
Uc:5-6wt%, Ds-ash:10-13µm
1700
Melting point [oC]
3.5
3.0
2.5
2.0
1.5
1600
1500
1400
1300
1.0 4
6
8
10
12
14
1200
16
0.0
Ash content [wt%]
1.0
2.0
3.0
4.0
5.0
Al2O3/(Fe2O3+CaO+MgO+Na2O+K2O) [-]
Figure 13 Relationship between index of ash fusibility and melting point
Figure 11 Influence of ash content on shape of ash particles
4.0
Furthermore, it was found that the unburned carbon concentration UC [wt%], which is affected by the fuel ratio and the combustion conditions such as the coal particle size and two-stage combustion ratio, affects SAb/SAd, as shown in Figure 10. The particles containing unburned carbon have a complex shape. In ash with a high unburned carbon concentration, SAb may not be correctly measured because the Blaine method cannot be efficiently applied to the particles with greatly different shapes. It was also found that the ash content affects the shape of ash particle as shown in Figure 11. The sphericity may decrease as the ash content Wash [wt%] increases because the number of coalesced particles of minerals affects the particle shape. On the other hand, SAb/SAd is affected by the fusibility of ash. As shown in Figure 12, it was confirmed that the melting point is correlated with SAb/SAd. In addition, to estimate SAb/SAd from the ash composition without measuring the melting point, an index that correlates with the melting point was investigated. It was also confirmed that Al2O3 / (Fe2O3+CaO+MgO+Na2O+K2O) (5) can be used for this purpose as shown in Figure 13. Moreover, it was found that the index is strongly correlated with SAb/SAd as shown in Figure 14.
Data from our test furnace 3.5 3.0
SAb/SAd [-]
SAb/SAd [-]
2.0
2.5 2.0 1.5 1.0 0.5 0.0 0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
Al2O3/(Fe2O3+CaO+MgO+Na2O+K2O) [-]
Figure 14 Relationship between index of ash fusibility and shape of ash particles
Correlation equation for SAb/SAd The correlation equation for SAb/SAd was set up as follows:
4.0 Our test furnace (R:0.95) Boiler A (R:0.93) Boiler B (R:0.93) Boiler C (R:0.89) Boiler D (R:0.99)
3.5
( S Ab S Ad ) est = ks1 ×(Wash
/ 100) ks2
× (U C
/ 100) ks3
3.0
Findex =
Al2O3 (Fe 2O3+CaO+MgO+Na2O+K2O)
Estimated SAb/SAd [-]
× Findex ks4 × Ds −ash ks5
(4)
where Ksi is the coefficient calculated by the least-squares method. The correlation between the actual SAb/SAd and the estimated SAb/SAd is shown in Figure 15. It was clarified that the factors affecting the shape in this study is correct, because of the strong correlation between the actual SAb/SAd and the estimated SAb/SAd. Moreover, the correlation between the measured SAb and the estimated SAb was calculated from the estimated SAb/SAd by equation (4) and SAd is shown in Figure 16. This indicates that SAb can be estimated from the estimated SAb/SAd and SAd .
From these results, it was found that the ash particle size distribution, ash composition and unburned carbon concentration are important factors in developing a method for predicting the density and specific surface area of fly ash. Reference (1)” Improvement of Pulverized Coal Combustion Technology for Power Generation”, CRIEPI Review No.46; Central Research Institute of Electric Power Industry: 2002. (2) Yan, L.; Gupta, P. R.; Wall, F. T. Fuel, 81, 337–344, 2002 (3) Barta, L. E.; Toqan, M. A.; Beer, J. M.; Sarofim, A. F. 24th Int. Conf. Comb., 1135–1144, 1992 (4) Shirai, H.; Tsuji, H.; Ikeda, M.; Kotsuji T., Energy & Fuels 2009, 23, 3406–3411 (5) NEDO report NEDO-C-9940, 2000
-15%
2.5 2.0 1.5 1.0 0.5 0.0 0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Actual SAb/SAd [-]
Figure 15 Correlation between actual SAb/SAd and estimated SAb/SAd 10,000 +15% 8,000 -15%
Estimated SAb [cm2/g]
4. Conclusion In this study, the influences of combustion conditions and coal properties on the ash particle size, density and specific surface area were investigated on the basis of experimental results obtained using our combustion test facility and the ash data from the boiler of an actual electric power plant. (1) Ash particle density The density was affected by the ash particle size, the true density of the component materials and aluminum content which is closely related to the fusibility. (2) Specific surface area of ash The specific surface area was strongly affected by the shape of the ash particles. The shape was affected not only by the ash particle size but also by the unburned carbon concentration in the ash, the ash fusibility and the ash content in the coal.
+15%
6,000
4,000 Our test furnace (R:0.84) Boiler A (R:0.91) Boiler B (R:077) Boiler C (R:0.75) Boiler D (R:0.92)
2,000
0 0
2,000
4,000
6,000
Measured SAb [cm2/g]
8,000
10,000
Figure 16 Correlation between measured SAb and estimated SAb
2010 International Pittsburgh Coal Conference
DEVELOPMENT OF COMMERCIAL CFBC BOILR FOR REFUSE DERIVED FULE Dowon Shun*, Dal-Hee Bae, Jaehyeon Park, Seung Yong Lee Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA *[email protected] Introduction Korea officially started to promote waste fuel utilization by proclaiming the regulation regarding quality standards and use of recycled plastics fuel in year 2003. Since then continuing enactment of related law activated the business of waste energy recovery. Regulations defining refuse derived fuel(RDF) which is made of municipal solid waste, wood chip fuel(WCF), sludge derived fuel(SRF) are proclaimed successively. So far the regal background of waste fuel business is established and business sectors are about to start(MOE, 2006; 2009). The main structure of waste fuel business is in thermal utilization, i.e., combustion. Among various combustors circulating fluidized bed combustion (CFBC) is considered to be the optimum (Shun, et al., 2004; 2006). CFBC can handle the fuel with varieties of shape, heating value, moisture and mineral content. It also can control sulfur dioxide, NOx, HCl inside the combustor, thus it can alleviate the burden of flue gas cleaning system. The design concept of CFBC for RDF is different from conventional CFBC burning coal. The RDF burning CFBC should handle HCl instead of SO2 and above all it must show durability during the operation in the corrosive environment. The plant must be equipped with HCl removing system from the flue gas. Steam temperature seldom goes up above 723K to reduce the affect of high temperature corrosion. Often the flue gas temperature is lowered below 923K at the upstream of the final superheater. The density of fuel is much lighter than that of coal and its ash is finer. Cyclone collection efficiency must be adjusted according to the particle size of fly ash. A new design of CFBC is conceived for Korean RDF and its pilot plant was constructed with the rate of 8ton of steam per hour. Feasibility of the design is evaluated and design parameter was determined by burning RDF and RPFs. Emission level for the design of flue gas treating system was measured. With the evaluation of all those design and operating variables, a pilot scale (1MWe equivalent) CFBC for RDF was built and its performance was investigated. Experimental data collected from the plant were analyzed and a design of 10MWe CFBC for RDF was made. Design of a pilot scale CFBC boiler for RDF A circulating fluidized bed combustion (CFBC) boiler with 8ton/h steam rate was designed and constructed. The maximum steam quality is 450 deg.-C and 38ata. The boiler circuit was arranged as economizer → drum → wall evaporator → drum → super heater 1 → super heater 2. The system consists of a combustor with evaporator, a hot cyclone, a convection pass with superheater, economizer, and air heater. Flue gas was cleaned by dry absorber gas cleaner in series with a bag house. After the bag house there was a wet scrubber to remove HCl and cleaned gas passed through a stack to environment. Figure 1 shows the schematic of the facility. The produced steam was totally condensed and returned to the feed water tank again.
To Steam Generator
Drum
Cyclone
S/H3
S/H1
Combustor
Scrubber
RDF Hopper SDR
Econ. Water A/H
A ir
ID Fan Stack
FD Fan
Figure 1. The diagram of the CFBC pilot plant for RDF Operation and performance of the boiler Table 1 presents the characterization of RDF fuel. RDF from municipal waste is RDF 5 equivalent. It is fabricated in pillar type pellet with 30mm of diameter and 100mm of length. Table 1. Analysis of waste derived fuel Sample Moisture Volatile Dried Industrial Sludge 5.80 51.8 Agricultural PE film 0.6 97.2 RDF from municipal waste 6.7 69.2
Ash F.Carbon Sulfur Carbon HCl HHV(kcal/kg) 42.3 0.00 2.2 30.7 2773 0.14 2.1 0.2 10550 16.2 7.9 0.2 52.6 1.0 5710
Figure 2 shows variation of temperature of major combustor sector , feed rate, and steam rate. Combustor temperature varied according to feed rate. However the final steam rate could be maintained almost invariably even without de-superheating. The mission at the boiler exit and the plant exit was presented in table 2. The emission of HCl was further cleaned by flue gas treatment and particulates were captured in the bag filter. The emission level of pollutants was maintained below of the limitation of Korean regulation(MOE, 2009). Figure 3 shows fluidized bed pressure variation during the operation. Except start up and shut down period the pressure maintained stable values. Air box pressure was in between 700-800mmaq and combustor exit pressure was almost atmospheric. Table 2. Emission level during the plant operation Mean value at given O2 level At the combustor exit SO2, ppm 3.3 HCl, ppm 236 NO, ppm 74 O2, % 8.5 Particulates, mg/Sm3 12,832
At the stack 3.3 9 74 8.5 14.7
Regulation 30 20 70 12 40
1000 900
400 630
800
Temperature(C)
700
14130
600
Fuel
500
s team
400 300 200
400
100
630
14130
Steam
Fuel
0
Time(min) Figure 2. Combustor condition during operation
1,000 Air box
900
L400
L630
14130
800
Pressure(mmAq)
700
Start up
Air box Shut down
600 500 400 400
300 200
630
100
14130
0 -100
Time(min) Figure 3. Combustor pressure difference during operation
The combustion was very stable during operation. Since the fuel consists mostly of volatiles the combustion efficiency is always over 99%. Figure 4 shows the proximate analysis of fly ash. Combustible content in fly ash was too low and almost undetectable. 100%
80%
Weight percent
F.Carbon Ash Volatile
60%
Moisture 40%
20%
0% 1
2
3
4
Samples
Figure 4. Proximate analysis of the flyash Design of a 10MWe CFBC boiler for RDF A 10MWe scale CFBC for RDF was designed based on the collected data from pilot plant operation. The combustor consists of one combustor with water wall, one hot cyclone, and one convection pass. The combustor configuration is 4m x 4m x 24m. The boiler includes 2 super heater banks, one economizer and banks of air heater is located in the convection pass. Steam pressure and temperature are 45ata and 723K respectively. Maximum continuous steam rate of the boiler is designed as 60 ton/hour. Table 3 presents the design specifications of 10MWe CFBC for RDF. Table 3. The technical data of 10MWe CFBC Specifications Net heat output, MWth Steam rate, ton/h Output steam temperature, deg.-C Output steam pressure, ata Feed water temperature, deg.-C Boiler load variation, % MCR Fuel flow range, ton/h
Values 46 60 455 47 143 50-100 5.4-10.6
The 10MWe CFBC shares the flow diagram with the pilot plant thus it is same with the diagram of Figure 1. Figure 4 shows the front view of the boiler lay out. Considering that RDF is a low density fuel the combustor has 4 surge bins each has 4hours of service term. The project to build a 10MWe CFBC boiler
for RDF will be launched this year. This plant will be the first commercial scale CFBC co-generation boiler for RDF in Korea.
Figure 4. The layout of 10MWe CFBC for RDF Conclusion A pilot scale CFBC boiler with 8ton/h of steam rate is designed, constructed and performed the combustion experiment. The RDF in CFBC was burned stably and thoroughly. The temperature and pressure profiles in the combustor were stable. Combustion efficiency was over 99.5%. Emission was below the Korean emission regulation except HCl which required flue gas treatment. A 10MWe scale commercial CFBC for RDF is designed from this experience and the construction project has started. References Shun, D., Bae, D-H, Lee, S. Y, Kim, and M-S,: Development of CFB RDF combustion and co-generation demonstration plants, The ninth Asian Conference on Fluidized-Bed and Three-Phase Reactors, pp.79-84 (2004). MOE, Revision of the Regulations relative to the Application of the law of regarding Savings of Resources and Recycling Promotion, legislation of MOE (2006), MOE, Tables, Revision of the Law of Conservation of the Environment, Ministry of Environment, Korea (2009). Shun, D., Bae, D-H, Lee, S-Y, and Jo, S.,”Circulating Fluidized Bed Combustion of Refuse Derived Fuel,” J. of Korean Inst. of Resources Recycling, 15(1), PP.58-65 (2006).
SIMPLIFIED QUANTIFICATION OF TETRAFLUOROBORATE ION IN FLUE GAS DESULFURIZATION EFFLUENT FOR MANAGEMENT OF FLUORINE EMISSION Seiichi Ohyama, Hiroyuki Masaki, Shinji Yasuike, and Kazuo Sato Central Research Institute of Electric Power Industry, 1646 Abiko, Abiko-shi, Chiba 270 1194, Japan ABSTRACT Boron compounds in industrial wastewaters mainly exist as boric acid and tetrafluoroborate ion (BF4-). BF4contributes to the emission regulations of both boron and fluorine. In coal-fired power plants, BF4- is formed in a flue gas desulfurization (FGD) unit, depending on the type of FGD and coal. BF4- is quite stable and slow to decompose, which may lead to increased values of especially fluorine in discharged effluents. Thus, in some plants, a dedicated treatment to decompose BF4- is employed to manage BF4- emission. The conventional measurements of BF4-, i.e., spectrophotometry and ion chromatography requires a tedious pretreatment or a stationary equipment and are difficult to carry out onsite. The onsite measurement of BF4- is helpful for the efficient treatment of BF4-, which may lead to a reduced amount of sludge. We propose a simple and rapid analytical determination method of BF4- for properly managing the BF4- emission in FGD effluents. We build up a compact flow measurement system using BF4- ionselective electrode (ISE) mounted in a flow cell. By measuring samples in a flow mode, the stability and reproducibility of measurements is largely improved compared to the conventional static batch measurement. The system can measures BF4- in the range of 1-100 mg/L and the measurement completes in less than 10 min. The proposed method is an easier and more rapid determination compared with the conventional analytical methods. The application of the system to the FGD effluents is discussed. INTRODUCTION Boron compounds in industrial wastewaters mainly exist as boric acid and tetrafluoroborate ion (BF4-). In Japan, the Environmental Quality Standard (EQS) for water pollution, which is a target level to be achieved in public water to protect human health, is set at 1 mg/L for boron and at 0.8 mg/L for fluorine.1 Reflecting these establishments, the National Effluent Standard (NES) for boron is set at 10 mg/L and 230 mg/L for effluents discharged to terrestrial water bodies and to coastal water bodies, respectively.2 Similarly, NES for fluorine is 8 mg/L in noncoastal areas and 15 mg/L for coastal areas. There is a large difference in the coastal effluent standards of boron and fluorine. Thus, in coastal areas where almost all of Japanese power plants are located, more stringent regulation is enforced on fluorine emission compared to boron’s. Based on the elemental composition, BF4- contributes to the emission regulations for both boron and fluorine, but considering a gap between boron and fluorine in the coastal areas, it is more responsible for fluorine emission than for boron emission. In coal-fired power plants, BF4- is formed in a flue gas desulfurization unit depending on a type of FGD unit and coal. Once it is formed, BF4- is fairly stable and slow to decompose, which is difficult to precipitate in ordinary coagulation sedimentation treatment. (1) BF4- + 3H2O → H3BO3 + 4F- + 3H+ Thus, in some power plants, dedicated treatment to decompose BF4- is employed to manage BF4- emission. The conventional measurements of BF4-, i.e., spectrophotometry and ion chromatography requires a tedious pretreatment or a stationary equipment, which are impossible to carry out onsite. In addition, since F- is involved in the formation and decomposition of BF4-, the onsite and simultaneous measurement of BF4- and F- would be helpful for the efficient management and treatment of BF4-, which would lead to a reduced amount of sludge in the BF4- treatment. This study aims to propose a simple quantification method of BF4- in flue gas desulfurization (FGD) wastewater, which is carried out onsite with a commercial ion-selective electrode. EXPERIMENTAL We built up a compact flow measurement system that uses BF4- and F- ion-selective electrodes (ISE) mounted in the dedicated flow cells (FLC-12), as shown in Figure 1. In the measurement of BF4-, the flow cell housed a BF4- ISE (EL7464L) and a reference electrode (4401L). The BF4- ISE is a liquid membrane electrode, which disperses tetrafluoroborate trioctylammonium as an ionophore, in the polyvinyl chloride (PVC) membrane. The ISE can measure the concentration of BF4- anion in the range of 0.1-1000 mg/L as boron, but in practice, 1-100 mg/L in the case of high matrix samples such as process effluents. In contrast, the F- ISE (F2021) is a solid membrane electrode, which can measure 0.02-20000 mg/L as fluoride. Buffer solution (pH7AB) is added to the sample in 10% volume and
the sample flowrate was controlled at 5 mL/min with a peristaltic pump. The potentiometric data of ISEs were collected on a multifunction ion meter (MM-60R) and stored on a linked PC. All the apparatus, equipments and chemicals described here were obtained form DKK-TOA Corporation. Actual FGD wastewater was obtained from full-scale power plants.
Figure 1 Schematic diagram of compact flow measurement system for BF4- and F- ions. RESULTS AND DISCUSSION Easy and rapid determination of BF4-. We confirmed that measurement in a flow mode on this system achieves improved stability and reproducibility in data collection, compared to the conventional static batch measurement. The system can measure BF4- and F- in the range of 1-100 mg/L and the measurement completes in less than 10 min. Figure 2 shows calibration curves of BF4- and Fin simultaneous measurement. BF4- ISE. Each ISE gave a linear calibration to its anion, respectively. However, F- ISE also showed a potentiometric response in BF4- standard solution, which indicates the decomposition of BF4- solution according to reaction (1). This means that F- is formed in small amount immediately after the BF4- standard solution is prepared. We should take care of the decomposition of BF4- standard solution, especially in the case of long-term storage.
Figure 2 Simultaneous measurement of BF4- and F- in (a) BF4- standard solution and (b) F- standard solution. Table 1 summarize the results for actual FGD wastewater samples. Since we were not able to obtain samples containing BF4- (sample A in Table 1), we added BF4- and/or F- to the FGD sample, which are spiked samples B-D in
Table 1. The measured values were obtained with high reproducibility on the proposed system (compact flow), which were consistent with the conventional measurement data (ICP-AES and IC). In the F- spiked samples (sample C and D), the recovery of F- is significantly lower, which is due to the reaction with Ca2+ in the FGD samples to precipitate CaF2. Thus, the compact flow measurement system can be successfully applied to actual FGD wastewaters. The proposed method is easier and more rapid determination compared with the conventional analytical methods. Table 1 Compact flow measurement of BF4- and F- in FGD wastewater. Concentration (mg/L) Conventional Method
Compact Flow
Sample Target
Total Boron
Method
BF4-
F-
BF4-
F-
ICP-AES
IC
IC
ISE
ISE
A
FGD wastewater
89.4
ND
5.4
0.2 (14.6)
6.0 (0.1)
B
Sample A + BF4- (50 m-Bg/L) addition
138
47.7
8.2
47.2 (0.2)
9.2 (0.5)
C
Sample A + F- (50 mg-F/L) addition
85.7
ND
26.0
0.2 (0.7)
27.8 (1.8)
D
Sample A + BF4- (50 mg-B/L) + F- (50 mg-F/L) addition
131.5
50.2
35.7
49.9 (1.7)
24.9 (0.3)
The values by compact flow method are an average of 2 measurements. Parentheses indicate relative standard deviation in %. IC: Ion Chromatography; ISE: Ion Selective Electrode; ND: not detected.
Preparation of stable BF4- standard solution. As describe before, the BF4- standard solution decompose to generate F- immediately after its preparation. Thus, we calculated the chemical equilibrium of the BF4- decomposition according to the procedure reported by Mizoguchi et al.,3 considering the following fluoroborate-related reactions, (2) BF4- + H2O → BF3(OH)- + F- + H+, (3) BF3(OH)- + H2O → BF2(OH)2- + F- + H+, (4) BF2(OH)2- + H2O → BF(OH)3- + F- + H+, (5) BF(OH)3 → H3BO3 + F , and fluoride-related reactions, (6) HF → H+ + F- , (7) 2HF → H2F2 , (8) HF + F- → HF2- . Figure 3 (a) shows the equilibrium fraction of BF4- standard solution. The BF4- fraction depends on the pH and the initial concentration of BF4-. BF4- is stable and predominant at higher initial concentrations of BF4- and lower pH values, whereas it decomposes at the lower initial concentrations and at higher pHs. Thermodynamically, to suppress the decomposition of BF4-, an excess amount of F- should be added to a sample solution and control it at lower pHs. When F- is added to a sample that amounts to 0.1 mol/L and control the sample pH at below 3, the decomposition of BF4- in the sample would be below 1%, as shown in Figure 3 (b).
Figure 3 Equilibrium fraction of BF4- at different concentrations. (a) BF4- alone; (b) BF4- with 100 mmol/L F-.
Based on the findings, we prepare the BF4- standard solutions, in which 0.1 mol/L NaF is added and its pH is controlled at less than 2 with H2SO4, i.e., the modified BF4- standard. Figure 4 compares the BF4- concentration during storage between the ordinary and the modified BF4- standards. As we expected, the BF4- concentration in the ordinary standard showed a monotonous decline and then reached a constant value in the end. The decomposition and time to reach a constant concetration depends on the initial BF4- concentration. In contrast, the modified BF4- standard showed a stable concentration of BF4- during 120 day storage. Since we obtain stable BF4- standard solution, we can eliminate the standard sample preparation step that used to be carried out whenever we employ the compact flow measurement of BF4-.
Figure 4 Time courses of BF4- concentration in BF4- standard solution. Open circle: BF4- standard; closed circle: BF4standard with NaF and H2SO4. CONCLUSIONS We propose a simple and rapid determination method of BF4- in order to manage BF4- emission in FGD effluents. The proposed system employs a flow mode measurement, which achieves improved stability and reproducibility compared to the conventional static batch measurement. The system can measure BF4- in the range of 1-100 mg/L and the measurement completes in less than 10 min. The proposed method is an easier and more rapid determination compared with the conventional analytical methods. We also develop the modified BF4- standard solution that can be subjected to long-term storage. REFERENCES (1) http://www.env.go.jp/en/water/wq/wp.pdf (2) http://www.env.go.jp/en/water/wq/nes.html (3) Katagiri, J.; Yoshioka, T.; Mizoguchi, T. Basic study on the determination of total boron by conversion to tetrafluoroborate ion (BF4-) followed by ion chromatography, Anal. Chim. Acta, 2006, 570, 65-72.
ENHANCING THERMAL EFFICIENCIES IN STEAM POWER PLANTS BY UTILIZING THE “W2” PRIME MOVER AS AUXILIARY EQUIPMENT
Jerry F. Willis Admiral Air, Inc.
1
Abstract: Coal fired power plants are the primary source of electrical power today. There are huge reserves of relatively inexpensive coal resources around the world. The coal fired electrical plant is here to stay. The thermal efficiency of the typical coal fired electric plant is understood to be approximately 35%. This thermal efficiency leaves a lot of room for improvement. Thermal efficiency today is improved by incorporating new boiler and steam turbine technology that produces only minor improvements in thermal efficiency. Increasing thermal efficiencies in power plants will reduce emissions, create potential for carbon credits, and extend coal reserves well into the future. This paper describes a method to substantially improve the thermal efficiencies in existing power plants by adding the new “W2” Prime Mover as auxiliary equipment to steam turbines.
2
System Turbine Characteristics
•
Many variations of steam powered turbines have been used to produce work
•
They are all driven by the flow and expansion characteristics of steam under pressure
Schematic outlining the difference between an impulse and a reaction steam turbine.
3
System Turbine Characteristics
Generally, the thermal efficiency of a coal fired electrical generating plant is 35%. Approximately 15% of the electrical power generated in a typical power plant will be consumed by operating equipment required to run the plant. Adding an auxiliary W2 driven generator designed to operate utilizing the static steam pressure at the end of the last stage of the steam turbine assumed at 300 psi will make up the 15% loss (31.5 MW). In this case, let’s look at Plant Mitchell in the United States. Plant Mitchell has boilers capable of producing 1.4 million pounds of steam an hour, burning 71 tons of coal an hour. Production is 210 MW of electricity per hour.
4
W2 Prime Mover Characteristics
Thermal efficiency of the W2 Prime Mover is in excess of 90%. A W2 Prime Mover requires only the static pressure of steam to function. Steam is consumed at the rate that will keep the system in the desired temperature range and replace steam lost due to small mechanical clearances in the W2. Adding a W2 powered electric generating unit will substantially increase the thermal efficiency of the power plant.
5
W2 Prime Mover Characteristics
The operational action of the W2 Prime Mover is very simple. A wobble plate is attached to the crank arm of a drive shaft with bearings. A wobble plate rotates freely on the crank arm. The drive shaft with the crank arm is attached to a mounting plate with bearings. The drive shaft rotates freely on the mounting plate. The edge of the wobble plate engages the mounting plate at a point notes as “Fulcrum” in the front view of the illustration in Figure 1. Applying a force in the downward direction as shown in “F1” will result in a lateral force at “F2” against the crane arm. The wobble plate will roll down toward the mounting plate at the “F1” location. This action forces the crank arm and drive shaft to rotate in the direction of the wobble plate rotation. The W2 Prime Mover is designed so that high pressure is continually applied to the “F1” side of the wobble plate, resulting in a continuous rotation of the drive shaft.
6
[Figure 1] Drawing illustrating the operational action of the “W2 Prime Mover”.
7
W2 Prime Mover
8
Side View of W2 Prime Mover 3D Prototype
9
Front View of W2 Prime Mover 3D Prototype
10
There has never been a better time to reduce energy requirements at power generating plants. “A typical company spends about 65% of its revenue or more on fuel. While in part the problem is leadership that is pre-occupied with other issues, an even bigger share is tied to the fact that innovation in the energy sector has been lacking.” Editor-In-Chief, Ken Silverstein, noted this premise in an article in Energy Biz Insider (2006). Mr. Silverstein emphasized further, “Consider that for every BTU of coal consumed to drive a steam turbine only 35 percent is converted into useful power. The other 65 percent is lost forever in the form of waste heat that is discharged into the environment”. The steam circulating system for the W2 is much smaller than that required for a conventional steam turbine of the same output.
Advantages of the W2 Prime Mover
less system heat loss
reduced fuel requirements
higher thermal efficiencies
lower carbon dioxide emissions
The “W2” Prime Mover will: •
Conserve revenue for electric utilities
•
Carbon footprint reductions
•
Generate carbon offsets
•
Replacing steam turbines with the “W2” Prime Mover will generate carbon offsets by directly and indirectly reducing the emissions of carbon to the atmosphere
11
COMPARING EFFICIENCIES OF THE STEAM TURBINE VERSUS THE “W2” PRIME MOVER
Jerry F. Willis Admiral Air, Inc.
1
Abstract: Coal fired power plants operate today using equipment which was designed during an era when fossil fuels were plentiful and relatively inexpensive. Greenhouse gas emissions and thermal efficiencies were of little concern when steam turbines were first developed. This paper compares the system characteristics of a typical coal fired power plant utilizing a steam turbine versus the new concept “W2” Prime Mover. The “W2” Prime Mover will be shown to increase the thermal efficiency of coal fired electric power plants from approximately 35% to values approaching 60%. This efficiency is achieved by utilizing static steam pressure to drive the “W2” Prime Mover whereas steam flow characteristics drive the steam turbine.
2
System Turbine Characteristics
•
Many variations of steam powered turbines have been used to produce work
•
They are all driven by the flow and expansion characteristics of steam under pressure
Schematic outlining the difference between an impulse and a reaction steam turbine.
3
System Turbine Characteristics
Generally, the thermal efficiency of a coal fired electrical generating plant is 35%. The thermal efficiency of a steam turbine can exceed 90%. What causes the thermal efficiency of a steam powered electric power plant to be 35% when the steam turbine is operating at 90%? The answer lies in the systems required to generate and deliver the massive volumes of steam required to run the steam turbines. In this case, let’s look at Plant Mitchell in the United States. Plant Mitchell has boilers capable of producing 1.4 million pounds of steam an hour, burning 71 tons of coal an hour. Production is 210 MW of electricity per hour.
4
W2 Prime Mover Characteristics
Thermal efficiency of the W2 Prime Mover is in excess of 90%. A W2 Prime Mover requires only the static pressure of steam to function. Steam is consumed at the rate that will keep the system at the desired temperature and replace steam lost due to small mechanical clearances in the W2. In this case, look at 210 MW production being driven by the W2 Prime Mover. What would be the estimated steam and coal requirements? .4 Million pounds of steam and twenty tons of coal would produce the 210 MW of power per hour if using the W2 Prime Mover.
5
W2 Prime Mover Characteristics
The operational action of the W2 Prime Mover is very simple. A wobble plate is attached to the crank arm of a drive shaft with bearings. A wobble plate rotates freely on the crank arm. The drive shaft with the crank arm is attached to a mounting plate with bearings. The drive shaft rotates freely on the mounting plate. The edge of the wobble plate engages the mounting plate at a point notes as “Fulcrum” in the front view of the illustration in Figure 1. Applying a force in the downward direction as shown in “F1” will result in a lateral force at “F2” against the crane arm. The wobble plate will roll down toward the mounting plate at the “F1” location. This action forces the crank arm and drive shaft to rotate in the direction of the wobble plate rotation. The W2 Prime Mover is designed so that high pressure is continually applied to the “F1” side of the wobble plate, resulting in a continuous rotation of the drive shaft.
6
[Figure 1] Drawing illustrating the operational action of the “W2 Prime Mover”.
7
W2 Prime Mover
8
Side View of W2 Prime Mover 3D Prototype
9
Front View of W2 Prime Mover 3D Prototype
10
There has never been a better time to reduce energy requirements at power generating plants. “A typical company spends about 65% of its revenue or more on fuel. While in part the problem is leadership that is pre-occupied with other issues, an even bigger share is tied to the fact that innovation in the energy sector has been lacking.” Editor-In-Chief, Ken Silverstein, noted this premise in an article in Energy Biz Insider (2006).
Mr. Silverstein emphasized further,
“Consider that for every BTU of coal consumed to drive a steam turbine only 35 percent is converted into useful power. The other 65 percent is lost forever in the form of waste heat that is discharged into the environment”. The steam circulating system for the W2 is much smaller than that required for a conventional steam turbine of the same output.
Advantages of the W2 Prime Mover •
less system heat loss
•
reduced fuel requirements
•
higher thermal efficiencies
•
lower carbon dioxide emissions
The “W2” Prime Mover will: •
Conserve revenue for electric utilities
•
Reduce carbon footprint
•
Generate carbon offsets
•
Replacing steam turbines with the “W2” Prime Mover will generate carbon offsets by directly and indirectly reducing the emissions of carbon to the atmosphere
11
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 OPTIMIZATION OF FUEL PROPERTIES WITH UTILIZATION OF BIODEGRADABLE MUNICIPAL WASTES FOR COMBUSTION UNITS JUCHELKOVA Dagmar, HAJKOVA Martina, RACLAVSKA Helena, TARARIK Luboš1 VSB – Technical University Ostrava, 17. listopadu 15, Ostrava, Czech Republic [email protected] 1 Frydecka skladka, a.s., Panské Nové Dvory 3559, Frydek-Mistek, Czech Republic
Introduction Alternative fuel is created by combination of two or more various material compounds and their optimized weight shares. A suitable choice and structure of the compounds enable attainment of needed qualitative parameters under complying with limit parameters of contaminants, required calorific value under parallel utilization of energy content of the all suitable shares. As the first material, brown coal of lower quality from Most Lignite Coal Basin was used (Fig.1). The second material is represented by solid municipal waste (MSW) from small settlements up to 2,000 citizens with accent on utilization of biodegradable waste (BMW). Effort in utilization of BMW as a fuel is conditioned by need of minimization of waste landfilling. New strategies in municipal solid waste (MSW) management, i.e., a separate collection of the organic fraction (EU Directive 1999/31/EC, EU Directive 2008/98/EC) and a reduction of the biodegradable MSW fraction allocated in landfills (EU Directive 2003/33/EC), have favoured their energetic utilization and composting. Fig.1 Lignite Coal Basins in the Czech republic With Libous Open Pit Mine
The production of municipal waste in the Czech Republic currently represents merely 306 kg per an inhabitant and a year. From this amount 83 % will be deposited in landfills, only 13 % of waste is incinerated, 2 % recycled and 2 % are used for production of compost (Report EEA). The aim of the investigation was verification of possibilities of fuel preparation while meeting conditions of the Czech legislative (specific sulphur content and calorific value). The fuel will be prepared from lignite of lower quality and from biodegradable waste and at the same time, the fuel will meet required mechanical properties expressed by PDI index (pellet durability index). Material for alternative fuel preparation Solid Municipal Waste (MSW). Composition of the solid municipal waste keeps changing in different localities in all states (Amlinger at al. 2010) and it is dependent on many factors as regional
differences, climatic conditions, a frequency of separate waste collection, a season of the year, cultural customs (Tchobanoglous et al., 1997), development trends in community, living standards of citizens, settlement's size, built-up area character, and a kind of heating. A share of biodegradable waste in the solid municipal waste is 57% in FRG, 62% in Italy and 52% in Hungary (Herczeg M. at al., 2009). Average structure of the MSW in EU is the following: paper and cardboard - 35%; kitchen and garden waste - 25%; plastics - 11%; glass - 6%; metals - 3%; textiles 2%; others - 18% (Calabrò 2009). Estimation of individual fractions in the municipal waste according to Brebu et al. 2010 might be considered as a very realistic one. It specifies 20 % share of inorganic fraction, up to 65 % share of ligno-celluloses materials (paper, cardboard, cartons, dietary residue of vegetal and animal origin) and about 20% share of plastics. Table 1: Abundance of individual compounds in municipal waste (%) Waste 1. Plastics 2. Glass 3. Textiles 4. Metals 5. Inert Material 6. Paper & Cardboard 7. Garden & Assorted BMW Ʃ BMW (3+6+7) Number of Samples
Frycovice Annual Average 15,3 ± 7,0 6,2 ± 5,0 6,5 ± 4,0 5,0 ± 0,9 21,0 ± 9,9 17,8 ± 7,5 28,3 ± 2,7 52,56 6
FM Municipalities (Jul. – Oct. 2009) March 2010
20,0 ± 9,5 11,5 ± 7,7 5,9 ± 7,3 4,8 ± 3,2 8,5 ± 8,4 17,9 ± 9,1 31,0 ± 13,6 54,80 11
14,4 ± 7,2 4,7 ± 3,9 3,1 ± 2,4 3,5 ± 2,3 27,6 ± 7,7 11,5 ± 6,2 35,7 ± 5,6 50,27 6
EU Calabro, 2009 11 6 2 3 18 35 25 62
Results of municipal waste assortment in the municipalities up to 2,000 citizens with village buildings are stated in Table 1. Composition of the municipal waste in Frycovice municipality was monitored within a year (2 months cycle of samples collection). The samples were collected at the municipality waste landfill. Abundance of the individual fractions in the municipal waste from FM municipalities was calculated after separation of waste from containers of 110 liters capacity which were transported to Frýdek Místek AG Land field. Comparing results (Calabro et al. 2009), the waste from village built-up area has considerably lower content of paper fraction and considerably higher content (up to by 10 %) of inert material. The municipal waste collected in March 2010 had plastics depleted by materials with high calorific power (PET, PVC) due to the „winter season“. Especially flexible foils (HDPE, LDPE and PP) stood in for plastics. Also quite high amount of children napkins was a part of the MSW. Current disposable nappies consist of a pervious foil (made of polypropylene mainly), a moisture absorbing napkin (cellulose filled up by gel absorber on polyacrylate base by reason of improvement of the absorptivity), an impervious foil for cloth protection (made of polyethylene), adhesive strips or burfasteners and elastic bands. For purposes of BMW share calculation, the children napkins were classified from 50 % as biodegradable waste and from 50% as plastics. For the fuel production, the utilization of separated BMW from the solid municipal waste was supposed. From the proximate analysis (Table 2), it is obvious that separated part of BMW contains still considerable share of inorganic fractions which will decrease BMW's calorific value. For achievement of the specific sulphur content limit in the proposed fuel, in the case of brown coal of lower quality utilization, it is necessary to ensure in order that the second fraction was of higher calorific value. However, BMW does not meet this requirement. For achievement of the calorific power improvement, also separated flexible plastics were again added to the BMW sample (Table 3)
Table 2: Ash content and calorific value separated BMW in dry matter FM Municipalities Unit Jul. – Oct. 2009 March 2010 o Ash (815 C) % 15.52 ± 12.35 15.56 ± 7.88 Net Calorific Value MJ/kg d.m. 16.23 ± 5.24 13.68 ± 5.57 Gross Calorific Value MJ/kg d.m. 13.35 ± 4.28 11.75 ± 6.75
Frycovice Annual Average 48.44 ± 28.69 11.58 ± 6.54 10.87 ± 7.69
Fig.2 Percental abundance of fraction in municipal waste after elimination of inert material (metals and glass). Brown Coal. Coal sample was removed from Libouš Coal Pit. Libouš Coal Pit is the largest coal pit in Most Lignite Basin (the Czech Republic). The Most Lignite Basin forms a relic of tertiary sedimentary deposit. Libouš Coal Pit is controlled by the North Bohemian Lignite Mines, AG, Chomutov. The brown coal bed is mined at Tušimice-Libouš coal deposit with thickness from 25 to 35m and average ash content of 36.8 % in anhydrous state and with the sulphur content of 2.7 % and calorific power of 10.40 MJ/kg in the original state. Total exploitable coal deposit is about 286 million tons. The exploited coal is utilized for production of energy fuel mixtures. Table 3: Ultimate and proximate analysis of waste and fuels MW a
W Ad Std St Cd Hd Nd Qs d Qi r
% % % % % % % MJ/kg MJ/kg
As
mg/kg d.m.
Cd
mg/kg d.m.
Cr
mg/kg d.m.
Hg Ni
mg/kg d.m. mg/kg
3,38 14,30 0,14 0,13 47,73 6,95 1,00 25,38 23,86 0,461 ± 0,067 0,347 ± 0,046 19,7 ± 2,2 0,006 ± 0,001 3,88 ±
BMW 2,68 6,30 0,08 2,60 36,24 4,54 1,24 13,68 8,56 0,116 ± 0,021 0,116 ± 0,018 12,6 ± 1,21 0,003 ± 0,0015 2,14 ±
Coal 23,54 41,93 2,62 2,02 33,76 3,75 0,84 15,80 9,13 33,8
MW80+Coal20
MW60+Coal40
MW40+Coal60
19,826 0,636 0,51 44,936 6,31 0,968 23,46 20,91 7,12
25,352 1,132 0,886 42,142 5,67 0,936 21,54 17,96 13,79
30,878 1,628 1,264 39,348 5,03 0,904 19,632 15,02 20,46
2,25
0,72
1,10
1,48
126
40,96
62,22
83,48
2,08
0,42
0,83
46,8
12,46
21,04
1,25 29,63
d.m.
0,71 0,44 35,2 ± 24,8 ± 1040 236,16 638,08 mg/kg Pb 7,0 4,6 437,12 d.m. 2,15 ± 1,88 ± 65,6 14,84 27,53 mg/kg V 0,27 0,12 40,22 d.m. 1596 ± 564 ± 66 3150 1906,8 2217,6 mg/kg Chlorine 160 2528,4 d.m. Explanations: Ad – Ash in anhydrous state, Qsd – Calorific value in anhydrous state, Qir – Calorific value in original state, Std – Sulphur in anhydrous state
Fuel for middle-sized energy sources According to Direction No.13/2009, Definition of Requirements on Fuels Quality for Stationary Sources from Air Conservation Point of View, fuel with specific sulphur content < 0.95 g.kJ-1 (in original state) and calorific value higher than 10 MJ.kg-1 might be burnt off in big and small sources. The direction does not specify which fractions or their parts can be used for the fuel preparation. Meeting the condition of specific sulphur content will be ensured by preparation of the fuel with minimal content of 40 % MSW and 60 % coal. The original aim was creation of the fuel which would utilize only a separated fraction BMW from MSW for improvement of CO2 emission limits. Considering that separated fraction BMW from MSW in municipalities up to 2,000 citizens has too low calorific value of 4 – 8 MJ/kg (in supplied state) and coal from Libouš Coal Field has a high content of sulphur, the BMW fraction with additive of flexible plastics would have to be used from MSW for meeting the specific sulphur content condition. Ballast fractions were separated (glass, metals, inert material). The condition of specific sulphur content is satisfied also in the case of the fuel prepared from 60% of coal and 40% of MW. For production of the mixed fuel, the coal with size of 0 – 40 mm was grained to size lesser than 5 mm. This size appears to be optimal for usage of JGE 150 laboratory pelletizing press (Pest Control Corporation Company, Vlcnov, the Czech Republic). MSW was dried and further grained in 9FQ 50 hammer grinding mill with power of 800 kg/h (Pest Control Corporation Company, Vlcnov, the Czech Republic) to size by 35 % > 1 mm. For measurement of pellets mechanical durability (PDI index – pellet durability index) the Holmen tester was used. The PDI is calculated as the ratio of weight after tumbling and the weight before tumbling, multiplied by 100. A sample is circulated in the Holmen tester for 60 s. The remaining pellets are collected, sieved with sieve size of about 80 % of pellet diameter, weighted, and the PDI is calculated. In the case of humidity higher than 10 %, the PDI value might be influenced by the humidity (Arshadi et al.2008). The PDI value of all samples was measured at the same humidity (4 – 5%).
Table 4: Pellet durability index (PDI - %) MW BMW Coal MW80+Coal20 88 96 78 92
MW60+Coal40 86
MW40+Coal60 82
It is obvious from the table that PDI index value keeps decreasing with growing of coal content in pellets. Biodegradable waste has high PDI index, however, it might not create function of binding agent for the coal fraction. For solution of this problem CaO, Ca(OH)2, CaCO3 or another material with high content of lignine (compost) can be used as a plasticizing agent. At the same time, it supports self-desulphuring effect of the proposed fuel.
Conclusions Separated municipal waste (BMW and flexible plastics) might form an important energy fraction in preparation of the fuel for middle-sized energy sources that utilize brown coal of lower quality with a specific high content of the sulphur. The specific sulphur content parameter that is required by Direction No. 13/2009 is met in the fuel mixture consisting of 40% of separated MW and 60% of coal. However, the prepared fuel in such way does not comply with requirements on mechanical durability, a share of the biodegradable material is insufficient in order that so called "solid bridges", which bind quite heterogeneous material, were created among the individual grains. Successful utilization of the fuel on the base of municipal waste base and coal might be recommended not earlier than after suitable binding material was found.
Acknowledgement This work was supported by research projects of Ministry of Education, Youth and Sport of the Czech Republic: INTERVIRON No. 2B06068. References Arshadi M., Gref R., Geladi P., Dahlqvist S.A., Lestander T. (2008): The Influence of Raw Material Characteristics on the Industrial Pelletizing Process and Pellet Quality. Fuel Processing Technology, V.89, 1442-1447 Brebu M., Ucar S., Vasile C., Yanik J., 2010: Co-pyrolysis of Pine Cone with Synthetic Polymers, Fuel doi: 10.1016/j.fuel.2010.01.029 Diverting Waste from Landfill, Effectiveness of Waste-Management Policies in the European Union, EEA Report, Copenhagen, Denmark, 2009
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COMBUSTION SULFUR RETENTION IN THE ASH DURING COMBUSTION OF TUNCBILEK BRIQUETTES
Ayfer Parlak, Turkish Coal Enterprises (TKI), GLI, Tuncbilek, Kutahya, Turkey [email protected] Bekir Zühtü Uysal, Department of Chem. Eng., Faculty of Engineering and Clean Energy Research and Application Center, Gazi University, Ankara, Turkey Mustafa Ozdingis, Turkish Coal Enterprises (TKI), Ankara, Turkey H. Köksal Mucuk, Turkish Coal Enterprises (TKI), Ankara, Turkey Selahaddin Anac, Turkish Coal Enterprises (TKI), Ankara, Turkey
Abstract: In this study, it was aimed to briquette the mixture of Tuncbilek lignite in the size range of 0.5-10 mm and Tuncbilek Coal Washing Plant’s slurry waste involving 0.1-0.5 mm particles with molasses and dolomite bindings and to achieve retention of sulfur in the ash during combustion. In this way coal fines with high sulfur content was converted by briquetting into a strong and uniform fuel giving rise to low sulfur emission upon combustion. In addition, Coal Washing Plant’s slurry waste which is currently not being utilized has been converted into a usable form in industry by the addition to coal fines.
Different amounts of dolomite and waste were used in briquetting and their effects on the efficiency of retention of sulfur in the ash were examined. A pilot equipment with 100 kg capacity was used for briquetting. Briquettes were burned in a TGA analyzer and in a commercial stove, and sulfur analyses were made in the ash. A mixture of 67% lignite, 20% slurry, 7% molasses and 6% dolomite gave the best result. Sulfur retention in the ash was found as 74.6% and 60.2% by TGA and stove tests, respectively.
Types of sulfur in the original Tuncbilek lignite were also determined experimentally. Since original sulfate in the coal is not affected in combustion, the retention efficiency of sulfur in the ash was also evaluated on the basis of combustible sulfur and was found to be 90.8 % and 73.2 % by TGA and stove tests, respectively. Thus, briquetting provided a decrease in sulfur dioxide emission not less than 70%. Proximate analyses and heating values of the briquettes were also made and the results indicated that the briquettes obtained were marketable.
Development of an Analytical Solution for Jet Diffusion Flame Equations ´ Felipe Norte Pereira∗ , Alvaro Luiz de Bortoli∗ , Nilson Romeu Marcilio∗ ∗
Federal University of Rio Grande do Sul (UFRGS), Department of Chemical Engineering. Porto Alegre, Brazil. E-mail: [email protected], [email protected], [email protected]
Abstract—Exact solutions of nonlinear equations help us to understand the mechanism of nonlinear effects. Any exact solution for flame propagation must make use of basic equations of fluid dynamics modified to account for the liberation and conduction of heat and for charges of chemical species within the reaction zone. The present work develops an analytical solution for a jet diffusion flame. The model is based on the solution of the mixture fraction for turbulent fluid flow with no constant eddy viscosity and the results are found to compare favorably with data in the literature.
I. I NTRODUCTION Combustion theory is an important area of classical phenomenology, which treats a wide range of natural phenomena. In general, combustion is fast compared to molecular mixing, occurring in layers thinner than the typical length scales of turbulence, especially for high Damk¨ohler values. Flamelets are thin reactive diffusive layers embedded within an otherwise nonreacting flow field. The basic idea is to assume that a small instantaneous flame element embedded in a turbulent field has a structure of a laminar flame. The flamelet concept can be employed in most practical situations, when the chemistry is fast. To solve nonpremixed jet flames many approximate models are found in literature: Williams (1985), Li˜na´ n (1991), Peters and Donnerhack (1981), Peters (2000), Warnatz et al. (2001), Veynante and Vervish (2003), Fern´andez-Tarrazo et al. (2006). The analytical developments present in this paper were motivated by the following works: — Peters and Donnerhack (1981), which calculates the mixture fraction of an axisymmetric flame, considering the turbulent viscosity constant throughout the entire jet, using a similarity transformation. — Agrawal and Prasad (2003), which developed analytical expressions for the eddy viscosity for turbulent jets, plumes and wakes, using a semi-empirical equation for the axial velocity. The current work aims to determine an expression for the axial velocity and mixture fraction of a confined jet diffusion flame, using the similarity transformation applied by Peters and Donnerhack (1981), but considering the eddy viscosity, obtained by Agrawal and Prasad (2003), which varies along the length and radius of the jet. The expression obtained for the mixture fraction is then compared with experimental results for a hydrogen flame.
Fig. 1.
Schematic representation of a vertical jet flame
II. T URBULENT J ET D IFFUSION F LAMES INTO S TILL A IR It was considered a diffusion flame in which the fuel, delivered from a round nozzle, with diameter d and exit velocity u0 , mixes with the surrounding air by convection and diffusion. The jet flame is chosen because it seems to be a representative of the class of nonpremixed flames. The effects of buoyancy and pressure gradients in the flame have been neglected. The Fig. 1 shows the structure of a diffusion flame in cylindrical coordinates. This lead to a two-dimensional axisymmetric problem governed by equations of Continuity (Eq. (1)), Momentum in x-direction (Eq. (2)) and Mixture Fraction (Eq. (3)) in steady state. ρv˜r) ∂(¯ ρu ˜r) ∂(¯ + ∂x ∂r ∂u ˜ ∂u ˜ ρ¯u ˜r + ρ¯v˜r ∂x ∂r
=
0
(1)
=
∂ ∂u ˜ ρ¯ν˜t r ∂r ∂r
(2)
∂ Z˜ ∂ Z˜ ρ¯u ˜r + ρ¯v˜r ∂x ∂r
∂ ∂r
=
ρ¯ν˜t r ∂ u ˜ Sc ∂r
(3)
The boundary layer assumption was considered in Eq. (2) and Eq. (3). Also, in Eq. (2), the viscous stress was neglected compared to the Reynolds stress component, which was modeled as show in Eq. (4). ∂u ˜ (4) ∂r The equations (1), (2) and (3) are subjected to the following boundary conditions (Eq. (5)).
The variable Uc corresponds to the centerline velocity, which for an axisymmetric jet varies as x−1 . The variable c corresponds to the spread rate of the jet. To simplify the calculation, it was considered that the mixture fraction is a product of two functions: the centerline mixture fraction Z˜cl and the variable ω, as shown in the Eq. (12).
00 00 −¯ ρug v = ρ¯ν˜t r
Z˜ = Zcl (5) ∂u ˜ r → ∞ u ˜ = 0, = 0, Z˜ = 0 ∂r The Continuity Equation (1) is automatically satisfied by introducing the stream function (ψ) (Eq. (6)). r
=
0
r¯2 = 2
Zr
ρ¯ rdr, ρ∞
ζ = x + x0
C=
C=
∂ ∂η ∂ r¯ ∂ = ∂r ∂ r¯ ∂r ∂η
(8)
Introducing the adimensional stream function F (Eq. (9)), the axial and radial velocity may be written as a function of F (η) (Eq. (10)), as follows: F =
ψ ρ∞ νtr ζ
(9)
1 ∂ (νtr F ) ∂ (νtr F ) u ˜= , ρ¯v˜r = −ρ∞ νtr F − η ηζ ∂η ∂η (10) The variable νtr present in Eqs. (9) and (10) corresponds to the eddy viscosity. To quantify the variation of the eddy viscosity, it was employed a semi-empirical expression obtained by Agrawal and Prasad (2003) for an axisymmetric jet diffusion flame (Eq. (11)). νtr
Uc c4 = x 4
1 − exp(−ξ 2 ) ξ2
,
ξ=
r c·x
(11)
(13)
(ρ0 ρst )1/2 ρ∞
(14)
Substituing the equations (10) and (12) into equations (2) and (3), and applying the chain rule described in equation (8), results the equations (15) and (16) ∂ (νtr F ) ∂(νtr F ) = ∂η η ∂η ∂ ∂ 1 ∂(νtr F ) = Cνtr η ∂η ∂η η ∂η ∂ C ∂ω ∂ ηνtr − ((νtr F )ω) = ∂η ∂η Sc ∂η −
0
∂ ∂ ∂η ∂ = + , ∂x ∂ζ ∂ζ ∂η
ρ¯2 νt r2 ρ¯2∞ νtr r¯2
According to Peters and Donnerhack (1981), the ChapmanRubesin parameter may be written as:
(7)
The Eq. (7) contains a density transformation leading to the density weighted radial coordinate r¯ . The new axial coordinate ζ corresponds to the sum between the axial coordinate x and the flame dislocation from the nozzle x0 . With the similarity transformation, the coordinate system (x, r) is transformed to (ζ, η) using the chain rule show in Eq. (8).
(12)
One introduces the Chapman-Rubesin parameter, C, in equation (13), and assumes it as a constant in the entire jet.
v˜ = 0,
∂ψ ∂ψ , ρ¯v˜r = − (6) ρ¯u ˜r = ∂r ∂x A similarity transformation is employed, following the work of Peters and Donnerhack (1981), in order to solve the equations analytically (Eq. (7)). r¯ η= , ζ
Z˜ = Z˜cl · ω
(15)
(16)
subject to the following bondary conditions (Eq. (17)): η r
= →
0: ∞:
F = 0,
ω=1 0
(νtr F ) = 0,
(17) 00
(νtr F ) = 0,
ω=0
To solve the two differential equations (Eq. (15) and (16)) analytically, the exponential term of equation (11) was expanded in a Taylor Series, truncated in the second term. III. C ALCULATION OF THE A XIAL V ELOCITY AND M IXTURE F RACTION Expanding the exponential term of equation (11) in a Taylor series truncated in the second term, results: exp(−ξ 2 ) ≈ 1 − ξ 2
(18)
With this consideration, the eddy viscosity can be written as shown in Eq. (19). Uc c4 x (19) 4 Now the eddy viscosity becomes independent of ζ and η, since Uc ∝ x−1 . Thereby, the equations (15) and (16) can be simplified, resulting in: νtr =
∂ − ∂η
∂ ∂ 1 ∂F = Cη ∂η ∂η η ∂η ∂ ∂ C ∂ω − (F ω) = η ∂η ∂η Sc ∂η F ∂F η ∂η
(20)
(21)
Solving the equation (20), one obtains: F =
C(γη)2 1+
(γη)2 4
Z 0
∞
−2 (γη)2 1+ 4
∞ ∂u ˜ ∞ ρ¯u ˜2 rdr + [¯ ρu ˜v˜r]0 = ρ¯νtr r ∂r 0
(23)
(24)
where ρ0 is the density of the fuel stream and R is the radius of the jet noozle. The substitution of equation (23) into equation (25), results in an expression to the constant γ (Eq. (26)). 3 64
ρ0 ρ∞ C 2
u ˜0 d νtr
2 (26)
In order to quantify the variation of the Schmidt Number (Sc), it was considered that Sc varies linearly with the mixture fraction, as shown in Eq. (27) Sc = (ScF − Scox )Z + Scox
(28) A2 = Scox
Substituiting the equations (22) and (28) into equation (21), it is possible to integrate the equation (21) to determinate ω (Eq. (29)). "
(γη)2 (A1 + A2 ) 1 + 4
#−1
2A2 − A1
(29)
By combining the mixture fraction equation (Eq. (3)) with the continuity equation (Eq. (1)) and integrating from r = 0 to r → ∞, in a similar way as in Eq. (24), one finds that the integrated mass flow rate must be independent of x (Eq. (30)). Z ∞ 2 ˜ = ρ0 u0 R ρ¯u ˜Zdy (30) 2 0 The combination of equations (12), (29), (23) in (30) results in the mixture fraction along the jet centerline Zcl (Eq. (31)) as Z˜cl θ
=
1 32A2 Z ∞
=
u0 d νtr 2
(ζ/d)−1 θ 2A2
(µ ((A1 + A2 )µ
(31) −1
− A1 ))
−1 dµ
1
With the bondary conditions given by Eq. (5), the last two terms in Eq. (24) are zero. It follows that the first integral denoting the jet momentum is independent of x and therefore constant, equal to the momentum in x = 0. Assuming a constant exit velocity across the orifice with the radius R, the jet momentum can be written as in Eq. (25) Z ∞ ρ0 u0 2 R 2 ρ¯u ˜2 dy = (25) 2 0
γ2 =
A1
= (ScF − Scox )Z˜cl
ω = A2
The Eq. (23) shows that the axial velocity in the centerline varies as x−1 , which agrees with the previous consideration made for Uc in the equation (11). Combining the momentum equation with the continuity equation and integrating from r = 0 to r → ∞ (Eq. (24)), one obtains: ∂ ∂x
= A1 ω + A2
(22)
where the variable γ is a constant (concerning to η) obtained in the integration of Eq. (20). The substitution of Eq. (22) into Eq. (10) results in an expression for u ¯ in terms of (ζ, η), Eq. (23), as: 2νtr Cγ 2 u ˜= ζ
Sc
(27)
where ScF and Scox correspond to the Schmidt Number for the fuel and the oxidizer, respectively. To express the variation of the Schmidt Number as dependent of (Z˜cl , ω) (Eq. (28)), the equation (12) was inserted into equation (27).
where µ is the variable of integration in the equation for θ. It was considered that A1 A1 + A2 in order to simplify the equation (31). The difference between the Schmidt Number of fuel and oxidizer is small, and when this value is multiplied by Zcl , it became even smaller, which justifies the simplification. Thus the equation (31) can be written as (32). Z˜cl
u0 d 1 ρ0 (ζ/d)−1 (32) 32 ρ∞ C νtr −1 −1 (2Scox + 1) − (ScF − ScOx ) Scox
= Scox
The Mixture Fraction can now be obtained by multiplying the equations (29) and (32). IV. C OMPARISON WITH E XPERIMENTAL DATA The analytical results for the mixture fraction was compared with experimental results from Barlow (2010) for a turbulent 50/50% hydrogen/nitrogen jet flame (where ρ0 /ρst = 0.24, ScF = 0.45 and Scox = 0.90). The figure 2 shows the comparison of the analytical and the experimental mixture fractions along the flame centerline. The mixture fraction result corresponds to the axial decreasing behavior of a commom free jet. The figures 3, 4 and 5 show the comparison of the analytical and the experimental mixture fractions for x/d equal to 20, 40 and 60 respectively.
The analytical solution for the mixture fraction along the jet centerline is valid for x/d > 10 as seen in figure 2. The expression for the mixture fraction varies with x in a way similar to x−1 . Thereby the solution tends to infinity for small values of x/d, which makes the solution unable to represent the jet entrance. For the mixture fraction along the jet radius (Fig. 3, 4 and 5), the analytical solution has a behavior similar to the experiment. Overall, the results show a greater discrepancy at the beginning of the flame, which tend to decrease as the radius value increases. Fig. 2. Comparison between the analytical and experimental mixture fraction along the jet centerline
V. C ONCLUSION Overall, the results were satisfactory. The analytical expression obtained for the mixture fraction seems to represent well a hydrogen jet, except at the beggining of it (small values of x). In terms of the radial variation, the solution tends to have lower discrepancies as the radius value increases. The radial profile of the mixture fraction may be reasonably approximated using a Gaussian function (Agrawal and Prasad, 2003, De Bortoli, 2009). Therefore, the next step of this study is to improve the eddy viscosity formula to decrease the jet spreading along its streamwise direction. VI. N OMENCLATURE
Fig. 3. Comparison between the analytical and experimental mixture fraction at x = 20d
C
=
c =
Parameter in the eddy viscosity equation
d =
Nozzle diameter
F
=
Adimensional stream function
r
=
Coordinate along the jet radius
r¯ =
Fig. 4. Comparison between the analytical and experimental mixture fraction at x = 40d
Density weighted radial coordinate
ScF
=
Schmidt number of the combustible
Scox
=
Schmidt number of the oxidant
u0
=
Exit velocity from the nozzle
u ˜ =
Favre averaged axial velocity
v˜ =
Favre averaged radial velocity
00
u ,v
00
=
x = Z˜ = Z˜cl =
Fluctuating components of velocity Coordinate along the axis Favre averaged mixture fraction
ρ0
=
Favre averaged mixture fraction along the jet centerline Fuel density
ρ∞
=
Oxidant density
ρ¯ =
Fig. 5. Comparison between the analytical and experimental mixture fraction at x = 60d
Chapman-Rubesin parameter
Reynolds averaged density
νt
=
Viscosity
νtr
=
Eddy viscosity
ψ
=
Stream function
ζ, η
=
Coordinates used in the similarity transformation
ω
=
Normalized mixture fraction
ACKNOWLEDGMENT The authors gratefully acknowledge the finantial support from CNPq (Conselho Nacional de Desenvolvimento Cientifico e Tecnol´ogico) under process 303007/2009-5. R EFERENCES A. Agrawal and A. Prasad. Integral solution for the mean flow profiles of turbulent jets, plumes, and wakes. Journal of Fluids Engineering, 145, September 2003. R. Barlow. H2 /N2 jet flames, Consulted on March 2010 at www.ca.sandia.gov/TNF. A. L. De Bortoli. Analytical/numerical solutions for confined jet diffusion flame (Sandia Flame C). Latin American Applied Research, (34):157–163, 2009. E. Fern´andez-Tarrazo, A. L. S´anchez, A. Li˜na´ n, and F. A. Williams. A simple one-step chemistry model for partially premixed hydrocarbon combustion. Combustion and Flame 147, pages 32–38, 2006. A. Li˜na´ n. The structure of diffusion flames in fluid dynamical aspects of combustion theory. Longman Scientific and Technical, UK, 1991. N. Peters. Turbulent Combustion. Cambridge University, Press, 2000. N. Peters and S. Donnerhack. Structure and Similarity of Nitric Oxide Production in Turbulent Diffusion Flames. 18th Symp. (Int.) on Combustion, The Combustion Institute, 1981. D. Veynante and L. Vervish. Turbulent combustion modeling. Lecture Series du Von K´arman Institute, 2003. J. Warnatz, U. Maas, and R. W. Dibble. Combustion: Physical and Chemical Fundamentals, Modeling and Simulation, Experiments, Pollutant Formation. Springer-Verlog, 3 edition, 2001. F. A. Williams. Combustion Theory. Addison-Weskey, Redwood City, CA, 2 edition, 1985.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COMBUSTION (FLUE GAS CLEAN UP) EXACT TITLE OF PAPER: IONIC LIQUIDS WITH AMINE FUNCTIONAL GROUP: A SHORCUT TO IMPROVE THE PERFORMANCE OF IONIC LIQUIDS FOR CO2 SCRUBBING Presenting Author: Jelliarko Palgunadi, Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1 Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected]; Telp: +82-2-961-0431; Fax: +82-2-966-3701 Co-Authors: Jin Kyu Im; Antonius Indarto, Dr; Hoon Sik Kim, Prof.; and Minserk Cheong, Prof., Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1-Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected] (H.S. Kim); [email protected] (M. Cheong); Telp: +82-2-961-0432; Fax: +82-2-966-3701
Abstract: One of the global environmental problems of today is the increase of the greenhouse gases concentrations in the atmosphere. This problem partly corresponds to the increase of carbon dioxide (CO2) emission from the burning of fossil fuels for power generations. To response this challenge, carbon capture and storage (CCS) using liquid scrubber receives great attentions because there is potential for retrofit to existing power plants without changing the existing process. Within this framework, ionic liquids (ILs) have been proposed as alternative media for scrubbing CO2 from post-combustion emission where SOx, NOx, and tiny particulates are also inevitably co-produced. Due to the ionic nature of these low-melting point salts, the problem associated with the solvent lost during cycled absorption-desorption processes might be minimized. Experimental results from our group combined with numerous published data demonstrated that the CO2 solubility in many conventional ILs at the pressure close to atmosphere is very low. Thus, these conventional ILs are not feasible compared to the molecular amine-based scrubbers, i.e. monoethanol amine (MEA) for capturing CO2 contained in a post-combustion stream. Regular solution theory and quantum chemical calculations demonstrate that the CO2 solubility in common ILs is merely controlled by weakly physical interactions (i.e. van der Waals interactions). To improve the performance of ILs, various task-specific ILs dissolved in a simple room temperature IL or in a non-volatile organic solvent were evaluated for CO2 capture at low pressures. Cost-effective ILs containing an amine moiety were prepared by quaternarization of commercially available diamines. Similar CO2 loading capacity as found in a molecular amine system was observed likely through the formation of a carbamate salt-like structure. The CO2 solubility measurements, the computational calculations of the CO2-IL complexes, and some factors associated with the optimum absorption conditions are discussed.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COMBUSTION (FLUE GAS CLEAN UP) EXACT TITLE OF PAPER: ABSORPTION OF SULFUR DIOXIDE IN TASK SPECIFIC IONIC LIQUIDS CONTAINING SO2-PHILIC GROUPS ON THE CATION Presenting Author: Sung Yun Hong, Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1 Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected]; Telp: +82-2-961-0431; Fax: +82-2-966-3701 Co-Authors: Jelliarko Palgunadi; Hoon Sik Kim, Prof.; and Minserk Cheong, Prof., Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1-Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected] (H.S. Kim); [email protected] (M. Cheong); Telp: +82-2-961-0432; Fax: +82-2-966-3701
Abstract: Fossil fuel burning power plant is one of major producers of sulfur dioxide emissions worldwide. To mitigate the emissions of SO 2, scrubbing process employing liquid absorbents is considered as an alternative method in addition to the flue gas desulphurization (FGD). Room temperature ionic liquids (RTILs) have been demonstrated to absorb SO2 effectively. Tunable psychochemical properties derived from the combinations of tailored ionic components and non-volatility of RTILs resulted from the coulombic force stabilization are the key features making these materials more attractive than volatile organics for SO2 capture. Some literatures suggest that the formation of specific interactions such as Lewis acid-base interactions control the SO2 solubility. Thus, the presence of SO2-philic groups on the molecular structure of RTILs is required to achieve high SO2 solubility. In our group, imidazolium-based cations containing various pendant groups with capabilities to form specific yet reversible interactions with SO2 combined with [MeSO3]- as the anion were synthesized. The solubility measurements for SO2 are presented and the effects of the cation structure on the absorption-desorption processes are discussed.
2010 International Pittsburgh Coal Conference
REACTION CHARACTERISTICS OF NEW OXYGEN CARRIERS FOR CHEMICAL LOOPING COMBUSTION Ho-Jung Ryu1,*, Jaehyeon Park1, Gyoung-Tae Jin1, Moon-Hee Park2 1
Korea Institute of Energy Research, 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA 2
Hoseo University, 185, Saechoolee, Baebangmyun, Asan, 336-795 Korea
*[email protected] Abstract In this paper, natural gas combustion characteristics of new oxygen carrier particles were investigated in a batch type fluidized bed reactor (0.052 m ID, 0.7 m high). Three particles, OCN703-1100, OCN7051100, and OCN708-1300 were used as oxygen carriers. Natural gas and air were used as reactants for reduction and oxidation, respectively. To check feasibility of good performance, inherent CO2 separation, and low-NOx emissions, CH4, CO, CO2, O2, H2, NO concentrations were measured by on-line gas analyzer. Moreover, the regeneration ability of the oxygen carrier particles was investigated by successive reduction–oxidation cyclic tests up to the 10th cycle. All three oxygen carrier particles showed high gas conversion, high CO2 selectivity, and low CO concentration during reduction and very low NO emission during oxidation. Moreover, all three particles showed good regeneration ability during successive reduction-oxidation cyclic tests up to the 10th cycle. These results indicate that inherent CO2 separation, NOx-free combustion, and long-term operation without reactivity decay of oxygen carrier particles are possible in the natural gas fueled chemical-looping combustion system. However, OCN708-1300 represented temperature and pressure fluctuations during reduction and slightly decay of oxidation reactivity with the number of cycles increased. Introduction Carbon dioxide, a major greenhouse gas, is produced in large quantities from combustion of fossil fuels, much of this related to electric power generation. In a conventional power generation system, fuel and air are directly mixed and burned; therefore it is not easy to separate CO2 from flue gas because CO2 is diluted by N2 in air. Chemical-looping combustion (CLC) is a novel combustion technology with inherent separation of the greenhouse gas CO2 and no NOx emission. The chemical-looping combustor consists of two reactors, an oxidation reactor and a reduction reactor. A fuel and an air go through the different reactors. Equations (1) and (2) explain a basic concept of the chemical-looping combustion system. In the reduction reactor, gaseous fuel (CH4, H2, CO or CnH2n+2) reacts with metal oxide according to the Eq. (2), and release water vapor and carbon dioxide from the top and metal particles (M) from the bottom. The solid products, metal particles, are transported to the oxidation reactor and react with oxygen in the air according to Eq. (1), and produce high-temperature flue gas and metal oxide particles. Metal oxide particles at high temperature are again introduced to the reduction reactor and supply the heat required for the reduction reaction. Between the two reactors, metal (or metal oxide) particles play an important role in transportation of oxygen and heat, therefore the looping material between the two reactors is named as an oxygen carrier particle. Oxidation: exothermic reaction, M + 0.5 O2 = MO
(1)
Reduction: endothermic reaction, CH4 + 4MO = CO2 + 2H2O + 4M
(2)
It is important for the exhaust gas from the reduction reactor to contain only highly concentrated CO2 and water vapor. Therefore, CO2 can be easily recovered by cooling the exhaust gas without any extra energy consumption (energy penalty) for CO2 separation. Another advantage of CLC is that NOx formation can be thoroughly eliminated because the oxidation reaction occurs at a considerably lower temperature (~900oC) without a flame; therefore there is no thermal NOx formation [1, 2]. Moreover, the efficiency of chemical-looping combustion system is very high. Wolf et al. [3] reported that natural gas fueled chemical-looping combustor achieves a thermal efficiency between 52–53% and is 5 percent points more efficient than an NGCC system with state-of-the-art technologies for CO2 capture. The previous results on chemical-looping combustion technology have concentrated on improvement of the oxygen carrier particles and most of the studies used methane or hydrogen as a reduction gas [4]. In order to understand the performance and the feasibility of the chemical-looping combustion system, it is necessary to check the performance of the oxygen carrier particles by using natural gas as a reduction gas because the NGCC, as a practical application of the CLC system, uses natural gas as a fuel. Moreover, confirmation of reduction of NOx emission in the oxidizer and high CO2 selectivity in the reducer are prerequisite for application of the chemical-looping combustion system in commercial plants. There are many reports on the reactivity of oxygen carriers for a chemical-looping combustor, but most of the previous works were performed with a TGA or fixed bed reactor and the data of gas concentrations from oxidizer and reducer were infrequent. Moreover, there is no report on NOx emissions during oxidation at all. In this paper, natural gas combustion characteristics of new oxygen carrier particles were investigated in a batch type fluidized bed reactor (0.052 m ID, 0.7 m high). Three particles, OCN703-1100, OCN7051100, and OCN708-1300 were used as oxygen carriers. Natural gas and air were used as reactants for reduction and oxidation, respectively. To check feasibility of good performance, inherent CO2 separation, and low-NOx emissions, CH4, CO, CO2, O2, H2, NO concentrations were measured by on-line gas analyzer. Moreover, the regeneration ability of the oxygen carrier particles was investigated by successive reduction–oxidation cyclic tests up to the 10th cycle. Experimental Multi-cycle tests were carried out in a bubbling fluidized bed reactor. A schematic of the reactor is shown in Figure 1. The major components consist of a gas input system, the fluidization column, a hot gas filter, a condenser, a gas cooler, and a gas sampling/analyzing unit. The fluidization column is 0.6 m high with an internal diameter of 0.05 m. A perforated gas distributor plate separates the fluidization column and the air box. Reactant gas was fed to the air box. An electric heater could be controlled by a thermocouple and a heater controller. Thermocouple measurements of temperature and pressure transducer data were recorded by a data acquisition system. The exit stream from the fluidized bed reactor was sampled at the outlet of the reactor. The CH4, CO, CO2, H2, NO and O2 concentrations were determined using an on-line gas analyzer and recorded by the data acquisition system. Further details of the reactor system are available elsewhere [5]. The three oxygen carrier particles tested were OCN703-1100, OCN705-1100, and OCN708-1300. Prior to the start of each experiment, the oxygen carrier particle was sieved to ensure that all particles were initially between 106 and 212 mm in size. The static bed height was 0.4 m in all cases, and the experiments were carried out batchwise for the particles, i.e. no particles were added during the run. The fluidized bed reactor operated with a total inlet gas flow of 2.0 Nl/min in all cases, corresponding to superficial gas velocities of 0.07 m/s at 900oC. Particles were oxidized in air as the bed temperature was increased from room temperature to 900oC. Once the particle was fully oxidized, the particle was exposed to a gas mixture. The gas mixture contained 10% natural gas and 90% N2 for all reduction tests. The inlet concentrations were measured by bypassing the reactor. Between the cyclic oxidation and reduction tests, nitrogen was used as a purge gas. A rapid decrease in the exit gas concentration marked the end of purge. For each particle, ten cycles of reduction–oxidation were carried out. The experimental conditions and the composition of natural gas are summarized in Tables 1 and 2, respectively.
Table 1 Summary of reaction conditions OCN703-1100 Particles OCN705-1100 OCN708-1300 Particle size range [mm] 106 – 212 Metal oxide content [%] 70 Temperature [oC] 900 Inert gas (purge) N2 (2 l/min) Reduction gas
NG (0.2 l/min) + N2 (1.8 l/min)
Oxidation gas
Air (2 l/min)
Table 2 Composition of natural gas
Figure 1. Schematic of BFB reactor.
Components CH4 C2H6 C3H8 i-C4H10 n-C4H10 i-C5H12 n-C5H12 N2
Content [vol%] 88.4857 6.8617 2.9631 0.6991 0.7222 0.0337 0.0089 0.2256
Results and discussion The effects of reduction–oxidation cycling on the fuel conversion of three oxygen carrier particles during reduction are shown in Figure 2. The fuel conversion is defined as (moles of reacted hydrocarbons)/(moles of input hydrocarbon). The moles of reacted hydrocarbon are calculated by measured CO and CO2 concentration. The input hydrocarbon concentration was measured before each cycle by bypassing the reactor. The fuel conversions for all three particles were maintained constantly during 10 cycles. The average values of fuel conversion during ten cycles for OCN703-1100, OCN7051100, and OCN708-1300 particles were high, the values being 95.67, 96.06, 96.11 %, respectively. OCN708-1300 particles showed slightly higher fuel conversion than other particles. The effects of reduction–oxidation cycling on the CO2 selectivity of three oxygen carrier particles during reduction are shown in Figure 3. CO2 selectivity indicates the portion of CO2 per carbon in consumed hydrocarbons ([CO2]/[Carbon in consumed hydrocarbon]). The CO2 selectivities for all three particles were maintained constantly during 10 cycles. The average values of fuel conversion during ten cycles for OCN703-1100, OCN705-1100, and OCN708-1300 particles were high, the values being 95.67, 95.29, 94.13 %, respectively. OCN703-1100 particles showed slightly higher CO2 selectivity than other particles. Figure 4 represents the effect of reduction-oxidation cycling on the NO emission of three oxygen carrier particles during oxidation. For all particles, NO concentrations were very low and the average values for OCN703-1100, OCN705-1100, and OCN708-1300 particles were 2.73, 0.11 and 0.07 ppm, respectively. Based on the results, we could conclude that NOx-free combustion is realizable in the natural gas fueled chemical-looping combustion system with OCN703-1100, OCN705-1100, and OCN708-1300 particles.
100
100
80
80
60
60
40
40
Fuel conversion [%]
95.67 96.06 96.11
OCN703-1100 OCN705-1100 OCN708-1300
20
0 1
2
3
4
5
6
7
8
9
Average
2 selectivity [%] CO
Average
10
OCN703-1100 95.67 OCN705-1100 95.29 OCN708-1300 94.13
20
0 1
2
3
4
5
6
7
8
9
10
Number of cycles [-]
Number of cycles [-]
Figure 2 Fuel conversion versus the number of cycles.
Figure 3 CO2 selectivity versus the number of cycles.
100
80
60
40 Average OCN703-1100 OCN705-1100 OCN708-1300
20
NO concentration [ppm]
2.73 0.11 0.07
0 1
2
3
4
5
6
7
8
9
10
Number of cycles [-]
Figure 4 NO concentration versus the number of cycles. Conclusion All three oxygen carrier particles showed high gas conversion, high CO2 selectivity, and low CO concentration during reduction and very low NO emission during oxidation. Moreover, all three particles showed good regeneration ability during successive reduction-oxidation cyclic tests up to the 10th cycle. These results indicate that inherent CO2 separation, NOx-free combustion, and long-term operation without reactivity decay of oxygen carrier particles are possible in the natural gas fueled chemical-looping combustion system. However, OCN708-1300 represented temperature and pressure fluctuations during reduction and slightly decay of oxidation reactivity with the number of cycles increased. Acknowledgement This work was supported by the Power Generation & Electricity Delivery program of the Korea Institute of Energy Technology Evaluation and Planning(KETEP) grant funded by the Korea government Ministry of Knowledge Economy
References 1. Ishida, M. and H. Jin; “CO2 Recovery in a Power Plant with Chemical Looping Combustion,” Energy Convers. Mgmt., 38, S187-S192 (1996). 2. Jin, H., T. Okamoto and M. Ishida; “Development of a Novel Chemical-Looping Combustion: Synthesis of a Looping Material with a Double Metal Oxide of CoO-NiO,” Energy & Fuels, 12, 12721277 (1998). 3. Wolf, J., M. Anheden and J. Yan; “ Comparison of Nickel- and Iron-based Oxygen Carriers in Chemical-Looping Combustion for CO2 Capture in Power Generation,” Fuel, 84, 993-1006 (2005). 4. Ryu, H. J., G. T. Jin, N. Y. Lim and S. Y. Bae; “Reaction Characteristics of Five Kinds of Oxygen Carrier Particles for Chemical-Looping Combustor,” Trans. Korean Hydrogen Energy Society, 14, 2434 (2003b). 5. Ryu, H. J., D. H. Bae and G. T. Jin; "LNG Combustion Characteristics of Oxygen Carrier Particles for Chemical-Looping Combustor", 31st KOSCO Symposium, pp. 141-147, Pusan Korea (2005).
COMBUSTION REACTIVITY OF CHAR DERIVED FROM SOLVENT EXTRACTED COAL Jeonghwan Lima, Hokyung Choia, Sangdo Kima, Youngjoon Rhima, Woosik Parkb, Sihyun Leea*, a
Clean Fossil Energy Research Center, Korea Institute of Energy Research, Republic of Korea b Dept. of Chemical & Biological Engineering, Hanyang University *E-mail: [email protected]
Abstract This study produced char from ash-free coals and investigated its reactivity with air. Ash-free coal was manufactured by using the solvent extraction technique. Three different ranks of coal were used as samples: Australian lignite coal, Indonesian subbituminous coal, and American bituminous coal. 1-MN and NMP were used as solvents for extraction. The FT-IR analysis showed that the 3490 cm-1 absorption band, which usually appears in high high-rank coal, appeared in the 1-MN-extracted coal regardless of the rank of the original coal. Furthermore, the 1011–1095 cm-1 band of the extracted coal decreased greatly due to the reduction of ash. The FT-IR data of the residual coal samples was not much different from that of the original coal. The 1-MN-extracted coals showed the same burning profile regardless of the rank of the original coal, and had a higher range of burning temperatures than the bituminous coal. On the other hand, the 1-MN-extracted coal had the same burning speed as that of the original coal. The NMPextracted coals all showed lower burning temperatures than the original coal, whereas the residual coals showed a similar range of burning temperature to that of the original coal. Key words: coal, ash-free, solvent extraction, 1-MN, NMP 1. Introduction The energy market is currently unstable. Oil and coal prices are fluctuating, and the predominant outlook is that both will increase in the long term. Furthermore, because coal is the primary source of CO2, which is a greenhouse gas, the general opinion is that CO2 emissions must be reduced to solve global environmental problems. To cope with the unstable energy supply, prepare for high oil and coal prices, and minimize CO2 emissions from the use of coal, many researchers are actively studying clean technology for using coal. Among the clean technologies for using coals, ash-free coal manufacturing technology manufactures solid fuels with a similar quality to oils by removing ash or extracting carbon from coal. Ash-free coal is now being developed as Ultra Clean Coal in Australia and as Hypercoal in Japan. Ultra Clean Coal, developed by the White Energy Co., is manufactured by leaching and removing ash from coal using alkalis such as NaOH and KOH [1,2]. Hypercoal, developed by Kobe Steel Ltd., extracts carbon from coal using solvents with a high affinity to coal [3,4]. Hypercoal produces higher quality coal than Ultra Clean Coal because of the high heating value of coal. Ash-free coals can be used as fuels in various applications such as thermal power generation, direct coalfired gas turbines, DCFC (direct carbon fuel cell), and IGFC (integrated gasification fuel cell). Interest in the use of ash-free coal as a means to solve global environmental problems is increasing because this increases power generation efficiency, leading to reduced CO2 emissions. As people’s expectations of
ash-free coal rise, interest in the reactivity of ash-free coal is also increasing. This study produced char from ash-free coal at 900 °C, using the solvent extraction technique, and investigated its reactivity with air. 2. Experimental 2.1. Sample preparation Ash-free coal was manufactured using the solvent extraction technique. The details of the manufacturing process for ash-free coal are described in other papers [5]. Three different ranks of coal were used as samples: Australian lignite coal, Indonesian subbituminous coal, and American bituminous coal. They were extracted at a temperature of 370 °C and a pressure of 10 atm using two solvents, 1-MN (1methylnaphthalene) and NMP (N-Methylpyrrolidone). The extraction time was 30 minutes. Char was produced from the extracted coal under the following conditions. In nitrogen atmosphere, a proximate analyzer was used to raise temperature from normal to 900 °C at a rate of 10 °C/min. Once they reached 900 °C, the samples were maintained at this temperature for 30 minutes before being cooled to normal temperature. 2.2. Analysis For the proximate analysis of the coal samples, the TGA-701 Thermogravimeter from LECO was used. The moisture, ash, volatile matter, and fixed carbon were measured for PC, EC, and RC in accordance with the ASTM standard. For ultimate analysis, the CHN-2000 Elemental Analyzer from LECO was used to measure carbon (C), hydrogen (H), and nitrogen (N). For analysis of sulfur (S), the SC-432DR Sulfur Analyzer from LECO was used. Among the ultimate analysis items, the oxygen (O) content was determined by subtracting the weights of carbon, hydrogen, nitrogen, sulfur, and ash contents from the total weight. For heating value analysis, the Parr 1261 Calorimeter from PARR was used. For FT-IR analysis, the Nicolet 6700 FT-IR spectrometer from Thermo Fisher Scientific was used. For DTG analysis, the SDT600 thermogravimetric analyzer from TA Instrument was used.
3. Results and discussion 3.1. Proximate analysis, ultimate analysis, and heating value Table 1 shows the results of proximate analysis, ultimate analysis, sulfur analysis, and heating value analysis of the original coals. Lignite coal is a low rank coal that has a low content of fixed carbons and a high content of volatile matters because it has high water content and low carbonization degree. Thus, it exhibits a low heating value. Bituminous coal is a high rank coal that has a high content of fixed carbon and a high heating value because it has low water content and high carbonization degree. Subbituminous coal is ranked in between lignite coal and bituminous coal, being accordingly characterized. Table 1 Proximate, ultimate analysis and high heating value of original coal samples. Proximate (wt%) Samples: Original coals Moisture Volatile matter
Ash
Fixed C carbon
H
N
O
S
Lignite 12.11 Subbituminous 6.68 Bituminous 2.88
10.06 3.6 12.32
28.19 36.8 52.19
4.65 5.36 5.08
0.69 0.65 1.55
28.31 21.73 6.25
0.11 0.06 0.53
49,64 52.92 32.61
Ultimate (wt%, dry and ash free)
55.98 68.76 73.5
High heating value (kcal/kg) 5,210 5,540 6,530
The anallysis results ffor the extraccted coal (EC C) and residuual coal (RC) by solvent aand coal typee are shown in Table 2. Residual coal refers to the coal remaining r in the solvent as solid thatt was not exttracted. As shown inn Table 2, thhe moisture content in tthe extractedd coal and reesidual coal was lower than t in the original coal. The ashh content in tthe extractedd coals was m much lower thhan in the oriiginal coal: itt was 0.05– 0.12 % w when extractted by 1-MN N and 0.76–0..16 % when extracted byy NMP. Thuss, the ash conntent in the NMP-exxtracted coal was higher than t that in thhe 1-NM-exttracted coal. The ash conttent in the reesidual coal was equaal to or a littlle higher thann that in the ooriginal coal.. Table 2 P Proximate, uultimate analyysis and highh heating valuue of extractiion products.. Samples: Extractionn products Liignite
Proximate (w wt%) Moisture
EC 1.49
U Ultimate (wt%,, dry and ash ffree)
Voolatile maatter
Ash
Fixed carbon
C
H
N
O
S
32.71
0.05
65.74
822.55
6.22
0.63
100.06
0.04
High heating Y Yield value (% %) (kcal/kg) 8,340 31
45.2
13.08 36.65
655.85
4.21
1.1
144.63
0.02
6,430
-
Suubbitum EC 4.43 1-MN innous RC 5.16
61.33
0.12
34.12
866.15
6.24
1.04
6.551
0.04
8,400
30
48.3
4.36
42.18
800
4.56
1.53
9.663
0.03
7,130
-
Biitumino EC 0.54 uss RC 1.58
42.71
0.06
56.69
855.75
6.68
2.15
4.997
0.25
8,460
34
37.9
12.24 48.28
71
4.33
2.08
100.09
0.26
6,890
-
EC 1.82
56.96
1.16
755.1
7.4
6.79
9.552
0.02
7,710
64
RC 2.16
43.88
24.28 29.68
600.45
4.45
4.04
6.443
0.35
5,920
-
Suubbitum EC 1.56 NMP innous RC 1.98
54.02
0.78
43.65
81.25
6.2
5.97
5.226
0.07
7,870
72
41.99
8.53
47.51
722.25
4.73
3.78
100.64
0.07
6,590
-
Biitumino EC 1.19 uss RC 1.46
50.59
0.76
47.47
788.3
6..74
6.71
7.224
0.22
7,860
68
32.78
23.26 42.4
622.3
3.82
2.76
5.112
0.28
6,090
-
Liignite
RC 5.08
40.05
The heatting value off the extracteed coal increeased greatly, but that of the residual coal was eqqual to or a little higher than thatt of the originnal coal. Thee heating valuue of the resiidual coal increased desppite the fact that the aash content w was equal to or higher thaan that in the original coaal because waater was remooved in the extractioon process. Fuurthermore, tthe oxygen aand sulfur conntents in the brown coal aand subbitum minous coal decreaseed. The meaan extraction yield was arround 30 % ffor 1-MN exttraction and aaround 65 % for NMP exxtraction. 1MN, whhich is a non--polar solvennt, only extraacts coal com mponents thaat are thermallly relaxed; in i contrast, NMP, w which is a polar solvent, uses u both theermal relaxaation and relaaxation by soolvent [6,7]. Therefore, the yieldd of the NM MP extractioon is higher than thatt of the 1-MN N. 3.2. FT-IIR analysis The FT--IR spectra ffor the originnal coal, 1MN exxtraction pproducts, aand NMP extractioon products aare comparedd in Fig. 1, Fig. 2 aand Fig. 3, rrespectively. FT-IR has been useed often for thhe analysis oof functional groups of o coal such as C-H, C-O O, and O-H. The smaall absorptioon band of 3050–3030
Fiig. 1 FT-IR sspectra for thhe original cooals
cm-1 indicates an aroomatic CH sttretch band, which show ws a cross-linnked structuree. Because oof this, it is generallyy reported thhat it does noot appear in thhe low-rank coals, and only appears in high-rank coals. The two absoorption bandss of 2920 cm m-1 and 2850 cm-1 represeent the aliphaatic CH3, CH H2, and CH ggroups. The absorptioon bands thaat appear aroound 1600 cm m-1 indicate aromatic a ringg vibration ddue to the binding with oxygen iin quinone, ether, and hheterocyclic ccompounds. The 1500 cm-1 absorptioon band represents the C=C binnding in the benzene b ringg. The absorpption band of 1450–14400 cm-1 repressents the CH H2 group, as -1 does thee 1380–1375 cm band. The 1370 cm m-1 and 14660 cm-1 absoorption bandss represent m methyl and methylenne, respectivvely. The 10000–1300 cm m-1 section caan be generaally divided iinto the alipphatic ether group att 1000–12000 cm-1 and thhe aromatic ether groupp in 1200–13300 cm-1. Thhe 900–700 cm-1 band represents the C-H ggroup combiined with thee aromatic compounds. c The sectionss 500–600 cm m-1, 1000– -1 -1 m , and 3600––3700 cm represent the bands of the minerals conntained in cooal [8,9]. 1100 cm As show wn in Fig. 2, 2 the absorpption band ccan be seen at 3490 cm m-1 for 1-MN N-extracted coals; this correspoonds to the aaromatic CH H stretch thaat appears inn high-rank coal. Another differencee from the original coals is thatt it is possiblle to clearly observe the CH2 or metthyl/methylenne group, aroomatic and d aliphaticc ether groupps, and the disappearanc e of the 10111–1095 cm--1 band due to the decrease in ash. Howeverr, the FT-IR of the residuual coal did not n show mucch difference from that off the original coal. In the caase of NMP extraction, aas shown in Fig. 3, the aabsorption baands of the extracted e coaal does not show muuch differencce from thosee of the originnal coal, exceept for the reemoval of ashh.
(a) 1-M MN extracted coals
(bb) 1-MN residdue coals
R spectra for tthe coals prooduced by 1-M MN extractioon. Fig. 2 FT-IR
(a) NM MP extracted ccoals
(bb) NMP residdue coals
Fig. 3 FT-IR R spectra for the coals prooduced by NM MP extractioon.
3.3. DTG G analysis Fig. 4 coompares the DTG burninng profiles off coal char saamples: the cchar of originnal coal, extrracted coal and residdual coal froom 1-MN annd NMP exttraction. As shown in Fiig. 4, the oriiginal coal sshowed the typical bburning charaacteristics thaat the higherr the rank, the higher the burning tem mperature. In the case of brown cooal, two peakks appeared because of thhe low fixedd carbon conttent and the hhigh content of volatile matters. That is, the volatile mattter burned firrst, and the ffixed carbons burned nexxt. Two peakks appeared because the contentss of volatilee matters andd fixed carbbons are sim milar. This phhenomenon sometimes appears in bituminouus coals as w well, in whichh case the buurning of the volatile mattters before tthe burning of the fixxed carbons appears as a small shoulder. In termss of DTG peaak height, orr the maximuum burning rate, the reactivity off the subbitum minous coal was the highhest, and the bituminous coal had a bbroad range of burninng temperatuures. The 1-M MN-extractedd coal show wed peculiar characteristiics. The 1-M MN-extractedd coals hadd the same burning profile, regarrdless of the rank of the ooriginal coal,, and had a w wider range oof burning tem mperatures G peak, the burning speeed of brownn coal was thhe highest, than bituuminous coaals. In terms of the DTG followedd by subbitum minous coal, and finally bituminous b cooal. On the other o hand, thhe 1-MN-exttracted coal had the ssame burningg speed as thaat of the origginal coal.
((a) Original coals c
(b) 1-M MN extractedd coals
(cc) 1-MN residue coals
(d) NM MP extractedd coals
(ee) NMP residdue coals
Fiig. 4 DTG buurning profilees of coal sam mples.
As show wn in the FT-IIR results abbove, 3490 cm m-1 was obseerved in the 11-MN-extractted coal, whiich appears in high-rrank coal; thhus, this indiccated an incrreased rank. The results of the burninng profile also indicate the samee trend becauuse all the burrning temperratures increaased. Unlike thhe 1-MN-exttracted coal, the NMP-exttracted coals all showed llower burninng temperaturres than the original coal, whereaas the residuual coals show wed a similaar range of bburning temperatures as tthe original milar burning temperatures to that of tthe original coal. Thuus, we can cconclude thatt residual coaals show sim coal regaardless of sollvent, but thee extracted cooals show low wer burning ttemperaturess than the original coal. Fig. 5 shhows the connversion rates of each coaal by temperrature. The reesidual coal conversion rrates of the coals exttracted by eaach solvent aare similar too that of the original coall. The subbittuminous coaals showed similar cconversion raates by tempperature, although the tem mperature vaaried slightly. In compariison, the 1MN-extrracted coals sshowed simillar conversioon rates, regaardless of rannk. In the casee of the NMP P-extracted coals, thhe subbituminnous coal shhowed an infflection pointt at conversiion 0.6, and the bituminoous coal at conversion 0.8, whicch seems to bbe due to chaanges in the burning b reacttion mechaniisms. This phhenomenon can be allso seen in thhe DTG burnning profile inn Fig. 4.
(aa) Original ccoals
(b) 1-M MN extractedd coals
(cc) 1-MN residdue coals
(d) NM MP extracted coals
(ee) NMP residdue coals
o coal samplles. Fig. 5 Conveersion rates of
4. Conclusions Ash-free coal was manufactured from coal using the solvent extraction technique. Char was produced from the extracted ash-free coal at 900°C in an inert atmosphere and the reactivity between the char and air was examined. The FT-IR analysis showed that the 3490 cm-1 absorption band, which usually appears in high-rank coal, appeared in the 1-MN-extracted coal regardless of the rank of the original coal. Furthermore, the 1011– 1095 cm-1 band of the extracted coal decreased greatly due to the reduction of ash. The FT-IR spectra of the residual coal samples were not much different from those of the original coals. The 1-MN-extracted coals showed the same burning profile regardless of the grade of the original coal, and had a higher range of burning temperatures than the bituminous coal. On the other hand, the 1-MNextracted coal had the same burning speed as that of the original coal. The NMP-extracted coals all showed lower burning temperatures than the original coal, whereas the residual coals showed a similar range of burning temperature to that of the original coal. References 1. Steel KM, Patrick JW, The production of ultra clean coal by chemical demineralization. Fuel 2001;80:2019-2023 2. http://www.det.csiro.au/science/lee_cc/ultra_clean_coal.htm 3. Okumura N, Komatsu N, Shigehisa T, Kaneko T, Tsuruya S, Hyper-coal process to produce the ashfree coal. Fuel Process Technol 2004;86:61-72 4. Takanohashi T, Shishido T, Kawashima H, Saito I, Characterization of hypercoal from coals of various ranks. Fuel 2008;87:592-598 5. Kim SD, Woo KJ, Jeong SK, Rhim YJ, Lee SH, Production of low ash coal from LRC (low rank coals) by thermal extraction with NMP (N-methyl-2-pyrrolidinone). KJChE 2008;25(4): 758-763 6. Li C, Takanohashi T, Saito I, H Aoki, K Mashino, Elucidation of mechanisms involved in acid pretreatment and thermal exstraction during ashless coal production. Energy Fuels 2004;18:97-101 7. Kashimura N, Takanohashi T, Saito I, Effect of noncovalent bonds on the thermal extraction of subbituminous coals. Energy Fuels 2006;20:1605-1608 8. Gupta R, Advanced coal characterization: a review. Energy Fuels 2007;21:451-460 9. Xuguang S, The investigation of chemical structure of coal macerals via transmitted-light FT-IR microspectroscopy. Spectrochimica Acta Part A 2005;62:557-564
th
27 International Pittsburgh Coal Conference October 11–14, 2010, Istanbul, Turkey
FORCED FLAME RESPONSE MEASUREMENT IN A GAS TURBINE COMBUSTOR WITH HIGH HYDROGEN FUEL K.T. Kim*, J.G. Lee, B.D. Quay, D.A. Santavicca Center for Advanced Power Generation Department of Mechanical and Nuclear Engineering The Pennsylvania State University, University Park, PA, 16802
ABSTRACT The forced response of swirl-stabilized lean-premixed turbulent flames to acoustic oscillations in a hydrogen enriched laboratory-scale gas turbine combustor was experimentally investigated. Nonlinear flame transfer function measurements were taken to investigate the flame’s heat release response to upstream acoustic perturbations. This analysis shows that the dynamics of natural gas-air premixed flames are characterized by several regimes: the linear, transition, and first and second nonlinear regimes, depending upon steady-state flame geometry, modulation frequency, and amplitude of excitation. The present results show that the flame geometry changes from a dihedral V flame to an enveloped M flame with an increase in hydrogen mole fraction, and the changes in steady-state flame structures have a significant impact on the flame’s response to acoustic modulations. The present results suggest that the M flame, unlike the V flame, has the unique dynamic characteristic of acting as a damper of flow perturbations. The response of the M flame remains in the linear regime, irrespective of the shedding of a vortex-ring structure, because the interaction between the large-scale structure and the flame is not strongly coupled. INTRODUCTION Combustion instabilities are characterized by large-amplitude pressure oscillations that are driven by unsteady heat release. It is by nature a self-excited oscillation, involving complicated physical phenomena such as unsteady combustion, acoustic fluctuations, heat transfer, and the vorticity field. These highamplitude pressure oscillations can substantially reduce the performance of a system and in extreme cases can cause structural damage to engines. Over the last two decades, a large number of studies using experimental [1-4], theoretical [5-8], and CFD simulation [9-11] approaches have been performed to investigate instability mechanisms, controlling parameters, and active or passive control methodologies to suppress the instability intensity level to acceptable ranges. Recently, Huang and Yang [12] have provided comprehensive reviews on this issue. Ultimately, to accurately predict instability characteristics at the development stage, a complete understanding of the nature of combustion instabilities will be required and sophisticated theoretical models must be developed to
incorporate the complex flame dynamics into the prediction model. Development of a theoretical model capable of predicting the formation of thermoacoustic pulsations is contingent on a physical understanding of the linear and nonlinear responses of a flame to periodic disturbances. When there are no equivalence ratio oscillations in the mixing plenum, the response of a premixed flame is primarily governed by acoustic velocity fluctuations. To quantitatively describe the response of a flame to acoustic forcing, the nonlinear flame transfer function (FTF) is introduced, defined as the normalized ratio of heat release and velocity fluctuations:
FTF ( f , A) =
Q '( f ) Q V '( f ) V
(1)
where Q is the time-averaged heat release rate, V is the mean velocity of the mixture in the mixing section, Q '( f ) and
V '( f ) are their corresponding amplitudes at the forcing frequency, f, and A is the magnitude of V '( f ) / V . The flame transfer function can be obtained by experimental [2-4, 13-14], theoretical [6-8], or numerical [5] methods. The flame transfer function provides an insight into the response of a flame to inlet disturbances, and it can be mathematically formulated and used in the thermoacoustic network modeling as a source term to predict self-induced combustion instability [14]. In the present article, the effects of fuel composition on the forced response of a swirled lean-premixed flame are experimentally investigated. Development of fuel-flexible gas turbine engines capable of operating with highly variable and potentially low quality fuels, such as coal-derived syngas fuels, while producing minimal air pollutants, is a key issue in the gas turbine combustion community [15-16]. Most previous studies on the effects of fuel composition on applications in gas turbines have dealt with static stabilities, pollutant emissions [17-20], flame propagation speed [21-23], and steady-state flame structures [24-25]. It was found that the addition of hydrogen to hydrocarbon fuels results in a significant increase in OH radical concentration, extending the lean stability limits and increasing flame propagation speeds. However, there have been relatively few experimental investigations on the influence
*
Corresponding author: University of Cambridge, Engineering Department, Trumpington Street, Cambridge, CB2 1PZ, UK. E-mail address: [email protected].
1
ICCD camera CH* inteferance filter, 432 nm
choked inlet for self-excited instability measurement
PT5 PT3 PT4
PT1 PT2
siren
forced air+fuel swirler
38.1
T2
T3
plug
centerbody
19.1
T1
1500
exhaust
762 ~ 1524
76.2
T4
109.2
PTc
333.5
334.8
mixing section
quartz combustor
steel combustor
CH*(432nm), OH*(307nm), CO2*(365nm) filters PMTs
Figure 1. Schematic of a swirl-stabilized, lean-premixed, model gas turbine combustor. Dimensions in millimeters.
of fuel composition (particularly, hydrogen-enrichment) on combustion dynamics [26-27]. It is not possible to make general statements about the impact of fuel composition variation upon a combustor’s propensity to become unstable– the effects can be either stabilizing or destabilizing. In the present paper, we focus on the linear/nonlinear dynamics of H2-enriched, swirl-stabilized, lean-premixed flames using a phase-resolved analysis and nonlinear flame transfer function measurements. Two key mechanisms of nonlinearity, i.e., shear layer rollup and unsteady flame liftoff for dihedral V flames, are examined. In particular, the dynamics of H2enriched enveloped M-shaped flames are compared with pure natural gas flames with inverted dihedral V flame geometry. It is shown that modification of steady-state flame structures significantly affects the linear and nonlinear dynamics of a laminar premixed flame [28]. The present results suggest that the M flames do not exhibit the strong overshoot behavior which is observed in the case of dihedral V flames. This study extends previous experimental and analytical results from laminar to turbulent premixed flames. EXPERIMENTAL METHODS Lean Premixed Gas Turbine Combustor A schematic of the experimental setup is shown in Figure 1. This facility consists of an air inlet section, a siren, a mixing section, an optically-accessible quartz combustor section, a steel combustor section, and an exhaust section. The air can be heated to a maximum temperature of 400 °C by a 30 kW electric heater. A siren-type modulation device (rotor + stator) is used to provide acoustic modulations. The siren is driven by a
variable-speed DC motor, providing capabilities for changing forcing frequency (~ 500 Hz). Two high-temperature globe valves are used to control the bypass flow rate, i.e., the inlet velocity fluctuation amplitudes. The mixing section is 0.333 m long and has an annular cross-section that is defined by a 19.1 mm O.D. centerbody and a 38.1 mm I.D. mixing tube. The centerbody is centered in the mixing tube, and it is positioned such that its downstream end is flush with the combustor dump plane. A 30o flat-vane axial swirler is mounted in the mixing tube 76.2 mm upstream of the combustor dump plane. At the entrance to the mixing section the flow is choked. This provides a well-defined acoustic boundary condition for self-excited instability measurements. For forced flame response measurements, the choking plate is removed and mounted upstream of the siren. Fuel (natural gas + H2) is injected and mixed upstream of the choked inlet in order to ensure that the reactant mixtures are spatially and temporally homogeneous before they enter the reaction zone. The combustor consists of a stainless steel dump plane, to which an optically accessible fused-silica combustor with a diameter of 109.2 mm and length of 334.8 mm is attached. The downstream end of the quartz combustor is connected to a stainless steel variable-length combustor section. The length of the combustor can be continuously varied between 762 mm and 1524 mm by moving a water-cooled plug along the length of the steel combustor section. For forced response measurements, the combustor length was maintained at the minimum to minimize the influence of system acoustics on upstream acoustic excitation.
2
3
360 V'/Vmean=9%(Linear)
(B)
V'/V mean=15%(Transition)
270
V'/V mean=25%(Nonlinear I)
Gain
Phase (degree)
2
1
180
90 V'/Vmean=59%(Nonlinear II)
(A)
0
0
0.2
0.4
0.6
0
0.8
0
0.2
0.4
0.6
0.8
V'/V mean
V'/V mean
Figure 2. (A) The gain and (B) phase of flame transfer function as a function of the magnitude of inlet velocity perturbation at a modulation frequency of 200 Hz. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00. Instrumentation and Test Conditions High frequency-response, water-cooled, piezoelectric pressure transducers (PCB 112A04) were used to measure pressure perturbations in the mixing (denoted by PT1−PT4 in Figure 1) and the combustor sections (PTc and PT5). The pressure signals were conditioned by amplifiers, digitized by an analog-todigital converter, and stored in microcomputer memory for processing at a sampling rate of 8192 Hz. A total of 16,384 data points were taken during each test, resulting in a frequency resolution of 0.5 Hz and a time resolution of 0.122 msec. Spectral analysis of the signals was performed using the fast Fourier transform (FFT) technique. Two pressure transducers located at 12.7 mm (PT4) and 50.8 mm (PT3) upstream of the combustor dump plane were used to estimate the inlet velocity fluctuations using the two-microphone method. To calibrate the two-microphone method, direct measurements of velocity fluctuations were performed under cold flow conditions with a hot wire anemometer (TSI 1210-20). A photomultiplier tube (PMT, Hamamatsu model H7732-10) was used to measure the global CH* (432 ± 5 nm) chemiluminescence emission intensities from a whole flame. An ICCD camera (Princeton Instruments model 576G) with a CH* band pass filter centered at 430 nm (10 nm FWHM) was used to record the flame images. For phase-averaged imaging, the ICCD camera was synchronized with the combustor pressure signal and fifty averaged images were taken over a cycle of oscillation with a phase interval of 30°. The phase angle φ = 0° corresponds to the positive-to-negative zero transition, and φ = 270° to the maximum combustor pressure. Because CH* chemiluminescence images are line-of-sight integrated images,
a three-point Abel deconvolution scheme was used to extract two-dimensional information from the line-of-sight images. All tests were performed at a mean pressure of 1 atm and at mean equivalence ratio of 0.60. Mean velocity at the nozzle was 60 m/s and the inlet temperature was kept constant at 200 °C, giving a Reynolds number of approximately 33,000. Forcing frequencies were varied from 100 to 400 Hz ( ∆f = 25 Hz). The frequency range of 100−400 Hz includes two distinct frequencies at limit cycle oscillations observed in the lean premixed gas turbine combustor [27]. H2-blended (based on volume) natural gas fuels with X H 2 = 0.00 and 0.30 were used to examine the impacts of fuel composition variation upon the response of the flame to acoustic excitation. RESULTS AND DISCUSSION Linear and Nonlinear Dynamics of Inverted Dihedral V Flames Experiments were performed wherein the heat release response of hydrogen-enriched natural gas-air premixed flames to acoustic forcing was determined for various levels of inlet velocity oscillations. Under the conditions investigated here, the highest perturbation levels are not determined by flame blowoff, but by limitation of the modulating device. The magnitude of upstream forcing which can be achieved by the device is contingent on modulation frequency. It was observed that the maximum amplitude can be attained at a modulation frequency of near 200 Hz, inasmuch as the combustor mixing section with upstream plenum acts as a resonator with peak responses around that frequency. The frequency, f ≈ 200 Hz, also corresponds to one of the self-excited instability frequencies observed in the test rig [27]. In the present study, the flame
3
Figure 3. Phase-synchronized deconvoluted CH* chemiluminescence images at a modulation frequency of 200 Hz for (A) V ' Vmean = 9%(top) and (B) V ' Vmean = 15%(bottom). The intensity is displayed in linear pseudo color scale that white denotes the highest intensity and black means the lowest intensity. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00.
transfer function measurements were performed at a modulation frequency of 200 Hz for different fuel compositions. It is believed that the response of the flame at f = 200 Hz will help to illuminate the flame dynamics under limit-cycle oscillations at a frequency near 200 Hz. Figure 2 presents the amplitude dependence of the gain and phase of the flame transfer function (see Eq. 1) at a modulation frequency of 200 Hz. These measurements were made at an inlet temperature of 200 oC, mean nozzle velocity of 60 m/s, mean equivalence ratio of 0.60, and hydrogen mole fraction of 0.00. The global CH* chemiluminescence intensities were used as an indicator of heat release rate oscillations, based on experimental observation that CH* chemiluminescence intensity increases linearly with fuel flow rate at a given overall equivalence ratio. The coherence function between inlet velocity and CH* chemiluminescence intensity was measured to close to unity at the forcing frequency of 200 Hz, enabling flame transfer functions to be accurately determined. Also, a high velocity perturbation magnitude of up to approximately 60% can be achieved at this forcing frequency. At these inlet conditions, the flame is stabilized in the inner shear layer and it is attached to the centerbody, exhibiting an inverted dihedral V structure [14]. The responses of the dihedral V flame to acoustic perturbations are divided into several regimes, as shown in Figure 2. In the linear regime ( V ' Vmean < 10%), the normalized heat release response increases linearly with the amplitude of excitation; thus, the gain is almost constant. When the magnitude of inlet velocity fluctuation is greater than 10%, the gain decreases with an increase in perturbation amplitude, representing typical nonlinear behavior. It is noteworthy that when the modulation amplitude is greater than 40% of the mean value, the normalized CH* chemiluminescence intensity levels
off again. Similar characteristics of acoustically forced swirlstabilized flames were also observed by Thumuluru and Lieuwen [29]. They state that the flame’s heat release response exhibits multiple saturating behaviors and a non-monotonic dependence upon amplitude. Phase-resolved flame imaging measurements in each regime will elucidate the amplitudedependent dynamic characteristics of acoustically forced swirl flames. Figure 2 (B) shows that the phase measured between the CH* chemiluminescence signal and the inlet velocity fluctuation is essentially independent of the modulation amplitude up to V ' Vmean = 0.60. This is consistent with previous studies on laminar premixed flames [28], and indicates that the frequency-dependent convection times for a velocity perturbation to reach the flame are not a function of the modulation amplitude, irrespective of linear and nonlinear regimes. To identify the key physical processes associated with the linear/nonlinear response of the dihedral V flame to flow forcing, phase-averaged chemiluminescence emission from the whole flame was recorded using an intensified CCD camera. Figure 3 presents a sequence of phase-synchronized CH* chemiluminescence images at a modulation frequency of 200 Hz during a period of oscillation for excitation amplitudes of (A) V ' Vmean = 9% and (B) 15%. They represent the response of the flame in the linear and transition regimes, respectively. Note that only the upper half of these deconvoluted images is shown, because the reconstructed images are axisymmetric. The direction of flow is from left to right. It can be observed from Figure 3 (A) that the flame moves back and forth due to the velocity oscillation in the annular jet region, and the reaction zone is located in the inner shear layer. Reaction occurs in the corner recirculation zone (CRZ) at φ = 90 ~ 180° when the
4
Figure 4. (A) Phase-synchronized deconvoluted CH* chemiluminescence images at modulation frequency of 200 Hz and V ' Vmean = 25% (top). (B) Line-of-sight integrated, background-corrected CH* chemiluminescence images at φ = 60, 150, 240, and 330° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00.
flame moves upstream. Evolution of flame surface area by shear layer rollup is not seen because the minimum excitation level for the formation of a large-scale coherent structure is not achieved. Figure 3 (B) shows the forced flame response in the transition regime. The flame shapes at each phase are similar to those of the linear flame behavior, displayed in Figure 3 (A). The formation and reaction of large-scale, coherent vortex structures are, however, observed at φ = 180 ~ 270°. They are convected by the mean flow, and the heat release reaches its maximum at φ = 270°. The vortices periodically entrain a large amount of combustible material into the reaction zone at the modulation frequency, and therefore the temporal variations of flame surface area are determined by the interplay between the coherent structure and the flame. Deconvoluted and line-of-sight integrated phase-averaged CH* chemiluminescence images in the first nonlinear regime are shown in Figures 4 (A) and (B), respectively. All inlet conditions are the same as for Figure 3, except for the modulation amplitude. The most distinct difference in flame behavior between the transition and the first nonlinear regimes is that in the first nonlinear regime, the flame front bends toward the inner recirculation zone at φ = 270°, as compared to the deconvoluted flame image at φ = 270° in Figure 3 (B). The interaction between a vortex-ring structure and the flame is significant at φ = 180° ~ 300°. The line-of-sight integrated images at φ = 150° and φ = 240° clearly show the vortex rollup and convection processes, respectively. It has been reported that the nonlinear response of premixed flames is related to the shear layer rollup [2, 4, 14, 29]. The shear layer rollup shortens the flame length, which in turn decreases the flame area in a nonlinear manner. Hence, the gain of the flame transfer function decreases in the nonlinear regime, as shown in Figure 2. Figure 5 shows phase-resolved flame images at a modulation amplitude of 59%, corresponding to the second nonlinear regime. Note that the normalized CH* chemiluminescence intensity is constant with respect to the
amplitude of excitation when V ' Vmean is greater than 40%. It is expected that the nonlinear dynamics of the flame is influenced by different controlling physics than in the first nonlinear response regime. The deconvoluted and line-of-sight integrated images at φ = 330° clearly show the unsteady flame liftoff from the attachment point. The high level of inlet velocity perturbation causes the flame attachment point to move off of the centerbody to the downstream inner recirculation zone. This indicates that when the inlet velocity perturbation magnitude is greater than 40% of the mean value, the dynamics of swirlstabilized premixed flames subjected to upstream acoustic forcing are less influenced by the flame-vortex interaction. The unsteady flame liftoff (flame holding) plays a dominant role in the flame dynamics, due to the high perturbation amplitude. At V ' Vmean > 40%, a vortex ring structure is shed (see φ = 150° in Figure 5), and it is convected with the mean flow in the annular jet region, but the main reaction region is located in the inner recirculation zone. Therefore, the response levels off when the magnitude of acoustic forcing is greater than 40%, as shown in Figure 2. To quantitatively describe the spatial distribution of the flame’s heat release depending on excitation amplitudes, coordinates of maximum CH* chemiluminescence intensity locations at φ = 270° for perturbation levels of V ' Vmean = 9, 15, 25, and 59% are plotted in Figure 6. It clearly shows that with increasing modulation amplitude, the intense reaction region moves toward the inner recirculation zone. In consequence, the key mechanism of nonlinearity is modified from shear layer rollup to unsteady flame liftoff, depending on perturbation amplitude. The amplitude-dependent dynamics of inverted V flames suggest that limit-cycle pressure oscillation magnitudes are also governed by different mechanisms of flame area modulation under self-sustained oscillations, depending on the magnitude of inlet velocity fluctuation. An experimental investigation of the dynamics of self excited flames and a
5
Figure 5. (A) Phase-synchronized deconvoluted CH* chemiluminescence images at modulation frequency of 200 Hz and V ' Vmean = 59% (top). (B) Line-of-sight integrated, background-corrected CH* chemiluminescence images at φ = 60, 150, 240, and 330° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00.
60
50
Y (mm)
40
30
20 V'/V mean = V'/V mean = V'/V mean = V'/V mean =
10
0
0
40
80
120
9% 15% 25% 59%
160
X (mm)
Figure 6. Coordinates of maximum CH* chemiluminescence intensity location at phase of φ = 270° (maximum CH* intensity) for inlet velocity magnitudes of V ' Vmean = 9, 15, 25, and 59%.
comparison between acoustically forced flames and flames under limit cycle oscillations are left for future studies. Forced Response of H2-Enriched Enveloped M Flames In previous work [27], the authors have provided details on the modification of stable flame configurations, depending upon H2 volume fraction in composite fuels. It has been shown that the flame geometry changes from an inverted dihedral V structure to an enveloped M geometry with increasing H2 mole fraction. The M flames have unique characteristics that allow them to damp flow perturbations, as compared to V flames. Thus, the
flame transfer function gain of enveloped M flames is much smaller than that of V flames for a given forcing frequency and amplitude. This is qualitatively consistent with theoretical and experimental studies on the response of laminar premixed flames by Schuller et al. [28, 30-31]. They described V flame behavior as an amplifier of flow perturbation in a certain range of frequencies, with the gain of laminar V flames being greater than that of conical and M-shaped flames. In this section, we will discuss the manner in which the combustion instability characteristics of a given gas turbine combustor are impacted by hydrogen levels in the fuel. The amplitude dependence of the normalized CH* chemiluminescence emission intensity at a modulation frequency of 200 Hz for a H2-enriched M flame with an enveloped geometry is shown in Figure 7. All inlet conditions are the same as for Figure 2, except for H2 mole fraction. Hydrogen blended natural gas fuel (30% H2 + 70% natural gas) was used and the overall equivalence ratio was kept constant at φoverall = 0.60. It can be observed from Figure 7 that the normalized CH* chemiluminescence intensity increases linearly with the amplitude of excitation up to approximately V ' Vmean = 50%, indicating that the flame behaves linearly with respect to flow perturbations. The fact that the response of the flame remains in the linear regime even at a high perturbation amplitude of 50% is an unexpected result. As previously shown in Figures 4 and 5, excitation of a large scale coherent structure plays an important role in inducing nonlinearity when the normalized acoustic velocity amplitude is greater than 25% at a frequency of 200 Hz. Considering that the formation and convection of the structure are governed by fluid dynamic conditions, not by fuel composition, it can be supposed that the H2-enriched M flame may feature significantly different response characteristics in comparison with the dihedral V flame. In order to elucidate the underlying physics, twelve successive phase-resolved CH* chemiluminescence images
6
1 X H2 = 0.00 X H2 = 0.30
CH*'/CH*mean
0.8
0.6
0.4
0.2
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
V'/V mean
Figure 7. The normalized heat release response as a function of modulation amplitude at a modulation frequency of 200 Hz. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00, 0.30.
taken during a single cycle at a modulation frequency of 200 Hz and amplitude of V ' Vmean = 25% are illustrated in Figure 8. Line-of-sight integrated, background corrected images at the phases 90, 150, 210, and 270° are also presented for comparison. The figure shows that with H2-enrichment ( X H 2 = 0.30), the spatial distribution of the reaction zone becomes very compact in both axial and radial directions. It is obvious from Figure 8 (A) that the flame angle variation is substantial during a modulation cycle. In the case of V flames, however, the flame oscillates in the flow direction, which means that the overall flame length changes over a period of perturbations, as shown
in Figures 3 and 4. This result demonstrates that the dynamics of M flames is significantly different from that of V flames. The flame angle and CH* chemiluminescence intensity variations over one period of oscillation are shown in Figure 9. It can be seen that the variations in flame angle and CH* chemiluminescence intensity are out of phase. As the flame moves back (largest flame angle, see the phase 180°), the flame brush is almost vertical with respect to the flow direction and the CH* chemiluminescence intensity is low due to decreased flame area. These oscillations in the flame position (angle) result from fluctuations in the outer shear layer. It has been reported that if the shear layers in the swirl-stabilized gas turbine combustor are acoustically excited to form a coherent vortex ring structure, the inner shear layer rolls up inward and the outer shear layer rolls up outward [2, 32]. Note that even though periodic formation and convection of vortex structures occurs at φ = 210−270°, the flame response remains in the linear regime, as shown in Figure 7. The formation and convection of large scale structures are determined by fluid dynamic conditions, such as inlet velocity, forcing frequency, and amplitude. As already shown in Figures 2 (A) and 4, for a given set of fluid dynamic conditions the inverted V flame with long flame length exhibits strong interaction with a large scale structure, leading to highly nonlinear behavior. In contrast, an M flame is less influenced by the Kelvin-Helmholtz instability, since the relative ratio of convective time to the acoustic forcing period is very small, τ M − flame < τ V − flame . Furthermore, the M flame oscillates such that the flame angle varies significantly over a period of oscillation. The convection of a vortex structure is mainly determined by the mean flow, implying that the flame and the vortex structure exist in different paths for most of a given period. Hence there is not enough time for the coherent structure to interact with the flame. This further suggests that shedding of a large scale structure is not a sufficient condition for the nonlinear response of swirlstabilized flames to acoustic excitation; the interaction between the structure and the flame plays a more important role.
Figure 8. (A) Phase-synchronized deconvoluted CH* chemiluminescence images at a modulation frequency of f = 200 Hz and V ' Vmean = 25% (top), (B) line-of-sight integrated CH* chemiluminescence images at φ = 90°, 150°, 210°, and 270° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30.
7
0.6
0.6
50 CH* intensity Flame angle
100 150 200 250 300
0.5
0.4
30
0.4
CH*'/CH*mean
40
Flame angle (degree)
CH* intensity (a.u.)
0.5
Hz Hz Hz Hz Hz
0.3
0.2
0.3 0.1
0.2
0
120
240
0
20 360
0.1
0.2
0.3
0.4
0.5
0.6
V'/V mean
Phase (degree)
Figure 9. Time traces of global CH* chemiluminescence intensity and flame angle over one period of oscillation. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30, f = 200 Hz, and V ' Vmean = 25%.
0
Figure 10. The normalized heat release response plotted as a function of forcing amplitude for a range of forcing frequency from 100 to 300 Hz. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30.
inducing nonlinearity.
Figure 11. (A) Phase-synchronized deconvoluted CH* chemiluminescence imaging at a modulation frequency of f = 300 Hz and V ' Vmean = 30% (top). (B) Line-of-sight integrated, background-corrected CH* chemiluminescence images at φ = 90, 150, 210, and 270° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30.
Figure 10 shows the normalized CH* chemiluminescence intensity plotted as a function of forcing amplitude for the range of frequency 100−350 Hz. Unlike the cases without H2enrichment at the same inlet flow condition, the flame tends to remain in the linear regime even at high forcing frequency (f = 300 Hz) and amplitude ( V ' Vmean = 0.351). This result shows good agreement with data from a theoretical study by Lieuwen [33]. He reported that the response of the conical flame is nearly linear over the entire velocity disturbance amplitude,
while the V flames exhibit substantial nonlinearity, which is manifested as saturation. This evidence strongly suggests that the flames with enveloped M configuration are more stable than inverted V flames in terms of combustion instability. The modification of the steady-state flame configuration from a V to an M geometry by H2-enrichment was found to reduce the limitcycle pressure oscillation amplitude significantly [34]. It is interesting to note that in contrast with V flames, the normalized heat release response of M flames is insensitive to the forcing
8
6
3 X H2 =0.00(V structure) X H2 =0.30(M structure)
5
4
Gain
Phase (rad)
2
3
2
1
1
0
0
100
200
300
400
0
500
0
100
200
300
400
500
Frequency (Hz)
Frequency (Hz)
Figure 12. The gain and phase of the flame transfer function as a function of modulation frequency at a constant forcing amplitude, V ' Vmean = 0.100. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00, 0.30.
frequencies, as shown in Figure 10. The measured CH* chemiluminescence intensity for a range of frequencies 100−300 Hz overlaps on the same straight line. This can be attributed to the fact that the dynamics of the M flame is less influenced by the formation and convection of large-scale structures, which is by nature dependent on perturbation frequency. The normalized CH* chemiluminescence intensity starts to decrease when the modulation frequency is greater than 350 Hz (not shown). Figure 11 shows phase-resolved CH* chemiluminescence images at a forcing frequency of 300 Hz and amplitude of 30% at the same inlet conditions as those shown in Figure 8. It can be concluded from Figures 8 and 11 that the dynamics of acoustically forced flames appear very similar, regardless of the forcing frequencies and amplitudes. The reason is that the response of the flames is linear in both cases and the forced response is independent of forcing frequencies. Under the conditions investigated here, shear layer rollup does not play an important role in inducing nonlinearity. The flame transfer functions in frequency domain are presented in Figure 12. The velocity modulation amplitude was maintained at a constant value of V ' Vmean = 10%. At this condition, the forced flame response is linear below the level of acoustic velocity perturbation of V ' Vmean = 10%. As reported by recent studies [2, 27, 28], the global flame response can be described as a low-pass filter, in that the gain decreases as the frequency increases. These results also show that with increasing X H 2 , the maximum gain of the flame transfer function decreases significantly and the phase decreases at a given forcing frequency, indicating that the gain depends strongly on flame length and steady-state flame configurations.
For the inverted V flames, the gain of the flame transfer function exceeds unity at the modulation frequency of 100 ≤ f ≤ 375 Hz, indicating that the flame amplifies flow perturbation at certain frequencies. In the case of M flames, however, the overshoot behavior is not observed, i.e., the gain is less than unity, suggesting that M flames damp upstream flow perturbations. Also, the gain is almost independent of forcing frequency when the frequency is less than 300 Hz. At high forcing frequencies, the gain asymptotically approaches zero. The phase of V flames is greater than that of M flames, due to the fact that the convection time for velocity disturbances to reach the flame increases with increasing the flame length. The phase increases almost linearly with frequency, confirming that the influence of flow disturbances on the flame’s heat release is purely convective. In the present article, the combustion response of turbulent premixed flames to acoustic oscillations was experimentally determined. The effects of spatio-temporal fluctuations of equivalence ratio nonuniformities were not considered, however. In a real gas turbine engine environment, mixture ratio oscillations can be induced in the mixing section, since fuel is generally injected at the swirl vanes or near the inlet of the combustion chamber to avoid auto-ignition and improve mixing of the fuel and air mixture. In this situation, the flame responds to both perturbations of acoustic velocity and equivalence ratio. From a scientific point of view, an experimental investigation of the response of a partially premixed flame raises many challenging questions. First, the linear/nonlinear response of a partially premixed flame subjected to acoustic velocity and fuel/air ratio modulations is expected to be appreciably different, both quantitatively and qualitatively, as compared to
9
the response of a premixed flame. The combined effects of two inlet disturbances may induce partial extinction of the reaction zone at certain inlet flow conditions, particularly when the combustor is operated near lean blowoff limits. Also, a systematic parametric study on the response of a partially premixed flame should be performed to illustrate the role of inlet disturbances in controlling the flame’s response characteristics and to elucidate the controlling physics. This work is left for future studies. CONCLUSIONS Nonlinear features of acoustically excited swirl-stabilized leanpremixed flames are characterized by nonlinear flame transfer function measurements and a phase-resolved analysis. Two distinct mechanisms of nonlinearity are investigated. The first mechanism is shear layer dynamics, which plays a key role in causing the nonlinearity of inverted dihedral V flames when the modulation amplitude is less than 40% at a modulation frequency of 200 Hz. H2-enriched enveloped M flames are found to be more stable in terms of combustion instability. Their interaction with a large-scale coherent structure is negligible, which leads to a linear response even at high forcing frequency and amplitude. This implies that steady-state flame configuration is one of the important controlling parameters determining linear and nonlinear dynamics. The second mechanism of nonlinearity is unsteady flame liftoff which causes rapid reduction in the flame surface area due to a merging of two flame branches. This mechanism is manifested when the forcing amplitude is greater than 40%. A systematic investigation of forced flame response measurements with special emphasis on the effects of fuel composition variation will be useful to develop and validate theoretical tools to predict combustion instability phenomena in fuel-flexible gas turbine engines. ACKNOWLEDGMENTS Funding for this research was provided by the Department of Energy University Coal Research Program through Contract # DE-FG26-07NT43069 and the National Science Foundation through Award #0625970. REFERENCES [1] Lieuwen, T. and Yang, V., 2005, Combustion Instabilities in Gas Turbine Engines. Progress in Astronautics and Aeronautics, 210, AIAA, Washington, DC. [2] Balachandran, R., Ayoola, B.O., Kaminski, C.F., Dowling, A.P., and Mastorakos, E., 2005, Experimental investigation of the nonlinear response of turbulent premixed flames to imposed inlet velocity oscillations. Combust. Flame, 143, 37−55. [3] Bellows, B.D., Neumeier, Y., and Lieuwen, T., 2006, Forced response of a swirling, premixed flame to flow disturbances. J. Propul. Power, 22, 1075−1084. [4] Kulsheimer, C. and Buchner, H., 2002, Combustion
dynamics of turbulent swirling flames. Combust. Flame, 131, 70−84. [5] Armitage, C.A., Balachandran, R., Mastorakos, E., Cant, R.S., 2006, Investigation of the nonlinear response of turbulent premixed flames to imposed inlet velocity oscillations. Combust. Flame, 146, 419−436. [6] You, D., Huang, Y., and Yang, Y., 2005, A generalized model of acoustic response of turbulent premixed flame and its application to gas-turbine combustion instability analysis. Combust. Sci. Tech., 177, 1109−1150. [7] Preetham, S.H. and Lieuwen, T., 2007, Response of turbulent premixed flames to harmonic acoustic forcing. Proc. Combust. Inst., 31, 1427−1434. [8] Fleifil, M., Annaswamy, A.M., Ghoneim, Z.A., and Ghoniem, A.F., 1996, Response of a laminar premixed flame to flow oscillations: a kinematic model and thermoacoustic instability results. Combust. Flame, 106, 487−510. [9] Roux, A., Gicquel, L.Y.M., Sommerer, Y., and Poinsot, T.J., 2008, Large eddy simulation of mean and oscillating flow in a side-dump ramjet combustor. Combust. Flame, 152, 154−176. [10] Sengissen, A.X., Van Kampen, J.F., Huls, R.A., Stoffels, G.G.M., Kok, J.B.W., and Poinsot, T.J., 2007, LES and experimental studies of cold and reacting flow in a swirled partially premixed burner with and without fuel modulation. Combust. Flame, 150, 40−53. [11] Staffelbach, G., Gicquel, L.Y.M., Boudier, G., and Poinsot, T., 2009, Large eddy simulation of self excited azimuthal modes in annular combustors. Proc. Combust. Inst., 32, 2909−2916. [12] Huang, Y. and Yang, V., 2009, Dynamics and stability of lean-premixed swirl-stabilized combustion. Prog. Energy Combust. Sci., 35, 293−364. [13] Kim, K.T., Lee, J.G., Quay, B.D., and Santavicca, D.A., 2010, Response of partially premixed flames to acoustic velocity and equivalence ratio perturbations. Combust. Flame, 157, 1731−1744. [14] Kim, K.T., Lee. J.G., Quay, B.D., and Santavicca, D.A., 2010, Spatially distributed flame transfer functions for predicting combustion dynamics in lean premixed gas turbine combustors. Combust. Flame, 157, 1718−1730. [15] Lieuwen, T., McDonell, V., Santavicca, D., and Sattelmayer, T., 2008, Burner development and operability issues associated with steady flowing syngas fired combustors. Combust. Sci. Technol., 180, 1167−1190. [16] Richards, G.A., McMillian, M.M., Gemmen, R.S., Rogers, W.A., and Cully, S.R., 2001, Issues for low-emission, fuelflexible power systems. Prog. Energy Combust. Sci., 27, 141−169. [17] Strakey, P., Sidwell, T., and Ontko, J., 2007, Investigation of the effects of hydrogen addition on lean extinction in a swirl stabilized combustor. Proc. Combust. Instit. 31, 3173−3180. [18] Schefer, R.W., 2003, Hydrogen enrichment for improved
10
lean flame stability. Int. J. Hydrogen Energ., 28, 1131−1141. [19] Jackson, G.S., Sai, R., Plaia, J.M., Boggs, C.M., and Kiger, K.T., 2003, Influence of H2 on the response of lean premixed CH4 flames to high strained flows. Combust. Flame, 132, 503−511. [20] Cozzi, F. and Coghe, A., 2006, Behavior of hydrogenenriched non-premixed swirled natural gas flames. Int. J. Hydrogen Energ., 31, 669−677. [21] Kido, H., Nakahara, M., Nakashima, K., and Hashimoto, J., 2002, Influence of local flame displacement velocity on turbulent burning velocity. Proc. Combust. Instit., 29, 1855−1861. [22] Sarli, V. and Benedetto, A., 2007, Laminar burning velocity of hydrogen-methane/air premixed flames. Int. J. Hydrogen Energ., 32, 637−646. [23] Mandilas, C., Ormsby, M.P., Sheppard, C.G.W., and Wooley, R., 2007, Effects of hydrogen addition on laminar and turbulent premixed methane and iso-octane-air flames. Proc. Combust. Instit., 31, 1443−1450. [24] Schefer, R.W., Wicksall, D.M., and Agrawal, A.K., 2002, Combustion of hydrogen-enriched methane in a lean premixed swirl-stabilized burner. Proc. Combust. Instit., 29, 843−851. [25] Wicksall, D.M., Agrawal, A.K., Schefer, R.W., and Keller, J.O., 2005, The interaction of flame and flow field in a lean premixed swirl-stabilized combustor operated on H2/CH4/air. Proc. Combust. Instit., 30, 2875−2883. [26] Wicksall, D.M. and Agrawal, A.K., 2007, Acoustics measurements in a lean premixed combustor operated on hydrogen/hydrocarbon fuel mixtures. Int. J. Hydrogen Energ., 32, 1103−1112.
[27] Kim, K.T., Lee, J.G., Lee, H.J., Quay, B.D., and Santavicca, D.A., 2010, Characterization of forced flame response of swirl-stabilized turbulent lean-premixed flames in a gas turbine combustor. J. Eng. Gas Turb. Power, 132, 04. [28] Durox, D., Schuller, T., Noiray, N. and Candel, S., 2009, Experimental analysis of nonlinear flame transfer functions for different flame geometries. Proc. Combust. Inst., 32, 1391−1398. [29] Thumuluru, S.K. and Lieuwen, T., 2009, Characterization of acoustically forced swirl flame dynamics. Proc. Combust. Inst., 32, 2893−2900. [30] Schuller, T., Durox, D., and Candel, S., 2003, Self-induced combustion oscillations of laminar premixed flames stabilized on annular burners. Combust. Flame, 135, 525−537. [31] Schuller, T., Durox, D., and Candel, S., 2003, A unified model for the prediction of laminar flame transfer functions: comparisons between conical and V-flame dynamics. Combust. Flame, 134, 21−34. [32] Cala, C.E.C, Fernandes, E.C., and Heitor, M.V., 2002, Analysis of oscillating shear layer. 11th Symposium on Application of Laser Techniques and Fluid Mechanics, Lisbon, Portugal. [33] Lieuwen, T., 2005, Nonlinear kinematic response of premixed flames to harmonic velocity disturbances. Proc. Combust. Inst., 30, 1725−1732. [34] Kim, K.T., Lee, H.J., Lee, J.G., Quay, B., and Santavicca, D., 2009, Flame transfer function measurement and instability frequency prediction using a thermoacoustic model. ASME paper GT2009−60026.
11
Manuscript Not AVAILABLE
SUITABILITY OF A SOUTH-AFRICAN HIGH ASH CONTENT AND HIGH ASH FLOW TEMPERATURE COAL SOURCE FOR ENTRAINED FLOW GASIFICATION JC van Dyk* and R Stemmer** * Sasol Technology, R&D, Sasolburg, South-Africa, [email protected], +27169604375 (tel), +27115224806 (fax) ** Corus Technology, RD&D, Ijmuiden, The Netherlands, [email protected], +31251492073 (tel), +31251470489 (fax) In slagging gasifiers the ash flows down the gasifier walls and drains from the gasifier as molten slag. Coals selected for slagging gasifiers should thus have an ash flow temperature (AFT) below the operating temperature of the gasifier, and in practice can be lowered by the addition of a flux, such as limestone. As the mineral matter start to melt and become a liquid (between the softening temperature and the flow temperature of the mineral matter in coal), it will have a specific composition and a related viscosity (ease of flow). During entrained flow gasification the viscosity has to be low enough for the slag to flow and drain from the gasifier. As a consequence, the viscosity of the slag, which depends on the slag composition, is one of the most critical factors in the operation of slagging gasifiers. The pilot gasifier at Corus in The Netherlands, as used to produce reduction gas for the steel industry, was used for the test. The objective of the experiment was to investigate the gasifiability of high ash coal in an oxygen blown, atmospheric pressure entrained flow pilot gasifier, and to quantify / qualify where possible and within the known limits of such a pilot test unit, the slag, fly-ash, water and gas characteristics. The coal was fed pneumatically with N2 at ±50kg/hr through the oxygen coal burner. The outlet velocity of the burner was set at 110m.s-1 (oxygen flow rate of 20Nm3hr-1). The header temperature was controlled at maximum 1600oC, resulting in a temperature at the bottom part of the gasifier between 1200 and 1300oC. Due to this rapid heat loss over the gasifier length, the coal was over-fluxed in order to maintain a running slag and avoid blockages at the bottom temperature of 1200oC, which is almost 150oC lower than normal slagging operating conditions. The refractory hot face showed limited wear or penetration of minerals into the refractory. The slag viscosity was thus fluxed adequately for the specific setup to flow gently over the refractory and protecting the SiC. The measured and equilibrium simulated carbon balances corresponded well and could be closed with a high degree of accuracy. During stable and high load periods, the CO content varied between 50 and 60 vol%, H2 between 10 and 20 vol% and CO2 <10%. Carbon conversion varied between 79% and 96%. The calculated residence times were 1.9 to 2.2 seconds. The testing facility at Corus Ijmuiden was able to maintain conditions for high-ash coal gasification at relatively long periods of stable operation and high production rates. High coal conversions were achieved and high ash flow temperature coal, blended with
limestone as fluxing agent, can be regarded as successful within the limits of the test unit. IP Approval 1 February 2010, PP001004
Continuous Experiments of Hot Gas Desulfurization Process Using ZnBased Solid Sorbents in a Pressurized Condition Sung-Ho Jo1†, Young Cheol Park1, Ho-Jung Ryu1, Chang-Keun Yi1, and Jeom-In Baek2 1
2
Korea Institute of Energy Research, Daejeon 305-343, Korea Korea Electric Power Research Institute, Daejeon 305-380, Korea ABSTRACT
The high temperature desulfurization technique is one of the elemental technologies of syngas purification having both higher thermal efficiency and lower emissions compared with conventional wet cleanup processes. The high temperature desulfurization is a novel method to remove H2S and COS efficiently in the syngas with regenerable solid sorbents at high temperature and high pressure condition which are preferable to gasification condition. In this study, we developed 5 Nm3/h of hot gas desulfurization process composed of a riser type desulfurizer, a fluidized-bed type regenerator, and a loopseal. We reported several hydrodynamics testing results of the proposed process in the previous Pittsburgh Coal Conference. In order to investigate the desulfurization performance, we used Zn-based solid sorbents which were provided by Korea Electric Power Research Institute (KEPRI). The solid sorbents was manufactured by a spray-drying method, so the shape of the sorbents is spherical, which is adequate for the fluidization. We performed high-temperature and high-pressure tests to check feasibility of stable operation with solid circulation and to investigate the desulfurization performance using simulated syngas with 2,000 ppm of H2S inlet concentration. We are now developing a bench-scale hot gas desulfurization system which handles 100 Nm3/h of syngas treatment at 20 bar of operating pressure and this system will be integrated with a coal gasifier which was installed in the Institute for Advanced Engineering (IAE). Key Words: Desulfurization, Pressurized condition, Zn-based solid sorbents, Continuous experiments, Fluidized process
INTRODUCTION The Integrated Coal Gasification Combined Cycle (IGCC) plants have currently used cold-gas cleanup technologies such as Selexol™ and Rectisol® in order to get rid of sulfur compounds in the syngas. However, the existing conventional cold-gas cleanup technologies require cooling and reheating of the gas stream, which results in significant reduction in thermal efficiency of the system. The hot gas desulfurization (HGD) technique is one of the elemental technologies having both high thermal efficiency and very low emissions [1]. The HGD is a new method to efficiently remove H2S and COS in the syngas with regenerable sorbent at high temperature and high pressure condition [2]. Research Triangle Institute (RTI) International and Eastman Chemical Company had reported successful demonstration of a transport reactor-based desulfurization technology with regenerable sorbent suitable for warm (250oC to 600oC) syngas [3]. In this study, we used transport reactor-based desulfurizer and bubbling fluidized-bed regenerator. The proposed system was designed to operate up to 20 atm. We reported the solid circulation rate and voidage in the desulfurizer through the cold-mode tests in the previous Pittsburgh Coal Conference [4]. We tested the performance of Zn-based dry sorbents by continuous operation at high temperature and 10 atm gauge †
To whom correspondence should be addressed. E-mail: [email protected]
in order to check feasibility of stable operation.
EXPERIMENTAL Fig. 1 [4] shows a schematic representation of the HGD system. The typical feature of the process includes a fast fluidized-bed desulfurizer, a bubbling fluidized-bed regenerator, two cyclones, and a standpipe. Detail dimensions were illustrated at the reference [4].
Fig.1. Schematic representation of the HGD system: ① desulfurizer; ② regenerator; ③ cyclone; ④ cleaned syngas; ⑤ to gas analyzer (SO2, O2); ⑥ valve In order to check feasibility of stable operation, we continuously operated the HGD system using Znbased solid sorbents. The temperature of desulfurizer and regenerator was maintained at 550oC and 600oC, respectively. The operating pressure was around 10 atm gauge. The inlet H2S content was 2,000 ppmv.
RESULTS AND DISCUSSIONS Fig. 1 showed the 6-hr continuous operation results. In Fig. 1(a), the temperature of the desulfurizer and the regenerator was maintained at 550oC and 600oC, respectively during continuous operation. Fig. 1(b) showed the stable solid circulation between the desulfurizer and the regenerator. Fig. 1(c) showed the pressure of the desulfurizer and the regenerator. Both reactors had been maintained at 10.1 atm gauge. In Fig. 1(c), the pressure of both reactors was almost the same so that the solid was stably circulated between two reactors. Fig. 1(d) showed the inlet and outlet H2S contents. We supplied 2,000 ppmv of H2S (N2-balance) and the outlet H2S contents was measured by a gas analyzer. In Fig. 1(d), the outlet H2S content was maintained below 0.1 ppmv. More than 99.99% of supplied H2S was removed in the desulfurizer.
Temperature Profiles
Desulfurizer - Differential Pressure Profiles 200
800
o
Temperature [ C]
700
Differential Pressure [mmH2O]
Desul-Bottom Desul-Middle Desul-Top Regen
600
500
dP-Bottom dP-Middle1 dP-Middle2 dP-Top
150
100
50
0
400 0
1
2
3
4
5
0
6
1
2
5
6
H2S Content
12
2500
10
2000
H2S Content [ppmv]
Pressure [atm gauge]
4
(b)
(a) System Pressure
8 Regen. Pressure Desul. Pressure
6
3
Time [hr]
Time [hr]
4
1500 H2S-Inlet H2S-Outlet
1000
500
0
2
0 0
1
2
3
4
5
6
0
1
2
3
4
5
6
Time [hr]
Time [hr]
(d)
(c) Fig. 1. 6-hr continuous operation results.
CONCLUSIONS We used transport reactor-based desulfurizer and bubbling fluidized-bed regenerator. Through the 6-hr continuous operation, we acquired a stable operation with solid circulation between two reactors in a high pressure of 10 atm gauge. The H2S removal was more than 99.99%. We have developed a bench-scale hot gas desulfurization system which handles 100 Nm3/h of syngas treatment at 20 bar of operating pressure and this system will be integrated with a coal gasifier which was installed in the Institute for Advanced Engineering (IAE).
ACKNOWLEDGEMENT This research was supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Ministry of Knowledge Economy of Korea.
REFERENCES [1] Choi, Y.C., Park, T.J., Kim, J.H, Lee, J.G., Hong, J.C., and Kim, Y.G., Korean J. Chem. Eng., 18:493 (2001). [2] Yi, C.K., Jo, S.H., Lee B.H., Lee S.Y., Son, J.E., and Jin G.T., Korean J. Chem. Eng., 18:1005 (2001). [3] Gupta, R., Turk and B., Lesemann M., 2009 Gasification Technologies Conference, USA (2009). [4] Jo, S.H., Park, Y.C., Ryu, H.J., Baek, J.I., and Yi, C.K., 26th Pittsburgh Coal Conference, USA (2009).
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
BIOMASS GASIFICATION IN DUAL FLUIDIZED REACTORS: PROCESS MODELING APPROACH
Thanh D. B. Nguyen Lab. FACS, RCCT, Department of Chemical Engineering, Hankyong National University Kyonggi-do, Anseong-si, Jungangno 167, 456-749 Korea. Tel.: +82 10 2358 0501; E-mail address: [email protected] Young-Il Lim (corresponding author) Lab. FACS, RCCT, Department of Chemical Engineering, Hankyong National University Kyonggi-do, Anseong-si, Jungangno 167, 456-749 Korea. Tel.: +82 31 670 5207 (office); Mobile phone: +82 10 7665 5207; Fax: +82 31 670 5445; E-mail address:[email protected] Byung-Ho Song Department of Chemical Engineering, Kunsan National University Gunsan, Jeonbuk, 573-701 Korea. E-mail address: [email protected] Won Yang, Uendo Lee, and Young-Tai Choi (presenter) Manufacturing System Division, Korean Institute of Industrial Technology Cheonan, Korea. Jae-Hun Song, Gi-Chul Myoung, and Yong-Soo Cho SeenTec Co., Ltd. No. 608, Office Plaza, 7-2 Yongho-dong, Changwon, Gyeongnam, Korea.
Abstract A process model including chemical reactions and mass/heat balances is investigated for biomass steam gasification in a dual fluidized bed (DFB) system. Gasification in the DFB gasifier is calculated using a two-stage thermodynamic equilibrium model in which the carbon conversion of biomass is predicted using solid-gas reactions, while the composition of syngas is calculated in gas-phase reactions. The drying and devolatilization of biomass are considered to be completed within the gasifier. Complete combustion of unconverted char and additional fuel is assumed in the combustion chamber (riser). Heat required for the endothermic gasification reactions is provided by the circulating bed material (silica sand). The model proposed in this paper is first validated using experimental data taken from published works. Then, the effects of gasification temperature and steam to fuel ratio on the carbon conversion of biomass, the composition of product gas and the circulation rate of bed material are examined.
1. Introduction In this study, a process model including a two-stage equilibrium model [1] incorporated with the pyrolysis stage is developed to predict the composition of the producer gas in dual circulating fluidized bed (DFB) biomass gasifier. The principle of the DFB gasification process is illustrated in Fig. 1. The pyrolysis gas is calculated from an empirical model which is established on the basis of the experimental data taken from the literature [2]. In order to calculate the composition of the producer gas, the extent of species reactions [3] is applied to the char-gas (first stage) and gas-phase (second stage) reactions of the two-stage equilibrium model. 1
Raw syngas:
Flue gas:
Heated silica sand
CO, H2, CH4
O2, N2, CO2, H2O
CO2, H2O
Biomass
Gasifier
Riser
Gasification (~850oC, 1atm)
Combustion (~950oC, 1atm)
Additional Fuel (Biomass)
Cooled silica sand Steam
Char and Ash
Air
Fig. 1. Principle of the DFB biomass gasification process. The proposed model is first validated with experimental data from the literature [4]. Then, the process performance in terms of carbon conversion, heat demand, solid circulation ratio, additional fuel ratio, and heat recovery is evaluated from the simulation results of the model.
2. Model description The process model for biomass gasification proposed in the present work is briefly introduced below where the three main steps (pyrolysis, char-gas reactions, and gas-phase reactions) are taken into account. The model includes four empirical sub-models. The two empirical equations for the pyrolysis are derived on the basis of experimental data conducted at a wide range of temperature (400-900 oC) using three kinds of biomass (wood, coconut shell and straw) [2]. The rest of the model for the char gasification is modified from the two-stage stoichiometric equilibrium model reported in the previous works [1,5].
2.1. Pyrolysis model The pyrolysis is considered to be completed within the gasifier with the assumption that the composition of the pyrolysis gas consist of CO, CO2, CH4, H2 and H2O [6]. The formations of CO2 and H2 from CO and CH4 respectively during the pyrolysis are expressed as the empirical models called CO2 and H2 formations (CO and CH) which are derived from experimental data [2]. The models have the form of an exponential function of the pyrolysis temperature,
CO
nCO 4413.3 3.17 102 exp nCO2 T
CH
nCH 4 nH 2
(1)
4239 1.16 10 2 exp T
(2)
Atomic balances are used to determine the composition of the pyrolysis gas from the physical properties of biomass (proximate and ultimate analyses) and Eqs. (1-2)
2.2. Two-stage equilibrium model In the first-stage, the major three reactions (Boudouard reaction, primary char-steam reaction and secondary char-steam reaction) are assumed to take place and to reach the equilibrium state in this study: C(s) + CO2(g) 2CO(g)
(R1)
C(s) + H2O(g) CO(g) + H2(g)
(R2)
C(s) + 2H2O(g) CO2(g) + 2H2(g)
(R3)
2
The steam amount involved in the reactions (R2) and (R3) is estimated from Eq. (3) which was proposed in our previous work [1],
nH 2 O,0 nH 2 O, tot
7542.8 51.4 exp T
(3)
where nH 2 O , 0 is the moles of water involved in the solid-gas reactions in the second stage, and nH 2 O , tot is the total moles of water present in the system after the pyrolysis is completed. In the second-stage, it was reported that the water-gas shift reaction dominates in biomass gasification at the temperature higher than 670 oC (940 K) [4,7,8]: CO(g) + H2O(g) CO2(g) + H2(g)
(R4)
The methanation reaction C(s) + 2H2(g) CH4(g) in the first stage and the steam reforming reaction CH4(g) + H2O(g) CO(g) + 3H2(g) in the second stage of the equilibrium model are supposed not to take place, because the equilibrium constant of the methane formation reaction tends to zero and that of the steam reforming reaction tends to infinity at the elevated temperatures [9]. Hence, only water-gas shift reaction (R4) is assumed to take place in the second-stage. Experiments
conducted
on
a
pilot-scale
biomass
fluidized
bed
gasifier
show
that
the
values
of
( yCO2 yH 2 ) /( yCO yH 2 O ) (called experimental equilibrium constant) determined from the experimental data are far from those of the theoretical equilibrium constant in the range of 650-850 oC [7,10]. Thus, in this study a ratio called non-equilibrium factor () accounting for the deviation from the equilibrium state of the water-gas shift reaction in biomass gasification is estimated. This factor is express as function of gasification temperature,
9 1012 exp(2.22 102 T )
(4)
The composition of the producer gas is calculated from the equilibrium constants of the reactions in the two-stage model using the reaction coordinate approach: yi
ni , 0 j i , j j
(5)
n0 j j j
where ni,0 and n0 are the moles of species Ai and total moles of all species in the system at the initial state, respectively.
j is the total stoichiometric number of the jth reaction which is defined as: GT0, j K j exp RT
(6)
where GT0 is calculated from the standard Gibbs energy of formation G 0f ( H 0f TS 0f ) , standard enthalpy of formation H 0f and molar heat capacity cp of each component in the related reaction.
3. Model validation
The process model with the four empirical equations proposed in this study is validated with experimental data taken from Wei et al. [4] where the steam gasification of Legume straw was conducted at the temperature varying from 750 to 850 oC and the steam to fuel ratio ranging from 0.0 to 1.0. Simulation results of the producer gas composition at various gasification temperatures and steam to fuel ratios are presented in Fig. 2 in comparison with experimental data. The numerical results show a similar trend with experimental data. The model predicts relatively well the compositions of H2 and CH4 at a wide range of gasification temperatures and steam to fuel ratios.
3
60
60 Steam to fuel ratio = 0.6 kg/kg
Gasification temperature = 800oC 50 Producer gas composition (vol.%)
Producer gas composition (vol.%)
50 CO-model CO-exp. CO -model
40
2
CO -exp. 2
30
CH -model 4
CH -exp. 4
H -model
20
2
H -exp. 2
10
CO-model CO-exp. CO -model
40
2
CO -exp. 2
30
CH -model 4
CH -exp. 4
H -model
20
2
H -exp. 2
10
0 740
760
780 800 820 Gasification temperature (oC)
840
0
860
0
0.2
0.4 0.6 Steam to fuel ratio (kg/kg)
0.8
1
Fig. 2. Producer gas composition vs. gasification temperature and steam to fuel ratio compared to exp. data.
4. Parametric study
The present model is used to investigate the effects of operating parameters (temperature and steam to fuel ratio) on the biomass gasifier performance in terms of the gas composition, carbon conversion, solid circulation ratio, heat demand and heat recovery. Hereafter, the effects of the gasification temperature and steam to fuel ratio on the gasification performance are examined for a DFB gasifier using pine wood chips as a feedstock.
4.1 Effect of gasification temperature Fig. 3 shows the carbon conversion (C), char residue (C,res), the gas yield (Vyield) and the producer gas composition of the wood chip gasification within 750 Tgasifier 850 oC at = 1.3. Carbon conversion increases with an increase in gasification temperature due to the endothermic reactions of the char-steam and char-CO2 (see Fig. 3a). Since the gasification reactions proceed forward with increasing temperature, an increase in gasification temperature results in the increase in the gas yield. Contrastingly with the variation of carbon conversion, char residue decreases with an increase in gasification temperature. In the studied range of gasification temperature, the H2 and CO2 formations increase, while CO and CH4 decrease, as shown in Fig. 3b. The trends agree well with the results reported by other researchers in a fluidized bed and a DFB [1112]. 1.5
(a)
0.9
1.4
0.8
1.3
0.7
1.2
0.6
1.1
0.5 740
760
780 800 820 Gasification temperature (o C)
(b)
50 Producer gas composition (vol.%)
Carbon conversion Char residue Gas yield
840
Gas yield (Nm3/kg)
Carbon conversion, Char residue (-)
1
CO
2
CH
4
H
2
40
30
20
10
0 740
1 860
CO
760
780 800 820 Gasification temperature (o C)
840
860
Fig. 3. Effect of carbon conversion, char residue, gas yield, producer gas composition on gasification temperature.
As the gasification temperature increases, the heat demand increases. That leads to an increase in the circulation rate of the bed material to meet the needs of the elevated heat. Thus, as shown in Fig. 4a, the solid circulation ratio (rcir) increases with an increase in the gasification temperature. However, additional fuel is not needed below 830 oC. The heat recovery (e) and the heat value of the producer gas (LHVproduct) are considered as the criteria to evaluate the energy efficiency of a gasification system. Fig. 4b shows the effect of the gasification temperature on the heat recovery and on the heat value of the producer gas from the wood chips. The heat value of the producer gas decreases with the 4
rise of temperature, because the un-combustible component CO2 increases at elevated temperature. The heat recovery, however, increases with the temperature in the gasifier due to the endothermic gasification reactions and to the high increasing rate of the syngas yield. 50
0.015
0.9
14
0.0125
40
0.01
35
0.0075
30
0.005
25
0.0025
20 740
760
780 800 820 Gasification temperature (o C)
e (-)
45
0.875
13
0.85
12
0.825
11
0 860
840
0.8 740
760
780 800 820 Gasification temperature (o C)
LHV of producer gas (MJ/Nm3)
(b) Additional fuel ratio (kg/kg)
Circulation ratio (kg/kg)
(a)
10 860
840
Fig. 4. Effect of circulation ratio, additional fuel ratio, heat recovery, LHVproduct on gasification temperature.
4.2 Effects of steam to fuel ratio Carbon conversion (C), char residue (C,res), syngas yield (Vyield) and producer gas composition are presented in Fig. 5 as a function of the steam to fuel ratio (). As an increase in the steam amount, the reaction of char-steam advances forward to produce CO and H2, which results in an increase in the carbon consumed and a decrease in the char residue (see Fig. 5a). As a result, the syngas yield increases with the rise of steam to fuel ratio due to the increase of carbon conversion. An increase in steam to fuel ratio leads to the increase of H2 and CO2 concentrations, while CO and CH4 decrease (see Fig. 5b). This behavior is attributed to the fact that the forward water-gas shift reaction rate increases, as the steam to fuel ratio increases. (a) 1.75
0.8
1.5
0.7
1.25
0.6
1
0.5
0.75
0.4
0
0.2
0.4
0.6 0.8 1 Steam to fuel ratio (kg/kg)
(b)
50 Producer gas composition (vol.%)
0.9
2 Carbon conversion Char residue Gas yield
1.2
1.4
Gas yield (Nm3/kg)
Carbon conversion (-), Char residue (-)
1
CO
2
CH
4
H
2
40
30
20
10
0
0.5 1.6
CO
0
0.2
0.4
0.6 0.8 1 Steam to fuel ratio (kg/kg)
1.2
1.4
1.6
Fig. 5. Effect of carbon conversion, char residue, gas yield, producer gas composition on steam to fuel ratio.
Fig. 6a shows the variations of the solid circulation ratio (rcir) and the additional fuel ratio (radd) required to provide the heat to the gasifier. Since the heat demand in the gasifier increases with increasing the steam to fuel ratio, an increase in the steam to fuel ratio results in the increase of the solid circulation rate in order to transport more heat from the riser to the gasifier. However, since the heat is released from the forward water-gas shift reaction due to the increase of the water amount, the circulation ratio shows a minimum around = 0.2. No additional fuel is required for the steam to fuel ratio up to 1.1 as observed in Fig. 6a. As shown in Fig. 6b, the heat recovery (e) increases with increasing the steam to fuel ratio due to the increase of the syngas yield. The heat value of the producer gas (LHVproduct) decreases with increasing the steam to fuel ratio because the heat value of CO is higher than that of H2 [11].
5
47
0.04
0.9
16
0.03
45
0.02
44
0.01
43
0
0.2
0.4
0.6 0.8 1 Steam to fuel ratio (kg/kg)
1.2
1.4
e (-)
46
0.875
15
0.85
14
0.825
13
0.8
12
0.775
11
0 1.6
0.75
0
0.2
0.4
0.6 0.8 1 Steam to fuel ratio (kg/kg)
1.2
1.4
LHV of producer gas (MJ/Nm3)
(b) Additional fuel ratio (kg/kg)
Circulation ratio (kg/kg)
(a)
10 1.6
Fig. 6. Effect of circulation ratio, additional fuel ratio, heat recovery, LHVproduct on steam to fuel ratio. 5. Conclusions
A process model including pyrolysis and gasification reactions is developed to predict the producer gas composition, carbon conversion, heat demand and heat recovery in a DFB gasifier. By considering the important role of biomass pyrolysis, the present model shows better results than the previous equilibrium models for biomass gasification. The novel model using the reaction coordinate can give in detail the degree of each gasification reaction. The simulation results of the model show not only good agreement but also similar trend with experimental data at various gasification temperatures and steam to fuel ratios. The effects of gasification temperature and steam to fuel ratio on the gasification performance are also examined from the simulation results.
Acknowledgements
This work is supported by Bilateral International Collaborative R&D program under the Ministry of Knowledge Economy, Korea.
Nomenclature
cp
molar heat capacity (kJ/kmol-K)
GT0
Gibbs free energy change of reaction (kJ/kmol)
G 0f
standard Gibbs energy of formation (kJ/kmol)
h
molar enthalpy (kJ/kmol)
H 0f
standard enthalpy of formation (kJ/kmol)
K
equilibrium constant (-)
LHVproduct
lower heat value of producer gas (MJ/Nm3)
n
number of moles (kmol)
radd
additional fuel ratio (kg/kg)
rcir
solid circulation ratio (kg/kg)
R
gas constant (kJ/kmol-K)
Vyield
gas yield (Nm3/kg biomass)
Tgasifier
gasification temperature (K)
x
mass fraction (-)
y
mole fraction (-)
Greek letters
6
steam to fuel ratio (kg/kg)
reaction coordinate (kmol)
non-equilibrium factor (-)
e
heat recovery (-)
fraction of formation (-)
C
carbon conversion (-)
C,res
char residue (-)
stoichiometric number (-)
Subscripts i
species index
j
reaction index
C
carbon
H
hydrogen
N
nitrogen
O
oxygen
CO
carbon monoxide
CO2
carbon dioxide
CH4
methane
H2
hydrogen
H2O
water
References
[1] Nguyen TDB, Lim, YI., Song BH, Kim, SM. and Joo YJ, Ahn, DH. Two-stage equilibrium model applicable to the wide range of operating conditions in entrained-flow coal gasifiers. Fuel 2010;in press. [2] Fagbemi L, Khezami L and Capart R. Pyrolysis products from different biomasses: application to the thermal cracking of tar. Applied Energy 2001;69:293-306. [3] Smith JM, Van Ness HC and Abbott MM. Introduction to chemical engineering thermodynamics. 7th ed. New York: Mc Graw-Hill; 2005. [4] Wei L, Xu S, Zhang L, Liu C, Zhu H and Liu S. Steam gasification of biomass for hydrogen-rich gas in a free-fall reactor. International Journal of Hydrogen Energy 2007;32:24-31. [5] Yoshida H, Kiyono F, Tajima H, Yamasaki A, Ogasawara K and Masuyama T. Two-stage equilibrium model for a coal gasifier to predict the accurate carbon conversion in hydrogen production. Fuel 2008;87:2186-2193. [6] Sadaka SS, Ghaly AE and Sabbah MA. Two phase biomass air-steam gasification model for fluidized bed reactors: Part I-model development. Biomass and Bioenergy 2002;22:439-462. [7] Herguido J, Corella J and Gonzalez-Saiz J. Steam gasification of lignocellulosic residues in a fluidized bed at a small pilot scale. Effect of the type of feedstock. Industrial & Engineering Chemistry Research 1992;31:1274-1282. [8] Walawender WP, Hoveland DA and Fan LT. Steam gasification of pure cellulose. 1. Uniform temperature profile. Industrial & Engineering Chemistry Process Design and Development 1985;24:813-817. [9] Sharma AK. Equilibrium modeling of global reduction reactions for a downdraft (biomass) gasifier. Energy Conversion and Management 2008;49:832-842.
7
[10] Rapagnà S, Jand N, Kiennemann A and Foscolo PU. Steam-gasification of biomass in a fluidised-bed of olivine particles. Biomass and Bioenergy 2000;19:187-197. [11] Lv PM, Xiong ZH, Chang J, Wu CZ, Chen Y and Zhu JX. An experimental study on biomass air-steam gasification in a fluidized bed. Bioresource Technology 2004;95:95-101. [12] Pfeifer C, Puchner B and Hofbauer H. Comparison of dual fluidized bed steam gasification of biomass with and without selective transport of CO2. Chemical Engineering Science 2009;64:5073-5083.
8
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: COAL SCIENCE EXACT TITLE OF PAPER: The role of O2/COG ratio on non-catalytic partial oxidation process of coke oven gas Presenting Author (Haizhu Cheng, Dr,Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology 030024, Taiyuan, China ,Tel: +86-13099071054,E-mail:[email protected] ) (Sufang Song,Dr, Institute of Chemical and Biological Technology, Taiyuan University of Science and Technology, 030021, Taiyuan, China ,E-mail:[email protected]) (Yongfa Zhang, Dr,Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology 030024, Taiyuan, China, E-mail:[email protected]) Co-Authors (Haizhu Cheng,Dr,Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology 030024, Taiyuan, China ,Tel: +86-13099071054,E-mail:[email protected]) Abstract: Coke oven gas(COG)was the main high-quality hydrocarbon resource , and its major component were H2 57~60%, CH4 25~28% and CO ~ 6% by volume. It was an important way for the coke oven gas to converse into synthetic gas, and then for the production of methanol and other chemical products. In this study the influence of O2/COG ratio on the content of synthesis gas was researched based on the small-scale single-hole nozzle experiment. The main equipments are single-hole nozzle and quartz tube reactor of which the size was 700 × 70mm. The contents of O 2, N2, CH4, CO, CO2, C2H4 and C2H6 in the synthesis gas were determined by chromatography GC-950 and the content of H2 by chromatography GC9890A. The experimental precision was determined by Content Subtraction of H2. The experimental data was analyzed based on Excel. The reaction process and the temperature distribution in the reactor were researched and the results showed that: methane
conversion increased with the increasing O2/COG ratio. CH4 conversion rate reached 95%~97% when the O2/COG ratio increased from 0.22 to 0.26, and the content of CH 4 in the synthesis gas was ≤ 1%. When the O2/COG ratio varied from 0.18 to 0.40, the H 2/CO ratio changed from 2.0 to 2.8, H2 65% to 58%, CO 24% to 28%, CO2 5% to 7.5% , O2 fluctuated in 0.15%~1.0%, N2 fluctuated in 2.5% ~7%, C2H4≦1.7% and C2H6 not detected. The temperature at the bottom of the reactor increased with increasing O2/COG ratio from 750~850℃ to 950~1050℃, the central temperature slightly higher from 600~680℃ to 650~750℃, the temperature at the top dropped from 450~ 550 ℃to 420~ 500 ℃. Key words: COG, O2/COG ratio, synthetic gas, methane conversion rate
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 manuscript Submission
PROGRAM TOPIC: COAL SCIENCE EXACT TITLE OF PAPER: Numerical simulation of Carbon catalytic reforming reactor Presenting Author (Haizhu Cheng, Dr,Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology 030024, Taiyuan, China ,Tel: +86-13099071054,E-mail:[email protected] ) (Yongfa Zhang, Dr,Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology 030024, Taiyuan, China, E-mail:[email protected]) (Sufang Song,Dr, Institute of Chemical and Biological Technology, Taiyuan University of Science and Technology, 030021, Taiyuan, China ,E-mail:[email protected]) Co-Authors (Haizhu Cheng,Dr,Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology 030024, Taiyuan, China ,Tel:+86-13099071054, E-mail:[email protected])
Numerical simulation of Carbon catalytic reforming reactor Haizhu Cheng*, Yongfa Zhang, Sufang Song (Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology 030024, Taiyuan, China ) Co-author (Haizhu Cheng, e-mail: [email protected])
Abstract: Coke oven gas (COG) was the main high-quality hydrocarbon resource , and its major component were H2 57~60%, CH4 25~28% and CO ~ 6% by volume. It was an important way for the coke oven gas to converse into synthetic gas, and then for the production of methanol and other chemical products. Based on the experiment of a small-scale single-hole nozzle reactor, coke oven gas carbon catalytic reforming reactor have been designed. CH4 and CO2 reforming catalyst is traditionally a precious metal catalyst. Because of carbon deposition the catalyst was easily inactivation. The prior study showed that carbon is a catalyst in CH 4 and CO2 reforming. Therefore, for using carbon catalyst to prevent effectively carbon deposition in the test reactor, it has provided the experimental foundation for the continuous industrial production. In this paper the Computational Fluid Dynamics (CFD) software was used to simulate reactor for the analysis of temperature distribution, flow field, and synthesis gas composition. Simulation results show that the reforming reactor demonstrated good flow field and temperature distribution. The H2/CO ratio was about 1.8-2.8 in the synthesis gas, which is the raw material for methanol synthesis. Key Word: carbon catalyst, numerical simulation, reforming, synthetic gas
1
Introduction
Methane (CH4) in COG reforming is the process of using carbon dioxide (CO2) to transform CH4 into synthesis gas, which can be used to produce liquid fuel, ammonia, methanol, etc. It is an important method in methane utilization and is referred to as the gas technology of liquefaction (GTL) [1]. Conversion of CH4 and CO2 into synthesis gas can effectively reduce greenhouse gas emissions and meet the demands of many synthesis processes in the chemical industry. High contents of CH4 in coke oven gas (COG) and high contents of CO2 in gasified coal gas can produce synthesis gas through reforming, which have several potential applications, such as in dual gas-head multi-combined technology [2]. Many types of catalyst (e.g., Ni-based, precious metal-based, and carbon catalysts) has been tested on CH4-CO2 reforming reactors with fixed and fluidized beds at a temperature range of 800-1200º C. Although a precious metal-based catalyst has better activation, selectivity [3,4], and a lower capacity for carbon deposition resistance than Ni-based and carbon catalysts, its industrial application is restricted due to higher cost [5,6,7]. Therefore, new and appropriate catalysts that have lower costs are still being investigated. One promising candidate is carbon. Research has shown that carbon can catalyze CH4-CO2 reforming [1,8]. Fucheng, et al. [9,10,11] have studied non-catalyzed partial oxidation in natural gas and discovered that there were three main zones in the oven, with the main reaction varying according to the zone. Moreover, the maximum amount of the synthesis gas output and the optimal proportion of oxygen (O2) and natural gas were subject to the model of the oven used. Currently, no simulation research has yet been reported for carbon as a catalyst in the CH4-CO2 reforming reaction. Based on a small experimental device, this study analyzed the impact of the ratio of oxygen flow to COG on the CH4 conversion rate and compositions of the main products. Simulation analyzed temperature distribution, flow field, and synthesis gas composition. Simulation results show that the reforming reactor demonstrated good flow field and temperature distribution. The H2/CO ratio was about 1.8-2.8 in the synthesis gas, which is the raw material for methanol synthesis.
2
Experimental study of small-scale reactor
2.1
Experimental devices
Figure 1 is coke oven gas catalytic partial oxidation into syngas of flow chart, The reactor main body is reactor and nozzle, etc.
F-Flow,T-Temperature,A-Analyzing,S-Sampling. 1-Reactor,2-Cooler,3-Reservoir for Product Gas,4-Cooling Water Inlet, 5-CO2 Inlet,6-Mixing Gas Inlet,7-O2 Inlet,8-Cooling Water Outlet Fig. 1. Conversion flow chart for the catalyzed partial oxidation in coke oven gas (COG)
2.2
Experimental process
The main equipment was a 700×70 MM single-hole nozzle quartz-tube reactor (Fig. 6). Pure oxygen supplied from the oxygen cylinder was injected into the reactor together with the COG for conversion and were regulated using a ZLB-15 gas glass rotor flow meter. The O2 , nitrogen (N2), CH4, carbon monoxide (CO), CO2, ethene (C2H4), and ethane (C2H6) contents of the synthesis gas was measured with a GC-950 chromatographic instrument (Partial pressure: 0.2Mpa; column temperature: 40 ℃; Bridge stream: 60mA; 13x molecular sieve * 1m), while the hydrogen (H2) content was determined using a GC9890A chromatographic instrument (Partial pressure: 0.4Mpa; column temperature: 80 ℃; Bridge stream: 60mA; 13x molecular sieve * 1m). The data were evaluated by a subtraction method. Three temperature measuring points were provided at 23 cm (bottom), 36 cm (middle), and 59 cm (top) from the nozzle. A platinum-rhodium thermocouple was used to measure the temperature in the reactor. Sampling was also conducted from the synthesis gas outlet, before it was delivered to the chromatograph for analysis. The reactor was water-cooled after the experiments. The outer quartz tube was coated with a nano- adiabatic material (thermal conductivity = 0.025 W/m.K), and the carbon catalyst was placed in the middle of the quartz tube reactor (3 kg).The composition of the COG is shown in Table 1. Table 1. Composition of coke oven gas (COG) and low heat value (LHV).
Content%
2.3
CH4
H2
CO
CO2
N2
O2
CmHn
LHV MJ/Nm3
24~28
54~59
5.5~7
3~5
1~3
0.3~0.7
2~3
17
Analysis of reaction process
The key chemical reactions that occurred during the partial oxidation of the COG are represented by the following equations: H2+0.5O2=H2O △H0298K = -241.84 KJ MOL-1 (1) CH4+0.5O2=CO+2H2 △H0298K = -27.32 KJ MOL-1 (2) CH4+2O2=CO2+2H2O △H0298K= -802.60 KJ MOL-1 (3) CO+0.5O2=CO2 △H0298K= -282.96 KJ MOL-1 (4) CH4+H2O=CO+3H2 △H0298K=206.15 KJ MOL-1 (5) CH4+CO2=2CO+2H2 △H0298K=247.27 KJ MOL-1 (6) CO2+H2=CO+H2O △H0298K=115.98 KJ MOL-1 (7) CH4=C+2H2 △H0298K=71.56KJ MOL-1 (8) Reactions (1)-(7) essentially reflected gasification. Reactions (1)-(4) represented the combustion reaction of the COG, which were strong exothermic processes that comprised the primary set of reactions, which occurred rapidly, with the complete reaction finishing within several dozens of milliseconds. Reactions (5)-(7) represented the control procedure of the whole process. These were endothermic reactions collectively called the secondary set of reactions. During the catalyzed partial oxidation, the conversion temperature could be maintained within
1200-1300 K. The oxidation process could be finished within 1-2 s [12]. To ensure that the secondary reaction could be completed, the appropriate H/D ratio had to be provided for the hearth. Reaction (8) played a secondary role in the catalyzed partial oxidation reaction in the presence of COG. Prior research has shown that a higher carbon content in the hydrocarbon make it easier for the splitting-decomposition reaction to take place. When the carbon contents were equal, unsaturated hydrocarbons underwent splitting-decomposition and carbon-crystallization reactions faster than saturated hydrocarbons. Among the hydrocarbons, CH4 was found to have the best stability for this kind of reaction [12,13]. The conversion reactor can be divided according to the main reactions into an upper local area where strong oxidation occurs and a lower-and-middle area where deoxidization-conversion occurs. The upper local area with strong oxidation maintained an average temperature of 1150-1350ºC. Meanwhile, the partial combustion of H2, CH4, and CO provided sufficient heat for the conversion of CH4 and CmHn. The unconverted methane in the lower and middle areas, where deoxidization-conversion occurs, facilitated the reforming reaction with H2O and CO2 under high temperatures, generating synthesis gas, such as H2 and CO, as well as a small amount of CO2, N2, and CH4.
3
Experimental results and discussion
3.1
Temperature distribution in the small reactor
In this study, O2/GAS represented the ratio of oxygen flow to the COG flow. Figure 2 shows the temperature distribution in the reactor when O2/GAS was increased from 0.18 to 0.40. As the O2/GAS increased, the bottom temperature rose from 1250-1350 K to 2060-2160 K, the middle temperature rose from 1100-1200 K to 1730-1830 K, and the top temperature rose from 1050-1150 K to 1470-1570 K.
2200 2100 2000
Top Middle Bottom
Temperature/K
1900 1800 1700 1600 1500 1400 1300 1200 1100 0.15
0.20
0.25
0.30
0.35
0.40
O2/GAS
Fig.2. Influence of O2/GAS on the reactor temperature (experiment)
3.2
The conversion rate of CH4
Figure 3 shows that the methane conversion rate increased along with O2/GAS. When the oxygen flow was at 1.4, the O2/GAS ratio was 0.22-0.26, while the overall CH4 conversion rate was at 95-97%. When the CH4 content was less than 1%, the reaction reached optimized working conditions. The synthesis gas to meet the requirements of CH4 content.
105 100
90 85 80
3
O2 FLOW 1.4 m /h
75
O2 1.3
70
O2 1.2
4
CH Conversion rate %
95
O2 1.1
65
O2 1.0 O2 0.9
60 55 0.15
0.20
0.25
0.30
0.35
0.40
O2/GAS
Fig.3. Dependence of the CH4 conversion rate on the different levels of O2/GAS
3.3
H2/CO ratio in synthesis gas and analysis of the effective synthesis gas content
Fig.4.
(a) CO and (b) H2 content in the synthesis gas
When the O2/GAS ratio increased from 0.18 to 0.40, the CO content rose from 20% to 29%, and the H2 content dropped from 65% to 42%. According to Fig. 4(a,b), H2/CO=1.8-2.8 and H2/CO=2.0 are theoretically required to synthesize methanol. Therefore, the H2/CO ratio in synthesis gas could meet the requirements for methanol synthesis.
88 86
2
CO+H Content %
84 82
O2 FLOW
80
3
1.4 m /h 1.3 1.2 1.1 1.0 0.9
78 76 74 72 0.15
0.20
0.25
0.30
0.35
0.40
O /GAS 2
Fig. 5.
H2+CO content of the effective synthesis gas
Figure 5 indicates that the effective synthesis gas content (H 2+CO) is 75-87%. When the oxygen flow was 1.2, the effective synthesis gas content was at its maximum, thereby presenting a parabola distribution. It attained its maximum value when O2/GAS was at 0.26.
4
Numerical simulation of the small reactor
4.1 4.1.1
Establishment of the geometric model: theory and methods Geometric model
The device illustrated in Fig. 6 is a geometric solid with axial symmetry measured in millimeters. Therefore, it can be simplified into a two-dimensional axial-symmetry calculation.
700mm
Fig. 6. Geometric model of the small reactor
4.1.2
Turbulence model
The standard k-ε model equation is the most popular turbulence model applied in hydromechanics simulation. For uncompressed fluids, the standard k-ε model equation is as follows[14]:
( k ) ( kui ) t xi x j
( ) ( u i ) t xi x j
t k Gk k x j
t
C1 2 G C k 2 k k x j
t C
k2
where the model constants are:
4.1.3 Methods of simulation The model was established using GAMBIT. The calculation regions were divided through a quadrilateral mesh. The small-sized oxygen entrance and the large flow rate combined into a more crucial initial reaction. Therefore, the encrypted grid was adopted in the region near the central axis entrance. There were a total of 50381 grids. Control equations were dispersed using the finite volume method.
4.2
Simulation results and discussion
Figure 7 represents the simulation nephogram of the temperature distribution in the reactor. The different colors in the nephogram represented the different temperatures in the reactor. In contrast with the true reactor experimental image (Fig. 8), the simulation flame and experimental flame profiles were similar, and the flame lengths were approximately equal. This confirmed that the simulation theory and the model were reasonable.
Fig.7. Temperature distribution map of the single-hole nozzle reactor (simulation)
Fig.8. Experiment flame image Figure 9 illustrates the maximum temperature and outlet temperature simulation results in the reactor (O2 flow = 1.2 m3/h, wall surface temperature = 1100 K) when the COG flow was changed and the O2/GAS changed from 0.18 to 0.40. The maximum temperature had a linear distribution and rose slightly from 2926 to 3080 K as the O2/GAS increased. The approximately equal maximum temperatures were due to the good coupling between the exothermic combustion and the endothermic reforming reactions. As O2/GAS increased, the various temperatures also rose: outlet temperature (1139 to 1488 K), bottom temperature (1262 to 2144 K), the middle
temperature (1211 to 1848 K). The top temperature only slightly increased from 1162 to 1584 K. The simulation results were consistent with the experimental results.
3000 2800 2600
Temperature/K
2400 2200 2000
Outlet Top Middle Bottom Maximum
1800 1600 1400 1200 0.15
0.20
0.25
0.30
0.35
0.40
O2/GAS
Fig. 9. Influence of O2/GAS on the reactor temperature (simulation) Figure 10 shows the simulation results of the main components in different reaction temperatures. The volume fraction of H2 decreased from 66.17% to 43.29%, CO decreased from 23.25% to 20.2%, CO2 increased from 1.79% to 6.7%, and CH4 decreased from 8.705% to 0.105 %. The effective synthesis gas (H2 + CO) decreased from 89.4% to 63.5%, and H2/CO decreased from 2.8 to 2.14. The simulated data is in good agreement with the experimental data.
Fig. 10. Contents of the main outlet gas under different reaction temperatures Figure 11 is the simulation nephogram of the flow field distribution. The different colors represent the different flow velocities. The flow field was attributed to the turbulent jet. Because the reactor diameter was maintained even without the sudden expansion and contraction of the variable cross-section, no vortex was observed. The combustion reaction occurred mainly close to the entrance and away from the nozzle. It relied on jet turbulent mixing, which was consistent with the laws of hydrodynamics. The carbon catalyst can be placed in
the middle of the high temperature area to promote CO2-CH4 reforming. The reactor had a big H/D ratio. Therefore, the reactants had a long residence time, guaranteeing that CO2-CH4 and H2O-CH4 reforming reactions could be completed.
Fig. 11. Flow field distribution map of the single-hole nozzle reactor Figure 12 is the simulation map of the axis temperature distribution in the reactor. It showed a rapid rise of the temperature of the system in the initial stage of the reaction and a nearly vertical growth condition. This explained the dramatic increase in the temperature after the COG and O2 entered the reactor and the reaction conditions reached the ignition temperature. The temperature distribution indicated that there was initially an increase in the temperature, representing a mainly exothermic reaction. In other words, H2, CH4, C2H4, and C2H6 had a combustion reaction, which enabled the temperature to increase dramatically to the maximum temperature. These combustion reactions emitted heat, which was advantageous for CH4 reforming. It enabled the CH4 reforming reaction to gradually occupy a dominant position. Since the reforming reaction was endothermic, the temperature gradually dropped. The simulation result is consistent with theoretical thermodynamic analysis. T/K
Fig. 12. Axes temperature distribution of the single-hole nozzle reactor Figure 13 is the axis distribution of the major components. A comparison of b and c was conducted for the H2 and CO values, and the H2/CO ratio was measured at about 2.5. The CH4 content also tended towards 0, with a CO2 content that was lesser than 4% at the operating condition. According to Reaction (9), H2 and CO can theoretically synthesize methanol by: 2H2 + CO = CH3OH (9) Reaction (9) shows that theoretical requirements for methanol synthesis are: a H2/CO ratio of about 2.0.The CH4 and CO2 concentrations, as well as the H2/CO ratio, were control the main parameters. Experimental and simulation results showed that the test system using the carbon-catalyzed CO2-CH4 reforming from COG to generate synthesis gas had reasonable a H2/CO ratio, CH4 concentration, and CO2 concentration, thereby making the conversion method feasible.
a
b
d
c
Fig. 13.
5
Axial distribution of the major components: (a) CO2, (b) CO, (c) H2, and (d) CH4.
Comparison between the experimental results and simulations
The simulation results of the single-hole nozzle under 1.3 m3/h of oxygen and 3.663 m3/h gas are shown below (Tables 2-3). Table 2 Comparison between the experimental result and simulation (temperature) Temperature/K Bottom
middle
top
Experiment
value
1510
1409
1312
Simulated
value
1524
1425
1326
Table 3 Comparison between the experimental result and simulation (component) Component mol%
H2
CO
CO2
CH4
C2H4
C2H6
N2
O2
Simulation Experiment
59.11 61.37
22.62 21.55
3.37 3.32
1.32 1.34
1.21 1.47
0.99 0.03
4.66 4.75
2.02 0.43
Table 2-3 shows that the maximum temperature was at 3090 K and the experimental value matched the simulated value very well. This shows that the simulation theories and the models are correct and suitable for this kind of experiment.
6
Conclusions
Experimental research and simulation results on a small reactor showed that the outlet H2 gas content was 50-70% and the CO content was 20%, giving a H2/CO ratio of 1.8~2.8 in the synthesis gas. An H2/CO ratio of about 2.0 was needed to synthesize methanol. The simulation results showed that when the wall temperature in the reaction area was at 1100 K, the molar fraction of the effective synthesis gas (CO+H2) in the outlet gas was
up to 90% and above, and CH4 remained at just 1%. Temperature simulation on the reforming reactor showed that carbon catalyst exited within a temperature range of 900-1400o C. However, relevant research has shown that carbon-catalyzed CO2 reforming of CH4 started at 600o C and that catalysis was optimal at 1200oC. Therefore, the reforming reactor was designed with an appropriate temperature field.
Acknowledgments This work was supported by the National Basic Research Program of China (2005CB221202), the Innovation Team Funds Scheme, and the School Funds of the Taiyuan University of Technology (TYUT).
References [1]Guojie Z., Yongfa Z., Wei Z., CO2 reforming of CH4 in coke oven gas to syngas over coal char catalyst, Chem. Eng. J. 139 (2009) 733–738. [2] Li Y., Yue D., Yongfa Z., Kechang X., The status and trend of China's coke oven gas utilization, Shanxi Energy and Conservation. 1 (2006) 1–4. [3] Xusheng W., Sibudjing K., Rh/Ce-SBA-15: Active and stable catalyst for CO2 reforming of ethanol to hydrogen, Catal Today. 142 (2009) 241-249. [4] Qi W., Yi C., Yong J., Dry reforming of methane in an atmospheric pressure plasma fluidized bed with Ni/γ-Al2O3 catalyst, Catal Today. 141 (2009) 321-329. [5]Nagai M., Nakahira K., CO2 reforming of methane on Rh/Al2O3 catalyst, Chem. Eng. Sci. 89 (2007) 4998–5000. [6] Haghighi M., Zhiqiang S., Jinhu W., On the reaction mechanism of CO2 reforming of methane over a bed of coal char, P. Combust. Inst. 31 (2007) 1983–1990. [7] Jing Q., Lou H., Syngas production from reforming of methane with CO 2 and O2 over Ni/SrO–SiO2 catalysts in a fluidized bed reactor, Int. J. Hydrogen Energy. 29 (2004) 1245–1251. [8] Yongfa Z., Yujie L.,Wei Z., Function of semi-coke for generating synthesis gas with methane-rich gas conversion, Coal Conversion. 28 (2005) 1-5. [9] Fucheng W., Zhenghua D., Technological comparison of generating synthesis gas with non-catalyzed partial oxidation and catalyzed partial oxidation in coke oven gas, Coal Chemical Engineering 2 (2006) 4-9. [10] Fucheng W., Weifeng L., Research on process of generating synthesis gas with natural gas non-catalyzed partial oxidation, Petroleum Chemical Engineering 1 (2006) 47-51. [11] Fucheng W., Zhenghua D., Numerical simulation of natural gas partial oxidation oven, Chemical Engineering 3 (2005) 13-16. [12]Onozaki M., Watanabe K., Hydrogen production by the patial oxidation and steam reforming of tar from hot coke oven gas, Fuel. 85 (2006) 143-149. [13] Sreethawong T., Thakonpatthanakun P., Chavadej S., Partial oxidation of methane with air for synthesis gas production in a multistage gliding arc discharge system, Int. J. Hydrogen Energy. 56 (2006) 850-856. [14] Fujun W., Analysis on calculation of hydromechanics, Beijing. Tsinghua University Press, 2004.
A STUDY ON THE TEMPERATURE PROFILE AND HEAT TRANSFER COEFFICIENTS IN UNDERGROUND COAL GASIFICATION CAVITIES Sateesh Daggupati1, Ramesh Naidu Mandapati1, Sanjay Mahajani1, Anuradda Ganesh2, Pal A.K.3, Sharma R.K.3, and Preeti Aghalayam1* 1
Department of Chemical Engineering, IIT Bombay, Powai, Mumbai-400076, India 2
Energy Systems Engineering, IIT Bombay, Powai, Mumbai-400076, India
3
UCG Group, IRS, ONGC, Chandkheda, Ahmedabad -380005, Gujarat-, India
*Corresponding Author: [email protected], Ph.No: +91(22)2576 7295, Fax: +91(22) 2572 6895 Abstract Underground Coal Gasification (UCG) is the potential in-situ method of converting un-mined coal into combustible syngas which can be served as a fuel for power generation, industrial heating or as chemical feedstock. UCG process provides a source of clean energy with minimal greenhouse gas emissions, when it is compared with the conventional coal mining and gasification. As gasification proceeds underground cavity is formed as the coal burns, and its shape and size changes with time as the coal is consumed. The shape and rate of growth of this cavity will strongly depend on the temperature profile inside the cavity. As underground coal gasification cavities are of irregular three-dimensional shapes, computational fluid dynamics studies (CFD) are essential in order to understand the temperature distribution inside the UCG cavities. A complete knowledge of the transport phenomena inside the UCG cavity is important for both cavity growth modeling and process modeling as it determines the quality and rate of production of the product gas. In the present study, CFD simulations are performed to study the convective heat transfer characteristics of UCG cavities at specified boundary conditions. By changing the feed flow rate, surface heat transfer coefficients are obtained over a wide range of Reynolds number for four different cavity sizes. The methodology of determining heat transfer coefficient through FLUENT is validated by performing simulations for a circular pipe over a wide range of Reynolds number. The predicted heat transfer coefficients are consistent with the correlations. Keywords: UCG, Cavity, CFD
Introduction
The general features of the UCG cavity are thin char layers on the periphery of the cavity, a porous bed of mainly ash, char and coal overlying the injection point and a void space between these zones and cavity roof as shown in figure 1. The amount/volume of these porous media depends on the spalling rate (i.e. thermo-mechanical failure), which in turn depends on the properties of coal and other geological aspects. The reactant gas from the injection well is expanded in the cavity resulting in a drop of velocity and further flow occurs in the above said porous medium and in the void space. Due to the high permeability of the void space relative to that of the ash, char and coal pieces, the majority of the flow between the injection well and production well is expected to occur in the void
space. The dimensions of the cavity in the vicinity of the injection point are of great interest. Since the fluid is injected low in the seam, it has a tendency to flow upwards and outwards from the injection point due to the buoyancy forces arises from the temperature gradient and concentration gradients which are generated from the chemical reactions. The UCG cavity involves stationary solid phase and flowing fluid phase. Moreover, the average temperature in the cavity is in the range 10000C-14000C.
Hence, all the modes of heat transfer i.e convection and
conduction, radiation are expected to be important in a growing UCG cavity,
The convection flow determines the transport of the gases and heat to and from the roof wall. As regards to the reaction taking place at the roof of the cavity, its extent is mainly determined by the local wall temperature and the rate of mass transfer of the reactant and product species to and from the roof wall. As can be seen from figure 1, which shows different ways of heat transfer to and from the roof wall, there exist four different heat sources or sinks. They are heat of reaction due combustion and gasification, heat transfer by conduction to the coal seam which results in evaporation of moisture and pyrolysis of coal, further, the heat transfer by convention to the gas flowing in the cavity towards the production well, heat transfer by radiation to the other walls of the cavity. The temperature of the roof wall is thus primarily determined by the net effect of the heats associated with these processes. It is therefore necessary to first analyze separately each mode of heat transfer and then appropriately incorporate it in the global process model. In the present work, we focus our attention on the convective heat transfer from roof wall to the cavity. Kuyper et al.1 showed that the natural convection is also important as the reactant gas is expanded from the injection well to the cavity of larger volume and gas velocity is greatly reduced, however, the model works well for the gasification of thin coal seams (<3 m thick). It can be concluded from his studies that, the natural convection flow is dominant in the cavity in which Grashoff number is high and the natural convection flow is turbulent under the conditions studied. In contrast, a cavity growth model developed by Yang et al.2 and Throsness et al.3 proved that the convective heat transfer is the significant factor influencing the UCG process. Harloff
4
showed that the forced
convection governs the heat transfer in UCG cavity, however this model assumed that the combustion occurs in the open channels. On the basis of model experiment, Ding et al.5 have modelled the convection diffusion for gasification agent in the UCG process. In earlier investigations (Harloff4 and Thorsness et al.6), smooth pipe correlations are used to determine the heat transfer coefficient. In reality, at the roof, the heated coal shrinks and cracks making the surface rough. Further, oxygen has to permeate in the cracked porous layers to reach the fresh coal, thus the heat transfer coefficient values are expected to be larger than the ones obtained by smooth pipe correlations. In addition to this, UCG cavity grows three dimensionally, in a non-linear fashion. The open literature there is not information on the heat transfer coefficient data for such irregular UCG cavities under the different set of operating conditions. It is with this background that we have undertaken a detailed study the convective heat transfer characteristics of UCG cavities by CFD under the absence of thermo-mechanical failure (i.e. spalling). Four different cavities at different times of UCG
process are considered. Twenty independent simulations are performed for each cavity over a wide range of inlet flow rates /Reynolds number.
Figure 1: Schematic of UCG cavity
Figure 2: Grids used CFD studies on heat transfer
CFD Simulations for wall to bulk heat transfer coefficients Geometry and grid independence Four cavities of different sizes are chosen based on the literature (Harloff4; Dagguapti et al.7). Meshing was done in GAMBIT 2.3.16 of commercial CFD FLUENT package. The shapes of different cavities and the grids used for simulation are shown in figure 2. The grid size is decided based on the grid independency test reported elsewhere8. The dimensions of cavities used for the simulations are given in Table 1. It also indicates total number of grid faces, nodes and cells of meshed geometry.
Input specifications and convergence criteria A „velocity_inlet‟ boundary condition was used at the inlet. A wide range of velocities from 0.2 to 16 m/s was considered for the simulations. At the outlet, a gauge pressure of 5 atm with „pressure outlet‟ boundary condition was specified. Oxygen is used as the feed gas at 673 K inlet temperature and the physical properties are considered to be varying with temperature. The presence of reaction was indirectly considered by choosing a constant roof temperature at 1273 K and rest of the cavity walls were assumed to be loosing heat to the surrounding at a constant heat flux of 25 W/ (m2-K). No slip condition was assumed at all the walls. The realizable κ-ε model with standard
wall functions is used in these computations to model the turbulence in the cavity. This scheme is suitable for flow involving recirculations (Shih et al.9; Fluent10). The boundary conditions used for this study are given in Table 2
Table 1: Dimensions of the UCG cavities at different times Cavity of size 1 (1 day)
2 (3 days)
3 (7 days)
4 (14 days)
9.1
9.1
9.1
9.1
1
2
2.5
3
Distance between the injection and production wells (m) Length of the cavity behind the inlet (backward direction) (m) Length of the cavity dome from inlet (forward direction) (m) Height of the cavity (m)
2.5
4
6
7.8
1.32
2.2
3.2
4.7
Width of the cavity (m) at the injection well
0.9
1.5
2.2
3.3
Width of the cavity (m) at the outflow channel
0.3
0.42
0.52
0.7
0.25
0.3
0.4
0.6
0.2
0.25
0.3
0.4
0.1
0.1
0.1
0.1
Grid faces
1919925
3009937
4697616
7846400
Grid nodes
174135
267232
411344
8321528
Grid cells (tetrahedral and hexagonal)
983334
17854211
2825016
4682295
Width of the cavity (m) at the back side of injection well Width of the cavity (m) at the outlet Inlet & Outlet diameter (m)
Table 2: Boundary conditions and other specifications used for CFD studies Inlet velocity Carrier fluid Outlet pressure, atm Viscous model Cavity roof temperature Cavity other walls
0.2 m/s to 16 m/s oxygen 5 Realizable κ-ε 1273 K Heat flux, 25 W/m-K
The conservations equations 1, 2 and 3 for the mass, momentum and energy for these meshed geometries were solved to obtain the flow patterns under the specified boundary conditions by using the computational software, FLUENT v6.3.26. The SIMPLE (Semi-Implicit Method for Pressure-Linked Equations) algorithm is used to resolve the pressure-velocity coupling and a steady state second order upwind accurate discretization model is used for all independent variables. The solution was judged to be converged when mass flow difference between in and out was also verified and which was of the order of 10 -8. It was also verified that the scalar residuals of each transport equation goes below 1×10-7. By changing the flow rates, eighteen independent simulations are performed under regions of both laminar and turbulent. This procedure is repeated for all the cavities. FLUENT has a facility to
provide the heat transfer coefficient for a specific surface and volume weighted average temperature for a selected volume. For a simulated steady state, the volume weighted average temperature and surface heat transfer coefficient are obtained from FLUENT, which are given in the Table 3.
Continuity Equation
.v 0 t
.............1
Momentum Equation
v .vv p .( ) g F t
..............2
Energy Balance Equation
E .vE p . k eff T h j J j eff .v t j where E h
p
..........3
v2 2
Table 3: Heat transfer coefficient data for all cavities Cavity Size
1
2
Hydraulic Diameter, m 2.27
V, m/s 0.2
TVWA, K 1228.58
2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 2.27 3.70 3.70 3.70 3.70 3.70 3.70 3.70 3.70 continued
0.25 0.5 1 1.5 2 2.5 3 3.5 4 5 6 7 8 10 12 16 32 0.2 0.25 0.5 1 1.5 2 2.5 3
1226.95 1208.98 1165.51 1124.96 1099.93 1079.48 1061.71 1045.00 1029.85 1001.24 949.11 928.34 917.87 854.23 824.66 800.18 752.16 1248.60 1243.59 1238.80 1214.42 1189.96 1160.02 1143.01 1133.95
HTC, W/m-K 0.171 0.176 0.288 0.509 0.636 0.759 0.868 0.972 1.080 1.187 1.698 2.128 2.513 2.878 3.292 3.470 4.340 7.608 0.066 0.082 0.113 0.208 0.307 0.392 0.465 0.536
NRe 2646 3316 6838 14708 23587 32761 42333 52279 62657 73374 96025 125255 151071 175573 243117 306067 424723 919491 4164 5250 10586 22070 34497 48347 62156 75706
NPr 0.7169
NNu 4.470
0.7169 0.7168 0.7165 0.7161 0.7159 0.7157 0.7155 0.7153 0.7152 0.7148 0.7142 0.7139 0.7138 0.7130 0.7127 0.7125 0.7120 0.7171 0.7170 0.7170 0.7168 0.7167 0.7164 0.7163 0.7162
4.630 7.644 13.926 17.898 21.748 25.247 28.682 32.270 35.902 52.550 68.787 82.733 95.630 116.04 125.93 161.49 297.99 2.760 3.450 4.789 8.981 13.469 17.524 21.044 24.412
Cavity Size
3
Cavity Size 4
3.70 3.70 3.70 3.70 3.70 3.70 3.70 3.70 3.70 3.70 Hydraulic Diameter, m 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 5.37 7.21 7.21 Hydraulic Diameter, m
3.5 4 5 6 7 8 10 12 16 32 V, m/s 0.2 0.25 0.5 1 1.5 2 2.5 3 3.5 4 5 6 7 8 10 12 16 32 V, m/s
1126.61 1115.14 1094.58 1082.33 1071.73 1026.46 994.93 972.49 958.47 954.63 TVWA, K 1258.43 1254.56 1245.28 1233.80 1221.74 1208.85 1197.04 1184.77 1175.45 1162.90 1130.02 1115.46 1101.78 1099.92 1079.24 1054.58 1044.19 924.66 TVWA, K
0.603 0.715 0.869 1.084 1.215 1.419 1.784 2.030 2.022 2.916 HTC, W/m-K 0.033 0.049 0.074 0.111 0.149 0.189 0.228 0.268 0.303 0.357 0.485 0.597 0.702 0.778 0.933 1.060 1.173 2.117 HTC, W/m-K
89393 104099 134555 164707 195484 240309 315962 393019 535916 107844 NRe 5944 7480 15201 31006 47474 64697 82495 101052 119741 139732 184395 226607 270341 309896 400591 500244 678251 164283 NRe
0.7162 0.7161 0.7159 0.7157 0.7156 0.7151 0.7148 0.7145 0.7143 0.7143 NPr 0.7171 0.7171 0.7170 0.7170 0.7169 0.7168 0.7167 0.7166 0.7166 0.7165 0.7162 0.7161 0.7159 0.7159 0.7157 0.7154 0.7153 0.7139 NPr
27.601 33.006 40.738 51.291 57.936 70.086 90.354 104.75 105.60 152.77 NNu 2.021 2.995 4.550 6.848 9.265 11.899 14.424 17.120 19.462 23.156 32.178 40.032 47.509 52.739 64.186 74.337 82.943 165.34 NNu
7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21 7.21
0.2 0.25 0.5 1 1.5 2 2.5 3 3.5 4 5 6 7 8 10
1267.53 1266.23 1260.99 1249.46 1239.86 1230.87 1222.69 1214.96 1207.72 1197.11 1188.36 1177.62 1166.96 1157.84 1145.24
0.010 0.013 0.026 0.056 0.081 0.105 0.127 0.148 0.168 0.203 0.220 0.249 0.280 0.323 0.363
7862 9849 19879 40557 61848 83742 106145 129055 152419 177332 224944 274816 326360 378659 448271
0.7171 0.7171 0.7171 0.7171 0.7170 0.7170 0.7169 0.7169 0.7168 0.7167 0.7167 0.7166 0.7165 0.7164 0.7163
0.847 1.092 2.120 4.589 6.719 8.765 10.645 12.454 14.176 17.225 18.854 21.483 24.344 28.235 32.012
Results and Discussions Figure 3 shows the temperature contours for all the cavities for a fixed velocity (10m/s). It indicates that the temperature gradients are prominent for the smaller cavity sizes. The volume of the cavity size 4 is more as compared to the other cavities and we observe a significant drop in velocity. Figure 4 shows the comparison of path
lines with their respective temperatures. It can be seen that most of the upper portion in the case of large cavities is occupied by the hot gas which is almost stagnant. The gas entering the cavity does not reach the roof and finds a shortest possible route (or bypass) to the outflow channel. Independent virtual tracer studies on residence time distribution also support the fact that there exits a dead zone. We would like to mention here that such situation is not favorable for UCG as the reaction in such case will be strongly governed by mass transfer limitations because of which the roof temperature will drop down significantly and the basic assumption of high roof temperature may not be valid in such case. In reality however, radiation effects would also be significant and there will be a radiation driven flow which may in turn alter the convective heat transfer characteristics. Further work is necessary in this direction. As expected, the surface heat transfer coefficient increases with flow rate and at the same time the temperature gradient increases with the feed flow rate as can be seen in Table 3. Figure 5 shows the variation in heat transfer coefficient with respect to the flow rate for all the cavities. Criterion for determining the hydraulic radius of UCG cavities is not easy as it is of highly irregular shape. It is known that the cavity dimensions in backward, forward and transverse directions decrease with increase in height. In order to calculate the average dimensions in these directions, cavity is divided in ten segments in the vertical directions. The average dimensions in each direction is taken at each and every segment. The flow area is calculated by considering the cavity as a hemisphere with a radius calculated as average of all the dimensions in forward, backward, transverse, and vertical directions. Thus, the hydraulic radius is obtained for all the cavities of interest, which are given in Table 3. All the flow related dimensionless numbers such as Reynolds, Prandtl and Nusselt numbers are calculated under these conditions. It is difficult to get the appropriate relation between these dimensionless groups, as the geometry is highly nonlinear in nature. Figure 6 shows the variation of Nusselt number with a wide range of Reynolds numbers in the logarithmic scale. Figure 7 gives the logarithmic relation between the Prandtl number and Nusselt number. Prandtl number is calculated at the volume weighted average temperature. There is a large variation in this average temperature that is causing Prandtl number to vary significantly. It would be advisable to determine this number at the roof temperature. Nevertheless, the results presented here, may be useful as valuable inputs to develop a complete process model for UCG.
Figure 3: Comparison of temperature contours for
Figure 4: Comparison of path lines colored by
all cavities (v=10 m/s)
temperature contours for all cavities 120
3.000
100
Cavity Size-2
2.400
80
Cavity Size-4
1.800
Cavity Size-2 Cavity Size-3
Cavity Size-3
Nnu
Heat Transfer Coefficient, W/m2-K
Cavity Size -1
Cavity Size -1
Cavity Size-4
60
1.200 40
0.600 20
0.000 0
2
4
6
Velocity, m/s
Figure 5: Variation of heat transfer coefficient w.r.t. to the flow rate for all cavities
8
0 1000
100000
Nre
Figure 6: Nusselt number vs. Reynolds number
100 Cavity Size-1 Cavity Size-2 75
Nnu
Cavity Size-3 Cavity Size-4 50
25
0 0.715
0.7154
0.7158
0.7162
0.7166
0.717
NPr
Figure 7: Nusselt number vs. Prandtl number
Methodology Validation for Heat Transfer Coefficient Determination Heat transfer in cylindrical tubes It is necessary to validate the methodology that we have adopted to calculate the heat transfer coefficient. We consider a standard case of heat transfer in an empty tube, shown in figure 8. In order to meet the condition of fully developed flow, the length of the tube is taken as twenty times to its radius. Grid independency was checked by varying the grid cell size nearer the wall. Baten and Krishna11 have performed similar studies for determination of mass transfer coefficient. It is found that average cell size of about 4 mm within the pipe geometry is sufficiently small for solving the velocity and temperature profiles, and surface heat transfer coefficient. Five independent CFD simulations are performed to solve the equations under the conditions of both laminar and turbulent flows. The simulation results for the surface heat transfer coefficient in empty circular tube are obtained are as a function of gas phase Reynolds number which are shown in Figure 10. For laminar flow conditions (NRe <2100) the CFD simulations results (i.e. Nusselt number) are compared with the results obtained from the classical Graetz solution for heat transfer to laminar flow, given by equation 4. The agreement between the calculation of Nusselt number form equation 6.5 and CFD simulation are shown in figure 9. It can be concluded that the results predicted with the help of CFD fall in line with the results calculated from the standard correlations which are available in literature. The percentage error between the heat transfer coefficients determined from CFD and that from standard pipe correlations lies below the ±6 %. Hence, it is proved that the methodology of obtaining the heat transfer coefficients from FLUENT is valid.
N Nu
T TI N Gz ln S TS TVWA
...........4
N Nu 0.023 N Re N Pr 0.8
0.3333
w
0.14
...............5
Figure 8: Geometry used for determining heat transfer in an empty circular tube
Heat transfer coefficient, W/m-K
40
Correlation 30
Simulation
20
10
0 100
1000
10000
100000
Nre
Figure 9: Comparison of heat transfer coefficients obtained from FUENT and correlations
Heat transfer experiments on lab-scale aluminum cavity set-up In earlier section, the methodology of heat transfer coefficient determination through FLUENT was validated for a regular geometry. It is important to know how the heat transfer characteristics differ in the case of an irregular geometry. For this purpose, a lab scale cavity set-up was designed and fabricated out of aluminum. The dimensions of this cavity size were chosen on the basis of the data available in literature. The primary objective of this part of the study is to validate the heat transfer profiles obtained by CFD simulations. The dimensions of the lab-scale aluminum cavity are given in Table 3. Top portion of the cavity roof is subjected to a high temperature by means of a heating tape/coil. In order to maintain the adiabatic conditions the whole cavity is insulated with the glass wool. Cold oxygen gas is sent to the cavity through an inlet. Figure 10 shows the schematic diagram of aluminum cavity set-up. At the start of the experiment, oxygen gas was passed to the cavity at a given flow rate. The flow rate was controlled by an automated mass flow controller. The roof of the cavity is heated to maintain the specified cavity roof temperature. The supply to the heating element was controlled based on the temperature at a specific position. The temperature at the outlet was continuously monitored with the help of a thermocouple. The temperature at the cavity outlet gives the information on how much heat is carried by oxygen gas from the cavity roof to the outlet. Once a steady state is attained, temperature at the outlet is noted down against the feed flow rate and cavity roof temperature. The procedure is repeated for different sets of feed flow rate and roof temperatures.
Table 4: Dimensions of the Lab-scale aluminum cavity Distance between the two wells (cm) Backward Length (cm) [ '-' ve X - direction] Forward Length (cm) ) [ '+ 've X - direction] Height (cm) ) [Z - direction] Width (cm) [Y - direction] Outflow channel width (cm) Outflow channel length (cm) Total volume (cm3) Inlet and Outlet diameter (cm) Faces, Lakhs Nodes, Lakhs Grid cells, Lakhs
80 12 38 25 30 3.5 420 13993 0.4 27.35 2.433 13.61
Figure 10: Schematic of lab-scale aluminum cavity set-up for heat transfer studies
Virtual heat transfer experiments by CFD The aim of this part of the study was to validate the methodology of obtaining heat transfer profiles from CFD simulations. The geometry corresponding to the experimental cavity described above was created and meshed by using pre-processing software GAMBIT. The procedure mentioned in earlier section is used to obtain the steady state outlet temperature by providing the experimental conditions described above, as inputs. Steady state outlet temperatures are obtained for all the cases studied in experiments. Figure 11 shows the comparison of the experimental outlet temperature with that of predicted from simulations under similar conditions. By keeping the cavity roof at a 500 K, the steady state outlet temperature was measured by changing the feed flow rates. The CFD simulations over predict the outlet temperatures due to inability of perfect insulation in the experiment. From this exercise, it can be concluded that the proposed methodology of characterizing the heat transfer phenomena in UCG cavities using CFD simulations, is valid.
Outlet temperature, K
360
345
330
Experiment Simulation
315
300 200
400
600
800
1000
Flow rate, ml/min
Figure 11: Comparison of experimental outlet temperature with that of simulation
Conclusions
CFD simulations are performed to study the convective roof-to-bulk heat transfer characteristics of UCG cavities in the absence of thermo-mechanical failure (i.e. spalling). By changing the feed flow rate, surface heat transfer coefficients are obtained over a wide range of Reynolds number for four different cavity sizes. The relationship between the Reynolds number, Prandtl number and Nusselt number is presented. It is observed that the surface heat transfer coefficient increases with flow rate. The methodology of obtaining the heat transfer coefficients from FLUENT is validated by performing CFD simulations on an empty tube. The comparisons of heat transfer coefficients which are determined from CFD and standard pipe correlations is reasonably good (i.e. % error is <5). The values of heat transfer coefficient in the cavity are extraordinarily low because of the near stagnant pool of gas in the upper part of the cavity. Such a case may not exist in reality as radiation is likely play an important role. Future work may be undertaken to model heat transfer due to radiation and its effect on the convective heat transfer coefficient. To sum up, apart from showing the importance of considering radiation, this work demonstrates a methodology to determine of heat transfer characteristics of the complex systems like UCG. In a longer run such work may help a process engineer to develop a reliable phenomenological model for UCG.
Nomenclature ρ
Density, kg/m3
t
Operation time, min
v
Velocity tensor
p
Static pressure
Stress tensor
ρg
Gravitational body force
F
External body force
hj
Enthalpy of species j
Jj
Diffusion flux of species j
TVWA
Volume weighted average temperature, K
HTC
Heat transfer coefficient, W/m-K
TS
Surface/cavity roof temperature, K
TI
Inlet temperature, K
μ
Viscosity at average temperature, Kg/(m-s)
μw
Viscosity at roof, Kg/(m-s)
NRe
Reynolds number, (DHV ρ/ μ)
NPr
Prandtl number, (Cp μ/k)
NNu
Nussult number (hDH/k)
NGz
Greatz number, (m*Cp/k-L)
V
Inlet velocity, m/s
DH
Hydraulic diameter, m
h
Heat transfer coefficient, W/m2-K
m
mass flow rate, kg/s
L
Characteristic length, m
References 1.
Kuyper, R.A., Van Deer Meer, H. and Hoogendoorn, C. J. “Turbulent Natural
Convection Flow Due to
Combined Buoyancy Forces during Underground Gasification of Thin Coal Layers”. Chemical Engineering Science. 1994, 49(6). 851-861.
2.
Yang, L.; Liu, S. Numerical Simulation on Heat and Mass Transfer in Process of Underground Coal Gasification. Numerical Heat Transfer Part A 2003, 44, 537-557.
3.
Thorsness, C.B. and Charles, B., General-purpose, Packed-bed Model for Analysis of Underground Coal Gasification Processes”, Lawrence Livermore National Laboratory Report, UCID-20731, 1985.
4.
Harloff G.J., Underground coal gasification cavity growth model. J Energy 1983;7:410.
5.
Ding G.X., Yang S.Q., and Zhang J.L., , Journal of China University of Mining and Technology, 1995, 24 (1) 46.
6.
Thorsness CB, Greens EA, Sherwood AA, One dimensional model for in-situ coal gasification, Report No. UCRL-52523. U.S. DOE, Livermore, CA: Lawrence Livermore National Laboratory;1978.
7.
Daggupati, S.; Mandapati, R.N.; Mahajani, S.M.; Ganesh, A.; Mathur, D.K.; Sharma, R.K.; Aghalayam, P.; Laboratory studies on combustion cavity growth in lignite coal blocks in the context of underground coal gasification. Energy 2010, 35, 6, 2374-2386.
8.
Daggupati, S.; Mandapati, R.N.; Mahajani, S.M.; Ganesh, A.; Mathur, D.K.;
Sharma,
R.K.;
Aghalayam, P.; Compartment modeling for flow characterization of underground coal gasification cavity. (submitted to IECR) 9.
Shih, Tsan-Hsing, Liou, W. W., Shabbir, A., Yang, Z., and Zhu, J. A new k−ε eddy viscosity model for high Reynolds number turbulent flows. Computers and Fluids, 1995, 24(3), 227.
10. FLUENT 6.3. Users Guide. Fluent Inc. Lebanon NH; December 2005 11. Baten J.M. and Krishna R. Gas and liquid phase mass transfer within KATAPAK-SR structures studied using CFD simulations. Chemical Engineering Science, 2002, 57, 1531-1536.
Manuscript Not AVAILABLE
ABSTRACT 2. Gasification (UCG)
SOME RESULTS of UCG EX-SITU TRIALS from HBP Company POINT of VIEW Peter CICMANEC*, Frantisek VERBICH*, Jaroslav BELACEK*, Karol KOSTUR** *Hornonitrianske bane Prievidza, a.s., Prievidza, Slovak Republic [email protected], [email protected], [email protected] **Faculty of Mining, Ecology, Process Control and Geotechnology Technical University of Kosice, Kosice, Slovak Republic, [email protected]
The Hornonitrianske bane Prievidza, a.s., Company (www.hbp.sk) as a private company which core business is deep coal mining and mining powered supports production has an interest on underground coal gasification at coal deposits in Slovak Republic for a certain time period. In 2007 year the company started a national grant research project in cooperation with the Technical University in Kosice aimed towards gaining knowledge on underground coal gasification especially for conditions of the Slovak Coal Deposits. The presentation briefly describes Coal Deposits in Slovakia as well as chosen quality parameters of coal seams. It evaluates in more details coal properties from the Novaky and the Handlova Coal Deposits and also the deposits conditions in which the company has mining licenses. Ex-situ research results are evaluated from possible practical usage point of view mainly in electro energy and heat energy sectors, because there is a power plant close to the deposits and the company. In a context of used research methods basically “UCG via Channels” and by means of “Fissured Coal Seam” it analyzes practical questions of future application of vertical boreholes or directed ones respectively of course in connection with mining and geological conditions of the coal deposits. The paper also describes Company´s plans with the underground coal gasification in future times.
THE COAGULATION IN ELECTRIC FIELD OF THE ARGILLACEOUS SUSPENSIONS FROM THE WASTEWATER RESULTED IN COAL PROCESSING Authors: Prof. PhD. eng. ROMULUS SÂRBU 1 Candidate for PhD. ecol. DIANA MARCHIŞ 1 PhD. eng. ADRIAN CORUI 2 1 2
University of Petrosani, ROMANIA SC AQUATIM SA Timisoara, ROMANIA
Abstract: The water resulted from mining industry and coal processing is characterized by a high concentration in argillaceous colloidal suspensions, which do not precipitate free not even in weeks. Waste water from U.P.Coroesti washing technological process has a lot of argillaceous colloidal substances and is recalcitrant in cleaning. Because the dispersion grade is high, these particles have a large specific surface, this explaining the high value of surface energy and their high capacity for absorption the ions from water. Because they will have the same electrical sign they will also have a high gravitational stability. So the flocculation process is a complex one, by electrical, chemical and mechanical nature where the cations tied by anionic group of flocculants challenge the inversion of the solid particles charged and in this way they lose the water layer adherent at their surface. The most efficient process is electrical discharge of the particles. Some research show that between electrodes happen similar phenomena to those from water electrolyze, liberating H+ ions, which are absorbed at the hydrated part of the micelle, changing its sign, or liberating Al3+ ions from consumables electrodes, who in their way to cathode, meeting some minerals particles negatively charged, they partially neutralize this particles and provoke their coagulation by decreasing the electro kinetic Zeta potential (Zp). The galvanic chemical process for cleaning permits to reach the necessary cleaning level, based on the utilization of galvanic elements formed by electrodes pairs, placed in the solution that must be cleaned, by applying a current from an exterior source, without using chemical coagulants reactive. The purpose of this study is to replace Zetag reactive – with coagulation role, used in present time at Coroesti processing plant, with electro coagulation in continuous electric field with consumable anode. Key words: electrocoagulation, electrolyzor, electric field, coagulation, flocculation, double electric layer, colloidal solid suspension, flocculation reactive, consumable anode, electro kinetic Zeta potential, mass balance, solid concentration phase, aluminum ions. 1. The stability of mineral suspensions in water Using previous research, regarding the mineralogical composition of the granulometric fractions, there was emitted the hypothesis of screening the organic compounds with argillaceous pellicles adsorbed to their surface, phenomenon present especially in -3 μm fraction. For this, there were made some destabilization trials on these suspensions with high colloidal character by increasing the electro kinetic Zeta potential by treating the wastewater in electric field. The forming of the colloidal suspensions, through dispersion, is a phenomenon which consists in increasing the separation surface between the two phases of the system: the disperse phase and the dispersal medium. The stability of the particles can be generally explained using the double electric layer from the surface of the particles. This layer is made during the process of the forming of colloids, due to the presence of some reactive used for dispersion. It is noticed that in the system there is, besides the two components which form the disperse phase and the dispersion medium, a third compound, called stabilizer its molecules and ions been fixed at the surface of the particles. Because the stabilizer is soluble in the dispersion medium , its molecules and ions bring to surface the particles and a solvatation sheath. Therefore, there can be said that the double electric layer and the solvatation sheath wrap the particles protecting them of the coagulation process.
The central problem of the colloidal systems stability and coagulation is that of the mechanism used by the superficial pellicles to protect the particles. The double electric layer can be defined as an inhomogeneous interfacial region with a finite thickness, in which there is an appreciable variation of charge density this creating a difference of electric potential at its extremities. [Sârbu R 2008.]. The sign of the zeta potential is adopted conventionally identical with the sign of the charges from the solid surface, but it can be calculated using the following formula: 4 d (1) Zp D where: D – the dielectric constant of the dispersion medium; - the global charge density from Stern layer. The thickness (d) of the mobile part of the diffuse layer is given below: D R T d (2) 8 F 2 c z2 where: - R- gases constant; - T- absolute temperature; - F- Faraday’s number; -c- coagulant electrolyte concentration; -z - the valence of the ions from the electrolyte. 2. Stability and the destruction of colloidal systems The stability of colloidal systems is provided by two factors which independently or synergic act for impeding or delaying the union of the particles from the disperse phase: electrostatic factor steric factor. Assuring the stability with electrostatic factor is realized because of the forming of the double electric layer on the surface of the colloidal particles, in this way assuring the electrostatic rejections between the charged particles. The intervention of the steric factor is the resultant of the absorption of some amfifile molecules or some polymeric substances on the surface of the colloidal particles which act like a mechanical barrier for particle collision. Adding electrolytes at the disperse systems produces coagulation because of the effect of contrary sign charge of the colloidal particle. The effect of the electrolytes is characterized by the minimum concentration at which takes place the rapid coagulation, called critical coagulation concentration (CCC). A flocculent polymeric agent can be characterized by the concentration at witch the flocculation is maximum. Specialized literature recommends many substances with the role of flocculent agents. The coagulants are generally salts of some polyvalent metals Fe3+, Al3+, Ca2+, etc, which by hydrolysis or by dissociation liberates metallic ions which cancel the negative electric charge of colloidal particles from water also producing the coagulation of disperse faze and its rapidly sedimentation in the cleaning process. Electro coagulation consists in the introducing in water of some metallic ions needed for coagulation, using the electrolyses process. There are used electrolyses cellules with metallic anodes of aluminum, iron or copper. As a result of the chemical reactions the pH of the water suffers some modifications, decreasing when the tension increases. Some research show that between electrodes happened similar phenomena like in water electrolyze, liberating H+ ions: H2O + e- → H+ads + OH (3)
This fact proves that the hydroniu ion is responsible for the coagulation of colloidal suspensions. This ion due to the existing electric field, passes from anode to cathode and in its way collides with the colloidal particles negatively charged, practically neutralizing their charge and coagulating them. [Sârbu R, Corui A. 2009]. Another hypothesis proved by Rojanski, is based on the liberation of trivalent ions of Al3+ or Fe3+ from the consumable anodes, which in their way to the cathode, meeting some minerals particles negatively charged, partially neutralize this particles and provoke their coagulation by decreasing the electro kinetic Zeta potential (Zp). Some research showed the benefic effect of the wastewater treatment in electric field and the influence of the electric parameters on the clearing speed, such as: - clearing speed depends only on the quantity of electric current that passes through the suspension. It does not depend on the tension or the intensity of the current only if these parameters are used to calculate the quantity of current; - there is an optimum quantity of current that produces a maximum sedimentation speed. The electrochemical water treatment permits to reach the cleaning level necessary, based on the utilization of galvanic elements formed by electrodes pairs, placed in the solution that must be cleaned, by applying a current from an exterior source, without utilizing chemical coagulant reactive. [Sârbu R 2008]. The constant for the speed of coagulation process is determined by:
K
1 t lg
c c0
(4)
where: K - constant for the speed of coagulation process [s-1]; t - processing, [s]; co – initial concentration of wastewater [g/dm3]; c – final concentration of wastewater [g/dm3]. In order to understand the process there is necessary to describe the modalities of metal solubilization (in our case aluminum). The general reaction of anodic dissolution of aluminum, forming hydrated ions, is shown below:: [Me] + H2O = Meaqz+ + ze(5) The process speed for this kind of electrode vm can be determined with the next formula:
d m (6) dt According to the first law of Faraday Δm= ke . Q, where: Q represents the quantity of electricity equal with the product between the current intensity I and time t, and ke – is a proportionality coefficient called electrochemical equivalent. If we report the first law of Faraday to a gram equivalent, E, then the quantity of current is: Q = I t = 1F (7) this means that: E = ke F (8) respectively E (9) ke F An equivalent gram of Al is: E = 26,9815/3 = 8,993 So, EIt (10) m [g ] F Or m k I t[g ] (11) 3+ where: k – electrochemical equivalent, k for Al is: Al = 0,09316 vm
According to Faraday’s second law, it is proportional with the chemical equivalent A/z, where A is the atomic mass of the element. The modification of the electrode mass during electrochemical dissolution can be determined if a is the ion mass and N – the number of those, that means Δm=a.N. The ion mass is equal with the atomic mass of the element divided to Avogadro number – Na; the number of ions N which pass through the solution is equal with the proportion between the total electric charge, Q, that passes through the system and the charge of the ion: Q (12) N ze Results that: Q A Q A (13) m ze N a F where F is Faraday’s number: F = Na . ze = 96486.7 C/molls. The equation that binds the two laws of Faraday shows that the specific speed of electrochemical dissolution of the substance is proportional with its mass and with the quantity of electricity passing through the system.. Because between the metal quantities dissolved in anode and electricity exists a direct proportionality, based on the above relations it can be written: d m dk e Q vm ke I (14) dt dt In this way, the speed of the electrochemical reaction is proportional with the current intensity I, and it can be expressed in equivalent gram of substance, if the current intensity is reported to the Faraday unit I/F, or in gram unit gram/ion, taking into account also the charge of the particle I/zF. The speed depends on the volume of the separation surface between the electrode-electrolyte phases and because of that, it has to be related to that surface, determining a current density: I (15) I s A/m2 S The quantity of electricity necessary to obtain an equivalent gram from a substance can be determined from the value of F, and the electric energy consumption from the relation: F.E. For evaluating the quantity of electricity involved in the electrode reactions the following were considered: At the cathode: 2H2O + O2 + 4e = 4 OH(16) By passing 4 electrons mol through the galvanic pair, 4 hydroxyl ions will be formed; 1 electron mol is 1 faraday and 1 hydroxyl ion is a gram atom, respectively 17 g. In this way the quantity of electricity necessary is 96486 C. Because 1C =1A x s results that through galvanic pair pass 96486 C, if for example current intensity in 10 A then the passing time will be 9649 s. At the anode: Al = Al3+ + 3e(17) 3 farads dissolve 26.9815 atoms-gram of Al. At the anode can be also possible the following reaction: 2H2O = O2 + 4H+ +4e(18) Results that by passing 4 farads of electricity are liberated 32 g of oxygen. We must specify that changing the electrodes polarity, a rational consumption of Al can be attained. The purpose of these experiments was to establish the coagulant effect of aluminum ions found in solution as an effect of anodic dissolution.
For the first set of trials we maintained the same intensity I = 0,4 A and tension U = 24 V changing only the time parameter and using as a response function the electro kinetic zeta potential the results are shown in the table no.1. Table 1 Anode Time mass No. [s] [mg] Δm1 1 2 3 4 5 6 7 8
15 20 30 60 120 180 300 480
0,7 0,8 1,3 2 4,7 6,8 10,8 18,9
Cathode mass [mg] Δm2 0,1 0,2 0,1 0,2 0,2 0,1 0,1 0,1
Al quantity found in water Δm1 - Δm2 [mg] 0,6 0,6 1,2 1,8 4,5 6,7 10,7 18,8
Unconsumed Al, determined with Al, Faraday’s law [mg] [mg] 0,032 0,008 0,016 0,056 0,208 0,44 0,75 0,96
0,55 0,745 1,12 2,24 4,48 6,7 11,1 17,9
ZETA Potential [mV] -4,85 -2,61 -0,81 0,79 1,12 3,02 7,13 8,71
Figure 1. Correlation between the Al quantities dissolved at the anode, Al determined with Faraday’s law and residual Al from water, all in function of treatment time. According to figure 1 there can be observed that if the solid phase is present the law of anodic dissolution is respected, the solved aluminum mass gravimetric determined and that calculated according to Faraday’s law, being approximately the same. Another conclusion is that from the Al quantity found in water Δm1 - Δm2,, a significant part is adsorbed at the surface of the particles provoking the coagulation process, this being also highlighted by the value of zeta potential. There can be also seen that at this concentration the Al ion is exceeding remaining an unconsumed quantity of Al emphasized in table 1.
Figure 2. Variation of zeta potential by the treatment time Source: [Treatment in electic field of wastewater from the processing of coal in order to clarify them, Doctoral thesis A. Corui , R. Sârbu 2009]
From the dependence between the stationary time in the electrolyzor and the zeta potential can be observed that, in the first few seconds, the unconsumed Al has the value of 0,032 mg and the zeta potential - 4,85 mV which means that the solution is in an accommodation period and afterwards there begins the properly coagulation process, in which the Al ions are adsorbed at the surface of the particles and the zeta potential decreases. The visualization of the coagulant effect of Al ions is shown in figure 3.
Figure 3. The coagulant effect of Al ions So, in this conditions, at t = 30 s there is obtained the lowest value of zeta potential of – 0.81 mV, which corresponds to a strong domain of agglomeration (Sarbu R. 2008). Any increase if the time over the values of 30-35 s produces the restabilisation of the system and the zeta potential takes positive values. Based on these conclusions, further on we treated the wastewater in electric field using an electrolyzor (figure 4 - 1) with cylindrical concentric Al electrodes, alternatively charged to + and – in continuous current. The assemblage used for the treatment of wastewater in continuous flux is shown in figure 4.
Figure 4. The assemblage used for the treatment of wastewater in continuous flux 1-electrolyzor; 2rectifier; 3-input recipient. In order to electrically destabilize the wastewater, we count on decreasing zeta potential to values that favor coagulation using different work schedule(table 2), on one hand and on the other hand we count on decreasing the current density gradient and the speed used to pass the interspaces between electrodes(table 3). We must specify that the alimentation is made from bottom to top, therefore the electrolyzor functions in “drown” regime. Table 2. Variation of sedimentation speed according to the time that the solution stays in the electrolyzor. Trial Alimentation Time, Medium Sedimentation speed flow, [s] dimension at 20 min, [x10-6m3/s] Dm, [μm] x10-3[m/h] 0 0,000 38,318 24 1 359 0.448 146,050 54 2 210 0,766 23,975 15 3 81 1,987 28,020 18 4 42 3,830 83,399 39 -3 3 Working parameters: U=20V; I=2,5 A; Ri=30Ω; c= 50 g/l; Vu= 0,161 x10 m . The results can be visualized in figure 5. We observed that by treating the wastewater in electric field with the dissolution of the anod appears the coagulation process and afterwards there takes place a restabilization of the suspension, phenomenon emphasized by the variation of the sedimentation speed.. Table 3 Variation of electric and technique parameters from inside the electrolyzor. Electrode Lateral area of the Interspaces Current density Speed between 2 2 electrode section [m ] [A/m ] interspaces [x10-3m2] [m/s] 1 2,19 0,691 1141 0,237 2 4,24 0,640 590 0,256 3 6,28 0,461 398 0,355 4 8,16 0,154 306 1,064 For I=2,5A and Q=0,164 x10-3m3/s.
Figure. 5 Variation of medium dimension and sedimentation speed by time From figure 5 results that in the interval 0-0,5 s stationary time, appears the coagulation phenomenon and the sedimentation speed increases but afterwards it decreases dramatically because the suspension stabilizes (Sârbu 2008) The transpose of the results of this research in industrial condition must take into account the stationary time of the wastewater in electric field so that the quantity of dissolved Al ion may be optimum.
CONCLUSIONS By treating the wastewater contaminated with argillaceous mineral suspensions in electric field and by obtaining the dissolution of the aluminum anode and the transfer of Al3+ ions in solution, the coagulation process can be achieved without using coagulant reactive, such as polyelectrolyte, which are very expensive.
REFERENCES
1
Corui A. Sârbu R.
2
Sârbu R.
3
Sârbu R. ,Bekesi C.
4
Valentin A., Solojenkin P. Krausz S
Treatment in electric field of wastewater from the processing of coal in order to clarify them, Doctoral thesis, Petrosani 2009 Treating waste water procedures and equipment Ed. Focus Petrosani 2008 Waste water treatment in electric field ROPET 2001 scientific symposium, Petroşani Theoretical ads practice base for galvanic chemical waste water cleaning, Ed. Universitas Petroşani 2004
İzmir University of Economics, Balçova, İzmir
Geopolitics of Coal and Global Environment Volkan Ş. Ediger, Ph.D. Professor of Energy Economics & Director of Research and Graduate Policies
27th Annual International Pittsburgh Coal Conference Plenary Session-3; Hilton Hotel, İstanbul, 14 Oct. 2010
GROWING COAL USE IN THE WORLD Secure, Abundant Domestic Reserves and Relatively Low Prices Ming Dynasty, 1637
3500 Energy Security
Coal Production, MToe
3000 2500
Stagnation
2000 Great Depression
1500
Global Climate Concern
II.WW
PostOil Crises
I.WW
1000 500
Post-War Growth
0 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
DECARBONIZATION EVOLUTION OF H/C RATIO
WOOD 1/10=0.1
COAL 1/1=1.0
C
OIL 2/1=2.0
HC
GAS 4/1=4.0
H
H 100.0
Sheikh Ahmed Zaki Yamani, Oil Minister of Saudi Arabia in the 1970s
Yamani said in 2000: “Stone Age did not come to an end because we had a lack of stones, and the oil age will not come to an end because we have a lack of oil.”
“Saudi Dove in the Oil Slick,” The Observer, January 14, 2001, 7.
WORLD ENERGY SUBSTITUTION 100
1913
90
Concentration, %
80
Wood Oil Hydropower
70
Coal Natural Gas Nuclear
1973
60
1881
50 40 30
?
1965
20
2000
10
0 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Wood Era
Coal Era
Oil Era
NG Era
THE COAL QUESTION W. Stanley Jevons (1835-1882)
An Inquiry Concerning the Progress of the Nation, and the Probable Exhaustion of our Coal-mines
First Edition, 1865 Second Edition, Revised 1866 Third Edition, Revised 1906
GEOPOLITICS OF COAL BETWEEN PAX BRITANICA & PAX AMERICANA 1200 1100
1913
Coal Production, Mtons
1000 900 800 700 600 500
1890’s
400 300 200 100 0 1740 1760
1780 1800
1820 1840
1860 1880 1900
1920 1940
1960 1980
2000 2020
GEOPOLITICS German Geopolitik Friedrich Ratzel (1844-1904); German Geographer-Ethographer 1897: Politische Geographie;
“Lebensraum” (living space).
Johan Rudolf Kjellén (1864-1922); Swedish Political Scientist-Politician 1900: Introduction to Swedish Geography. 1916: The State as a Living Form; 5 key concepts of German geopolitik.
Sir Halford John Mackinder (1861-1947); English Geographer 1904: “The Geographical Pivot of History”; The term “Geopolitics” 1919: Democratic Ideals and Reality; The Heartland Theory.
SIR WINSTON CHURCHILL’S DIVERSITY 1911-1916: First Lord of Admiralty
British Parliament; 17 July 1913: “On no one quality, on no one process, on no one country, on no one route and on no one field must we be dependent. Safety and certainty in oil lie in variety and variety alone.”
LIMITS TO GROWTH (1972)
It was among the first ones which pointed to the finite nature of fossil fuel: “If the rate of resource use is increasing, the amount of reserves cannot be calculated by simply taking the current known reserves and dividing by the current yearly usage, as is typically done to obtain a static index (exponential growth).” The word “sustainability” was first used in this book: “If current trends are allowed to persist, the breakdown of society and the irreversible disruption of the life-support systems on this planet, possibly by the end of the century, certainly within the lifetimes of our children, are inevitable.”
SUSTAINABLE DEVELOPMENT
G. Bruntland, Ed., 1987, Our Common Future, WCED (The World Commission on Environment and Development)
Mrs Gro Harlem Bruntland Prime Minister of Norway
Item 27: Humanity has the ability to make development sustainable to ensure that it meets the needs of the present without compromising the ability of future generations to meet their own needs. "We have not inherited the world from our fathers -- we have borrowed it from our children”
UNFCCC United Nations Framework Convention on Climate Change 21 March 1994
KYOTO PROTOCOL Int. Agreement Linked to the UNFCCC 16 February 2005
While the Convention encouraged industrialized countries to stabilize GHG emissions, the Protocol commits them to do so. (GHG Emissions Reduction of -5%)
EMISSIONS FROM THE CONSUMPTION OF FOSSIL FUELS
Share in Total Emission, %
60 50
42.5%
40
37.0%
30 20
20.6%
10 0 1980
OIL
1985
1990
COAL
1995
Data from US DOE IEA
GAS
2000
2005
2010
EMISSIONS FROM THE COAL CONSUMPTION
CO2 Emissions from Coal, Bill Metric Tons
6
5
4
China India Japan Germany Australia
41.7%
USA RF S. Africa S. Korea Poland
3
16.5% 2
8.0%
1
3.5% 0 1980
1985
1990
1995
2000
2005
2010
WORLD THERMAL ELECTRICITY NET GENERATION
Thermal Electricty Generation, BillkWh
3500 USA 3000 2500 2000
23.0%
China Japan
20.6%
RF India
1500 1000
5.0% 500 0 1980
1985
1990
1995
2000
2005
World Total Electricity Generation: 18,778.67 Bill. Kwh World Thermal Electricity Generation: 12,739.98 Bill. kWh World: 67.8%, USA: 71.1%, China: 81.2%
2010
WORLD FOSSIL FUEL PRODUCTION 10
Production/Consumption, Bill. Toe
9 8
Coal
7
Oil Gas
6
Fossil Fuel
5
OIL COAL GAS
4 3
1965
2
2000
1 0 1850
1860
1870
1880
1890
1900
1910
1920
1930
1940
1950
1960
Data from Grübler (2003) and BP (2010)
1970
1980
1990
2000
2010
COAL CONSUMPTION AND OIL PRICES 3500
120
II
III
100 90
2500
80 70
I
2000
60
1500
50 40
1000
30 20
500
10 0 1900
0 1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
2010
2020
Oil Prices, US $ 2009
Coal Production, MToe
3000
110
COAL PRODUCTION 1800 1600 Coal Production, MToe
1400 1200 1000
China
USA
FSU
Australia
India
EU
45.6%
800
15.8%
600 400
5-7%
200 0 1980
1985
1990
1995
2000
2005
2010
BIGGEST TEN IN COAL, 2009 Reserve, %
Production, %
Consumption, %
1
USA
28.9
China
45.6
China
46.9
2
RF
19.0
USA
15.8
USA
15.2
3
China
13.9
Australia (C: 1.6)
6.7
EU
8.0
4
Australia
9.2
India
6.2
India
7.5
5
India
7.1
EU
4.6
Japan (R: ?, C: ?)
3.3
6
Ukraine (P: 1.1, C: 1.1)
4.1
Indonesia (R:0.5, C: 0.9)
4.6
S. Africa
3.0
7
Kazakhstan (C:1.0)
3.8
S. Africa
4.1
RF
2.5
8
S. Africa
3.7
RF
4.1
Germany (R: 0.8, P: 1.3)
2.2
9
EU
3.6
Poland
1.7
S. Korea (R: ?, P: ?)
2.1
10
Poland
0.9
Kazakhstan (C: 1.0)
1.5
Poland
1.6
%
94.2
%
94.9
%
92.1
BIGGEST TEN IN OIL, 2009 Reserve, %
Production, %
Consumption, %
1
Saudi Arabia
19.8
RF
12.9
USA
21.7
2
Venezuela (C: 0.7 )
12.9
Saudi Arabia
12.0
EU (R:0.5, P: 2.6)
17.3
3
Iran (C: 2.2)
10.3
USA (R:2.1)
8.5
China
10.4
4
Iraq (C: ?)
8.6
Iran
5.3
Japan (R: ?, P:?)
5.1
5
Kuwait (P: 3.2, C: 0.5)
7.6
China (R: 1.1)
4.9
India (R:0.4, P: 0.9)
3.8
6
United Arab Emirates (C:0.6)
7.3
Canada (R:2.5)
4.1
RF
3.2
7
RF
5.6
Mexico (R:0.9, C:2.2)
3.9
Saudi Arabia
3.1
8
Libya (P: 2.0, C: ?)
3.3
Venezuela (C: 0.7)
3.3
Germany (R: ?, P:?)
2.9
9
Kazakhstan (P: 2.0, C: 0.3)
3.0
United Arab Emirates
3.2
Brazil (R:1.0, P: 2.6)
2.7
10
Nigeria (P: 2.6, C: ?)
2.8
Iraq
3.2
South Korea (R:?, P:?)
2.7
61,3
%
73
%
75,6
%
BIGGEST TEN IN GAS, 2009 Reserve, %
Production, %
Consumption, %
1
RF
23.7
USA
20.1
USA
22.2
2
Iran
15.8
RF
17.6
EU
15.6
3
Qatar (C:0.7)
13.5
EU (R:0.5)
5.7
RF
13.2
4
Turkmenistan (P:1.2, C: 0.7)
4.3
Canada (R:0.9)
5.4
Iran
4.5
5
Saudi Arabia
4.2
Iran
4.4
Canada
3.2
6
USA
3.7
Norway (R: 1.1, C:0.1)
3.5
China
3.0
7
UAE (P:1.6, C: 2.0)
3.4
Qatar
3.0
Japan (R: ?, P:?)
3.0
8
Venezuela (P: 0.9, C: 1.0)
3.0
China (R:1.3)
2.8
UK (R:0.2, P:2.0)
2.9
9
Nigeria (P:0.8, C:?)
2.8
Algeria
2.7
Germany (R:?, P:0.4)
2.6
10
Algeria (C: 0.9)
2.4
Saudi Arabia
2.6
Saudi arabia
2.6
%
76.8
%
67.8
%
72.8
HYDROCARBON PRODUCTION 1400
1972
1989
1200
HC Production
1000
RF USA
800 600
1999
400
EU
200 0 1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
2015
STRUGGLE OVER OIL AND GAS 600 500
USA
Production, MToe
RF 400 300
OIL GAS
200
EU
100 0 1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
2015
70
8
60
7 6
50
1998-99
5
40 4 30 3 20
2
Share of Net Import Domestic Production
10
1
Total Consumption 0 1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
Production/Consumption, Bill. bbls
Share of Net Import, %
USA’S OIL DEPENDENCY AND ENERGY SECURITY
0 2015
European Union, 2009 (In 2030, 70% in Gas, 90% in Oil, 100% in Coal): “to guarantee uninterrupted supplies of traditional energy sources while working to mitigate their environmental impact and investing in infrastructure that can help secure the supply of fossil fuels and exploit existing renewable technologies.”
FOSSIL FUEL RESERVE BILLLIONAIRES Coal
USA
115.8
6.1
3.4
125.3
16.4
92.4
RF
68.7
39.2
10.2
118.0
15.5
58.2
China
58.9
2.2
2.0
63.1
8.3
93.4
Iran
-
26.1
18.9
45.0
5.9
-
S. Arabia
-
7.0
36.3
43.3
5.7
-
Venezuela
0.3
5.0
24.8
30.2
4.0
Qatar
-
22.4
2.8
25.2
3.3
-
UAE
-
5.7
13.0
18.7
2.4
-
12.7
2.1
0.8
15.6
2.1
81.0
Total
256.3
115.8
112.3
484.5
63.5
52.9
World Total
412.4
168.7
181.7
762.8
100.0
54.1
EU
Gas
Oil
FF
%
% Coal in FF
Btoe
1.1
Qmax 1 a e bt
Qt
HUBBERT CURVE OF COAL 2056 4,467.6 MToe
Coal Production, MToe
Oil
45.7 yrs.
4000
Natural Gas
62.8 yrs.
Coal
119 yrs.
3500 3000
1 b
ln
URR: 701 BToe
5000 4500
t max
2500 2000 1500 1000 500 0 1750
1850
1950
2050
2150
2250
Liu Lin (2010)
China
~2025
Evans Mohr (2009)
World
2026/2034
Aleklett Höök (2009)
USA
2030
Total Shafiee (2010)
World
2042
Croft Patzek (2010)
World
2047
2350
1 a
“UNPRECENTED UNCERTAINTY” Our energy future has never been so uncertain
2013
1913 Supremacy in Energy
1901
2001
?
1914-45
? WOOD International Energy Regime
COAL
Geopolitics
OIL
Energy Security
NG
H2
Interdependence
VEnergy’0809
Climate Change and Energy Insecurity: The Challenge for Peace, Security and Development Felix Dodds, Andrew Higham and Richard Sherman, Eds., October 2009
“We must treat climate as a security issue, the most important threat to global security we will ever face. Energy is at the heart of this transition. Climate security and energy security are two sides of the same coin: one cannot be achieved without the other…” Maurice Strong Secretary General of the Rio and Stockholm UN
CARBON INTENSITY 3,0
World United States India United Kingdom Iran
Carbon Intensity using PPP
2,5 2,0 1,5 1,0
China Russia Canada Korea, South
Metric Tons of C02 per Thousand Year 2005 U.S. Dollars
0,5 0,0 1980
1985
1990
1995
2000
2005
2010
LONG WAVES IN DOMINANT ENERGY SOURCE AND BASIC INNOVATIONS 1880-90
1930-40
A.K. Graham and P.M. Senge, 1980
CLEAN COAL TECHNOLOGIES
IGCC
UCG
FBC
Integrated Gasification Combined Cycle Underground Coal Gasification Pulverized Coal Combustion Fluidized Bed Combustion Carbon Capture & Storage Coal Bed Methane
PCC
CCR
CBM
The Future of Coal An Interdisciplinary MIT Study, 2007
Options for a carbon constrained world This MIT study believes that coal use will increase under any
foreseeable scenario because it is cheap and abundant. We conclude that CO2 capture and sequestration (CCS) is the critical enabling technology that would reduce CO2 emissions significantly while also allowing coal to meet the world’s pressing energy needs.
KING COAL IS BACK?
INTERNATIONAL COOPERATION ON NEW CLEAN COAL TECHNOLOGIES FOR AN INTERDEPENDENT SUSTAINABLE FUTURE
Thar Coal Mining Challenges
Dr. Farid A. Malik EMR-Consult & FCC University - Lahore Pakistan
Dr. Muhammad Hanif EMR-Consult. Lahore Pakistan
Contents
-
Abstract
-
Introduction
-
Discussion
-
References
-
Appendix
ABSTRACT:
Despite tremendous potential, the mining sector remains seriously underdeveloped in Pakistan. Thar Coal Field, with huge reserve of 175 billion tons of lignite having coal seams as thick as 25 meters offers a tremendous opportunity to turn the grossly energy deficient country into energy surplus one. However, there may be equally big challenges, 80 % of which may be related to mining and 20 % to subsequent processing and utilization. The paper concentrates on mining challenges related to depth, weak nature of ground and ground water. These challenges can be successfully met by using proper technologies of ground excavation and ground water control. Hundreds of exploratory boreholes in Thar Coal Field have helped in delineation and evaluation of the deposit. The authors believe that there is dire need of starting a pilot open pit so as to procure large coal samples for pilot scale tests and to establish the best mining practices on sound technical and commercial basis depending on the ground and water conditions. INTRODUCTION: The mining sector in Pakistan remains seriously deficient. From Coal to Chromite the mining, beneficiation and refining practices are either outdated or nonexistent. In the seventies Resource Development Corporation (RDC) was created for the development of the Mineral sector. Saindak Copper and Gold Mine Development was the first major project of RDC.
Lack of field experience and mismanagement not only consumed RDC, it also resulted in the lease out of the Saindak Metals Ltd (SML) to Chinese contractors (Metallurgical Construction Company). Despite huge investment (Rs.14.5 billion) an opportunity was missed to nurture local mine development process capability.
Chromite mineral has been mined in Hindubagh District of Baluchistan since decades but exported without subsequent processing or value addition. In case of Thar Coal Field the Asian Development Bank (ADB) in their experts review report has declared 80% mining challenge and only 20% thereafter. Unfortunately this single world’s largest coal deposit has remained un-tapped due to lack of mining focus. Like most mineral deposits, Thar poses its unique mining and hydrology challenges that can be overcome by modern mining techniques, technologies and know how.
Food and energy security have become very important in a global, de-regulated world. For centuries the rich alluvial plains of East and West Punjab have been the food baskets of India/Pakistan. Thar coal Field with its estimated reserve of 175 billion tons ranks as one of the world’s largest single deposit and can transform Pakistan from a fuel deficient to surplus nation, as such both these vital areas can adequately be covered.
Each natural deposit is unique and requires a dedicated process of extraction. Mineability has to be established. The Asian Development Bank (ADB) report has characterized the challenge as 80% mining and 20% utilization (1). Since the early 1990s no significant development has been carried out mainly because of lack of mining focus. Commercial approach has also been missing (2).
The ADB report strongly suggests that a test-pit be dug for the required information on utilization, economic viability and mining challenges. In line with this recommendation a test-pit has been designed (Appendix I). Soil and hydrology data has been used to calculate the slope angles. A 10 meter wide coal seam will be exposed after digging to a depth of 150 meters from which bulk samples will be obtained to evaluate the gasification and liquefaction characteristics of the Thar Lignite and to establish the mine ability of the deposit. DISCUSSION: Four areas will be discussed to understand the mine ability of the deposit, Soil (Figure 1), Hydrology, Test-Pit Design (Figure 2), Challenges; Soil: The top dune layer is of variable thickness, but averages around 68 meters, followed by alluvium, Mudstone, Sandstone and Hydrostatics Barrier (equators) whose exact nature and extent is not fully understood (3,4). Design to a depth of 150 meters with a slope angle of 300 and stripping ratio of 7.75:1M3:t, 22 million cubic meters of soil will be removed and 54000 tons of coal mined, which will be used for pilot level evaluation and process development. The existence of pressurized aquifers above and below the lignite have to be fully investigated at the Test-Pit level in order to understand their behavior in terms of capacity, hydraulic head and recharge potential. This understanding is essential in determining the feasibility, cost of mining and will eventually effect the choice of mining system for large scale production (1).
Hydrology: There are discontinuous aquifers in and above sub-recent deposits in relatively low permeability ground which can have insignificant and transient effect on mining operations. The aquifers in Bara formation are continuous, under high hydrostatic pressure which can release enormous amount of water under pressure into any excavation through weak or cohesion-less gritty sandstone of high permeability. Because of weak ground, almost impossible ground water control and thick coal seams, underground mining will not be possible. As for open pit mining, dune sand and most of sub-recent formation can be mined by usual surface mining methods, such as bucket wheel, scraper-dozer, dragline, tower dragline, shovel and truck or front-end loader and belt conveyor. The actual method will be decided by detailed feasibility study. The slope angle will depend upon geo-mechanical and hydrological investigation. It is felt that attempt to dewater up to the coal seam bottom will be very difficult. The egress of water under pressure out of high permeability, cohesion-less gritty soil will be damaging. It is possible that rather than taking water out of the pit, it may be more advantageous to live with it. Coal and some of the ground just above it may be removed by dredging or dragline dredging. The dredging depths may be prohibitive. Accordingly, special custom made dredges may have to be built(7). Trial Pit Before the start of actual surface mining operation, a pilot scale trial pit will be required to be excavated for the following reasons: 1)
To ascertain slope design of the actual production pit based on the stability and deformation behavior of the prototype pit.
2)
To assess mining equipment performance by possible field trials.
3)
To collect bulk samples of coal for pilot scale tests related to various processes such as combustion, gasification production of fertilizers, liquid fuels and petrochemicals. A serious effort to obtain bulk samples by underground mining miserably failed due to very weak cohesion-less soil and violent egress of water under excessive pressure.
4)
The trial pit will serve as the first cut for the actual production pit. Hence every cubic meter of ground excavated for the trial pit will be accounted for as positive investment.
5)
Unless a trial pit is excavated, a reliable bankable feasibility will not be possible.
Test Pit Design: Over 64% of the world coal is mined by the Open Pit Method. Reaching the deposit economically will establish the mine ability and subsequent evaluation of the bulk samples. The top of coal seam at depth of 150 M will be 100 M in diameter. The slope angle may be initially assumed as 300 for upper dune sand, 350 for recent alluvial deposits and 400 for Bara formation and coal. The overburden to be removed to a depth of 150 M will be around 15 million M3. With a 30M wide ramp which will remain at 20M after sloughing, this may be enough for the pilot pit. However for the production mine the ramp will be at-least 40M. The conceptual design is attached in Annexure I. The pit will be circular in horizontal sections having an inverted conical shape. Every single cubic meter removed for the test pit will be discounted for in the cost of the production mine. The importance of coal as a fuel cannot be undermined as it is abundant, lower in cost, and located in areas away from Middle East which is prone to instability (5). The National Coal Council of America has stated that coal production be doubled from the 2005 level of 1.1 billion tons/year to 2.4 billion tons/years by 2025(6) Challenges: Thar coal deposit is a huge energy and petrochemical source which offers a great opportunity to solve Pakistan’s otherwise un-resolvable energy shortfalls, in terms of electric power cuts, gas rationing and import of 70% of petroleum consumed. These problems are bound to become unbearable due to unrestricted population growth and urbanization. In Thar coal deposit hundreds of exploration holes have been drilled for regional reconnaissance and exploration as well as for specific target investigation. Several large blocks have been extensively explored. The exploration effort has now reached a stage where pilot scale investigation is required for coal mining and subsequent utilization processes of gasification, power generation, coal to petro-chemicals etc. The Thar coal deposit has all the qualities of an excellent mineral deposit in terms of huge quantity, simple geological formation and being ideal for gasification. It is minable by simple mechanical operations of earth moving machinery. No drilling and blasting is required. This is a huge advantage.
However, Thar coal mining will not be without challenges, these are depth, weak ground and ground water. These parameters will together control mining method or in-situ coal gasification process(7). CONCLUSIONS: 1.
With estimated reserves of 175 billion tons and value of about 20 trillion dollars, Thar coal is an important energy resource that must be developed expeditiously.
2.
80% challenge is in mine development followed by 20% for utilization where established technologies can be utilized under the IGCC (Integrated Gasification Combined Cycle) processing approach.
3.
Production mines require huge capital investments therefore the Test Pit approach is most appropriate to establish the mineability of the deposit.
4.
The bulk samples obtained from the Test Pit can then be utilized for evaluation of the gasification and liquefaction potential of coal.
5.
The proposed design of the test can be implemented after detailed evaluation by a team of experts comprised of geologists, geotechnical and mine planning engineers. Hydrological studies are crucial for selection of digging sites and development of extraction methodologies.
6. 7.
Information gathered from the test pit will form the basis of a bankable feasibility and Master Plan for development
REFERENCES 1.
Pakistan Coal Resource Project Final Report Asian Development Bank, Ref: Z7Y26AIA May 26, 2007.
2.
F. Malik, R.BAQRI; Proper Exploitation of Thar Coal, The Dawn, May 07, 2005
3.
S. Jafferi; Geological Survey of Pakistan, Private Commission, November 2008
4.
I. Khan; Sind Coal Authority, Private Communication, November 2008
5.
EMR-Consult; Thar Coal Project Development Road Map, Internal Document, May 2010
6.
National Coal Council of America; Coal Strategy upto 2025
7.
M. Hanif; Thar Coal Development Strategy, EMR Internal Document, July 2010
Appendix I
Figure 1:
Bore Hole Data
Figure 2:
Test Pit Design
Figure 1: Bore Hole Data
PILOT TEST PIT FOR THAR COAL PROJECT
DUNE SAND
SUB-RECENT BARA FORMATION COAL BARA FORMATION GRANITE
Figure 2: Test Pit Design
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission PROGRAM TOPIC: SUSTAINABILITY AND ENVIRONMENT (Energy Production and Water Use - Conservation and Recycle) EXACT TITLE OF PAPER: SUBMERGED SEQUENCING-BATCH MEMBRANE BIOREACTOR TO TREAT THE COKE WASTEWATER Wen-Ying LI, Prof., Key Laboratory of Coal Science & Technology (Taiyuan University of Technology), MOE and Shanxi province Contact Information: Tel.&Fax: +86 351 6018453, Email: [email protected] Jingwen Wu, Taiyuan University of Technology Baojie Zhao, Taiyuan University of Technology Jie Feng, Prof., Taiyuan University of Technology; Email:[email protected]
Abstract: The coke wastewater contains toxic substances and higher ammonium nitrogen, it is difficult for biological treatment. Removal of COD and NH4+-N from the coke-plant wastewater in a laboratory scale by the Submerged Sequencing-Batch Membrane Bioreactor (SSBMB) was investigated in this article. The reactor was fed with coke wastewater in descending dilution multiples until finally original coke wastewater was supplied. Results showed in the “anoxic-aerobic” process, most of the organic compounds were removed in the aerobic stage, the COD removal efficiency of the two stages were 24% and 65% respectively; in the “anoxic–aerobicanoxic” process, the removal efficiency of organic compound in the former anoxic stage was increased to 32%. Removal mechanisms indicated that there existed the membrane filtration and biological degradation occurred simultaneously on the membrane surface, but the former was the dominant.
OPTIONS OF CO2 CAPTURE IN OXYFUEL COAL COMBUSTION TECHNOLOGIES
J. Rodrigo –Naharro, C. Clemente-Jul Escuela Técnica Superior de Ingenieros de Minas. Universidad Politécnica de Madrid. Alenza 4. 28003 Madrid. Spain A review of the projects that are developing the oxyfuel coal combustion technology around the world has been undertaken based on their available progress. The evaluation carried out is focused on different projects, such at small and large pilot scale as demonstration projects. All these studies and projects undertaken and in process show that fundamental understanding of the principles and basis of oxycoal combustion with flue gas recirculation have been well established during the past 20 years of R&D activities, but there are still some gaps in knowledge that are trying to be solved. The last objective of all of these projects is to demonstrate the oxyfuel technology in a commercial scale, taking into account another competitive alternatives, such the post-combustion for retroffiting existing power plants or the IGCC option. he conclusion of this study is that the oxyfuel combustion technology is suitable for retrofitting pf boilers and achieves clean coal combustion, lowering NOx, increasing SOx removal, possibly lower mercury emissions and obtaining a CO2 concentration for sequestration.
Keywords: CO2 capture, oxyfuel, coal, combustion
Introduction
Nowadays it is assumed by nearly everyone that the earth is warming. Since the moment that has been recorded the temperatures, by 1861, the earth’s overall temperature has risen by 0.74 ºC. Other changing climatic parameters have been observed over recent years : heat-waves and periods of drought, flooding, glaciers melting, and a rise in sea level. In the opinion of the scientific community, climate change is very probably related to human activities and the large amounts of so-called greenhouse gases (GHG) that are accumulating in the atmosphere. The main GHG that causes the climate change is CO2, which emissions to the atmosphere has been increased so rapid that it appears our planet is capable of absorbing only half of the emissions now generated by human activities. This fact will provoke a disastrous impact on global warming, unless we reduce CO2 emissions by 50%-80% by 2050, according to IPCC report.1 Yet with world energy demand predicted to increase by 60% between 2004 and 2030, and renewable energies to make up only a third of the energy mix by 2050, the immensity of the challenge becomes clear and evident. Improving energy efficiency will help enormously, but in itself will not solve the problem.2-3 There are different possible alternatives that are taking into account in order to reduce the CO2 emissions: improving energy efficiency, implementing renewable energies, changing lifestyle, increasing the use of biofuels and biomass, improving technologies for cleaner energy production from fossil fuels, etc. The last alternative cited is commonly called CCS (Carbon Capture and Storage) technologies, that represents a safe and efficient method of capturing and storing billions of tonnes of CO 2 underground for thousands of years. Furthermore, CCS involves three stages: capture, transport and storage. CO2 is captured at a large, fixed source then concentrated and transported to a suitable storage site. For each of these stages, various techniques are either already available or being evaluated. Among the capture techniques, there’s the capture by
oxyfuel combustion, which involves combustion of the fossil fuel in pure oxygen rather than air so that the flue gas stream has a very high CO2 content, 90% or even 95%. This paper aims to show the actual situation of the worldwide projects that are trying to demonstrate the oxyfuel technology using coal as the fuel, making a discussion among them and focusing on its main characteristics.
Pilot Scale Projects Pilot-scale studies on oxyfuel have used standard pf test facilities with additions of oxygen supply, capability to recycle flue gas, and possibly CO2 compression capability. These typically have thermal inputs from 0.02 to 30 MWt and include those of EERC/ANL, IFRF, IHI, CANMET, AIR LIQUIDE, ENCAP and E.ON UK. The earliest study of coal oxy-fuel combustion in a pilot-scale furnace was carried out by the Energy and Environmental Research Corporation (EERC) for Argonne National Laboratory (ANL) in their 3 MWt pilot facility located in the USA to test performance. The focus of the study is to demonstrate the technical feasibility of the CO2 recycle into the boiler, to determine the ratio of recycle gas to O2 for achieving heat transfer performance similar to air firing, to quantify the observable operational changes such as flame stability, pollution emissions, and burnout and to provide a basis for scaling experimental results to commercial scale.4 The main achievements of the study are that the oxyfuel combustion had a similar infurnace gas temperature profile as the normal air-fired combustion and also it was found that oxyfuel combustion had lower NOx (50%) and SOx emissions, and a high carbon burnout compared to air firing. The International Flame Research Foundation (IFRF) Study was carried out at their 2.5 MWt Facility located in the Netherlands, and aims to evaluate the combustion of pulverised coal during oxy-fuel combustion for retrofitting existing pulverised coal fired
boilers to maximise the CO2 concentration in flue gas. Besides, tries to optimise oxy-fuel combustion conditions to yield similar radiative and convective heat transfer performance to air firing and to evaluate the impact of oxy-fuel combustion on furnace performance, including flame ignition and stability, heat transfer, combustion efficiency, pollutant emissions, compared to air operation. The furnace used has internal square cross-section of 2x2 m and 6.25 m long air-staged swirl burner.4 The results of the tests were similar to those from ANL/EERC tests, with the additional conclusions that the recycle ratio depended on the coal used and air leakage resulted in max CO2 concentration to 91%. The IHI Study was carried out at their IHI’s 1.2 MWt Facility located in Japan. The combustion-test furnace has a horizontal cylinder furnace with 1.3 m inner diameter and 7.5 m length and a swirl burner. The focus of the study was to test the combustion characteristics of pulverised during oxyfuel combustion with primary consideration to reduce NOx and SOx pf boiler. The conclusions were that oxygen concentration should be high to raise the adiabatic flame temperature during oxy-fuel combustion to match that in air combustion and that the NOx and SOx emissions decreased.
4
The CANMET Study is being performed by the Energy Technology Center, that is a division of Natural Resources Canada, in their Vertical Combustor Research Facility (VCRF) in Canada. This facility, which was built in 1994, has a long history in experimental results and represents CANMET’s most modern pilot facility for combustion research. The facility has got a nominal thermal input of 0.3 MWt and is capable of firing pulverized coal and or natural gas using air or oxygen. The furnace is a cylindrical, down-fired and adiabatic vertical combustor with an inner diameter of 0.60 m and a length of 6.7 m. It is one of the few research facilities operating in the world that can fire pulverized coal in a mixture of recycled flue gases and oxygen in order to simulate O 2/CO2 combustion techniques. These combustion techniques can be used to enrich the concentration of CO 2 in the products of combustion of fossil-fueled power plants in order to facilitate the capture
of CO2 for use and or sequestration. Furthermore, the objectives are to understand pulverised coal combustions behaviours in various O2/FGR mixtures, compared with air combustion and to demonstrate the technical factors on the combustion performance. The tests proved the technical feasibility of the technology and the effect of oxygen purity recycle and different oxygen concentration in burner was also investigated. Besides, the emissions of NOx and SOx were lower (decrease in SOx by higher retention in ash).4 The Air Liquide and Babcock&Wilcox Study, with the sponsorship of the U.S. Department of Energy, was carried out at their facility of 1.5 MWt boiler with air staged combustion system located in the USA. The focus of the study is to demonstrate the technical feasibility of conversion from air firing to oxy-fuel combustion for large scale boiler and to highlight the impacts of oxy-fuel combustion process on pollutant (NOx, SO2 and Hg) emissions and boiler efficiency. The results were that overall oxy-coal combustion characteristics were comparable to the air-firing case even with the change in oxidant composition from air to oxygen-enriched flue gas; the NOx emissions from oxy-coal combustion
were
significantly
lower
(65%
less)
than
the
air-fired
case;
the
thermodynamics and heat transfer in the furnace and the convection pass changed only modestly; substitution of combustion air with oxygen and recycled flue gas increased the CO2 concentration from 15% to 80% at the boiler exit; and the flue gas volume exiting the boiler is reduced by 70% relative to air-fired operation.5 The ENCAP Project is being developed in the context of an European project. The “Enhanced Capture of CO2” (ENCAP) is a five-year integrated project (IP) within the EU sixth framework research programme.6 This Project is structured in 6 Sub-Projects, and ENCAP SP3 is focused on oxyfuel technologies where combustion using almost pure oxygen and recycle of flue gas to moderate the combustion temperature is carried out. 8 In general the oxyfuel process used for power generation of large scale coal fired boilers is a new technique. Furthermore, the work in ENCAP SP3 is only related to oxyfuel
combustion of solid fuel such as bituminous and lignite coal and pet coke and it’s focused on PF and CFB technology. The goal of the work in ENCAP SP3 is to show that oxyfuel combustion of bituminous coal, lignite can reach a CO2 avoidance cost of 20 € per ton of CO2 by applying innovative design and utilise process integration opportunities. The objectives are to develop and validate oxyfuel combustion based power plants concepts for bituminous coal and lignite for a greenfield plant using PF and CFB boiler technology, to provide a conceptual boiler design and suggest its integration with a power generating plant to provide an economically competitive technology, and to examine special issues related to oxyfuel combustion in laboratory and pilot scale to be able to accommodate for those in the plant design and mitigate risk. There are several work packages (WP) within this sub-project. The WP 3.1 aims to understand the fundamentals in oxyfuel combustion that has been throughout the extensive experimental work performed in the 0.1 MWt gas-fired test unit of Chalmers and in the 0.02 MWt coal-fired unit of IVD at the University of Stuttgart. The WP 3.3 and WP 3.4 aims to define the conceptual design of greenfield oxyfuel PF and CFB plant respectively, based on advanced supercritical power plant. The WP 3.5 aims to perform combustion tests in semi-technical scale (0.5 MWt). The WP 3.6 aims to validate the technology of oxyfuel in Vattenfall’s 30 MWt oxyfuel pilot plant in Schwarze Pumpe. In summary, a number of activities related to experimental investigations of oxyfuel fundamentals in the small-scale test rigs and the conceptual development of the full-scale oxyfuel PF and CFB plants have been finished. Combustion tests related to investigation of NOx emission behaviour with staged combustion under simulated oxyfuel conditions have been investigated. Similar reduction behaviour as under air firing has been concluded. The investigated ash and slagging behaviour under oxyfuel and air firing conditions with ENCAP coals have also found to be similar. In the work package related to
conceptual design of oxyfuel PF plants, and economic assessment including transport and storage scenarios has been concluded. The layout of the system used for compression and processing of CO2 has been reported together with identification of processes to be included to reach the CO2 purity levels defined for the Enhanced Oil Recovery (EOR) and ship transport case in the ENCAP guidelines. A RAM analysis has been performed. The conceptual design of the oxyfuel CFB plant has been finished. Activities to reconstruct a 0.5 MWt test rig at the University of Stuttgart for oxyfuel operation has mainly been finished and the functionality of the rig demonstrated. A decision to focus the ENCAP Phase II Large Scale Testing activities on tests in Vattenfall 30 MWt Oxyfuel PF Pilot Plant in Schwarze Pumpe power station in Germany has been taken.7 In short, the main achievements and conclusions of ENCAP SP3 are, after analysing the combustion characteristics of oxyfuel combustion of 2 lignites and 2 bituminous coals in small combustion facilities (0.02-0.5 MWt), for PF plant, to achieve similar temperature profile as in air-firing requires higher overall O2-level (25-27%(v)) but will result in higher gas radiation intensity; for CFB tests have been performed up to 90%(v) O 2 in oxidant; NOx reduction is possible through flue gas recycle and staged combustion; no major impact on burnout and CO level at furnace outlet and no major impact on composition of ash and deposits detected. With regard to the oxyfuel power plant concept development, we can conclude that the conceptual designs of PF and CFB oxyfuel plants have been established (1st generation oxyfuel plants) with the characteristics that the design for the furnace and boiler is more compact than for air, the design of a CO 2 compression and purification unit to reach the required <4% non-condensable components in CO2 product, the economic analysis indicate that the target of 20 €/ton CO2 avoided (year 2004 level) can possibly be reached for a nth plant, the procedures for start-up and shut-down and load change have been suggested, the safety aspects, technical risks and resulting expected availability for a nth oxyfuel plant and no show-stoppers have been identified.8
The E.ON UK Study is carried out at their Facility located in the UK and property of E.ON Company. This study is being developed under a Project called “Assessment of Options for CO2 Capture and Geological Sequestration (ASSOCOGS)” which is looking for ways of capturing CO2 produced from oxyfuel combustion plants with advanced supercritical boiler and turbine technology. It’s a project in association with the Research Fund for Coal and Steel (RFCS) and the work package which is related to the oxyfuel technology is WP2 “Oxyfuel combustion”. The Combustion Test Facility (CTF) is already demonstrating oxyfuel firing technology and besides E.ON is collaborating with other industrial partners to demonstrate large-scale oxyfuel burners. The objectives of the WP2 are to review design of CTF for oxyfuel combustion, to develop revised operational procedures, to prepare detailed redesign, to construct and commission CTF, to refine operational procedures, to perform parametric testing and to review findings and implications.9 The E.ON UK’s 1 MWt CTF was commissioned in 1993, has got a time-temperature scaled and it can operates with varieties of fuel: coal, biomass, oil, orimulsion, gas, others. Besides, has got a full combustion staging (overfire air, reburn and flue gas recycle) and is highly instrumented and controllable. The thermal input is 1 MWt (0.8 – 1.2 MWt), the furnace is horizontally fired, refractory lined, water cooled and balanced draft, with dimensions of 1m x 1m x 1m.10 The conclusion of the tests carried out demonstrated a stable flame without burner O2 addition, some reduction in NOx concentrations, similar composition of the deposits and the ashes. Table 1 provides a summary of the studies about oxyfuel technology performed at a laboratory scale and at a small pilot-scale that have been explained previously.
Large Scale Pilot and Demo Projects Large Pilot-scale and Demo Projects studies on oxyfuel that have used standard pf test facilities with additions of oxygen supply, capability to recycle flue gas, and possibly CO 2 compression capability. These typically have thermal inputs from 30 MWt and include those of Jupiter, Ciuden, B&W, Vattenfall, OxyCoal-UK, Callide, Jamestown, Pearl Plant and Youngdong. The US Coal Jupiter Oxyfuel Project is carried out at a Facility of 15 MWt in Indiana (USA) by the Jupiter Oxygen Corporation, who has patented fossil fuel combustion technology using a high temperature flame with 95% to 100% pure oxygen for combustion while air is excluded. Therefore, there is no airborne nitrogen. Nitrogen uses energy during combustion, thereby reducing the amount of energy available to be transferred to the water tubes for the making of steam. Lower nitrogen means that more combustion energy is available for making steam. The oxy-coal flame is at a much higher temperature and generates far more radiant energy transfer than an air-coal flame (also true using other fuels). Radiant heat transfer is far more efficient than convective heat transfer (the higher flame temperature can be safely used for power plant retrofits with existing materials and existing process temperatures). Furthermore, Jupiter's combustion technology operates at or very close to stoichiometry and creates longer residence times. This means that there is a more efficient burn of the fuel and more time for the energy generated to be absorbed as the heated gas passes through the boiler. These four factors (reduction of nitrogen at combustion, greater radiant heat transfer, stoichiometric oxygenfossil fuel ratio and longer residence times) mean that Jupiter Oxygen's technology makes boiler combustion become much more efficient and use significantly less fuel to create the desired megawatts. In addition, exit gas emission clean-up with back-end technologies are more efficient and less costly because of lowered NOx at combustion with air excluded,
lower pollutants due to the fuel reduction, reduced exit gas volumes and the changed composition of the exit gases. The Jupiter Oxygen patented technology also results in more concentrated CO2, resulting in less expense capture and sequestration. Initial test results using carbon and natural gas showed a tenfold increase in CO2 concentration.
11
The CIUDEN Spanish Project consists in the development of the oxyfuel technology by means of the construction of a Test Facility with pulverised coal technologies (20 MWt) and coal fluidised bed technologies (20 MWt). Some of the technical characteristics of the Experimental Facility are that it has got a Pilot PC boiler with 2+2 burners in order to see the flame interactions. Besides, it has a flue gas cleaning section for NOx, SOx and particles and a CO2 Capture Section, including several technologies, such as chemical absorption (amines), flue gas compression and carbonation/calcination eventually. The initial focus is low volatile fuels (anthracites and petroleum coke), which is a differential factor as regards to the other international projects. It has fully integrated preheating train adaptable to any operating scenario, from air-fuel to oxyfuel combustion.12 The B&W Clean Environment Development Facility (CEDF) for CO 2 capture Project is carried out in Alliance, Ohio (USA) and was managed and funded by Babcock&Wilcox, American Liquide and Utility Advisory Group. The CEDF was modified to use and mix oxygen, added WFGD and other auxiliary equipment for oxycoal and the boiler type is pulverised coal. The Utility Advisory Group provided end user design feedback for commercial applications. The test campaigns underway include lignite, sub-bituminous coal and bituminous coal. The plant was built in 1994 with support of DOE and others, and it’s in operation since 2007. The input of coal is 30 MWt. The results of the tests proved that the emissions of NOx were significantly reduced (>50%), the SO 2 removal was not significantly different than with air and the oxy flame was bright and stable.13 The German Vattenfall Project is cited above, within the last stage of the ENCAP Project, and is carried out by Vattenfall, who has taken a decision to build a 30 MWt
Oxyfuel PF pilot plant near to the existing power plant at Schwarze Pumpe. The investment decision was assumed in May 2005, and it’s in operation since August 2008. Pilot testing using lignite will result in validation and tuning of the more or less commercially available technologies included in the oxyfuel concept to allow launch of a demonstration project of the technology in commercial scale by 2015 (≈ 600 MWt). The objectives of these tests are to define optimal operating technologies for oxyfuel conditions in a large-scale facility for the entire process, to identify critical issues for further R&D and to gain operating experience in the oxyfuel field.
14
The OxyCoal-Uk Project consists in developing and demonstrating a competitive oxyfuel firing technology suitable for full plant application post-2010. To reach this goal, the Project is structured in three stages. The phase 1 aims to understand the fundamentals and underpinning technologies, the phase 2 aims to demonstrate an Oxyfuel Combustion System of a type and size (40 MWt) applicable to new build and retrofit advanced supercritical oxyfuel plant and the last stage completes, thanks to the previous stages, the foundation for the development of an oxyfuel boiler reference design, demonstration plant installation and operation.15 The Callide Oxyfuel Project is a project performed by Australia and Japan and consists in retrofitting 30 MWe (90 MWt) oxyfuel coal-fired power plant for capture and storage demonstration. The Callide Oxyfuel Project will take the technology to a larger scale in order to demonstrate that it can be applied to existing and new coal-fired power stations to achieve very significant reductions (up to 90% of CO2) in emissions.16 The objectives of this Project are to establish the design and operating requirements for large scale (commercial) retrofit and new-build oxyfuel plants, including geological storage and to estimate capital and operating costs for the next generation of oxyfuel near zero emission technology electricity generating plants.17
The Jamestown Oxy-Coal Demonstration Project is a demonstration-scale 50 MWe oxyfuel coal plant with CCS and is located in Jamestown, New York (USA) and carried out by the New York Oxy-Coal Alliance and other companies, Praxair among them. The OxyCoal Alliance also provides the opportunity to explore geologic carbon sequestration and a new carbon capture technology (“oxy-fuel,” developed by New York-based Praxair). The objectives of the Project are to demonstrate the fully integrated CCS project, to employ advanced technologies (CFB boiler, ASU and CO2 processing unit and system integration), to prove reliability and availability, to operate with typical load factor variations and to learn transient modes of operations. These goals will enable direct scale-up to a commercial CCS operation. The project is estimated to start up in 2013.18 The Pearl Plant Project is carried out in a Facility with has a 22 MWe (66 MWt) PC boiler and is located in the USA. The main fuel is bituminous coal and it’s estimated to start up in 2009. The Youngdong 100 MWe (300 MWt) Oxy-Combustion Power Station Project Demonstration is carried out by the Korea Electric Power Research Institute, with the support of the Korea Institute of Science and Technology and is located in Youngdong (Korea). The Project aims to replace the Youngdong 125 MWe unit #1 that is currently being operated with domestic anthracite and it will be decommissioned by 2013. The objectives of the Project are to demonstrate the oxy-fuel operation by 2018, to improve the fuel flexibility, using sub-bituminous coal, and to optimize the power generation cost.19
Table 2 provides a summary of the different large scale pilot and demo projects developing the oxy-coal technology. Figure 1 shows the oxy-coal combustion projects in operation and planned to start up in the world-scene.
Oxy-Coal Combustion Projects 1000 Vattenfall Youngdong
Thermal Input (MWt)
100 B&W
10 EERC/ANL
IFRF AIR LIQUIDE
IHI
1
Jamestown Callide Pearl Plant Oxy-coal UK Vattenfall Ciuden Jupiter
E.ON UK ENCAP CANMET
0,1 1980
1990
2000
2010
2020
Year
Figure 1. Oxy-coal combustion projects in operation and planned to start up in the worldscene.
Discussion
It’s obvious the rising importance that the oxy-combustion is acquiring in the context of the climate change as a potential greenhouse gas (GHG) mitigation option for fossil fuel power plants. Proof of this, is the number of projects that have been doing up to now and those in progress. After initial introduction in the early 80’s, oxyfuel combustion for pulverised coal combustion was researched as a means to produce relatively pure CO2 for Enhanced Oil Recovery (EOR). Despite
these research efforts, the technology did not pick up on a large scale for this application. However, the awareness and the concern for the greenhouse gas emissions into the atmosphere has taken up the interest again in this technology, with a two-fold focus: the generation of a concentrated stream of CO2, which is need for sequestration, and the potential to reduce pollutant emissions, in particular NOx. This research into this technological option for capturing CO2 has not been limited to the construction and development of new plants that have the advantage of smaller flue gas cleaning equipment due to the lower volume, but has also included the retrofiting of current plants that are in operation. This renewed interest in oxyfuel combustion has impulsed to many laboratory-scale and small and large pilot-scale studies to develop and demonstrate this technology. The main technical considerations and key points in the development of oxy-coal combustion with CCS are: the combustion section (furnace/boiler), the flue gas cleaning section, the CO2 capture section and the O2 supply section. The issues that need further investigations detail to obtain a more fundamental understanding of the changes between oxyfuel combustion and conventional air-fired combustion are: heat transfer performance of new and retrofitted plant and the impact of oxygen feed concentration and CO2 recycle ratio; the gas cleaning required; assessment of retrofits for electricity cost and cost of CO2 avoided; the combustion of coal in an O2/CO2 atmosphere, including ignition, burn-out, and emissions. Regarding the boiler, most of them are pulverised coal (PC) boilers. The advantages of the circulating fluidized bed (CFB) regarding the PC boilers are that in an oxygen-fired CFB system, with the proper amount of recirculated gas, it is possible to obtain bed conditions within the furnace that are very similar to air firing using the appropriate systems that can be designed to have the same flue gas flow to coal flow ratio as with air firing and the same volumetric fraction of oxygen in the flue gas leaving the bed as that of the air-fired case. Furthermore, a critical advantage for oxygen-fired CFB type furnaces is the capability to reduce flue gas flow for a given coal input, by minimizing the recirculated flue gas flow, while maintaining the furnace temperature. With an oxygen-fired CFB, a recycling loop of solid materials is used, by adjusting the heat pick up, to control the furnace temperature at an optimum operation level for combustion efficiency without any potential risk of agglomeration. The concept scheme offers significant opportunities for size and cost reduction in
14
greenfield design. Moreover, due to the inherent fuel flexibility associated with the CFB boiler, multiple fuels can be used such as coal, petroleum coke, biomass and a range of opportunity fuels. 20 From the studies done so far, a list with the main advantages and disadvantages regarding the oxyfuel technology extracted are shown in Table 3.
Conclusions All these projects explained in this review show the rising importance of the oxyfuel technology as a technological option to capture CO2. The pilot-scale studies have demonstrated the feasibility of pulverised coal oxy-fuel combustion. Besides, no major technical barriers in pilot-scale studies have been found and furthermore, the oxyfuel combustion technology is suitable for retrofitting pf boilers and achieves clean coal combustion, lowering NOx, increasing SOx removal, possibly lower mercury emissions and obtaining a CO2 concentration suitable for its sequestration. Technology has been demonstrated only at a laboratory and pilot-scale, not at a practical scale, but demonstrations are planned. The ability to capture CO2 from power plants is feasible in advanced modes of current technology and with new technologies under development with significant industrydriven R&D underway. Oxy-combustion technology requires introducing some new equipment such as the ASU and the CO2 CPU to power plants. Furthermore, new coal fired power plants shall be designed for highest efficiency to minimize CO2 and other emissions. Oxy-combustion is complementary with conventional boiler and steam power plant technology, including efforts towards ultra-supercritical conditions (for efficiency), and environmental control developments. Therefore, to demonstrate this technology, is needed performance optimization (management of incondensibles like Ar, N2 or O2, combustion with low NOx and excess O2, boiler start up and optimization); CO2 treatment (purification, compression and utilization of the captured CO2); optimize the power generation cost (minimize the efficiency penalty, CO2 capture cost), and scale-up and validation. The opportunities of the deployment of oxy-combustion technology are due to the low technological risk option, it’s possible a large power plant size and the repowering and retrofitting, all boiler
15
technologies are adaptable with the use of flexible fuels, the increase of the steam cycle is also possible, apart from the potential reduction of the size of the boiler ant the supply of advanced O2. The challenges that has to face this technology are those related to the time (on time development), technology (scale-up validation, adaptation to installed base, innovation), regulatory policy (air and CO2 injection permits), government budget to support CCS demonstration and value for CO 2. Besides, another important challenges are the cost (cryogenic oxygen, CO 2 quality, CO2 compression, heat flow optimisation, integration) and the public acceptance of this technology. Particularly, the challenges that have to face the demonstration projects are that all pilot projects will proceed with industrial and governmental financial support. The success factors for the demonstration of this technological option are those engaged with financial, regulatory, legal, geological and technical issues. It’s also important for the final result the proximity to a suitable geologic CO2 storage site and the receptivity by stakeholders. The possible future uncertainties for the sucess of this technology are mainly, the international CCS regulations and weak storage resources. Work to date has demonstrated oxyfuel technology as a technically viable solution for CO 2 Capture that combines well proven, commercially available components with minimal risk.
16
References (1) Intergovernmental Panel on Climate Change, “Third Assessment Report - Climate Change 2001”, Cambridge University Press (2) The IEA World Energy Outlook, 2007 (3) Shell’s Long-Term Energy Scenarios (4) Buhre, B.; et al. Progress in Energy and Combustion Science 2005, 31, 283–307. (5) Farzan, H.; Vecci, S.; McDonald, D.; McCauley, K.; Pranda, P.; Varagani, R.; Gautier, F.; Tranier, J.P.; Perrin, N. State of the art of oxy-coal combustion technology for CO2 control from coal-fired Boilers. Presented to Third International Technical Conference on Clean Coal Technologies for Our Future (May 15 - 17, 2007. Sardinia, Italy) (6) ENCAP Periodic Activity Report Year 3 (7) Andersson; Klas; et al. Public summary report of ENCAP deliverable D3.1.4 Final report on combustion fundamentals, slagging and fouling and analysis of impact on flue gas treatment equipment [online]. Available from Internet: www.encapco2.org (accessed on November 10, 2008) (8) Anheden, M. ENCAP SP3: Oxyfuel Boiler Technologies. Presented to CASTOR-ENCAPCACHET-DYNAMIS Common Technical Training Workshop. 22-24 January 2008, IFP-Lyon (9) Goh, B.; 1 MWt Oxyfuel Combustion Test Facility. Presented to Coal Research Forum and Combustion Division meeting (April 17 2007, Imperial College London). (10) Goh, B.; E.ON UK’s Pilot Scale Oxyfuel Combustion Experiences: Development, Testing and Modelling. Presented to the 3rd Workshop of the IEA GHG Oxy-Combustion Network (March 5-6, 2008. Yokohama, Japan) (11) Patrick, B.; Jupiter Oxygen Corporation. Presented to the 3rd Workshop of the IEA GHG OxyCombustion Network (March 5-6, 2008. Yokohama, Japan)
17
(12) Cortes-Galeano, V.; Test Facilities for Advanced Technologies for CO2 Abatement and Capture in Coal Power Generation. Presented to the 3rd Workshop of the IEA GHG Oxy-Combustion Network (March 5-6, 2008. Yokohama, Japan) (13) Farzan, H.; McDonald, D.; McCauley, K.; Varagani, R.; Periasamy, C.; Perrin, N. Oxy-Coal Combustion Pilot. Presented to the 3rd Workshop of the IEA GHG Oxy-Combustion Network (March 56, 2008. Yokohama, Japan) (14) Anheden, M. Vattenfall’s Schwarze Pump Oxyfuel Pilot – An update. Presented to the 3rd Workshop of the IEA GHG Oxy-Combustion Network (March 5-6, 2008. Yokohama, Japan) (15) Fitzgerald, F.; Current Oxyfuel Combustion Activities at Doosan Babcock Energy Limited. th
Presented to Coal Research Forum 19 Annual Meeting (April 10, 2008. University of Nottingham, UK) (16) Van Puyvelde, D.; CCS in Australia. Presented to RITE – The current status and future issues for the development of CCS technology (September 26, 2008. Tokyo, Japan) (17) Spero, C. Oxy-combustion Technology in the World Scene. Presented to the Coal 21 Conference (September 2007) (18) Bonaquist, D.; Victor, R.; Shah, M.; Hack, H.; Hotta, A.; Leathers, D. Oxy-Coal Combustion Demonstration Project. Presented to the 3rd Workshop of the IEA GHG Oxy-Combustion Network (March 5-6, 2008. Yokohama, Japan) (19) Kim, J.S.; Choi, S.; Kim, Y. ; Kim, S.C. Oxy-Combustion Research Activities in Korea. An overview to the Youngdong 100 MWe Oxy-Combustion Power Station Project Demonstration. Presented to the 3rd Workshop of the IEA GHG Oxy-Combustion Network (March 5-6, 2008. Yokohama, Japan) (20) Béal, Corinne et al. Public summary report of ENCAP deliverable D3.4.4 Feasibility study of an integrated
445
MWe
oxy-fuel
CFB
supercritical
boiler
[online].
Available
from
Internet:
www.encapco2.org
18
Table 1. Studies about Oxyfuel Technology performed at a Laboratory Scale and at a Small PilotScale organism
MWt Facility
EERC/ANL
3
locatio
focus of the study
n
results compared to air-fired combustion
USA
CO2 recycle boiler
similar temperature profiles
determine ratio FGR
lower NOx, SOx emissions
19
IFRF
2.5
Holland
quantify operational changes
high carbon burnout
provide a basis for scaling
higher CO2 concentration
retrofitting of existing pc
similar temperature profiles
optimise
lower NOx, SOx emissions
combustion
conditions
high carbon burnout
evaluate
furnace
performance
FGR depends on coal and air leakage higher CO2 concentration
IHI
1.2
Japan
reduce NOx and SOx pf boiler
same heat flux lower NOx, SOx emissions higher CO2 concentration
CANMET
0.3
Canad a
pc combustion in O2/FGR mixtures
lower NOx emissions lower SOx by higher retention in
combustion performance
ash higher CO2 concentration
AIR LIQUIDE
1.5
USA
retrofitting from air-fired to oxyfuel evaluate
emissions
NOx,
UE
emissions
(65%
fundamentals
similar temperature profiles similar thermodynamics
determine boiler efficiency
0.5
NOx
less)
SO2, Hg
ENCAP
lower
of
oxyfuel
higher CO2 concentration higher [O2] to achieve similar
20
combustion
temp profiles for PF
evaluate oxyfuel issues at lab scale
similar carbon burnout
develop oxyfuel PF and CFB boiler provide
lower NOx emissions
similar
ash
and
deposit
composition conceptual
boiler
similar CO levels
designs boiler and furnace is more compact higher CO2 concentration E.ON UK
1
UK
demonstrate technology
oxyfuel
stable flame without burner O2 addition similar
ash
and
deposit
composition lower NOx emissions
Table 2. Different Large Scale Pilot and Demo Projects developing the Oxyfuel Technology
21
project
MWt
location
start up
boiler type
main fuel
Jupiter
15
USA
2007
industrial
natural gas coal
Ciuden
20+20
Spain
2010
pilot PC + pilot CFB
anthracite petroleum coke
B&W
30
USA
2007
pilot PC
bituminous coal sub-bituminous coal lignite
Vattenfall
30
Germany
2008
pilot PC
lignite
Oxy-coal UK
40
UK
2008
pilot PC
coal
Pearl Plant
66
USA
2009
pilot PC
bituminous coal
Callide
90
Australia
2010
pilot PC
bituminous coal
Jamestown
150
USA
2013
pilot CFB
bituminous coal
Youngdong
300
Korea
2018
100 MWe PC
coal
Vattenfall
600
Germany
2015
200 MWe PC
lignite bituminous coal
Table 3. Main advantages and disadvantages of the oxyfuel technology advantages
disadvantages
22
operation
similar
to
air-fired
boiler,
high
availability
significantly reduced efficiency because of ASU and CO2 compression, by 9%
ability to operate in oxygen or air mode.
reduced sent out electricity by 18-20%
applicable for existing (retrofitting) and new PF
may require SOx removal
plants and CFB plants lower emissions of SOx, NOx, Hg
large air separation unit is required compared to other near-zero technologies
flue gas volume lower than air-fired combustion
not applicable to capture-ready plants
variety of feedstocks can be used does
not
require
CO2
capture
prior
to
compression
23
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
PROGRAM TOPIC:
Carbon Management FIXED-BED ADSORPTION OF CARBON DIOXIDE-NITROGEN MIXTURES ONTO ACTIVATED CARBON: CHARACTERISTICS OF CO2 ADSORPTION AND MODELING Prof. Dr. Regina F.P.M. Moreira, Department of Chemical and Food Engineering – Federal University of Santa Catarina – Campus Universitário – Trindade – 88040-900 – Florianópolis – SC – Brazil. e-mail: [email protected] Prof. Dr. Tirzhá L.P. Dantas, Department of Chemical and Food Engineering – Federal University of Paraná – Centro Politécnico o – Jardim das Américas – 81531-980 – Curitiba – PR – Brazil. e-mail: [email protected]
Prof. Msc. Francisco Murilo T. Luna; Prof. Dr. Ivanildo J. Silva Jr.; Prof. Dr. Diana C. S. de Azevedo, Department of Chemical Engineering – Federal University of Ceará – Campus do Pici – 60455-760 – Fortaleza – CE – Brazil e-mail: [email protected] Dr. Carlos A. Grande; Prof. Dr. Alírio E. Rodrigues, Department of Chemical Engineering – Faculty of Engineering – University of Porto – Rua Dr. Roberto Frias s/n – 4200-465 – Porto – Portugal e-mail: [email protected]
Abstract: The atmospheric content of the most abundant greenhouse gas, CO2, has risen from preindustrial levels of 280 parts per million (ppm) to present levels of over 365 ppm. The main sources of CO2 emissions are the combustion of fossil fuels, such as coal, natural gas and petroleum, and industrial processes, such as oil refinement and the production of cement, iron and steel. The reduction of carbon dioxide emissions from flue gases can be achieved using post-combustion technologies such as adsorption. Different adsorbents, such as activated carbon, zeolites, MCM-41, mesoporous silica material SBA-15 and several enriched-amine sorbents, have been tested. Good recovery and product purity have been accomplished with very high energy consumption. An ideal sorbent should offer high adsorption and selectivity for carbon dioxide as well as economically feasible regeneration. However, if the 1
affinity of the adsorbent for carbon dioxide is too high, the regeneration step can negatively affect the cost of the process. In this paper, we report an experimental and theoretical study on the separation of carbon dioxide and nitrogen on activated carbon in a fixed bed. The breakthrough curves were obtained at different temperatures using CO2/N2 mixtures. A model based on the Linear Driving Force (LDF) approximation for the mass transfer was used, taking into account the energy and momentum balances, to satisfactorily reproduce the breakthrough curves.
1. Introduction Global warming is widely attributed to an increase in greenhouse gases in the atmosphere and, especially, to an increase of CO2 release to the atmosphere. The capture and storage of CO2 is a feasible option to reduce emissions to the atmosphere (Wong and Bioletti, 2002). The pathways of CO2 separation through pre-combustion decarbonation (that is, CO2 capture after the gasification process and before the combustion stage), O2/CO2 recycle combustion (using O2 instead of air combustion) or post-combustion separation (CO2 capture from the exhaust gases) can be readily carried out using known technologies and processes. However, these technologies still present a series of problems related to their application, including the great volume of exhaust gases and the cost-benefit ratio. Thus, the goal of reducing carbon dioxide emissions on an industrial scale requires the development of low cost capture methods. Many studies have been conducted worldwide in the field of CO2 capture by adsorption, indicating that this technique is attractive as a post combustion treatment of flue gas: adsorbents must have a high adsorption capacity, high selectivity, good mechanical properties, ease of regeneration and remain stable over repeated adsorption/desorption cycles (Yong et al., 2002). Different adsorbents, such as activated carbon (Guo et al., 2006; Grande and Rodrigues, 2008), zeolites (Zhao et al., 2007; Zhang et al., 2008), MCM-41 (Glover et al., 2007), mesoporous silica material SBA-15 (Liu et al., 2007) and several enriched-amine sorbents (Drage et al., 2007; Gray et al., 2008) have been tested. Activated carbons are efficient adsorbents that offer acceptable CO2 adsorption capacity without modification, due to their high surface area and adequate pore size distribution (Maroto-Valer et al., 2005). In this study, the adsorption of a CO2/N2 mixture onto a commercial activated carbon was measured and the surface changes caused by the CO2 adsorption onto the solid were investigated. A model using the Linear
2
Driving Force (LDF) approximation was used to describe the breakthrough curves for the carbon dioxide and nitrogen adsorption onto the activated carbon, taking into account the energy and momentum balances.
2. Experimental 2.1. Materials The gases used in this study were provided by Air Liquide S.A (Portugal) or by White Martins Ltda (Brazil). Carbon dioxide N48 (99.998% purity), helium N50 (99.99% purity) and nitrogen N45 (99,995% purity) were provided by Air Liquide S.A (Portugal). These gases were used in the determination of the adsorption equilibrium isotherms. Helium (99.995%) and a standard CO2/N2 mixture (20% CO2/N2 v/v) were provided by White Martins S.A (Brazil), and used for the breakthrough curves. The adsorbent used was a commercial activated carbon NORIT R2030 supplied by Norit (Netherlands). The textural characterization was performed by nitrogen sorption measurements at normal boiling point, using an automatic sorptometer, Autosorb 1C (Quantachome, USA), and the particle density and particle porosity were determined by mercury intrusion, Table 1.
Table 1: Properties of the activated Carbon Norit R2030 Particle density, p
1138 kg m-3
Particle diameter, d p
0.0038 m
Particle porosity, p
0.46
Particle tortuosity,
2.2
Solid specific heat, Cs
880 J kg-1K-1
BET surface area
1053 m2.g-1
Average pore radius, ro
1.23 nm
2.2. Adsorption equilibrium isotherms Before obtaining the adsorption equilibrium isotherms, the solid samples were pre-treated for 12 hours at 150oC under vacuum. The equilibrium of pure CO2 and N2 adsorption onto commercial activated carbon was measured at 301K and 423K using the static method in a Rubotherm magnetic suspension microbalance (Bochum, Germany) at up to approximately 5 bar. The adsorption equilibrium isotherms were described using the Toth model, Eq. (1): 3
q
qm K eq P
1 ( K
n eq P )
(1)
1/ n
where qm is the maximum adsorbed concentration, i.e., the monolayer capacity; K eq is the equilibrium adsorption constant and n is the heterogeneity parameter of the activated carbon. The temperature dependence of the adsorption equilibrium can be described by the Van't Hoff equation, Eq. (2):
K eq K o .e
H ads R.T
(2)
where K o is the constant for adsorption equilibrium at the infinity dilution and
H ads
is the isosteric heat of
adsorption at zero coverage.
2.3. CO2/N2 Breakthrough curves: Fixed-bed CO2/N2 adsorption The activated carbon was pretreated in a helium flow at 30 mL.min-1 and 150oC for 2 hours. The breakthrough curves were obtained by passing a flow of the standard gas mixture (20% CO2/N2 v/v) at different temperatures – 28, 50, 100 and 150oC. The total gas flow-rate was 30 mL.min-1 and it was controlled by a mass flow unit (Matheson, USA). A gas chromatograph, model CG35 (CG Instrumentos Científicos, Brazil), equipped with a Porapak-N packed column (Cromacon, Brazil) and a thermal conductivity detector (TCD) with helium as the reference gas, was used to monitor the carbon dioxide concentration at the bed exit. The column was located inside a furnace with controlled temperature. The experimental system (column and furnace) was considered adiabatic because it was isolated with a layer of 0.10 m of fiberglass and with a refractory material. The fixed-bed properties are given in Table 2.
Table 2: Properties of the fixed bed used in the CO2/N2 adsorption. Bed length, L
0.171 m
Bed diameter, d int
0.022 m
Bed weight, W
0.0352 kg
Bed voidage fraction,
0.52
Column wall thickness, l
0.0015m
Column wall specific heat, C p , w
440 J kg-1K-1
4
Wall density, w
7280 kg m-3
The adsorbed equilibrium concentration of carbon dioxide on the activated carbon was estimated as a function of the feed concentration from the mass balance in the fixed bed. For each experimental breakthrough curve, the adsorbed concentration was given by: q
C F QF t st C F V V 1
(3)
where C F is the feed concentration, V is the bed volume, QF is the feed volumetric flow rate and t st is the stoichiometric time (Ruthven, 1984).
3. Mathematical modeling of the fixed bed adsorption 3.1. Fixed-bed model The model used to describe the fixed-bed experiments is derived from the mass, energy and momentum balances. The flow pattern is described with the axially dispersed plug flow model and the mass transfer rate is represented by a Linear Driving Force model – LDF. It was assumed that the gas phase behaves as an ideal gas and the radial concentration and temperature gradients are negligible. Based on these assumptions, the fixedbed model is described as detailed below. The mass balance for each component is given by (Ruthven, 1984), Eq. (4):
Ci (uCi ) 2 Ci q DL (1 ) p i 2 t z t z
0
(4)
where ε is the bed void fraction, C i is the gas phase concentration of component i, q i is the average amount adsorbed of component i, D L is the axial dispersion coefficient, u is the superficial velocity, and p is the particle density. The rate of mass transfer to the particle for each component is given by Eq. (5):
qi K L,i (qi* qi ) t
(5)
where K L is the overall mass transfer coefficient of component i and q i* is the amount adsorbed at equilibrium of component i, i.e., qi* f (Ci ) given by the isotherm. The concentration Ci is given by: 5
Ci
yi P RTg
(ideal gas)
(6)
where yi is the molar fraction of each gas in the gas phase, P is the total pressure, Tg is the gas temperature and R is the universal gas constant. The Ergun equation considers the pressure drop and velocity changes:
g (1 ) 2 P (1 ) 150 u 1 . 75 gu2 2 3 3 z dp dp
(7)
where g is the gas viscosity, g is the gas density, and d p is the particle diameter.
The energy balance is (Ruthven, 1984):
CCv , g
Tg
CC p , g
t
(1 ) p H i i
(uTg )
L
2Tg
(1 ) p C s
z z qi 4hw (T Tw ) t d int g 2
Ts t (8)
where Cv , g is the molar specific heat at constant volume for the gas phase, C p , g is the molar specific heat at constant pressure for the gas phase, L is the effective axial thermal conductivity, C s is the solid specific heat,
H i
is the adsorption heat for the i component at zero coverage, hw is the coefficient of the internal
convective heat transfer between the gas and the column wall, d int is the bed diameter, and Tw is the wall temperature.
The solid phase energy balance is expressed by:
p Cs
Ts 6h f qi (Tg Ts ) p H i t dp t
(9)
where h f is the coefficient of the film heat transfer between the gas and the adsorbent. For the column wall, the energy balance for an adiabatic system can be expressed by:
6
wC p,w
Tw w hw (Tg Tw ) t
(10)
where w is the column wall density, C p , w is the column wall specific heat, w is the ratio of the internal surface area to the volume of the column wall, U is the external overall heat transfer coefficient, and T is the furnace external air temperature. The boundary conditions are as follows: z = 0: DL .
z = L:
Ci z
z = 0: L .
z = L:
Ci z
z
Ci
z
(11)
)
0
(12)
z
Tg z
Tg z
u (Ci z
uCC p , g (Tg z
z
Tg
z
)
0
(13)
(14)
z
z = 0: uC z uC z
(15)
The initial conditions for the adiabatic system are: t = 0: P P0 ; Tw Tg Ts Ti and Ci ( z,0) qi ( z,0) 0
(16)
The mathematical model was solved using the commercial software gPROMS (Process System Enterprise Limited, UK). The orthogonal collocation method on finite elements was used with six finite elements and three collocation points in each element of the adsorption bed.
3.2. Estimation of model parameters The LDF overall mass transfer coefficient – K L – considers all of the resistances to the mass transfer, i.e., intra- and extraparticle resistances. The value of the LDF mass transfer coefficient was estimated using the expression proposed by Farooq and Ruthven (1990), which considers film resistance, and macropore and micropore resistances:
7
rp qo rp 2 qo rc 2 1 K L 3k f C o 15 p D p,i C o 15Dc,i
(17)
where r p is the particle radius, k f is the film mass transfer coefficient, q 0 is the value of q (concentration on the solid phase) at equilibrium with C o (adsorbate concentration in the feed at feed temperature Ti ) and expressed in appropriate units, D p ,i is the macropore diffusivity of component i, P is the particle porosity, rc is the radius of activated carbon microparticles, and Dc,i is the micropore diffusivity of component i. The micropore diffusivity values were those reported by Cavenati and coworkes (2006) and it has been demonstrated previously that these micropore resistance values are suitable for use with this kind of adsorbent (Dantas et al., 2010). The axial dispersion coefficient and the film mass transfer coefficient were evaluated using the correlations of Leitão and Rodrigues (1995) and Seguin and coworkers (1996), respectively. The mass and heat transport parameters were estimated according to correlations reported in the literature (Dantas et al., 2010).
4. Results and Discussion 4.1. Carbon Dioxide and Nitrogen adsorption equilibrium isotherms The CO2 and N2 adsorption equilibrium isotherms are shown in Figures 1a and 1b, respectively. 6.0
q, 10 mol/g
(a)
(b)
4.0
-3
4.0
-3
q, 10 mol/g
6.0
2.0
0.0
0
1
2
3
P, bar
4
5
2.0
0.0
0
1
2
3
4
5
P, bar
Figure 1- Adsorption equilibrium isotherms for (a) carbon dioxide and (b) nitrogen on activated carbon at different temperatures: ∆ 28oC; □ 50oC; ○ 100oC and x 150oC. Solid lines: Toth Model.
8
The Toth model described quite well the experimental data and the parameters of the model are shown in Table 3. Carbon dioxide gas is preferentially adsorbed onto the activated carbon in comparison with nitrogen. The value for the isosteric heat of adsorption also indicates a higher affinity of CO2 than N2 for adsorption onto the activated carbon. However, the maximum capacity was almost the same for the two gases. This can be explained through the CO2 isotherm profile being more abrupt and the maximum capacity for this gas can be reached at lower pressures than those observed for nitrogen. The results shown are consistent with findings reported in the literature for studies at lower temperatures.
Table 3: Parameters of adsorption isotherms for carbon dioxide and nitrogen used for fitting of the Toth model Gas
qm, 10-3mol/g
n
Ko, 10-5bar-1
-∆Hads, KJ/mol
CO2
10.05
0.678
7.62
21.84
N2
9.74
0.518
6.91
16.31
4.2. CO2/N2 Breakthrough curves The mass balance in the columns after saturation of the adsorbent was performed for each component in the CO2/N2 mixture in order to evaluate possible competitive interactions between the adsorbates (Table 4). We observed that the amount of CO2 adsorbed in the dynamic tests is the same as that predicted by the single component Toth isotherm using the fitted parameters.
Table 4: Experimental conditions and adsorbed carbon dioxide concentrations predicted by the single component Toth model and observed experimentally through the mass balance of the breakthrough curves. T, oC
P, bar
q, 10-3 mol g-1(a) q, 10-3 mol g-1(b)
28
1.02
0.734
0.743
50
1.02
0.450
0.466
100
1.02
0.163
0.170
150
1.02
0.072
0.071
(a) Values calculated from the mass balance; (b) values calculated using Toth model parameters.
9
The amount of CO2 adsorbed onto the activated carbon from the CO2/N2 mixture (20%/80%) is the same as that adsorbed from pure carbon dioxide at the same partial pressure. This suggests that there are distinct active sites for N2 and CO2 adsorption. Moreover, considering that the nitrogen adsorption under the experimental conditions of the breakthrough curves is very low, the N2 adsorption can be neglected in this study. This assumption is in agreement with Delgado and coworkers (2006) who observed that the adsorption of nitrogen can be neglected when it is mixed with carbon dioxide. Figure 2 shows a comparison between the experimental and theoretical curves obtained for CO2 adsorption. It can be observed that the model reproduces the breakthrough curves very well. As expected, when the temperature is increased, the breakthrough curves are steeper due to the exothermic character of the adsorption. The LDF overall mass transfer coefficients used for the breakthrough curves can be found in Table 5. 1.0
C/Co
0.8 0.6 0.4 0.2 0.0
0
50
100
150
200
time, minutes
Figure 2- Breakthrough curves for CO2 adsorption onto activated carbon. Symbols: experimental data (∆ 28oC; □ 50oC; ○ 100oC and x 150oC). Solid lines: LDF model.
Table 5: LDF global mass transfer coefficients for CO2 adsorption onto activated carbon. T, oC
KL, s-1
28
0.0025
50
0.0042
100
0.0138
150
0.0323
10
5. Conclusions In this study, the preferential fixed-bed adsorption of carbon dioxide from CO2/N2 mixtures onto activated carbon was verified. The adsorption dynamics were determined at several temperatures and under different conditions. It was demonstrated that activated carbon adsorbed carbon dioxide to its total capacity, indicating that the equilibrium of CO2 and N2 adsorption from CO2/N2 mixtures can be very well described through the adsorption equilibrium behavior of the single components. A model based on the LDF approximation for the mass transfer was able to adequately describe the adsorption kinetics of carbon dioxide. The commercial activated carbon used in this study has high selectivity for CO2 and is suitable for CO2/N2 separation processes.
References Cavenati, S.; Grande, C.A.; Rodrigues, A.E. Separation CH4/CO2/N2 mixtures by Layered Pressure Swing Adsorption for upgrade of natural gases. Chemical Engineering Science, 61, p. 3893-3906, 2006. Dantas, T.L.P.; Amorim, S.; Luna, F. M. T.; Silva Jr., I.J.; de Azevedo, D.C.S.; Rodrigues, A.E.; Moreira, R.F.P.M. Adsorption of Carbon Dioxide onto Activated Carbon and Nitrogen-Enriched Activated Carbon: Surface Changes, Equilibrium, and Modeling of Fixed-Bed Adsorption. Separation Science Technology, 45, p.73-84, 2010. Delgado, J.A.; Uguina, M.A.; Gómez, J.M; Sotelo, J.L.; Ruíz, B. Fixed-bed adsorption of carbon dioxidehelium, nitrogen-helium and carbon dioxide-nitrogen mixtures onto silicalite pellets. Separation and Purification Technology, 49, p. 91-100, 2006. Drage, T.C.; Arenillas, A.; Smith, K.M.; Pevida, C.; Piipo, S.; Snape, C.E. Preparation of carbon dioxide adsorbents from the chemical activation of urea-formaldeyde and melamine-formaldeyde resins. Fuel, 86, p. 22-31, 2007. Farooq, S.; Ruthven, D.M. Heat Effects in Adsorption Column Dynamics. 2. Experimental Validation of the One-dimensional Model. Industrial & Engineering Chemical Research, 29, p. 1084-1090, 1990. Glover, T.G.; Dunne, K.I.; Davis, R.J.; LeVan, M.D. Carbon–silica composite adsorbent: Characterization and adsorption of light gases. Microporous and Mesoporous Materials, 111, p. 1-11, 2008. Gray, M.L.; Champagne, K.J.; Fauth, D.; Baltrus J.P.; Pennline, H. Performance of immobilized tertiary amine solid sorbents for the capture of carbon dioxide. Int. J. Greenhouse Gas Control, 2, p. 3-8, 2008.
11
Leitão, A.; Rodrigues, A.E. The simulation of solid-liquid adsorption in activated carbon columns using estimates of intraparticle kinetic parameters obtained from continuous stirred tank reactor experiments. Chemical Engineering Journal, 58, p. 239-244, 1995. Liu, X.; Zhoua, L.; Fu, X.; Sun, Y.; Su, W.; Zhou, Y. Adsorption and regeneration study of the mesoporous adsorbent SBA-15 adapted to the capture/separation of CO2 and CH4. Chemical Engineering Science, 62, p. 1101-1110, 2007. Maroto-Valer, M.M.; Fauth, D.J.; Kuchta, M.E.; Zhang, Y.; Andresen, J.M. Activation of magnesium rich minerals as carbonation feedstock materials for CO2 sequestration. Fuel Processing Technology, 86, p. 1627-1645, 2005. Ruthven, D. M. Principles of adsorption and adsorption processes, John Wiley & Sons, New York, 1984. Wong, S.; Bioletti, R. Carbon dioxide Separation Technologies. Carbon & Energy Management, Alberta Research Council, Edmonton, Alberta, Canadá, 2002. Yong, Z.; Mata, V.; Rodrigues, A.E. Adsorption of carbon dioxide at high temperature – a review. Separation and Purification Technology, 26, p. 195-205, 2002. Zhao, Z.; Cui, X.; Ma, J.; Li, R. Adsorption of carbon dioxide on alkali-modified zeolite 13X adsorbents. Int. J. Greenhouse Gas Control, 1, p. 355-359, 2007. Zhang, X.; Schubert, S.; Gruenewald, M.; Agar, D.W. Studies on the kinetics of carbon dioxide absorption with immobilised amines (IA). The Chemical Engineering Journal, 107, p. 97-102, 2005.
12
Installation and Operation of 0.5 MW-Scale Dry Sorbent CO2 Capture Pilot Plant Integrated with Real Coal-Fired Power Plant Chang-Keun Yi1†, Sung-Ho Jo1, Young Cheol Park1 and Chong Kul Ryu2 1
2
Korea Institute of Energy Research, Daejeon 305-343, Korea Korea Electric Power Research Institute, Daejeon 305-380, Korea ABSTRACT
Korea Institute of Energy Research (KIER) and Korea Electric Power Research Institute (KEPRI) have developed the dry sorbent carbon dioxide (CO2) capture system since 2002. The principle of this technology is the reversible reaction between potassium carbonate, CO2 and water vapor to form potassium bicarbonate in a thermal-swing process. Based on this reaction, we developed a bench scale unit (BSU) which treated 100 Nm3/h of flue gas in late 2006. The BSU consisted of a transport fluidizedbed carbonator, cyclones, a loop-seal, and a bubbling fluidized-bed regenerator. We have tested the BSU facility using a slip stream of the real flue gas from 2MW coal-fired circulating fluidized bed combustor located in KIER in order to check the performance of CO2 removal, the stability of operation and the possibility of deactivation of dry sorbent by contaminants such as SOx and NOx in the real flue gas. The results showed that more than 80% of CO2 removal had been maintained for more than 50 hour continuous operation. Those results indicated that the CO2 capture system using developed solid sorbents, which was supplied by KEPRI, could be applied to the real coal-fired power plant regardless of the presence of several contaminants. We have finished the detail design of pilot plant based on the long-term operation and the several tests. The pilot plant consists of a transport type carbonator and a bubbling type regenerator, which configuration is the same as the BSU facility. The pilot plant has been constructed at Hadong coal-fired power plant in Korea Southern Power Company. It can treat 2,000 Nm3/h of flue gas (0.5 MW scale). The construction has been done in late 2009 and the operation has been performed in 2010. The utilities such as cooling water for maintaining the carbonator temperature and steam for the regeneration energy have been supplied by the power company. It is expected that we can analyze the exact energy consumption of the dry sorbent CO2 capture system after the operation of the pilot plant. We plan to start the next scale-up in late 2011 up to 10 MW-scale which will be reflected in the operation results of a pilot scale unit. Key Words: Carbon dioxide capture; Dry sorbent; Fluidized bed reactor; Pilot plant; Power plant
INTRODUCTION It is expected that CO2 Capture and Storage (CCS) technology can reduce 19% of CO2 emissions in 2050 by BLUE Map scenario [1]. For CO2 capture, several methods such as wet absorption, adsorption, and membrane separation have been developed. As a breakthrough technology, Korea Institute of Energy Research (KIER) and Korea Electric Power Research Institute (KEPRI) have developed the dry sorbent carbon dioxide (CO2) capture system since 2002. KIER reported the experimental results of bench-scale unit (100 Nm3/h of gas treatment) using slip stream of 2 MW coal fired power facility in a previous Pittsburgh Coal Conference [2]. The pilot-scale unit had been constructed in late 2009 at Hadong coalfired power plant in Korea Southern Power Company. The pilot-scale unit was installed after Gas Gas Heater (GGH) and has been tested using a slip stream of 500 MW power facilities. †
To whom correspondence should be addressed. E-mail: [email protected]
PILOT-SCALE UNIT Fig. 1(a) showed the location of the pilot-scale unit and Fig. 1(b) was a real photo of the pilot-scale unit. In order to prevent a deactivation of dry sorbents by sulfur in the flue gas and to meet the sorption temperature, the pilot-scale unit was installed after Flue Gas Desulfurization (FGD) and Gas Gas Heater (GGH). The pilot-scale unit has been tested using K-based dry sorbents supplied by KEPRI. We first measured the solid circulation rate according to the valve opening area. We also checked the cyclone efficiency by 18-day continuous operation with solid circulation. The solid inventory was not changed due to the high efficiency of the cyclone. We controlled the temperature of carbonator and regenerator using cooling water and steam, respectively. By changing the flow rate of cooling water, the temperature of carbonator has been maintained at the target sorption temperature while that of regenerator maintained at the target temperature by changing the flow rate of steam. We have investigated the performance of the pilot-scale unit by continuous operation as well as found out the value of scale-up factor.
(a) (b) Fig. 1. The pilot-scale unit (2000 Nm3/h of gas treatment).
ACKNOWLEDGEMENT This research was performed for the Carbon Dioxide Reduction & Sequestration Center, one of 21st Century Frontier R&D Programs funded by the Ministry of Education, Science and Technology of Korea.
REFERENCES [1] International Energy Agency, Energy Technology Perspectives 2008. [2] Park Y C, Jo S, Lee S, Ryu C K, and Yi C, 2009, 26th PCC, USA.
2010 International Pittsburgh Coal Conference
REACTION CHARACTERISTICS OF WATER GAS SHIFT CATALYSTS FOR SEWGS PROCESS IN A BUBBLING FLUIDIZED BED Seung-Yong Lee*, Ho-Jung Ryu, Dowon Shun, Dal-Hee Bae Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA *[email protected] Abstract As a next generation hydrogen production technology from syngas with in-situ CO2 capture, SEWGS (Sorption Enhanced Water Gas Shift) process has been developing. In this paper, the best operating conditions of three WGS catalysts for SEWGS process have been investigated in a bubbling fluidized bed reactor. The commercial low temperature WGS catalyst (MDC-7) produced by Süd-chemie and new WGS catalysts (PC and RSM) produced by KEPRI (Korea Electric Power Research Institute) by means of spray-drying method were used. MDC-7 catalyst has pellet shape and we crushed the pellet to 106~212 m. However, PC and RSM catalysts have spherical shape and the same particle size range was prepared. The RSM catalyst was reformed using fine powder of MDC-7 catalyst. Reaction temperature, steam/CO ratio, and gas velocity were considered as experimental variables. Moreover, long-term operation results of WGS catalysts were compared as well. The best operating temperature and steam/CO ratio showed different results depend on the WGS catalysts. For MDC-7 catalyst, high CO conversion up to 99.4% was observed in the range of 220~240oC at 4.0 of steam/CO ratio. For RSM catalyst, 95% of CO conversion observed more than 250oC at 4.0 of steam/CO ratio. However, for PC catalyst, 90% of CO conversion achieved even at higher temperature (>350oC) at the same steam/CO ratio. The effect of steam/CO ratio on CO conversion showed different results for three WGS catalysts. For MDC-7 and RSM catalysts, CO conversion increased slightly as the steam/CO ratio increased up to 2.0, and maintained. However, CO conversion of PC catalyst increased continuously as the steam/CO ratio increased up to 5.0. The reactivity of MDC-7 catalyst was maintained more than 8 hours but that of PC catalyst decreased as the reaction time increased. As a conclusion, MDC-7 and RSM catalysts showed better reactivity and PC catalyst should be improved its reactivity. However, MDC-7 catalyst generated much fines during operation, and therefore, attrition resistance should be improved.
Introduction Hydrogen production is the most fundamental part of the hydrogen energy system, and has always been the object of intense and vigorous research and development. World hydrogen production has been growing rapidly at 8-10% per annum for many years [1]. At present, hydrogen is produced mainly from fossil fuels, water and biomass. However, more than 90% of the hydrogen is produced from fossil fuels [2]. Series of gasification of coal, water gas shift, and CO2 separation is the predominant production route to hydrogen from coal for commercial scale application. However, this process needs multiple-steps such as high- and low-temperature water gas shift reaction as shown in eq. (1) to improve hydrogen yield, and CO2 separation process to separate almost pure hydrogen from the gas mixture of CO, CO2, and H2. To separate CO2 from the exhaust gas, additional energy and equipments are required. More than 22% of hydrogen generation cost comes from CO2 separation process for purifying hydrogen[3]. Although the
previous process has been used for many years, there are some areas for improvement. The previous process requires many reactors and many kinds of catalysts and/or sorbents. Therefore, it will be extremely desirable if new concepts can be developed which can reduce the capital and operating cost of the conventional process[4]. To overcome these disadvantages, SEWGS (Sorption Enhanced Water Gas Shift) system has been developed. Equation (2) and (3) explain concept of SEWGS system. The thermodynamic equilibrium in the shift reaction can be enhanced to give more hydrogen yield by adding a CO2 absorbent into the shift reactor. Carbon dioxide is then captured as a solid carbonate as soon as it formed, shifting the reversible water-gas shift reactions beyond their conventional thermodynamic limits as shown in eq. (2). Regeneration of the sorbent releases pure CO2 suitable for sequestration as shown in eq. (3). It is important that the gas composition of the exhaust gas from the SEWGS reactor contains only highly concentrated H2 and excess water vapor. Therefore, H2 can be easily recovered by cooling the exhaust gas without any extra energy consumption for H2 separation. Moreover, the exhaust gas from the regeneration reactor contains only carbon dioxide and water vapor if we use steam as fluidization gas. After water condensation, almost pure carbon dioxide can be obtained with little energy loss for component separation. Water gas shift reaction CO + H2O H2 + CO2 SEWGS reaction CO + H2O + MO H2 + MCO3 Regeneration reaction MCO3 MO + CO2 where, MO: metal oxide, MCO3: metal carbonate
(1) (2) (3)
In this study, the reaction characteristics of three WGS catalysts for SEWGS have been investigated in a bubbling fluidized bed reactor, as a preliminary research. The commercial low temperature WGS catalyst (MDC-7) produced by Süd-chemie and new WGS catalysts (PC and RSM) produced by KEPRI (Korea Electric Power Research Institute) by means of spray-drying were used as bed materials. Reaction temperature, steam/CO ratio, and gas velocity were considered as experimental variables. Moreover, long-term operation results of WGS catalysts were compared as well.
Experimental The reactivity tests were carried out in a bubbling fluidized bed reactor. A schematic of the reactor is provided in Figure 1. The major components consist of a gas input system, a fluidized bed, a condenser, a hot gas filter, a gas sampling/analyzing unit, and water feeding pump. The fluidization column is 0.7 m high with an internal diameter of 0.05 m. A perforated gas distributor plate separates the fluidization column and air box. Reactant gas was fed to the air box. An electric heater could be controlled by a thermocouple and a heater controller. Temperature and pressure data were recorded by a data acquisition system. The exit stream from the fluidized bed reactor was sampled at the outlet of the reactor. The CH4, CO, CO2, H2, NO, and O2 concentrations were monitored using an on-line gas analyzer and recorded by a data acquisition system. Further details of the reactor system are available in our previous paper[5]. Three water gas shift catalysts, the commercial low temperature WGS catalyst (MDC-7) produced by Süd-chemie and new WGS catalyst (PC, RSM) produced by KEPRI (Korea Electric Power Research Institute) were used. MDC-7 catalyst had pellet shape and we crushed the pellets to 106~212 m. However, PC and RSM catalyst has spherical shape and the same particle size range was prepared. Figure 2 shows photos of three WGS catalysts. The PC and RSM catalysts show spherical shape and the MDC-7 catalyst shows irregular shape. The static bed height was 0.4 m in all cases, and initial solid masses were 0.57 kg for PC catalyst (b=724.7 kg/m3), 0.64 kg for RSM(b=747.3 kg/m3), and 0.88 kg for MDC-7 catalyst (b=1117 kg/m3), respectively. The fluidized bed reactor operated with a total inlet gas flow of 2.0 Nl/min in all cases, except for tests to check effect of gas velocity. The total inlet gas contained 50% of syngas and 50% of nitrogen. The syngas composition was 60.5% of CO, 27.2% of H2, 9.9% of CO2 and N2 as a balance.
Figure 1. Schematic of a bubbling fluidized bed reactor.
Figure 2. Photos of PC, RSM and MDC-7 catalysts. Results and discussion Prior to the start of each experiment, catalysts were reduced by H2 gas (57%, N2 balance) at 400oC. Figure 3 shows breakthrough curves of hydrogen concentrations during pretreatment (reduction) of catalysts. A breakthrough of hydrogen concentration marked the end of reduction. As shown in Figure 3, since MDC-7 and RSM catalysts showed sharper breakthrough curve than PC catalyst, and therefore, we could expect that the MDC-7 and RSM catalysts would show better reactivity than PC catalyst.
70 PC RSM (reformed) MDC-7
H2 concentration [%]
60 50 40 30 20 10 0 0
20
40
60
80
Time [min]
Figure 3. H2 breakthrough curves during pretreatment of catalysts. Figure 4 shows effect of temperature on CO conversion of catalysts. The CO conversion to H2 and CO conversion to CH4 were calculated by mass balance based on the output gas concentration and the tie component (N2). For MDC-7 catalyst, high CO conversion up to 99.4% was observed in the range of 220~240oC at 4.0 of steam/CO ratio. For RSM catalyst, lower CO conversion than MDC-7 catalyst was observed. Moreover, for PC catalyst, lower CO conversion than MDC-7 catalyst was observed even at higher temperature (380~400oC).
100
CO conversion [%]
90
100
PC catalyst
90
100
RSM
MDC-7
90
80
80
80
70
70
70
60
60
60
50
50
50
40
40
40
30
30
30
20
20
20
Total CO conversion CO to H2 conversion
10
10
10
CO to CH4 conversion
0 200
250
300
350
400 o
Temperature [ C]
Steam/CO ratio = 4
0 0 450 140 160 180 200 220 240 260 280 300 320 140 160 180 200 220 240 260 280 300 320
Temperature [oC]
Temperature [oC]
Figure 4. CO conversion versus reaction temperature. Figure 5 shows effect of steam/CO ratio on CO conversion of two catalysts. For MDC-7 catalyst, CO conversion increased slightly as the steam/CO ratio increased up to 2.0, and maintained at high level thereafter. For RSM catalyst, CO conversion increased as the steam/CO ratio increased up to 2.0, and maintained thereafter. However, CO conversion of PC catalyst increased continuously as the steam/CO ratio increased up to 5.0. MDC-7 catalyst showed higher CO conversion at the same steam/CO ratio and at lower temperature. Moreover, PC catalyst generated higher CH4 at lower steam/CO ratio.
CO conversion [%]
100
100
100
90
90
90
80
80
80
70
70
70
60
60
PC series o 360 C
50
60
RSM 200 oC
50
MDC-7 200 oC
50
40
40
40
30
30
30
20
20
20
Total CO conversion CO to H2 conversion
10
10
10
CO to CH4 conversion
0
0 1
2
3
4
5
6
0 1
2
Steam/CO ratio [-]
3
4
5
6
1
Steam/CO ratio [-]
2
3
4
5
Steam/CO ratio [-]
Figure 5. CO conversion versus steam/CO ratio. Figure 6 shows long-term test results of two catalysts. The reactivity of MDC-7 catalyst was maintained more than 8 hours but that of PC catalyst decreased as the reaction time increased. As a conclusion, MDC-7 catalyst showed better reactivity than PC and RSM catalysts from the viewpoints of reaction temperature, seam/CO ratio, CO conversion, and long-term durability. 100 90
CO conversion [%]
100
gas analyzer trouble
90
80
80
70
70
60
60
50
50 (a) PC catalyst at 380oC steam/CO ratio = 5
40 30
(b) MDC-7 catalyst o at 200 C steam/CO ratio = 4
40 30
20
Total CO conversion CO to H2 conversion
20
Total CO conversion CO to H2 conversion
10
CO to CH4 conversion
10
CO to CH4 conversion
0 0
100
200
300
Time [min]
400
0 500 0
100
200
300
400
500
Time [min]
Figure 6. CO conversion versus time. To check effects of syngas concentration and gas velocity, supplementary tests were performed using MDC-7 catalyst in the same reactor and the results are provided in Figure 7. The CO conversion of MDC7 catalyst decreased slightly as the syngas concentration increased, but increased as the gas velocity decreased and steam/CO ration increased. However, these values are much higher than the results from the fixed bed with the same catalyst[6].
100 99
CO conversion [%]
98 97 96 95 94
Steam/CO ratio=2, U=0.05 m/s Steam/CO ratio=4, U=0.05 m/s Steam/CO ratio=4, U=0.025 m/s Steam/CO ratio=3, U=0.078 m/s Steam/CO ratio=3, U=0.039 m/s
93 92 91 90 0
20
40
60
80
100
Syngas concentration [%]
Figure 7. CO conversion to H2 versus syngas concentration (MDC-7 catalyst). Conclusion The best operating temperature and steam/CO ratio showed different results depend on the WGS catalysts. For MDC-7 catalyst, high CO conversion up to 99.4% was observed in the range of 220~240oC at 4.0 of steam/CO ratio. However, for PC catalyst, 90% of CO conversion achieved even at higher temperature (>350oC) at the same steam/CO ratio. For MDC-7 and RSM catalysts, CO conversion increased slightly as the steam/CO ratio increased up to 2.0, and maintained thereafter. However, CO conversion of PC catalyst increased continuously as the steam/CO ratio increased up to 5.0. The reactivity of MDC-7 catalyst was maintained more than 8 hours but that of PC catalyst decreased as the reaction time increased. For RSM catalyst showed better reactivity than PC catalyst but worse reactivity than MDC-7 catalyst. As a conclusion, MDC-7 catalyst showed better reactivity than PC and RSM catalysts from the viewpoints of reaction temperature, seam/CO ratio, CO conversion, and long-term durability. However, MDC-7 catalyst generated much fines during operation, and therefore, attrition resistance should be improved. Acknowledgement This work was supported by the Energy Efficiency and Resources program of the Korea Institute of Energy Technology Evaluation and Planning(KETEP) grant funded by the Korea government Ministry of Knowledge Economy References 1. Kothari, R., Buddhi, D. and Sawhney, R. L. (2004). “Sources and Technology for Hydrogen Production: a Review”, Int. J. Global Energy Issues, 21(1/2), 154-178. 2. IEA report (2005). “Prospects for Hydrogen and Fuel Cells”, IEA Books, 49-55. 3. Maurstad O. (2008), “An Overview of Coal Based Integrated Gasification Combined Cycle (IGCC)”, MIT report, Publication No. LFEE 2005-002 WP, 1-43. 4. Ryu, H. J. (2009). “Selection of Process Configuration and Operating Conditions for SEWGS System”, Trans. of the Korean Hydrogen and New Energy Society, 20(2), 168-178. 5. Ryu, H. J., Shun, D., Bae, D. H., and Park, M. H. (2009). “Syngas Combustion Characteristics of Four Oxygen Carrier Particles for Chemical Looping Combustion in a Batch Fluidized Bed Reactor”, Korean J. of Chem. Eng., 26(2), 523-527. 6. Chen, W. S. and Jheng, J. G. (2007). “Characterization of Water Gas Shift Reaction in Association with Carbon Dioxide Sequestration”, J. of Power Sources, 172, 368-375.
CARBON DIOXIDE CAPTURE OF FLUE GASES FROM COAL-FIRED POWER PLANT USING ENZYMES ORIGINATED MARINE LIFE Soonkwan Jeonga, Kyungsoo Lima, Jeongwhan Lima, Daehoon Kimb, Mari Vinobaa,c, S.H. Leea,* a
Climate Change Technology Division, Korea Institute of Energy Research 102, Gajeong-ro, Yuseong-gu, Daejeon, 305-343, Korea b Department of Chemical and Biological Engineering, Korea University, Seoul 136-713, Korea c Department of Chemical Engineering, Anna University-Chennai, Chennai 600 025, India
*E-mail: [email protected]
ABSTRACT The focus of this study is the separation and storage of green house gas, CO2, and the use of enzymes from marine life in development of technology to provide novel method for CO2 capture. Carbonic anhydrase has recently been used as a biocatalyst to accelerate an aqueous processing route to carbonate formation. In this study, we compared soluble proteins of HDS (Hemocyte from Diseased Shell) and EPF (Extrapallial Fluid) extracted from crassostrea gigas with HCA (Human Carbonic Anhydrase) and BCA (Bovine Carbonic Anhydrase) on their ability to promote CO2 hydration and the production of calcium precipitates. HCA, BCA, HDS, and EPF have shown promising results for use as promoter to accelerate CO2 hydration and increase the rate of precipitation of carbonate mineral with Ca2+ ions. The ideal temperature and pH of operation was found to be 40℃ and 6~7, respectively. Lineweaver-Burk relationship was employed to estimate Michaelis-Menten kinetic parameters for the enzymes. BCA showed the fastest rate constant and followed HCA, HDS and EPF. In previous efforts to use CO2 mineralization as a method for CO2 sequestration the slow rate of hydration of CO2 to carbonic acid has been limiting factor of CO2 mineralization. In the presence of enzymes, rate determining step is eliminated, and therefore the overall reaction rate is enhanced dramatically. Precipitated CaCO3 was all calcite and the particles size was below 100nm. These nano-particles could be use in other industrial processes such as paper, ink, or building materials. Therefore it can be reduced the operating cost of CO2 capture, which increase feasibility to install CO2 capture process in coal-fired power plant. This result suggests that enzymes mentioned above may be involved not only in CO2 hydration but also in CO2 mineralization.
1. INTRODUCTION The contribution of greenhouse gases, especially carbon dioxide, to global warming is well recognized [1]. An energy source option for the humankind energy system has become too numerous, and the thirst for more and more energy targets on usage of fossil fuels in coal-fired power plants, steel industries, and cement contributes to an increasing amount of CO2 emission. The flue gases from conventional coal-fired power plants in Korea typically contain 12~14% CO2 with the balance consisting mainly of nitrogen and small quantities of oxygen and impurities. The carbon dioxide concentration of flue gases is insufficient for direct compression
to transpport and stoorage of carbbon dioxidee. In order to t get conceentrated CO2 from flue gases, we have to install CO O2 capture pprocess suchh as amine or ammonnia absorptioon. The am mine-based solventss are widely used for the CO2 ccapture proccess. Despiite of the ggood CO2 aabsorption capacityy and fast absorption ab raate, they haave some drrawbacks suuch as corroosion, loss oof solvent, generatiion of heat stable salts,, and high ppenalty for eenergy conssumption duuring regeneeration [24]. If ann alternativee solvent syystem havingg less energgy for regenneration couuld be develloped, it is very atttractive opption to innstall CO2 capture faacility. Sincce recentlyy, emergingg ex-vivo applicattions of carbbonic anhyddrase for its potential usse in CO2 caapture technnologies are attracting attentionns [5-6]. The obbjective of tthe present study was tto investigatte the feasibbility of usinng enzymess extracted from maarine life ass a biocatalyyst for hydrration of CO O2, as well as a its precippitation in thhe form of calcium m carbonate. We have bbeen investigated a kineetic analysiss and the eff ffects of conncentration of enzyyme, pH, teemperature oon hydratioon of CO2. In additionn, an enzym matic precippitation of calcium m carbonate was w studiedd.
2. EX XPERIMENTAL O2 capture aand sequesttration procedures 2-1. CO The eexperiment w was conducteed in a glass reactor (1 L)) with a conttrol system, aas shown in F Fig. 1. The reaction temperature was controllled by a wateer bath (JEIO OTECH CW-10G) and C CO2 was introoduced into the bottoom of the reactor. The sttirring rate and a temperatture in the reeactor were ccontrolled byy a control system, and the pH was measurred using a ppH meter. C CaCl2 (Sigmaa C1016) waas used as thhe calcium source. T The bovine ccarbonic anhhydrase II ussed in this stuudy was obttained from S Sigma (St. L Louis, MO, U.S.A.). HDS and EP PF were extraacted from Crassostrea C ggigas (C. gigaas). microscope The miicrostructure of precipitattes was studied using a JJEOL JSM-8840A scanninng electron m with AN N-10000/85S (LINK systeem) energy diispersive speectrometer (E EDS) capabiliity.
Figure1. Experimenttal illustratioon for CO2 sequestratioon using byy enzyme. T The experim ments were approxim mately divideed into two parts (dotteed line), reacctor part (a)) (i: reactor, ii: electric stirrer, iii: thermal couple, iv: ppH meter, v: CO2 line) annd control syystem part (bb). (i: controller of tempeerature and
stirring rate, ii: flow mass). And the other parts were equipped with computer (c), water bath (d), and CO2 gas (e).
2-2. Separation and purification of enzymes The diseased area in oyster shell, Crassostrea gigas, was defined in previous study. The diseased area was separated using a knife and ground to a fine powder in an electric mill. The powder (25 g) was extracted in 150 ml of 0.2 M EDTA (pH 8.0) at 4 oC with continuous stirring. After 2 days the soluble extract was obtained by centrifugation at 25,000 × g for 20 min. The supernatant was diluted with an equal volume of deionized water. The diluted solution was concentrated by ultrafiltration using a minimodule. The concentrated fraction was dialyzed against 1 L of deionized water at 4 C for 3 days. The dialyzed fraction was concentrated to a volume of 20 ml and subjected to ethanol precipitation at 25 C for 1 week. The precipitate was dissolved with continuous stirring in 3 ml of 50 mM NaHPO2·H2O (pH 7.18) at 5 C, and dialyzed for 3 days against deionized water at 4 C. The dialyzed solution was lyophilized. The soluble protein extracted from the diseased area was termed HDS (hemocyte from diseased shell). Specimens of C. gigas were collected (Namhae, Korea) and the EPF was removed immediately by prying the valves open, inserting the needle of a sterile hypodermic syringe under the mantle edge into the central extrapallial space, and withdrawing the fluid. The EPF samples were maintained on ice in a refrigerator at 4 C.
3. RESULTS and DISCUSSION Fig. 2 demonstrates the pH change at room temperature with buffer and without buffer system. Carbonic anhydrase catalyzes the hydration of CO2 and consequently hydrogen ions are generated. This results in a change of pH. Therefore, measuring pH is a viable method to monitor the progress of this enzymatic reaction [6]. Addition of CO2 into the reactor at time zero tended to drop immediately without buffer solution, but the presence of buffer solution pH value leveled off. BCA was widely used as a biocatalyst for hydration of CO2 and showed very fast reaction rate. HDS and EPF extracted from C. gigas showed almost same activity. This result suggests that HDS and EPF can apply a biocatalyst for CO2 capture. 7.0
7.2 o
Temp : 25 C
6.5
7.0
Blank(Buffer) BCA (Buffer) Blank(Water) BCA (Water)
5.5
6.8
pH
6.0
pH
blank BCA HDS EPF
6.6
5.0
6.4
4.5
6.2
4.0
6.0
0
200
400
Time (sec)
600
800
0
200
400
600
Time (sec)
Figure2. pH experiments at T = 25oC with or without buffer.
800
Though enzyme shows high catalytic activity for hydration of CO2, it has certain limitation in its application due to the short lifetime of enzyme. There are different methods to improve catalytic stability of enzyme such as enzyme immobilization, enzyme modification and genetic modification. Among the different methods, immobilization of the enzyme into the solid support is used because of its proper applications. We made several kinds of solid supports such as SBA15, SBA-16, and KIT-6. Pore expanded SBA-15 that has long chain pore shows high activity. SBA-15 has a 2D hexagonal structure and 14nm of pore diameter. Enzyme was immobilized over large pore expended meso-porous silica through cross linked enzyme aggregation with covalent bonding. Fig. 3 shows comparison of free and immobilized enzyme on hydration of CO2. Bio-catalytic activity of free enzyme demonstrates slightly high due to loss of enzyme during immobilization. The catalytic activity of immobilized enzyme did not change during cyclic experiment. The stability of immobilized enzyme indicated the enzyme attached as cross linked aggregates with covalent bond within the porous silica did not lose its activity by either denaturation or detachment from the structure. 7.0
7.0
CO2 in
CO2 in
CO2 in
Immoblized BCA
Free BCA 6.8
6.8
6.6
pH
pH
6.6
6.4
6.4
6.2
6.2
0
100
200
300
Time(sec)
400
500
600
CO2 out N2 in
CO2 out N2 in
6.0
6.0 0
2000
4000
BCA Immobilization 6000
8000
Time(sec)
Figure3. Comparison of the catalytic activity for free and immobilized enzyme. Determination of Km and Kcat values of enzyme was carried out by measuring the activity in the presence of p-nitrophenyl acetate. Km and Kcat were determined using the Lineweaver-Burk double reciprocal plot, in which the reciprocals of the initial velocities of the enzyme activity were plotted against the reciprocals of the concentration of p-nitro phenyl acetate used. Km and Kcat value of BCA were 16.8mM and 2.1s-1 with p-NPA as substrate. However, Km and Kcat value of EPF were 8.8mM and 0.1s-1. HDS and EPF are soluble biopolymer. They consist of active site for hydration of CO2 and inert site, and then their activities are lower than those of BCA. As bivalve shells are composed mainly of calcium carbonate, they provide a good model for the study of CO2 sequestration because the shell is derived from the calcium ions and CO2 in seawater. In the precipitation reaction, the onset time for precipitation with enzyme is approximately 4 times faster as compared to the carbonation reaction in the absence of enzyme. It means that enzyme accelerates the formation of bicarbonate or carbonates ions. Precipitated calcium carbonate has calcite structure by XRD and size of it is below 100nm. In previous efforts to use CO2 mineralization as a method for CO2 sequestration, the slow rate of hydration of CO2 has been limiting factor of CO2 mineralization [7]. A calcium precipitate was identified in experiments using enzyme such as BCA, HCA, HDS and EPF. This result suggests that all
enzymes in this study may be involved not only in CO2 hydration but also in CO2 mineral carbonate conversion. Calcium carbonate is a common and thermodynamically stable mineral found in rocks worldwide, and is the main component of shell of marine organism, snails and eggs. If the widespread transformation of CO2 to CaCO3 is possible, it will represent a stable process for long-term CO2 capture and storage. The study presented here is only the beginning for the capture and storage of CO2 using enzyme, especially HDS and EPF. Additional studies on various conditions which include cloning of enzyme, operating conditions, and scale-up factor are underway.
COMCLUSIONS Enzyme extracted from marine life has been successfully applied for the hydration and mineralization of carbon dioxide. HDS and EPF accelerate the rate of CO2 hydration and mineralization. Enzyme immobilization over pore expanded silica enhances the stability of enzyme as compared to free enzyme. Kinetic parameters of enzyme were also evaluated from Lineweaver-Burk plot. The Km was 16.8mM and Kcat was 2.1s-1 for BCA, whereas of the EPF Km was 8.8mM and Kcat was 0.1s-1. Therefore, HDS and EPF need to purify to improve catalytic activity. HCA, BCA, HDS, and EPF have shown promising results for use as promoter to accelerate CO2 hydration and increase the rate of precipitation of carbonate mineral with Ca2+ ions.
Reference 1. IPCC, Special Report on Carbon Dioxide Capture & Storage, (2005). 2. B.P. Mandal, M. Guha, A.K. Biswas, S.S. Bandyopadhyay, Chem. Eng. Sci., 56 (2001) 6217. 3. P.Y. Chung, A.N. Soriano, R.B. Leron, M.H. Li, J. of Chemical Thermodaynamics, 42 (2010) 802. 4. G.Gao, G.H. Liang, H. Wang, Corrosion Science, 49 (2007) 1833. 5. G.M. Bond, J. Stringer, D.K. Brandvold, F.A. Simsek, M.G. Medina, G.E. Egeland, Energy and Fuels 15 (2001) 309. 6. P. Mirjafari, K. Asghari, N. Mahinpey, Ind. Eng. Chem. Res., 46 (2007) 921. 7. S.W. Lee, S.B. Park, S.K. Jeong, K.S. Lim, S.H. Lee, M.C. Trachtenberg, Micron, 42 (2010) 273.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
CARBON MANAGEMENT PREPARATION AND CHARACTERISTICS OF FORMED ACTIVE CARBONS FOR NATURAL GAS STORAGE Grzegorz Łabojko, Ph.D., Institute for Chemical Processing of Coal Zamkowa stree,t 1, 41-803 Zabrze, Poland [email protected] Tel: + 48 32 271-00-41 int. 133, Fax. + 48 32 271-08-09
Aleksander Sobolewski, Ph.D., Institute for Chemical Processing of Coal Zamkowa street, 1, 41-803 Zabrze, Poland [email protected] Tel: + 48 32 271-00-41 int. 220, Fax. + 48 32 271-08-09 Leszek Czepirski, prof. , AGH – University of Science and Technology, Faculty of Fuels and Energy, al. Mickiewicza 30, 30-059 Cracow, Poland [email protected] Tel.+48 12 617-46-36, 617-20-66; Fax. +48 12 617-45-47
Abstract: Natural gas as an automotive fuel presents various advantages versus petrol and gasoline including reduced vehicle emissions, lower maintenance and savings in fuel costs. Main problem in using natural gas as fuel for vehicles is low volumetric energy density of methane at room temperature. Compressed Natural Gas (CNG) is a solution used worldwide (more than 4 millions vehicles), but heavy and expensive cylinders have to be used for its storage at pressures about 200 bar. Adsorbed Natural Gas (ANG) can overcome issues with heavy storage systems by lowering operating pressure down to 35-40 bar. Key to the success of adsorptive storage is the choice of suitable adsorbent and operating conditions. Porous carbonaceous materials are well known as one of the best adsorbents for gases. Preparation procedure of carbonaceous adsorbents for natural gas storage is described. Based on physicochemical parameters Polish hard coal type 34.1 from KWK Marcel was selected as precursor material for adsorbents for gas storage. Four grindings were performed on selected coal and particle size distribution was determined. Fullers distribution was calculated as providing best packing of precursor material. Comparison between optimal and obtained distributions allowed to select the best way of grinding coal providing best packing of precursor material. Thermo prepared coal tar and phenol-formaldehyde resol type resin were selected as binders, thermoreological study of binders and coal-binder mixtures were performed. Granules and pastilles from coal-binders mixtures were prepared and pyrolysed. Influence of binder amount in mixtures on mechanical resistance of granules were studied. Influence of steam activation time of granules on sorption parameters (Iodine number) of obtained active carbon monoliths was studied. High pressure adsorption of methane and low pressure adsorption of nitrogen was performed on selected samples of obtained active carbons monoliths.
The kinetics of the CO2 reforming of CH4 over carbonaceous catalyst Fengbo Guo, Yongfa Zhang* , Guojie Zhang, Bingmo Zhang Key Laboratory of Coal Science and Technology of Shanxi Province and the Ministry of Education ,Taiyuan University of Technology , Taiyuan 030024 , China Abstract: The CO2 reforming of CH4 on carbonaceous catalyst was performed in a Plug Flow Reactor (PFR) at temperatures range from 1223 K to 1323K, the ratio of CH4/CO2=1 and residence time 3 ~ 30s under normal pressure. The outlet gas was analyzed by Gas Chromatogram (GC-960TCD and GC-950TCD), the carbonaceous catalyst was analyzed with element analyzer and specific surface area analyzer. The experimental results show that the conversion of CH4 and CO2 increases with the increase of the reaction temperature and residence time. The reaction temperature remains the chief influential factors on the conversion of CH4 and CO2. Under the conditions of the temperature of 1323K and residence time over 20s, the conversion of CH4 and CO2 can be expected over 90%. In addition, the conversion of CO2 was significantly higher than the conversion of CH4, which indicate that the gasification reaction of the carbonaceous catalyst and carbon dioxide was occurred during the reforming process. A mechanism of the CO2 reforming of CH4 on carbonaceous catalyst has been proposed based on the experimental results. Based on the mechanism, a kinetic model was developed. The kinetics of the CO2 reforming of CH4 on carbonaceous catalyst was described in a rate law, the experimental data was analysed in non-linear regression. The apparent activation energy Ea and the pre-exponential factor A were solved. The rate constant was as asfollows: k
= 1765
exp (
−
109.9
± 10.1KJ RT
)
A comparison is made between calculation data and experimental data of the CH4 conversion, which illustrates the rationality of the kinetic model. Keywords:
Kinetic; Carbon dioxide reforming of methane; carbonaceous catalyst
1. Introduction Methane is the main composition of natural gas , coal bed gas and coke oven gas. The CO2 reforming of CH4 provides a practical method for effectively combines CH4 utilization with CO2
*Corresponding author. Tel. : +86 0351 6018676 E-mail address: [email protected]
transforming, since both methane and carbon dioxide are the main greenhouse gas, and the reaction gets raw gases for Fischer-Tropsch, methanol and carbonyl synthesis [1-3]. In recent years, the CO2 reforming of CH4 has received considerable attention word wide. There are four major kinds of catalysts, which are transitional metals catalysts, Composite metal oxide catalysts, supported metal sulfide catalysts and supported noble metal catalysts[4-6], but carbonaceous catalyst has attracted the attention of most researchers who are studied CO2 reforming of CH4 reaction , carbonaceous catalyst is both cheap and abundantly available in the word, and which cannot lead to catalyst poisoning . Li Yan-bing[7] studied the influence of different chars and operation parameters on reactant gas conversion, and the results show that Tongchuan char with the lowest ash concentration exhibits the highest activity, and all the char samples show similar behaviors. A higher conversion is always achieved at the beginning, and then it decreases and levels off at a stable value after 30 min. Guojie Zhang [8] studied the effects of the coal char catalyst pretreatment and the ratio of CO2/CH4, experimental results showed that the coal char was an effective catalyst for production of syngas, and the product gas ratio of H2/CO is strongly influenced by the feed ratio of CO2/CH4. Zhang Hua wei[9] studied the effect of coke on the steam and carbon dioxide reform ing reaction s of methane at 700℃-1300℃, the results show that the conversion of methane to produce syngas obviously increased with the present of coke in the reactor. In the present work, the CO2 reforming of CH4 over the DaTong carbonaceous catalyst was performed in a Plug Flow Reactor. The effect of reaction parameters on the carbonaceous catalyst activity for CO2 reforming of CH4 were studied , and the kinetic behavior was investigated as functions of temperature and residence time .
2. Experimental 2.1 Experimental Equipment The CO2 reforming of CH4 over the DaTong carbonaceous catalyst was performed in a Plug Flow Reactor at temperatures range from 1223 K to 1323K,
the ratio of CH4/CO2=1 and residence
time 3~30s under normal pressure. The reactor (internal diameter 20mm; length 100mm)was horizontal heated in a furnace , A weighed amount (15g) of carbonaceous catalyst was loaded in the middle of the reactor . The CH4 and CO2 gas flow rates were measured and controlled by Mass flow
controllers (flow control range 0~200mL/min).The outlet gas was analyzed by Gas Chromatogram (GC-960TCD and GC-950TCD), the carbonaceous catalyst was analyzed with element analyzer and specific surface area analyzer. The schematic diagram of the Experimental equipment is given in Fig. 1.
Fig. 1 Schematic diagram of the experiment 1.2. cylinder 3. N2 cylinder 4. Mass flow controllers 5. quartz reactor6. carbonaceous catalyst 7. Temperature controller 8. Heat oven 9. Sample sampler 10. Gas chromatograph
2.2 Catalyst preparation and characterization The carbonaceous catalyst was prepared by pyrolysis of Da tong coal at 800℃ for 2h , crushing the catalyst mass to 80~100 mesh-size particles . The carbonaceous catalyst was dried in muffle furnace for 12h at 900℃ in nitrogen . The Proximate analysis and ultimate analysis of carbonaceous catalyst were shown in table 1. Table 1 The Proximate analysis and ultimate analysis of carbonaceous catalyst proximate analysis/ w %, ad
Ultimate analysis/ w %, daf
Sample M
A
V
C
H
N
S
O(diff)
DT Coal
3.10
12.20
29.00
87.70
4.96
1.27
0.42
5.36
C- catalyst
1.20
13.30
4.50
94.60
1.47
0.99
0.17
2.36
3. Result and discussion 3.1 Effect of residence time and reaction temperature on CH4 and CO2 conversion The carbonaceous catalyst activity was investigated at different temperatures range from 1223
K to 1323K with CH4/CO2=1. The residence time was varied by changing the feed gases flow. The effect of residence time on CH4 and CO2 conversion at different temperatures is shown in Fig. 2. 100
100
a
90
CO2 Conversion (%)
CH4 Conversion (%)
90
1223K 1273K 1323K
b
1223K 1273K 1323K
80
70
60
50
80
70
60
40 50 0
5
10
15
20
25
30
35
0
5
10
Residence time (s)
Fig.2
15
20
25
30
35
Residence time(s)
Effect of residence time and reaction temperature on CH4 and CO2 conversion
As can be seen from Figure 2, with the increase of residence time the conversion of CH4 and CO2 have increased rapidly, then leveled off. When the residence time from 3s to 30s in reaction temperature 1223K, the conversion of CH4 can be increased from 40% to 83% , but the conversion of CO2 can be increased from 60% to 93% . The result showed that the conversion of CO2 was significantly higher than the conversion of CH4 at the same condition of residence time and reaction temperature, which indicate that the gasification reaction of the carbonaceous catalyst and carbon dioxide was occurred during the reforming process. Increase temperature is in favor of CH4-CO2 reforming reaction, which promotes the conversion of CH4 and CO2.When the reaction temperature arrived 1323K under the residence time 30s, the conversion of CH4 and CO2 are about 92.8% and 98%. And the temperature is main factor affecting reaction rate 2.2.
Mechanism The CO2 reforming of CH4 on carbonaceous catalyst is not a independent reaction, duing
reforming of CO2 - CH4 the following reaction may take place The Carbon dioxide reforming of methane Carbon dioxide gasification
C + CO 2 → 2 CO
Methane decomposition
CH
4
CH
→ C + 2H
4
2
+ CO
2
→ 2 CO + 2 H 2
Water-gas shift reaction
H
2
+ CO
→ CO + H 2 O
2
The Carbon dioxide reforming of methane is the dominating reaction, carbon dioxide gasification, carbon dioxide gasification and methane decomposition are Subordinate reaction。 On the condition of carbonaceous catalyst,the oxygen-bearing functional group is the active center of CO2 reforming of CH4 , which could change the reaction course and reduce the activation energy . After a thorough analysis of experimental data and literature data [10-11], reaction mechanism of CO2 reforming of CH4 over carbonaceous catalyst as follows: CH
→ C
* 4
x [ CH
(1 +
+ * →
4
y )[ C O
2
+ 2H 2]
(2)
+ * → C O *2 ]
(3)
+ * →
y[C
*
*
C
(4)
]
y [ CO
* 2
+ C
2 x[ H
2
+ 2* → 2 H
* 2
CO
+ H
(1
−
CO
*
→
→ CO
*
+ OH
*
x
)[ C
*
→ CO + *
H
* 4
*
2 CO
*
( 4 x − 1 )[ H
(1)
* 4
CH
+
*
*
(5)
]
]
+ OH
(6) *
(7)
→ H 2O + 2 * ] O H
*
→
(8)
C O
*
+
5 2
H
2
+ * ]
(9) (10)
Where * denotes the active site on the carbonaceous catalyst Which corresponds to the over reaction stoichiometry: CH
4
+ (1 + y
)C O
2
+ yC →
4 -x -z H 2
2
x ⎞ ⎛ + 2 ⎜1 + y⎟ C O + zH 2O + xC 2 ⎠ ⎝
2.3. Kinetic model The CO2 reforming of CH4 is very complex reaction course, reaction mechanism is not the same with the different types of catalyst, the kinetic model was created in conversion of CH4 . The basic assumption of Kinetic model are as follows: ⅰ
The CO2 reforming of CH4 is the dominating reaction;
ⅱ The reaction temperature and content in feed of CH4 and CO2 were considered, nothing to do with the intermediate . Molar content in feed and concentration distribution of CH4 and CO2 are same in CO2 reforming
ⅲ
of CH4; The conversion of CH4 defined as: x
CH
4
= 1 −
[ CH [ CH
]
4
]
4
(11)
0
Where [CH 4 ]0 is content of CH4 in feed, [CH 4 ] is content of CH4 in outlet . The global rate for the transformation of CO2 reforming of CH4 can be expressed as follows: d [ CH dt
−
4
]
= k [ CH
4
] a [ CO
2
(12)
]b
where k is reaction rate constant; a and b are the reaction orders for CH4 and CO2, respectively.; Taking into account Eqs. (11) and (12) the global rate for CH4 conversion is represented in Eq. (13) d[CH4 ]0 (1 − xCH4 ) dt
(
= k[CH4 ]0a+b 1 − xCH4
)
( a +b )
(13)
Eq. (13) integral obtained: x CH
4
= 1 − [1 +( n - 1) k [ CH
4
]
n −1 0
t]
1 1− n
Where,n=a+b is the overall reaction order. Using some basic mathematic models, after adjusting the experiment data, the kinetic equations are created in origin multi-parameter curve fitting. Fig.3 shows curve fitting of different reaction temperatures. The result as Fig.3 shown,the coefficient of multiple determination (R2)was above 0.92 , indicating the goodness of fit . Assuming an Arrhenius temperature dependence, the reaction rate constant can be expressed as:
k = A exp( −
Ea ) RT
where A is the pre-exponential factor, Ea is the apparent activation energy (J mol−1), R is the universal gas constant (8.314 J mol−1 K−1) and T is the temperature (K).
1.0
0.8
C H 4 C on version
0.6
k=0.03481 n=2.04865 2
R =0.92809
0.4
k=0.05394 n=2.11935
0.6
2
R =0.98792 0.4
0.2 10
20
0
30
10
20
Resident time (s)
Resident time (s)
a T=1223k
b T=1273k 1.0
CH4 Conversion
0
0.8
k=0.07866 n=2.06985
0.6
2
R =0.98242
0.4 0
10
20
30
Resident time (s)
c Fig.3
T=1323k
kinetic curves of methane conversion of different temperatures
-2.5
lnK
C H 4 C o n v e r sio n
0.8
-3.0
-3.5
0.000090
0.000093
0.000096
1/RT
Fig.4 The lnk vs.1/RT relation
0.00009
30
The kinetic parameters were estimated by linear regression of the experimental data using the least squared method , the pre-exponential factor A and the apparent activation energy Ea are obtained , Fig.4 shows the lnk vs.1/RT relation . As can be seen from Fig.4,the activation energy and the pre-exponential factor were 109.9kJ/mol and 1765min·L/mol respectively.
k = 1765 exp( −
109.9 ± 10.1KJ ) RT
2.4 Comparison between calculation data and experimental data In order to compare between calculation data and experimental data , the conversion of CH4 and CO2 were calculated in different residence time and reaction temperatures. Fig.5 shows a comparison between calculation data and experimental data of conversion of CH4 and CO2. 90
Measured Calculated
80
Experiment Calculation
90
CH4 Conversion(%)
CH4 Conversion(%)
80 70
60
50
40
30
70
60
50
40 0
5
10
15
20
25
30
35
0
5
Residence time(s)
15
20
25
30
35
25
30
35
Residence time(s) 100
100
Experiment
Measured Calculated
Calculation
90
CO2 Conversion (%)
90
CH4 Conversion(%)
10
80
70
60
80
70
60
50 50 0
5
10
15
20
Residence time(s)
25
30
40 0
5
10
15
20
Residence time (s)
100
100
Experiment Calculation
Measured calculated
90
CO2 Conversion (%)
CO2 Conversion (%)
90
80
70
60
50
80
70
60
40
50 30 0
5
10
15
20
25
30
35
Residence time (s)
0
5
10
15
20
25
30
Residence time (s)
Fig.5 Comparison between calculation and experiment of CO2 reforming of CH4 in different temperature
The result as Fig.5 shown,the calculation data of CH4 conversion are close to the experimental data, the maximum error of which is 7.3%. which illustrates the rationality of the kinetic model. However, the calculation data of CO2 conversion of the simulation had a gap when compared to the experimental data and the maximum error of which is up to 20%, one of the reasons for the error is caused by dynamic equation which is the established in an ideal condition, and the experiment conditions were assumed, which ignored the effect of single particle in the diffusion and mass transfer on reforming reaction, the second reasons for the error is caused by the gasification reaction between carbonaceous catalyst and carbon dioxide.
4. Conclusion (1)With the residence time and reaction temperatures increasing , the conversion of CH4 and CO2 increases ,
and the temperature is main factor affecting reaction rate .
(2)the conversion of CO2 was significantly higher than the conversion of CH4, which indicate that the gasification reaction of the carbonaceous catalyst and carbon dioxide was occurred during the reforming process. (3)The kinetics of the CO2 reforming of CH4 on carbonaceous catalyst was described in a rate law, the experimental data was analyzed in non-linear regression. The apparent activation energy Ea and the pre-exponential factor A were solved. The rate constant was as as follows:
k = 1765 exp( −
109.9 ± 10.1KJ ) A comparison is made between calculation data RT
and experimental data of the CH4 conversion, which illustrates the rationality of the kinetic model.
Acknowledgements This work was supported by the National Basic Research Program of China (2005CB221202) and Shanxi Provincial Natural Science Foundation (2010011014-1).
References [1] Zhang Yongfa. A new technology of three products from coal based on a L T integrated apparatus [C]. Proceedings of Workshop on Coal Gasification for Clean and Secure Energy for China. Beijing, 2003: 235-246. [2] Zhang Yongfa et al. patent , ZL 031527973, 2005. [3] I. Suelves, M.J. Lázaro, R. Moliner et al. Hydrogen production by methane decarbonization: Carbonaceous catalysts[J]. Hydrogen Energy. 32 (2007) 3320 – 3326 [4] Seok Seung-Ho, Han Sung Hwan, Lee Jae Sung. The role of MnO in Ni/MnO-Al2O3 catalysts for carbon dioxide reforming of methane[J]. Applied Catalysis A: General, 2001, 215 (1 /2) : 31-38. [5] NowosielskaM, JozwiakW K, Rynkowski J. Physicochemical characterization of Al2O3 supported Ni– Rh systems and their catalytic performance in CH4 /CO2 reforming [J]. Catal. Lett, 2009, 128: 83-93. [6] Guo Hai - jun, Yang Min, Zhou Le et al. Progress in methane catalytic reforming with carbon diox ide to synthesis gas[J]. Guangzhou Chemical Industry. 2009, 37(5): 2-5. [7] Li Yan-bing , Xiao Rui , Jin Bao-sheng et al. Carbon dioxide reforming of Methane to produce syngas with Coal Char[J]. Journal of Combustion Science and Technology. 2009, 15(3): 238-242 [8] Guojie Zhang, Yue Dong , Meirong Feng et al. CO2 reforming of CH4 in coke oven gas to syngas over coal char catalyst[J]. Chemical Engineering Journal. 2010, 156: 519-523. [9] Zhang Huawei. Experimentalstudy ofoven gas reforming to producesynthesis gas over the coal char holding in the reactor [D]. Taiyuan university of technology , 2005. [10]Nandini A, Pant K K, Dhingra S C. Kinetic study of the catalytic carbon dioxide reforming of methane to synthesis gas over Ni - K/CeO2 -Al2O3 catalyst[J]. App l. Catal. A, 2006, 308: 119 - 127. [11]A. Nandini, K.K. Pant , S.C. Dhingra. Kinetic study of the catalytic carbon dioxide reforming of methane to synthesis gas over Ni-K/CeO2-Al2O3 catalyst[J]. Applied Catalysis A: General, 2006,308: 119–127.
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 CARBON MANAGEMENT:
A Study on the absorption characteristics of CO2 with a vortex tube type absorber KEUN-HEE HAN, WOO-JUNG RYU, JONG-HO PARK*, WON-KIL CHOI, JONG-SUB LEE, BYOUNG-MOO MIN Korea Institute of Energy Research, Jungang Vacuum Co. Ltd* 102, Gajung-ro, Yuseong-gu, Daejeon 305-343, KOREA [email protected] Abstracts : In this study, the CO2 removal characteristics of the vortex tube type absorption apparatus were investigated to enhance the compactness of CO2 absorption process and to reduce the amount of absorbing solution for the process. The vortex tube with the diameter of 12 mm and the length of 200 mm was introduced in the experimental apparatus 3
to treat 20 Nm /hr of CO2 containing flue gas. The flue gases for the experiments containing 11-15 vol% of CO2 were supplied from the coal-firing CFBC power plant with 12 ton/hr of steam producing capacity. EDA and DETA based on MEA were used as the absorbing solution. The absorption experiments were executed under the various conditions like the absorbing solution concentrations in the 30 wt%, the flow rate of CO2 containing 3
3
flue gases in the range of 6 to 12 m /hr and the flow rate of absorbing solution in the 0.18m /hr and the range of operation pressure in the 1.4 to 5.5 kgf/cm2. As a results, the CO2 removal efficiency increases with the operation pressure but deceases with the flow rate of flue gas. However, the development of an additional process to improve the efficiency of CO2 absorption is required. Key words: Vortex Tube; Absorption; Flue Gas; Carbon Dioxide; Alkanol Amine; 1. Introduction One of the major known causes of global warming is the CO2 emitted by the use of fossil energy. CO2 is mostly emitted by the steel making, chemical, cement and power plant industries, which form the mainstay of the national economy. In particular, coal-fired thermal power plants generate a significant amount of CO2 and have a major impact on the global environment. Measures aimed at restricting the emission of CO2 include reducing energy consumption, increasing the efficiency of equipment, using more natural sources of energy, and retrieving emitted CO2. Many technologies and processes are being developed to remove the emitted CO2 while ensuring the stable operation of existing energy production facilities. All such measures pursue the effectiveness, low cost and high efficiency. The CO2 separation processes can be divided into absorption, adsorption, cryogenics (distillation), oxygen enriched combustion, and membrane separation, and etc. The most suitable process for recovering the emitted CO2 from large-scale sources such as thermal power plants is the chemical absorption method, which is appropriate when the CO2 concentration and pressure in the flue gas are relatively low1,2). Alkanol amines are popularly used as the absorbent. A vortex tube with a simple structure can separate the compressed gas into low temperature portion and high temperature portion using the vortex generated when high pressure gas is sprayed into a tube
3,4)
. When a gas/liquid
mixture is passed through a tube at high speed, the gas and liquid can be separated by using the difference in gravity
- 1 -
5)
caused by the centrifugal force of the mixture . The mixture is formed of micro droplets as it is moved at high speed. Its principal benefit is the processing treatable 6)
a large volume of mixed gas from a small device . Using such characteristics, it can be applied to CO2 absorption from the flue gas. The most important factor of CO2 absorption from the flue gas is the contact time between the flue gas and the absorbent. This study used MEA (Monoethanolamine), EDA (Ethylenediamine), and DETA (Diethylenetriamine), which are alkanol amine type absorbents. The absorption characteristics of CO2 in the flue gas generated by the burning of coal were invetigated by varying the flue gas flow rate, the aqueous solution flow rate and the operating pressure of the vortex tube type absorber. 2. Experimental equipment and procedure Fig.1 shows the equipment used for this study. In a vortex tube, the compressed gas is needed to create a vortex. 2
3
For this purpose, the flue gases were charged to the flue gas storage tank (pressure: 5~6 kgf/m , volume:0.75m ) using a compressor. Flue gases were generated by combustion of coal in the circulating fluidized bed combustor (CFBC). The vortex tube for CO2 absorption composed of 12 mm in diameter (ID) and 200 mm in length(L). Mixing of the flue gas and the aqueous solution was controlled in the spray chamber before supplying to the Gas
vortex tube. As the flue gas and the aqueous
CO2 Analyzer Condenser
solution were sprayed into the vortex tube, the
T P
T P
Regulrator
MFC
CO2 in the flue gas was absorbed through the vortex flow. A CO2 analyzer (Horiba, VA-3000) P
was installed in the gas exit line of the vortex
Vortex Tube
tube
Liquid
reactor
to
measure
the
change
in
the
concentration of CO2, which was monitored and
Pump
recorded in a personal computer. Flue gas Rich Absorbent
Fresh Absorbent
Flue Gas Storage Tank
The experimental conditions are described in Table 1. MEA, EDA and DTEA were used as the
Compressor
absorbent. The aqueous solution was supplied at Fig. 1 Schematic Diagram of Vortex Tube System for CO2
3.0 liter/min as a fixed value, while the flue gas
Absorption
flow rate and pressure of the tube in which the absorption occurs were varied. The flue gas and the aqueous solution were flowed through the tube at the end of which the liquid/gas separation
Table 1. Experimental Conditions
occurs by centrifugal force due to the difference in Parameter
Conditions
gravity. The concentration of CO2 was measured
A kind of Absorbent
MEA, EDA, DETA
for monitoring the CO2 removal at the exit of the
Concentration of Absorbent (wt %)
30
flue gas. The gas analyzer measured the reduction
3
Flow Rate of Absorbent (m /hr) 3
6.0~12.0
2
1.4~5.5
Flow Rate of Flue Gas (m /hr) Operation Pressure (kgf/cm )
in the concentration of CO2 as the absorption took
0.18
place. The equilibrium was assumed when there was no further change in the concentration of CO2. The removal efficiency and CO2 loading were calculated based on the equilibrium state
assumed.
- 2 -
3. Results and Discussion The removal efficiency and CO2 loading of the vortex tube equipment can be expressed by the following equations.
×
× ×
× ×
× ×
× ×
here : CO2 removal efficiency, %
: CO2 loading, mole/mole_solution
: mole of CO2 in flue gas at inlet, -
: mole of CO2 in flue gas at outlet, -
: mole of aqueous solution, : flow rate of inlet flue gas, ℓ/min,
: flow rate of out gas, ℓ/min
: CO2 mole fraction of flue gas, -
: CO2 mole fraction of out gas, -
2
: absorber pressure, kgf/cm
: outlet pressure, kgf/cm2
: temperature of absorber, ℃
: temperature of out gas, ℃
Absorbents Fig. 2 and 3 shows the effect of G/L ratio on CO2 loading and the removal efficiency. The concentrations of aqueous solutions are 30 wt% in the water. The gas/liquid ratio was varied to observe CO2 loading and removal 2
efficiency at the operating pressure of 1.4 kgf/cm . As the G/L ratio increased, CO2 loading also increased while the removal efficiency decreased. The reason for the increase of CO2 loading was because the vortex velocity increased with the G/L ratio. The higher vortex velocity makes the droplets of the aqueous solution smaller which increases the 7)
interfacial area to react with CO2 in the flue gas per unit volume . Furthermore, the reason for the removal efficiency decrease was the higher gas flow per unit volume which increases the concentration of CO2 to be absorbed. Different CO2 loadings were observed for different aqueous solutions. In the molecular structure of alkanol amine, the capacity to absorb CO2 depends on the number of amine groups formed. In the case of DETA, the structure of 2-NH2 and 1-NH has more amine groups to absorb CO2 than that of MEA.
Fig. 2 Effect of G/L ratio on CO2 loading for each
Fig. 3 Effect of G/L ratio on CO2 removal efficiency
absorbent
for each absorbent
- 3 -
Operating Pressure Fig. 4 and 5 show the effect of operation pressure on CO2 removal efficiency and the CO2 loading. The operating pressure was adjusted by using the opening of the throttle valve at the end of the vortex tube absorber. The pressure of the flue gas ranged from 1.4~5.5 kgf/cm2. The aqueous solution was 30 wt% MEA in the water. As the operating pressure increases, the CO2 removal efficiency and CO2 loading per mole of absorbent increased. The higher pressure 7)
means higher CO2 partial pressure and higher CO2 partial pressure increases the absorbing capability . In addition, at higher operating pressure, the absorbent may produce the physical absorption by the forced hydrolysis of CO2 in the water and amine.
Fig. 4 Effect of Operation Pressure on CO2 Removal
Fig. 5 Effect of Operation Pressure on CO2 Loading at
Efficiency at G/L Ratio.
G/L Ratio.
Absorption rate Fig. 6 shows the absorption rates of three absorbents (MEA: 4.91mol/ℓ, EDA: 4.90mol/ℓ, DETA: 2.90mol/ℓ) with the same concentration 30 wt% at 30℃. The absorption rate is measured by the time between the CO2 in the flue gas gets in contact with the sprayed aqueous solution to begin absorption and the CO2 does not increase any longer. The difference in absorption rates is due to the number of amine groups in the absorbent. EDA, which has 2-NH2 that can quickly absorb CO2 was the fastest. It also showed the highest CO2 removal. If the absorbents with the same number of moles are used, DETA which has one more amine group (NH) than EDA would have shown the highest removal efficiency. Fig. 6 Effect of CO2 Concentration Change on Operation time
4. Conclusion The absorption rate was in the order of EDA, DETA and
MEA. EDA, which had 2-NH2, showed higher rate than MEA, which had 1-NH2. DETA, which had 2-NH2 and 1-NH, showed similar rate to EDA but had higher CO2 loading.
- 4 -
As the flue gas flow rate increased, CO2 loading increased and removal efficiency decreased. That is because the higher flue gas flow rate caused the droplets of the absorbent to be smaller as the vortex velocity increased, and the smaller droplets increased the contact surface. As the operating pressure increased, the CO2 loading and removal efficiency increased. Higher pressure means higher CO2 partial pressure and that increased the number of CO2 moles that can be absorbed per unit mole of absorbent. Furthermore, increased pressure enables the physical absorption due to the forced hydrolysis of the water not reacting with CO2 and amine. Reference 1. Chakma A., and Tontiwachwuthikul P., “Designer Solvents for Energy Efficient CO2. Separation from Flue Gas Streams”, Greenhouse Gas Control Technologies, 35~42(1999). 2. Han K. H. et al., "A Study on the CO2 Removal Efficiency with Aqueous MEA and Blended Solutions in a Vortex Tube Type Absorber", Korean Chem. Eng. Res., 47(6), 197~202(2009). 3. Rangue, G. J., "Method and Apparatus for Obtaining from Fluid under Pressure Two Currents of Fluids at Different Temperatures", US Patent No. 1,952,281 (1934). 4. Hilsch R., “The Use of Expansion of Gases in a Centrifugal Field as a Cooling Process”, The Review of Scientific Instruments, Vol. 18, No. 2, pp. 108~113(1947). 5. Eiamsa, S. and Promvonge, P., "Review of Ranque-Hilsch Effect in Vortex Tube", Renew. Sust. Energ. Rev., 12, 1822~1842(2008). 6. Han K. H. et al., "Absorption Equilibrium of CO2 in the Sterical Hindered Amine, AMP Aqueous Solution", Korean Chem. Eng. Res., 45(2), 795~800(2007). 7. A. Chakma et al, "Absorption of CO2 by aqueous triethanolamine(TEA) solutions in a high shear jet absorber", Gas Separation & Purification, Vol. 3, June (1989).
- 5 -
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: CARBON MANAGEMENT Pd-FREE COMPOSITE MEMBRANE FOR PRE-COMBUSTION CAPTURE Jung Hoon Park, PhD., Korea Institute of Energy Research [email protected], Tel +82-42-860-3766, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea Sung Il Jeon, Master, Korea Institute of Energy Research [email protected], Tel +82-42-860-3796, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea Young Jong Choi, Master, Innowill Corporation [email protected], Tel +82-42-862-7500, Fax +82-42-862-6500 533,Yongsan–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea
Abstract: Pre-combustion CO2 capture technology is recently focused on one of reduction methods of carbon dioxide from power generation system in view of environmental (carbon management) and sustainable (Hydrogen economy) point. Separation of hydrogen from water-gas shift reactors through dense hydrogen transport membranes, while retaining CO 2 produces essentially pure hydrogen in the permeate and CO2 at high pressure and high concentration in retentate, which is ideal for efficient sequestration of CO2. Moreover, the combination of a hydrogen selective membrane with a water-gas shift catalyst in a single reactor would allow a high degree of CO conversion, despite a low equilibrium constant at high temperature, due to the continuous depletion of H 2. This equilibrium shift can provide more hydrogen productivity, higher concentration of CO 2 and lower impurity such as CO. Nowadays, many researches have been studied on various membranes for low cost separation of hydrogen. Pd and Pd alloy membranes have been studied for separating hydrogen from pre-combustion capture process but there is limitation for large scale application without reduction of Pd layer and improvement of membrane stability, due to the price and embrittlement of noble metal. Recently, Pd-free membranes using V, Ta, ans Nb which has higher permeability have been researched to improve its high cost and low stability. In this work, metal alloy and composite membranes have been developed to separate hydrogen from mixed model gases, particularly product streams generated during coal gasification and/or water gas shift reaction. The powder mixture for fabricating the cermet membranes was prepared by mechanically mixing 60 vol.% vanadium with Y 2O3-stabilized ZrO2 (YSZ). The powder mixture was pressed into disks, which were then sintered in vacuum at 1600 ℃ for 2 h. As-sintered membrane was dense and mounted to a stainless steel ring with brazing filler. Hydrogen fluxes of V/YSZ membrane have been measured in the range of 200~350℃ with 100% H2. The crack was formed in the both sides of membrane at 350 ℃ and pressure of 0.5 bar. During permeation experiment, vanadium of V/YSZ membrane reacted with hydrogen to form V2H which was the origin of crack formation. To improve the membrane stability, we prepared metal alloy membranes with vanadium. Hydrogen fluxes of metal alloy membranes have been measured in the range of 350~500℃ with 100% H2 and the mixture gas of H2 and He (or CO2). In addition, the stability of membrane was investigated according to operating temperature and hydrogen partial pressure.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
PROGRAM TOPIC: CARBON MANAGEMENT COMPOSITE CERAMIC MEMBRANE FOR OXYGEN SEPARATION Jung Hoon Park, PhD., Korea Institute of Energy Research [email protected], Tel +82-42-860-3766, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea Jong Pyo Kim, Master, Chungnam National University [email protected], Tel +82-42-862-4200, Fax +82-42-862-6500 220, Gung-dong, Yuseong-gu, Daejeon 305-764, Republic of Korea Soo Hwan Son, Master, Korea Institute of Energy Research [email protected], Tel +82-42-860-3796, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea
Abstract: Oxygen and nitrogen are used in many industrial processes. Pure oxygen is used in the production of metals and in the integrated gasification combined cycle (IGCC) for partial oxidation. Recently, oxy-fuel combustion CO2 capture process as one of carbon capture technologies has been developed and this process also needs large scale oxygen separation unit. It has been predicted that the total market would grow significantly if pure oxygen could be produced at lower cost. The technology used for commercial separation of oxygen varies according to the scale and requirements for oxygen purity. For example, the cryogenic distillation method that was started in 1902 is used for large-scale production of pure oxygen, and the simultaneous production of nitrogen, argon and helium. However, high investment costs and energy consumption make it difficult to integrate this process with other power generation. Over the last decade, membrane technology for gas separation has developed rapidly. The interest in dense ceramic membranes for the transport of oxygen has grown considerably. As a result, the knowledge of the intrinsic properties of the membrane materials is now overwhelming, and new reports are being published frequently. The use of a dense mixed-conducting, perovskite-type ceramic membrane is a new technology for the production of pure oxygen. An obvious advantage of perovskite membranes is their 100% selectivity for the permeation of oxygen. However, ceramic membrane process has some drawbacks; the crack formation of membrane under high pressure and temperature condition, low stability of perovskite structure under ambient air, the difficulty of heat exchange integration with flue gas. In this work, the effect of carbon dioxide in ambient air was studied using the composite membrane with La0.6Sr0.4Ti0.3Fe0.7O3-δ (coating layer, denoted as LSTF-6437) and Ba0.5Sr0.5Co0.8Fe0.2O3−δ (bulk permeation body, denoted as BSCF-5582). BSCF-5582 and LSTF-6437 powder have been synthesized using polymerized complex method. BSCF-5582 powders were compressed into disks of 20 mm in diameter and 1.0 ~ 2.0 mm of thickness in a stainless steel mold under a hydraulic load by unilateral press. The green disk sintered at 1353 K for 5 hr. The sintered disk was polished to smooth the surface and to control the thickness of disk with 600 grit SiC. The surface of membrane was modified by coating of LSTF-6437 slurry. The optimum coating condition was evaluated according to coating time, rate and number. The coating membrane was sintered again to obtain composite dense membrane. The phase of the powder and the disk before and after sintering was characterized with an X-ray diffraction. Prior to oxygen permeation test, the cell part is purged with He gas to remove the air in permeation cell tube and to confirm sealing of the assembly for 20 hr. The leakage through membrane during oxygen permeation test was also measured for all runs at each temperature and the oxygen permeation fluxes were corrected on the basis of the measured leakage. Permeation study was performed in the temperature range of 750~950 ℃
and pressure of 1~3 atm with synthetic air (21 vol.% O2+79 vol.% N2) and ambient air model gases (CO2 300~700 ppm). Oxygen permeation flux was increased as temperature increased irrespective of membrane coating. In the case of LSTF coating membrane, it reached 1.9 ml/cm2·min at 950 ℃ exposed to flowing air (Ph =0.21 atm, feed side of membrane) and helium (Pl =10 -5 atm, permeated side of membrane). The oxygen permeation of BSCF-5582 membrane in the condition of air and CO2 (300~700 ppm) in feed stream decreased more than 43% in comparison with air feed stream while that of LSTF coating membrane maintained almost same flux irrespective of CO2 concentration.
Reactions of Coal Structures with Polymers Leading to Hydrogen Production P. Straka Institute of Rock Structure and Mechanics, v.v.i., Academy of Sciences of the Czech Republic, V Holešovičkách 41, 18209 Praha 8, Czech Republic E-mail address: [email protected]
Abstract Thermal reactions of polystyrene, acrylonitrile-butadiene-styrene and styrene-butadiene rubber with chosen coal fraction were investigated from point of view of hydrogen production. As method a two-stage copyrolysis of coal/polymer mixtures was selected. Experiments were carried out on pyrolysis laboratory unit with a vertical quartz reactor (the first stage) and a horizontal cracking oven (the second stage). Thus, coal/polymer mixtures contained 30 wt.% of polymer were heated and products further cracked. From the results, the process conditions leading to maximum hydrogen production were defined and the yields of products determined. If the mixtures with 30 wt.% of polymers are considered, the heating rate of 5 K/min and final temperature of 900oC at vertical reactor and 1200oC at horizontal cracking oven are sufficient for achievement of a hydrogen-rich gas with 77–79 vol.% H2. On the basis of solid-state NMR, FTIR and GC analyses of coal fraction and cokes, obtained tar, and obtained gas, respectively, reactions of coal structures with polymers leading to hydrogen production were described and, using the results of isoconversional analysis, discussed. Introduction Other than standard technologies, hydrogen can be obtained by non-traditional methods using waste polymers, in particular their mixtures. Studies have been published dealing with the transformation of polymer structures into gaseous and liquid products and solid carbon during the carbonization process [1], while others have been published on the possibility of utilizing the liquid products of pyrolysis as a substitute for fuels [2-4]. Possibility of hydrogen obtaining from the coal/waste-tyre mixture including the use of solid carbonaceous residue is discussed in the work [5]. Our preliminary study on the hydrogen distribution during copyrolyses of coal/polymers mixtures containing 15– 60 wt.% of polymer showed that a significant portion of hydrogen is distributed just in tar. Subsequently, the tar as a part of raw gas can be led into cracking oven where it will be decomposed into hydrogen and pyrolytic carbon. Therefore, a two-stage process involving the thermal decomposition of the mixture in the first stage and the subsequent splitting of liquid and gaseous products in the second stage can be used to obtain hydrogen. To control such a process should be identified the ongoing chemical reactions and determined the process conditions. The aim of this study is on the basis of solid-state NMR, FTIR and GC analyses of the used coal fraction and obtained products and kinetic evaluation describe reactions of coal structures with polystyrene polymers leading to hydrogen production, and the optimal process conditions determine. Experimental Experiments were carried out on pyrolysis laboratory unit with a vertical quartz reactor (an inner diameter of 6 cm and a length of 50 cm) as the first stage, and a horizontal oven for cracking of raw gas components in a quartz tube (with an inner diameter of 4.4 cm and a length of 60 cm) which worked as the second stage. 100 g of sample was pyrolysed/copyrolysed; coal/polymer mixtures contained 30 wt.% of polymer. As polymers, polystyrene (PS), grain size 3–4 mm, acrylonitrile-butadiene-styrene (ABS), grain size 2–3 mm, and styrene-butadiene rubber (SBR) with grain size -3 mm were tested, further, coal
1
fraction from the An Tai Bao coal (Shanxi, China) was used (apparent density of 1.30–1.35 g/cm3; for size analysis and proximate and ultimate analyses see Table 1). For description of Table 1 Parameters of the coal fraction used. Size analysis Proximate and ultimate analyses (wt.%) Grain size Share (wt.%) water 2.85 C (daf) 75.00 +3 0.45 ash (d.b.) 10.17 H (daf) 6.46 2–3 19.23 total S (d.b.) 1.70 N (daf) 0.45 1–2 44.81 VM (daf) 37.48 S+O(daf) 18.09 -1 35.50 Qs (daf) 32.20 (MJ/kg) reactions in question the coal fraction and obtained cokes were characterized by the solid-state 13 C CP/MAS NMR structural parameters; tars were analyzed by FTIR spectroscopy; gas components were during copyrolyses continuously controlled by infrared analyzers and finally determined by GC method. Further, kinetic measurements by TG/DSC method were carried out and parameters of Arrhenius plot determined. Finally, an isoconversional analysis of thermal degradation process of coal fraction and the mixtures in question by the Friedman, Ozawa, and Flynn and Wall methods [6] was performed. Results and Discussion In all the cases, the yield of tar from one-stage copyrolyses was significantly higher in comparison with that from pyrolysis of coal alone (Table 2). Moreover, at the one-stage Table 2 The yields of the products from one-stage pyrolysis of coal fraction alone and copyrolysis with 30 % of polymer (wt.%). Tar Reaction water Gas Losses Initial sample Coke Coal 67.5 6.2 11.8 11.9 2.6 Coal+PS 48.4 30.0 9.4 8.3 3.9 Coal+ABS 50.0 31.5 5.6 8.9 4.0 Coal+SBR 58.4 19.4 8.3 8.5 5.4 process the content of non-methanic hydrocarbons in obtained gas was higher with copyrolyses (4 vol.%, Table 3, coal+ABS and coal+SBR) as against pyrolysis (3 vol.%, Table 3). Due to this, the gas from two-stage copyrolyses contained always quite high amount of hydrogen (Table 4). Table 3 Parameters of gas from the one-stage pyrolysis and copyrolyses with polymers. Coal Coal+PS Coal+ABS Coal+SBR 3 3 3 vol.% Ndm vol.% Ndm vol.% Ndm vol.% Ndm3 H2 54.85 12.38 55.13 8.88 53.80 8.98 55.53 9.22 CH4 25.43 5.75 25.82 4.16 26.65 4.45 25.12 4.17 C2H4 0.43 0.10 0.52 0.08 0.77 0.13 0.78 0.13 C2H6 2.02 0.46 1.98 0.32 2.37 0.40 2.33 0.39 C3H6 0.30 0.07 0.30 0.05 0.48 0.08 0.49 0.08 C3H8 0.48 0.11 0.44 0.07 0.55 0.09 0.65 0.11 0.09 0.02 0.06 0.01 0.10 0.02 0.10 0.02 ΣC4 N2 1.76 0.40 1.57 0.25 1.65 0.28 2.16 0.36 CO 8.51 1.92 8.65 1.39 8.39 1.40 7.83 1.30 CO2 6.13 1.39 5.53 0.89 5.24 0.88 5.01 0.83
2
Table 4 Parameters of gas from the two-stage pyrolysis and copyrolyses with polymers. Coal Coal+PS Coal+ABS Coal+SBR 3 3 3 vol.% Ndm vol.% Ndm vol.% Ndm vol.% Ndm3 H2 69.43 30.83 77.32 40.52 78.71 42.66 77.82 41.09 CH4 3.64 1.62 4.80 2.52 5.52 2.99 5.34 2.82 C2H4 0.02 0.01 0.05 0.03 0.06 0.03 0.07 0.04 C2H6 0.01 0.00 0.01 0.00 0.02 0.01 0.01 0.01 C3H6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 C3H8 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ΣC4 N2 3.46 1.54 3.15 1.65 1.99 1.08 1.97 1.04 CO 18.83 8.36 12.83 6.72 12.00 6.50 13.16 6.95 CO2 4.61 2.05 1.84 0.96 1.69 0.92 1.62 0.90 From comparison with one-stage process it follows that the products and their yields from these two processes differ substantially. Due to cracking of raw gas components in the second stage as products of two-stage pyrolysis or copyrolysis the coke, glossy and amorphous carbons, and gas were formed, further, the reaction water was found as a minor product. Quantitatively, if coke, glossy and amorphous carbon are considered as a carbon material, this process is a two-products one as the yield of reaction water is low (Table 5), often lower than losses. Table 5 The yields of the products from two-stage pyrolysis of coal fraction alone and copyrolysis of coal with 30 % of polymer (wt.%). Initial sample Coke Glossy and amorphous Gas Reaction water Losses carbons Coal 69.7 4.1 20.2 5.6 0.4 Coal+PS 48.8 22.4 17.8 3.1 7.9 Coal+ABS 51.4 20.7 17.2 4.1 6.6 Coal+SBR 59.7 13.9 17.4 4.3 4.7 Presented results were achieved under the process conditions leading to maximum hydrogen production. If the mixtures with 30 wt.% of polymer are considered, the heating rate of 5 K/min and final temperature of 900oC at the first stage reactor and 1200oC at cracking oven in the second stage are sufficient for achievement of a hydrogen-rich gas with 77–79 vol.% H2 and the yields of the gas 17–18 wt.%. Probably, these conditions can be applied even if the polymers content is 15–60 wt.% [7]. For exploring of thermal decomposition of mixtures in question the dependences of both effective activation energy of decomposition (Eα) and the pre-exponencial factor on the extent of conversion (α) were followed. These dependences for coal alone and coal/polymer mixtures showed that in the case of coal alone the Eα is substantially higher than that of decomposition of coal/polymer mixtures. Generally, on heating, both polymers and coal degrade forming low molecular products; the degradation involves breaking of C–C bonds whose bond energy is around 350 kJ/mol and occurs above 200–300oC with polymers in question and above 300oC with coal. The effective activation energy varies throughout the process as at earlier stages the weak links are initiated and cleaved, but at later stages the stronger bonds are splitted. From TG/DSC curves in the key temperature range of 300–550oC the dependences of Eα on α were calculated and variability of Eα evaluated by applying an advanced isoconversional method [6]. The results are shown in Figs. 1–5. In the case of coal
3
alone, Eα 300–350 kJ/mol was found in the range of α 10–90 % (Fig. 1). Contrary, for coal/PS mixture only 200–250 kJ/mol (Fig. 2) and 180–290 kJ/mol for coal/ABS mixture (Fig. 3) was registered in the same range of α. It means that thermal degradation of mixtures with PS and ABS proceeds more easily in comparison with coal alone. The reason is that the mixtures contain substantially more aliphatic C–H bonds that are easier to cleave.
Fig. 1 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal fraction in the range of 300–550oC.
Fig. 2 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/PS mixture in the range of 300–550oC. 4
Fig. 3 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/ABS mixture in the range of 300–550oC. Another case is the coal/SBR mixture. Unlike PS and ABS, this mixture is decomposed in two temperature intervals, at 300–400oC and further at 400–550oC (Fig. 4).
Fig. 4 DTG curves of coal/SBR mixture at 3 heating rates. Due to this, with coal/SBR mixture two dependencies of Eα on α were followed. The first, for interval of 300–400oC, the second, 400–550oC were considered. In the first interval the weak bonds were splitted, thus, Eα varied between 150–100 kJ/mol and decreased with an increasing α (considered again between 10–90 %) (Fig. 5). In the second interval of 400– 550oC, the stronger bonds were cleaved. Therefore, in the range of α 10–70 %, Eα varied between 200–300 kJ/mol and increased with an increasing α (Fig. 6). In the range of α 70–90 % the Eα increased up to 700 kJ/mol as strong bonds were cleaved. In all the cases,
5
simultaneously with degradation the polycondensation reactions leading to new aromatic formations occurred. Therefore, in the range of α 90–100 % the Eα increased up to 420 or 700 or 2,400 kJ/mol together with pre-exponential factor. On the whole, in the temperature range of 300–550oC the prevailing Eα values were lower than those at decomposition of coal fraction alone. It means that splitting reactions in the mixtures proceed more easily than those with coal alone; due to this the hydrogen production is higher.
Fig. 5 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/SBR mixture in the range of 300–400oC.
Fig. 6 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/SBR mixture in the range of 400–550oC. 6
Further, tars from pyrolysis and copyrolyses were analyzed by FTIR method. No significant differences between tar from pyrolysis of coal fraction alone and copyrolyses with considered polymers were found (Fig. 7), but the aromaticity index of the tar from pyrolysis of coal (0.136) was higher in comparison with those for tars from copyrolyses (0.095 and 0.089 at PS/coal and SBR/coal, respectively). It means that the tars from copyrolyses contain more aliphatic C–H bonds than pyrolysis coal tar. This feature is favorable for hydrogen production in the second stage. A
B
C
Fig. 7 FTIR spectra of the tars from pyrolysis of coal fraction (A) and copyrolyses with SBR (B) and PS (C). Further, coal fraction and cokes resulting from pyrolysis and copyrolyses were investigated by solid-state 13C CP/MAS NMR method. The initial coal fraction had lower aromaticity (0.6259) and worse coking properties. Its 13C CP/MAS NMR spectrum is shown in Fig. 8. From spectrum analysis was found that a share of phenolic carbons from total carbons was of 8.22 %, alkylated aromatic carbons 9.3 %, protonated aromatic carbons 36.91 % and bridgehead carbons 6.92 %; further, the share of carbons in CH2 and CH3 groups was 26.90 % and quaternary carbons 6.69 %. From these data it follows that together with quite high content of hydrogen (Table 1) the coal fraction has the good preconditions for hydrogen production because of high amounts of protonated aromatic carbons and CH2 and CH3 groups. 13 C CP/MAS NMR parameters of resulting cokes are summarized in Table 6, spectra of cokes are shown in Figs. 9 (coke from coal fraction) and 10 (coke from coal/PS mixture). Spectra of cokes from coal/ABS and coal/SBR were quite similar to that from the coal/PS mixture. Table 6
13
C CP/MAS NMR parameters of cokes from pyrolysis of coal fraction alone and copyrolyses (%). protonated Bridgehead C Coke from phenolic C alkylated aromatic C aromatic C Coal fraction 3.66 3.98 46.43 23.76 Coal/PS 2.37 4.76 41.06 24.48 Coal/ABS 4.32 7.73 51.28 22.69 Coal/SBR 0.97 3.29 49.34 24.92
7
Fig. 8
13
C CP/MAS NMR spectrum of the coal fraction used.
Fig. 9
13
C CP/MAS NMR spectrum of coke from pyrolysis of coal fraction.
8
Fig. 10
13
C CP/MAS NMR spectrum of coke from copyrolysis of coal fraction with 30% PS.
From the data in Table 6 is obvious that in the case of cokes from coal/ABS and coal/SBR mixtures the share of protonated aromatic carbons is higher (with the coal/ABS mixture significantly higher) as compared with coke from coal fraction alone. It can be deduced that during copyrolysis the similar structures as at pyrolysis are formed but with a higher content of bound hydrogen. On the whole, during copyrolysis, polymers promote the creation of hydrogen-rich structures, consequently, together with coals with worse coking properties the waste polymers can be used for production of hydrogen. From presented data a following picture of reactions of polymers with coal can be outlined. In the one-stage copyrolysis, hydrogen is released during thermal splitting of C–H bonds, mainly from alicyclic rings, methylene and ethylene bridges and polymer chains, further, from aromatic rings during their polycondensation. Decomposition of styrene polymers yields e.g. styrene, ethyl benzene, toluene, isopropyl benzene, 1,3-diphenyl propane and α-methyl styrene which enrich the coal tar. Simultaneously, the new aromates arise by aromates-aromates reactions as alicyclic parts of coal clusters are cleaved during copyrolysis and formed aromatic structures react with aromatic products of decomposition of polystyrene polymers. Thus, essential constituents of copyrolysis tar are formed. In this way, a significant portion of hydrogen from coal and polymers is transferred into tar. Formed H2 molecules and aliphatic C1–C4 hydrocarbons are transferred into the gas. Nitrogen containing compounds are predominantly converted to ammonia, and sulfur containing structures to hydrogen sulfide. In the two-stage copyrolysis, the tar components, and gaseous hydrocarbons are thermally decomposed up to pyrolytic carbon and hydrogen. Oxygen containing compounds are splitted and elemental oxygen is generated, which reacts with carbon and hydrogen to carbon monoxide and reaction water. Through endothermic gasifying reactions between water and carbon, and through water-shift reaction with carbon monoxide, again hydrogen is produced. Finally, ammonia is splitted to hydrogen and nitrogen. On the whole, through twostage copyrolysis the polystyrene polymers with coal are converted to hydrogen-rich gas and pyrolytic carbon.
9
Acknowledgements The work was supported by the Czech Science Foundation (GAČR) as the project No. 105/07/1407 and conducted under the Institutional Research Plan AVOZ30460519, Academy of Sciences of the Czech Republic. References 1. Kidena K., Murata S., Nomura M.: Studies on the chemical structural change during carbonization process, Energy and Fuels 10, 1996, 672-678. 2. Ucar S., Karagoz S., Yanik J., Saglam M. Yuksel M.: Copyrolysis of scrap tires with waste lubricant oil, Fuel Processing Technology 87, 2005, 53-58. 3. Walendziewski J.: Engine fuel derived from waste plastics by thermal treatment, Fuel 81, 2002, 473-481. 4. Rodriguez I.M., Laresgoiti M.F., Cabrero M.A., Torres A., Chomon M.J., Caballero B.M.: Pyrolysis of scrap tyres, Fuel Processing Technology 72, 2001, 9-22. 5. Kříž V., Brožová Z., Přibyl O., Sýkorová I.: Possibility of obtaining hydrogen from coal/waste-tyre mixture, Fuel Processing Technology 89, 2008, 1069-1075. 6. Vyazovkin S., Sbirrazzuoli N.: Isoconversional Kinetic Analysis of Thermally Stimulated Processes in Polymers, Macromolecular Rapid Communications 27, 2006, 1515-1532. 7. Bičáková O., Straka P.: The resources and methods of hydrogen production, Acta Geodynamica et Geomaterialia 7, No.2(158), 2010, 175-188.
10
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
COAL-DERIVED PRODUCTS:
COAL SUPPLY AGREEMENTS AND COMPETITION: Değer Boden Akalın, LLM Contact Information: Abdülhakhamit Cad. No:70 K:4 D:11Taksim,İstanbul, Turkey Tel: +90 212 253 09 28 Fax: +90 212 253 09 29 [email protected] www.boden-law.com
Abstract: Coal is used for many different purposes. It is primarily used as a solid fuel to produce electricity and heat. It can also be converted into liquid fuels such as gasoline or diesel. It can be used to produce synthetic natural gas, synthesis gas and hydrogen through gasification. Coal supply agreement is critical for the usage of coal for different purposes. Under a coal supply agreement, the supplier agrees to sell coals to a purchaser at pre-determined prices with minimum/maximum annual quantities. Duration, quality of coal, quantities, delivery, risk and ownership, price determination, termination and force majeure clauses are important clauses of a coal supply agreement. In drafting coal supply agreements antitrust laws should be reviewed. This paper aims to analyze how a coal supply agreement may violate competition.
CHEMICALS FROM TURKISH LIGNITES Vedat Mihladız TKI HISTORY Coal was used in Turkey to produce chemicals (ammonia, nitric acid and ammonium nitrate) since year of 1961 to 2000. For this purpose two type of gasifiers (both atmospheric) were used successfully; - fludize bed (Winkler) - entrained bed (Koppers-Totzek) TEST RESULTS Various coal types were tested in these gasifiers and the data collected from these tests were carried to date. Soma coal gasification test results are given in table 1 and table 2. Seyitömer lignites were used in both gasifiers and operation results are given for K-T gasifiers in table 3. Table 1. Soma coal gasification test results K-T GASIFIER
WINKLER GASIFIER
Consumption
Consumption 10.000
Kg/h
Coal
7.700
Kg/h
C
43,7
%
C
40.1
%
H
3,2
%
H
2.9
%
O
15.0
%
O
13.7
%
N
0,8
%
N
0.8
%
Coal
S
0,3
%
S
0.3
%
Ash
34.7
%
Ash
36.6
%
H2O
2.2
%
H2O
5.6
%
Oxygen (95%)
3.950
Nm3/h
Oxygen (95%)
2.300
Nm3/h
Quench water
2.500
kg/h
Steam
2.600
kg/h
8.000
Nm3/h
Products Gas
Products 11.750
Nm3/h
Gas
CO2
11.4
%
23.5
%
CO
57.8
%
33.0
%
H2
27.8
%
37.8
%
N2
2.8
%
2.7
%
O2
0.2
%
0.2
%
CH4
NA
2.8
%
70.4
%
CO+H2 Slag C in slag Flue ash C in flue ash
85.6 2.430 0 1.140 8.0
% kg/h % w/w Kg/h % w/w
CO+H2 Slag C in slag Flue ash C in flue ash
1.400 8,5 2.000 18.0
kg/h % w/w kg/h % w/w
Table 2. Seyitömer coal gasification test results Depending on these tests, formulations were created and now conceptual design of gasifiers and following process stages are being calculated (Table 3) to utilize in feasibility studies.
Table 3. Soma coal gasification calculation results K-T GASIFIER Consumption Coal
10.000
Kg/h
C
43.76
%
H
3.19
%
O
14.96
%
N
0.84
%
S
0.3
%
Ash
37.74
%
H2O
2.21
%
Oxygen (98%)
4,060
Nm3/h
Quench water
3,195
kg/h
10,123
Nm3/h
CO2
11.76
%
CO
64.90
%
H2
22.48
%
N2
0.2
%
H2S
0.66
%
CO+H2
87.38
%
Products Gas
K-T GASIFIER Consumption Coal
11.959
Kg/h
C
39.5
%
H
3.8
%
O
15.4
%
N
0.5
%
S
3.2
%
Ash
29.6
%
H2O
8.0
%
Oxygen (98%)
4.329
Nm3/h
Quench water
3.710
kg/h
Products 12.689
Nm3/h
CO2
12.10
%
CO
55.27
%
H2
28.87
%
N2
1.07
%
H2S
1.91
%
COS
0.20
CO+H2
85.6
%
1.238
kg/h
Gas
Slag
GASIFIER SELECTION Most important criteria in selection of gasifier type is ash melting point. This should be made under reducing atmosphere (CO+CO2). Atmospheric gasification can easily be made and problems of pressure gasification can be avoided. CONCEPTUAL DESIGN A block scheme for SNG production is given in figure 1 and consumption and production figures are given in Table 3. and same for the production of ammonia is given in figure 2 and table 4.
Figure 1. SNG production block scheme
COAL
GASIFIERS
HRSG
AIR SEPERATION
AIR
METHANE SYNTHESIS
ASH
METHANE
Table 4. SNG from coal Consumptions Coal Electric Fresh water
: 600 t/h : 132,000 kwh/h : 4,700 m3/h
Production Ammonia
: 800,000,000 Nm3/year
GAS WASHING COOLING
GAS PURIFICATIO N
CO2
H2S REMOVAL
CO CONVERSION
Figure 2. Ammonia production block scheme
COAL
GASIFIERS
HRSG
AIR SEPERATION
AIR
AMMONIA SYNTHESIS
ASH
AMMONIA
GAS WASHING COOLING
GAS PURIFICATIO N
H2S REMOVAL
CO CONVERSION
CO2
Table 5. Ammonia from coal Consumptions Coal Electric Fresh water Steam
: 100 t/h : 66,000 kwh/h : 1,030 m3/h : 60,000 kg/h
Production Ammonia
: 1,000 t/d
CONCLUSIONS Production of chemicals from coal is nowadays economic when it is compared with natural gas as feedstock. Ammonia production from coal was stopped 10 years ago because of high cost of production when compared to imported ammonia and to ammonia from natural gas. Now Turkey has no ammonia production ability with his natural resources. Care should be taken urgently to build an coal based ammonia plant to produce fertilizer and explosives.
Methane Cracking over De-ashed Coal Chars and the Effect of the De-ashing Conditions †‡
†
†
†
†
Ling Wei , Yisheng Tan* , Yizhuo Han , Hongjuan Xie , Jiantao Zhao , §
Jinhu Wu*, and Dongke Zhang †
Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China ‡
Graduate School of Chinese Academy of Sciences, Beijing 100039, China
* Qingdao Institute of Bioenergy and Bioprocessing Technology, Chinese Academy of Sciences, Qingdao 266101, PR China §
Centre for Energy (M473), The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia
Abstract: Methane cracking over different de-ashed coal chars derived from the same parent coal (Xiao-long-tan lignite) was studied using a fixed-bed reactor operating at atmospheric pressure and 1123 K. The first set of samples of Xiao-long-tan lignite chars were prepared by heating the coal 1173 K in nitrogen for 30 minutes and then washing the resulting char with 5 N HCl (char 1), 5 N HCl and 29 N HF (char 2), respectively. The second set of samples of Xiao-long-tan lignite chars were prepared by washing the coal with 5 N HCl, 5 N HCl and 29 N HF, respectively and then pyrolyzing the de-ashed coal samples in nitrogen, also at 1173 K for 30 min (char 3 and char 4). Comparing to blank experiments using quartz particles, the chars were shown to have a significant catalytic effect on methane cracking. Hydrogen was the primary gas-phase product of methane cracking. Different chars also showed different catalytic activities in methane cracking. Chars 1 and 2 were shown to have higher catalytic activities than chars 3 and 4. These observations were also confirmed by additional experiments using a temperature-programmed desorption coupled with mass spectroscopy (TPD-MS), using methane, instead of nitrogen. Chars 1 and 2 were found to be more porous than chars 3 and 4, which is speculated as the major cause of the difference in their catalytic effect in methane cracking.
Keywords: methane cracking; coal char; de-ash; hydrogen
* Corresponding author: Prof. Tan, Tel.: +86-351-4044287; E-mail address: [email protected]
1. Introduction Methane cracking to produce hydrogen over a bed of coal char may offer an effective means for utilization of coal-bed methane (CBM) as coal and coal-bed methane are co-located and coal char has been observed to exert a catalytic effect on methane cracking 1
. There are several conventional methods of hydrogen production from methane such as
steam reforming of methane (SRM) and partial oxidation reforming (POR). SRM is a highly endothermic process and requires substantial external heating. The POR is an exothermic reaction and requires pure oxygen supply. Both SMR and POR produce large amounts of CO (and CO2) which has to be converted to CO2 and removed using the water-gas shift. The direct catalytic decomposition of methane has been considered as an attractive alternative. Compared with other reforming of methane, it is simple, energy-efficient and without the need of CO2 and CO removal. In methane cracking, high purity hydrogen can be produced and the produced clean carbon may also have a reasonable commercial value.8, 9 Methane can undergo thermal decomposition without a catalyst but this process requires a high operating temperature in the order of 1500-2000 K8, 10 . Catalytic decomposition can reduce the required temperature of methane cracking. Transition metals (such as Ni, Fe, Co, Pb) showed superior catalytic activity in decomposition of methane, but they can be rapidly deactivated owing to carbon from methane cracking accumulating on the catalyst surface. The catalysts have to be re-generated by burning off the deposited carbon.4, 8, 11, 12 The catalytic decomposition of methane and other hydrocarbons over carbon has been proposed as a viable method compared with the other methods, offering advantages including (i) high temperature resistance, (ii) no metal carbides formed, (iii) production of a marketable byproduct carbon which could substantially reduce the net cost of hydrogen production, and (iv) no need of regeneration of the carbon catalyst.10,
13
Muradov
investigated various kinds of carbon, such as activated carbon, carbon blacks, glassy carbon, graphite, diamond, carbon fibers and carbon nanotubes as the catalyzers for methane cracking9, 10, 13-15 . Our previous work has also found that the coal chars, rather than the ash, have a profound catalytic effect on methane cracking. In the reaction, methane decomposition over coal chars showed a reaction order of 0.5 and the activation energies from 89 to 105 kJ mol-1 which is much lower than the methane C-H bond
dissociation energy about 440 kJ mol-1,16, 17. In order to ascertain the catalytic effect of the different de-ashed coal chars and the process of methane decomposition, the present work studied methane cracking over de-ashed coal chars prepared by different de-ashing processes.
2. Experimental A Chinese coal (Xiao-long-tan lignite) was employed in this work. The Xiao-long-tan lignite char was prepared by devolatilising the air-dried Xiao-long-tan lignite in nitrogen at 1173 K for 30 minutes. In order to study the effect of ash removal on the effect of char on methane cracking, two sets of de-ashed char samples were prepared. The first set of samples of chars were prepared by heating the coal 1173 K in nitrogen for 30 minutes and then washing the resulting char with 5 N HCl (char 1), 5 N HCl and 29 N HF (char 2), respectively. The second set of samples of chars were prepared by washing the coal with 5 N HCl, 5 N HCl and 29 N HF, respectively and then pyrolyzing the de-ashed coal samples in nitrogen, also at 1173 K for 30 min (char 3 and char 4). The chars prepared were sieved to a size fraction of 0.355-0.63 mm for the experiment. The methane cracking experiments were carried out in a vertical quartz tube fixed bed reactor as shown in Figure 1. The reactor has a diameter of 25 mm and a height of 620 mm, housed in an electrically heated furnace. The flow rates of methane and nitrogen were controlled using mass flow controllers. The methane and nitrogen gases used in the experimentation were of high purity analytical grade gases. In a typical experimental run, a char of about 10 g was weighed and placed into the quartz tube and heated in nitrogen. The nitrogen was continuously flowed through the reactor at 160 ml min-1 while the furnace was heated to a desired temperature of 1123 K. When the temperature of the reactor reached the desired value and stabilized, the gas was switched to the reactant gas mixture of methane and nitrogen at a volumetric ratio of 1:4 (CH4:N2). The flow rate of nitrogen was also 160 ml min-1, but the flow rate of methane was 40 ml min-1. Allowing 3 min after the gas switching, gaseous samples of the exit stream were taken periodically and analyzed using a gas chromatograph (GC-14C) fitted with a molecular sieve column and a thermal conductivity detector (TCD), using Ar as the carrier gas for H2 measurement and H2 for hydrocarbons, respectively.
The methane adsorption and desorption on the char was also investigated using a temperature-programmed desorption coupled with mass spectroscopy (TPD-MS). The amount of char sample used in each experiment was ~0.1 g. The furnace was heated to 1073 K and then reduced to 373 K with Ar used as the protective gas. After 30 min, the Ar was replaced by mixture gas (CH4:N2=4:1). Then the char was heated in the gas mixture to 1073 K at 7 K min-1. The MS
measured the composition of the gases desorbed from
the char, which may contain H2, CH4, H2S, CO2, SO2 and so on. In addition, the superficial morphologies of the coal chars before and after methane decomposition were examined using a NOVA NANO, SEM 430 scanning electronic microscope (SEM), the samples were observed under an acceleration voltage of 20.0 kV. The amplificatory multiple is 5000. And the specific surface area and pore structure properties are determined using a TriStar 3000 physical adsorption apparatus using N2 sdsorption at 77 K. The conversion of methane can be calculated according to the following equation:
X CH 4
VinCCH 4 ,in Vout CCH 4 ,out VinCCH 4 ,in
and the hydrogen yield is expressed as:
YH 2
Vout CH 2 ,out 2VinCCH 4 ,in
where “C” is the conversion of CH4 or H2 and the subscripts “in” and “out” refer to reactor inlet and outlet conditions, respectively.
Figure 1. A schematic diagram of the experimental system: (1) mass flow controllers, (2) mixing chamber, (3) temperature controller, (4) electrically heated furnace, (5) quartz tube reactor.
3. Results and discussion 3.1 Methane Decomposition over Different Coal Chars It is well known that methane molecules are highly stable and could only crack or decompose at high temperatures, usually greater than 1303 K.18 In a blank experiment, methane cracking on a bed of quartz particles of the same size fraction as the char was studied. The results are shown in Figure 2. The highest value of methane conversion achieved over quartz was less than 4.3 % at 1123 K. However, the methane conversion over the chars was observed to be higher than 70%, as also shown in Figure 2. Figure 2 also presented the results of methane cracking over different de-ashed coal char. It can be clearly seen that the methane conversions were different on the different coal char. The best coal chars were char 1and char 2 at the same temperatures. It proved that the ash of the coal char did not have a catalytic effect on methane cracking. Figure 3 displayed the results of hydrogen yield over the different de-ashed coal chars at the same temperatures. It can be seen that the hydrogen yield decreases with increasing the time in the stream. The trend of hydrogen yield was similar to that of methane conversion. The highest hydrogen yield over quartz was 0.13 %, but over the coal chars, about 52 %.
100
Xiao-long-tan lignite char char 1 char 2 char 3 char 4 quartz
CH4 conversion (%)
80
60
40
20
0 0
20
40
60
80
100
120
140
Time (min)
Figure 2. Methane conversion on the different coal char and quartz(T =1123 K, CH4:N2 = 1:4, total flow rate 200 ml/min)
100
Xiao-long-tan lignite char char 1 char 2 char 3 char 4 quartz
H2 yield (%)
80
60
40
20
0 0
20
40
60
80
100
120
140
Time (min)
Figure 3. Hydrogen yield on the different coal char and quartz(T =1123 K, CH4:N2 = 1:4, total flow rate 200 ml/min)
The methane decomposition on the coal char was also investigated using the temperature-programmed desorption coupled with mass spectroscopy. Figure 4 shows the results of methane cracking over the quartz in TPD-MS. The figure displays that the signals of H2, CO2, H2S and SO2 were about zero. The content of methane was still stable. The methane cannot crack over the quartz below 1073 K. The hydrogen signal on the MS was the same as that in methane decomposition in the fixed bed experiments. The methane started cracking from 850 K over the chars, which is much lower than the temperature of the methane cracking without a catalyst. So the coal chars are effective catalysts for methane decomposition. The signals of the hydrogen increased with increasing reaction temperature and the chars 1 and 2 provoked greater hydrogen yields. It proved that the best catalysts were char 1 and char 2. The results of the methane decomposition on the different de-ashed coal chars were shown in the Figure 5.
2.00E-009
H2
Gas singal
1.50E-009
CH4 H2S 1.00E-009
CO2 SO2
5.00E-010
0.00E+000 300
400
500
600
700
800
900
1000
1100
Temperature (K)
Figure 4. Gas signals of the methane cracking over the quartz on TPD-MS
1.20E-010
char 1 char 2 char 3 char 4
1.00E-010 8.00E-011
H2
6.00E-011 4.00E-011 2.00E-011 0.00E+000 300
400
500
600
700
800
900
1000
1100
Tempreature (K)
Figure 5. Hydrogen signals of methane cracking over the different coal chars in TPD-MS
Figure 6 shows the CO2 signals of methane cracking on the different coal chars. It can be seen that the carbon dioxide was produced at temperature above 500 K. With the temperature increasing, the content of carbon dioxide was firstly enhancive, and then moderative. After 1073 K, the content of carbon dioxide was about zero over the bed of the chars 1 and 2. But it was much more on the chars 3 and 4.
8.00E-010
CO2
6.00E-010
char 1 char 2 char 3 char 4
4.00E-010
2.00E-010
0.00E+000 300
400
500
600
700
800
900
1000
1100
Temperature (K)
Figure 6. Carbon dioxide signals of the methane caracking over the different coal char on TPD-MS
3.2 Char Characterization The above experimental results have demonstrated that the coal chars possess the catalytic effect, promoting methane cracking. At the same time, the results presented so far consistently show a decreasing trend with increasing reaction time. The phenomena indicated that the chars rapidly lose the catalytic activity in methane cracking. In order to speculate the mechanisms of the char deactivation, a series of scanning electron microscopy image analyses were performed on the chars before and after the reaction. Figure 7 shows the SEM images of different chars before and after the methane cracking. It is evident that the fresh chars have a very clear surface and clean porous structure with clear edges, but the deactivated char after methane cracking is covered with carbon deposits. The surfaces of the char 1 and char 2 have more porous structures than the surfaces of char 3 and char 4. The BET surface areas and pore characteristics of the fresh coal chars are detailed in Table 1. It can be seen that the total surface area decreased in the order from char 1 to char 4. The charge in the char surface area is consistent with the methane conversion over the char. It appears that the catalytic effect increases with increasing total surface area of the char and has no correlation with the micropore area. It is speculated that the total area surface is the main reason of the catalytic effect on methane cracking.
Figure 7. SEM images of chars before (A, B, C and D) and after (a, b, c and d) 123 min being subjected to methane cracking on the chars from the different de-ashed processes (N2:CH4 = 4:1, total flow rate 200 ml min-1, A and a: char1, B and b: char 2, C and c: char 3, and D and d: char 4) Table 1. Variations in the surface properties of the different coal chars
Sample BET Area, m2/g Micropore Area m2/g Total Volume, cm3/g Micropore Volume cm3/g Average Pore Diameter nm
Char 1
Char 2
Char 3
Char 4
Before reaction
Before reaction
Before reaction
Before reaction
44.6
38.5
26.6
14.9
16.5
12.5
22.15
4.95
0.048
0.050
0.0225
0.0205
0.007
0.006
0.01
0.002
4.2
5.2
3.3
5.4
4. Conclusions Methane cracking experiments have been performed over coal chars prepared using different de-ashing methods and quartz in a fixed-bed reactor at 1123 K and temperature-programmed desorption coupled with mass spectroscopy. It has been clearly demonstrated that the chars has a profound catalytic effect and methane does not crack on quartz. Methane conversion and hydrogen yield decrease with increasing reaction time, but increase with the reaction temperature. The best catalyst is char 1. The ash in the coal char is not responsible for the catalytic effect. Carbon deposition on the char surface
during methane cracking is responsible for the loss of catalytic activity of the char. The catalytic activity of the char in methane cracking is consistent with the total surface area of the char.
References 1.
Sun, Z.; Wu, J.; Zhang, D., CO2 and H2O gasification kinetics of a coal char in the presence of methane.
Energy & Fules 2008, 22, (4), 2160-2165. 2.
Haghighi, M.; Sun, Z.; Wu, J.; Bromly, J.; Wee, H.; Ng, E.; Wang, Y.; Zhang, D., On the reaction
mechanism of CO2 reforming of methane over a bed of coal char. Proceedings of the Combustion Institute 2007, 31, 1983-1990. 3.
Li, Y.; Xiao, R.; Jin, B.; Zhang, H., Experimental study of the reforming of methane with carbon dioxide
over coal char. International Journal of Chemical Reactor Engineering 2008, 6, 15. 4.
Chen, J.; He, M.; Wang, G.; Li, Y.; Zhu, Z., Production of hydrogen from methane decomposition using
nanosized carbon black as catalyst in a fluidized-bed reactor. Int. J. Hydrog. Energy 2009, 34, (24), 9730-9736. 5.
Song, Q.; Xiao, R.; Li, Y.; Shen, L., Catalytic carbon dioxide reforming of methane to synthesis gas over
activated carbon catalyst. Industrial & Engineering Chemistry Research 2008, 47, (13), 4349-4357. 6.
Xu, Z; Wu, J.; Wang, Y.; Zhang, D., Methane cracking over lignite char. Journal of Fuel Chemistry and
Technology 2009, 37, (3), 277-281. 7.
Sun, Z.; Wu, J.; Haghighi, M.; Bromly, J.; Ng, E.; Wee, H.; Wang, Y.; Zhang, D., Methane cracking over a
bituminous coal char. Energy & Fules 2007, 21, (3), 1601-1605. 8.
Yoon, S. H.; Park, N. K.; Lee, T. J.; Yoon, K. J.; Han, G. Y., Hydrogen production by thermocatalytic
decomposition of butane over a carbon black catalyst. Catal. Today 2009, 146, (1-2), 202-208. 9.
Muradov, N. Z.; Veziroglu, T. N., From hydrocarbon to hydrogen-carbon to hydrogen economy. Int. J.
Hydrog. Energy 2005, 30, (3), 225-237. 10. Lee, S. Y.; Ryu, B. H.; Han, G. Y.; Lee, T. J.; Yoon, K. J., Catalytic characteristics of specialty carbon blacks in decomposition of methane for hydrogen production. Carbon 2008, 46, (14), 1978-1986. 11. Gac, W.; Denis, A.; Borowiecki, T.; Kepinski, L., Methane decomposition over Ni-MgO-Al2O3 catalysts. Appl. Catal. A-Gen. 2009, 357, (2), 236-243. 12. Kim, M. H.; Lee, E. K.; Jun, J. H.; Kong, S. J.; Han, G. Y.; Lee, B. K.; Lee, T. J.; Yoon, K. J., Hydrogen production by catalytic decomposition of methane over activated carbons: kinetic study. Int. J. Hydrog. Energy 2004, 29, (2), 187-193. 13. Muradov, N.; Smith, F.; T-Raissi, A., Catalytic activity of carbons for methane decomposition reaction. Catal. Today 2005, 102, 225-233. 14. Muradov, N., Catalysis of methane decomposition over elemental carbon. Catal. Commun. 2001, 2, 89-94. 15. Muradov, N., Hydrogen via methane decomposition: an application for decarbonization of fossil fuels. Int. J. Hydrog. Energy 2001, 26, (11), 1165-1175. 16. Bai, Z.; Chen, H.; Li, W.; Li, B., Hydrogen production by methane decomposition over coal char. Int. J. Hydrog. Energy 2006, 31, (7), 899-905. 17. Zhang, Y.; Wu, J.; Zhang, D., Cracking of simulated oil refinery off-gas over a coal char, petroleum coke, and quartz. Energy & Fules 2008, 22, (2), 1142-1147. 18. Gueret, C.; Daroux, M.; Billaud, F., Methane pyrolysis: Thermodynamics. Chem. Eng. Sci. 1997, 52, (5), 815-827.
CeO2-K2O promoted Co-Mo sulfur-tolerant shift catalyst for the shift reaction of CO in coke oven gas Yuqiong ZHAO, Yongfa ZHANG () , Guojie ZHANG Key Laboratory of Coal Science and Technology of Shanxi Province and Ministry of Education, Taiyuan University of Technology, Taiyuan 030024, China Abstract To avoid energy and H2 consumption in the process of CH4-CO2 reforming (CH4 in coke oven gas and CO2 in gasification gas), a method was proposed, which through CO shift reaction makes CO in coke oven gas converted and simplifies the separation of CH4 and H2. In this work, CeO2-K2O promoted Co-Mo sulfur-tolerant shift catalyst for the shift reaction of CO in coke oven gas is investigated. The results indicate Ce and K have a synergy effect on promoting the catalytic activity, and the Co-Mo-Ce-K/γ-Al2O3 catalyst with 3.0 wt-% CeO2 and 6.0 wt-% K2O exhibits the highest activity. Moreover, CeO2 contained Co-Mo catalysts have higher catalytic activity with low steam/COG ratio, demonstrating that introduced CeO2 increases the water adsorb ability of catalyst. Keywords coke oven gas, water gas shift, sulfur-tolerant catalyst, cerium dioxide
1.Introduction In the “dual-gas resources” polygeneration system based on coke oven gas (CH4) reforming of gasification gas (CO2) [1, 2], a large number of H2 in coke oven gas (COG) is heated from ambient temperature to above 950℃, which consumes substantial energy of compression, transportation and heating [3]. Moreover, a quantity of H2 is wasted due to the oxidation of H2 [4]. Through CO shift reaction (CO+H2O→CO2+H2) converting the CO in COG into H2 and CO2, the product gas will be composed mainly of H2 and CH4, which can be easily separated into H2 and CH4-rich gas. The CH4-rich gas reforming of CO2 would save the energy and reduce H2 consumption. Based on the ideas above, the CO shift catalysts and shift properties of CO in COG are investigated. Coke oven gas contains high concentration of tar and multiple forms of sulfur, which requires the CO shift catalyst should be more active, more E-mail: [email protected]
sulfur-tolerant and more anagotoxic. Compared to the Fe-Cr and Cu-Zn catalysts, Co-Mo catalysts are more active and can be applied to a wider temperature range without sulfur poisoning [5]. CeO2 with a unique electronic structure has excellent oxygen-storage capacity (OSC). Gorte R. J. et al. [6] found a close relationship between CO shift activity and OSC. To further increase the conversion of CO in COG, CeO2 promoted Co-Mo sulfur-tolerant shift catalyst is investigated.
2. Experimental 2.1.
Catalyst preparation CoMo-CemKn, CoMo-Cem and CoMo-Kn catalysts were prepared by incipient
wetness impregnation of γ-Al2O3 with aqueous solution of metal salts according to certain sequence, where m and n represent mass ratio of promoter/catalyst respectively. The impregnation was followed by drying at 120℃ for 2 h and calcinating at 450℃ for 2 h. All obtained catalysts contain 3.0 wt-% of CoO and 10.0 wt-% of MoO3. FB301 is a commercial catalyst. 2.2.Catalytic activity measurement Catalytic activity tests were carried out with 15.0 ml of catalyst (diameter 3.0~ 4.0 mm) in a fixed-bed reactor. Prior to activity tests, the samples were sulfided in situ at 300℃ for 4h with a mixture of CO/H2/H2S (45.0/50.0/5.0, v/v/v). The reaction was conducted under the conditions: feed gas composition CO/CH4/H2/H2S = 9.7/30.0/60.0/0.3 (vol-%), steam-to-COG ratio = 0.5, COG space velocity = 2000 h−1, reaction pressure = 0.1 MPa. CO content was analyzed by using a gas chromatograph equipped with a thermal conductivity detector (a 13X molecular sieve column, H2 as the carrier gas). The conversion of carbon monoxide XCO is given by: XCO = (Vco-Vco')/Vco(1+Vco')×100% Where Vco and Vco' are the volume fraction of CO in the gas at the reactor inlet and outlet, respectively.
3.Results and discussion 3.1.Effect of CeO2 content on the catalytic activity of Co-Mo-Ce/γ-Al2O3
Conversion of CO/%
80
250OC 350OC
70 60 50 40 30
0
1
2
3
4
CeO2/%
5
6
7
8
Fig.1 Influence of CeO2 content on the catalytic activity of Co-Mo-Ce/γ-Al2O3
Figure 1 shows the catalytic performances of Co-Mo-Ce/γ-Al2O3 catalysts. It can be seen that CeO2 has a promoting effect on the activity of CoMo-Cem catalysts. When the CeO2 content is raised to 3.0 wt. % at 350℃, the conversion of CO reaches a maximum value, 77.6%. Hereafter, the catalytic activities decrease with CeO2 content increasing. As shown in Table 1, when the content of CeO2 on the CoMo-Cem catalyst is replaced by the same content of K2O, the obtained CoMo-Kn catalysts exhibit lower activities than that of the CoMo-Cem catalysts (m and n≦3). Furthermore, the Ce/K atom ratio in CoMo-Cem and CoMo-Kn catalysts is only 0.265, indicating that adding a little CeO2 can increase the catalytic activity of CoMo catalyst. This phenomenon may be due to valence variability and excellent oxygen-storage capacity of CeO2 [6, 8] which strengthens the transfer process of S-O and promotes CO shift reaction. Although the role of CeO2 is significant, the promotion of CeO2 is limited. The fact suggests that CeO2 can not serve as the main additive used in the catalyst, and should coordinate with other additives to further increase the conversion of CO in COG. Table 1 Comparison of CoMo-Cem and CoMo-Kn catalytic activity Content (wt%) Sample
CO conversion
(%)
K2O
CeO2
250℃
350℃
CoMo-Ce1
0
1
45.80
62.60
CoMo-K1 CoMo-Ce3
1 0
0 3
40.80 57.87
58.86 77.60
CoMo-K3
3
0
54.27
75.80
3.2 Effect of CeO2-K2O on the catalytic activity of Co-Mo-Ce-K/γ-Al2O3 90
Conversion of CO /%
80 70 60 50
CoMo-Ce3K6 CoMo-Ce1K8 CoMo-K9 CoMo-K8 CoMo-K6 CoMo-Ce3
40 30 20 200
250
300
350
400
Temperature/C O
Fig. 2
Figure
Influence of CeO2-K2O on the catalytic activity of Co-Mo-Ce-K /γ-Al2O3
2
shows
the
influence
of
CeO2-K2O
on
the
activity
of
Co-Mo-Ce-K/γ-Al2O3 catalysts under different reaction temperatures. It can be seen that the activities of all CoMo-CemKn catalysts increase rapidly with reaction temperature increasing,and CO conversion reaches an optimum value at 300℃. Hence, the catalytic activities decrease with the increase of reaction temperature, which may be limited by thermodynamic equilibrium. CeO2 and K2O promote the catalytic activity better comparing with each along, especially at the lower temperature. When CeO2 is added into CoMo-K6 and CoMo-K8 catalysts respectively, the conversion of CO greatly increases. CoMo-Ce3K6, CoMo-Ce1K8 and CoMo-K9 catalysts have the same content of total promoter (m + n = 9). However, CoMo-Ce3K6 and CoMo-Ce1K8 catalysts exhibit higher activities than that of CoMo-K9 catalyst. The fact indicates that there is a synergy effect between Ce and K in the Co-Mo-Ce-K/γ-Al2O3 catalyst, which improves the catalytic activity obviously. The specific synergy mechanism is not clear yet now and the further research is needed. In addition, the optimum values of total promoter content and CeO2/K2O mass ratio are found. When the total mount of promoter CeO2-K2O reaches 9.0 wt-% and the mass ratio of CeO2/K2O reaches 0.5, the conversion of CO reaches 91.72% at 300℃, being 15.51% higher than that of CoMo-K9 catalyst.
Besides, no methane is detected in the product gas when the feed gas dose not contain CH4, demonstrating that CO methanation reaction dose not occur and CoMo-CemKn catalysts have good selectivity. 3.3 catalytic activity of CoMo-Ce3K6 under different atmosphere 90
Conversion of CO/%
80 70 60 50
a b c d
40 30
200
250
300
0 Temperature/ C
350
400
Fig.3 The catalytic activity of CoMo-Ce3K6 under different atmosphere a. CoMo-Ce3K6:H2O/Gas~0.25; b. CoMo-Ce3K6:Semi-water gas c.FB301:H2O/Gas~0.25; d. FB301:Semi-water gas
The activities of CoMo-Ce3K6 and FB301 catalysts with low steam/COG ratio and semi-water gas are presented in Figure 3. It shows that CoMo-Ce3K6 catalyst has better activity than FB301 catalyst at the two kinds of atmosphere, which demonstrates that Co-Mo-K-Ce/γ-Al2O3 shift catalyst used in energy-saving technology and ammonia synthesis also obtain good results. With the ratio of steam/COG decreasing, the activity of Ce3K6-CoMo and K9-CoMo catalysts have a slip. However, the descend degree of Ce3K6-CoMo catalytic activity is lower than that of K9-CoMo catalyst. X.M.Миначев [9] reported that a large number of oxygen which can combine with hydrogen at a certain temperature adsorbed on the surface of CeO2. Therefore, CeO2 contained Co-Mo catalyst may have higher adsorption capacity of H+ in H2O, which makes OH- enrich on the surface of catalyst and maintain a high conversion of CO with low steam/COG ratio.
Table 2 Influence of CeO2 on activity of the catalyst with a low steam gas ration Sample
CO conversion/%(H2O/Gas~0.25)
CO conversion/%(H2O/Gas~0.5)
300℃
400℃
300℃
400℃
CoMo-K9
70.27
64.47
76.60
81.37
CoMo-Ce3K6
86.22
77.63
91.72
84.91
4. Conclusion CeO2 has a promoting effect on the activity of Co-Mo/γ-Al2O3 catalyst. However, the promotion of CeO2 is limited, indicating that CeO2 coordinate with other additives would promote the catalytic activity better. In the Co-Mo-Ce-K/γ-Al2O3 catalyst, Ce and K have a synergy effect on promoting the shift reaction of CO in COG. The catalytic activity reaches a maximum when the total promoter mount of CeO2-K2O reaches 9.0 wt-% and the mass ratio of CeO2/K2O reaches 0.5. Moreover, with the ratio of steam/COG decreasing, the descend degree of CoMo-Ce3K6 catalytic activity is lower than that of CoMo-K9 catalyst, which can be interpreted by that the introduced CeO2 increases the water adsorb ability of catalyst.
Acknowledgements This work was supported by the National Basic Research Program of China (2005CB 221202) and Shanxi Provincial Natural Science Foundation (2010011014-1).
References [1] Xie K C, Zhang Y F. CN Patent, 1974732A, 2007-06-06. [2] Xie K C, Zhang Y F, Zhao W. The basic research of dual-gas resources multi-generation coke oven gas to syngas. Shanxi Energy and Conservation, 2008, 49: 10-12. [3] Tian C S, Zhang Y F. Optimization of process conditions for syngas production by non-catalytic partial oxidation of methane. In: 17th International Symposium on Alcohol Fuels. Taiyuan: 2008, 350. [4] Wang J C, Zhang S Y, Bai T Z. Technical evaluation on the process of methanol
syngas production based on coke oven gas. Coal Chemical Industry, 2006, 126: 48-51. [5] Li S Y, Zhou X Q. Latest progress in CO shift catalysts. Coal Chemical Industry, 2007, 129: 31-34. [6] Gorte R J, Zhao S. Studies of the water-gas-shift reaction with ceria-supported precious metals. Catalysis Today, 2005, 104: 18-24. [7] Yang Y Q, Fang W P. CN Patent, 1559679A, 2005-01-05. [8] Ye B H, Jiang L L. Effects of CeO2 content on structure and properties of Ni-Mn-K/Al2O3 multiple-metal catalysts for water-gas shift reaction. Journal of The Chinese Rare Earth Society, 2006, 24: 661-664. [9] X.M.Миначев et al.《Application of rare earth in catalysts》. Beijing: Science Press, 1987: 1-25.
Effects of preparation conditions on Ru/Al2O3 catalyst for coal-based syngas methanation reaction Liping Wang, Yongfa Zhang* ,Yaling Sun, Xianglan, Li Key Laboratory of Coal Science and Technology of Shanxi Province and the Ministry of Education ,Taiyuan University of Technology , Taiyuan 030024 , China Abstract: Effects of preparation conditions on Ru/Al2O3 catalyst for coal-based syngas methanation reaction have been studied. The catalysts have been characterized by XRD technique. Using of ultrasound impregnation method can significantly decrease impregnated time and improve catalytic activity of the catalyst. The catalyst prioritizing preparation conditions are:Ru loading of about 2%, calcination temperature of 500℃, and H2 reduction temperature of 400℃. Under the optimum catalysts conditions, the conversion of CO and the CH4 selectivity reach 97.18% and 83.29%, respectively. Besides, the catalyst washed with deionized water and diluted ammonia can remove chlorine ions and increase catalytic activity. Key words:Ru catalyst;coal to methane;methanation;preparation conditions
1. Introduction Since Sabatier and Senderens first discovered methane formation by the reaction of carbon monoxide and hydrogen over a nickel catalyst in 1902[1], the methanation reaction has been developed and widely used for the purification of hydrogen in ammonia and hydrogen plants. In recent years, the interest for the reaction has grown significantly as a result of recent advancements in clean coal technology and the demand for development of converting coal source into substitute natural gas(SNG). This was stimulated by a shortfall of natural gas supplies. The traditional catalyst for methanation is Ni supported on aluminum oxide[2,3] and Ru-based catalysts, they are also smaller loadings
[4–8]
. Many studies
[9,10]
have reported Ru or Ni catalyst
research for CO and CO2 methanation for purifying fuel cell or ammonia synthesis feed gas, but very few results for Ru-based catalyst for high CO concentrations about *Corresponding author. Tel. : +86 0351
6018676
E-mail address: [email protected]
coal-based syngas to natural gas. In the present work, a number of Ru catalyst with different metal loading supported on γ-Al2O3 were prepared and studied in the CO methanation reaction. Ru/γ-Al2O3 catalysts were prepared by wet incipient and ultrasonic impregnation method from RuCl3.nH2O precursor and their performance in CO methanation was investigated. From the catalysis test results, the best preparation conditions are found.
2. Experimental 2.1. Catalyst preparation The γ-Al2O3(Shanxi Fenyang Catalyst Plant produce)used as Ru carrier was prepared through crushed and sieved to 40-60 mesh, the material was then calcined in static air by heating to 500℃. The temperature was maintained for tree hour before cooling to room temperature. Two different methods of preparation catalysts were applied by using RuCl3.nH2O as precursor. One was that the materials xRu- Al2O3, with x = 1%, 1.5%, 2.0% or 2.5%, were prepared by impregnating Al2O3 with the aqueous solutions of RuCl3.nH2O of required concentrations for 24 hour. The other was prepared by the ultrasonic impregnation method. The xRu-Al2O3 was treated with a constant ultrasonic frequency of 40 kHz for 20 min, and then impregnated 12 hour. All solids were again dried and calcined under air at 500℃ for 3 hour. 2.2. Activity measurements A known mass of catalyst was loaded into a quartz tube reactor and capped with a quartz wool pulg. The catalyst was heated to 500℃ under flowing N2 (1000 ml. min -1) and H2 (50 ml. min -1) with the consumption of hydrogen monitored by an on-line gas chromatograph (GC950). Once hydrogen consumption was complete, signifying reduction of the ruthenium particles, the reaction was initiated by passing CO (75 ml. min -1) and H2 (25 ml. min -1) over the catalyst in the temperature range of 180~500℃. The mixtures of reactant gases and products were periodically analysed on-line using a TCD chromatography which contained a 1m column packed with 5A molecular sieve. Prior to analysis, the effluent was passed through a water-trap in
order to remove reaction water. The calculation equation of CO conversion、CH4 and CO2 selectivity was followed: X CO =
S CO2 =
FCOin × YCOin − FCOout × YCOout × 100% FCOin × YCOin FCO2 out × YCO 2out FCOin × YCOin − FCOout ×Y COout
S CH 4 =
FCH 4 out × YCH 4out FCOin × YCOin − FCOout × YCOout
× 100%
× 100%
where X CO : CO conversion, S : the selectivity of the CH4 or CO2, F : the gas flow rate of in and out, ml. min -1, Y : different fraction volume percentage, Superscript in : the inlet, superscript out : the outlet.
2.3 Catalyst characterization The crystalline structure of mixed oxides was determined by powder X-ray diffraction using Shimadzu XRD-6000 apparatus equipped with a monochromator for the Cu radiation at 40 kV and 30 mA. The 2θ scans covered the range from 20°to 80°.
3. Results and discussion 3.1 Effect of different impregnation methods on the catalysts activities
CH4 Selectivity, %
60 50
1# 2#
40 30 20 10 0
400
440
Tempreture, °C
480
Fig. 1 Catalytic activity of Ru/Al2O3 catalysts prepared with different impregnation methods 1#-0.5%Ru/Al2O3(ultrasonic impregnation),2#-0.5%Ru/Al2O3(wet impregnation)
As can be seen from Fig.1, in the same experimental conditions, the catalyst catalytic activity prepared by ultrasonic impregnation method was higher than that prepared by wetness impregnation. CH4 selectivity was 53.9% over the 2#-0.5%Ru/Al2O3 catalyst at 480 ℃ , CH4 selectivity was reached 63.1% over 1#-0.5%Ru/Al2O3 catalyst, increased by 9.2%. The results indicate that the ultrasonic wave has an important influence on the catalyst activity. It may be the effects of
ultrasonic cavitation of ultrasonic wave to impregnating solution. The shock waves generated by cavitation form a certain pressures in the catalyst surface and internal. These pressures act on catalyst carrier surface, which make Ru metal uniformly and quickly distributed in the matrix. The ultrasonic impregnation can not only shorten the impregnation time but also can improve the catalytic activity which has also been reported by other authors [11, 12] 3.2 Effect of calcination temperature on the catalysts activities 100
(B) 100
80
80
CH4 selectivity, %
Convertion of CO, %
(A)
60 40 20 0 200
250
300
350
300°C 400°C 500°C 600°C
calcination calcination calcination calcination
400
450
60
40
300°C 400°C 500°C 600°C
20
0
500
200
250
Tempreture,℃
300
350
calcination calcination calcination calcination
400
450
500
Tempreture,℃
Fig.2 Effects of calcination temperature of Ru/γ-Al2O3 on catalyst performance
RuO2 γ-Al2O3
300℃
400℃
500℃
600℃ 0
10
20
30
2θ, 40
50
60
ã
70
80
90
2θ, ° Fig. 3 XRD patterns of Ru/Al2O3 catalysts under different calcined temperatures
Figure 2 shows that the different calcination temperature has effect on the Ru/γ-Al2O3 catalyst performance. It can be seen from Fig. 2 (A) that CO conversion increases and then decreases rapidly with the calcination temperature increasing. When the reaction temperature was increased from 300℃ to 600℃, CO conversion was decreased rapidly from 98.4% to 26.5%, decreased by 71.9%. This may be that the higher calcination temperature easily leads to the catalyst pore volume reduced. As
a result, the metal grain growth, Ru dispersion was decreased, the catalytic activity was decreased. It also can be seen the XRD spectra of the different calcination temperature Ru/γ-Al2O3 catalyst, Fig. 3. With the increasing of calcination temperature, the characteristic peaks of γ-Al2O3 gradually increase. It indicates that more γ-Al2O3 explored in surface area, that is, the surface active sites amount is reduced. Because of high temperature calcinations, the metal catalysts have a certain degree of sintering, the catalysts active sites and catalytic activity decreased. The effect of calcination treatment on catalysts had been previously reported in the literature [13]. Figure 2 (B) shows the effect of calcination temperature on the CH4 selectivity. When the calcinated temperature was increased from 300℃ to 500℃, the CH4 selectivity increased significantly from 75.6% to 98.3%, with a growth of 22.7%. In the lower calcination temperature, the catalyst can not form a suitable active phase, detriment of reaction. It also may be due to RuCl3 decomposition incomplete, the residual Cl-1 affected the catalyst activity, which was proved by PH test paper after experiment. However, when the temperature was further up to 500℃, the CH4 selectivity decreased. Comprehensive consideration CO conversion and methane selectivity, the optimum calcination temperature is 500℃ which is different from literature reported [14]. 3.3 Effect of reduction temperature on the catalysts activities (B) 100 80
80
60
40 300°C reduction 400°C reduction 500°C reduction
20
0
200
250
300
350
400
Tempreture,℃
450
500
CH4 selectivity, %
Convertion of CO, %
(A) 100
60
40 300°C reduction 400°C reduction 500°C reduction
20
0
200
250
300
350
400
450
500
Tempreture,℃
Fig. 4 Effects of reduction temperature on Ru/Al2O3 catalyst for CO methanation
Four different Ru/Al2O3 catalysts samples were reduced by hydrogen at 300、400 and 500℃, respectively. The reduction temperature effecting on the catalytic performance can be seen from Fig. 4(A) and (B). As shown in Fig. 4, the reduction temperature has obvious influence on catalytic performance of catalysts. The reduction temperature was heated from 300 to 400℃, the conversion of CO increased from 66.8% to 90.2%, while the selectivity of CH4 gradually increased from 60.6% to 80.3%. However, the reduction temperature further increased to 500℃, CO conversion gradually decreased to 54.5%. It also found that, with the reduction temperature increasing, CO2 concentration in export gas increases. The optimum reduction temperature is 400℃. 3.4 Effects of Ru metal loading on catalytic performance The effects of metal loading on catalytic activity have been investigated over catalysts with variable Ru content (1.0~2.5wt%), the results are shown in Fig. 5. 100
80
CH4
Selectivity, %
80
60
60
40
40 20
20
CO2
RuO2 γ-Al2O3
Convertion of CO, %
100
CO
5.0%Ru 2.5%Ru 2.0%Ru 1.5%Ru 1.0%Ru 0.5%Ru
0
0
1
2
3
4
5
0
Ru loading, %
Fig. 5 CO methanation over Ru/Al2O3 catalysts with different Ru loading
10
20
30
40
50
60
70
80
90
2θ, °
Fig. 6 XRD patterns of Ru/Al2O3 catalysts with different Ru loadings
As can be seen from Fig. 5, with the Ru loading increases, CO conversion and methane selectivity increased significantly. When the Ru content arrived at 2.0%, CO conversion and methane selectivity were 98.33% and 81.30% respectively. As the Ru content increased continuously, CO conversion did not change significantly. While methane selectivity decreased slightly with by-product of the selectivity of CO2 nearly unchanged. It is believed this phenomenon is due to the excess of ruthenium content,
uneven distribution on the carrier surface, and thus the catalytic activity decreased caused by metal crystal pile-up [15]. Figure 6 shows the Ru/γ-Al2O3 catalyst XRD spectra of different ruthenium loadings. It can be seen from Fig. 6, increasing the ruthenium metal content from 0.5% to 5.0 wt%, the intensity of the RuO2 peak is increased. When the ruthenium loadings of γ-Al2O3 was 5%, from the XRD curve, two peaks were found to at 57.3° and at 59.2 ° respectively, which could be assigned to the overload RuO2 accumulation in the carrier surface. Based on these results, the 2% Ru/Al2O3 catalyst is thought to be optimal when the thermal stability and metal dispersion are comprehensively considered, which can attractive more CO molecules to the catalyst active center by adsorption, dissociation and activation from all directions [16]. A result was slightly different from that previously published [17]. 3.5 Effect of Cl- on the catalysts activities The influence of washing in different ways on residual chlorine on catalyst activity has been studied. Sample 1 is named Ru/Al2O3(L), the sample without any treatment to elucidate the effect of residual chlorine; Sample 2 is named Ru/Al2O3(M), the sample washed with deionized water to remove chlorine ions; Sample 3 is named Ru/Al2O3(P), the sample washed with deionized water and diluted ammonia to remove chlorine ions. Then all catalysts samples were dried at 110℃ for 12h and reduced in H2 flow at 400℃ for 2h. Tab. 1 the effect of Cl- on Ru /Al2O3 catalyst activity Output concentration of CH4 (%) Sample
washing agent 300℃
Ru/Al2O3(L) Ru/Al2O3(M)
none deionized water
Ru/Al2O3(P) deionized water and diluted ammonia
350℃
400℃
4.0
5.1
18.4
3.4
11.7
41.1
2.7
46.6
50.0
As seen from tab. 1, the catalyst Ru/Al2O3(P) exhibits the highest activity and selectivity, and the catalyst Ru/Al2O3(L) shows lowest catalytic activity. The results indicate that the existence of Cl- did not favor the improvement of the activity of the
catalyst, which was in agreement with previously reported in the literature [18, 19].
4. Conclusion The ultrasonic impregnation method not only promote the catalytic activities of Ru/Al2O3, but also speed up the proceeding of impregnation. By comprehensive consideration, calcination temperature 500℃, H2 reduction temperature 400℃, and Ru loading 2%, an ideal Ru/Al2O3 catalysts for CO methanation can be obtained. The catalyst leached and washed with distilled water and diluted ammonia is an effective way to remove chloride ions and improve Ru /Al2O3 catalysts activity.
Acknowledgements This work was supported by the National Basic Research Program of China (2005CB221202) and Shanxi Provincial Natural Science Foundation (2010011014-1).
References [1] Jan Kopyscinski, T. J. Schildhauer, Frédéric Vogel, et al. Applying spatially resolved concentration and temperature measurements in a catalytic plate reactor for the kinetic study of CO methanation [J]. Journal of Catalysis, 2010, 271(2): 262-279. [2] Geoffrey Colin Bond. Metal-catalysed reactions of hydrocarbons [M]. American, spring science, 2005, 441-448. [3] A.L. Kustova, A.M. Freya, K.E. Larsena, et al. CO methanation over supported bimetallic Ni–Fe catalysts: From computational studies towards catalyst optimization [J]. Applied Catalysis A: General, 2007, 320: 98-104. [4] Zhang Cheng. Research progress of methanation of carbon monoxide and carbon dioxide [J]. Chemical industry and engineering process, 2007, 26(9): 12169-1273. [5] M A. Vannice. The catalytic synthesis of hydrocarbons from H2/CO mixtures over the Group VIII metals(V):The catalytic behavior of silica supported metals[J]. Journal of Catalysis, 1977, 50 (2): 282-236.
[6] Takashi Amano, Atsusshi Takumi, Shugou Zhang, et al. Carbon monoxide removing catalyst and production process for the same as well as carbon monoxide removing apparatus [P].US:20060160697 A1,2006. [7] Karin Yaccato, Ray Carhart, Alfred Hagemeyer, et al. Competitive CO and CO2 methanation over supported noble metal catalysts in high through put scanning mass spectrometer [J]. Applied Catalysis A: General, 2005, 296: 30-48. [8] Huang Zhongtao. Handbook of industrial catalysts [M]. Beijing China: Chemical Industry Press, 2001, 705-707. [9] S Takenaka, T Shimizu, K Otsuka. Complete removal of carbon monoxide in hydrogen-rich gas stream through methanation over supported metal catalysts [J]. Inter J Hydrogen Energy, 2004, 29 (10) : 1065-1073 [10] V P Londhe, V S Kamble, N M Gupta. Effect of hydrogen reduction on the CO adsorption and methanation reaction over Ru/TiO2 and Ru/Al2O3 catalysts[J]. J Mol Catal A, 1997, 121 (1): 33-44. [11] Kenneth Suslick, Taeghwan Hyeon, Mingming Fang, et al. Cichowlas. Sonochemical synthesis of nanostructured catalysts [J]. Materials Science and Engineering A: General, 1995, 2004 (1/2): 186-192. [12] T. J. Mason, A. Newman, J. P. Lorimer, et al. Hutt. Ultrasonically assisted catalytic decomposition of aqueous sodium hypochlorite [J]. Ultrasonics Sonochemistry, 1996, 3(1) : 53-55. [13] V. Ragaini, R. Carli, C.L. Bianchi, et al. Fischer-Tropsch synthesis on alumina-supported ruthenium catalysts II. Influence of morphological factors [J]. Applied Catalysis A: General, 1996, 139(1/2) : 31-42 [14] LU Hong-xuan, QIN Bang-hui, SUN kun-peng,et al. Effect of pretreatment on catalytic performance of Ru / Al2O3 for methanation [J]. Natural Gas Chemical Industry (Chinese), 2004, 29(4) : 1-9. [15] Di Li, Nobuyuki Ichikuni, Shogo Shimazu, et al. Catalytic properties of sprayed Ru/Al2O3 and promoter effects of alkali metals in CO2 hydrogenation [J]. Applied Catalysis A: General, 1998, 172(2): 351-358. [16] C H Bartholomew, R B Pannell. The stoichiometry of hydrogen and carbon
monoxide chemisorption on alumina and silica-supported nickel [J]. Catalysis, 1980, 65(2): 390-401. [17] P Panagiotopoulou, D.I. Kondarides, X. E. Verykios. Selective methanation of CO over supported Ru catalysts [J]. Applied Catalysis B: Environmental, 2009, 88(3/4): 470 - 478. [18] T Narita, H Miura, M Ohira,et al. The effect of reduction temperature on the chemisorptive properties of Ru/Al2O3: Effect of chlorine [J]. Applied Catalysis A: General, 1987, 32:185-190. [19] Shuzo Murata, Ken-Ichi Aika. Removal of chlorine ions from Ru/MgO catalysts for ammonia synthesis [J]. Applied Catalysis A: General, 1992, 82(1/2): 1-12.
Manuscript Not AVAILABLE
THE PRODUCTION OF ORGANIC FERTILIZERS FROM GÖYNÜK, ILGIN AND ELBİSTAN LIGNITES WİTH H2SO4 OXİDATİON M.Çöteli, A.Güntürk, A.Yavuz, A.Köker, G.Yıldırım, S.Atlıhan General Directorate of Mineral Research and Exploration ,Üniversiteler Mahallesi Dumlupınar Bulvarı No:139 06800- Çankaya/ANKARA
[email protected] Abstract Various organic fertilizer production processes related with alkali and oxidation of coals, with HNO 3 has been developed. In spite of this, these products are controversial for agriculture due to their high pH content of alkali oxidation products, high temperature and pressure comprising control difficulties. It is known that production with HNO 3 is difficult and has not found any application in time. Except pH, due to similar reasons and chemical structures of the produced products are not clearly known, therefore, nitro humates have not found any application areas. In this study, partial molecular disintegration of high humus, with H 2SO4 oxidation, comprising of Ilgın, Göynük and Elbistan lignite were studied, the neutralization reaction was carried out with ammonia by increasing acid ratios and a new type of organic fertilizer production with a total chemical work base and a new molecular structure was produced. While the used material was chosen randomly from the material’s beds and Ilgın contains of 42.02%, Göynük of 33.19 % and Elbistan of 53.00 % humic compounds. The samples cleaned with “acid leaching” in case of metal pollution and stoichiometric amounts of H2SO4, H3PO4 and Mardin phosphate were added through these compounds to substantiate of required amount for theoretical N-P2O5-K2O. Moreover, rested oxidized coal was neutralized with liquid ammonia and KOH. As a final product, fertilizer was produced in theoretical form of (8-6-1) + S. Humic acid ratio in produced fertilizers varied between 33.43% and 37.40 %; nitrogen amount, between 7.35 % and 8.46; P 2O5 amount, between 5.84 % and 6.46 %; K2O ratio, between 0.96 % and 1.12 %. Organic material changes between 70.92 % and 74.58 %. This is a cheap fertilizer which may be evaluated in the framework of the National Organic Fertilizer Regulation. Cheap raw material resource is a type of fertilizer which fixation of direct industrial chemicals’ usage and “industrial humification”, being used as an alternative of long time taking natural humidification as well as theoretical N, P 2O5, K2O like macro and micro nutrition elements, can be adjusted as desired ratios. Keywords: Humic acid fertilizers, Organomineral fertilizer, Organic sourced fertilizers
Modeling, Scaleup and Optimization of Slurry Bubble Column Reactors for Fischer-Tropsch Synthesis Laurent Sehabiague and Badie I. Morsi Chemical and Petroleum Engineering Department, University of Pittsburgh, 1249 Benedum Hall, 3700 O'Hara Street, Pittsburgh, PA 15261
1. ABSTRACT The gas holdup and mass transfer coefficients for syngas in actual Fischer-Tropsch wax were measured in a pilot-scale (0.3 m ID and 3-m height) Slurry Bubble Column Reactor (SBCR) operating under conditions typifying those of the actual Fischer-Tropsch (F-T) synthesis (T up to 530K, P up to 30 bar, UG up to 0.3 m/s, CS up to 20 vol.% ). Two new correlations were developed to predict the gas hold up and mass transfer coefficients experimental data with an Absolute Average Relative Errors (AARE) of 15.7% and 17.6%, respectively. These new correlations along with four different kinetic expressions, taken from the literature, were used in our simulator to predict the performance of a conceptual commercial-scale (7m ID, 30-m height) SBCR operating with iron catalyst. The simulator was used to study the effects of various operating variables (H2/CO ratio, CS, and UG) on the syngas conversion, the space time yield (STY) and catalyst productivity; and to delineate the “optimal” operating conditions for this a conceptual reactor. The simulation results led to the following conditions which could be considered “optimal” for operating this conceptual SBCR: H2/CO ratio =1, superficial gas velocity = 0.3 m/s and iron catalyst concentration = 20 vol%. Thus, this conceptual SBCR operating with iron catalyst could be integrated with a Coal-To-Liquid (CTL) process. It should be remembered, however, that the severe problems associated with iron catalyst attrition and separation from viscous F-T products could hamper the performance of such a conceptual SBCR and therefore a cobalt-supported catalyst could be used instead.
2. INTRODUCTION The Fischer-Tropsch (F-T) synthesis (originally known as “Synthol”) was developed in the 1920’s in Germany at the Kaiser Wilhelm Institute by Franz Fischer and Hans Tropsch, with the intent of producing synthetic hydrocarbons [1, 2]. Their work was based on the 1902 discovery of Sabatier and Senderens [3] that methane can be produced from H2 and CO in the presence of nickel catalyst. In the Fischer-Tropsch process, the syngas (H2+CO) reacts in presence of a solid catalyst to produce a wide spectrum of hydrocarbon, such as paraffins, olefins, and oxygenated chemicals (alcohols, aldehydes, acids, ketones, etc…). The F-T synthesis is a combination of oligomerization reactions [4], which can be summarized as follows: •
n-Paraffins synthesis:
+ 2 +1 •
+
(1)
1-Olefins synthesis:
+2 •
→
→
(2)
+
Alcohols synthesis:
+2
→
+
−1
(3)
A simplified expression of the above F-T overall reactions can be written as: + 1+
1 2
+
(4)
It has been reported that the main products of the F-T synthesis are linear paraffins with a typical H2/CO usage ratio ranging from 2.06 to 2.16 [4, 5]. The syngas is conventionally produced through steam reforming of natural gas (methane) or gasification of coal and/or biomass. Considering the huge quantities of coal and natural gas available worldwide (Figure 1), the F-T synthesis appears to be a promising technology which could help alleviating the dependence on oil imports and securing more environmentally friendly and sustainable alternative energy sources. The F-T process could play an important role in national security not only in the US, but also in developing countries, such as China and India.
The F-T reactiions has b been carrie ed out for o over 50 yea ars mainlyy in fixed-bed reactorrs (FBRs); and Sasoll (Sou uth Africa) pioneered d low-tempe erature F-T T synthesiss in slurry bubble column reacttors (SBCR Rs) in earlyy 1993 3. In these e reactors, the solid-p phase conssists of miccron-size ccatalytic pa articles whiich are susspended in n the lliquid-phasse (F-T pro oducts) by the synga as usually ssparged fro om the botttom of the e reactor th hrough the e slurrry-phase (ssolid catalyyst+ F-T liq quid products). The a advantagess of SBCR Rs over FBR Rs are [4, 6 6-8]: • • • • • •
Betterr temperatu ure control//removal Lowerr capital co ost (~25% o of that of a multi-tubu ular reactor) due to th heir relative ely simple design Low pressure drrop (4 times less than n in fixed b bed reactorr) Ability of using fine catallyst particlles (<100 μ) allowin ng huge ssurface are ea and be etter masss er transfe Higher yield per reactor vo olume nd removed continuo ously, allow wing longe er runs since no shu utdown are e Catalyyst can be added an necessary.
The use of SB BCRs in F F-T synthe esis, howevver, is ham mpered byy their inhe erent comp plex hydro odynamics,, nty mass transfer da ata and diffficulties asssociated w with catalysst separatiion from th he viscouss F-T liquid d scan prod ducts. The present sstudy aims, on one h hand, at de eveloping novel corrrelations fo or predictin ng the hydrodynamicc nsfer param meters under typical F-T operrating cond ditions; and d, on the o other hand d, at using g and mass tran along with different F F-T reactio on kineticss available in the lite erature to p predict the e thesse new corrrelations a effeccts of supe erficial gass (UG), syn ngas comp position (H H2/CO ratio o) and cata alyst conce entration (C CS) on the e perfo ormance of SBCRs. The simulator d developed at the University of Pittsb burgh, bassed on a com mprehensive er model, w was used to o carry outt these predictions [9 9]. e compute
Figure 1: Order of Magnitude of Energ gy Resourrces (EJ = 1018 J) [10, 11]
3. NOVE EL HYDRO ODYNAMIC CS AND MASS TRAN NSFER CO ORRELATIIONS
o novel correlations to estimatte the hydrodynamicc and mass transfer parameters in the F F-T SBCR R Two were e develop ped using the expe erimental data obta ained in tthe pilot-scale SBC CR availab ble in ourr labo oratories a at the Univversity of Pittsburgh h. The ran nges of th he operatin ng conditio ons emplo oyed were e temp perature ((373-530K K), pressurre (7-30 bar), superficial ga as velocityy (0.1-0.3 m/s) and catalystt conccentration (0-20 vol.% %), which cover the typical tho ose of a tyypical F-T conditionss [12]. An actual F-T T waxx and deacctivated iro on catalystt as well a as alumina, frequentlly utilized as supporrt for coba alt catalyst,, were e also used d in the exxperiments. The empirical correlations for gass holdup (εG) develo oped by Behkish et al [13] an nd that for the masss ed by Lemo oine et al. [14] are ccompared with our e experimental data in n transsfer coefficcient (kLa)) develope Figu ures 2 and 3, and the e Absolute e Average Relative E Errors (AARE) using the two co orrelation are 21.8% % and 23.6%, resspectively..
Figure e 2: Compa arison Bettween Measured an nd Predictted εg Valu ues Usin ng Behkish et al [13 3]
Figure 3: Compa arison Bettween Mea asured and Predicte ed kla Valu ues U Using Lem moine et all. [14] Corrrelation his study, tthe effect o of the solid d concentrration as w well as the foaming ttendency o of the liquid d-phase iss In th emp phasized. T The latter was accou unted for b by introduccing the te erm “X” wh hich equalss either 0 for single-com mponent an nd non-foa aming liqu uid mixture es or 1 fo or foaming g liquid mixtures. Th he new co orrelationss deve eloped are e:
.
.
= 0.0068
.
.
× exp − −2.5
.
.
.
+1
− − 10 0
− 0.157 7
Γ + 0.323
.
(5)
.
= 948.5762 ×
. .
.
−7
.
− 2464.7
.
.
.
.
−
+ 0.3
.
(6)
+ 0.291
Figures 4 and 5 show a comparison between the experimental data and predicted values data; and the AARE corresponding to these new correlations are 15.7% and 17.6%, respectively. This means that the new correlations presented in this study provide an improvement over those by Behkish et al. [13] and Lemoine et al. [14] for predicting the gas holdup and mast transfer, respectively.
Predicted
1
0.1 0.1
1 Experimental
Figure 4: Comparison Between Measured and Predicted εg Values Using Equation (5)
Predicted
3
0.3
0.03 0.03
0.3
3
Experimental
Figure 5: Comparison Between Measured and Predicted kla Values Using Equation (6)
4. CATALYSTS AND KINETICS Two types of catalysts are commonly used in F-T synthesis, namely iron or iron-based and cobalt-based catalysts. The comparison between the two catalyst reveals the following: iron catalyst is cheaper; possesses higher Water-Gas-Shift (WGS) activity; borne to strong attrition; the F-T water produced decreases its activity [4, 15]; and is well suited for syngas produced through gasification units with low H2/CO ratio. The cobalt-based catalyst, on the other hand, is more expensive [4]; has negligible WGS activity; is much more resistant to attrition; and is well suited for syngas produced from natural gas reforming units with high H2/CO ratio. In this study, a conceptual F-T SBCR, to be integrated in a Coal-ToLiquid (CTL) process, is simulated; and therefore only iron catalyst is considered. The iron catalyst activity was reported to increase with H2 partial pressure and decrease with H2O partial pressure [16, 17], suggesting a strong competition between CO and H2O for the adsorption on the catalyst active sites. Indeed, iron catalyst has high WGS activity and therefore the rate of the WGS reaction and that of the F-T synthesis should be taken into account for simulation purposes. In the WGS reaction, the H2O produced by the F-T reaction reacts with CO in the syngas to form H2 and CO2: +
(ΔH0 = -41.2 kJ/mol at 298 K [18]).
+
(7)
In this study, the four kinetic rate expressions for the F-T reactions, listed in Table 1, were taken from the literature and each was used independently in the reactor model along with the WGS reaction given in Table 2. The constants in these rate expressions are expressed as function of the operating temperature as given in Tables 3 and 4. It is important to note that these F-T kinetic rate expressions rates were here selected due to the fact that they were obtained in slurry reactors and are simple enough to be directly integrated in our simulator.
Table 1: Kinetics Studies for the F-T Synthesis on Iron Catalyst Operating Conditions Reference
Catalyst
Reactor
Equation T,C
P, MPa
H2/CO feed
Huff and Satterfield [19]
Reduced Fused Fe/K2O/CaO/SiO2
Slurry
232-263
0.4-1.5
0.5-1.8
Ledakowicz et al. [20]
Reduced Prec. Fe/K
Slurry
220-260
1.0
0.5-0.6
Deckwer et al.[21]
Reduced Prec. Fe/K
Slurry
220-260
-
0.5-2.0
Zimmerman and Bukur [8]
Prec. Fe/Cu/K & Reduced Fe/Cu/K/SiO2
Slurry
235-265
1.5-3.0
0.6-1.0
rFT = k FT
PH22 PCO PCO PH 2 + aPH 2O
rFT = k FT
rFT = k FT
PH 2 PCO PCO + aPCO 2
PH22 PCO PCO PH 2 + aPH 2O
rFT = k FT
PH 2 PCO PCO + aPH 2 O
Table 2: Kinetics Studies for the WGS Reaction on Iron Catalyst Operating Conditions Reference
Catalyst
Reactor T,C
Zimmerman and Bukur [8]
Prec. Fe/Cu/K & Reduced Fe/Cu/K/SiO2
Slurry
235265
P,
H2/CO
MPa
feed
1.5-3.0
0.6-1.0
Equation
rWGS = kWGS
⎛ P P ⎜ PH O PCO − CO2 H 2 ⎜ 2 K eq ⎝ PCO + aPH 2O
⎞ ⎟ ⎟ ⎠
Table 3: F-T Kinetics Constants Values Used in the Simulation Reference
Constant
Units ,
= 3.714
mol/kgCatalyst/Pa/s
Huff and Satterfield [19]
,
= 1.786. 10 ,
= 1,593 Ledakowicz et al. [20] = 0.0241
.
= 0.0185
Zimmerman and Bukur [8]
= 6.763
mol/kgCatalyst/Pa/s .
-
,
= 13.142 Deckwer et al.[21]
Pa
mol/kgCatalyst/Pa/s
,
Pa ,
mol/kgCatalyst/Pa/s
= 5.3
-
Table 4: WGS Kinetics Constants Values Used in the Simulation Reference
Constant
Units ,
= 9.2538. 10 Zimmerman and Bukur [8]
= 21 = 0.01317
mol/kgCatalyst/Pa2/s -
,
-
5. REACTOR MODEL
In this study, a conceptual commercial-scale SBCR for F-T synthesis ( 7-m ID and 30-m height) was simulated using the axial dispersion model in conjunction with the two-class gas bubbles model [22-26] since visual observations and photographic methods pointed out to the coexistence of distinctly two-class gas bubbles (small and large) in BCRs operating in the churn-turbulent flow regime [25, 27-30]. The reactor model was built with the following features: • • • • •
The SBCR is operated in the churn-turbulent flow regime The gas is sparged at the bottom of the reactor trough a multiple orifices gas distributor. The catalyst used is iron catalyst. The slurry is moving upward. The total heat of reaction is removed using saturated water flowing in a bundle of cooling tubes.
The reactor model was also based on the following key assumptions: • • • • •
The mass transfer resistance in the bulk gas-phase and liquid-solid interface are negligible when compared with that in the liquid film. The slurry superficial velocity is constant. The slurry temperature is constant The catalyst suspension behavior follows the sedimentation-dispersion model [31-35]. The reactor is operating in a steady-state.
The reactor model details can be found elsewhere [9]. All the differential material and energy balance equations for all components along with the equation parameters and boundary conditions were incorporated in our Windows compatible simulator with graphical user interface (see Figure 6). The set of equations are numerically solved using the finite elements method with the free FORTRAN compiler g95.
Figure 6 6: Graphical User Interface off the Simulator
6. SIMUL LATION RESULTS
d to studyy the effeccts of different opera ating variables, inclu uding supe erficial gass The simulatorr was used ocity, H2/CO O inlet rattio and catalyst conccentration on the pe erformance e of a SBCR operatting in the e velo churrn-turbulen nt flow reg gime with iiron catalyyst. The ra anges of a all operatin ng variable es in addittion to the e sparrger type/geometry, catalyst particle ssize/densityy, slurry superficial velocity, and cooling pipess num mber/size used in the simulator are given iin Table 5.
Table 5: Operatting Variab bles Used d in the Sim mulator fo or F-T SBC CR System
Reacctor and Sp parger Geometryy
Cooling pipe es
Operrating Varia ables
Cata alyst H2/C CO L, m dR , m NO , dorf., m Sparge er Type Nu umber of co ooling pipe es Cooling g pipes outtside diame eter, m T,K P,M MPa UG , m/s USL , m/s dP , µm ρP , kg/m3 CS , V Vol.%
Iron 0 0.01-3 30 7 103 33 - 1367 0.06 65 - 0.075 M M-ON 600 0 0.112 493 3 0 0.1-0.5 0.01 32 3000 0-50
nce of the reactor w was evaluated in term ms of the ssyngas conversion, tthe space time yield d The performan Y) and cattalyst produ uctivity. Th he STY is d defined acccording to o de Swart [22] as the mass (kg g) of liquid d (STY hydrrocarbons produced per reacctor volume e (m3) pe er unit time e (hour). The catalyyst producctivity wass defin ned as the e mass (kg g) of liquid hydrocarb bons (C5+) produced per masss (kg) of ca atalyst in tthe reactorr per unit time (h hour).
6.1 Effect of inlet H2/CO ratio Figure 7 shows the effect of the inlet H2/CO ratio on syngas conversion; and as can be seen in the figure, there is a maximum at H2/CO ratio at about 0.65 when using the kinetic rate expressions by Deckwer et al. [21], Huff and Satterfield [19] and Zimmerman and Bukur [8]. This maximum, however, occurs at about H2/CO ratio of 1.3 when using the kinetic rate expression by Ledakowicz et al. [20]. This difference in behavior can be related to the absence of the effect of H2 and H2O partial pressures on F-T reaction kinetic expression by Ledakowicz et al. [20], which was replaced by a term representing the effect of CO2 partial pressure as it is related to the H2O presence through the WGS reaction. In this study, regardless of the kinetic rate expression used, the simulation results indicate that iron catalyst could be used in a SBCR which can be fully integrated in a CTL process, where a H2/CO ratio of less than 1 would achieve a syngas conversion between 73% and 81%. It is important to note that decreasing the H2/CO ratio below 0.5 will lead to a sharp decrease of the syngas conversion in all cases, while increasing the H2/CO ratio above 1 will lead to a much slower decrease in the syngas conversion. It appears therefore more appropriate to operate the SBCR with a H2/CO ratio slightly above the 0.65 not below such a value. That is why throughout the following simulations, a H2/CO ratio equals 1 was used along with the four 4 different kinetic rate expressions.
Deckwer et al.
100%
Huff and Satterfield
90%
Ledakowicz et al.
Syngas conversion , %
80%
Zimmerman and Bukur
70% 60% 50% 40% 30% 20% 10% 0% 0
0.5
1
1.5
2
2.5
3
H2/CO ratio, -
Figure 7: Effect of Inlet H2/CO Ratio on Syngas Conversion (CS = 20 vol% and UG = 0.2 m/s)
6.2 Effect of catalyst concentration The effect of catalyst concentration on the syngas conversion at a constant inlet H2/CO ratio of 1 in the SBCR is presented in Figure 8; and as can be seen the reactor performance greatly increases with the addition of catalyst, then levels off quickly, and reaches a maximum at a catalyst concentration of about 20 - 25 vol.%. Also, a further increase of the solid concentration above 30 vol.% leads to a dramatic decrease of the syngas conversion. In this region, the high catalyst concentrations decrease the mass transfer coefficients, which in turn drive the SBCR to be mass transfer-controlled. It is important to note that this appears to be true for the four kinetic rate expressions used, suggesting that the transition takes place independently of the reaction kinetics. The increase of catalyst concentration in the SBCR, however, appears to decrease the catalyst productivity, expressed in kg of hydrocarbon produced per kg of catalyst per unit time as can be seen in Figure 9. This decrease of catalyst productivity is more significant when using the kinetics rate expressions by Deckwer et al. [21] and Huff and Satterfield [19]. When increasing the catalyst concentration above 20 vol%, however, the decrease of the catalyst productivity is reduced and even becomes constant. It should also be remembered that in this mass transfer-controlled region, the catalyst productivity is independent of the kinetics rate expressions used.
Deckwer et al.
100%
Huff and Satterfield
90%
Ledakowicz et al. 80%
Zimmerman and Bukur
Syngas conversion , %
70% 60% 50% 40% 30% 20% 10% 0% 0%
10%
20%
30%
40%
50%
Cs , vol.%
Figure 8: Effect of Catalyst Concentration on Syngas Conversion (H2/CO ratio = 1, and UG = 0.2 m/s)
Deckwer et al.
0.6
Catalyst Productivity , kgHC/kgcatalyst/hr
Huff and Satterfield Ledakowicz et al.
0.5
Zimmerman and Bukur 0.4
0.3
0.2
0.1
0 0%
10%
20%
30%
40%
50%
Cs , vol.%
Figure 9: Effect of Catalyst Concentration on Catalyst Productivity (H2/CO ratio = 1, UG= 0.2 m/s) Since the mass transfer appears to influence the performance of the SBCR under the conditions studied, the relevance of the mass transfer resistance was quantified using the following expression [36, 37]:
β =
1
R
R
=
1
+
1
(8)
Where ki is the pseudo-kinetics constant, which can be expressed as: =
,
(9)
The β ratios for CO and H2 are plotted as a function of the catalyst concentration in Figures 10 and 11, respectively, and as can be observed the resistance due to mass transfer becomes significant (β >50%) in the case of CO for solid concentration >26 vol.% while in the case of H2 the resistance to mass transfer
does not become significant over the range of operating conditions used, except when using Ledakowicz et al. [20] rate expression at a catalyst concentration of 35 vol.%. Thus, CO appears to be the limitingreactant and could explain why in Figure 8 at a solid concentration of about 26 vol.% the syngas conversion reaches a maximum value, indicating that the reactor behavior is changing from a kineticcontrolled to a mass transfer-controlled regime. During the kinetics-controlled regime, the reactor performance greatly increases with the addition of catalyst, then it level off quickly; whereas in the mass transfer-controlled region, the addition of solids leads to a dramatic decrease of the reactor performance.
Deckwer et al.
100%
Huff and Satterfield
mass transfer controlled regime
90%
Ledakowicz et al.
80%
Zimmerman and Bukur
70%
β CO , %
60% 50% 40% 30%
kinetics controlled regime
20% 10% 0% 0%
10%
20%
30%
40%
50%
Cs , vol.%
Figure 10: Effect of Catalyst Concentration on Mass Transfer Resistance of CO (H2/CO ratio = 1, and UG = 0.2 m/s)
Deckwer et al.
100%
Huff and Satterfield
mass transfer controlled regime
90%
Ledakowicz et al.
80%
Zimmerman and Bukur
70%
β H2 , %
60% 50% 40% kinetics controlled regime
30% 20% 10% 0% 0%
10%
20%
30%
40%
50%
Cs , vol.%
Figure 11: Effect of Catalyst Concentration on Mass Transfer Resistance of H2 (H2/CO ratio = 1, and UG = 0.2 m/s) Figure 8 also shows that increasing the solid concentration from about 20 to 26 vol%, does not lead to significant increase in the syngas conversion for the four kinetic rate expressions used. Considering that the highest syngas conversion occurred at catalyst concentration of about 20 vol%, and a H2/CO ratio of 1, these values were used in the following simulations.
6.3 Effect of superficiial gas veelocity The effect of tthe superficial gas vvelocity on the synga as converssion and STY is pressented in F Figures 12 2 bserved, in ncreasing tthe gas ve elocity was found to d decrease tthe syngass and 13, respecctively. As can be ob nd increase e the hydrocarbons yield, exp pressed in n kg of hyd drocarbons produce ed per unitt convversion an reacctor volume per unitt time. Un nder the operating cconditions used in th he simulattion, the lo owest gass velo ocities allow w the highe est syngass conversio on (see Figure 9) du ue primarilyy to the lon ng residen nce time off the gaseous re eactants in n the reacttor; whereas the highest gas vvelocities a allow the g greatest hyydrocarbon n d (see Figure 10). Itt can also be observved that th he hydroca arbon yield d seems tto slowly level off att yield supe erficial gass velocitiess greater th han 0.3 m//s. Increassing the su uperficial ga as velocityy from 0.1 m/s to 0.3 3 m/s appears to o more tha an double the STY w while furthe er increase es from 0.3 m/s to 0 0.5 m/s only leads to o ut 30% inccrease in th he STY forr all the kin netics rate e expression n used. Thus, considering the lo ow syngass abou convversion and high ope erating cosst associate ed with run nning the S SBCR at high superfficial gas ve elocities, itt coulld be concluded that an optima al superficial gas velo ocity would d be aboutt 0.3 m/s fo or such a cconceptuall mmercial-sccale SBCR R operating with iron ccatalyst. com
Deckw wer et al.
100%
Huff and Satterfieeld
90%
Ledako owicz et al. 80%
Zimmeerman and Bukur
Syngas Conversion , %
70% 60% 50% 40% 30% 20% 10% 0% 0.1
0.2
0.3
0.4
0.5
UG, m/s
Figure 12: Effectt of Superfficial Gas Velocity o on Syngas s Convers sion (Catalys st concenttration = 2 20 vol%, H2/CO ratio o = 1)
Figure 13 3: Effect o of Superfic cial Gas V Velocity on n STY (Catalys st concenttration = 2 20 vol%, H2/CO ratio o = 1)
7. CONCLUDING REMARKS
The simulations results of a conceptual commercial-scale F-T SBCR (7-m ID, 30-m height) operating with iron catalyst considering mass transfer and using different kinetic rate expressions, taken from the literature, led to the following remarks: • • • •
•
•
At a superficial gas velocity of 0.2 m/s and a catalyst concentration of 20 vol%, the maximum syngas conversion occurred when using inlet H2/CO ratio of about 0.65. At H2/CO ratio of 1, CO is the limiting reactant and the regime in which the SBCR is operating is therefore dependent on the β ratio of CO. At a H2/CO ratio of 1 and a superficial gas velocity of 0.2 m/s, the maximum syngas yield occurred at a catalyst concentration of about 26 vol%. At a H2/CO ratio of 1 and a catalyst concentration of 20 vol%, no optimal superficial gas velocity could be clearly found as increasing the gas velocity led to the decrease of the syngas conversion and to the increase of the space time yield. Under the operating conditions studied, a superficial gas velocity about 0.3 m/s appears to be a good compromise. These above findings indicate that the following conditions which could be considered “optimal” for operating the conceptual SBCR: H2/CO ratio =1, superficial gas velocity = 0.3 m/s and iron catalyst concentration = 20 vol%. Thus, this conceptual SBCR operating with iron catalyst could be fully integrated with a CTL process. It should be remembered, however, that considering the sever attrition problems and the difficulties associated with separating the micron-size catalyst from the viscous F-T products inherent to the iron catalyst, a cobalt-supported catalyst could be used instead for such a conceptual reactor size.
8. NOMENCLATURE a
coefficient in kinetic rate expressions, dimensionless or Pa
aL
Gas-liquid interfacial area per unit liquid volume, m-1
Ci,L
Component i concentration in the liquid phase, mol/m3
CS
Solid concentration in the slurry phase by volume, vol %
dR
Reactor diameter, m
dorf.
Diameter of the orifices of the gas distributor, m
dP
Diameter of the solid particles, m
kLaL
Volumetric liquid-side mass transfer coefficient (based on liquid volume), s-1
L
Reactor length, m
MW
Molecular weight, kg.kmol-1
NO
Number of orifices of the gas distributor, -
P
Pressure, Pa
PS
Saturation vapor pressure, Pa
r
Reaction rate, mol.kg-1 catalyst.s-1
U
Superficial velocity, m s-1
X
foaming coefficient, -
xS
Solid concentration in the slurry phase, kg/m3
Greek symbols ε
holdup, -
Γ
Gas sparger coefficient, -
μ
Viscosity, kg m-1 s-1
ρ
Density, kg m-3
σ
Surface tension, Nm-1
Acronyms R
Ring
SBCR Slurry Bubble Column Reactor
Subscripts FT
Refer to the Fischer-Tropsch reaction
G
Refer to the gas phase
L
Refer to the liquid phase
P
Refer to the solid particles
S
Refer to the solid particles suspension
SL
Refer to the slurry phase
9. REFERENCES
1. 2. 3. 4. 5. 6. 7.
8. 9.
10. 11. 12. 13.
14.
15. 16. 17. 18.
19.
20.
21.
22. 23.
24.
25.
Dry, M.E., Commercial conversion of carbon monoxide to fuels and chemicals. Journal of Organometallic Chemistry, 1989. 372(1): p. 117-127. Stranges, A.N. Germany’s Synthetic Fuel Industry 1927-45. in AIChE 2003 Spring National Meeting. 2003. New Orleans, LA. Sabatier, P. and J.D. Senderens, Nouvelles Syntheses du Methane. Comptes Rendus, 1902. 134: p. 514. Dry, M.E., The Fischer-Tropsch process: 1950-2000. Catalysis Today, 2002. 71(3-4): p. 227-241. Steynberg, A. and M. Dry, Fischer-Tropsch Technology. Studies in Surface Science and Catalysis, ed. G. Centi. Vol. 152. 2004: Elsevier Science. Nigam, K.D.P. and A. Schumpe, Three-Phase Sparged Reactors. 1996, Amsterdam, The Netherlands: Gordon and Breach Publishers. Satterfield, C.N. and G.A. Huff, Product Distribution from Iron Catalyst in Fischer-Tropsch Slurry Reactors. Industrial & Engineering Chemistry Process Design and Development, 1982. 21(3): p. 465-470. Zimmerman, W.H. and D.B. Bukur, Reaction kinetics over iron catalysts used for the FischerTropsch synthesis. Canadian Journal of Chemical Engineering, 1990. 68(2): p. 292-301. Sehabiague, L., et al., Modeling and optimization of a large-scale slurry bubble column reactor for producing 10,000 bbl/day of Fischer-Tropsch liquid hydrocarbons. Journal of the Chinese Institute of Chemical Engineers, 2008. 39(2): p. 169-179. WEC, 2004 SURVEY OF ENERGY RESOURCES. 2004, Amsterdam, The Netherlands: Elsevier. WEC, 2007 SURVEY OF ENERGY RESOURCES. 2007, London, United Kingdom: World Energy Council. Sehabiague, L. and B.I. Morsi. Scaleup of SBCRs for Fischer-Tropsch Synthesis. in AIChE Annual Meeting. 2009. Nashville, TN. Behkish, A., et al., Novel Correlations for Gas Holdup in Large-Scale Slurry Bubble Column Reactors Operating under Elevated Pressures and Temperatures. Chemical Engineering Journal, 2006. 115(3): p. 157-171. Lemoine, R., et al., An Algorithm for Predicting the Hydrodynamic and Mass Transfer Parameters in Bubble Column and Slurry Bubble Column Reactors. Fuel Processing Technology, 2008. 89(4): p. 322-343. Schultz, H., Short history and present trends of Fischer-Tropsch synthesis. Applied Catalysis A: General, 1999. 186: p. 3-12. van der Laan, G.P., Kinetics, Selectivity and Scale Up of the Fischer-Tropsch Synthesis. 1999, University of Groningen: Groningen. Satterfield, C.N., et al., Effect of water on the iron-catalyzed Fischer-Tropsch synthesis. Industrial & Engineering Chemistry Product Research and Development, 1986. 25(3): p. 407-414. Callaghan, C.A., Kinetics and Catalysis of the Water-Gas-Shift Reaction: A Microkinetic and Graph Theoretic Approach, in Department of Chemical Engineering. 2006, Worcester Polytechnic Institute. p. 400. Huff, G.A. and C.N. Satterfield, Intrinsic kinetics of the Fischer-Tropsch synthesis on a reduced fused-magnetite catalyst. Industrial & Engineering Chemistry Process Design and Development, 1984. 23(4): p. 696-705. Ledakowicz, S., et al., Kinetics of the Fischer-Tropsch Synthesis in the Slurry Phase on a Potassium-Promoted Iron Catalyst. Industrial & Engineering Chemistry Process Design and Development, 1985. 24(4): p. 1043-1049. Deckwer, W.D., et al., Kinetic studies of Fischer-Tropsch synthesis on suspended iron/potassium catalyst - rate inhibition by carbon dioxide and water. Industrial & Engineering Chemistry Process Design and Development, 1986. 25(3): p. 643-649. de Swart, J.W.A., Scale-Up of a Fischer-Tropsch Slurry Reactor. 1996, University of Amsterdam: Amsterdam, Netherlands. de Swart, J.W.A. and R. Krishna, Simulation of the transient and steady state behavior of a bubble column slurry reactor for Fisher-Tropsch synthesis. Chemical Engineering and Processing, 2002. 41(1): p. 35-47. de Swart, J.W.A., R.E. van Vliet, and R. Krishna, Size, Structure and Dynamics of "Large" Bubbles in a Two-Dimensional Slurry Bubble Column. Chemical Engineering Science, 1996. 51(20): p. 4619-4629. Lemoine, R., A. Behkish, and B.I. Morsi, Hydrodynamic and Mass Transfer Characteristics in Organic Liquid Mixtures in a Large-Scale Bubble Column Reactor for the Toluene Oxidation Process. Industrial & Engineering Chemistry Process Design and Development, 2004. 43(19): p. 6195-6212.
26. 27.
28. 29.
30.
31.
32. 33. 34. 35. 36. 37.
Rados, N., M.H. Al-Dahhan, and M.P. Dudukovic, Modeling of the Fischer–Tropsch synthesis in slurry bubble column reactors. Catalysis Today, 2003. 79-80: p. 211-218. Behkish, A., et al., Gas holdup and bubble size behavior in a large-scale slurry bubble column reactor operating with an organic liquid under elevated pressures and temperatures. Chemical Engineering Journal, 2007. 128(2-3): p. 69-84. Grund, G., A. Schumpe, and W.D. Deckwer, Gas-Liquid mass transfer in a bubble column with organic liquids. Chemical Engineering Science, 1992. 47(13-14): p. 3509-3516. Krishna, R., M.I. Urseanu, and A.J. Dreher, Gas holdup in bubble columns: influence of alcohol addition versus operation at elevated pressures. Chemical Engineering and Processing, 2000. 39(4): p. 371-378. Vermeer, D. and R. Krishna, Hydrodynamics and mass transfer in bubble columns in operating in the churn-turbulent regime. Industrial & Engineering Chemistry Process Design and Development, 1981. 20(3): p. 475-482. Behkish, A., Hydrodynamic and Mass Transfer Parameters in Large-Scale Slurry Bubble Column Reactors, in Chemical and Petroleum Engineering Department. 2004, University of Pittsburgh: Pittsburgh, PA. p. 331. Kato, Y., et al., The behavior of suspended solid particles and liquid in bubble columns. Journal of Chemical Engineering of Japan, 1972. 5(2): p. 112-118. Kojima, H., H. Anjyo, and Y. Mochizuki, Axial mixing in bubble column with suspended solid particles. Journal of Chemical Engineering of Japan, 1986. 19(3): p. 232-234. O’Dowd, W., et al., Gas and solids behavior in a baffled and unbaffled slurry bubble column. American Institute of Chemical Engineering Journal, 1987. 33(12): p. 1959-1970. Smith, D.N. and J.A. Ruether, Dispersed solid dynamics in a slurry bubble column. Chemical Engineering Science, 1985. 40(5): p. 741-754. Deckwer, W.-D., et al., Chemical Engineering Science, 1981. 36: p. 765. Deckwer, W.D., et al., Chemical Engineering Science, 1981. 36: p. 765.
BIOGASIFICATION OF SOMA LIGNITE (A Preliminary Study) Mustafa Baysala, Sedat İnanb, Yuda Yürüma a
Faculty of Engineering and Natural Sciences, Sabanci University, Orhanli, Tuzla, Istanbul 34956, Turkey TÜBİTAK Marmara Research Centre, Earth and Marine Sciences Institute, Gebze-Kocaeli, Turkey Corresponding author: e-mail: [email protected] b
ABSTRACT In this project, the bacterial gasification on the coal samples which were evacuated from Soma basin in Turkey and gas adsorption mechanism of these samples were analyzed. It is known that coal can be solubilized chemically (alkaline solutions) and biologically by using wood-rotting fungi species. Chemical solubilization of coal samples was investigated. For this purpose, coal samples were solubilized in the different Lewis base solutions. For biogasification process, solubilization at moderate pH (9≥ pH ≥5) level is an important factor for the conserve bioactivity of the microorganisms. We found that carbonate and oxalate systems can be solubilized coal at moderate pH and also these Lewis bases was used in biogasification process to solubilized coal samples and increase gasification efficiency. To understand gas adsorption on the coal surface, high pressure gas adsorption experiments were conducted. Keywords:
Coal, biogasification, solubilization, adsorption
1. INTRODUCTION Within the increasing interest of the renewable energy source and clean energy, coal utilization has become a more important subject in today’s world. Natural gas is a relatively clean energy with respect to the combustion of the coal. We know that coal bed methane is a considerable large energy source for most of the countries. For example in America, coal bed methane covers more than %10 of the natural gas demand [1]. Coal bed methane (CBM) can be arise from both thermogenic and biogenic activity on the coal beds and can be adsorb on the porous structure of the coal. Biogenic methane gas, due to the bacterial activity of the coal beds, can be called primary biogenic methane, and more than %20 of the coal bed methane is originated from these activity. Coal can be utilized by using biological processes. Last two decades, lots of coal utilization methods have been developed. After the first report published by Cohen & Gabrielle in 1982 [2], lots of biological methods have been developed by scientists like microbial coal liquefaction, methane production. Methane can be produced by using laboratory incubation experiment, called secondary biogenic methane. Methane production due to the biological treatment is a relatively clean process compared to the coal combustion. Biological treatment of the coal takes place under mild conditions at low temperature and pressure unlike the classic processes. Production rate changes from the rank of the coal, usually low rank coals can be converted to the methane more easily under the proper environmental conditions [3]. Although the mechanism of the process have not been fully discovered, the pathways of the microbial treatment of the coal are common in most of the studies. Isbiste and Baric [4] reported two main approaches for the synthesis of the methane from coal. First one is the direct gasification of the coal. In this process, solid coal samples are treated with the methanogenic bacteria to produce methane from coal without any solubilization/depolymerization step. In the second approach, which is called indirect gasification, first, lignite is depolymerized/solubilized biologically or chemically (using aqueous alkali). By this way, complex polymer structure of the coal is broken down to smaller organic molecules. Then, acetogenic
bacteria ferment these smaller organic molecules (fatty acids, alcohols and aromatic acids) to acetate, hydrogen and carbon dioxide. After that, methanogens used to convert acetate, hydrogen and carbon dioxide to methane. The goal of this study is the investigation of the biogenic natural gas potential and microbial gasification process of Soma lignite in Turkey by laboratory experiments. We tried to improve the efficiency of gasification process and investigate biogasification suitability of Soma lignite. Also, gas potential of lignite was explored by using surface adsorption and desorption isotherms.
2. EXPERIMENT AND RESULTS In this study, solubilization of the coal samples was investigated as a beginning. Studies show that, coal can be solubilized biologically or chemically. In biological methods, wood-rotting fungi species are able to solubilize/depolymerize low rank coals by secreting oxalate ions. Bumpus et al. shows that chemically, coal macromolecules are solubilized in an alkaline medium, like aqueous solution of oxalate ions, instead of using any fungi species [5]. In the light of this point, we used different Lewis bases to reach maximum solubilization capacity for our coal samples. Effect of pH on the ability of Lewis bases to solubilize coal was investigated. Also, characterization of the coal samples which were excavated in different depths from Soma Basin in Turkey was performed. Coal samples which come from Soma basin were broken and grinded into small particles, and then passed through the sieve to reduce particle size to 80 µm. Resulting samples were preserved at nitrogen atmosphere. For the determination of the nitrogen, sulfur and carbon content in the coal, ultimate and proximate analysis of the samples were performed. Table1. Ultimate and proximate analysis of the Soma lignites Samples Sample Depth No 690.45 K1 626,20 K2 726,20 K3 728,20 K4 779,60 K5
Original sample
Dry sample
C%
H%
N%
S%
C%
H%
N%
S%
O%
44,5350 24,8250 59,4530 56,1470 68,1460
3,9394 2,9924 5,7984 4,7999 5,6492
1,2000 0,4089 1,8942 0,9614 1,6289
0,6966 2,0001 1,2361 1,4146 1,3910
47,5490 26,1840 63,6270 59,8460 73,1650
3,4486 2,5437 5,4198 4,3790 5,2411
1,2812 0,4313 2,0272 1,0247 1,7489
0,7438 2,1095 1,3229 1,5078 1,4934
9,4874 10,4815 12,6931 11,2225 10,5216
Samples
Original sample
Sample No
Depth
Moisture %
Volatile Matter %
Ash %
K1 K2 K3 K4 K5
690.45 626,20 726, 20 728,20 779,60
6,34±0,03 5,19±0,02 6,56±0,03 6,18±0,03 6,86±0,03
26,21±0,03 29,58±0,04 37,63±0,05 35,12±0,05 37,94±0,05
35,12±0,04 55,23±0,06 13,93±0,01 20,66±0,02 7,29±0,01
Dry sample Fixed Carbon % 32,33 10,00 41,88 38,04 47,91
Volatile Matter % 27,98 31,2 40,27 37,43 40,73
Ash % 37,49 58,25 14,91 22,02 7,83
Fixed Carbon % 34,53 10,55 44,82 40,55 51,44
Furthermore, aromatic groups of the coal samples are degraded into aliphatic chains by microorganisms during the biogasification process. Therefore, the functional groups of the sifted coal samples were determined by the Bruker Equinox 55 FTIR spectrometer.
140
T R 120 A N S 100 M I T 80 T A N 60 C E [%] 40 20 0 4000
3500
3000
2500
K1
2000
1500
1000
500
Wavenumber cm-1 K2 K3 K4 K5
Fig.1. FTIR spectrometry for Soma lignites
Figure1 shows that FT-IR spectra of the lignite samples which were evacuated from different depths of Soma basin. Most of the peaks are common for all samples. The broad band at 3400 cm-1 is due to the O-H and N-H groups. The peaks at 2900-2800 represent C-H stretching vibration of aliphatic and alicyclic methyl, methylene and alkyl groups. The intensity of the peak at 2900 cm -1 is greater than intensity of the peak at 2800 cm-1. This shows the presence of long aliphatic chains in the samples. Also, the peak at 1600 cm-1 and the shoulder at 1700 cm-1 indicate the high concentration of conjugated carbonyl structures and conjugated aromatic structures. Relatively strong band at 1400 cm-1 is due to the C-H bend in methyl, methylene or aromatic C-H bending. Usually, peaks between 1100 cm-1 - 400 cm-1 belong to inorganic mineral matter in the coal. For the chemical solubilization experiments, sodium oxalate, sodium carbonate and sodium phosphate were used as Lewis bases. 10 ml, 0.1 M aqueous solutions of sodium oxalate, sodium carbonate and sodium phosphate at varying pH values (4-12) were prepared. Then 50 mg of coal was added to these solutions. Mixtures were shaken at 200 rpm, for 24 hours on a rotary shaker. After that time, aliquots were centrifuged at 5000 rpm for 20 min and absorbance intensities of final solutions were determined by using UV-visible spectrometer.
Fig.2. Absorbance vs. pH dependence chart
This figure shows the comparison of the absorbance intensity at 275 nm of the solubilized coal in the aqueous solution of the different Lewis bases at varying pH. At high pH values (pH ≥10), solubility of the coal is more in carbonate and phosphate anions systems than oxalate anions. On the other hand, at lower pH levels between 9≥ pH ≥5, oxalate anion system solved coal more effectively. This result is important, since in the microbial gasification processes of the coal, harsh conditions reduce microorganism’s activity. Therefore, higher solubilization of the coal at moderate pH levels is a very important parameter to increase the efficiency of the biogasification processes due to the decomposition of the complex structure of the coal into the smaller organic substances which are more easily converted to methane by microorganism. For the future prospect of this study, Lewis bases will be used for the direct and indirect biogasification of the lignites samples to increase efficiency of these processes. Gas adsorption experiments were performed by using a Hiden isochema intelligent gravimetric analyzer IGA-001. Mechanism of the system is based on measuring the weight change of the sample against reference as a function of time, pressure and temperature. High pressure sensor (vacuum to 10bar) and temperature controller allow us to make a wide variety of sorption measurements and gives information about kinetics of the sorption processes. The precision of the measurement can be controlled by PC with long term stability of microbalance was 0.1µg with a weighting resolution 0.2 µg. To understand the mechanism of the coal adsorption, we start with nitrogen adsorption isotherm at room temperature with temperature stability is 0.1oC. Coal sample was outgased at 105oC until the constant weight was achieved at 10-6 mbar. For the biogasification experiments, anaerobic media which contain minerals and vitamins were prepared to isolate microorganisms from coal in order to understand the mechanism of the methane generation due to the endemic population. After this procedure, optimization of the methane generation will be done by adjusting pH, temperature and coal amount.
Acknowledgements We would like to thank Fırat Duygun from TUBITAK Marmara Research Center, Earth and Marine Sciences Institute for ultimate and proximate analyses.
References [1] S.H. Harris et al, International Journal of Coal Geology 76, 46–51(2008) [2] Cohen, M.S., Gabriele, P.D., Applied and Environmental Microbiology 44, 23-27 (1982) [3] Rice, D.D., Claypool, G.E., American Association of Petroleum Geologists Bulletin 65, 5–25 (1981) [4] Isbister, J. D., Barik, S., Crawford, D. L., Ed.; CRC Press: Boca Raton, FL,139-156 (1993) [5] Bumpus et al, Energy & Fuels, Vol. 12, No. 4, 664-671 (1998)
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 PROGRAM TOPIC: COAL-DERIVED PRODUCTS (FISCHER-TROPSCH) CATALYTIC PERFORMANCE IN FIXED-BED AND BUBBLING FLUIDIZED-BED REACTOR DURING FISCHER-TROPSCH SYNTHESIS ON THE IRON-BASED CATALYSTS Jong Wook Bae* Petroleum Displacement Technology Research Center, Korea Research Institute of Chemical Technology (KRICT), P.O. Box 107, Sinseongno 19, Yuseong, Daejeon, 305-600, Korea *Presenting author: Tel.: +82-42-860-7383 Fax: +82-42-860-7388; E-mail address: [email protected]. Ki-Won Jun, Yun-Jo Lee and Kyoung Su Ha Petroleum Displacement Technology Research Center, Korea Research Institute of Chemical Technology (KRICT), P.O. Box 107, Sinseongno 19, Yuseong, Daejeon, 305-600, Korea Abstract: Fischer-Tropsch synthesis (FTS) for olefin production from syngas was investigated on the four different iron-based catalysts in a fixed-bed and a bubbling fluidized-bed reactor. The catalysts were prepared by wet-impregnation using Al2O3, SiO2 and iron ore (FeOx) with active components of Fe, K and (or) Cu, and K/FeCuAlOx catalyst was prepared by co-precipitation method. The impregnated K/FeOx catalyst is found to be one of the promising catalysts to be applied in bubbling fluidized-bed reactor for high temperature FTS reaction due to its high resistance to catalyst attrition with high catalytic performance. Keywords: Fischer-Tropsch synthesis, iron-based catalyst, fixed-bed reactor, bubbling fluidizedbed reactor. 1. Introduction FTS is an important technology in the production of liquid fuels and valuable chemicals from syngas derived from the gasification of coal or biomass and the reforming of natural gas or other carbon-containing components.1-3 In FTS process, the reactor is chosen appropriately according to the targeted products such as light olefin, gasoline, diesel or heavy hydrocarbons including wax etc., and it also requires the appropriate selection of catalyst preparation method depending on the types of reactor such as plug flow (fixed-bed or supercritical), fluidized-bed or slurry bubble column reactor.4-6 In the present investigation, FTS reaction with the various iron-based catalysts for the production of light hydrocarbons from syngas was carried out in fixed-bed reactor (FBR) and bubbling fluidized bed reactor (BFBR). 2. Experimental Iron-based FTS catalysts adopted in the present study were prepared by using the conventional coprecipitation for K/FeCuAlOx catalyst and wet-impregnation method for FeCuK/γ-Al2O3,
FeCuK/SiO2, and K/FeOx catalyst. Catalytic performance on the four different iron-based FTS catalysts was studies in FBR (I.D. = 10.7 mm) with a catalyst of 0.3 g possessing a particle size of 53 - 120 µm. In addition, catalytic performance was also carried out in BFBR (0.034 m I.D. × 1.5 m height) equipped with the perforated type of gas distributor. The catalyst possessing same particle size distribution adopted in FBR were loaded with 100 g in BFBR equipped with a perforated plate containing 11 evenly spaced holes of 1.0 mm diameter, which served as a reactant distributor. Prior to the reaction in both reactors, the catalyst was reduced at 450 oC for 24 h in a flow of 50 % H2 balanced with helium. After reduction, the synthesis gas (H2/CO = 2) was fed into the reactor. The FTS reaction was carried out subsequently under the following reaction conditions; T = 300 oC, P =1.0 MPa and Ug (linear velocity of syngas) = 0.2, 0.4, 1.0, 2.0, 4.0, 8.0 and 10.0 cm/s for BFBR and T = 300 oC, P =1.0 MPa and space velocity (SV) = 2000 ml/gcat/h for FBR. 3. Results and discussion The textural properties of four different catalysts prepared using different preparation methods are summarized in Table 1. The larger surface area was observed on the impregnated catalysts such as FeCuK/Al2O3 and FeCuK/SiO2 around 192.1 and 157.8 m2/g separately. The lower average pore diameter on the impregnated FeCuK/Al2O3 and co-precipitated K/FeCuAlOx catalysts was observed around 4.4 and 5.9 nm respectively. The lower average pore diameter on FeCuAl/Al2O3 catalyst with high surface area is mainly attributed to the presence of bimodal pore structure with the co-existence of micro and mesopores in the regions of 3-5 nm and 10-20 nm. The apparent and bulk density of four different catalysts is in the following order; K/FeOx > K/FeCuAlOx > FeCuK/Al2O3 > FeCuK/SiO2. Table 1. Textural properties of the iron-based FTS catalysts catalysts
a
FeCuK/Al2O3 FeCuK/SiO2 K/FeCuAlOx K/FeOx
preparation method (wt. ratio of catalyst) Impregnation (20/2/4/100) Impregnation (20/2/4/100) Coprecipitation (4/100/6.6/15.7) Impregnation (6/94)
surface area (m2/g)
pore volum e (cm3/g)
average pore diameter (nm)
apparent density (g/cm3)
192.1
0.264
4.4
1.53
157.8
0.678
13.0
0.84
89.5
0.171
5.9
2.43
0.6
0.010
14.4
3.81
a
FeCuK/Al2O3 and FeCuK/SiO2 catalysts were prepared by conventional impregnation method at a fixed weight ratio of Fe/Cu/K as 1/0.1/0.2 on γ-Al2O3 abd SiO2 support with a fixed amount of Fe/support at 20/100. K/FeCuAlOx catalyst was prepared by coprecipitation method using K2CO3 (Fe:K = 100:4 wt%) and K/FeOx catalyst was made by wet-impregnation of K2CO3 on iron ore (FeOx) with K/Fe weight ratio of 6/94.
The summarized results of C2 - C4 yield with respect to superficial gas velocity in FBR and BFBR are shown in Figure 1. The hydrocarbon selectivity in BFBR is found to be shows a superior than that of FBR. The beneficial effects are the main advantages of BFBR for the production of C2 - C4 olefins. The higher formation rate of active iron carbide species with facile reducibility of iron oxides and appropriate acidity plays an important role to obtain high catalytic performance in FBR and BFBR.
Figure 1. The yield to C2 - C4 olefin with respect to the superficial gas velocity in FBR and BFBR 4. Conclusions FTS reaction for the production of C2 - C4 light olefins from syngas was carried out in FBR and BFBR for HTFT reaction with four different iron-based catalysts. Under similar operating conditions such as temperature and pressure, the C2 - C4 selectivity and olefin selectivity in BFBR are higher than that in FBR on the K/FeCuAlOx and K/FeOx catalysts. K/FeOx catalyst which is simply impregnated with potassium promoter shows a high resistance for catalyst attrition with high catalytic performance and this catalyst can be one of the best catalysts for applying HTFT reaction using BFBR. References 1. 2. 3. 4. 5. 6.
A.Y. Khodakov, W. Chu, P. Fongarland, Chem. Rev. 107 (2007) 1692-1744. B.H. Davis, Top. Catal. 32 (2005) 143-168. M.E. Dry, Catal. Today 71 (2002) 227-241. D.B. Bukur, X. Lang, L. Nowicki, Ind. Eng. Chem. Res. 44 (2005) 6038-6044. S.H. Kang, J.W. Bae, K.J. Woo, P.S. Sai Prasad, K.W. Jun, Fuel Process. Technol. 91(4) (2010) 399-403. S.H. Kang, K.J. Woo, J.W. Bae, K.W. Jun, Y. Kang, Korean J. Chem. Eng., 26(6) (2009) 1533-1538.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
Abstract Submission
COAL-DERIVED PRODUCTS: (COAL-TO LIQUIDS)
Operation of slurry reactor for Fischer-Tropsch synthesis
Ho-Tae Lee, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3662
Jung-Il Yang, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3795 Jung Hoon Yang, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3695 Dong-Hyun Chun, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3071 Hak-Joo Kim, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3654 Heon Jung, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3663
ABSTRACT A slurry bubble column reactor with a capacity of 0.03 bbl/day, was designed and operated for Fischer-Tropsch reaction. Active Fe based catalysts were prepared and tested at 2.5 MPa and H2/CO=1. The average CO conversion higher than 80% was observed. The effects of the reaction temperature and the superficial velocity of synthesis gas on the conversion, the product selectivity and the oil productivity were investigated. The liquid oil productivities increased with the increasing superficial velocity and the increasing temperature.
Coal-Derived Products Process Simulation of Steam Hydrogasification to Produce F-T Products and Electricity Xiaoming Lu, Chan S Park*, Joseph M Norbeck Bourns College of Engineering Center for Environmental Research and Technology Department of Chemical and Environmental Engineering, University of California, Riverside *Corresponding author: Tel: 1-951-781-5771; Fax: 1-951-781-5790; E-mail address: [email protected] Abstract: A new process for the co-production of synthetic fuels and electricity based on steam hydrogasification is being developed at the College of Engineering – Center for Environmental Research and Technology (CE-CERT) at the University of California, Riverside. One of the key benefits of this new process is the enhanced conversion of carbonaceous material to synthesis gas compared to other thermochemical processes. Another benefit is that it does not require the use of oxygen from a cryogenic air separation unit or ASU, thus reducing capital costs. One current coal-based application being considered for this technology (called the CE-CERT process) is to co-produce Fischer Tropsch products and electricity. The results of a process simulation model of an integrated conceptual design of a 4000 TPD Sub-Bituminous coal conversion to Fischer-Tropsch liquids and electricity is presented in this paper. A circulating fluidized bed with a regenerator setup for providing the heat to the Steam Hydrogasification Reactor (SHR) is modeled. The process simulation model involves the major steps of: 1) simulation of the steam hydrogasification reactor using gasification units based on built-in Aspen reactor blocks with the determination of the equilibrium composition of the gaseous components in the reactor by means of Gibbs free energy minimization; 2) empirical simulation of a warm gas cleanup system; 3) a steam methane reforming process simulated by using a built-in REQUIL equilibrium block; 4) empirical simulation of a hydrogen separation process for syngas ratio adjustment and excess hydrogen recycle to the steam hydrogasifier; and 5) a Fischer-Tropsch diesel synthesis by means of an empirical expression. The regenerator is essentially a combustor where residue char is burnt in the presence of air to heat circulating sand. The hot sand flows back into the SHR to provide the process heat for the main reactor. The material and energy balance of the whole process was developed using Aspen Plus. The overall process performance and the optimum F-T liquids/electricity output from the 4000 TPD coal plant is determined. The optimum process thermal efficiency is 50.25% with 37.1% of coal carbon in F-T liquids. 1.
Introduction
Petroleum was widely used in the 20th century and dominated the energy production and chemical industries. Fuels from crude fossil oil supply about 96–98% of the worldwide energy for transportation (cars, ships, airplanes). More than 55% of the petroleum extracted is refined to produce fuels. Estimates of remaining petroleum availability at the present rate of consumption span from 40 to 60 years [1]. Not surprising the price of oil rose from $12/barrel to more than $130/barrel during the period from 1945 to 2008 [2]. The price is expected to increase in future years. Conversely, the currently known reserves of coal exceed those of crude oil by a factor of 25 [3]. Production of syngas from coal and the successive conversion to a range of fuels and chemicals become of increasing interest as the reserves of crude oil are depleted and the global price per barrel increases. The University of California Riverside’s College of Engineering Center for Environmental Research and Technology (CE-CERT) is developing a new process for the production of F-T products and electricity based on steam hydrogasification. This process is also called the CE-CERT process. The details of this technology have been published elsewhere [4-7] but a short description of the important unit operations are presented below for completeness. Initially, the feedstock is made into a slurry with water and is fed
along with H2 into the Steam Hydrogasification Reactor (SHR). The SHR produces methane, along with carbon monoxide and carbon dioxide. A regenerator is setup to provide the process heat for SHR by combusting the leftover char using circulating sand [8]. The methane rich gas from the SHR is then subjected to cleanup in order to remove contaminants, primarily sulphur species. The methane is then converted into a synthesis gas, a mixture of H2 and CO, in the Steam Methane Reformer (SMR) [9]. The SMR is followed by a fuel synthesis reactor, usually a F-T reactor. Alternatively, the syngas can be used for power generation if desired. The excess H2 from the syngas is separated and is fed into the SHR as feed thus eliminating the need for a separate source of hydrogen. The F-T reactor operating parameters and catalyst can be varied to obtain a maximum yield of diesel, gasoline or kerosene or other fuels such as jet fuel. The basic chemical reactions taking place during the different stages of the process are given below. Steam Hydrogasification (SHR): C + H2O + 2H2 → CH4 + H2O + others (CO, CO2, C2+, etc)
(1)
Steam Methane Reforming (SMR): CH4+ H2O → 3H2+ CO
(2)
F-T Reaction (FTR): CO + 2H2 → -(CH2)- + H2O 2.
(3)
Features of CE-CERT process
A schematic of the CE-CERT process is shown in Fig.1. The main features are summarized below. 1.
Enhanced conversion of carbonaceous material to synthesis gas compared to other thermochemical processes.
2.
No need for oxygen from a cryogenic air separation unit or ASU and a moderate operation temperature and pressure compared with partial oxidation gasification, thus reducing capital costs.
3.
Slurry feed, wet feedstock can be used directly which reduces cost of drying the feedstock and offers the potential of more efficient handling of feedstock.
4.
Closed-loop H2 cycle, no external H2 is needed and make the process self sustainable.
5.
A high rate of methane is produced in SHR and be used as a source of clean synthetic natural gas if liquid fuel is not desired.
6.
A flexible control, optimum H2 to CO ratio for efficient downstream production of fuel products can be achieved by controlling the input H2O/feed ratio in the SHR.
Fig.1 Schematic of the CE-CERT Process
The technologies for the commercial production of F-T diesel using synthesis gas are currently considered to be mature. As mentioned earlier, the versatile nature of the CE-CERT process allows the product syngas H2/CO ratio to be controlled in a relatively simple manner by varying the H2 to carbon and H2O to coal ratios of the SHR feed. This enables the production of F-T products in an efficient manner irrespective of the moisture content of the coal and catalyst used in F-T reactor. 3.
Aspen Plus Simulation Details and Assumptions
Peng–Robinson equation (PR) is used to estimate all physical properties for the gasification and downstream unit operations. The SOLIDS property option is used for the coal crushing and screening section. The enthalpy model for both COAL and ASH, the nonconventional components, is HCOALGEN and the density model is DCOALIGT. The HCOALGEN model includes a number of empirical correlations for heat of combustion, heat of formation and heat capacity. All other values used were retrieved from the Aspen Plus database. For these simulations the H2/CO ratio of the synthesis gas feed to the F-T reactor will be set at 2.0 since the F-T model reactor is assumed to use a cobalt catalyst. The temperature of the SHR is set to be 750 ºC and the SMR is at 850 ºC for the simulation results shown below. The F-T reactor is operated at 220 ºC. The operating pressure is set at 400 psi for all the reactors. The average operation condition for the model is a H2O/Coal mass ratio of 2.1and a H2/C molar ratio of 1.0. This was selected to be consistent with the previous simulation results of this process [6]. Material and energy balances were accounted for and solved for every process unit, although no detailed chemical kinetic models were considered in the simulation. Detailed operation units used in the simulation of the CE-CERT process is given in Table 1. Table.1 Representive unit operations used in the simulation
Unit operation
Aspen plus model
Specifications
Feedstock crushing
Crusher
Rigorous simulation of particle size distribution
Feedstock particles screening
Screen
Rigorous simulation of the separation efficiency of the screen
Feedstock gasification
RGIBS
Specification of the possible products: H2O, H2 ,Cl2, HCl, C, CO, CO2, CH4, COS, H2S, CS2
Regenerator
Rstoic
Rigorous simulation of char combustion
Solid removing
Sep
Simplified simulation of gas/solid separation
Warm gas cleanup
Absorber, Rstoic
Rigorous simulation of the H2S, trace metal chloride removal
SMR furnance
Rstoic
Rigorous simulation of fuel gas combustion
H2 separation
Sep, Split
Simplified simulation of gas separation and split
FT reactor
RYield
Empirical simulation of F-T production distribution
HRSG
Counter current multiple
Rigorous simulation of the steam cycle with heat
stream heat exchanger
recovery
CO2 Recovery
Absorber, Rstoic
Rigorous simulation of the CO2 amine unit
Gas and steam turbines
Compressor
Calculate power produced
and
The main assumptions made in the process simulation are listed as below. 1.
No tar generated in the steam hydrogasification process.
2.
Process is steady state and isothermal, no heat and mass loss during the whole process.
3.
The product chain growth rate in F-T reaction is 0.9 with one single syngas pass.
4.
The byproducts such as sulfur and ammonia are not considered as credits.
In F-T reactor, the -(CH2)- is a basic structure for hydrocarbons with long chain. A main characteristic regarding the performance of the F-T synthesis is the process liquid selectivity. The liquid selectivity is determined by chain growth probability. This is the chance that a hydrocarbon chain grows with another -(CH2)- group instead of terminating. A high liquid selectivity (or C5+ selectivity: SC5+) is necessary to obtain a maximum amount of long hydrocarbon chains. The yield in the C1–C4 range decreases with increasing SC5+ and any C1–C4 in the outlet may efficiently be used for power generation. The F-T reactor block used an external model called through a FORTRAN module. This external model was empirically developed by Hamelinck et al [10], to predict the selectivity of the F-T process and can be expressed as below.
S C5+ = 1.7 + 0.0024T + 0.088
[H 2 ] + 0.18([ H 2 ] + [CO]) + 0.0078 pTotal [CO]
(4)
Where, Sc5+ – Mass fraction of hydrocarbons in the product with 5 or more carbon atoms [H2] and [CO] - Concentrations of H2 and CO expressed as fraction of the feed gas T – Temperature (K) P – Pressure (bar) According to Hamelinck et al, a least sum of squares fit of the above model with proprietary data resulted in the following equation, which was also found to be in accord with experimental results on cobalt catalysts reported [3]. This equation has been used to simulate the F-T reaction in this paper. The F-T liquids produced in the reactor is naphtha, diesel and wax. Waxes are finally converted into high quality diesel by hydrocracking. The products of hydrocracking are diesel (80%), naphtha (15%) and gaseous compounds, such as methane, ethane, propane and butane (5%) by weight [11]. The whole simulation is controlled using FORTRAN routines (calculator blocks) and design specifications to reduce the number of independent specifications and to adjust automatically those associated variables, i.e. The dependent variables are automatically adjusted when independent input variables are modified by the a calculator block or a design specification. The main functional relationships of the simulation are: the amount of H2O input as a function of the feedstock mass input, the amount of H2 input as a function of the carbon in the coal, the recycled H2 depends on the H2O input. The coal used in the simulation is Utah bituminous coal. The physical properties of the coal are given in Table 2. A heat exchange net is built based on maximization of heat recovery [12]. The results from Aspen Plus are directly imported into an excel worksheet and the mass and energy balance are calculated using excel. 4.
Simulation results
The simulations are based on a 4000 TPD coal plant with 65% CO2 capture to product F-T products and electricity using CE-CERT steam hydrogasification technology with H2/CO ratio of 2.0. The plant is operated under a capacity factor of 90%. The schematic diagram of the simulation process is shown in Fig 2. The total plant performance is compared with the simulation results in
Table.2 Feedstock analysis of Utah bituminous coal Ultimate Analysis Carbon
Dry, %
As Received, %
68.85
58.40
Hydrogen
4.74
4.02
Nitrogen
1.04
0.88
Sulfur
1.18
1.39
Ash
10.57
9.94
Oxygen
11.39
13.43
Proximate
Dry Basis, %
Moisture
As Received, %
--
15.18
Ash
10.55
8.95
Volatile Matter
40.00
33.93
Fixed Carbon
49.45
41.94
3.07
2.89
12,077
10,244
Sulfur Btu Content
National Energy Technology Laboratory (NETL) report within the Department of Energy (DOE). NETL has reviewed the CE-CERT process through a Cooperative Research And Development Agreement (CRADA) by validating the equilibrium model, process flow sheet for producing F-T products using CE-CERT technology with H2/CO ratio of 1.0 in 2008. The performance comparison of the two plants is given in Table 3. Table.3 Performance of 4000 TPD coal plants using different H2/CO ratio H2/CO ratio
2.0
1.0
Plant input (TPD)
4000
4000
Net Plant Power (MW)
74
107
9,990
7,143
50.25%
53.40%
CO2 Recovered (lb/hr)
321,742
306,507
Total Carbon fixed in the F-T liquids
37.10%
28.80%
F-T Production (BPD) Net Plant Effective Thermal Efficiency (HHV)*
*-ETE=(Higher Heating Value of F-T products + Electrical equivalent)/Thermal input
Fig.2 Schematic diagram of CE-CERT simulation process 5.
Conclusion A process simulation based on CE-CERT steam hydrogasification technology for co-production of F-T products and
electricity was developed using Aspen Plus simulation. The process includes an internal heat regenerative step where char from the gasifier is sent to a combustor and recirculation of sand is used to heat the main reactor. Detailed simulation methods and operation conditions were introduced. H2O/Coal mass ratio of 2.1and H2/C molar ratio of 1.0 was used as optimal operation condition in the process with Utah bituminous as feedstock. The overall process performance for a 4000 TPD coal plant with 65% CO2 capture were determined. Based on the simulation results, the plant had a F-T liquids production of 9,990 BPD along with 74MW electricity available for export to the grid with a process thermal efficiency of 50.25%. The total carbon fixed in the F-T liquids was 37.1% of coal carbon.
Acknowledgement Thanks to the NETL for their dedicated CRADA support and partial funding support from Viresco Energy, LLC.
Reference 1.
Ifp,
Innovation
Energy
Environment.
Crude
oil
supply
and
oil
availability
estimations.
See
also:
http://www.ifp.fr/IFP/en/files/cinfo/IFP_Panorama05_06-CarburantsalternatifsVA.pdf 2.
Federal Reserve bank of St. Louis. Advancing economic knowledge through research & data. Oil price available.
See also:
http://research.stlouisfed.org/fred2/series/OILPRICE/98/10yrs. 3.
JR. Anderson and M. Boudart, Berlin, Dry ME, The Fischer–Tropsch synthesis, Catalysis: science and technology, Germany, Springer, 160-253, 1981.
4.
JM. Norbeck and CE. Hackett, United States patent number 7, 208,530 B2
5.
SK. Jeon, CS. Park, CE. Hackett and JM. Norbeck, Characteristics of steam hydrogasification of wood using a micro-batch reactor, Fuel, 86, 2007, 2817-2823.
6.
CS. Park, SP. Singh and JM. Norbeck, Steam hydrogasification of carbonaceous matter to liquid fuels, 24th Annual International Pittsburgh Coal Conference, Johannesburg, South Africa, 2007
7.
ASK. Raju, CS. Park and JM. Norbeck, Baseline technical and economic assessment of a small scale steam hydrogaisification process with Fischer-Tropsch liquids facility. 25th Annual International Pittsburgh Coal Conference, PA, USA, 2008.
8.
Masuo Hasegawa and Junzo Fukuda, Gasification of solid waste in a fluidized bed reactor with circulating sand, Conservation & Recycling, 3, 1979, 143-153.
9.
ASK. Raju, CS. Park, JM. Norbeck, Synthesis gas production using steam hydrogasification and steam reforming, Fuel Processing Technology, 90, 2009, 330-336.
10. CN. Hamelinck, APC. Faaij, H. Uil and H. Boerrigter, Production of FT transportation fuels from biomass; technical options, process analysis and optimization, and development potential, Energy, 29, 2004 11. Tijmensen MJA, et al. Exploration of the possibilities for production of Fischer-Tropsch liquids and power via biomass gasification. Biomass and Bioenergy, 2002, 23:129-52. 12. The Cost and Performance Baseline for Fossil Energy Power Plants, Volume 1: Bituminous Coal and Natural Gas to Electricity” DOE/NETL 2007/1281 (May 2007).
Further development of the PSRK model for the prediction of the Vapor-Liquid-Equilibria of Direct Coal Liquefaction System at High Temperatures and High Pressures Xue-feng Mao*, Shi-dong Shi , Wen-bo Li, Zhen-nan Gao (China Coal Research Institute, 100013, Beijing, China) Abstract: Since Huron and Vidal (1979) developed the basic idea of so called G E mixing rules, similar models have been proposed by different authors. The aim of all recent developments of G E mixing rules is to combine the successful GE models or group contribution methods with equations of state to enable the description of Vapor-liquid equilibria at high temperatures and pressures including supercritical compounds. The system involved in direct coal liquefaction was a three-phase state of solid, liquid, and gas phases at high temperatures and high pressures. One of the challenges in the analysis of the Coal liquefaction process was due to the fact that the contents of both hydrogen and methane are very high and both of them were in the supercritical state. Besides, when small, spherical hydrogen and methane molecules were mixed with liquefaction oil, the huge difference between their molecule weights causes extreme asymmetry of the system, bring about the difficulty in analysis and modeling of the liquefaction process. To Predict the vapor-liquid equilibria of direct coal liquefaction system at high temperatures and high pressures, a group contribution equation of state called PSRK was proposed , wherein the group contribution equation of state PSRK(predictive Soave-Redich-Kwrong) as suggested by Holderbaum and Gmehling (1991) combines the mod UNIFAC model (Hansen et al.1991) with the SRK equation of state. In this work, the range of applicability of the PSRK method was extended by the introduction of additional gases and the determination of missing interaction parameters between the following gases: CH4, CO2, CO, H2S, H2, H2O and the original UNIFAC structural groups. A computational method to solve the flash model for coal liquefaction system at high temperatures and high pressures has been developed. The numerical code has two-cycle iteration, including the inner cycle of β iteration and the outer cycle of K iteration, and the results of fast convergence can be achieved. Generally, after about 5 times of iteration (outer cycle), the program can reach the convergence precision. Based on the model of PSRK, the composition of 24 components of the gas-liquid phase and the gas-liquid equilibrium constant has been calculated. Calculated results were consistent with the literature results and the vapor-liquid equilibria for the coal liquefaction reactor were calculated. The comparison of the results showed that the capability of the PSRK model to describe and predict the vapor-liquid equilibria of coal liquefaction systems at high temperatures, high pressures, strong asymmetry and strong polar.
1. Introdcution Since the pioneer work of Vidal[1],who matched the excess Gibbs free energy GE,from an equation of state (EOS) at infinite pressure with that from an existing GE model,the so-called EOS/ GE method has received great attention. One of them is the PSRK model proposed by Holderbaum and Gmethling [2],wherein the GE-mixing rule the UNIFAC model[3] is combined with the SRK equation of state, the PSRK method is one of the most successful models,whose wide practical applicability has identified by many investigations due to its good predictive
accuracy and the large parameter matrix.However,as pointed out by many researches[4,5-7]. A lot of attempts have been made to improve it to allow PSRK model to describe complex systems accurately over wide temperature and pressure ranges[8-12], Among the various efforts, it does not work well for highly asymmetric systems,For example, The system involved in direct coal liquefaction was a three-phase state of solid, liquid, and gas phases at high temperatures and high pressures. One of the challenges in the analysis of the Coal liquefaction process was due to the fact that the contents of both hydrogen and methane are very high and both of them were in the supercritical state. Besides, when small, spherical hydrogen and methane molecules were mixed with liquefaction oil, the huge difference between their molecule weights causes extreme asymmetry of the system, bring about the difficulty in analysis and modeling of the liquefaction process. To Predict the vapor-liquid equilibria of direct coal liquefaction system at high temperatures and high pressures, a group contribution equation of state called PSRK was proposed. In this work, the range of applicability of the PSRK method was extended by the introduction of additional gases and the determination of missing interaction parameters between the following gases: CH4, CO2, CO, H2S, H2, H2O and the original UNIFAC structural groups. 2. The Research Content The direct liquefaction was carried out at 0.01t/d at BSU at CCRI using Shenhua Shangwan bituminous coal. In order to get the vapor-liquid equilibrium constants under the real liquefaction conditions, the Pseudo-components and existing gas,such as H2 and C2H6 were combined together ot establish the coal lquefaction oil flash distillation system. By using the flash-distillation system,the Vapor-liquid equilibrium composition and distribution were calculated at different reaction temperature and pressure. 24 components were studyed as the research object,include 14 narrow boiling-range fractions and 10 Light Hydrocarbons gas. The resulted liquid product was separated as 14 narrow boiling-range fractions (IBP~110℃、110~150℃、150~180℃、180~200℃、200~220℃、 220~240℃、240~260℃、260~280℃、280~300℃、300~320℃、320~340℃、340~360℃、 360~380℃、380~450℃) successively by atmosphere distillation. 10 Light Hydrocarbons gas : H2、CH4、CO、CO2、C2H4、 C2H6、H2S、C3H8、C4、H2O 3. Thermodynamic model 3.1 The PSRK model The PSRK model was proposed by Holderbaum and Gmehling[2],which is a predictive SRK EOS with the MHV1 mixing rule[13] coupled with the UNIFAC model. And the constant A0=-0.64663 in the PSRK mixing rule is basically calculated using experimental liquid volumes in the saturated state of a large number of substances at atmospheric pressure. Details of the PSRK model are as follows. 3.1.1 The EOS The SRK EOS[14] is adopted in the PSRK model in the PSRK model:
p
RT a v b v (v b )
(1)
With the pure component parameters ai and bi obtained from the critical data Tc and Pc of the pure components
ai 0.42748
R 2Tci2 f (T ) pci
(2)
bi 0.08664
RTci pci
(3)
Enables the correct reproduction of pure component vapor pressures with the help of the Mathias –Copeman parameters c1,i c2,i and c3 i fitted to experimental vapor pressure data: 1
f (T ) [1 (0.48 1.574i 0.176i 2 )(1 Tr 2 )]2
(4)
or 1
1
1
f (T ) [1 c1,i (1 Tr ,i 2 ) c2,i (1 Tr ,i 2 )2 c3,i (1 Tr ,i 2 )3 ]2 1
f (T ) [1 c1,i (1 Tr ,i 2 )]2
Tr ,i 1
Tr ,i 1
(5) (6)
For nonpolar substances,Eq(4) is good enough,however,for polar substances,Eqs.(5) and (6) proposed by Mathias and Copeman[15] are much more accurate. Eqs.(5) and (6) are used in this work except for those large molecules,whose c1,c2,c3 cannot be found in the papers of Holderbaum and Gmehling and Fischer and Gmehling. In this case, Eq(4) is adopted.The critical properties and the acentric factors for those molecules whose values cannot be found in the book of Reid et al.[16] are listed in Table 1. Table 1 critical parameters and basic physical properties of the narrow boiling-range fractions Component
Tc(K) Pc(MPa)
i
Component
Tc(K)
Pc(MPa)
i
H2
33.2
1.3
0
150~180
649.9
31.718
0.342
CH4
111.6
4.6
0.008
180~200
685.1
32.201
0.370
CO
132.9
3.5
0.049
200~220
708.6
30.915
0.381
CO2
304.1
7.38
0.225
220~240
728.4
28.506
0.392
C2H4
282.4
5.04
0.095
240~260
747.9
26.372
0.417
C2H6
305.4
4.88
0.098
260~280
768.7
24.842
0.454
H2S
373.2
8.94
0.100
280~300
792.7
24.341
0.489
C3H8
369.8
4.25
0.154
300~320
816.0
23.707
0.535
C4
408.2
3.65
0.195
320~340
840.3
23.371
0.582
H2O
647.3
22.12
0.344
340~360
863.8
22.888
0.645
IBP~110
524.9
30.834
0.217
360~380
889.2
22.748
0.714
110~150
590.1
26.739
0.284
380~450
934.2
20.457
1.034
Applying Eos to mixtures,the parameters f (T ) and b can be calculated using the PSRK mixing rule. Therefore, the pure component parameters ai , bi and the excess Gibbs energy at a reference state( G0E ) are required. At the reference state (the liquid at atmospheric pressure) an optimized ratio of the inverse packing fraction ,the following relation is obtained. 3.1.2 The Mixing rule
The MHV1 mixing rule[14],so-called EOS/GE mixing rule,is used in the PSRK model:
GE a RT a b 0 xi i bi A1 i A1
b
x ln b i
i
i
(5)
b xi bi
(6)
i
Where a and b are the mixture co-energy and co-volume paramters,resperctively and GE is the molar excess Gibbs free energy,which can be calculated from a GE model.The constant A1 is set to be -0.64663,which is little different from the value in the original MHV1 mixing rule. The fugacity coefficient is given by
ln i
bi b b ( Z 1) ln Z (1 ) ln(1 ) b V V
(7)
ai 1 b bi ln i ln 1 A1 bi b bi RT
(8)
3.1.3 The Modified UNIFAC model For the modified UNIFAC model, interaction parameters between 45 subgroups and 21 main solvent groups have been determined by Lasen et al.(1987). In the present work we have added 10 new groups for systems involving gases. The group definitions for these new groups together with the structural parameters Rv and Qv (Sander et al.,1983) are listed in Table 2. Table 2 molecular Rv and Qv values and the main-group (MG) Definitions for the Gases Component
MG
Rv
Qv
H2
22
0.8320
1.1410
CH4
28
2.2440
2.3120
CO
25
2.0940
2.1200
CO2
26
2.5920
2.5220
C2H4
30
3.1482
2.9700
C2H6
31
3.6044
3.3920
H2S
27
2.3330
2.3260
C3H8
33
4.9532
4.4720
C4
34
6.3020
5.5520
H2O
35
0.9200
1.400
The modified UNIFAC parameters (aij)d etermined in this work for the interactions between the modified UNIFAC main groups 1-7, 9-10 (Larsen et al., 19871, and the 13 gases (main group 22-34) are given in Table III. In the expression used to describe the temperature dependence of the interaction parameters in modified UNIFAC we use only two terms,i.e.
aij aij ,1 aij ,2 (T T0 )
(9)
where To is a reference temperature, To = 298.15 K, and i and j represent the main groups. Further, all gas-gas interaction parameters have been assigned a value of zero. In Table 3 (a) we have given supplementary values of the interaction parameters between the CH2 group of an alcohol and the gases H2 and N2. It was not possible to describe the VLE behavior for N2 and H2 by using the same values for the interaction parameters (a CH3 of the CH2
groups resent in alkanes and in alcohols. The same problem appeared with the GC-EOS model for which special "water-soluble" CH2 groups in alcohols and ketones were introduced. The CH2 group present in an alcohol is here denoted CH2′.The a
CH2,gas values
in Table 3 should be used
when the CH2 group is in an alcohol molecule. Table 3 (a) Modified UNIFAC interaction Paramaters ( aij First Row Gives aij ,1 and Second Row aij ,2 (Eq ) i
j 22
25
26
27
28
30
31
33
34
1
509.0
58.78
123.9
188.3
-109.4
-133
-86.55
35.48
-22.08
CH2
1.440
0.000
-0.4065
0.0251
-0.8640
-0.8723
-0.7601
-0.0620
-0.5395
2
527.6
na
-28.36
na
na
-115.9
-173.9
-140.1
171.1
C=C
0.3671
na
1.169
na
na
0.0569
9.200
0.7277
0.5866
3
188.3
141
134.1
326.9
-3.522
214.9
1.951
2.323
2.323
ACH
3.334
-1.770
1.738
0.805
-0.4008
0.00
-0.2326
-0.6311
-0.6311
4
432.1
446.4
261.7
425.8
503.1
802
514.5
469.3
524.7
OH
-0.1257
-2.385
5.381
6.850
0.000
-5.435
-3.574
-4.203
-1.664
5
198.9
-19.85
-126.6
-132.8
-55.3
-70.84
-37.64
-67.57
-53.35
CH3OH
-2.138
-0.9319
-0.2024
0.000
-1.104
-0.8463
-0.8812
-0.7372
-0.8700
6
949.9
494
226.6
253.9
499.2
346.5
405
361.8
331.8
H2O
-0.3100
0.1390
-0.2410
-0.3050
-0.2550
-0.3326
0.0930
0.1271
-0.2550
7
273.7
na
81.51
na
769.5
216.2
-53.51
-9.088
na
CH2CO
-6.293
na
0.000
na
0.00
0.000
0.000
-1.490
na
9
319
159.0
-80.55
-0.1893
na
na
-80.97
na
na
CCOO
-5.3620
-1.039
-0.5914
0.1456
na
na
0.000
na
na
10
341.3
57.04
117.7
152.4
177.5
na
308.5
na
na
CH2O
-7.130
-3.694
5.759
-2.750
-1.502
na
0.000
na
na
9
10
(b) Modified UNIFAC interacton Parameters ( aij ) i
j 1
2
3
4
5
6
7
22
-228.1
-227
-54.49
1746
1540.90
1586
273.7
319
341.3
H2
-1.304
-1.305
-1.545
5.153
-5.567
3.924
-6.293
-5.362
-7.130
25
4.278
17.54
141.0
1170
1368
1455
na
159.0
57.04
CO
0.000
0.000
-1.770
-9.344
-2.928
-2.906
na
-1.039
-3.694
26
-55.69
-47.16
-16.56
583.3
727.9
1067
-0.0172
153.0
82.87
CO2
-0.4904
0.6526
-1.928
-4.940
-1.331
-0.4180
0.00
0.4509
-2.877
27
-93.27
na
-196.2
594.7
800.3
753
na
-0.1893
152.4
H2S
-0.6816
na
-0.7045
3.480
0.000
-0.8809
na
0.1453
-2.750
28
118.0
na
148.7
979.7
1532
1608
-11.9
na
198.8
CH4
0.3910
na
-0.2084
0.000
-4.137
-2.059
0.000
na
0.5265
30
149.1
152.7
-144.5
802
672.6
1354
76.87
na
na
C2H4
0.5241
0.2693
0.000
-5.453
-2.160
-1.542
0.000
na
na
31
99.22
185
70.69
663.1
760.9
1529
579.2
501.6
138.6
C2H6
0.4350
-5.864
-0.1656
0.0349
1.374
-3.081
0.000
0.000
0.000
33
-27.08
234.9
84.60
561.1
1138.0
1644
115.1
na
na
C3H8
-0.1935
-1.321
0.3180
-0.9400
1.18
-3.55
3.014
na
na
34
31.77
-88.39
84.60
524.7
689.8
1606
na
na
na
C4
0.4364
-0.4392
0.3180
-1.664
4.078
-2.221
na
na
na
The original UNIFAC model was adopted as the required GE model in the PSRK model,however, for the interaction parameters of gas-containin group pairs temperature dependent interaction parameters were intorduced,and the UNIFAC expression
nm exp(
anm ) T
(10)
anm bnmT cnmT 2 ) T
(11)
Is replaced by
nm exp(
Eqs.(1)~(11) constitute the PSRK model,and a variety of group interaction parameters are available,which makes the model have a good applicability. However, this temperature dependence is only used if necessary,e,g. for interaction parameters between gases and water,where a large temperature range is covered,or if a strong temperature dependence of the phase equilibrium behavior (e.g. for the Henry coefficients) is observed. With PSRK ,the results of the UVIFAC group contribution method can be reproduced,and furthermore ,the model can be used at higher temperatures and pressurres even at super-critical conditions. The optimization of a large number of interaction parameters between UNIFAC and new PSRK main groups was performed using the vapor-liquid equilibrium ,gas solubility and critical data sotored in the Dortmund Data Bank (DDB).These developments enable the calculation of the phase equilibrium behavior and other thermodynamic properties for various of systems[17~18]. Up to now,parameters for the 50 original UNIFAC main groups and 31 new PSRK groups (exoxy group and 30 different gases,such as CO2,CO,O2,N2,CH4,etc.)were established.As for the pure component values,the supplemtary material contains all available interactions parameters for the PSRK/UNIFAC method. The current status of the parameter matrix is shown in Fig.1.
Fig.1 Current status of the PSRK group interaction parameter matrix. 4. Results and discussion In most cases only two binary interaction parameters(aij,aji) were fitted. In a few cases ,linear(bij,bji) or even a quadratic(cij,cji) temperature dependent paameters were required to describle the correct temperature dependence of the phase equilibrium behavior,e.g. for the system CH4+H2O.the obtained parameters for 34 group pairs are given in Table 2 and Table 3. By using the flash-distillation system,the Vapor-liquid equilibrium composition and distribution were calculated at reaction temperature and pressure (T=675K /P=19MPa) ,the result was shown in Table 4. Table 4 The Vapor-liquid equilibria Data of high temperature Separator (T=675K/ P=19MPa)
yi
Ki
xi
H2
0.1508
0.7796
5.1689
150~180
0.0112
0.0022
0.2000
CH4
0.0228
0.0734
3.2171
180~200
0.0119
0.0017
0.1465
CO
0.0024
0.0092
3.803
200~220
0.0163
0.0019
0.1194
CO2
0.0072
0.0064
0.8858
220~240
0.0503
0.0046
0.0914
C2H4
0.0003
0.0007
2.6133
240~260
0.04
0.0027
0.0681
C2H6
0.0098
0.0274
2.7777
260~280
0.0446
0.0023
0.0508
H2S
0.002
0.0036
1.7452
280~300
0.0565
0.0022
0.0381
C3H8
0.0085
0.018
2.1105
300~320
0.0878
0.0024
0.0276
Component
xi
yi
Ki
Component
C4
0.0031
0.0054
1.7551
320~340
0.1175
0.0023
0.0194
H2O
0.0038
0.0487
12.7421
340~360
0.0651
0.0009
0.0135
IBP~110
0.002
0.0011
0.5436
360~380
0.0482
0.0005
0.0096
110~150
0.0072
0.0022
0.3007
380~450
0.2306
0.0008
0.0033
So through the change of temperature, The Vapor-liquid equilibria Data of coal liquefaction reactor were shown in Table 5. Table 5 The Vapor-liquid equilibria Data of coal liquefaction reactor (T=728K/ P=19MPa)
xi
Component
Ki
yi
xi
Component
yi
H2
0.1905
0.7679
4.0307
150~180
0.0084
0.0025
0.2977
CH4
0.0249
0.0725
2.9059
180~200
0.0088
0.002
0.2311
CO
0.0026
0.0091
3.4733
200~220
0.0121
0.0024
0.1951
CO2
0.0101
0.0063
0.6258
220~240
0.0376
0.0059
0.156
C2H4
0.0003
0.0007
2.4942
240~260
0.0305
0.0037
0.1217
C2H6
0.0101
0.0271
2.6882
260~280
0.0348
0.0033
0.0961
H2S
0.002
0.0035
1.7662
280~300
0.0457
0.0035
0.0761
C3H8
0.0082
0.0178
2.1643
300~320
0.0744
0.0044
0.0586
C4
0.0029
0.0054
1.8771
320~340
0.1057
0.0046
0.0439
H2O
0.004
0.048
12.038
340~360
0.062
0.002
0.033
IBP~110
0.0016
0.0011
0.6747
360~380
0.0482
0.0012
0.0257
110~150
0.0056
0.0023
0.4132
380~450
0.2693
0.0029
0.0107
0.74 0.72 H2 Solubility,mol/kg
Ki
0.70 0.68 0.66 0.64 0.62 13.0
13.5
14.0
14.5
15.0
15.5
16.0
16.5
17.0
H2 partial pressure,MPa Fig. 2 H2 partial pressure -dependence curves of the hydrogen solubility at high temperature
0.95
H2 Solubility,mol/kg
0.90
0.85
0.80
0.75
0.70 280
300
320
340
360
380
400
420
440
460
480
Temperaturre,℃ Fig. 3 Temperature-dependence curves of the hydrogen solubility in high temperature Separator
Also, the solubility of hydrogen in oil of coal liquefaction at high temperature and high pressure has been obtained. Further the solubility of hydrogen linearly increases with the rise of partial pressure of hydrogen as well as the temperature. But the linear fitting in the curve of hydrogen solubility-partial pressure departs from the origin of coordinate, which demonstrates that the solubility of hydrogen in coal liquefaction oil does not strictly comply with the Henry’s law. 5. Conclusions A computational method to solve the flash model for a coal liquefaction system at high temperature and high pressure has been developed. The numerical code has two-cycle iteration, including the inner cycle of β iteration and the outer cycle of K iteration, and the results of fast convergence can be achieved. Generally, after about 5 times of iteration (outer cycle), the program can reach the convergence precision. Based on the model of PSRK, the composition of 24 components of the gas-liquid phase and the gas-liquid equilibrium constant has been calculated.
References [1] M.Huron, J. Vidal,New mixing rule in simple equations of state for representing vapour-liquid equilibrim of strongly non-ideal mixtures. Fluid Phase Equilibria 3 (1979) 255–272. [2] T. Holderbaum, J. Gehling, PSRK: A group contribution equation of state based on UNIFAC, Fluid Phase Eqnilibria 70 (1991) 251-265. [3] H.K. Hansen, P. Rasmussen, A.a. Fredenslund, M. Schiller, J. Gmehling, Vapor-liquid equilibria by UNIFAC group contribution: 5. revision and extension, Ind. Eng. Chem. Res. 30 (1991)2352-2355. [4] K.Fischer,J,Gmehling. Further development status and results of the PSRK method for the prediction of vapor-liquid equilibria and gas solubilities. Fluid Phase Equilibria 121 (1996) 185–206. [5] M.L. Michelsen, A modified Huron-Vidal mixing rule for cubic equation of state.Fluid phase Equilibria 60(1990) 213-219. [6] S.Dahl.M.L. Michelsen. High pressure vapour-liquid equilibrium with A UNIFAC Based equation of state. ALCHE J (1990) 1829-1836. [7] D.S.H.Wong,S.I.Sandler,A theoretically Correct mixing rule for cubic equations of state, ALCHE J .38(1992) 671-680. [8] Qingyuan. Yang, Chongli,A modified PSRK model for the prediction of the vapor-liquid equilibria of asymmetric systems, Fluid Phase Equilib. 192 (2001)103–120. [9] Jurgen. Gmehling, Jiding.Li, Kai.Fischer,Further development of the PSRK model for the prediction of gas solubilities and vapor-liquid-equilibria at low and high pressures Ⅱ, Fluid Phase Equilib. 141 (1997)113–127. [10] J. Li, K. Fischer, J. Gmehling, Prediction of Vapor-Liquid Equilibria for Asymmetric Systems at Low and High Pressures with the PSRK Model, Fluid Phase Equilibria submitted, 1997. [11] C. Boukouvalas, N. Spoliotis, P. Coutsikos, N. Tzouvaras, D. Tassio, Fluid Phase Equilib. 92 (1994) 75–106. [12] Jiding Li,Kai Fischer,Jurgen Gmehling. Prediction of vapro-liquid equilibria for asymmetric systems at low and high pressures with the PSRK model , Fluid Phase Equilibria 143 (1998) 71–82. [13] M.L. Michelsen, A modified Huron-Vidal mixing rule for cubic equations of state ,Fluid Phase Equilib. 60 (1990) 213–219. [14] Soave G.Equilibrium constants from a modified Redlich-Kwong equation of state[J].Chem Eng Sic,1972, 42:381~387 [15] P.M. Mathias, T.W. Copeman, Extension of the Peng-Robinson equation of state to complex mixtures: Evaluation of the various forms of the local composition concept ,Fluid Phase Equilib. 13(1983) 91–108. [16] R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, 4th Edition, McGraw-Hill, Singapore, 1987. [17] S. Horstmann, K. Fischer, J. Gmehling, PSRK group contribution equation of state: revision and extension III, Fluid Phase Equilib. 167 (2000) 173–186. [18] J. Gmehling, K. Jiding Li, Fischer, Further development of the PSRK model for the prediction of gas solubilities and vapor-liquid -equilibria at low and high pressures II ,Fluid Phase Equilib. 141 (1997)113–127.
Arsenic and Mercury Removal by Using Iron Humate Prepared from Turkish Coal Based Humic Acid Hacer Dogan1, Murat Koral1, Tulay Inan1, Selahattin Anaç2, Zeki Olgun2 1
TUBITAK Marmara Research Center, Chemistry Institute, 41470 Gebze, Kocaeli TURKEY 2 TKI(Turkish Coal Enterprises)Hipodrom Cad. No:12 Yenimahalle 06330 Ankara
Abstract: Humic acid produced by TKI (Turkish Coal Enterprises) was used in the production of iron humate. Different iron sources (iron (II) sulfate, iron (III) sulfate and iron (III) chloride) were used in the ion exchange reactions. Iron (II) sulfate (FeSO4) was found to be most suitable iron source in terms of iron content and adsorption capacity for As, Pb, Ni and Cd. The Fourier transform infrared (FTIR) spectroscopy, thermal gravimetric analyzer (TGA) and scanning electron microscopy (SEM) were used to characterize the humic acid and iron humate samples. The efficiency of iron humate as adsorbent has been studied as a function of amount, contact time and initial arsenic (As) and mercury (Hg) concentration by a series of batch experiments. The adsorption capacity of iron humate for As and Hg was above 90 % and higher than that of humic acid. It was concluded that iron humate can be used as an effective sorbent for the removal of As and Hg. Its application on the sorption of cadmium (Cd), cobalt (Co) and nickel (Ni) was not successful.
1. Introduction Humic acids (HAs) are present in soils, natural waters, river, lake and sea sediments, peat, brown and brown-black coals and other natural materials as a product of chemical and biological transformations of animal and plant residues [1]. The principal properties of HAs and their subsequent potential applications depend strongly on their origin (source) as well as on the isolation procedure. HA is a chemically heterogeneous compound having different types of functional groups in different proportions and configurations. It contains carboxyl (-COOH), amine(-NH2), hydroxyl(-OH), and phenol(Ar-OH) functional groups [3]. Carboxyl and hydroxyl groups are capable of substituting their hydrogen atoms for ions of metals. When humic acids interact with multi valency metals, such as iron, zinc, copper and others they form new type of insoluble compounds [4,5]. Metal humates can be prepared in a relatively simple way by the precipitation of HAs with suitable metal compounds. Recent years, low cost and highly effective adsorban materials have been developed for heavy metal adsorption. Some of the reported low-cost sorbents include bark/tannin-rich materials, lignin, peat moss, iron-oxide-coated sand, leaf mould, seaweed/algae/alginate, dead biomass etc. Humic acid based materials are also shown as one of the best examples for low cost sorbents. Iron humate that is markedly less soluble in aqueous solutions has been used as a new low-cost sorbent for inorganic and organic pollutants [6-10].
Humic acid extracted from Turkish coal in Ilgın Konya by TKI was used without any purification in the ion exchange reactions. The effect of the iron resources was examined on the ion-exchange reaction and the ability of the reaction was evaluated in terms of the iron content in the final product and adsorption of the heavy metal mixture containing As, Pb, Cd and Ni in the amount of each 10 ppm. Adsorption experiments have been performed for arsenic (As) and mercury (Hg) removal and the kinetics of the metal adsorption has been investigated. 2. Materials and Methods 2.1. Humic acid characterization Humic acid sent by TKI was analyzed and the results are given in Table 1. Carboxyl groups, total acidity and the coefficient E4/E6 were then determined. The amount of carboxylic groups and the total acidity were determined using the calcium acetate and the barium hydroxide methods, respectively. The quantity E4/E6 is the ratio of absorbances at 465 and 665nm of humic acid solutions. Elemental analyses of carbon, hydrogen, nitrogen and sulfur were obtained with a Thermo Finnigan Flash 1112 Series EA elemental analysis instrument. E4/E6 ratio was found as 3,26 of humic acid obtained from coal. This value shows that humic acid has high aromatic groups.
2.2. Adsorbent preparation Humic acid solution (1 % wt.) in the glass jacketed reactor was prepared and converted to sodium humate with NaOH granules at 60°C and stirred for two hours. pH value of the solution was measured as 11.2. Then iron salt (iron III chloride FeCl3.6H20, iron II sulfate-FeSO4.7H2O, and iron III sulfate- Fe2(SO4)3.7H2O) was added to the solution at 60°C and it was stirred for 24 h. Resulting suspensions of Fe-HA (iron humate) complexes were separated by centrifugation and rinsed with distilled water until removal of chlorine and sulfate ions. Finally the solid complexes were dried at 70°C. The content of Fe in the dry Fe-HA was analyzed by ICP-AES. The solubility of adsorbent was studied as a function of pH.
Table 1. Properties of humic acid TKI Humic Reference Acid Humic Acid* Content (Wt. %) Moisture Ash (at 950°C) Organic matter Inorganic matter Humic acid Hexane extraction residue Methanol extraction residue Petroleum ether extraction residue Elemental Analysis (wt. %) C H N S O Elemental ratio N/C H/C O/C Functional Groups (meq/g) COOH Phe-OH Total acidity E4/E6
11.8 38.2 56.6 43.4 31.3 0.12 3.63 0.018 55.4 6.4 1.4 0.8 36
54.72 4.04 1.47 0.36 38.54
0.022 1.38 0.49
0.023 0.88 0.53
2.16 6.84 9.0 3,26
5.0 1.16 6.16
* International Humic Substances- IHSS- (http://ihss.gatech.edu/ihss2/index.html) Pahokee Peat
2.3. Batch adsorption experiments Batch adsorption experiments were carried out by shaking 1 g of the adsorbent with 25 ml of solution of As and Hg with desired concentration in PE bottles. The metal solution together with the adsorbent was agitated magnetically in the closed PE bottle for different mixing times. Then the solid phase was separated by centrifugation, the concentration of the metal in the solution was determined immediately by ICP-AES. The adsorption capacity was calculated using the following equations:
q(mg / g ) =
(Ci − Ct ) V m
where Ci and Ct are the concentrations of the metal ion in initial and final solutions respectively. V is the volume of the aqueous phase (mL), m is the weight of Fe-HA (g). The adsorption degree, AD, as a function of time was also determined from the experimental data using the following relationship: AD % = 1 − C(t)/C(i) *100
3. Results and Discussion
3.1. Iron humate Humic acid containing 43.4 wt. % inorganic matters was used without any purification in the production of humate. The potential acidity of the iron salt added to the humate solution lowers the pH, and a precipitate of chiefly iron humate. The change in the pH value with the addition of FeCl3.6H2O caused different iron content in Fe-HA samples (Table 2). Lower pH values gave higher iron content. The increase in the HA concentration decreased iron content in the final product with the addition of FeSO4.7H2O. Fe-HA samples obtained with the addition of FeSO4.7H2O and Fe2(SO4)3.7H2O have about 11 wt. % iron content. It can be concluded that FeSO4.7H2O and Fe2(SO4)3.7H2O salts are more appropriate in the formation of complexes. However, iron humate obtained from FeSO4.7H2O showed higher adsorption capacity for As, Ni, Pb and Cd mixture than others. FTIR studies (Fig. 1) also exhibited the differences between FeCl3.6H2O (Fe-HA1) and FeSO4.7H2O (Fe-HA 2) compared to TKI humic acid. Thermogravimetric analysis of Fe-HA sample showed less weight loss less than 200°C compared to TKI humic acid (Fig. 2). Table 2. Iron humate production conditions
FeHA-1 FeHA-2 FeHA-3 FeHA-4 FeHA-5 FeHA-6
HA Conc. % 1 1 1 1 1 1
FeHa-7 FeHA-8 FeHA-9
5 5 1
Sample Code
Fe source FeCl3.6H2O FeSO4.7H2O FeCl3.6H2O FeCl3.6H2O FeSO4.7H2O K4P2O7 and FeSO4.7H2O FeSO4.7H2O* FeSO4.7H2O Fe2(SO4)3.7H2O
* Amount of FeSO4.7H2O was taken two fold.
pH value
Fe amount in Fe-HA, wt.%
2.3 6.42 6.42 3.4 4.5 7.05
4.91 11.08 2.29 4,61 9.74 9.87
6.42 6.42 2.4
5.31 4.72 11.89
The solubility of FeHA-2 and the calcined form of the same sample at pH:11 were determined as 1.2 and 0.4 wt. %, respectively. Because of the low solubility at high pH, calcined iron humate samples were used in the adsorption studies. SEM photograph of FeHA-2 sample is depicted in Fig. 3.
Fig. 1. FTIR spectra of Fe-HA samples and TKI humic acid
Fig. 2. Thermogravimetric analysis of Fe-HA sample and TKI humic acid
Element CK OK Al K Si K KK Ca K Ti K Fe K
Weight % 26.68 33.66 8.17 11.00 0.81 0.36 0.32 19.00
Totals
100.00
Atomic% 41.16 38.99 5.61 7.25 0.38 0.17 0.12 6.30
Fig. 3. SEM micrograph of Fe-HA sample
3.2. Adsorption studies 3.2.1. Effect of contact time The effect of contact time on the adsorption of As and Hg ions on Fe-HA was investigated over time intervals from 2 up to 48 hours. Fig. 4 shows the adsorption capacity as a function of contact time. As seen from Fig. 4, it reached to a maximum at about 2 hours for Hg and about 24 hours for As. About 90% of the total metal ion sorption was achieved within 2 and 24 hours for Hg and As, respectively (Fig. 5).
4,5
Adsorption capacity, q (mg/g)
4 3,5 3 2,5 2 1,5 1
As
0,5
Hg
0 0
10
20
30
40
50
60
Tim e (hour)
Fig 4. Variation of adsorption capacity with contact time
120 2h
Adsorption Degree (%)
100
6 h 10 h 16 h 20 h 24 h
24 h 48 h 10 h
80
16 h
20 h
6h 2h
60
40
20
0 As
Fig. 5. Variation of adsorption percentage y with contact time
Hg
3.2.2. Adsorption isotherms Several equilibrium models have been developed to describe adsorption isotherm relationships, the two main isotherm models used in this work are the Langmuir and Freundlich models. The Langmuir isotherm described the monolayer coverage of adsorbate over specific homogeneous sites within an adsorbent. Linear form of the Langmuir model could be described by the following equation:
1 ⎛ 1 ⎞ 1 1 ⎟⎟ + = ⎜⎜ q e ⎝ K L qm ⎠ Ce qm Where Ce was the equilibrium concentration (mg/L), qe was the amount of metal ion adsorbed at specified equilibrium (mg/g), qm and KL were the Langmuir constants related to adsorption capacity and adsorption energy. In this study, Langmuir isotherm was applied to analyze relationship between As and Hg concentration and adsorption capacity of iron humate. As shown in Fig. 6, experimental data of iron humate were well fitted for As and Hg by the Langmuir plots. The Langmuir model effectively described the sorption data with all R2 values >0.99 (Table 1). Freundlich isotherm suggested that sorption energy exponentially decreased on the completion of sorptional centers of an adsorbent and described heterogeneous systems. The Freundlich adsorption isotherm was tested in the following linearized form:
1 ln qe = ln K f + ln Ce n Where Ce was the equilibrium metal ion concentration in solution (mg/L), qe was the amount of metal ion adsorbed at specified equilibrium (mg/g). Kf and n were the Freundlich constants characteristics of the system, indicating the adsorption capacity and adsorption intensity, respectively. As can be seen in Fig. 7, the Freundlich model well fitted for As and Hg. Because of the correlation coefficient, Langmuir isotherm was more applicable for As. Both isotherms are applicable for Hg.
4,5 4 3,5
1/qe (g/mg)
3 2,5 2 y = 40,653x + 0,0538 R2 = 0,9996
1,5 1 0,5 0 0
0,02
0,04
0,06
0,08
0,1
0,12
1/Ce (L/m g)
a) 0,45 0,4 0,35
1/qe (g/mg)
0,3 0,25 0,2 0,15 0,1
y = 40,69x - 0,0011 R 2 = 0,9998
0,05 0 0
0,002
0,004
0,006
0,008
0,01
0,012
1/Ce (L/mg)
b) Fig. 6. Langmuir adsorption isotherm of arsenic (a) and mercury (Hg) on Fe-HA
2 1,5 1
Lnqe
0,5 0 -0,5 -1
y = 0,9102x - 3,4589 R 2 = 0,9867
-1,5 -2 0
1
2
3
4
5
6
LnCe
a) 2,5
2
Lnq e
1,5
1
0,5
y = 0,9987x - 3,6946 R2 = 0,9998
0 3
3,5
4
4,5
5
5,5
6
LnCe
b) Fig. 7. Freundlich adsorption isotherm of arsenic (a) and mercury (Hg) on Fe-HA
Table 3. Isotherm parameters of As and Hg ion adsorption on Fe-HA Adsorption model Langmuir isotherm qm KL R2 Freundlich isotherm Kf n R2
isotherm Arsenic
Mercury
adsorption 18.59 0.01156 0.9996
-909.09 0.000027 0.9998
31.78 1.0986 0.9867
0.0248 1.0013 0.9998
adsorption
4. Conclusion Different iron sources (iron (II) sulfate, iron (III) sulfate and iron (III) chloride) were used in the ion exchange reactions. Iron (II) sulfate (FeSO4) was found to be most suitable iron source in terms of iron content and adsorption capacity for As, Pb, Ni and Cd. The solubility of the iron humate sample at high pH values (pH:11) was obtained to be as low as 0.4 wt. % as a result of the calcination treatments applied after obtaining Fe-HA. Calcined Fe-HA sample was successfully used for selective As and Hg removal and adsorption capacities were obtained to be over 90 % for both Hg and As in a period of 2 and 24 hours, respectively.
References 1. F.J. Stevenson, Humus Chemistry. Genesis, Composition, Reactions (2nd Edition ed.), Wiley, New York (1994). 2. J. Novák, J. Kozler, P. Jano , J. Cezíková, V.Tokarová and L. Madronová, Humic acids from coals of the North-Bohemian coal field: I. Preparation and characterisation Reactive and Functional Polymers. 47 (2001) 101-109. 3. S. Erdogan, A. Baysal, O. Akba, C. Hamamci, Interaction of Metals with Humic Acid Isolated from Oxidized Coal, Polish J. of Environ. Stud. 16 (2007) 671675. 4. Humates and Humic Acids. How do they work?http://www.simplicitea.com/humates_how_they_work.doc 5. E. Tipping, Cation binding by humic substances, Cambridege environmental chemistry series -12, Cambridge University Pres, (2002). 6. P. Janos, J. Fedorovic; P. Stanková; S. Grötschelová; J. Rejnek; P. Stopka, Iron Humate as a Low-Cost Sorbent for Metal Ions, Environmental Technology, 27 (2006)169-181. 7. P. Janoš, Iron humate as a multifunctional sorbent for inorganic and organic pollutants, Environ. Chem. 2 (2005) 31-34. 8. P. Janoš, L. Herzogová, J. Rejnek, J. Hodslavská, Assessment of heavy metals leachability from metallo-organic sorbent—iron humate—with the aid of sequential extraction test, Talanta 62 (2004) 497-501. 9. P. Janos, Sorption of Basic Dyes onto Iron Humate Environ. Sci. Technol. 37 (2003) 5792-5798. 10. P. Janos, V. Šmídová, Effects of surfactants on the adsorptive removal of basic dyes from water using an organomineral sorbent—iron humate, Journal of Colloid and Interface Science 291 (2005) 19–27.
Heavy Metal Adsorption of Turkish Coal Based Humic Acid/Epoxy Composites Emel Yıldız1, Hacer Dogan1, Murat Koral1, Tulay Inan1, Selahattin Anaç2, Zeki Olgun2 1
TUBITAK Marmara Research Center, Chemistry Institute, 41470 Gebze, Kocaeli TURKEY 2 TKI(Turkish Coal Enterprises)Hipodrom Cad. No:12 Yenimahalle 06330 Ankara Abstract: The objective of this study was to investigate the adsorption capability of Humic acid/epoxy based composites. Humic acid produced by TKI (Turkish Coal Enterprises) was used as a co-curing agent for epoxy resin system based on Bisphenol F. The stoichiometrical amount of humic acid as co-curing agent, Diethylene triamine (DETA) as curing agent and Bisphenol F based epoxy resin were mixed. The homogenous mixtures were cured at 190 °C into the preheated molds. The curing agent/humic acid compositions were optimized to investigate the effect of humic acid concentration to removal of As, Pb, Ni and Cd. The Fourier Transform Infrared (FTIR) Spectroscopy, Thermal Gravimetric Analyzer (TGA), Differential Scanning Calorimeter (DSC) and Scanning Electron Microscopy (SEM) were used to characterize the humic acid/epoxy based composite samples. 1. Introduction Humic acids (HAs) are present in soils, natural waters, river, lake and sea sediments, peat, brown and brown-black coals and other natural materials as a product of chemical and biological transformations of animal and plant residues [1]. The principal properties of HAs and their subsequent potential applications depend strongly on their origin (source) as well as on the isolation procedure. HA is a chemically heterogeneous compound having different types of functional groups in different proportions and configurations. It contains carboxyl (-COOH), amine(-NH2), hydroxyl(-OH), and phenol(Ar-OH) functional groups [3]. Carboxyl and hydroxyl groups are capable of substituting their hydrogen atoms for ions of metals. When humic acids interact with multi valency metals, such as iron, zinc, copper and others they form new type of insoluble compounds [4,5]. Recent years, low cost and highly effective adsorbents have been developed for heavy metal adsorption. Some of the reported low-cost sorbents include bark/tanninrich materials, lignin, peat moss, iron-oxide-coated sand, leaf mould, seaweed/algae/alginate, dead biomass etc. Humic acid based materials are also shown as one of the best examples for low cost sorbents. The immobilization of HAs on a support is a promising approach in order to investigation of adsorption characteristics [6-9].
The aim of the study is to prepare the immobilized HAs and to investigate the adsorption capacity of Humic acid/epoxy based composites. 2. Materials and Methods 2.1 Materials Diglycidylether of Bisphenol F (DER 354, Dow Chemical Company) was used as epoxy resin. Diethylene triamine (DETA) (Dow Chemical Company) was used as curing agent without further purification. The chemical structures of epoxy resin and DETA are shown in Scheme 1. The epoxy equivalent weight of DER 354 is 167-174 and amine value of DETA is 1626 mgKOH/g. The calculated ratio of DETA to DER 354 is 11.84 parts curing agent per hundred parts epoxy resin (parts curing agent per 100 parts epoxy resin will be abbreviated as pph amine). Humic acid was supplied by TKI and was used as co-curing agent. 2.2. Humic acid characterization Humic acid sent by TKI was analyzed and the results are given in Table 1. Carboxyl groups, total acidity and the coefficient E4/E6 were then determined. The amount of carboxylic groups and the total acidity were determined using the calcium acetate and the barium hydroxide methods, respectively. The quantity E4/E6 is the ratio of absorbances at 465 and 665nm of humic acid solutions. Elemental analyses of carbon, hydrogen, nitrogen and sulfur were obtained with a Thermo Finnigan Flash 1112 Series EA elemental analysis instrument. E4/E6 ratio was found as 3,26 of humic acid obtained from coal. This value shows that humic acid has high aromatic groups. Humic acid was dried at 70ºC for 8 hours before used. 2.3. The preparation of humic acid/epoxy based composites The epoxy resin used in this work is DER 354 , a bifunctional epoxy resin formulated from the reaction of epichlorohydrin and Bisphenol F. For the composite sample preparation, the stoichiometric calculations were made on the basis of the amount of amine hydrogen present in DETA. The humic acid was added, in stoichiometric proportions (0-100 w/w %) relative to the total amount of amine hydrogen in DETA to the mixtures and all formulations were based on 100 parts of epoxy resin. Epoxy resin, curing agent (DETA) and humic acid were mixed homogenously at room temperature. The resulting homogenous mixture was poured into a mold. The curing cycle was 2 hours at room temperature (25°C) and then 190°C for 4 hours. The cured samples were grounded and then extracted by Soxhlet extraction method by using chloroform in order to remove the uncured epoxy resin. The crosslinked densities of humic acid/epoxy based composite samples were given in Table 2.
Table 1. Properties of humic acid
Content (Wt. %) Moisture Ash (at 950°C) Organic matter Inorganic matter Humic acid Hexane extraction residue Methanol extraction residue Petroleum ether extraction residue Elemental Analysis (wt. %) C H N S O Elemental ratio N/C H/C O/C Functional Groups (meq/g) COOH Phe-OH Total acidity E4/E6
TKI Humic Reference Acid Humic Acid* 11.8 38.2 56.6 43.4 31.3 0.12 3.63 0.018 55.4 6.4 1.4 0.8 36
54.72 4.04 1.47 0.36 38.54
0.022 1.38 0.49
0.023 0.88 0.53
2.16 6.84 9.0 3,26
5.0 1.16 6.16
* International Humic Substances- IHSS- (http://ihss.gatech.edu/ihss2/index.html) Pahokee Peat
O O
OH O
O
O O
F3 C F3 C
CF3 n (a)
CF3
H2N
N H
NH2
(b) Scheme 1 Chemical structures of a- DER 354 epoxy resin, b- DETA as curing agent Table 2. The crosslinked densities of humic acid/epoxy based composite samples Sample 100 DER 354/0 HA/100 DETA
Crosslinked density (% by w/w) 98.56
100 DER 354/20 HA/80 DETA
98.79
100 DER 354/40 HA/60 DETA
97.24
100 DER 354/60 HA/40 DETA
95.88
100 DER 354/80 HA/20 DETA
92.14
100 DER 354/100 HA/0 DETA
91.56
2.4. The Characterization of the Composite Samples
Heat of Reaction Measurements of Composite Samples The heat of curing reaction of the Humic acid/epoxy based composite sample was determined using a Perkin Elmer Jade Differential Scanning Calorimeter (DSC) operated in a nitrogen atmosphere according to ASTM E 2160. All the measurements were performed in alumina pans with sample weights ranging from 4 to 10 mg. The sample was heated along with an empty sample holder at a heating rate of 10°C/min from room temperature to 400°C with a nitrogen flow rate of 30 mL/min. The total heat of curing reaction was estimated by the area of an exothermic peak in temperature scanning mode calculated on Perkin Elmer Jade DSC Software.
Glass Transition Temperature Glass transition temperatures of composite samples were determined by using a Perkin Elmer Jade DSC according to ASTM D 3417. DSC scans were run at a heating rate of 10°C/min from room temperature to 200°C under nitrogen purge at a rate of 30 cm3/min. The glass transition temperatures (Tg) of composites were obtained from the second heating after a quick cooling. Tg value was taken from the second scan as midpoint of the change in slope of the baseline. The analysis of the resulting curve was performed on Perkin Elmer Pyris 1 DSC Software. Thermal Gravimetric Analysis (TGA) TGA of the humic acid/epoxy based composite samples was performed using a Perkim Elmer Pyris 1 TGA according to ASTM D3850. A small sample (8-20 mg) of composite sample was placed into a ceramic sample holder and loaded into the heating chamber. The heating chamber was kept at room temperature for approximately five minutes to purge the chamber of ambient air. The sample was heated at a rate of 10°C/min from ambient temperature to 900°C in N2 atmosphere. The test equipment measured and recorded the sample weight over the course of the analysis. 2.5. Batch adsorption experiments Batch adsorption experiments were carried out by shaking 1 g of the adsorbent with 25 ml of solution of As, Cd, Ni and Pb with desired concentration in PE bottles. The metal solution together with the adsorbent was agitated magnetically in the closed PE bottle for different mixing times. Then the solid phase was separated by centrifugation, the concentration of the metal in the solution was determined immediately by ICP-AES. 3. Results and Discussion In order to investigate the effect of humic acid on the thermal degradation behaviour and glass transition temperatures of humic acid/epoxy networks, the parent and composite networks were subjected to TGA and DSC, respectively. Thermal decomposition behaviour of the Humic acid, neat epoxy resin and composite samples determined by TGA under N2 and are reported in Table 3. The thermal decomposition temperatures of composite samples are higher than neat epoxy network and Humic acid. As espected, the decomposition temperatures at 50% weight loss, the weight losses at 500ºC and residue values at 900ºC of composite samples increased as the increasing amount of Humic acid as co-curing agent. TGA thermograms are given in Figure 1.
Table 3. Thermal properties of humic acid/epoxy based composite samples. Sample Humic acid
5% Weight 50% Weight Residue at loss (°C) weight loss loss at 900°C (%) (°C) 500°C (%) 184.8 539.7 56.0 41.8
100 DER 354/0 HA/100 DETA
205.3
392.8
20.3
14.8
100 DER 354/20 HA/80 DETA
327.6
402.9
26.9
20.4
100 DER 354/40 HA/60 DETA
321.7
409.1
32.0
25.0
100 DER 354/60 HA/40 DETA
323.1
427.2
41.0
33.0
100 DER 354/80 HA/20 DETA
325.3
459.5
47.2
38.3
100 DER 354/100 HA/0 DETA
329.7
680.4
56.9
46.0
Figure 1. TGA thermograms of Humic acid, Neat epoxy resin and composite samples
The effect of humic acid concentration on the glass transition temperatures of composite samples were investigated using dynamic DSC experiments. All the samples have one glass transition temperature. The DSC results are presented in Table 4 and DSC thermograms are given in Figure 2. Table 4. The Glass Transition Temperatures of composite samples Sample
Tg (°C)
100 DER 354/0 HA/100 DETA
100.7
100 DER 354/20 HA/80 DETA
93.9
100 DER 354/40 HA/60 DETA
92.2
100 DER 354/60 HA/40 DETA
77.4
100 DER 354/80 HA/20 DETA
69.6
100 DER 354/100 HA/0 DETA
79.1
Figure 2. The glass transition temperature curves of neat epoxy resin and composite samples
As seen in Table 4 and Figure 2, the glass transition temperature values of composite samples decreased as the increasing amount of Humic acid. It is well known that the glass transition temperature of a polymer is closely correlated by the rigidity of the polymer backbone. Therefore, the glass transition temperature of neat resin cured by DETA only is higher than composite samples including Humic acid as co-curing agent. The dynamic DSC scans at 10 ºC/min heating rate for the curing reaction of neat epoxy resin and composite samples under N2 atmosphere are shown in Figure 3. From DSC thermograms, the onset temperature of curing (T 0), the exothermic peak maximum (Tmax), the final temperature of curing (Tf), and the heat of curing reaction are determined. As seen in Figure 3, there are two curing peak maximum, the first peak area between 50-150 ºC decreased as the decreasing amount of DETA, but the second peak area between 250-350 ºC increased as the decreasing amount of DETA. The structural characterizations of neat epoxy resin cured by DETA only and composite samples including Humic acid used as co-curing agent were performed by FTIR spectroscopy technique. The FTIR spectrum of epoxy resin cured by DETA and by Humic acid are given in Figure 4. The strong absorptive peak at 3435 cm-1 is due mainly to the stretching vibration of –OH group and there is no stretching absorption peak belongs to –NH group at 3304 cm-1 in FTIR spectrum of neat epoxy resin cured by DETA only. FTIR spectrum of epoxy resin cured by Humic acid shows a considerably smaller peak of epoxy ring at 910 cm-1 but there is no peak belongs to oxirane group in the FTIR spectrum of epoxy resin cured by DETA. It is verified by crosslinked density values of cured samples. And also there is no peak belongs to carbonyl group between 1725-1700 cm-1 in both of FTIR spectrum.
Figure 3. DSC thermograms of Neat epoxy resin and composite samples
100.0
Cured by DETA
95 90 85 80 75 70 65 60 55 %T 50 45 40 35 30
Cured by HA
25 20 15 10 3.2 4000.0
3600
3200
2800
2400
2000
1800 cm-1
1600
1400
1200
1000
800
600
450.0
Figure 4. FTIR spectra of epoxy resins cured by DETA and Humic acid
The effect of Humic acid content of composite samples on the removal of As, Cd, Ni and Pb ions are given at Figure 5. The uptake capacity for As, Cd, Ni and Pb ions depended on the Humic acid content of composite samples. As seen in Figure 5, the Humic acid/Bis F epoxy based composites are not effective sorbent for As removal. But they are effective for Pb removal and adsorption power depends on directly to the Humic acid content of composite samples.
Heavy metal content (ppm)
14 12 10
As
8
Cd Ni
6
Pb
4 2 0 0
20
40
60
80
100
Humic acid content (% w/w )
Figure 5. Removal of As, Cd, Ni, Pb by Humic acid/Bis F epoxy based composite 4. Conclusion Humic acid/epoxy based composites were used as sorbent in order to As, Cd, Ni and Pb removal. Humic acid produced by TKI (Turkish Coal Enterprises) was used as a co-curing agent for epoxy resin system based on Bisphenol F. DETA/humic acid compositions were optimized to investigate the effect of humic acid concentration on the adsorption capability for As, Pb, Ni and Cd. The uptake capacity for As, Cd, Ni and Pb ions depended on the Humic acid content of composite samples. Humic acid/Bis F epoxy based composites are effective for Pb removal and adsorption power depends on the Humic acid content of composite samples.
References 1. F.J. Stevenson, Humus Chemistry. Genesis, Composition, Reactions (2nd Edition ed.), Wiley, New York (1994). 2. J. Novák, J. Kozler, P. Jano , J. Cezíková, V.Tokarová and L. Madronová, Humic acids from coals of the North-Bohemian coal field: I. Preparation and characterisation Reactive and Functional Polymers. 47 (2001) 101-109. 3. S. Erdogan, A. Baysal, O. Akba, C. Hamamci, Interaction of Metals with Humic Acid Isolated from Oxidized Coal, Polish J. of Environ. Stud. 16 (2007) 671675. 4. Humates and Humic Acids. How do they work?http://www.simplicitea.com/humates_how_they_work.doc 5. E. Tipping, Cation binding by humic substances, Cambridege environmental chemistry series -12, Cambridge University Pres, (2002). 6. M. Klavins, L. Eglite, Immobilization of Humic Substances, Colloids and Surfaces A: Physicochemical and Engineering Aspects 203 (2002) 47-54 7. X. Wang, X. Shan, L. Luo, S. Zhang, B. Wen, Sorption of 2, 4, 6Trichlorophenol in Model Humic Acid-Clay Systems, Journal of Agricultural and Food Chemistry, 2005, 53, 3548-3555. 8. M. Klavins, L. Eglite, A. Zicmanis, Immobilized Humic Substances as Sorbents, Chemosphere 62 (2006) 1500-1506. 9. A. H. Rosa, A. A. Vicente, J. C. Rocha, H. C. Trevisan, A new Application of Humic Substances: Activation of Supports for Invertase Immobilization, Fresenius J. Anal. Chem. (2000) 368: 730-733.
Manuscript Not AVAILABLE
NUMERICAL SIMULATION OF SYNGAS PRODUCTION BY PARTIAL OXIDATION OF COKE OVEN GAS UNDER NON-PREMIXED CONDITION
Honggang Chen, Prof . ,National Engineering Laboratory for Biomass Power Generation Equipment, North China Electric Power University, Beijing 102206, China Email: [email protected] Kai Zhang, Prof. , National Engineering Laboratory for Biomass Power Generation Equipment, North China Electric Power University, Beijing 102206, China Email: [email protected] Hui Zhao Dr. ,State Key Laboratory of Heavy Oil Processing China University of Petroleum, Dongying 257061, China E-mail:[email protected] Yongfa Zhang, Prof. Key Laboratory of Coal Science and Technology of Shanxi Province and Ministry of Education Taiyuan University of Technology,Taiyuan 030024,China Email: [email protected]
Abstract: China is the largest coke producer in the world. The yield of coke oven gas (COG), a byproduct of the coking process, reached 1190 billion Nm3 per year in 2007. COG is a good feedstock for many chemical processes e.g. FT synthesis, methanol synthesis and ammonia synthesis, instead of only as a heating fuel. The main component of COG is hydrogen (~56-60 vol. %) but there are also other compounds such as methane (~25-30 vol. %), carbon monoxide, carbon dioxide, and nitrogen. There are several methods to convert the methane into hydrogen in order to fully make use of the COG resources to the larger extent. The non-catalytic partial oxidation of COG with oxygen is one of important routes for COG comprehensive utilization. In the present work non-premixed combustion of coke oven gas under low oxygen condition was investigated through computational fluid dynamic (CFD) simulation coupled with radiation heat transfer. The mathematical model was formulated to describe the fluid flow, heat transfer, mass transfer , and gas phase chemical reactions. The obtained model was numerical solved with finite volume method using commercial software FLUENT TM. Temperature profile and product distribution in the reactor were obtained from simulation. Calculated temperature profile indicates that oxygen react strongly with coke oven gas in region near the oxygen nozzle and the temperature in this region increase sharply. Significant influence of reactor wall temperature on product distribution could be found in simulation results. Higher selectivity of syngas could be archived at higher reactor wall temperature, while conversion rate of methane in coke oven gas was suppressed simultaneously. Reactor wall temperature should be treated as a key parameter and optimized well in this process. The simulation results are helpful the development and design of the COG non-catalytic partial oxidation reactor and process.
Modeling of Coal Drying in a Pneumatic Dryer Kyoungsoo Lim, Sangdo Kim, Soonkwan Jeong, Youngjun Rhim, Sihyun Lee*,
Clean Fossil Energy Research Center, Korea Institute of Energy Research Daejeon 305-343 South Korea *E-mal: [email protected]
ABSTRACT Operation of the pneumatic conveying system for drying coal was influenced by many parameters, in particular, gas velocity, gas temperature, coal residence time, and coal particle size. In this study, the pneumatic conveying drying system was simulated for drying wet coal. Numerical studies were conducted to examine the effect of these parameters on drying coal. As the gas temperature, the drying rate of wet coal was increased. In addition, coal drying increased with an increase of coal residence time. Particle size is one of the most important parameter on moisture removal efficiency of the pneumatic conveying dryer. The moisture removal increased in short drying time with a decrease in particle size due to the large surface area. KEY WORDS: Drying; coal; hot gas, moisture; residence time; gas temperature; particle size.
INTRODUCTION Pneumatic conveying drying is a combination of heat and mass transfer between hot gas and wet particles. This technique has been widely used in various industry processes, due to the reasons of maximum efficiency, minimum gas flows and power consumption, improved product quality and increased workplace safety. In pneumatic dryer system, a wet particle is rapidly dried as it comes into contact with hot gas, and the atomized feed sample travels concurrently with the drying-gas. Operation of the pneumatic conveying system for drying wet particles was influenced by many parameters, in particular, gas velocity, gas temperature, particle residence time, and particle size. Many experimental and numerical studies of a pneumatic conveying dryer have focused on an effect of these parameters on drying rate of wet particles. As for experimental approach, Yamazawa et al. (1972) presented an improvement of rice physical properties using a pneumatic conveying dryer. Tandano et al (1981) investigated pneumatic conveying dryer operation and clarified the drying time experimentally. Baeyens et al. (1995) presented a design procedure to predict moisture and temperature profiles in a pneumatic dryer. Rocha and Paixao (1996) developed a pseudo two-dimensional model for pneumatic drying and predicted axial and radial profile for a gas solid velocity and solids moisture content. Fyhr and Rasmuson (1997) presented a complex model for a pneumatic dryer considering a distribution of particle sizes for steam drying of wool chips. Levy and Borde (1999) studies the drying of wet PVC particle in a large-scale pneumatic dryer and compared the results with prediction of a numerical simulation taking into account particle shrinkage. More recently, many models for pneumatic conveying dryer have been developed and used to predict the behavior of a batch (Pelegrina and Crapiste, 2001; Narimatsu et al., 2004; Skuratovsky et al., 2004; Tanaka et al., 2006; Narimatsu et al., 2007). Pelegrian and Crapiste (2001) developed the model to analyze the effect of different variables as solid to air mass flowrates ratio, velocity and temperature of air and, size and shape of potato particles. Tanaka et al. (2008) investigated the influence of initial rice
particle size and moisture content distributions on the moisture propagation in a pneumatic conveying dryer. However, most previous studies had focused on drying the food particles. Although a few studies developed the models to analyze the coal drying in the pneumatic conveying dryer, the variables were limited. Levy et al. (1998) simulated a dispersed gas-coal solids flow, without mass or heat transfer in one-dimensional transport system. Ross et al. (2005) compared the experimental data for the final moisture content and gas temperature profile for a pressurized entrained flash dried at 800 oC and 10 bar to a one-dimensional mathematical model. This study primarily seeks to present the drying characteristics of wet coal particles in a pneumatic conveying dryer by exploring the effects of several variables numerically, such as coal particle diameter, gas temperature and dyer length, and so on. Numerical solutions are derived employing energy, momentum and mass balance using FLUENT program.
NUMERICAL METHOD The geometry and dimensions of calculation domain to study the influence of several important parameters on coal drying are shown in Fig. 1 and Table 1. There are three parts construing the calculation domain i.e., the vertical upstream pipe, U-bend and cyclone. The length of the vertical upstream pipe is 5.0 m for all variations. The radius of U-bend is 0.3 m and the horizontal pipe between the vertical upstream pipe and cyclone is 0.7 m. The diameter of cyclone body is 0.24 m and total height of cyclone is 0.8 m. The computational grids of the pipe and cyclone are approximately 61,500 cells. The grid was generated using Gambit, and the commercial FLUENT software was used for solving the set of governing equation. FLUENT provides flexibility in choosing discretization schemes for each governing equation. The discretized equation, along with the initial condition and boundary conditions, were solved using the segregated solution method to obtain a numerical solution. Using the segregated solver, the conservation of mass and momentum were solved sequentially and a pressure correction equation was used to ensure the conservation of momentum and the conservation of mass. The RNG k- model for turbulence model was employed to predict flow behavior in the pipes and cyclone, and wet combustion model was also employed for predict the evaporation and drying from wet coal particles.
Fiig. 1. Geomettry of the callculation dom main.
Table 1. Dimensions of the pneum matic dryer. Param meters Length of the verticcal upstream m pipe, Lv Diametter of the vvertical upsstream pipe, Dp Radius of U-bbend, Ru Inner R Length of the horizoontal pipe, Lhh D Diametter of the cyclone body, Dc Length of cyclone ccylinder, h1 Length of cyclone ccone, h2 Diametter of cyclonee exit, dc
Diimensions (mm) 5000 40 300 700 240 300 500 40
In the prresent investiigation, the steady state fl flow of wet soolid particless and a gas thhrough the vertical gassolid sysstem was described. The ssimulation w was conductedd consideringg the followinng assumptioons; 1) Gas and a solid phaase flow co-ccurrent in thhe upward diirection. 2) M Mass, momenntum, and heeat transfer occur onnly between tthe phases annd not within the particless. 3) Theree is no radiantt energy trannsfer betweenn the gas- sollid suspensionn and the surrroundings. 4) The soolid feed is uuniform in sizze and shape. The solid particles are sspherical.
5) The gas phase is considered as a mixture of gas and liquid vapor
RESULTS AND DISCUSSION The simulation of pneumatic conveying dryer was carried out at different drying conditions. Solid feeding was coal particles of 5.0 kg/hr at temperature Tcoal = 25oC and water content of coal X0 = 0.274 kgwater/kg-coal in most cases. The gas absolute humidity (Y0) was 0.007 kg-water/ kg-dry gas and gas velocity was 20 m/s. Effect of gas temperature, particle size and initial moisture content is discussed in this section. Gas temperature and absolute humidity profiles along the dryer for a typical drying simulation are shown in Figs. 2 and 3. Usually, after the entrance region, the velocity of coal particles is accelerated rapidly in the up-stream gas flow according to the momentum balance equation, and then reached almost constant velocity. During the residence of these coal particles in dryer, there are heat and mass exchange between particles and hot gas. The temperature of coal particles rises faster from its initial value up to the weltbulb temperature of the drying gas, while gas temperature decreases as shown in Fig. 3. As the initial gas temperature increases, the gas temperature drop is large and absolute gas humidity increases. It means that the drying rate of coal particles increases as the gas temperature increases because the water content of coal particles decreases and gas moisture content increases at the same rate. 200 oC 300 oC 400 oC 500 oC 600 oC 700 oC
Gas Temperature (oC)
800
600
400
200 Initial absolute humidity : 0.007 kg/kgdry air Initial coal particle diameter : 500 m Initial coal moisture content : 0.274 kg/kg coal 0
1
2
3
4
5
Dryer length (m)
Fig. 2. Gas temperature profiles at different initial gas temperature.
Absolute Humidity (kg/kgdry air)
0.05 200 oC 300 oC 400 oC 500 oC 600 oC o 700 C
0.04
Initial absolute humidity : 0.007 kg/kgdry air Inital coal particle diameter : 500 m Initial coal moisture content : 0.274 kg/kg coal
0.03
0.02
0.01 0
1
2
3
4
5
Dryer length (m)
Fig. 3. Absolute humidity profiles at different initial gas temperature. Fig. 4 shows the final moisture content of the dried coal product and moisture removal efficiency by evaporation. The moisture removal efficiency of coal particles in 500 m size is about 40% at gas temperature of 400 oC, and more than 90 % at 900 oC. The evaporation rate continuously increases and the moisture removal efficiency also linearly increases as gas temperature increases because of the higher driving force for heat transfer. However, the maximum temperature that the dried product can reach without having thermal damage limits initial gas temperature.
100 Initial moisture content : 0.35 kg water/kg coal Coal particle diameter : 500 m Dryer length : 5 m
0.25
80
0.20 60 Moisture content Moisture removal
0.15
40 0.10
Moisture Removal (%)
Moisture Content (kg water/kg coal)
0.30
20
0.05
0.00
0 100
200
300
400
500
600
700
800
o
Gas Temperature ( C)
Fig. 4. Moisture content and moisture removal efficiency at different initial gas temperature. The size of particles conveyed using a pneumatic system varies greatly, so it is important to study the influence of particle diameter on heat transfer behavior. The predicted change of gas temperature and absolute temperature at different particle size as function of dryer length is shown in Figs. 5 and 6. Simulation were carried out for particles in 100, 300, 500, 1000 m size As particle size decreases, the decline rate of gas temperature with dryer length is higher, with the corresponding absolute humidity increasing. The particles of 300, 500, 1000 m size continuously dry as they travel in dryer column. However, the particle of 100 m size dry up within 1.5 m in dryer length. In this simulation the initial moisture content of coal particles was set as constant value of 0.273 kg-water/kg-coal. Fig. 7 shows the final moisture content and moisture removal efficiency at different particle sizes. As particle size decreases, the moisture removal efficiency dramatically increases from 20% to 100%. An increase in particle diameter is the most dominant variable in the inter-phase momentum exchange coefficient. The
decrease in the inter-phase momentum exchange coefficient will result in an increase in slip velocity between gas an solid phases. Smaller particles have a higher surface area by unit weight of solid than larger particles, favouring the rate of heat and mass transfer between solid and gas. 500
Initial absolute humidity : 0.007 kg/kgdry air Inital gas temperature : 400 oC Initial coal moisture content : 0.274 kg/kg coal
Gas Temperature (oC)
450
100 m 300 m 500 m 1000 m
400
350
300
250
200 0
1
2
3
4
5
Dryer length (m)
Fig. 5. Gas temperature profiles at different particle sizes. 0.040
Initial absolute humidity : 0.007 kg/kgdry air o
Inital gas temperature : 400 C Initial coal moisture content : 0.274 kg/kg coal
Absolute Humidity (kg/kg dry air)
0.035
100 m 300 m 500 m 1000 m
0.030
0.025
0.020
0.015
0.010
0
1
2
3
4
5
Dryer length (m)
25
100
20
80
15
60 Moisture content Moisture removal
10
40
5
Moisture Removal (%)
Moisture Content (kg water/kg coal)
Fig. 6. Absolute humidity profiles at different particle sizes.
20 Initial moisture content : 0.274 kg water/kg coal Initial gas temperature : 400 oC Dryer length : 5 m
0 0
200
400
600
800
1000
0 1200
Particle Diameter (m)
Fig. 7. Moisture content and moisture removal efficiency at different particle sizes.
This result is consistent with observances made for other studies. Saravanan et al. (2007) carried out the simulation with two resin particle size to verify the effect of particle size on drying rate. They also described that the outlet solid moisture content increased with an increase in the solid particle size because of large particle surface area per unit mass of solid. Ross et al. (2005) compared the experimental data for the final moisture content and gas temperature profile for pressurized entrained flash drier at 800oC and 10 bar to mathematical model developed previously for the brown coal mill process. They conducted the simulation with five Yallourn lignite particle size. They presented that the small particles dried almost instantaneously (<150 m), while larger particles underwent marginal drying. Effect of initial moisture content on the final moisture content and moisture removal efficiency was simulated for the same gas and particle conditions, as presented in Fig. 8. For a given condition, the moisture removal efficiency decreases with an increase in initial moisture content. The moisture removal efficiency is about 90% at 0.15 kg-water/kg-coal of initial moisture content while that is about 10% at 0.62 kg-water/kg-coal. The initial drying rate increases with moisture content because of the higher water activity and the lower particle velocity caused by a higher density. However, the drying rate at the final stages of drying decrease significantly with feed moisture content due to the reduction in driving forces for heat and mass transfer.
100 Initial absolute humidity : 0.007 kg/kgdry gas Initial gas temperaute : 673 K Initial coal particle diameter : 500 m
0.6
80 0.5 60
0.4 Moisture content Moisture removal
0.3
40
0.2
Moisture Removal (%)
Final Moisture Content (kg water/kg coal)
0.7
20 0.1
0.0
0 0.1
0.2
0.3
0.4
0.5
0.6
0.7
Initial Moisture Content (kg water/kg coal)
Fig. 8. Moisture content and moisture removal efficiency at different initial moisture contents. As shown in Fig. 1 and Table 1, the dryer system consists of 5.0 m vertical dryer column, 0.3 m U-bend dryer column, 0.7 m horizontal dryer column, and cyclone with 0.24 m diameter and 0.8 m height. The results shown in Fig. 2 ~ Fig. 8 are the drying characteristics of coal particles only in 5.0 m vertical dryer column. Fig. 9 shows the moisture content and moisture removal efficiency of the dryer system from the vertical dryer column to the cyclone. The moisture removal efficiency of coal particles in 1,000 m size is about 35% in the vertical dryer column, at 673 K of initial gas temperature and 20 m/s of gas velocity. On the same condition, that is about 65 % in dryer system containing the cyclone. An increase on moisture removal efficiency results from longer residence time of coal particles in U-bend, horizontal dryer column and the cyclone. In particular, there is a swirling flow in a cyclone, which leads to longer residence time. Thus, the cyclone also plays an important role in the drying characteristics of pneumatic conveying dryer. In this simulation, it was assumed that all the coal particles moving the cyclone wall by centrifugal force were collected in the cone part of the cyclone and there is no more drying process between particles and gas. Usually, particles introduced into the cyclone are collected on the cyclone wall and go down the hopper through cone bottom by gravitation. There is sometimes re-entrainment of collected particles. For
this reason, it is expected that the particle residence time is longer and drying efficiency is higher in experimental conditions. Finally, simulation result shows that it is very important to design and set up the cyclone in the pneumatic conveying dryer.
110 100
0.20 90 80
Moisture content (Without cyclone) Moisture removal (Without cyclone) Moisture removal (With cyclone) Moisture content (With cyclone)
0.15
0.10
70
Initial absolute humidity : 0.007 kg/kgdry gas Initial gas temperaute : 673 K Initial coal moisture content : 0.274 kg/kg coal
60 50
Moisture Removal (%)
Final Moisture Content (kg water/kg coal)
0.25
0.05 40 0.00 0
200
400
600
800
1000
30 1200
Coal Particle Diameter (m)
Fig. 9. Moisture content and moisture removal efficiency without cyclone and with cyclone.
CONCLUSIONS A numerical investigation using FLUENT for pneumatic conveying dryer was performed to study the effects of several parameters; initial gas temperature, particle diameter, and initial moisture content, in incompressible and steady-state conditions. The predictions were used to illustrate the complex transport phenomena that take place during the process and the coupling effects among gas and solid conditions. The gas temperature decreased and the solid temperature increased along the axial direction of the dryer column due to the heat from the gas phase to the solid phase. The decline rate of gas temperature is higher as the initial gas temperature increases. The particle diameter is very important parameters on drying characteristics of pneumatic conveying dryer. An increase in particle diameter results in the inter-phase momentum exchange coefficient since the particle diameter is most dominant variables in the inter-phase momentum exchange coefficient. The decrease in the inter-phase momentum exchange coefficient will result in an increase in slip velocity between gas and solid phases. In addition, smaller particles have a higher value of the heat transfer coefficient because of a larger surface area by unit weight of solid than larger particles. The simulation results shows that the moisture removal efficiency dramatically increases from 20% to 100% with an decrease in particle size from 1,000 m to 100 m. The moisture removal efficiency decreases with an increase in initial moisture content of particle because of reduction of the driving forces for heat and mass transfer. The moisture removal efficiency is about 90% at 0.15 kg-water/kg-coal of initial moisture content while that is about 10% at 0.62 kg-water/kg-coal. Compared with the moisture removal efficiency in a vertical dry column without the cyclone, that with cyclone increases from 35 % to 62 % in particle size of 1,000 m. An increase on moisture removal efficiency results from longer residence time of coal particles in the cyclone with a swirling flow.
REFERENCES Baeyens, J, Van Gauwbergen, and Vinckier, I (1995). "Pneumatic drying: the Use of Large-scale Experimental Data in a Design Procedure," Powder Technol, Vol 83, No 2, pp 139-148. Fyhr, C, and Rasmuson, A (1997). "Mathematical Model of a Pneumatic Conveying Dryer," AIChE J, Vol 43, No 11, pp 2889-2902. Levy, A and Borde, I (1999). "Steady State One-dimensional Flow Model for a Pneumatic Dryer," Chem Eng Proc, Vol 38, pp 121-130. Levy, A and Mason, DJ (1998). "The Effect of a Bend on the Particle Cross-section Concentration and Segregation in Pneumatic Conveying System," Powder Technol, Vol 98, No 2, pp 195-103. Narimatsu, CP, Ferreira, MC, and Freire, JT (2004). "Drying of Porous Alumina Particles in a Vertical Pneumatic Dryer," Drying 2004, Proc the 14th International Drying Symp, São Paulo, Vol A, pp 549-556. Narimatsu, CP, Ferreira, MC, and Freire, JT (2007). "Drying of a Vertical Pneumatic Conveyor," Drying Technol, Vol 25, pp 291-302. Pelegrina, AH, and Crapiste, GH (2001). "Modelling the Pneumatic Drying of Food Particles," J Food Eng, Vol 48, pp 301-310. Rocha, SC, and Paixao, AE (1996). "Pseudo Two-dimensional Model for a Pneumatic Dryer," Drying ’96, pp 340-348. Ross, D, Doguparthy, S, Huynh, D, and McIntosh, M (2005). "Pressurised Flash Drying of Yallourn Lignite," Fuel, Vol 84, pp 47-52. Saravanan, B, Balasubramanian, N, and Srinivasakannan, C (2007). "Drying Kinetics in a Vertical GasSolid System," Chem Eng Technol, Vol 30, No 2, pp 176-183. Skuratovsky, I, Levy, A and Borde, I (2004). "Pneumatic Drying of Solid Particles," Drying 2004, Proc the 14th International Drying Symp, São Paulo, Vol A, pp 366-373. Tadano, T, Yokoya, K, Hirokawa, R (1981). "Studies on Pneumatic Drying of Okara," Bull Coll Agr Vet Med, Vol 38, pp 149-154. Tanaka, F, Morita, K, and Uchino, T (2006). "Mathematical Modeling of Pneumatic Drying of Food Particle," Proc IUFOST 13th World Cong Food Sci Technol, Nantes, pp 513-514. Tanaka, F, Uchino, T, Hamanaka, D, and Atungulu GG (2008). "Mathematical Modeling of Pneumatic Drying of Rice Powder,"J Food Eng, Vol 88, pp 492-498. Yamazawa, S, Yosizaki, S, Maekawa, Harasawa, T, and Harazawa, K (1972). "On Improvement of Unhulled Rice in Physical Properties by Pneumatic Conveying Drying," J Jpn Soc Agr, Vol 33, No 4, pp 418-421.
MODEL STRUCTURE OF HIGH SULPHUR N.E. REGION INDIAN COALS ATMA RAM SINGH and SUNIL KUMAR SRIVASTAVA Central Institute of Mining & Fuel Research PO- FRI, Dhanbad-828108, Jharkhand INDIA ABSTRACT As per Model structure of coal proposed by Shri B. K. Mazumdar, number 9 and 10 positions are available in phenanthrene unit, which is the mother nucleus of coal ( lignite to semi anthracite). In this model structure, one unit of coal is linked with other unit by certain linkages such as ether and methylene groups. Oxygen in coal has been distributed in the form of different functional groups such as –OH, -COOH, >C=O and ether. The functional groups –OH and –COOH are available at the site of phenanthrene nucleus either at 9 position or at any other place. Position 10 is available for linkage with other unit of coal. >C=O group is present in a cyclic ring. Assuming this model structure of coal to be correct, oxidation of such a coal should lead to carboxylic acid formation up to a maximum extent of 5-6% because oxidation of methyl group and carbonyl group are expected to yield of –COOH groups. The actual experimental results obtained on oxidation of the same coal by 3N dilute nitric acid yielded 18% of –COOH groups. Hence this model cannot explain this experimental fact. If one assumes that besides these two positions, oxidation occurs at several other sites also then more solubility of coal in alkali / organic solvents on oxidation is expected to occur but this would lead to the formation of low molecular weight products on oxidation, which is not actually obtained. Hence a modification is needed in this coal structure so that the new model structure of coal can explain these experimental results also. Thus a modified model structure of coal has been proposed. Mr. Mazumdar has not spelt out anything about model structure of high sulphur coals of North Eastern Region. With the help of the data generated by Dr. Srivastava’s group on disposition pattern of sulphur in NE region coal structure, a model structure of high sulphur NE coal has been proposed.
INTRODUCTION A number of model structures of coal have been proposed word wide. Based on specific reactions yielding known products, some group of workers considered aromatic cluster as the nucleus of the coal.
Yet another group of researchers found out the presence of aliphatic and hydro-
aromatic structures associated with the aromatic cluster of coal. Since coal is not a exactly a polymer but constitute a number of macromolecule attached to each other by certain linkage, the same was also studied by various researchers. Keeping all the above researches in mind, it was concluded that coal is a condensed poly aromatic mass associated with aliphatic & hydroaromatic groups and these units are linked with ether and methylene bridges. In eighties, Indian scientist Mr. B.K. Mazumdar proposed the coal systematics from Lignite to Anthracite covering all the physical and chemical properties of different ranks of coal. It is well known that the elementary carbon and hydrogen are distributed in coal into three basic parts viz. aromatic, aliphatic and hydroaromatic. Similarly oxygen is distributed in the form of functional groups such as hydroxyl, carboxyl, carbonyl and ether. The other elementary part nitrogen exists in his coal model as aromatic, but sulphur have not been covered by the author in his proposed coal model. It has been found recently [1] that this model is not able to explain some of the experimental results and hence there is an urgent need to modify this model structure of coal. The present authors have taken an attempt to fill these gaps.
EXPERIMENTAL Oxidation of Samla Coal by 3N Nitric Acid : Proximate and Ultimate analyses of raw Samla coal and the demineralised Samla coal were done following IS methods. The methods for determination of functional groups such as –OH, --COOH and -CH3 have been described in an earlier publication made from this laboratory [1]. Oxidation of coal was done as follows. Samla coal was taken in a flat bottom flask and to this was added 3N Ntric Acid as an oxidizing agent for carrying out oxidation reaction at desired temperature for desired time. After completion of the reaction, the oxy coal was washed with distilled water and filtered. The residual mass i.e. oxy coal was extracted using polar organic solvent. After solvent removal, the remaining solid was taken out followed by its characterization. The above data are summarized in table-1.
Disposition Pattern of Sulphur in N.E.Region Coals : N.E. Region coal and resorcinol (hydrogen-donating compound) mixture was heated from room temperature at the rate of 10°C / min in the presence of hydrogen gas in a reactor up to 620°C. The H2S evolved at different temperature ranges were absorbed in cadmium acetate solution to estimate sulphur gravimetrically for each temperature range[2]. The coal particle size used for this purpose was 212 µ. The total sulphur in the coal was determined using a Leco Sulphur Determinator (Model SC-132 from Leco Corporation, USA). The pyritic sulphur was estimated [3] by refluxing coal with HNO3 and determining the iron content while the sulphate sulphur was determined [4] by refluxing coal with HCI and estimating the sulphur as barium sulphate.
RESULTS AND DISCUSSION
Table – 1 Summarizes the proximate, elemental and functional groups analyses of raw and demineralised Samla coal as well as of oxy coal and soluble part of oxy coal in polar organic solvent. On careful examination of the above table, it is observed that the ash content has been drastically reduced from 14.6% to 1% and the moisture content of demineralised coal is reduced to 1.7% from 8.6%. Except these, practically no change in Carbon, Hydrogen and nitrogen percentages (dmmf basis) have occurred and hence in functional groups also. It is observed from Table-1 that after oxidation of samla coal, the carboxylic acid functional groups have enhanced from 1.2% to 18.3%. This order of enhancement in carboxylic acid functional groups at the periphery of aromatic rings / have come mainly from 9-10 position [1,5] in angle ring aromatic nucleus like phenanthrene. Methylene and methyl groups present in this coal is only about 3-4% and the reduction of methyl group from this level to 1% level alone cannot explain [1] the generation of carboxylic acid functional groups of the order of 18.3%. The model structure of samla coal (80-83 %C) as proposed by Mazumdar [6] is as follows and shown in (fig-1).
On careful examination of the above model structure of coal, it is crystal clear that 9,10 position of each phenanthrene unit is blocked either by functional groups or by other linkages. It is known that the mechanism of oxidation and hydrolysis of phenanthrene in which 9,10 position are available , gives a product generating carboxylic and hydroxyl groups as follows [7,8].
In some cases, it is known that oxidation facilitates at benzylic and 9,10 position of
phenanthrene units yielding a product containing more number of carboxylic acid functional groups as follows [9].
In the present case, as also appeared in an earlier publication [10], it was observed that samla coal was easily oxidized by oxidizing agents. In order to explain the experimental result that – COOH group has enhanced from 1.2% to 18.8%, it is concluded that 9,10 position of phenanthrene of each unit is not blocked. It is either vacant in coal structure or present in the form
of 9,10 dihydrophenanthrene. Many of the model structure proposed such as that of Lander’s [11] and Gibson [12], are ‘’two dimensional” although the presence of folded ring such as 9-10 dihydrophenanthrene means that the structure are not flat as shown in fig-2. Pitt [13] has proposed models of coal structure based on 9,10- dihydrophenanthrene structures (fig-3) and that these molecules exist in a tangled state in coal. Heredy and Wender (fig-4) proposed model structure of coal molecules (83-84%C) indicating availability of 9,10 dihydrophenanthrene and phenanthrene in coal units which are linked with each other by ether and methylene bridges. The angled aromatic rings like phenanthrene is the mother nucleus of coal. Based on its rank, association of different functional groups in different proportions with other units of coal by methylene and by ether linkages [14 -16] are reported in literature. It is also known that this linkage is less reactive towards oxidation hence it cannot explain the above result. By virtue of vast literature available on oxidation of condensed aromatic ring, it is obvious that the 9-10 position of phenanthrene is more reactive towards oxidation rather than other sites [14,15,17,20]. Thus, during oxidation preferably dissociation of this bond takes place. In a previous work [1], it was observed that not only the carboxylic acid functional groups are generated on oxidation of coal but subsequently the nitro groups are also substituted in the periphery of aromatic structure on coal matrix. Thus, creation of carboxylic acid functional groups in the coal matrix takes place first and then nitration occurs, so that nitro groups substitute at meta position with respect to carboxylic acid functional groups which is meta-directing as shown in Reaction- 1. The quantity of OH functional group on oxidation of demineralised coal to oxy-coal [18,19] remained practically the same (Table-1).. The above data clearly indicates the availability of 9,10 position of phenanthrene type of aromatic ring in coal which are responsible for generating large amount of carboxylic acid functional groups.
NO2 COOH
HNO3
-------(1)
OXIDATION
COOH
NO2
Table-2 Summarizes the proximate, ultimate, functional groups analysis of Tipong coal
and
distribution of carbon in different form.. Based on the experimental results the hydroxyl group recorded from 3--4.7% which is within the range. It is the trend the percent hydroxyl group gradually decreases with increasing the rank and caking behavior developed. In case of high sulphur coal the percent of oxygenated function groups replaced by sulphur in some extent shown in the unit-1 and unit-2. It is also observed the total sulphur present in coal exist in different form shown in above model structure proposed by two units. The factor of aromatic and hydroaromatic also reflect the behavior of coal within rang of caking even the sum of hydroxyl and thiol cover the same value as compare to original coal of this range. Proximate, Ultimate & functional groups analyses and distribution of carbon in different forms of Tipong coal are summarized in Table 2. It is observed from functional group analysis of several N.E.region coals that hydroxyl group ranged between 3 – 4.7 % which is within the range for caking variety of coal. It is well known that percent hydroxyl group gradually decreases with increase in rank and develop caking properties. In case of high sulphur N.E.region abnormal coal [21-24], it is known that part of the oxygen in coal is replaced by sulphur. It has already been established [2] that the organic sulphur in high sulphur N.E.region coals exist in five forms viz. disulphide, thiol, thioether, thioketone and thiophene. It is evident from table2 that aromaticity, hydroaromaticity,H / C aromatic and Non aromatic values are in the range of caking variety coals. Also the sum of –OH and –SH groups is equal to that of –OH group present in normal coals of same carbon rank. Temperature programmed reduction (TPR) studies on high sulphur coals of N.E.region of India [ ] to determine the sulphur functional groups have been reported by Attar and Dupuis [ ] and Majchrowicz et al. [ ] using l - 75 mg of coal in each experiment with continuous heating at a certain rate with gravimetric analysis of sulphur evolved as H2S and absorbed to yield precipitate. Table 3 presents the data on TPR studies on Tipong Coal. Our approach was - the C-S bond is broken and then the sulphur is removed as H2S. If sulphur is present in coal in more than one
form depending upon the nature of sulphur functionality / bond strength, the weakest bond will break first while the strongest at the last. The quantity of H2S evolved at five different temperature ranges indicate that the sulphur present in coal is not in one form rather it is in five forms. The evolution of H2S at different temperature ranges depends on the nature of C-S bond. The studies were done on Tipong coal [2] under different reaction conditions. Based on the experimental work it was concluded that the organic sulphur present in this coal are in five different forms viz. disulphide, thiol, thioketone, thioether and thiophene. It was also concluded that the Sulphur is interchangeable with oxygen in N.E. region coal structures at all places. The following were some of the findings of the above work. 1. During reduction, mercaptan SH will react with Hydrogen to form H2S which will be evolved in the temperature range 190-220°C. The disulphide will yield mercaptan on hydropyrolysis which will further react with Hydrogen to yield H2S in the same temperature range. If the disulphide is converted into sulphide and H2S, then the aliphatic sulphide will be transformed into complex thiophenic sulphur below the temperature at which H2S is evolved’. 2. The thiol group reacts with Hydrogen to form H2S in the temperature range 260- 2900C. 3. The thioether linkage or the aromatic sulphide reacts with Hydrogen to give H2S, directly or via -SH, which will be evolved in the temperature range 360-390°C. 4. Reaction of simple thiophene with Hydrogen will break the ring giving H2S in the temperature range 460-490°C. 5. During continuous pyrolysis, any C=S groups will cyclize in the presence of aliphatic chains which will lead to the formation of complex/ condensed thiophenic rings, which are difficult to decompose but small amounts of which may be removed as H2S above 500°C. 6. It is well known that the reaction: FeS2 + H2 = FeS + H2S is complete at 530°C and the reaction FeS + H2 = Fe + H2S starts well above 900°C. These two reactions are believed to be responsible the peak of H2S evolutions at 530°C and the shoulder at 620°C. On careful examination of Table 2 (Upper Part), it is evident that the molecular formula / molecular weight and the ultimate analysis data based on experimental results, when compared with the
corresponding data based on proposed model structure are in excellent agreement. Thus we can assume that the proposed model structure of high sulphur NE region Tipong coal is correct. But to prove this, other supporting evidences are a must. The proposed model structure of NE region of coal must explain the formation of the products on various coal conversion processes. It must explain the existence of aromatic, aliphatic, hydroaromatic and hetero-atoms present in NE region coals. The gaseous products obtained on pyrolysis of coals from aliphatic groups are [25] Carbon di oxide, carbon monoxide, methane, hydrogen, hydrogen sulphide, ammonia, other hydrocarbons etc. can be very well be explained by the proposed model structure of NE region coal supporting our model. The heteroatoms present in coal are oxygen, sulphur and nitrogen. Nitrogen appears to be evenly distributed and has no significant effect on coal reactivity [25], but oxygen functional groups play a very important role in coal conversion. Some of the heteroatoms containing units identified [17] are fluorenone, Anthraquinone, Benzanthraquinone and/or Naphthacenequinone, Dibenzofuran,
Benzonaphathofuran,
Xanthone,
Benzoxanthone,
Dibenzo-p-dioxin,
Benzothrophene, Dibenzothiophene, Pyridine, Quinoline and/or Isoquinolene, Carbazole, Acridone. These products were obtained on heating coal [17] at 2500C in 0.4 m sodium dichromate, and the aromatic acids produced were isolated. The proposed model structure of coal on its oxidation can easily explain the above formations. The products of pyrolysis also include aromatics like Benzene, Pyrene, Benzopyrene, naphthalene, phenanthrene etc. production of which can also be explained from our proposed model structure of coal. Thus these validate our proposed model structure. Similarly during solvent refining of coal, the hydroaromatics obtained [25]
are
Decahydropyrenes,
Hexahydropyrenes,
Dihydropyrenes,
Pyrenes,
Octahydrophenanthrenes, Tetrahydrophenanthrenes, Phenanthrenes, Tetralins, Napthalenes, Tetrahydroacenaphthenes, Acenaphthenes, Our proposed model structure of coal can very well explain their formations. Thus it can be concluded that the proposed model structure of high sulphur NE region coal is very much valid and can explain all the Physico-Chemical properties of coal.
REFERENCES
1. D.P. Choudhury, S.S.Choudhury, Raja Sen, J.Mukherjee, G.Ghosh and S.K.Srivastava, Energy and Fuels, 21,1006, 2007. 2. A.Kumar and S.K.Srivastava, Fuel, 71, 718, 1992. 3. IS : 6444 : 1979 4. IS : 1350 Part III 1969 5. G.Ghosh, A.Banerjee, B.K.Mazumdar, Fuel, 54, 294, 1975. 6. Atma Ram Singh, D.P.Choudhury, G.Ghosh and S.K.Srivastava, The Open Fuels and Energy Science Journal, 3, 8, 2010. 7. A.R.Singh, D.P.Choudhury and S.K.Srivastava, Composite Polymeric Materials from Coal, Unpublished work. 8. D.P.Choudhury, Raja Sen, G. Ghosh and S.K.Srivastava, Direct Sourcing of Coal: Solubilization of coal from N.E.region of India, Presented at 23 rd Annual International Pittsburg Coal Conference, Pittsburg, USA, 25-28 September, 2006 and its references. 9. P.H Given, Fuel, 39, 147, 1960. 10. R. Hayatsu, R.E.Winans, R.G Scott, L.P.Moore, M.H. Studier, Oxidation degradation study of coal and solvent refined coal, ACS Symp. Ser.,71, 108,1978. 11. J.N.Bhowmik, P.N.Mukherjee, A.Lahiri, Fuel, 38. 21, 1959. 12. L.A.Heredy,I.Wender, Flammability , Ignition and Electrostatic Properties, ACS Div. Fuel Chem. Preprints, 25, 1980.
13. D.W. Vankrevelen, Fuel, 29, 269, 1950. 14. W.R.Lander and C.E. Snape, Fuel, 57, 658, 1978. 15. J. Gibson, J. Inst. Fuel, 51, 61, 1978. 16. G.J. Pitt, G.R.Millward, in Coal and Modern Coal Processing; an introduction, structural analysis of coal, Academic Press, London, UK, 27, 1979. 17. R.Hayatsu, R.G.Scott, L.P.Moore, M.H.Studier, Nature, 257, 378, 1975. 18. R.Hayatsu, R.G.Scott, L.P.Moore, M.H.Studier, Nature, 77, 261, 1976. 19. R.Hayatsu, R.E.Winans, R.G.Scott, L.P.Moore, M.H.Studier, Nature, 275, 116, 1978. 20. R.Hayatsu,R.G.Scott, L.P.Moore, M.H.Studier, Nature, 57, 541, 1978. 21. A.Ali, S.K.Srivastava, R.Haque, Fuel, 71, 835, 1992. 22. S.K.Srivastava, A.Kumar, R.Haque, Fuel Sci. Tech., 11, 111, 1992. 23. S.K.Srivastava, A. Kumar , R.Haque, Proceedings of 5th Australian Coal Science Conference, Melbourne, Australia, 30 November-2 December,1992, p. 344. 24. A.Kumar, S.K.Srivastava, K.K.Singh, Fuel Sci. Tech., 13, 71, 1994. 25. Coal Science and Technology 2 : Fundamentals of Coal Beneficiation and Utilization, Edited by S.C.Tsai, Elsevier Scientific Publishing Company, Printed in The Netherlands, 1982.
Table – 1 Characterization of Raw and Demineralised Samla coal, Oxy coal & soluble part of oxy coal in polar organic solvent Analyses Moisture % *
8.6
Demineralised coal 1.7
Ash%*
14.6
1.0
6.0
0.0
Carbon % (dmf)
80.6
80.5
62.3
59.7
Hydrogen % (dmf)
4.5
4.6
2.9
4.5
Nitrogen % (dmf)
2.1
2.0
4.4
4.4
OOH % (dmf)
4.8
5.2
5.0
2.0
OCOOH % (dmf)
0.96
1.2
18.3
12.6
CCH3 % (dmf)
3.1
2.4
1.0
0.8
* (as received basis)
Raw coal
oxy coal 9.1
Organic soluble oxy coal 7.9
Table-2 Analysis of Tipong Coal on as received basis _________________________________________________________________________ | | | Experimental data | Data based on | | Proposed Structure ------------------------------------------------------------------------------------------------------------------------Proximate( %) Moisture 2.8 Ash 5.3 Molecular formula C87H77O7N1.5S2.5 Volatile matter 42.0 Fixed Carbon 49.9 Ultimate Analysis (%) (dmf) Carbon 71.19 78.3 78.3 Hydrogen 5.3 5.8 5.8 Nitrogen 1.22 1.4 1.6 Sulphur 5.7 6.3 6.0 Oxygen(By diff.) 8.49 8.3 8.4 _________________________________________________________________________ _______________________________________________________________________ Unit-1
Unit-2
AV
BKM
fa 0.62 0.64 0.63 0.67 fhar 0.29 0.27 0.28 0.22 fcCH3 0.045 0.034 0.039 0.047 Har/C 0.21 0.21 0. 21 0.22 H/C(non aromatic)` 1.7 1.8 1.75 1.66 %OOH 3.5 4.7 4.1 5.8 % SH 0.05 0.02 0.035 --_________________________________________________________________________
Table-3 Sulphur evolved from Tipong coal during TPR studies _________________________________________________________________________ Temperature Sulphur evolved Sulphur functionality 0 Range ( C) (%) assigned _________________________________________________________________________ 1 190—220 0.36 Mercaptain/disulphide 2 260—290 1.7 Thiol 3 360--390 0.51 Aromatic sulphide 4 460--490 0.88 Simple thiophene 5 510—540 0.35 Pyritic and part of complex thiophene 6 590--620 0.24 Part of complex thiophene _________________________________________________________________________
CH3
CH3
co
CH3
co o
0
CH2
OH
CH3
CH2
N
N H
H
COOH
0 HO
OH
OH
OH
OH
CH3
co
Fig-1
Unit-1
CH3
CO
CH3
CH3
O
CS
CH
O
CH3
OH
OH
SH
N H
O OH
Fig-2
Unit-2 CH3
CH2
CH3
COOH
CH3
O
N H
OH
CS
SH
OH
Fig-3
OH
CH2
N H
S OH
Influence of the Surface Treatment with O3 and NH3 on the Physical and Chemical Characteristics of Dried Low Rank Coal Gi Bo Han, Yongseung Yun, Changsik Choi*
Plant Engineering Center, Institute for Advanced Engineering, 633-2 Goan-ri, Baekam-myon, Cheoin-gu, Yongin-si, Gyeonggi-do Gyeonggi-do, Republic of Korea
Abstract In this study, the dried coal with the surface treatment was characterized to investigate the effect of the surface treatment on the chemical and physical properties of dried coal. The surface treatment was conducted by the dry method with 5 vol% NH3 at 200 oC and 15 g/m3 O3 at room temperature. As compared with the fresh dried coal, it was found that the dried coal obtained after the surface treatment was changed in aspect of chemical and physical properties such as the increase in the content of carbon, hydrogen and fixed carbon. From the change of the chemical and physical properties, the various effects were obtained as follows: 1) Removal of oxygen-contained functional group, 2) Elevation of calorific value, 3) Inhibition of spontaneous ignition potential, 4) Suppression of H2O adsorption of dried coal. 1. Introduction After 1980’s, the green house effect was caused with depletion of high rank coal and the consumption amount of coal was diminished in the whole world. In 2005, coal industry was dropped; however, it was in charge of electricity of 41% and the energy consumption of 28% all over the world. Recently, gas price is raised up to higher than $100/bbl, the consumption amount and price of coal steeply rising. In energy crisis of the 1970’s, the coal technologies such as an integrated gasification combined cycle (IGCC) and coal liquefaction process have been developed to utilize and convert coal to clean energy. Coal conversion technology such as IGCC system, coal liquefaction and coal-water mixture production process is an issue as a one of the coal technologies but, has not yet been commercialized due to huge expenses for development [1-13]. Recently, the concerns of environments have been increased to coal industry due to the air pollutant emission with coal utilization. The green house gas emission especially increased global warming. The development of the various clean coal technologies to reduce green house gas emissions and energy consumption is necessary for the sustainability of world economy. Upgrading of low rank coals is more focused due to the increase in energy consumption and demand. Coal is a heterogeneous natural polymer material which is composed of moisture, volatile matter, ash and fixed carbon. According to the degree of carbonization, the coal rank is classified into a lignite (or brown coal), (Sub-) bituminous coal. Lignite is not
coking, quickly burned and it is mainly used as a resource for the electric power generation. On the other hand, lignite coal contains the low content of sulfur and ash and can be converted into an alternative energy of petroleum with the development of technology for the elevation of calorific value and the inhibition of CO2 generation. To upgrade the coal rank of lignite, it is necessary to elevate the calorific value, inhibit the spontaneous ignition and drop the CO 2 generation. The various low rank coals including lignite, brown coal and sub-bituminous coal have generally characteristics of high moisture content which plays an important role in combustion, gasification and liquefaction. Drying process of low rank coals is very important for combustion technologies and boilers using low rank coals. The high content of moisture of low rank coal causes many problems in coal processes including storage and transportation. The high moisture content in low rank coals causes high energy consumption and emission of CO2. Spontaneous ignition associates with moisture adsorption. Upgrading process of low rank coal for the pretreatment of coal which results in the decrease in the water content, the increase in the calorific value and the prevention of self-heating and spontaneous combustion during transportation and storage has been presented in recent. In this paper, the surface treatment of dried coal by NH3 and O3 was conducted for upgrading the coal rank of lignite. Also, the chemical and physical properties of coal were investigated for the characterization of coal such as composition, calorific value and ignition temperature. 2. Experimental 2-1. Coal Sample and Drying Process The raw coal was an Indonesian lignite coal having high moisture content as shown in Table 1 and its particle size range was less than 3 mm. To reduce the H2O content of raw coal with high moisture content, the raw coal was dried at 150 oC under the air atmosphere. 2-2. Surface Treatment of Dried Coal The fixed bed reactor was horizontal-typed quartz tube with OD of 1/2 inch and length of 30 cm. The dried coal with the appropriate amount was packed and fixed in the reactor. The gas for the treatment was flowed into the reactor and then the flow rate was 100 ml/min. The surface treatment conditions were listed in Table 1. The treatment temperatures of the dried coal with O3 and NH3 were room temperature and 200 oC, respectively and then their concentrations were 15 g/m3 and 5 vol%, respectively. 2-3. Characterization of Coal with the Surface Treatment The composition of carbon, hydrogen, nitrogen, oxygen and Sulfur was analyzed by elemental analysis. Also, the content of fixed carbon, moisture, volatiles and ash was found by the proximate analysis. The relative comparison of oxygen-contained functional group like
carbonyl and carboxyl group was conducted by FT-IR spectra. Calorific value was measured with surface treatment. Under N2 atmosphere, thermal gravimetric analysis (TGA) for the raw and dried coals was conducted with raising the temperature at the ramping rate of 5 oC/min from room temperature to 800 oC. Ignition temperature of raw and dried coal before and after the surface treatment was measured by raising the temperature from room temperature to 900 oC at the ramping rate of 5 oC/min under air atmosphere after the coal sample was packed in ignition temperature tester (Krupp Flash Point Tester, Model: KRS-RG-9000). H2O adsorption for raw and dried coals before and after the surface treatment was conducted as follows: H2O content of coal sample absorbing H2O was measured by proximate analysis after the vapor of water of 200 ml boiled in the round flask was passed through the packed coal sample of about 2 g for 4 h. 3. Results and discussion 3-1. Drying Process for Dewatering Raw Coal Figure 1 shows the thermal gravimetric analysis result of raw and dried brown coals to investigate the temperature for the removal of water of raw coal. O2. The amount of raw and dried of coal was 8 mg and it was packed into bowl and then the temperature increased from room temperature to 800 oC at the ramping rate of 5 oC/min under the atmosphere of N2 of 100 ml/min. As compared with Figure 1(b), it was known that the water of raw coal was removed under 120 oC in Figure 1(a). Therefore, the temperature for the water removal of raw coal was fixed at 110 oC. Figure 2 shows the proximate analysis of raw and dried coals. In Figure 2, the moisture, volatiles, ash and fixed carbon were contained in the raw coal and then their contents were 34.4, 32.2, 1.8 and 31.6%, respectively. However, in the coal dried at 107 oC, the moisture of 1.4%, volatiles of 48.4%, ash of 2.7% and fixed carbon of 47.5% were contained and then it was known that the water content decreased and the contents of other components increased because the water was removed from the coal through drying process as compared to the raw coal. 60
0.008
100
Weight loss (%) dw/dt 80
Raw coal Dried coal
50
(b) Dried coal
48.4
47.5
0.004 40
Content (%)
60
dw/dt
Weight loss (%)
0.006
40 34.4
32.2
31.6
30
20
0.002 20
10 1.8
1.4 0
0.000 100
200
300
400
500
Temperature (oC)
600
700
800
2.7
0
Moisture
Volatiles
Ash
Component
Fixed carbon
Figure 1. Thermal gravimetric analysis of raw Figure 2. Proximate analysis of the raw and dried coals.
and dried coals.
3-2. FT-IR Spectra of Dried Coal with Surface Treatment To investigate the characteristics of dried coal Figure 3 shows the FT-IR spectra of the surface-treated coals with O3 and NH3 after drying at 107 oC. As shown in Figure 3, it was known that the coal dried at 150 oC under air atmosphere have oxygen-contained functional groups such as carbonyl and carboxyl croups corresponding to peaks observed in the range of 1200-1700 cm-1 and hydroxyl group corresponding to peak observed in the range of 2800-3000 cm-1. The various analysis results and molecular models of coal structures such as have been presented in the previous report and then it was presented that low rank coals such as brown coal have relatively high content of oxygen-contained functional groups like carbonyl and carboxyl groups as compared to high rank coals; therefore, their energy efficiency such as calorific vale were low [14]. In general, it was well-known that high oxygen content results in the low calorific value according to the various theoretical equations for the correlation between elemental analysis result and calorific value. The calorific value of the coal of low oxygen content was higher than that of high oxygen content because the oxygen content decreased and the carbon and hydrogen content relatively increased. As shown in Figure 3, in case of the dried coal treated with O3, the peak intensity of 1200-1700 cm-1 increased in comparison with the fresh dried coal and it might be due to the generation of carbonyl and carboxyl groups with partial oxidation of carbon structure surface by O3. On the other side, in case of dried coal treated by NH3, the peak intensity of carbonyl and carboxyl groups decreased because the carbonyl and carboxyl having acidic properties contained in dried coal might be decomposed by NH3 having basic properties.
In addition, the peak intensity of hydroxyl group corresponding -1
to the range of 2800-3000 cm decreased after the treatment with NH3 but increased after the treatment with O3 through the surface reaction between either NH3 or O3 and the hydrogen matrix contained in dried coal. It was known that oxygen-contained function group plays an role of precursor for the CO2 generation in combustion [15]. Therefore, the removal of oxygencontained functional group with the surface treatment of dried coal by NH3 leads to the inhibition of CO2 generation.
1.2
46.89 Fixed carbon
Transmittance (%)
80
Content (%)
B
A
C
4000
3500
0.96 Fuel ratio
49.33 Fixed carbon
1.03 Fuel ratio
60 2.9
2.8 Ash
2.89 Ash
Ash
1.0
44.83 Fixed carbon 0.85 Fuel ratio
0.6
40 48.39 Volatile
48.78 Volatile
1.82 Mositure
0.02 Mositure
0.4
52.37 Volatile
20
3000
2500
2000
1500
1000
0.8
0
0.2
A
-1
2.8 Mositure
B
Fuel ratio (FC(%)/VM(%))
100
0.0
C
Wavelength (cm )
Figure 3. FT-IR spectra of dried coal with the Figure 4. Proximate analysis of dried coals surface treatment (A: Fresh dried coal, B: with the surface treatment (A: Fresh dried NH3, C: O3).
coal, B: NH3, C: O3, VM: Volatile matter, FC: Fixed carbon).
3-4. Composition of Dried Coal with Surface Treatment Figure 4 shows the proximate analysis result of surface-treated coal after dried process. It was known that the rank of coal is related to the proximate analysis and fuel ratio (Fuel ratio = Fixed carbon (FC, %)/Volatile matter (VM, %)) and the rank of coal increases with increasing the fuel ratio [14]. In this section, the effect of the surface treatment with NH 3 and O3 on the proximate and fuel ratio for the dried coals was described. The fixed carbon of 49.33%, ash of 2.89%, volatile of 48.78% and moisture of 0.02% were contained in the dried coal treated by NH 3 and then the fuel ratio was about 1.03. In comparison to the fresh dried coal, both the content of the fixed carbon and fuel ratio increased because the oxygen-contained functional group carbonyl and carboxyl group might be decomposed by the surface reaction with NH3. On the other hands, the dried coal treated by O3 was composed of 44.83% fixed carbon, 2.8% ash, 52.37% volatile and 2.8% moisture and then its fuel ratio was about 0.85. The surface treatment with NH3 raised Fuel ratio but that with O3 did not. In a viewpoint of fuel ratio which is related to the energy resource efficiency, the rank of coal was improved by the surface treatment with NH 3 as compared with the fresh dried coal. 3-6. Ignition Temperature of Dried Coal with Surface Treatment Figure 5 shows the ignition temperature of the dried coal before and after the surface treatment by NH3 and O3. In this section, the effect of the surface treatment on the ignition temperature was investigated to compare the simultaneous ignition potential according to surface treatment method, relatively. The ignition temperature of raw coal was about 284 oC and that of dried coal was about 312 oC. This result was caused by decrease in the moisture content.
Therefore, it was estimated that the spontaneous ignition potential was dropped slightly. In the previous report, it was described that, in fact, it is not yet well understood how the moisture affects the tendency of a coal to selfheating and what is the mechanism of adsorption and desorption. In general, it has been found that a coal reacts with oxygen more rapidly when the coal is wet than when it is dry [16-17]. The ignition temperature goes up to 363 oC after the surface treatment with NH3; however, the ignition temperature was dropped to 293 in case of the dried coal after the surface treatment by O3. It was thought that this result comes from the variation of the content of volatile matter. It was previously referred that the content of volatile matter has influenced on the spontaneous ignition temperature and the higher volatile matter content, the higher spontaneous ignition potential [18]. Therefore, as compared with the dried coal, the ignition temperature increased from 312 to 363 oC in case of dried coal treated by NH3 and decreased from 312 to 293 oC in case of dried coal treated by O3. This result will contribute toward the safety coal stockpile and transportation. 400
Ignition temperature (oC)
363
300
312
293
284
200
100
Weight (g) & H2O content (wt%)
40 H2O content 35.0 wt%
30
20 H2O content 14.3 wt%
H2O content 11.7 wt%
10 W2 W1 2.2901 g ¥ÄW 2.0113 g 0.2788 g
W2 W1 2.2414 g ¥ÄW 2.0101 g 0.2313 g
W1 2.0108 g
W2 2.9607 g
¥ÄW 0.9499 g
0
0
A
B
C
A
D
B
C
Figure 5. Ignition temperature of dried coals Figure 6. Variation of weight and H2O content with the surface treatment (A: Raw coal, B: of the dried coal treated by NH3 and O3 with Fresh dried coal, C: NH3, D: O3).
H2O adsorption (A: Fresh dried coal, B: NH3, C: O3).
3-7. H2O Adsorption of Dried Coal with Surface Treatment Figure 6 shows the relative weight variation of the dried coal treated by NH3 and O3 with H2O adsorption to investigate the relationship between H2O adsorption and surface treatment. After adsorbing H2O, the weight of dried coal increased from 2.0113 to 2.2901 g with the H2O content of 14.3 wt%. In case of the dried coal treated by NH3, the weight increased from 2.0101 to 2.2414 g with H2O content of 11.7 wt%. In case of the dried coal treated by O3, the weight increased from 2.0108 to 2.9607 g with H2O content of 35 wt%. It was thought that these results are related to the removal and generation of oxygen-contained functional group such as carbonyl and carboxyl groups. In general, it was well known that carbonyl and carboxyl groups
are hydrophilic. As described in the previous section involving FT-IR spectra, the peak intensity corresponding to the oxygen-contained functional group decreased with the surface treatment by NH3 but increased with the surface treatment by O3. In case of the surface treatment by NH3, and then H2O adsorption capacity decreased with the removal of the oxygen-contained functional group. On the other hand, the H2O adsorption capacity increased with the generation of the oxygen-contained functional group. Therefore, it was expected that the surface treatment of dried coal by NH3 to prevent the H2O adsorption contributes toward the improvement of the coal stockpile and transportation. 4. Conclusion In this study, the influence of the surface treatment with NH3 and O3 on dried coal was investigated by characterizing the chemical and properties of the coal such as functional group, composition, H2O adsorption capacity, calorific value and ignition temperature. In case of the surface treatment with O3, as compared with the fresh dried coal, the content of carbon, hydrogen and fixed carbon decreased and that of oxygen and volatile matter increased with the treatment by O3. And then the H2O adsorption capacity relatively increased and the ignition temperature and the calorific value decreased with the variation of composition. As compared with fresh dried coal, the dried coal treated by NH3 has the low content of oxygen and volatile matter and the high content of the carbon, oxygen and fixed carbon. With this composition variation, the ignition temperature and calorific value increased and the H2O adsorption capacity decreased. Therefore, the surface treatment with NH3 after the drying process of raw lignite coal having the high content of moisture is the effective method for the safety storage and transportation for the long periods. Reference 1. Sakaguchi, M., Laursen, K., Nakagawa, H., Miura, K.: “Hydrothermal Upgrading of Loy Yang Brown Coal: Effect of Upgrading Conditions on the Characteristics of the Products,” Fuel Processing Technology, 89, 391 (2008). 2. Mahidin, Ogaki, Y., Usui, H., Okuma, O.: “The Advantages of Vacuum-Treatment in the Thermal Upgrading of Low-Rank Coals on the Improvement of Dewatering and Devolatilization,” Fuel Processing Technology, 84, 147 (2003). 3. Sato, Y., Kushiyama, S., Tatsumoto, K., Yamaguchi, H.: “Upgrading of Low Rank Coal with Solvent,” Fuel Processing Technology, 85, 1551 (2004).
4. Elliott, D.C.: “Decarboxylation as a Means of Upgrading the Heating Value of Low-Rank Coals,” Fuel, 59, 805 (1980). 5. Umar, D.F., Usui, H., Daulay, B.: “Change of Combustion Characteristics of Indonesian Low Rank Coal due to Upgraded Brown Coal Process,” Fuel Processing Technology, 87, 1007 (2006). 6. Friesen, W.I., Ogunsola, O.I.: “Mercury Porosimetry of Upgraded Western Canadian Coals,” Fuel, 74, 604 (1995). 7. Ruyter, H.P., Raam, L., Poel, H.: “Upgrading of Wakefield Browncoal from South Australia,” Fuel Processing Technology, 9, 163 (1984). 8. Ogunsola, O.I.: “Thermal Upgrading Effect on Oxygen Distribution in Lignite. Fuel Processing Technology,” 34, 73 (1993). 9. Ogunsola, O.I., Mikula, R.J.: “Effect of Thermal Upgrading on Spontaneous Combustion Characteristics of Western Canadian Low Rank Coals,” Fuel, 71, 3 (1992). 10. Petela, R., Petela, G.: “Indices for Coal Desulfurization and De-Ashing Processes,” Fuel, 75, 1259 (1996). 11. Friesen, W.I., Ogunsola, O.I.: “Principal Component Analysis of Upgraded Western Canadian Coals,” Fuel Processing Technology, 38, 139 (1994). 12. Lockhart, N.C.: “Dry Beneficiation of Coal,” Powder Technology, 40, 17 (1984). 13. Nandi, B.N., Macphee, J.A., Ciavaglia, L.A., Chornet, E., Arsenault, R.: “The Role of Mesophase in Upgrading Inert-Rich Oxidized Coal for Combustion,” Carbon, 20, 148 (1982). 14. Kabe, T., Ishihara, A., Qian, E.W., Sutrisna, I.P., Kabe, Y.: “Coal and Coal-Related Compounds,” Studies in Surface Science and Catalysis, 150, 85 (2004). 15. Sato, Y., Kushiyama, S., Tatsumoto, K., Yamaguchi, H.: “Upgrading of Low Rank Coal with Solvent,” Fuel Processing Technology, 85, 1551 (2004).
16. Küçük, A., Kadıoğlu, Y., Gülaboğlu, M.Ş.: “A Study of Spontaneous Combustion Characteristics of a Turkish Lignite: Particle Size, Moisture of Coal, Humidity of Air,” Combustion and Flame, 133, 255 (2003). 17. Bhattacharyya, K.K., Hodges, D.J., Hinsley, F.B.: The Mining Engineer. 126, 274 (1968). 18. Hesketh, R.P.; DAVIDSON, J. F. The Effect of Volatiles on the Combustion of Char in a Fluidised Bed. Chemrcal Engineering Science, 46, 12, 3101-3113, (1991).
Manuscript Not AVAILABLE
Energy, Natural Gas, Türkiye & Ankara İbrahim Halil KIRŞAN Başkent Doğalgaz Dağıtım A.Ş. [email protected] ABSTRACT Energy, Natural Gas, Türkiye & Ankara Natural gas; environment friendly, economic, confortable, safe as long as it is used properly, clean, harmless to nature, is an “environmentalist” energy source. Is a combustible gas formed in the lower layers of the earth and primarily consisting of methane and ethan hydrocarbons. Has been formed by decaying organic substances millions of years ago. It is a primary energy source ie: it can be used immediately after drilling out. Being lighter than air it can readily effuse into the atmosphere. Natural gas; when burned does not produce environmentally harmfull wastes such as sulphurdioxide and carbon particles. Natural gas; is a serious insurance to nature, environment hence the future of mankind. As a fossil energy source similar to petrol, however does not leave behind ash or slag when burnt and has no need for storage during use. Total natural gas reserves of the world is 178,7 trillion cubic meters and 3 Trillion cubic meters natural gaz is consumed annually throughout the world. May be all natural gas reserves will last 50 or 60 years. Our country in 2006, 19.6 billion cubic meters of natural gas consumed imported from Russia. This corresponds to 63% of total consumption. The amount of LNG imported from Nigeria and Algeria, 5.3 billion cubic meters. Our total reserves our calculated to be less than our annual consumption amount. Türkiye relies on foriegn countries for the supply of natural gas. Majority from Russia and Iran(through pipes), remainder from Algeria and Nigeria (as LNG). Başkentgaz is the second largest company with 2 billion cubic meters consumption per year and oldest distribution company in Turkiye. Keywords: Energy, Natural Gas, Reserves, Başkentgaz.
1 Characterization of Chars Made of Solvent Extracted Coals Hokyung Choia, Jiho Yooa, Sangdo Kima, Jeongwhan Lima, Thiruppatha Rajaa, Wantaek Job, Sihyun Lee a* a
Clean Fossil Energy Research Center, Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejeon, 305-343, Korea b Department of Chemical Engineering, Yonsei University, Seodaemun-gu, Seoul, Korea *Email : [email protected]
ABSTRACT Of contaminants contained in coal, ash was blamed for the serious issues in the power sector; a decrease in the power efficiency and discharge of fly ash into the air. Thermal extraction of coals with solvents has produced ash-free coals successfully, potentially solving the ash problem and bringing new applications such as direct coal feeding into gas turbine. Whereas the properties of the ash-free coals are well known, chars made of the extracted coals have not yet been characterized. In this study, the organic portion of a sub-bituminous coal (Roto south) was extracted at 370 °C using 1-methyl-naphthalene solvent. The extract/residue coal as well as its parent coal were pyrolyzed at 300 − 900 °C. The carbonized products were then characterized. Proximate and ultimate analysis were performed to study the compositional change. The difference in chemical structure was investigated using a solid-state 13C-NMR. Calorific value was also determined. The discussion focused on how the physical/chemical properties of the chars varied depending on the kind of the coals and the temperature pyrolyzed.
1. INTRODUCTION Coal is one of the most important energy sources, currently accounting for ~25% of worldwide energy consumption [1,2]. The use of coal is expected to increase, thanks to its abundance and economic advantage. However, there are various problems coupled with the combustion of coal such as ash deposition, emission of pollution and greenhouse gas. Coal ash decreases the power efficiency and discharges fly ash as an air pollutant [3−5]. Therefore, a lot of works have concentrated on the development of efficient methods to prepare ash-free coals [6−18]. Among them, thermal extraction with organic solvents has produced ashfree coals, namely “Hypercoal”, most successfully, potentially solving the ash problem and also bringing new applications such as direct coal feeding into gas turbine [19,20]. Once the direct combustion of ashless coal in gas turbines is achieved, an increase in the power output and reduction of CO2 emissions is expected. Economically, ash-free coals derived from low rank coals are even more attractive [21].
2 Recently, ash-free coals have found new applications; fuel for direct carbon fuel cell (DCFC) and catalytic gasification, binder/additive of metallurgical coke, and aluminum anode coke, mainly owing to their ashless character and high thermoplasticity [22−25]. The devolatilization of relatively small organic molecules happens in the initial stage of coal utilization. This early devolatilization is influential on the subsequent processes, normally homogeneous/heterogeneous reactions of char. Utilizing ash-free coals for the new applications, therefore, necessitates better understanding of the changes associated with conversion of the extracted coal into char. The devolatilization behavior of raw coals, including the structural/compositional transformation and associated thermoproperty, was studied previously [26−30]. Chemical and structural change of char, the size of aromatic rings and the nature of functional groups and cross-links, were determined by the spectroscopic methods [26]. Alonso et al. investigated the changes in char reactivity as a function of pyrolysis temperature and found a strong dependence of char reactivity on the pyrolysis temperature [27]. Partially carbonized samples at the intermediate temperature (290 − 630 °C) were also characterized, where the transformations taking place in the thermoplastic stage of coking were described [28]. However, only limited information is available on the carbonization characteristics of the solvent extracted coal and its insoluble by-product. In this study the solvent extracted coal originated from an Indonesian Roto south coal (a subbituminous coal), its insoluble by-product (residual coal), as well as its parent coals, were pyrolyzed at the temperature range, 300 − 900 °C. The effect of pyrolysis temperature was then analyzed by comparing with one another, based on the data obtained using proximate/ultimate analysis and solid-state 13C-NMR.
2. EXPERIMENTAL
2-1. Preparation of ashless coal (solvent extraction) The organic portion of a subbituminous coal (Indonesian Roto south coal) was thermally extracted using non-polar 1-methylnaphthalene (1-MN) solvent. As shown in Fig. 1, the extraction was processed via extraction, filtration, and drying step. A batch-type autoclave (0.5 liter volume) was adapted as an extractor. The raw coal was ground, meshed to < 74 μm (200 mesh), and then dried in a vacuum oven. Coal slurry was prepared by combining 20 g of the dried coal with 200 g of 1-MN. The mixture was added to the autoclave, which was then purged with N2 gas. While stirring with the magnetic agitator, the coal/solvent mixture was heated to 370 °C and held for an hour under 30 bar. The thermally treated sample was transferred to the filtering unit (glass or stainless steel filter) to separate the solvent extract from the residual matter. The solvent in both the extract and the residue was removed by being kept in a vacuum oven (~300 °C) for 3 − 4 hr under N2 atmosphere. After drying, solid materials were obtained, namely, extracted coal (EC) and residual coal (RC). Proximate/ultimate analysis, calorific value, BET surface area of the parent coal (PC), EC, and RC are tabulated in Table 1. Based on ASTM D3172 standard, the proximate analysis determined the content of moisture, ash, volatile matter, and fixed carbon, using TGA-701 Thermogravimeter (LECO Co., USA). Elemental composition (C, H, N, and O) was obtained using CHN-2000 Elemental Analyzer (LECO Co., USA), calorific value using Parr 1261
3 Calorimeter (PARR Co., USA), and BET surface area using ASAP 2420 (Micromeritics Co., USA). Agitator TC
Vent
Extractor
PG
N2 Filter Heater To dryer
Fig. 1. Schematic diagram of solvent extraction. (PG: pressure gauge, TC: thermocouple) Table 1. Proximate/ultimate analysis, calorific value, BET surface area of the parent coal (PC), extracted coal (EC), and residual coal (RC). Proximate analysis (wt%)
Sample
Moisture
Volatile matter Ash (dry) (daf)
Ultimate analysis (wt%, daf)
Fixed carbon
C
H
N
O
(daf)
BET (m2/g)
Calorific value (kcal/kg)
Parent coal (PC)
7.4
59.0
2.6
41.0
69.0
4.3
0.3
26.4
3.9
5,950
Extracted coal (EC)
-
33.6
< 0.1
66.4
87.9
4.3
0.0
7.9
0.2
8,160
Residual coal (RC)
-
39.5
4.6
60.5
87.7
4.2
0.0
8.0
4.1
7,490
2-2. Preparation of char The pyrolysis of the samples (PC, EC, and RC) was carried out at 300, 400, 500, 600, 700, 800, and 900 °C. The sample (~ 5 g) was heated at 10 °C/min ramp rate and kept at the predetermined peak temperature for 30 min under N2 sweep gas, which prevented coal oxidation and removed the pyrolyzed from the reaction zone. The product was then naturally cooled down to room temperature. The chars were named with the combination of the char origin (PC, EC, and RC) and the pyrolysis temperature; for example, ‘EC-600’ corresponds to the char made of EC at 600 °C.
4 2-3. Solid-state 13C CP/MAS NMR Solid-state 13C nuclear magnetic resonance (NMR) was performed using the CP/MAS NMR spectrometer (400 MHz, Bruker II+ model, Avance co.) at KBSI Daegue center. The measurements were made at the carbon frequency = 100.62 MHz. ~70 mg of sample was loaded in a sample rotor. All the spectra were acquired, employing contact time of 2 ms, repetition time of 3 sec, and 90° 1H pulse width of 3.8 μs. This was combined with a magic angle-spinning frequency of 13.5 kHz. The chemical shifts were calibrated with respect to tetramethylsilane (TMS) using the peak of the methyl group on hexamethyl-benzene at 0 ppm as the external standard. 3. RESULTS and DISCUSSION 3-1. Proximate/ultimate analysis and calorific values The proximate analysis data of the pyrolyzed PC, EC, and RC are given in Table 2. The ash content of PC and RC increased as the pyrolysis temperature increased. The ash content of PC changed from 2.6 wt.% for the untreated (RT) to 5.9 wt.% after the carbonization at 900 °C. For residual coal (RC), the ash content increased from 4.6 wt.% at RT to 7.9 wt.% after heat treatment at 900 °C. This increase resulted from the enrichment of ash due to the loss of volatile matters by raising temperature. During the solvent extraction, most of ashes in PC were retained in RC and the ash content in EC chars was less than 0.1 wt.%, regardless of the temperature. With the rise in the carbonization temperature, a marked decrease in volatile matter along with an increase in fixed carbon content was observed for all three samples, leading to an increase of fuel ratio. The change is most significant for the chars treated at 500 − 700 °C. Higher content of volatile matter in RC than in EC seemed to be attributed to devolatilization, which happened during the solvent drying step at 300 °C. EC is more susceptible to devolatilization than RC, since EC has more light components (low molecular weight) readily volatized or decomposed at the elevated temperatures [21]. Much higher fuel ratio (Fixed carbon wt.% / volatile matter wt.%) of EC than PC and RC agree with the above idea. More than 65% of the volatiles for PC/EC and ~50% for RC, were released by heating at 900 °C, compared to the chars made at 300 °C. Table 2. Results of proximate analysis and calorific values of pyrolyzed PC, EC, and RC. Ash (wt.%, dry)
Volatile matter (wt.%, daf*)
Fixed carbon (wt.%, daf)
Calorific value (kcal / Kg)
Fuel ratio
T (°C)
PC
EC
RC
PC
EC
RC
PC
EC
RC
PC
EC
RC
PC
EC
RC
300
2.0
0.0
4.7
52.6
32.0
34.5
47.4
68.0
65.5
0.9
2.1
1.9
6020
7610
7330
400
2.9
0.0
4.9
50.6
24.8
32.1
49.4
75.2
67.9
1.0
3.0
2.1
6150
7820
7350
500
3.7
0.0
5.3
49.0
24.3
32.5
51.0
75.7
67.5
1.0
3.1
2.1
6120
7940
7390
600
3.9
0.0
6.7
37.5
16.4
26.4
62.5
83.6
73.6
1.7
5.1
2.8
6910
8020
7410
700
4.8
0.0
7.2
23.3
16.4
22.2
76.7
83.6
77.8
3.3
5.1
3.5
7490
8120
7430
800
4.9
0.0
7.8
19.4
11.4
18.7
80.5
88.6
81.3
4.2
7.8
4.4
7520
7960
7480
900
5.9
0.0
7.9
17.7
8.9
15.3
82.3
91.1
84.7
4.6
10.2
5.6
7430
7980
7520
*daf: dry, ash free
5
The calorific values of PC increased notably by heating at 500 − 700 °C, from 6120 kcal/Kg for PC-500 to 7490 kcal/Kg for PC-700. Negligible changes were, however, shown at the other temperatures. The abrupt increase seemed to be due to a decrease of oxygen content coupled with an increase in carbon content. As shown in Table 3, data of the elemental analysis indicated that oxygen content was reduced from 23.7 wt.% at PC-500 to 5.2 wt.% at PC-700. It has been known that the calorific values of the oxidized coals are lower than those of the counterpart [31]. On the contrary, temperature dependence was not observed for EC and RC, showing 7610 − 8120 kcal/Kg for EC and 7330 − 7520 kcal/Kg for RC. These values are as high as those for usual bituminous coals. The heating values of EC were higher than those of PC and RC, especially for the chars prepared below 600 °C, partly as a result of its lower ash and oxygen content [31, 32]. Table 3. Results of ultimate analysis of pyrolyzed PC, EC, and RC. C (wt.%, daf)
H (wt.%, daf)
N (wt.%, daf)
H/C, O/C
O (wt.%, daf)
T
PC
EC
RC
PC
EC
RC
PC
EC
RC
PC
EC
RC
PC
EC
RC
300
70.8
84.1
84.2
3.3
3.9
3.4
0.2
0.0
0.1
25.6
12.0
12.3
0.05, 0.36
0.05, 0.14
0.04, 0.15
400
72.6
87.0
85.1
3.6
3.8
3.3
0.2
0.0
0.3
23.5
9.3
11.3
0.05, 0.32
0.04, 0.11
0.04, 0.13
500
73.4
87.3
85.8
2.4
2.2
2.0
0.5
0.0
0.4
23.7
10.5
10.7
0.03, 0.32
0.03, 0.12
0.02, 0.12
600
82.1
91.5
89.0
2.5
2.0
2.2
0.8
0.0
0.7
14.6
6.5
8.1
0.03, 0.18
0.02, 0.07
0.02, 0.09
700
92.7
93.9
91.7
1.3
1.2
1.4
0.8
0.0
0.7
5.2
4.9
6.1
0.01, 0.06
0.01, 0.05
0.01, 0.07
800
93.6
93.2
93.0
0.8
0.5
0.8
0.5
0.3
0.4
5.1
5.9
5.7
0.01, 0.05
0.01, 0.06
0.01, 0.06
900
93.7
94.6
95.5
0.6
0.6
0.6
0.2
0.2
0.1
5.4
4.6
3.8
0.01, 0.06
0.01, 0.05
0.01, 0.04
*daf: dry, ash free
As expected, the carbon content increased by increasing the heating temperature (Table 3), while the hydrogen and oxygen content decreased. No difference in the nitrogen content was observed on temperature variation. In addition, 0.01 − 0.1 wt.% of sulfur remained in the chars and its content was not affected by the different pyrolysis conditions (data is not shown). A decrease of the hydrogen content according to the temperature occurred either by devolatilization of alkyl groups or condensation of the aromatic molecules [32]. The oxygen was most likely released as water or CO2 by decomposition of oxygen containing functional group (such as carboxylic and etheric group) at the higher temperature. The carbonized products were generally of more condensed aromatic form than the raw coals and the oxygen contents of the chars pyrolyzed above 700 °C were 4 − 6%, which corresponded to the oxygen level of typical bituminous coal. The H/C ratio of the three chars and its temperature dependence were very similar to each other in spite of a large difference in their nature. In common, the H/C ratio varied from 0.05 at 300 °C to 0.01 at 900 °C, decreasing gradually with the increase of the pyrolysis temperature. Again, this resulted from devolatilizing low molecular weight fragments and dehydrogenation of the coals [33]. The O/C ratios of PC obtained at ≤ 600 °C were at least twice as large as those of
6 EC and RC, indicating the difference in the chemical compositions between PC and EC/RC. The PC-originated chars at ≤ 600 °C were richer in oxygen functional groups as well as aliphatic hydrocarbons [19]. The pyrolysis at the higher temperature (600 − 700 °C) devolatilized the majority of the oxygenated groups in PC, giving about the same level of O/C ratio at ≥ 700 °C as in EC/RC. 3-2. Solid-state 13C-NMR The 13C-NMR spectra of the chars made of PC, EC, and RC are shown in Fig. 2. The NMR spectra revealed the presence of a variety of carbon types and also estimated the ratio of aromatic and aliphatic carbon [34−37]. A weak peak appeared at chemical shift (δ) = 190 − 175 ppm, a characteristic peak of carboxylic carbons, for PC-300 and PC-400 in Fig. 2(a), but disappeared when heat-treated at the higher temperatures. That peak was absent for both EC (Fig. 2(b)) and RC (Fig. 2(c)) at all the temperatures.
7 0 0
5 9 5
(a) PC
(b) EC PC-800
EC-700
6 0 0
PC-700
Intensity (arb.)
Intensity (arb.)
4 9 5
5 0 0
PC-600 4 0 0
PC-500 3 0 0
EC-600 3 9 5
EC-500 2 9 5
EC-400
PC-400
1 9 5
2 0 0
PC-300 1 0 0
5 9
200
5 9 0
150
100
ppm
50
0
200
150
100
EC-300 50
ppm
(c) RC RC-700
Intensity (arb.)
4 9 0
RC-600 3 9 0
RC-500 2 9 0
RC-400 1 9 0
RC-300 0 9
200
150
100
ppm
50
0
Fig. 2. 13C NMR spectra of chars made of (a) PC, (b) EC, and (c) RC.
0
7 A phenolic carbon peak showed up at δ = 165 − 150 ppm. A phenolic peak was stronger in PC than in EC/RC. For all the samples the intensity became weaker with an increase in the temperature, such that it disappeared when the carbonization was done at ≥ 700 °C. The oxygencontaining functional groups were removed from the aromatic rings during the high temperature carbonization. Broad peaks covering 150 − 100 ppm was the strongest with the maximum positioned at ~128 ppm. This peak corresponds to the overlap of non-protonated (148 − 129 ppm) and protonated (129 − 93 ppm) aromatic carbon peaks [35, 36]. The protonated peaks in PC were relatively strong, compared to those in EC/RC. Intense peaks at δ = 50 − 0 ppm were shown for PC-300 and PC-400, whose intensity was comparable to the intensity of their aromatic components (Fig. 2(a)). Much smaller peaks were detected for EC and RC at the same condition (Fig. 2(b) and (c)). The peaks at 50 − 25 ppm can be assigned to unsaturated aliphatic carbons. The peaks in the 25 − 0 ppm region are due to the resonance of methyl groups attached either to straight-chain aliphatic groups or aromatic/alicyclic structures [34]. The strongest peak in PC-300 and PC-400 was found at ~31 ppm and known to be related to the methylenic carbon [34]. A lack of aliphatic peaks in EC/RC was most likely caused by the severe solvent drying (300 °C for 3 − 4 hr), indicating that the aliphatic molecules were devolatilized at the lower temperature than any other aromatic species [28]. Therefore, significant loss of volatile matters observed in Table 2 at the higher temperature seemed to come from small aromatic molecules. The aromaticity of the three components became higher with increasing temperature, in agreement with the ultimate analysis results.
CONCLUSION Chars made of three different coals (sub-bituminous coal, its solvent extracted and residual portion) have been characterized using proximate/ultimate analysis and solid-state 13C-NMR. The chemical analysis data agreed with the expected trends, which were an increase in carbon content/fuel ratio, a decrease in volatile matter/hydrogen/oxygen contents with increasing pyrolysis temperature. The change was more significant at 500 − 700 °C. An increase of the calorific value in PC was observed, which was due to a decrease of oxygen content and a concomitant increase in carbon content. The hydrogen content decreased with increasing carbonization temperature, as a result of devolatilization of alkyl groups and condensation of the aromatic molecules. The changes in the chemical structures were examined using solid-state 13CNMR results. With an increase of the pyrolysis temperature, aliphatic hydrocarbons were removed, such that only a negligible amount of the aliphatic groups existed above 500 °C.
8 Reference 1. International Energy Agency (IEA), World Energy Outlook 2007 China and India Insights, IEA, Paris, 2007, p663. 2. R. Kurose, M. Ikeda, H. Makino, M. Kimoto, T. Miyazaki, Fuel 83 (2004) 1777. 3. K. Steel, J. Besida, T. O’Donnell, D. Wood, Fuel Processing Technol. 70 (2001) 171. 4. K. Steel, J. Besida, T. O’Donnell, D. Wood, Fuel Processing Technol. 70 (2001) 193. 5. K. Steel, J. Patrick, Fuel 80 (2001) 2019. 6. Kershaw Jr., Fuel 76 (1997) 453. 7. K. Bartle, D. Jones, H. Pakdel, C. Snape, A. Calimli, A. Olcay, T. Tugrul, Nature 277 (1979) 284. 8. T. Takanohashi, T. Yanagida, M. Iino, Enegy Fuel 10 (1996) 1128. 9. L. Pang, A. Vassallo, M. Wilson, Org. Geochem. 16 (1990) 853. 10. B. Van Bodegom, J. Van Veen, G. Van Kessel, M. Sinnige-Nijssen, H. Stuiver, Fuel 63 (1984) 346. 11. N. Kashimura, J. Hayashi, T. Chiba, Fuel 83 (2004) 353. 12. K. Miura, M. Shimada, K. Mae, Y. Huan., Fuel 80 (2001) 1573. 13. K. Renganathan, J. Zondlo, E. Mintz, P. Kneisl, A. Stiller, Fuel Process Technol. 18 (1988) 273. 14. K. Miura, M. Shimada, K. Mae, S. Yoo, Fuel 80 (2001) 1573. 15. H. Shui, Z. Wang, G. Wang, Fuel 85 (2006) 1798. 16. C. Li, T. Takanohashi, T. Yoshida, Fuel 83 (2004) 727. 17. D. L. Khoury, "Coal Cleaning Technology", Noyes Data Corporation, Park Ridge, New Jersey (1981). 18. G. Andrews, M. Dorroch, T. Hansson, Biotechnol. Bioeng. 32 (1988) 81. 19. R. Ashida, M. Morimoto, Y. Makino, S. Umemoto, H. Nakagawa, K. Miura, K. Saito, K. Kato, Fuel 80 (2009) 1485. 20. N. Okuyama, N. Komatsu, T. Shigehisa, T. Kaneko, S. Tsuruya, Fuel Processing Technol. 85 (2004) 947. 21. T. Takanohashi, T. Shishido, H. Kawashima, I. Saito, Fuel 87 (2008) 592. 22. X. Li, Z. Zhua, R. De Marcob, J. Bradleya, A. Dicksa, J. Power Sources 195 (2010) 4051. 23. M. Casal, A. González, C. Canga, C. Barriocanal, J. Pis, R. Alvarez, M.A. Díez, Fuel Processing Technol. 84 (2003) 47. 24. A. Sharma, T. Takanohashi, I. Saito, Fuel 87 (2008) 2686. 25. R. Andrews, T. Rantell, D. Jacques, J. Hower, J. Steven Gardner, M. Amick, Fuel 89 (2010) 2640. 26. X. Li, J. Hayashi, C. Li, Fuel 85 (2006) 1700. 27. M. Alonso, A.Borrego, D. Alvarez, J. Parra, R. Menéndez, J. Anal. Appl. Pyrolysis 58–59 (2001) 887. 28. M. Maroto-Valer, C. Atkinson, R. Willmers, C. Snape, Energy Fuel 12 (1998) 833. 29. J. Yua, J. Lucasb, Terry F. Wall, Prog. Energy Combustion Sci. 33 (2007) 135. 30. P. Nelson, I. Smith, R. Tyler, J. Mackie, Energy Fuel 2 (1988) 391. 31. M. Sakaguchi, K. Laursen, H. Nakagawa, K. Miura, Fuel Processing Technol. 89 (2008) 391. 32. Y. Sato, S. Kushiyama, K. Tatsumoto, H. Yamaguchi, Fuel Processing Technol. 85 (2004) 1551. 33. H. Wachowska, M. Kozlowski, Fuel 75 (1996) 517. 34. S. Supaluknari, F. Larkins, P. Redlich, W. Jackson, Fuel Processing Technol. 23 (1989) 47. 35. M. Solum, R. Pugmire, D. Grant, Energy Fuels 3 (1989) 187. 36. P. Straka, J. Brus, J. Endrýsová, Chem. Pap. 56 (2002) 182. 37. B. Erdenetsogt, I. Lee, S. Lee, Y. Ko, D. Erdene, Intl. J. Coal Geology 82 (2010) 37.
An Alternative Application to the Centrifugal Dryer at a Coal Preparation Plant Ahmet Gitmez, Mustafa Yılmaz Western Lignite Establishment (GLI), Tavsanli-Kütahya, TURKEY
ABSTRACT In this study, compared with centrifugal dryer and dewatering screens substitution instead of centrifugal dryer used for drying of the fine clean coal product at the Omerler Coal Preparation Plant. This comparison was made to both rates of humidity of the product initial and investment and management costs. In the Omerler Coal Preparation Plant, fine clean coal washing at primary heavy medium cyclone and then dewatering. Fine clean coal had been feeding a centrifugal dryer after classification. The centrifugal conical sieve required change once a month and mechanical malfunction had increased, so it was thought that a different solution. The centrifugal dryer instead mounted two units dewatering screens manufactured by Omerler Coal Preparation Plant’s repair and maintenance team. The average moisture content in the range 18,5-19,5% at fine clean coal were obtained using centrifugal dryers. After the use dewatering screen, moisture content in the range 16,5-18% at fine clean coal were obtained. Initial investment cost of centrifugal dryer was $100,000. Initial investment cost of founded two units dewatering screen instead of centrifugal dryers was $50,000. These dewatering screens were mounted with machinery spare parts of Omerler Coal Preparation Plant. The annual operating cost of the centrifugal dryer has been $75,000. But the annual operating cost of two units dewatering screens have been $20,000. At the same time these dewatering screens was in operation a major contribution for dewatering process of fine clean coal. 1 INTRODUCTION Tavsanli is a town with a population of 62.000 people by Kütahya in the western part of Turkey in the Western zone. Tavsanli, the biggest town of Kütahya, continues its development as a lignite town. The town’s economy depends on coal mining which is run by the Western Lignite Establishment, a private companies and the Thermal Power Plant. Tavsanli’s general view and map are shown on Figure 1.
Figure 1. Tavsanli-Kütahya The lignite business in Turkey via the state administration began on 16/02/1938 with the Establishment of Degirmisaz enterprise of Etibank. Later, on 18.05.1939 Tunçbilek and on 23.09.1939 Soma Enterprises are operational. These three businesses combined on 01.01.1940 and established as "Western Lignite Establishment (GLI)" a subsidiary of Etibank and took place in Turkey Coal Enterprises (TKI) Authority as of 15.09.2007 which is established with the law numbered 6974. Approximately 4.6% of the lignite reserves of Turkey are in Tunçbilek. The total annual production of saleable lignite in Turkey is 50 million tons, whereas 7.5% are produced in Western Lignite Establishment (GLI) out of this. The Head Office of GLI enterprise has been operating in the open pit and underground facilities in the concession field numbered IR 4364 which is located in Kütahya province and district boundaries of Tavsanli. Approximately 80% of the scale of 7,000,000 tons 7 year run-of-mine lignite coal produced in opencasts and the rest is produced in underground businesses. The Western Lignite Establishment has removed 46 million m3 overburden, recovering 6.4 million tones of raw coal yielding 3.7 million tones coal in 2009. The Western Lignite Establishment has 30 million tones of reserves at open pit mine; 245 million tones at underground mine totaling over 275 million tones of coal reserve. The Western Lignite Establishment sells the coal to Thermal Power Station (365 MW) and individual domestic buyers.
2 DESCRIPTION of the OMERLER COAL PREPARATION PLANT The first performance tests for a 600 t/h capacity plant were completed at the end of 1993, and the final acceptance of this plant was confirmed by the end of 1994. One operator runs the processing plant with the SCADA PLC automation system. It consists of 3 sections; sizing of the run-of-mine (ROM) coal, preparation and stockpiling. The processing plant can be divided into 3 groups according to size and the required equipment. ROM coal sizes and their processing equipments can be summarized as: -150x18mm (heavy medium vessels), -18x0.5mm (heavy medium cyclones) and -0.5x0.1mm (spirals). 3 CENTRIFUGAL DRYER Centrifugal dryers in coal preparation are especially used to continuously separating mixtures of coal solids and liquids. Typical applications include coal fines, refuse fines, coarse slurry and middling. Centrifugal dryer at Omerler Coal Preparation Plant had been used to dewater the coal particles, from 10mm down to 0.5mm. Feeding capacity of centrifugal dryer had been 300 tons per hour. The photo of a centrifugal dryer was showed in Figure 2.
Figure 2. Centrifugal Dryer 4 DEWATERING SCREEN Most of the modern processes for mineral dressing and classification consume large amounts of water. Different types of machinery and equipment have been developed to recover the water used for processing and to have a final product that is easy to store. Considering its low investment and operating costs, its high capacity and ease of use, dewatering screen is the most efficient solution for recovering the water used in processing of aggregates like sand, gravel, salt and coal.
In a process, dewatering screens are mostly placed just after the hydro cyclones, monosizers, hydrosizers or spiral classifiers for a drip-free product. 4.1 Principle of Operation The screen incorporates a 45° sloping back deck section, fitted with slotted apertures across the direction of flow. Incoming slurry is fed uniformly along the top of this back section, which acts as a vibrating drainage panel. The main deck of the screen slopes upwards at 5° and is fitted with smaller slotted apertures (again slots across flow). At the lowest point of the screen, where the sloping back and main deck meet, a pool of partially dewatered slurry forms. Here solid particles bridge over the apertures and form a cake. This cake then acts as a filtration bed, only allowing very fine particles to pass through. The vibrating action conveys the cake along the screen and out of the pool where further dewatering takes place depending on the porosity of the cake. The cake is finally discharged over the adjustable weir into the product chute. Vibration is produced by two exciter motors operating at 980rpm. Alternatively geared exciters with external drive motor can be fitted to the larger screens. Easy adjustment of the amplitude of vibration, deck inclination, as well as the discharge weir plate are features incorporated to suit changes in process requirements. A high solids recovery is achieved when the screen underflow is kept in closed circuit and the only solid losses occurring would be the very fine material exiting in the cyclone overflow. With the benefits of high efficiency, high capacity, low headroom and reduced operating and maintenance costs, the screens are ideally suited for heavy-duty applications. The initial dewatered cut flows onto the top deck of a high frequency dewatering screen with a 0.5mm profile wire of Cr-Ni screen. Fine coal is taken the sieve over and bottom decks can be independently recovered. Moisture content varies from 15 to 18% depending on particle size. Screens are designed for high conductance and abrasion resistance. The operation principle of the dewatering screen was showed in Figure 3.
Figure 3. The operation principle of the dewatering screen
4.2 Operational Advantages and Gains of Application The -18x0.5 mm fine clean coals of primary heavy medium cyclone have been further classified to produce two size fractions a -18 mm by 10 mm stream and -10 mm by 0.5 mm stream. The fine clean coal has been drained, rinsed, and classified by three fine clean coal screens. The -10x0.5 mm fine clean coal has been collected on a common conveyor belt and dewatered by the centrifugal dryer until 2008. The basket screen it belongs to the centrifugal dryer has been amended once a month. At the same time this centrifugal dryer has been break downed and it was standstill in 2007. The centrifugal conical sieve required change once a month and mechanical malfunction had increased, so it was thought that a different solution. The centrifugal dryer instead mounted two units dewatering screens manufactured by Omerler Coal Preparation Plant’s repair and maintenance team in 2008. The average surface moisture content in the range 18,5-19,5% at fine clean coal were obtained using centrifugal dryers. After the use dewatering screen, surface moisture content in the range 16,5-18% at fine clean coal were obtained. These surface moisture contents belongs to the samples of the -18+0.5 mm fine clean coal it taken every shift. 5 RESULTS The centrifugal dryer instead used two units dewatering screens can be reduced the surface moisture content of the -18+0.5 mm fine clean coal from 19,5% to 18%. Initial investment cost of centrifugal dryer was $100,000. Initial investment cost of founded two units dewatering screens instead of centrifugal dryers were $50,000. These dewatering screens were mounted with machinery spare parts of Omerler Coal Preparation Plant. The annual operating cost of the centrifugal dryer has been $75,000. But the annual operating cost of the two units dewatering screens have been $20,000. Accordingly, the annual earnings are $55,000. At the same time, these dewatering screens were in operation a major contribution for dewatering process of fine clean coal.
REFERENCES CLI-Tekfen Consortium, 1991. Technical Specification of TKI Omerler Coal Washing Plant, Ankara, V-1/V-7, 1.1-4.34 http://www.linatex.com/en/process-equipment/dewatering-and-sizing-screens.html http://www.dermak.com.tr/EN/dewatering.htm http://www.kemtron.com/Products http://www.metso.com/miningandconstruction/mm_bulk.nsf http://www.elginindustries.com/sitecore/content/Solution/Industrial
The Evaluation of the Contributions to the Productivity of the Process Changes at Tuncbilek Coal Preparation Plant Ahmet Gitmez, F. Zehra Taksuk, Fatih Albayrak Western Lignite Establishment (GLI), Tavsanli-Kütahya, TURKEY ABSTRACT The most important business mission in coal washing plants, just as in all mineral processing plants, is to implement the processes and working methods in order to get access to extract the saleable products in targeted and desired quality which is theoretically possible. Coal is a non-recyclable energy source as all other underground sources. Therefore, the coal should be made best of it to ensure maximum benefit since it is produces in very difficult conditions. The process changes for the acquisiton of products for all of the additional revenue derived from the run-of-mine coal are identified in this study. At Tunçbilek Coal Preparation Plant under the direction of Western Lignite Establishment of (TKI) Turkish Coal Enterprise, the implementation regarding process changes and engineering application is discussed and an evaluation is made on the increase of production and productivity and decrease in operating costs and increase in profitability. In this assessment, the decreases in operating costs as well as increase in profitability are achieved as a result four years of implementation with respect to productivity. The approximate total income increase was annual $15,600,000, through process changes and engineering applications. 1 INTRODUCTION Tavsanli is a town with a population of 62.000 people by Kütahya in the western part of Turkey in the Western zone. Tavsanli, the biggest town of Kütahya, continues its development as a lignite town. The town’s economy depends on coal mining which is run by the Western Lignite Establishment, a private companies and the Thermal Power Plant. Tavsanli’s general view and map are shown on Figure 1.
Figure 1. Tavsanli-Kütahya
The lignite business in Turkey via the state administration began on 16/02/1938 with the Establishment of Degirmisaz enterprise of Etibank. Later, on 18.05.1939 Tunçbilek and on 23.09.1939 Soma Enterprises are operational. These three businesses combined on 01.01.1940 and established as "Western Lignite Establishment (GLI)" a subsidiary of Etibank and took place in Turkey Coal Enterprises (TKI) Authority as of 15.09.2007 which is established with the law numbered 6974. Approximately 4.6% of the lignite reserves of Turkey are in Tunçbilek. The total annual production of saleable lignite in Turkey is 50 million tons, whereas 7.5% are produced in Western Lignite Establishment (GLI) out of this. The Head Office of GLI enterprise has been operating in the open pit and underground facilities in the concession field numbered IR 4364 which is located in Kütahya province and district boundaries of Tavsanli. Approximately 80% of the scale of 7,000,000 tons 7 year run-of-mine lignite coal produced in opencasts and the rest is produced in underground businesses. The Western Lignite Establishment has removed 46 million m3 overburden, recovering 6.4 million tones of raw coal yielding 3.7 million tones coal in 2009. The Western Lignite Establishment has 30 million tones of reserves at open pit mine; 245 million tones at underground mine totaling over 275 million tones of coal reserve. The Western Lignite Establishment sells the coal to Thermal Power Station (365 MW) and individual domestic buyers. 2 DESCRIPTION of the TUNCBILEK COAL PREPARATION PLANT Coal preparation plant has two parallel circuits with the capacity of 350 tph each. Tunçbilek coal preparation plant general view is shown on Figure 2. Raw coal is fed from 600 X 300 mm grate and screened on a 150 mm opening screen. Oversize raw coal is crushed with a rotary breaker below 150 mm and combined with undersize raw coal. This prepared material is divided into two parallel circuits. Minus150 mm size material is fed to 18 mm screens.
Figure 2. A view of Tunçbilek preparation plant
Coal preparation plant section is considered in three sections: -150+18mm heavy medium drums as two parallel units each of capacity 200 ton/hour; -18+0,5mm primary heavy medium cyclones and secondary heavy medium cyclones; -0,5+0,1mm spirals. 3 The DEVELOPMENT PROCESS of TUNCBILEK COAL PREPARATION PLANT Tuncbilek Coal Preparation Plant was established in years 1952-1954 as a Jig Lavvar with the capacity 286 t/h. In 1958, Jig group was established to the fine coal part. Jig has a capacity of 150 t/h and used in 0.5-18 mm. coal dust enrichment work until 2005. During 1967-1968 revision, the Jig with the capacity 286 t/h, working with compressed air and water was dismantled due to the many failures and replaced by two pieces of magnetite working with a heavy media (2 * 200 t / h capacity, 3600 * 4700 mm in size) washing drum mounted. In 1968, the heavy medium cyclone circuit (Roberts and Schaefer) has been established to the fine coal part. Its capacity is 150 t/h and is used for enrichment of 0.5-18 mm until March 2006. In 1985 the treatment plants were established. Thereby, coals in 1000 t/h waste water previously flowing to Adranos Stream , are acquired and pumped to open pit ponds within the coal plants by the cyclones and so the solids sinked and overflowed clean water can be reused in lavvar. In 1996 thickener pool was established. Thus, 1000t/h waste water chemical precipitants (flocculants) were cleaned with 800t/h by feeding back part plant is used as wash water. 200 t/h waste water collapsed into the thickener floor is pumped to the open pit pond. In April 2006, the Jig circuit which was established in 1958 was dismantled, and then 1st heavy medium cyclone circuit to enrich 0.5 to 18 mm. coal was established instead of it. In April 2006, the current heavy medium cyclone circuit which was established in 1968 has been adapted as 2nd heavy medium cyclone circuit in order to obtain intermediate product from the separated part of 1st heavy medium cyclone circuit. In April 2006, the spiral circuit was established to enrich the 0.1 - 0.5 mm. slime coal 4 PROCESS CHANGES DONE and ITS ECONOMICAL YIELDS 4.1 Establishment of 1st Heavy Medium Cyclone Circuit by Dismantling of Current Jig Circuit Before 2006, 50% of the 0.5-18 coal obtained from the Lavvar was enriched in Jig circuit. It is determined by studies at different times, the ratio of clean coal in the Jig Schist is 10% of run-of-mine coal fed to Jig. This determination shows that 0.5 - 18 clean coal recovery as of April 2006 is 110.122 tones and the monetary value of it is approximately USD 7,693,733. The amount of washed coal and its monetary value between the years 2006 -2009 are shown in the following table 1.
Table 1. The amount of washed coal and its monetary value AMOUNT OF (0.5/18) COAL SOLD (TON)
AMOUNT OF 0.5/18 CLEAN COAL RECOVERY* %50*%10 (TON)
UNIT PRICE (USD/TON)
AMOUNT (USD)
2006
335,098
16,755
46,50
779,108
2007
579,574
28,979
66,32
1,921,887
2008
771,386
38,569
80,20
3,093,234
2009
516,376
25,819
73,57
1,899,504
2,202,434
110,122
69,87
7,693,733
YEARS
TOTAL
4.2 Adaptation of Current Heavy Media Cyclone Circuit as 2nd Heavy Fluid Cyclone Circuit In 2006, the current Heavy Media Cyclone Circuit in the fine coal circuit has been adapted as 2nd Heavy media cyclone circuit and by this revision the thin mixed product has been obtained. Since April 2006, obtained from the total amount of 245.915 tonnes of mixed product and the monetary value of approximately is 8,901,328 USD. The production and monetary value between the years are shown in table 2: Table 2. The production and monetary value RUN-OF-MINE COAL FED (TON)
AMOUNT OF GAINED MIXTURE
UNIT PRICE (USD/TON)
AMOUNT (USD)
2006
2.519.851
62.000
31,29
1.939.980
2007
2.730.464
61.365
36,22
2.222.640
2008
3.323.009
60.725
37,82
2.296.620
2009
2.838.541
61.825
39,50
2.442.088
TOTAL
11,411,865
245,915
36,20
8,901,328
YEARS
4.3 Establishment of Spiral Circuit By the establishment of Spiral Circuit in April 2006, 0.1-0.5 mm was begun to enriched, and approximately 2% of run-of-mine coal fed as clean coal has been obtained. This obtained product is offered to sale within the coal dust. Before the establishment of spiral cycle, the product obtained by classification only was given to the thermal power plants. The amount of coal is at a rate of 3% of the amount of run-of-mine coal fed in this case. The acquisition that would be obtained is shown in Table 4 if the coal production had been continued in this system.
Table 4. – Material gain if 0.5+0.1 mm coal were classified and sold YEAR 2006
RUN-OF-MINE COAL FED CLASSIFIED RUN-OF-MINE (TON) COAL * 3%
UNIT PRICE (USD/TON)
AMOUNT (USD)
2006
1.828.610
54.858
31,29
1,716,507
2007
2.730.464
81.914
36,22
2,966,925
2008
3.323.009
99.690
37,82
3,770,276
2009
2.838.541
85.156
39,50
3,363,662
TOTAL
11,411,865
342,356
34,52
11,817,370
After the establishment of spiral unit, material gains from the sales to the industrial sector instead of thermal power plant are shown in Table 5. Table 5. – Material gains from the sale of 0.5+0.1 mm coal after enrichment in the spiral: RUN-OF-MINE COAL FED (TON)
AMOUNT OBTAINED: RUN-OF-MINE COAL* %2
UNIT PRICE (USD/TON)
2006
1.828.610
36,572
46,50
2007
2.730.464
54,609
66,32
2008
3.323.009
66,460
80,20
2009
2.838.541
56,771
73,57
TOTAL
11,411,865
214,412
69,87
YEAR 2006
AMOUNT (USD)
14,980,966
By running the spiral enrichment unit, it could be possible to produce clean coal from 0.5+0.1 mm sized raw coal. Therefore more amount of clean coal could be sold to industrial consumers instead of selling raw coal to Thermal Power Plant. The difference is the additional income. 4.4 The Establishment of Coarse Mixture Product Circuit In 2010, in order to produce big mixture, the second heavy medium roller circuit will be established. With a second washing process, it would be possible to produce big mixture from the mixture and the schist which are removed from big coal washing unit. By running the system it is estimated that an average of 175,000 tons of big mixture will be produced per year. At least $7,000,000 income will be expected by selling this mixture to Thermal Power Plant with a price of $40 per ton. 4.5 The Income Obtained from Regaining of Coal from the Material which Goes to the Thickener In current system, the coal inside of the slime coal which is smaller than 0.1 mm can not be taken and therefore this slime coal is sent to the thickener then goes to waste dam. In order to regain the coal inside of the 0.1 mm sized slime coal a facility will be established in Tunçbilek Lavvar in 2010. By establishing this facility, %3 of the run of mine coal will be regained and 90000 tons of merchantable coal will be produced. It is
estimated that at least $3,600,000 income will be gained by selling this product to Thermal Power Plant with a price of $40 per ton. 5 RESULTS Western Lignite Establishment has made many changes and additions due to technological developments and changing market conditions since it was established. In this study, it is seen that the process changing lead to significant increments on incoming. After removing Jig circuit and establishing the heavy medium cyclone circuit, 110,122 tons of 0,5-18 clean coal has been regained since April, 2006. Its value is approximately 7,693,733 USD. Heavy medium cyclone circuit 245,915 tons of thin mixture has been regained since 2006. Its financial value is 8,901,328 USD. Establishment of the spiral circuit led to $3,163,596 of additional income between 2006 and 2009. The total additional income resulting from the process changing in a period of four years is $19,758,657. This means an annual average of $5,000,000 additional income. In 2010 with the establishment of second heavy medium roller system it is estimated that 175000 tons of big mixture will be produced and at least $7,000,000 income will be gained. Also 90000 tons of coal will be regained from thickener waste and this will provide $3,600,000 additional income. With these new systems, an annual total of $10,600,000 will be saved. The process changing in Tunçbilek Lavvar and the additions which will be made in 2010 will provide an annual average of $15,600,000 additional income. Like all mineral preparation plants, the most important goal of coal washing plants is realizing the appropriate process and working techniques in order to reach the desired and targeted quality. In this study the importance of regaining the merchantable coal from run of mine coal is clearly showed by the help of additional incomes. REFERENCES The Evaluation of the Performance's Tuncbilek Coal Preparation, The Evaluation of the Waste’s Plant, and the Final Report of the Process Assessment Project for Omerler Coal Preparation Plant, METU, The Applied Research Project Code Nr: 87.03.05.01.06, April 1989, Ankara, sh:2-5 GLI Annual Reports, 2006,2007,2008,2009, Tavsanli, Not published.
THE DUMP TRUCK REQUIREMENT PLANNING STUDİES OF TURKİSH COAL CORPORATİON
Mustafa ZIYPAK, Turkish Coal Corporation, Machine and Supply Department, Ankara
ABSTRACT Operation field of Turkish Coal Corporation (TKİ) is lignite production , most of which is achieved from open pits. The Corporation is realized important portion of overburden removal and coal production activities carried out at open pits by means of its own facilities (personnel and equipment). For overburden removal workings carried out at TKİ’s open pits, generally electrical excavator – dump truck combinations have been aplied. Economical life of electrical excavators, whose operation costs are low, has been going on. In addition, these excavators have idle capacities due to dump truck inefficiency. On the contrary, economical life of The Corporation’s dump trucks is expired to a great extent owing to high working hours and obsolete technology. Because of this, by scraping most of existing dump trucks of Turkish Coal Corporation, instead new dump truck investment is inevitable. In this study, by taking capacities and loading heights of existing excavators as a base, classes of trucks to be purchased have been determined; by considering estimated overburden removal to be performed by TKİ facilities, total capacity of excavators which is working at TKİ site, annual capacity of dump trucks whose economical life is not completed, annual capacity of dragline excavators that are working at TKİ sites, number of dump trucks to be invested have been calculated. Furthermore, comparison of operating costs of advanced technology dump trucks (electrical) with old technology mechanical trucks have been made; calculation of pay back period of trucks to be replaced has been made.
1. GENERAL INFORMATION Turkish Coal Enterprise (TKİ) which carries out activities of production, enrichment and marketing of lignite, that is one of the important energy sources of Turkey, is among the leader mining establishments in Turkey. TKİ, which was founded in 1957, is a state owned organization. For the year of 2009, sales income of the Enterprise is approximately $1.500.000.000, profit of that is $280.000.000. As end of 2009, TKİ recruited 8832 personnel in total. Lignite production activities of the Corporation have been conducted by 8 limited autorised companies which are spread in various regions. These are as followings: - ELİ: Aegean Lignite Establishment (Soma/Manisa) - GLİ: Western Lignite Establishment (Tavşanlı/Kütahya) - SLİ: Seyitömer Lignite Establishment (Seyitömer/Kütahya) - GELİ: South Aegean Lignite Establishment (Yatağan/Muğla) - YLİ: Yeniköy Lignite Establishment (Milas/Muğla) - ÇLİ: Çan Lignite Establishment (Çan/Çanakkale) - BLİ: Bursa Lignite Establishment (Orhaneli/Bursa) - ILİ: Ilgın Lignite Establishment (Ilgın/Konya) While 20% of total known lignite reserves in Turkey belongs to TKİ Corporation, the company realises 50-55% of lignite production in the country. 2009 lignite production of the corporation is 43 million tonnes.
80% of lignite produced by TKİ is used for power plant, remaning is used for heating and industry. TKİ suplies 8 power plant at 4209 MW which corresponds to approximately 10% of Turkey’s total electricity production capacity. TKİ Corporation produces Lignite by surface mining as well as by underground mining. For 2009 production 80% lignite from open pits, remaining by underground pits was obtained. However, it is estimated that underground rate will remain around last achieved level and open pit production will preserve its importance. 2. SURFACE MINING AT TKİ As known, overburden removal is an important activity of surface mining which is most costly. TKİ Corporation performs work of overburden removal by means of its own facilities (machinery, personnel) as well as outsourcing application. At recent years, The Corporation carries out 90-100 million m3 overburden removal (on average 42%) annually by its facilities, between 200-275 million m3 in total together with outsourcing. For overburden removal at TKİ’s open pits which is carried out by its facilities, generally electrical ekscavator (shovel) – dump truck method have been applied. In this method, firstly the area to be excavated and loaded are softened by means of drilling and blasting, after trucks are loaded by electrical excavators with 2-6 passes depending on compatibility. Trucks are unloaded at dumping site in 1-8 km distance according to projects. In the case of lacking of electrical excavators or owing to platform conditions, sometimes high capacity hydraulic excavators (backhoe type) may be used. In landslided areas, in addition to electrical excavators, high capacity loaders (12-14 yd3) and hydraulic excavators have also been used. For lignite production mostly hydraulic excavators and loaders, at intermittent coal seams fully hydraulic excavators are operated. Moreover, at some of TKİ’s open pits dragline machines that work integrated with excavator – truck combinations have been used. Draglines are usually used for digging overburden at first stage and dumping it to area where coal production has allready been completed. Since it does not require transportation, it is the most economical method at surface mining. 3. A GENERAL VIEW TO OPEN PIT MACHINERY OF TURKISH COAL CORPORATION Prior to 1980 open pit machinery of TKİ Corporation was limited, therefore coal production from surface mining was quite low. The eighties were TKİ’s progress years; huge machinery investments were realised (majority of coal fired power plant investment coincided during same period). Existing mining machinery of TKİ are comprised mostly the machines which purchased during the eighties which can be seen in table 3.1. For the works of TKİ’s overburden removal and coal production, various type, mark, model approximately 950 mining machinery (excavator, dump truck, loder, buldozer, grader, drilling machine etc.) have been used. In addition, The Corporation owns approximately 800 auxiliary machinery (fuel tanker, road truck, watering truck, forklift, mobile vinch etc.) From the nineties machinery investments have been carried out as replacement at low volumes. Outstanding investments from 1990 up to today have been 12-14 yd3 capacity loaders (18 pieces), 170 short ton electrical dump truck (20 pieces), 6.5 yd3 capacity hydraulic excavators (22 pieces). Purchasing of the mentioned loaders and excavators have been realised gradually. During last ten years, nearly 8-9% of mining machinery have been replaced; these are excavators, loaders, bulldoders, graders, auxiliary machines etc. As firstly dump trucks, due to high age and working hours, performances of majority of TKİ’s mining machinery have dropped and operating cost of those increased; therefore economical lives of the machinery have expired. However economical lives of the machines procured during nineties and millennium years and electrical machinery namely excavators, draglines have still been going on.
Table 3.1 According to Commissioning Years, Distribution of TKİ Machinery Machine Dragline Electrical exc. Hyraulic exc. Dump truck Bulldozer Grader Loader Drilling machine Total %
<1981 81-85 2 0 1 7 0 0 4 242 1 91 0 24 4 16
86-90 6 60 2 173 65 8 12
91-95 0 0 3 27 0 0 7
96-00 0 0 12 20 9 6 8
00-05 >2006 0 0 0 0 9 17 0 0 7 1 10 4 16 9
Total 8 68 40 466 173 50 71
0
24
28
1
0
2
0
55
12 1,3
404 43,1
354 37,7
38 4,0
55 5,9
44 4,7
31 3,3
931 100,0
4. DUMP TRUCKS AND EXCAVATORS As mentioned above, excavators and dump trucks are main machinery for overburden removel and coal production. As 170 short ton 20 electical, 85 short ton 372 rigid mechanical, 50 short ton 50 rigid mechanical, 120-150 short ton 24 trailer type, there are totally 462 dump trucks in TKİ’s Machinery park. Total capacity of these trucks is 40850 short ton. Transportation of almost all TKİ’s own overburden has been accomplished by means of 85 and 170 short ton dump trucks. 50 short ton dump trucks are usually used for coal transportation. 120-150 short ton trailer type dump trucks are also specially designed for coal transportation. Experiences show that transportation cost for a surface coal mining is approximately 50% of total overburden removal cost. Thus it is very important performance of dump trucks. Performance level of dump trucks are mainly monitored by working hours and transported material. In this context, operation and maintenance costs are also important parameters on which more attention should be paid. As shown in previous section, total working hours of 85 short ton dump trucks have increased considerably due to high age (table 4.1). In addition to high age, as being products of old tecnology causes high operating cost and therefore ending economical life. Athough working hours of 170 short ton electrical dump trucks have reached at certain level, they still fulfill high performance and thus have not completed their economical life. Table 4.1 Working Hours Distribution of TKİ Dump Trucks
Dump Truck Tonnage 85 s.ton ORAN (%) 170 s.ton ORAN (%)
Toplam 0(adet) 10.000
Total 10.000 20.000 30.000 40.000 50.000 60.000 > averega 70.000 working 20.000 30.000 40.000 50.000 60.000 70.000 hours
372
0
8
30
60
130
89
37
18
100
0
2,2
8,1
16,1
34,9
23,9
9,9
4,8
8
2
10
40,0
10,0
50,0
0,0
0,0
0,0
20 100
0,0
0,0
47317
35.925
Electrical excavators were put into operation during the eighties in the scope of the open pit projects that TKİ started. Movements of bom and bucket of electrical excavators, which are shovel type, are achieved by means of wire ropes. Total number of excavators, whose bucket capacities ranges from 10 to 20 yd3, are 68 and total capacity of those are 985 yd3. Average working hour of TKİ’s shovels is around 47500. Since electrical excavators are built from durable units and parts, reliabilities and lives of them are high. In fact, it is argued by experts that economical life of electrical excavators can reach at 100.000 hours. Therefore, it can be stated that TKİ’s excavators are at half life. Due to explained reasons, operating cost of electrical excavators is quite lower than that of diesel powered hydraulic excavators. Moveover, although ages of these machines are quite high, there is no any problem in procuring spare parts. 5. EXAMİNATION OF DUMP TRUCK OPERATION COSTS Direct cost of overburden removal and coal production activities carried out in the open pits consists of drilling and blasting, excavating-loading, transpotation and auxialliry machinery expenditures. Among these, dump truck, used for transportation, cost comes first; because truck costs constitute 45-50% of total overburden removal cost. Main items of dump truck operation cost (including maintenace cost) are fuel, spare part, tire, oil, driver and maintenace labour expences. For comparison, operation costs of 85 short ton mechanical and 170 short ton electrical dump trucks have been calculated by taking ELİ ve GLİ operating data as base. The costs heve been determined as $/hour and $/m3, and showed in table5.1 and 5.2 respectively. It is important to stress that there are similarities and differences in working conditions of ELİ and GLİ. Table 5.1 Dump Truck Operation Costs ($/hour) Driver
Maintenace
Tonnage
Fuel
85 S.ton
68,9
21,67
16,28
21,09
4,36
2,79
135,09
170 S.ton
91,5
17,42
11,39
11,98
12,79
1,48
146,56
Spare part Labour
Tire
Oil
Total
Table 5.2 Dump Truck Operation Costs ($/m3) Tonnage
Maintenance
Fuel
Driver
85 S.ton
0,7
0,23
0,16
0,24
0,05
0,03
1,41
170 S.ton
0,48
0,11
0,05
0,07
0,07
0,01
0,79
Spare part Labour
Tire
Oil
Total
Examining $/hour costs in table 5.1, although double capacity, it is understood that total expenditure of 170 short ton electrical trucks is only 8,5% higher than that of 85 short ton mechanical trucks; fuel expenditure of 170 short ton electrical trucks is only 32,8% higher than that of 85 short ton mechanical trucks. Looking at $/m3 costs in table 5.2, it is calculated that total expenditure of 170 short ton electrical trucks is 56,0% of that of 85 short ton mechanical trucks; fuel expenditure of 170 short ton electrical trucks is 68,6% of that of 85 short ton mechanical trucks.
6. TKİ’S DUMP TRUCK NECESSITY DETERMINATION 6.1 Importance of New Dump Truck Investment It can easily be stated that TKİ is urgently in need of dump trucks when considering TKİ’s open pit reverves, life of these reserves, envisaging of TKİ’s own operation continuation and necessity of this, expiring of most of the existing dump trucks, being more economical of new tecnology product dump trucks (electrical), continuation of economical life of existing electrical excavators and being idle capacity of these etc.; some of which have already been mentioned in previous sections. 6.2 Determination of Dump Truck Capacity As mentioned in section 4, economical life of electrical excavators has still been going on; thus when new truck investment is realised, existing excavators will be continued to use. Because of this, in determination of the dump truck capacity, capacities and dimensions (loading height) of existing excavators should be considered. In addition, for economical operation, 3 matters should be taken into consideration in surface mining. Firstly, it should be benefited from “scale of economics”, in other words as the capacity increases, it should be known that unit cost decreases. Secondly, capacities of excavators and trucks are fitted at integer multiples as close as poosible (e.g. excavator capacity: 20 m3, truck capacity: 61 m3, 3 pases). Lastly, in order to optimize total idle cost of excavator plus truck, trucks sould be loaded certain number of pases; this is determined as 3-6 pases in practice. The excavators with capacities of 10, 15, 17 and 20 yd3 exist in TKİ’s machinery park though there are differences from site to site. Moreover, there are 6 dump truck manufacturers in the world; they manufacture classes of 100, 150, 190, 240, 300 and 360 short ton standart trucks. Therefore by considering TKİ’s excavators and other matters mentioned above, it is necessery to select from these trucks. As a result of investigation that we made, the maximum capacity trucks which TKİ’s excavators can load are 190 short ton class trucks. What is more, dimensions of existing workshops provide maintenace of maximum 190 short ton class trucks. Because of these, trucks to be purchased must be 190 short ton or lower class. As a result of analysis carried out in relation to above explanations, it has been come out that 190 short ton class truck for ELİ, GLİ and ÇLİ mining sites, 150 short ton class for YLİ, 100 short ton class for SLİ are suitable. 6.3 Criterions for Dump Truck Need Determination and Quantity Calculation For determining truck need, TKİ’s own overburden removal programme has taken into consideration, envisaging on the assumption of performing coal production by existing trucks, coal production hes been neglected in this study. In this context, following paramaters have been taken as base in determinig number of trucks whose classes are chosen for every open pit site: 1. Annual planned quantity of TKİ’s own overburden removal 2. Annual capacity of the highest capacity excavators of every open pit site 3. Annual capacity of TKİ’s dump trucks whose economical life has not completed 4. Annual capacity of TKİ’s dragline machines As seen table 6.1 in determining dump truck need, 85 short ton mechanical trucks have not been considered. As dump truck investment is realised, It is envisaged that these trucks will be scrapped. Furthermore, these trucks may be used for coal production. In the context to these conditions, firstly dump truck which has economical life and dragline capacities have been deducted from planned overburden removal quantity, then the obtained value has been compared with excavator capacities, eventually small value has been used for calculating
Table 6.1 Calculation of Truck Requirement
20 yd3
5
2.250.000
20 yd3
5
Ann.o.burden Selected truck's Determination of annual truck capacity Cal.truck.num. removal Nominal D.body Average Undigged Annual prog. for Bucket capacity volume dumping density capacity Real Interger truck need number (s.ton) (m3) distance(m) (tonne/m3) (m3/truck) (m3) 20.089.880 16.000.000 190 77-107 3200 6 2,45 756.703 21,1 21 7.400.000 190 77-107 2750 6 2,2 830.413 20.536.320 8,9 9
11.000.000
6.000.000
15 yd3
9
28.140.939
5.000.000
SLİ
10.000.000
3.000.000
10 yd3
10
22.003.200
7.000.000
100
40-60
2000
6
2
636.127
11,0
11
ÇLİ BLİ
8.600.000 7.000.000
20 yd3
5
20.536.320
8.600.000
190
77-107
2000
6
2,2
950.353
9,0
9
1.000.000
15 yd3
3
7.677.384
6.000.000
YLİ
20.000.000
4.000.000
15 yd3
4
18.482.688
16.000.000
915.430 17,5
18
Ann. capacity of trucks that have econ. Life (m3)
ELİ
25.000.000
9.000.000
GLİ
15.650.000
6.000.000
GELİ
Production company
Annual own overborden removal prog. (m3)
Excavators' annual capacity of Nominal Total ann. draglines capacity Number capacity (m3) (m3) (m3)
New truck investment not needed
New truck investment not needed 150
Maneuver time (loading)………: 190 s.ton: 32 sn, 150 s.ton: 31 sn, 100 s.ton: 30 sn Maneuver time (unloading)……: 190 s.ton: 38 sn, 150 s.ton: 37 sn, 100 s.ton: 36 sn Unloading time…………...........: 190s.ton: 94 sn, 150 s.ton: 92 sn, 100 s.ton: 90 sn Average truck speed…………...: 190 s.ton: 25 km/hour, 150 s.ton: 26 km/hour, 100 ston: 27 km/hour Excavator cycle time…..………: 20 yd3: 30 sn, 15 yd3: 30 sn, 10 yd3: 28 sn Average dumping site distance..: GLİ-SLİ-YLİ: 2000m, ELİ: 2400m, ÇLİ: 2250m Undigged overburden density....: SLİ-YLİ: 2 t/m3, GLİ-ÇLİ: 2,2 t/m3, ELİ: 2,3 t/m3 Annual working time………….: 4800 hour Assumed productivity………....: GLİ-SLİ-ÇLİ-YLİ: %70, ELİ: %80
55-88
2000
6
2
truck number. The number of needed truck has been calculated by dividing the idendified value with the selected truck capacity. Excavator cycle time, dump truck maneuver time (for loading amd unloading), truck unloading time, average truck speed, average dumping site distance, undigged overburden density, annual working time and workplace productivity which are used in the calculation of capacity of the truck chosen for every open pit site have been given under table 6.1. As shown in table 6.1, according to given parameters and assumptions, number of necessary dump truck for each mining site has been calculated, i.e. - 190 short ton 21 dump trucks for ELİ, - 190 short ton 9 dump trucks for GLİ, - 190 short ton 9 dump trucks for ÇLİ, - 150 short ton 18 dump trucks for YLİ, - 100 short ton 11 dump trucks for SLİ. 6.4 Economical Evaluation of Dump Truck Investment In section 5, operation costs of TKİ’s two major dump trucks, namely 85 short ton mechanical, 170 short ton electrical have been compared in terms of $/hour and $/m3. It has been understood that operation cost of 170 short ton electrical truck is much lower and therefore more economical. For simplicity, dump truck investments for every mining site have been evaluated with payback period method. Regarding this, operation cost difference between 85 short ton mechanical truck and 170 short ton electrical truck of TKİ which is 0,62 $/hour has been taken as base. As shown in table 6.2, it has been calculated that payback period is 6.2 years for ELİ, 5.6 years for GLİ, 3.8 years for SLİ, 4.9 years for ÇLİ, 4.2 years for YLİ. It can be argued that calculated payback periods are reasonable; since economical life of new trucks is at least 20 years. Table 6.2 Economical Evaluation of TKİ Truck Investment Selected Production truck company tonnage
Neccesary Price Truck unit overburden difference price ($) removal for ($/m3) payback (m3)
Annual truck capacity (m3)
Payback period (year)
ELİ
190
2.900.000
0,62
4.677.419
756.703
6,2
GLİ
190
2.900.000
0,62
4.677.419
830.413
5,6
SLİ
100
1.500.000
0,62
2.338.710
636.127
3,8
ÇLİ YLİ
190
2.900.000
0,62
4.677.419
950.353
4,9
150
2.400.000
0,62
3.870.968
915.430
4,2
7. CONCLUSION 466 dump trucks are operated at TKİ’s open pits for overburden removal and coal production. Economical life of the most trucks has expired, operation cost of them increased. Therefore they must be replaced as soon as possible. TKİ’s dump trucks must be scrupped in short or medium term whether they are repaced or not. Consequently, as taking a strategical decision, direction of the corporation shoul be made clear regarding its own operation. If new dump truck investment will not be done, TKİ have to decrease own operations to minimum level by reducing machinery with respect to a certain Schedule.
If a positive decision is taken for new dump truck investment, according to the result achieved from this study, as 190 short ton 39 pieces, 150 short ton 18 pieces and 100 short ton 11 pieces, totally 68 dump trucks should be purchased. Total investment cost of these trucks is approximately 173.000.000$ and procurement of them can be realised step by step over years. Analysis conducted in this study indicates that considerable reduction in operation cost of trucks can be obtained and investment can pay back within 4-6 yeras. Owing to low operation cost, high operation productivity that is provided by high safety and confort, easiness of management and organization enabled by few trucks, it is estimated that cost of TKİ’s own overburden removal can compete with that of contractor easily.
REFERENCES TKİ INVESTİGATİON, PLANNİNG AND COORDİNATİON DEPARTMENT (2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010), Activity Reports of TKİ Enterprise for 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 and 2009, Ankara/Turkey TKİ MACHINERY AND SUPPLY DEPARTMENT (2010), The book of TKİ Machinery Park, Ankara/Turkey ZIYPAK, M. (2007), Dump trucks, Examination of Their Structures and Performances, Proceedings of the 1st Mining Machinery Symposium and Exibition, Kütahya/Turkey, p.335-346 ZIYPAK, M. (2009), A General View to Open Pit Machinery of Turkish Coal Corporation, Proceedings of the 2nd Mining Machinery Symposium and Exibition, Zonguldak/Turkey, p.117-127
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Coal Science: Coal Chemistry Investigation of Radioactive Contents Soma Coals İsmail DEMİR, İlgin KURŞUN Istanbul University Engineering Faculty Mining Engineering Department, Istanbul, Turkey [email protected], [email protected] Abstract: Coal, world's the most abundant, the most accessible and the most versatile source of fossil energy was brought to the forefront of the global energy scene by the industrial revolution of the 18th century. Like any fossil fuel, coal is associated with naturally occurring radioactive materials. This is due to their U, Th, and K contents. This certainly has radiological implications not only for the miners but also for the populace in the immediate environment of the mines and the users. In this study, the radioactive elements in Manisa-Soma coals and their ashes were carried out. In the experimental section, the coal and thermal power plant ashes which were taken from Manisa – Soma were used. Sieve, moisture, ash, calorific value, volatile amount, total carbon, total sulphur, major element and radioactive element analysis of the samples were carried out. The float and sink analysis and flotation tests of the samples which were taken from Manisa-Soma were carried out. Thus, radioactive elements changes and moving mechanisms were investigated with coal preparation and burning methods. Furthermore, the pre-investigation of the assessment of the thermal power plant ashes was carried out with the experiments on the ash samples which were taken from Soma thermal power plant. 1. General Information Social and technological development changes in direct proportion to the amount of energy that is consumed. As a result of the fast growth of world population, consumed energy naturally increases alongside. Especially, the fact that the population growth of Turkey is higher than the worldwide average means that the requirement for energy will increase day by day. In 2008, petroleum has the highest share in energy consumption in Turkey with 32.8%, which is followed by natural gas with 30.4%, coal with 28% and the remaining 10% is occupied by renewable resources including hydraulic (TEİAŞ 2008). A clean environment is needed for a healthy life and energy is needed for a comfortable life, which requires utilisation of resources by minimising their impact. The fact that even ashes of burned coal is usable is an important point both for economic benefit and environmental impact, and this may only be possible by fulfilment of the features of existing coal. When coal is combusted in thermal power stations, toxic trace elements in the coal like As, Cd, Ga, Ge, Pb, Sb, Se, Sn, Mo, Ti and Zn which have the potential of contaminating transfer to the waste products (cinder, ash and gas). Volatile ashes containing many poisonous elements may be collected in ash collection pools under furnaces or as piles. Because resoluble metal ions and compounds that leak from the ash pools or piles have the potential to contaminate soil, and surface and underground water, severe environmental problems may occur (Karayiğit and others 2000, Perçinel 2000, Esenlik 2005, Tuna and others 2005). When coal is combusted, toxic trace elements like arsenic (As), cadmium (Cd), lead (Pb), antimony (Sb), selenium (Se), stannum (Sn) and zinc (Zn) inside that are contaminants are transferred to waste products like cinder, ash and gases. When waste products are disposed of, contained poisonous (toxic) trace elements may be conveyed to the atmosphere, earth surface and oceans. These elements may be seriously threatening for living organisms by creating environmental, area use and health problems when the waste products are washed under rain and these elements are carried away with underground water to the soil, surface waters and underground waters (Baba 2001, Ateşok 2004). Some of the diseases occurring in time on people, who live near thermal power plants, due to the toxic elements spread in the neighbourhood are given below (Perçinel 2000): As: Anaemia, nausea, renal symptoms, ulcer, skin and pulmonary cancer, defective births. Be: Malfunction of respiration and lymph, lungs, spleen and kidneys, carcinogenic effects. Cd: Lung emphysema and fibrosis, kidney diseases, cardiovascular effects, carcinogenic effects. Hg: Nervous and kidney damages, cardiovascular effects, birth problems. Mn: Respiratory problems. Ni: Skin and intestinal diseases, carcinogenic effects. Pb: Anaemia, nervous and cardiovascular problems, delayed growth, gastric and intestinal problems, carcinogenic effects, birth problems. Se: Gastric and intestinal nausea, pulmonary and splenic damages, anaemia, cancer, teratogenic effects. V: Acute and chronic respiratory malfunction (Perçinel 2000). Radon gas forms in the area in which ashes of the thermal power plant collect (ash chambers) reach the air. Even if these ashes are buried in soil, radon gas infiltrates through the pores of the soil and blends in the air. Radon gas may transform into polonium and active lead in 3.8 days. Therefore, piles of ash emit radioactivity. Perhaps the most critical material that is disposed through the chimneys is uranium that is contained in lignite and revealed during combustion to spread around. Uranium is also a serious problem (Özyurt 2006).
1.1 Major and Trace Elements Contained in Coal and Coal Ash C, H, O, N and S contained in the structure of coal, contents of which are generally higher than 1000ppm, form the organic matrix and they are called as major elements. Al, Fe, Mg, As, Zn, Cu, F, Th, V, etc with a concentration that is generally less than 1000ppm are called as trace elements in coal (Ateşok 2004, Özyurt 2006). There are some elements in coal which are inorganic based, which may form inorganic or organometalic compounds, and which may be produced if they are at an economic level. Kin the sediments containing coal layers and in coal formations, Ge, Ga, U and Cu may be found at economic levels. Other than these, coal contains toxic trace elements like Be, Mo, V, Zn, W, Co, Cd, As, Pb, Se, and Cr which are contaminants (Kural 1998, Özyurt 2006). After combustion of coal, trace elements Pb, Cu, Zn, V, As and Th become volatile and concentrated in the furnace ash (Özyurt 2006, Riley 2008). When coals are combusted in high temperatures, their molecular structure demolished, and an important portion of Cl and F is disposed into air as gases together with smoke (Özyurt 2006). When dust coal is combusted in thermal power plants, carbon, nitrogen and sulphur contained in the coal structure oxidises and transforms into carbon oxide (CO x), nitrogen oxide (NOx) and sulphur oxide (SOx). Some water vapour forms during this transformation, too. Whereas cinder is collected under combustion furnaces, volatile ashes are caught by electro-filters and some are transported with the chimney gas. Researches show that trace elements mostly collect on volatile ashes (Karayiğit and others 2000, Özyurt 2006). In thermal power plants that use coal, combustion in the furnaces occurs at around 900-1400°C depending on the type of coal. Coal pieces heat up in the furnace, vaporizable materials convert into gases and combustion occurs. Minerals disintegrate and melt under heat, start decomposing and agglomerate (Kural 1998, Özyurt 2006). Hg, As, Se, Ni, Pb, Ce are Zn mostly related to sulphide minerals and organic substances. Combination (formation) of coal minerals or organic substances with trace elements may seriously affect vaporisation limit and consequently its ratio in the chimney gas disposed by the plant. Trace elements detected in chimney gases are mostly associated to sulphide minerals (Riley 2008, Shah and others 2008). Behaviour of trace elements in coal during and after combustion is shown in the Figure. During combustion of coal, some trace elements contained in coal like As, Cd, Ga, Ge, Pb, Sr, Mo, Zn, Ba transfer to the waste products (cinder, ash and gases). Especially volatile ashes of such wastes produce very convenient media for adhesion of elements in liquids and gases because they have clayish structure, endure high temperatures and have large surface area (Özyurt 2006). 2. Findings 2.1. Results of Particle Size Analysis Coal samples collected from the site were crashed in the laboratory type jaw crusher and roll crusher in Istanbul University, Department of Mining Engineering, Ore Preparation and Concentration Laboratory, and then dry sieve analysis were applied to determine the particle size distribution. Undersize and oversize curves were drawn according to the sieve analysis performed using 8mm, 4mm, 2mm, 1mm and 0.5mm laboratory type Retsch brand stainless steel sieves of square section and d 80 size were calculated.
Quantity (%)
Figure 1 shows undersize and oversize curves drawn according to the results of sieve analysis made after crashing of coal samples collected from Manisa-Soma region in roll crusher. The curves show that d80 size of roll crusher output of Manisa – Soma coal is 5.4mm.
Undersize Elek Altı Oversize Elek Üstü
Sieve Size (mm)
Figure 1: Manisa – Soma Sieve Analysis of the Roll Crusher Output
Quantity (%)
Figure 2 shows undersize and oversize curves of the coal samples collected from Manisa – Soma Region, which were fed to the float and sink test, according to the results of sieve analysis. The curves show that d80 size of Manisa – Soma coal that is used in float and sink test is 2100µ.
ea Undersize eü Oversize
Sieve Size (µ) Figure 2: Undersize and Oversize curves of Manisa – Soma float and sink test sample 2.2. Results of Chemical Analysis Ash samples and coal samples were brought to the Ore Preparation and Concentration Laboratory of Istanbul University, and humidity, ash, density, volatile substances, total sulphur and thermal value analyses were carried out on these samples. Evaluation and interpretation of the chemical analysis results of the coal and ash samples are summarised below: Humidity: Coal samples collected from the Turkish Coal Enterprise (TKI) Ege Lignite Enterprise (Manisa-Soma) were brought to the Ore Preparation and Concentration Laboratory of Istanbul University, and total humidity analysis was made on these samples. Analysis was made according to standard TS 690 ISO 598 (Method-C). Total humidity analysis was made after bringing samples to the laboratory in closed nylon bags without losing time. Due to the high temperature at electro-filters, humidity values of ash samples are mostly almost zero.
Humidity (%)
Figure 3 shows the percentage curve of the humidity that is lost in time when coal samples collected from Manisa-Soma region are heated in drying oven at 105ºC. It has been calculated that coal samples collected from Manisa-Soma contain 15.49% humidity. The analyses show that almost entire humidity contained in the samples may be removed in 3 hours.
0 2
1 3
Humidity loss (%)
Time (h) Figure 3: Manisa-Soma Coal – Humidity Loss % vs. Time Curve Volatile Amount: It has been observed that volatile substance content of the ash samples collected in the scope of the study is lower than the coals’. Whereas Manisa-Soma coal sample contains 31.32% volatile, it is 1.2% in the ashes of the thermal power plant. The ashes have volatile substance content because the unburned coal pieces. Ash: Ash analysis of the coal samples within the study were realised in Istanbul University, Department of Mining Engineering, Ore Preparation and Concentration Laboratory. The results of this analysis show that the coals collected from Manisa – Soma region have high ash ratio.
Theoretically, ash content of waste products formed as the result of combustion in thermal power plants (fly ash and bottom ash) must be 100%. However, depending on characteristics of combusted coal and conditions of combustion systems, it is always possible to find some unburned coal remnants in these wastes. As a matter of fact, Manisa – Soma thermal power plant ash contains 1.2% unburned pieces. Total Sulphur: Total sulphur content analysis in dry base of the the coal samples were carried out in Acme Analytical Laboratories Ltd.in Canada. Leco carbon sulphur device was used in total sulphur analysis. When coals used in Manisa-Soma thermal power plant and ashes formed by combustion are examined for total sulphur contents, it is observed that a little portion is disposed into the air by burning and the rest is concentrated in ash. Thermal Value: Upper and lower thermal values of the coal samples were analysed in the Environment and Fuel Analysis Laboratory approved by TÜRKAK belonging to Istanbul Metropolitan Municipality, Department of Environmental Protection & Development. IKA C7000 device was used for thermal value analyses. Density Analysis: Density analysis with a pycnometer was performed to determine the densities of the coal samples collected for the study and to compare optimum sorting density used in float and sink experiments. Density Analysis was performed in Istanbul University, Department of Mine Engineering, Ore Preparation and Concentration Laboratory. TS ISO 5072 Brown Coal and Lignite – Assessment of Real and Apparent Relative Density standard was utilized in the analysis method. It has been concluded that Manisa – Soma coals have high density. When density analysis and optimum sorting density used in float and sink experiments are combined, it is found that it would be fit to sort Manisa – Soma coals in the heavy medium adjusted to 1.6g/cm3. Table 1: Dry base chemical analysis results of Manisa – Soma coal samples and Manisa – Soma power plant ash samples in air
Sample Manisa-Soma Coal Manisa-Soma Ash
Amount of Volatile Substances (%) 31.32 1.2
Ash Content (%)
Total Sulphur Content (%)
Upper Thermal Value (Kcal/kg)
Lower Thermal Value (Kcal/kg)
Density (g/cm3)
35.88 98.80
0.67 1.26
2942
2761
1.553
Elementary Analyses: Elementary Analysis of the coal samples collected for the study were performed in Advanced Analysis Laboratory of Istanbul University. Dry base elementary analysis results of the analysed samples in air are given in Table 2. Table 2: Dry base elementary analysis of Manisa – Soma coal samples in air Manisa – Soma
C% 43.13
H% 2.75
N% 0.42
S% 0.28
Methods used in elementary analysis are explained below: 1DX Analysis: In 1DX analysis, 0.5g of the sample is leached with Aqua Regia heated up to 95ºC (Aqua Regia is generally obtained by mixing one third concentrated hydrochloric acid and nitric acid) and the solution placed in ICP-MS device to read the values. Elements detected by 1DX analysis are: Mo, Cu, Pb, Zn, Ni, As, Cd, Sb, Bi, Ag, Au, Hg, Tl, Se Leco TOT/C and TOT/S analysis: Total C and Total S analyses are performed with Leco carbon sulphur device. 4A Analysis: In 4A analysis, 0.2g coal and ash samples are applied lithium metaborate/tetraborate fusion and decomposed with diluted nitric acid, and then major oxides they contained were detected with the ICP-ES device. Minerals that are analysed with 4A are: SiO2, Al2O3, Fe2O3, MgO, CaO, Na2O, K2O, TiO2, P2O5, MnO, Cr2O3 4B Analysis: In 4B analysis, 0.2g coal and ash samples are applied lithium metaborate/tetraborate fusion and decomposed with diluted nitric acid, and then rare soil elements and refractor elements they contained were detected with the ICP-MS device. In addition, 0.5g samples were decomposed in royal water, and precious metals and base metals were detected with ICP-MS. Elements that are analysed with 4B are (nitric acid and ICP-MS): Ba, Be Co, Cs, Ga, Hf, Nb, Rb, Sc, Sn, Sr, Ta, Th, U, V, W, Y, Zr, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu
Elements that are analysed with 4B are (Royal Water and ICP-MS): Au, Ag, As, Bi, Cd, Cu, Hg, Mo, Ni, Pb, Sb, Se, Tl, Zn Combustible Efficiency Analysis: It is used in interpretation of combustible efficiency washing performance. Combustible efficiency has been calculated with the formulae as below: (1) Where; t: waste schist ash % f: raw coal ash % c: clean coal ash %. Float and Sink Experiments: Trace element analysis results of the products obtained by float and sink experiments are given in Attachment 1. Combustible efficiencies of samples according to the float and sink experiment results are given in Table 3. Table 1: Combustible efficiencies of the coal samples that floated and sank in float and sink experiments Float and Sink (-4mm+0,5mm) Soma +1,6 Floating Soma -1,6 Sinking Soma Fed
Quantity % 47.51 52.49 100.00
ΣC Content % 63.35 15.37 38.17
Ash % 9.90 59.40 35.88
Combustible Efficiency % 66.76 33.24 100.00
Combustible efficiencies of the flotation experiment results are given in Table 4. Table 4: Combustible efficiencies of the floating and sinking coal samples in flotation results Flotation (-0,5mm) Soma Floating Soma Sinking Soma Fed
Quantity % 61.21 38.79 100.00
ΣC Content % 31.44 30.81 31.20
Ash % 45.70 45.50 45.62
Combustible Efficiency % 61.12 38.88 100.00
Combustible efficiencies according to total coal dressing works, in which results of float and sink experiments and flotation experiments are evaluated together, are given in Table 5. Flotation Experiments: Trace element analysis results of the products obtained by flotation experiments performed in the study are given in Attachment 1 as tables. Combustible efficiencies of Manisa – Soma coal is low because it is high ash coal and lignite flotation is difficult. Table 5: Combustible efficiencies of floating and sinking coals in total after coal dressing processes Total Coal Dressing (-4mm) Soma ΣFloating Soma ΣSinking Soma ΣFed
Quantity % 50.92 49.08 100.00
ΣC Content % 53.80 18.41 36.43
Σ Ash % 45.70 45.50 45.60
Σ Combustible Efficiency % 50.83 49.17 100.00
Trace element analysis results of Manisa Soma Coal, Ash and Thermal Power Plant Ash, are given in Attachment 2.
Attachment: All Analysis Results And Averages Of Trace Elements
Float and Sink Coal Float and Sink Ashes Unsorted Coal Unsorted Coal Ashes Flotation Flotation Ashes
Soma -4+0.5mm +1.60 Floating Soma -4+0.5mm -1.60 Sinking Soma -4+0.5mm +1.60 Floating Ash Soma -4+0.5mm -1.60 Sinking Ash Soma Unsorted Coal
Method Analyte Unit MDL Coal Coal Ash Ash Coal
Soma Unsorted Coal Ashes
Ash
21.06
8.7
2.51
2.47
58.14
0.23
0.98
0.3
0.19
0.06
0.007
7.00
425.00
<1
6.8
13.8
Soma -0.5mm Floating Soma -0.5mm Sinking Soma -0.5mm Floating Ash Soma -0.5mm Sinking Ash
Coal Coal Ash Ash Coal Coal Coal Coal Coal Coal
19.49 17.99 41.37 N.A.
8.69 7.01 17.88 N.A.
2.09 2.95 4.43 N.A.
0.86 0.9 1.77 N.A.
13.15 15.33 26.5 N.A.
0.14 0.18 0.32 N.A.
0.89 0.73 1.93 N.A.
0.24 0.2 0.51 N.A.
0.07 0.08 0.16 N.A.
0.02 0.04 0.05 N.A.
0.009 0.005 0.013 N.A.
7.00 7.00 15.00 N.A. 1.00 1.00 10.00
298.00 287.00 604.00 584.00 200.00 20.00 1000.00
0.07 12.00
0.26 12.57
0.04 3.17
0.06 14.08
0.01 2.34
0.01 1.85
0.01 0.57
0.01 0.24
0.01 0.07
3.5 3.2 7.9 7.2 5.00 0.50 30.00 9.40 1.00 55.00
10.7 9.00 21.1 18.9 1.00 0.30 5.00
0.67 28.27
<1 <1 2.00 3.00 2.00 0.10 15.00 1.00 0.20 7.00
4A-4B Ga ppm 0.5 2.10 6.60 18.3 10.5 7.7
4A-4B Hf ppm 0.1 0.30 1.10 4.8 2.00 1.4
4A-4B Nb ppm 0.1 2.00 3.30 24.7 6.2 4.8
4A-4B Rb ppm 0.1 4.40 37.90 8.9 58.9 41.2
4A-4B Sn ppm 1 <1 <1 4.00 3.00 1.00
4A-4B Sr ppm 0.5 77.40 135.10 753.00 231.1 152.4
4A-4B Ta ppm 0.1 <0.1 0.20 1.00 0.4 0.4
4A-4B V ppm 8 51.00 45.00 617.00 83.00 74.00
4A-4B Th ppm 0.2 1.00 5.30 12.4 8.8 5.7
4A-4B U ppm 0.1 7.40 7.00 72.6 12.1 8.3
4A-4B W ppm 0.5 <0.5 0.80 4.9 1.4 0.7
4A-4B 4A-4B Zr Y ppm ppm 0.1 0.1 13.90 1.80 37.80 9.50 163.7 21.3 65.5 18.8 44.9 9.1
4A-4B La ppm 0.1 2.10 11.80 24.1 21.5 13.7
4A-4B Ce ppm 0.1 4.00 22.40 43.00 41.6 25.1
4A-4B Pr ppm 0.02 0.47 2.68 5.19 4.98 3.01
World Average* Lowest Highest Turkey Average Lowest Highest
Float and Sink Coal Float and Sink Ashes Unsorted Coal Unsorted Coal Ashes Flotation Flotation Ashes World Average* Lowest Highest Turkey Average Lowest Highest *(Swaine 1990)
4A-4B SiO2 % 0.01 1.71 21.81 N.A. 32.00 17.91
4A-4B AI2O3 % 0.01 1.05 6.01 N.A. 10.34 7.26
4A-4B Fe2O3 % 0.04 0.49 3.02 N.A. 5.56 2.33
4A-4B MgO % 0.01 0.23 1.11 N.A. 1.74 0.84
4A-4B CaO % 0.01 3.23 21.78 N.A. 33.79 13.08
4A-4B Na2O % 0.01 0.06 0.11 N.A. 0.26 0.11
4A-4B K2O % 0.01 0.07 0.65 N.A. 1.02 0.77
4A-4B TiO2 % 0.01 0.05 0.16 N.A. 0.28 0.2
4A-4B P2O5 % 0.01 0.03 0.08 N.A. 0.13 0.08
4A-4B MnO % 0.01 <0.01 0.04 N.A. 0.07 0.03
4A-4B Cr2O3 % 0.002 0.00 0.00 N.A. 0.008 0.005
4A-4B Sc ppm 1 <1 6.00 N.A. 11.00 6.00
4A-4B Ba ppm 1 113.00 196.00 1239.00 350.00 250.00
4A-4B Be ppm 1 <1 1.00 5.00 <1 <1
4A-4B Co ppm 0.2 1.10 2.50 16.8 4.8 3.3
4A-4B Cs ppm 0.1 2.80 9.50 5.1 14.7 10.4
Soma -4+0.5mm +1.60 Floating Soma -4+0.5mm -1.60 Sinking Soma -4+0.5mm +1.60 Floating Ash Soma -4+0.5mm -1.60 Sinking Ash Soma Unsorted Coal
Method Analyte Unit MDL Coal Coal Ash Ash Coal
Soma Unsorted Coal Ashes
Ash
8.8
1.9
6.3
48.4
2.00
426.4
0.4
75.00
8.2
11.5
0.9
68.1
13.1
18.8
35.00
4.14
Soma -0.5mm Floating Soma -0.5mm Sinking Soma -0.5mm Floating Ash Soma -0.5mm Sinking Ash
Coal Coal Ash Ash Coal Coal Coal Coal Coal Coal
8.7 7.9 19.00 15.1 5.00 1.00 20.00
1.6 1.4 3.5 2.8 1.00 0.40 5.00
5.00 4.4 10.5 8.9 5.00 1.00 20.00
47.6 39.4 100.9 79.1 40.00 2.00 50.00
<1 2.00 3.00 2.00 2.00 1.00 10.00
179.2 180.7 362.1 350.8 200.00 15.00 500.00
0.4 0.3 0.8 0.7 0.20 0.10 1.00
72.00 66.00 155.00 132.00 40.00 2.00 100.00 87.00 6.00 287.00
7.1 5.7 14.3 12.6 4.00 0.50 10.00 6.00 1.00 29.00
9.00 8.6 17.9 17.7 2.00 0.50 10.00 13.00 0.40 132.00
0.6 0.8 1.6 1.3 1.00 0.50 5.00
53.7 43.9 118.2 93.2 50.00 5.00 200.00
9.5 9.2 19.9 19.00
14.8 13.00 30.2 27.6 10.00 1.00 40.00
28.3 24.1 56.9 51.1 20.00 2.00 70.00
3.36 2.94 6.81 6.06
2.00 50.00
1.00 10.00
Float and Sink Coal Float and Sink Ashes Unsorted Coal Unsorted Coal Ashes Flotation Flotation Ashes
Soma -4+0.5mm +1.60 Floating Soma -4+0.5mm -1.60 Sinking Soma -4+0.5mm +1.60 Floating Ash Soma -4+0.5mm -1.60 Sinking Ash Soma Unsorted Coal
Method Analyte Unit MDL Coal Coal Ash Ash Coal
Soma Unsorted Coal Ashes
Ash
16.4
2.85
0.62
2.33
0.39
2.16
0.43
1.32
0.2
1.34
0.19
0.97
0.73
5.2
99.81
Soma -0.5mm Floating Soma -0.5mm Sinking Soma -0.5mm Floating Ash Soma -0.5mm Sinking Ash
Coal Coal Ash Ash Coal Coal Coal Coal Coal Coal
12.4 10.5 25.00 23.2 10.00 3.00 30.00
2.08 1.91 4.64 4.25 2.00 0.50 6.00
0.49 0.46 0.99 0.99 0.50 0.10 2.00
1.84 1.63 3.77 3.67
0.3 0.27 0.61 0.57
1.64 1.57 3.49 3.31
0.33 0.31 0.67 0.63
0.99 0.98 2.12 1.97
0.16 0.15 0.32 0.31
0.16 0.15 0.33 0.3
31.44 30.81 1.49 N.A.
0.76 0.81 1.52 N.A.
54.3 54.5 4.8 N.A.
99.9 99.88 99.77 N.A.
0.40 4.00
0.10 1.00
0.50 4.00
0.10 2.00
0.50 3.00
0.50 3.00
1.00 0.93 2.18 2.03 1.00 0.30 3.00
1DX Mo ppm 0.1 0.70 0.50 12.00 1.4 0.6
1DX Cu ppm 0.1 9.50 15.00 61.00 21.8 18.9
1DX Pb ppm 0.1 3.90 8.90 5.5 10.4 8.6
1DX As ppm 0.5 20.90 66.60 219.00 102.3 52.00
1DX Cd ppm 0.1 0.10 0.30 0.9 0.5 0.2
1DX Sb ppm 0.1 0.10 <0.1 1.7 0.2 0.2
1DX Bi ppm 0.1 <0.1 0.10 0.4 0.3 0.2
1DX Ag ppm 0.1 <0.1 <0.1 <0.1 <0.1 <0.1
1DX Au ppb 0.5 <0.5 0.70 8.3 <0.5 <0.5
1DX Hg ppm 0.01 0.06 0.06 <0.01 <0.01 0.07
1DX Tl ppm 0.1 0.70 0.90 <0.1 0.1 <0.1
World Average* Lowest Highest Turkey Average Lowest Highest
Float and Sink Coal Float and Sink Ashes Unsorted Coal Unsorted Coal Ashes Flotation Flotation Ashes World Average* Lowest Highest Turkey Average Lowest Highest *(Swaine 1990)
4A-4B Nd ppm 0.3 1.80 10.00 20.4 20.00 11.6
4A-4B Sm ppm 0.05 0.32 2.02 3.78 3.76 2.07
4A-4B Eu ppm 0.02 0.07 0.48 0.92 0.89 0.47
4A-4B Gd ppm 0.05 0.31 1.83 3.72 3.29 1.67
4A-4B Tb ppm 0.01 0.06 0.28 0.63 0.55 0.28
4A-4B Dy ppm 0.05 0.27 1.50 3.54 3.02 1.48
4A-4B Ho ppm 0.02 0.07 0.31 0.65 0.62 0.3
4A-4B Er ppm 0.03 0.18 0.85 2.12 1.81 0.93
4A-4B Tm ppm 0.01 0.03 0.14 0.27 0.3 0.14
4A-4B Yb ppm 0.05 0.21 0.96 2.13 1.82 0.95
4A-4B Lu ppm 0.01 0.03 0.15 0.36 0.28 0.15
2ALeco TOT/S % 0.02 0.70 0.65 N.A. 1.18 0.75
0.03 1.00
Soma -4+0.5mm +1.60 Floating Soma -4+0.5mm -1.60 Sinking Soma -4+0.5mm +1.60 Floating Ash Soma -4+0.5mm -1.60 Sinking Ash Soma Unsorted Coal
Method Analyte Unit MDL Coal Coal Ash Ash Coal
Soma Unsorted Coal Ashes
Ash
1.6
20.8
16.4
58.9
0.3
0.1
0.2
<0.1
1.00
<0.01
Soma -0.5mm Floating Soma -0.5mm Sinking Soma -0.5mm Floating Ash Soma -0.5mm Sinking Ash
Coal Coal Ash Ash Coal Coal Coal Coal Coal Coal
0.7 0.9 2.1 2.2 3.00 0.10 10.00
21.9 21.1 45.5 42.2 15.00 0.50 50.00
10.6 10.4 16.8 16.9 20.00 2.00 80.00 12.00 0.10 286.00
46.9 56.1 99.5 114.4 10.00 0.50 80.00 53.00 2.00 686.00
0.3 0.2 <0.1 0.1 0.50 0.10 3.00 1.00 0.01 45.00
0.2 0.2 0.3 0.3 1.50 0.10 10.00
0.2 0.2 <0.1 <0.1
<0.1 <0.1 <0.1 <0.1
<0.5 <0.5 0.9 <0.5
0.08 0.1 <0.01 <0.01 0.02 1.00 0.10 0.03 0.70
0.2 1
2.00 20.00
2A Leco TOT/C % 0.02 63.35 15.37 N.A. 0.22 33.93
1DX Se ppm 0.5
1DX Ni ppm 0.1 0.80 2.80 2.1 3.1 0.7
7.60 7.10 54.2 15.2 9.4
0.2
1.1
21.5
<0.1 <0.1 0.5 0.3
0.6 0.6 1.7 1.5 1 0.2 10 2 0.1 26
12.1 13.2 29.7 25.3 20.00 0.50 50.00 126.00 3.00 1700.00
4A-4B LOI % -5.1 90.10 40.60 N.A. 7.3 57.3
4A-4B Sum % 0.01 97.00 95.41 N.A. 92.48 99.89
Attachment 2: Trace Element Analysis Results Of Manisa Soma Coal, Ash And Thermal Power Plant Ash
Soma (-4mm) Coal Soma (-4mm) Coal Ash Soma Thermal Power Plant Ash
Method 4A-4B Analyte SiO2 Unit % MDL 0.01 17.91 Coal 21.06 Ash 44.97 Ash
4A-4B 4A-4B 4A-4B Al2O3 Al Fe2O3 Fe MgO % (%) % (%) % 0.01 0.04 0.01 8.37 7.26 3.84 2.33 1.63 0.84 9.85 8.70 4.61 2.51 1.76 2.47 21.02 21.25 11.25 4.10 2.87 1.67
Soma (-4mm) Coal Soma (-4mm) Coal Ash Soma Thermal Power Plant Ash
Method 4A-4B Analyte TiO2 Unit % MDL 0.01 0.20 Coal 0.30 Ash 0.73 Ash
Soma (-4mm) Coal Soma (-4mm) Coal Ash Soma Thermal Power Plant Ash
Method Analyte Unit MDL Coal Ash Ash
4A-4B 4A-4B Ga Hf ppm ppm 0.5 0.1 7.70 1.40 8.80 1.90 22.60 4.00
4A-4B 4A-4B 4A-4B Nb Rb Sn ppm ppm ppm 0.1 0.1 1 4.80 41.20 1.00 6.30 48.40 2.00 20.30 94.80 3.00
4A-4B 4A-4B 4A-4B Sr Ta Th ppm ppm ppm 0.5 0.1 0.2 173.60 0.40 5.70 426.40 0.40 8.20 381.10 1.50 26.20
4A-4B 4A-4B 4A-4B 4A-4B U V W Zr ppm ppm ppm ppm 0.1 8 0.5 0.1 8.30 74.00 0.70 44.90 11.50 75.00 0.90 68.10 26.50 275.00 3.80 167.60
Soma (-4mm) Coal Soma (-4mm) Coal Ash Soma Thermal Power Plant Ash
Method Analyte Unit MDL Coal Ash Ash
4A-4B 4A-4B Pr Nd ppm ppm 0.02 0.3 3.01 11.60 4.14 16.40 10.58 39.50
4A-4B Sm ppm 0.05 2.07 2.85 7.30
4A-4B 4A-4B 4A-4B Tb Dy Ho ppm ppm ppm 0.01 0.05 0.02 0.28 1.48 0.30 0.39 2.16 0.43 0.96 5.26 1.11
4A-4B 4A-4B 4A-4B 4A-4B Er Tm Yb Lu ppm ppm ppm ppm 0.03 0.01 0.05 0.01 0.93 0.14 173.60 0.15 1.32 0.20 1.34 0.19 3.05 0.46 3.34 0.48
Soma (-4mm) Coal Soma (-4mm) Coal Ash Soma Thermal Power Plant Ash
Method Analyte Unit MDL Coal Ash Ash
1DX Pb ppm 0.1 8.60 16.40 25.00
Si (%)
4A-4B 4A-4B P2O5 P MnO % (%) % 0.01 0.01 0.12 0.08 0.03 0.030 0.18 0.19 0.08 0.060 0.44 0.22 0.10 0.04
Ti (%)
1DX Zn ppm 1 28.00 38.00 81.00
1DX As ppm 0.5 52.00 58.90 144.90
4A-4B Cr2O3 % 0.002 0.023 0.005 0.046 0.007 0.031 0.02 Mn (%)
4A-4B 4A-4B Eu Gd ppm ppm 0.02 0.05 0.47 1.67 0.62 2.33 1.53 6.52
1DX Cd ppm 0.1 0.20 0.30 1.60
1DX Sb ppm 0.1 0.20 0.10 0.40
1DX Bi ppm 0.1 0.20 0.20 0.70
Mg (%) 0.51 1.49 1.01
Cr (%) 0.003 0.005 0.010
4A-4B CaO % 0.01 13.08 58.14 23.67 4A-4B Ni ppm 20 <20 27.00 37.40
Ca (%) 9.35 41.56 16.92 4A-4B Sc ppm 1 6.00 7.00 15.00
4A-4B Na2O % 0.01 0.11 0.23 0.54
4A-4B K2O % 0.01 0.08 0.77 0.17 0.98 0.40 1.29
Na (%)
K (%) 0.64 0.81 1.07
4A-4B 4A-4B 4A-4B 4A-4B Ba Be Co Cs ppm ppm ppm ppm 1 1 0.2 0.1 250.00 <1 3.30 10.40 425.00 <1 6.80 13.80 893.00 3.00 15.20 25.10
1DX 1DX 1DX 1DX 1DX 2ALeco 2ALeco Ag Au Hg Tl Se TOT/C TOT/S ppm ppb ppm ppm ppm % % 0.1 0.5 0.01 0.1 0.5 0.02 0.02 <0.1 <0.5 0.07 <0.1 0.70 33.93 0.75 <0.1 1.00 <0.01 0.20 1.10 0.97 0.73 <0.1 <0.5 0.02 0.30 2.20 0.71 1.26
4A-4B 4A-4B La Ce ppm ppm 0.1 0.1 13.70 25.10 18.80 35.00 50.90 95.40 1DX Mo ppm 0.1 0.60 1.60 2.40
Ash % 42.70 94.80 98.80
1DX Cu ppm 0.1 18.90 20.80 29.20
Conclusions Because of Manisa-Soma lignite is low in thermal value, they can only be used in thermal power plants. It has been observed that trace element sedimentations settling in coal during geological formation of the area is higher than worldwide coal average. In the Float and Sink Experiments performed on Manisa-Soma coal, which has a relative density of 1.553g/cm3, ZnCl2 solution of 1.6g/cm3 density has been used. It has been calculated that it is very easy to sort at this density, 47.51% of the coal fed with 35.88% ash content floats, and floating coal contains 9.90% ash and sinking coal contains 59.40% ash. When sorting is made at 1.6g/cm 3 density, combustible efficiency of floating coals has been calculated as 66.76%. As a result of flotation experiments applied on Manisa-Soma coal, it has been found that 61.21% of the coal fed with 45.62% ash content floats, and floating coal contains 45.70% ash and sinking coal contains 45.50% ash. At the end of flotation, it has been calculated that combustible efficiency of floating coal is 61.12%. It has been aimed to obtain a total coal preparation conclusion by combining the results of float and sink experiment and flotation experiment. According to these results, total floating ratio of the coal with 45.60% total ash content is 50.92% and ash content is 45.70%. Total combustible efficiency has been calculated to be 50.83%. According to the major and trace element results of the Float and Sink Experiment made on Manisa-Soma coal, whereas radioactive element content of Th is 3.26ppm in the feeding coal, it is 1.00ppm in floating coal and 5.30ppm in sinking coal. Furthermore, when post-combustion ashes of these coals are examined, the value found in Floating Coal ash is 12.40ppm, in Sinking Coal ash 8.80ppm and in Unsorted Coal ash 8.20ppm respectively. This shows that Th element collects in the sinking part after coal dressing, and gets concentrated in the ash after combustion. In the same way, when the samples are analysed in respect to U element, the values are 7.19ppm in Unsorted Coal, 7.40ppm in the floating part and 7.00ppm in the sinking part after float and sink experiment. When they are analysed with respect to ash, the value is 11.50ppm in the feeding coal ash, and they concentrate in Floating Coal ash as 72.60ppm and 12.1ppm in Sinking Coal ash. When samples are analysed with respect to air polluting elements, it was observed that As, Co and Mn concentrated in floating coals, Be, Cd, Hg and Ni concentrated in the ashes of these coals, and Se concentrated in sinking coal and its ash. Acknowledgement A part of this work was supported by Scientific Research Projects Coordination Unit of Istanbul University. Project number T-2324 References Ateşok. G., 2004. Coal Preparation and Technology. Publications of the Foundation for Development of Mining,. 376 s.. ISBN-9757946-22-2. İstanbul, Turkey. Baba. A., 2001. Effect of Yatağan (Muğla) Thermal Power Plant Waste Product Storage Area on Underground Water, Dokuz Eylül University Journal of Geology Engineering Vol: 25.2. İzmir, Turkey. Demir. İ., 2009. Investigation of Evaluation Possibilities of Turkish Coals In Terms of Their Trace Element Contents By Use of Coal Preparation Methods, Istanbul University, Institute of Science, Mining Engineering Department, Master’s Thesis,. 269 s. İstanbul, Turkey. Esenlik. S., 2005. Mineralogy, Petrography and Element Contents of Coal Combusted in Orhaneli Thermal Power Plant and Waste Products, Bursa, Turkey. Hacettepe University, Institute of Sciences, Geological Engineering Department, Master’s Thesis.116 s. Ankara, Turkey. Karayiğit. A. I., Gayer. R. A., Querol. X., Onacak. T., 2000. Contents of major and trace elements in feed coals from Turkish coalfired power plants. International. Journal of Coal Geology. Vol.44. p.169-184. Kural O.. 1998. (Editor) Properties, Technology and Environmental Impacts of Coal. ITU Faculty of Mining. 785 s. İstanbul, Turkey. Özyurt. Z., 2006. Environmental Effect of the Trace Elements in Thermal Power Plants’ Waste Products. Eskişehir Osmangazi University, Mining Engineering Department, Master’s Thesis. 75 s. Eskişehir, Turkey.
Perçinel. S., 2000. Effect of Coal Use in Thermal Power Plants on Human Health. Proceedings of the 12th Turkish Coal Congress. 2326 May 2000. Zonguldak, Turkey. Riley. K., 2008. Analysing for Trace Elements. http://www.ccsd.biz/products/traceelementdb.cfm. Shah. P., Strezov. V., Prınce. K., Nelson. P. F., 2008. Speciation of As. Cr. Se and Hg Under Coal Fired Power Station Conditions. Elsevier Fuel. Vol.87. p.1859-1869. Swaine. D. J., 1990. Trace Elements In Coal. Butterwarh. 278 p. London. TEİAŞ. 2008. Web Page of Türkiye Elektrik İletim A.Ş. http://www.teias.gov.tr / (Visiting Date: 12.09.2008). TSE. 1992. TS 10091/April 1992. Mineral Coal Foamy Flotation Experiment, Section 1, Laboratory Process, Turkish Standards Institute, Ankara, Turkey. TSE. 1999. TS ISO 5072/April 1999. Brown Coal and Lignite – Assessment of Real and Apparent Relative Density, Turkish Standards Institute, Ankara, Turkey.. TSE. 2002a. TS 690 ISO 589/March 2002. Assessment of Total Humidity in Mineral Coal, Turkish Standards Institute, Ankara, Turkey. TSE. 2002b. TS 711 ISO 562/March 2002. Assessment of Volatile Substances in Mineral Coal and Coke, Turkish Standards Institute, Ankara, Turkey. TSE. 2003. TS 3037 ISO 7936/March 2003. Definition and Display of Mineral Coal Float and Sink Experiment Requirements, General Definitions for Device and Processes, Turkish Standards Institute, Ankara, Turkey. TSE. 2006. TS ISO 1171+Tech Cor 1/November 2006. Solid Mineral Fuels – Quantity of Ash, Turkish Standards Institute, Ankara, Turkey. Tuna. A. L., Yağmur. B.. Hakerlerler. H.. Kılınç. R.. Yokaş. İ.. Bürün. B.. 2005. Research on Pollution Caused by the Thermal Power Plants in Muğla Region. Muğla University, Scientific Research Projects Final Report. 79 s. Muğla, Turkey.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
6. COAL SCIENCE, COAL PREPARATION EFFECT OF TRIBOELECTROSTATIC SEPARATION ON COAL DESULFURIZATION AND DEASHING Byoung-Gon Kim, Principal researcher, Korea Institute of Geoscience and Mineral Resources (KIGAM) [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Ho-Seok Jeon, PrincipalFossilfossil researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Sang-Ho Back, Graduate student, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Chong-Lyuck Park, Post doctoral researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea]
Abstract: Coal deposit among the fossil fuels is very plentiful in natural resources and has high economical efficiency but its application techniques are very inconvenient. Coal is not an expensive mineral compare with other natural energy resources, therefore a various researches for economical coal pre-preparation technique have been developed. Triboelectrostatic separation cost less in operation process than the others separation methods for coal purification. A principle of this separation process is use a difference of work function when mixed particles crush a surface of the other material, at this time, surface of particles is charged by positive and negative. A coal particle that have a small work function is moved to the negative electric rode because the particle is positively charged by loss of an electron, and an ash particle that has a big work function is moved to the positive rode because the particle is negatively charged by getting an electron when charged particles pass through electronic magnetic field by high current. In this study, we made a bench-scale's triboseparation equipment using electrostatic technology, and got an optimum condition of various factors for increasing recovery rate and purification in separation. Also, we used a copper that has middle work function between coal and ash by pipe's quality. An optimum condition in coal separation by this process is particle size of 20mesh, flow rate of air is 3kg/cm2, electric voltage of 30kV and using a coated screen rode by film, also rejection rate of ash and sulfur content is very different in each samples. Therefore we got a clean coal that recovery rate is 68.10%, rejection rate of ash and sulfur content is 31.23% and 28.33%.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
6. COAL SCIENCE, COAL PREPARATION ROMOVE OF ASH AND SULFUR MINERALS FROM COAL BY TRIBOELECTROSTATIC SEPARATION Ho-Seok Jeon, Senior researcher, Korea Institute of Geoscience and Mineral Resources (KIGAM) [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Byoung-Gon Kim, Senior researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Woo-Zin Choi, Professor, Dept. of Environmental Engineering, The University of Suwon [[email protected], San 2-2 Wau-ri, Bongdam-eup, Hwaseong-si, Gyeonggi-do, Republic of Korea]
Abstract: Electrostatic separation technique has received much attention in recent years as a method of removing pyrite sulfur and ash-forming minerals from coal. Recent efforts to use electrostatic methods for the cleaning of coal have focused on the use of triboelectric charging of the coal particles followed by separation of the dry charged material in a static high voltage field. These studies have demonstrated that it is feasible to use the triboelectric charging to achieve high separation efficiency, but the application of this technique at a commercial scale has yet to be demonstrated. In triboelectrostatic separation, the coal is entrained in a stream of nitrogen carrier gas and then collides with the surface of the charger under turbulent conditions. The coal and mineral matter acquire opposite charges (coal +, mineral -) and can be separated by a strong electric field according to polarity. In this study, triboelectrostatic separation process has been proven in bench-scale tests to be capable of better than 50% removal of pyritic sulfur and greater than 60% reduction of ash for coal from China.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
6. COAL SCIENCE, COAL PREPARATION RECOVERY OF VALUABLE METALLIC AND NON-METALLIC MINERALS FROM COAL MINE WASTES Sang-Bae Kim, Principal researcher, Korea Institute of Geoscience and Mineral Resources (KIGAM) [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Soo-Bok Jeong, Principal researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Ho-Seok Jeon, Principal researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Chong-Lyuck Park, Post doctoral researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea]
Abstract: Disposal methods of coal mine wastes were mainly suggested a landfill and a revegetation in Korea. Tailings and mine wastes generated from coal mines (mainly anthracite in Korea) contain at least a little iron sulfide. These wastes, like abandoned metal mines, are often conspicuous because of iron oxide staining and the acidity of the water that drains them. Hazards to adjacent areas are also associated with subsidence problems due to the abandoned space, and debris from mining and milling operations that do not involve sulfide minerals. In this investigation, reprocessing tests were carried out on various types of coal mine wastes in order to recover valuable metallic and non-metallic resources and to minimize long-term environmental damages. Reprocessing tests were carried out by employing various separation techniques such as gravity and magnetic separation, classification, etc. Tailings from coal preparation plants were crushed, pulverized and screened to obtain below 1.0 mm size fraction which was evaluated to utilize as ceramic materials. Especially, silica minerals were recovered from the coal mine wastes by applying the selective crushing and grinding, and screening methods, processes were also developed to recover valuable metals and non-metallic resources, which will be used for aggregates, ceramic materials, foundry sands, and other construction applications. The processes we developed should meet the major conditions such as process with large scale treatment, easy operation and low maintenance cost, no secondary pollution. As a result of this study, fundamental information is now available for future pilot and commercial scale testing of the proposed processes, and will contribute to the utilization of coal mine wastes and the improvement of the environment in old mine regions.
Manuscript Not AVAILABLE
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission
6. COAL SCIENCE, COAL PREPARATION MANUFACTURE OF FIRED CLAY BRICK FROM COAL-PREPARATION REFUSE ANS ITS CHARACTERISTICS Soo-Bok Jeong, Principal Researcher, Korea Institute of Geoscience and Mineral Resources (KIGAM) [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Ho-Seok Jeon, Principal Researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Chong-Lyuck Park, Post doctoral researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea] Byoung-Gon Kim, Senior researcher, Korea Institute of Geoscience and Mineral Resources [[email protected], KIGAM, 92 Gwahang-no, Yuseong-gu, Daejeon 305-350, Republic of Korea]
Abstract: Large amount of coal-preparation refuse (CPR) has been discharged from coal mines all over the world. This disposal of CPR is somewhat toxic to living organism since though CPR contains small amounts of heavy metals, sulfur, and organic materials, they contaminate the surface and ground water through the leachate generated at the dump-sites. Therefore, it is important to find a possibility for effective utilizing CPR generated from coal mines in terms of environmental and economic points of view. CPR generated from anthracite contains the SiO2/Al2O3 weight ratio of 2.0-4.0, which is nearly similar to that contained in fired clay bricks. In this study, a manufacture of fired clay bricks using CPR and its characteristics were investigated. The results indicated that firing shrinkage ratio and compressive strength decrease with increasing the addition amount of CPR, while water absorption ratio is almost constant under the entire addition amount ranges of that. It was also found that the properties f plasticity, firing shrinkage, water adsorption, and compressive strength of bricks manufactured using CPR satisfied them of the 1st grade clay bricks of Korea Standard L (ceramics) 4201. Therefore, it is expected that the use of CPR as a raw material to make fired clay bricks can save energy and decrease pollution.
UPGRADING OF LOW RANK COAL BY HYBRID FLASH DRYER Sangdo. Kim, Hokyung Choi, Kyoungsoo Lim, Sangyoung Lee, Sihyun Lee* Clean Fossil Energy Research Center, Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejeon, 305-343, Korea *E-mail : [email protected] Abstract: Deposit of low rank coal (LRC) represents nearly half of the estimated coal resources in the world. But, these cannot be economically utilized due to the characteristics of high moisture contents (30% or more), low heating value (5,000kcal/kg or less) and spontaneous combustion ability during transportation and storage. To avoid such problems, development of upgrading technology of low rank coal is needed. In this study, hybrid flash dryer system for upgrading coal was tested in the laboratory scale of 5kg/hr apparatus. In a flash dryer, a feed material is rapidly dried by direct contact with hot air while being transported by the air stream. Coal samples obtained from Mongolian coal (lignite) that has moisture contents of 29.74%, volatile matter contents of 27.83%, ash contents of 10.51% and heating value of 4,270kcal/kg (as received basis). Drying air is heated with a burner and inlet velocity of hot gas is in the range of 10-20m/sec, inlet air temperature ranges of 300-600C, particle size of raw coal is in the range of 0.1-2mm. Moisture content of raw coal was rapidly decreased to lower than 7% with increasing the inlet air temperature and the heating value was increased to more than 5,550kcal/kg. Also, the moisture contents decreased with decreasing gas inlet velocities.
1. Introduction Lignite and subbituminous coal, which are classified as low rank coal (LRC), constitute over 50% of coal deposits worldwide. However, the amount of LRC that is currently being produced resches as little as 15% of total coal usage. According to a report from the World Energy Council (WEC), as of 2007, the R/P(reserve/production) ratio for subbituminous coal and lignite were 416 years and 170 years, respectively, which is higher than high rank coals such as bituminous coal and anthracite coal, which have an R/P ratio of 85 years. Moreover, the R/P ratio of high rank coals rapidly decreases every year (1). Hence, studies that can make use of LRC with high R/P ratio are required. However, LRC has several disadvantages when it is used as a coking coal. The moisture content is very high—30–70 wt%—and its heating value is low—below 4,500 kcal/kg. Moreover, the oxygen content is especially high in LRC, making it difficult to storage and transportation due to the high risk of spontaneous combustion. Large carbon dioxide emission from the combustion of LRC is also problematic, causing environmental hazards. In order to solution these problems, there are currently a number of studies that focus on upgrading LRC. Technologies such as Binderless Coal Briquetting (BCB)(2), by Australian company White Energy, Upgrading Brown Coal (UBC)(3), by Kobe Steel in Japan, and Integrated Drying Gasification Combined Cycle (IDGCC)(4), by Australian company HRL, are demonstrated or commercialized. Among these, BCB is a technology that dries coals using the principle of flash drying. Flash drying refers to the drying of wet material with direct contact with hot gas; the residence time is very short, ranging from 0.5–10 seconds. Some of the advantages of flash drying include a short drying time, simple structure, ease of operation and control, and low maintenance cost. It has been widely used for drying wet
materials(5). But, the wet materials should have a relatively rapid drying speed because drying time is short in flash drying. A particle size of 1–2 mm is appropriate (5). The moisture of coal is divided into surface moisture and inherent moisture. While surface moisture dries easily, inherent moisture is difficult dry according to exist moisture inside the coal pore. Mujumdar argues that a single process is not appropriate for drying low rank coal(6). Hence, it is necessary to develop technology for a hybrid drying process with simple equipment and the ability to enhance the drying performance. This paper presents an experimental study examining the upgrading properties of LRC through the use of a hybrid flash dryer that combines a drying riser tube and cyclone dryer. The experiment, which was conducted with 5 kg/hr equipment, examines the properties of a hybrid flash dryer by exploring the effects of several variables, such as inlet temperature, particle size, and inlet velocity.
2. Experiment The following three types of lignite were used for the experiment: Mongolia coal from China Inner Mongolia, Samwha coal, and ECO coal from Indonesia. The results of proximate analysis, ultimate analysis, and heating value analysis are summarized in Table 1. The moisture of the coals was 29.74 wt%, 34.27 wt%, and 32.98 wt% for the Mongolia coal, Samwha coal, and ECO coal, respectively. The heating value of the coal was about 4,200 kcal/kg. Fig. 1 shows the experimental hybrid flash dryer equipment. The experimental equipment consists of seven parts including an LPG burner for generating hot gas, high-moisture coal feeder, drying riser tube, cyclone dryer, bag filter, and dried coal storage tank. Cyclone plays a role in separating the dried particles and gas and in enhancing the drying of the particles. As the residence time of the particles that flow into the cyclone increases due to the swirling flow, the drying rate consequently also increases. Hence, this study uses a name cyclone dryer instead of a viewing cyclone as a simple separator. The hybrid flash dryer is a device in which the drying riser tube and cyclone dryer are linked. The diameter and height of the drying riser tube in the hybrid flash dryer are 0.04 m and 5 m, respectively. The diameter and height of the cyclone dryer are 0.24 m and 0.8 m, respectively. The experiment was conducted as follows: First, hot gas was generated using an LPG burner. When it reached the necessary temperature, high-moisture coal was supplied. Coals containing a large amount of moisture were dried as they were transferred through the drying riser tube and cyclone dryer. The dried coal particles were separated by the cyclone dryer to be stored in the storage tank. The temperature of inlet hot gas was set at a range of 400–600 °C. The inlet temperature was measured right before feeding of coal. The inlet velocity of gas passing the drying riser tube was set at 10–20 m/sec. The particle size of the wet coal was prepared by grinding and sieving. A range of particle size was used 100–2,800 ㎛, according to the experimental conditions. After the experiment, the coal was sampled at the spot to be measured for its moisture content and heating value.
Coal Name
Table 1. Analyses of coal sample. Monggolia coal Samhwa coaal
ECO O coal
Proximate analyssis (wt%) Moisture
29.74
34.27
322.98
V Volatile matter
27.83
33.64
366.42
Ash
10.51
2.01
44.69
F Fixed carbon
31.92
29.99
225.9
Ulttimate analysiis (wt%) Carbon
622.73
70.5
668.8
Hydrogen
44.11
5.14
5.33
Nitrogen
00.95
0.99
00.4
Sulfur
00.28
0.03
00.12
Oxygen
21.42
21.33
200.66
Heatinng value analyssis (kcal/kg) Ass received basee
4,270
4,192
4,,196
Dry base
5,730
6,070
5,,880
Fig. 1. Experimeental hybrid fflash dryer eqquipment
3. Results and Discussion Fig. 2 shows the simulation results of the flash dryer; the simulation was conducted in two cases: one case with only a drying riser tube and another case with an additional cyclone in the rear of the drying riser tube. When the particle size was 1,000 ㎛, the moisture removal was approximately 35 wt% without the cyclone, while it increased to 65 wt% when the cyclone was added. When the particle size was 500 ㎛, the moisture removal was about 90 wt% without the cyclone, and it increased to 100 wt% when the cyclone was added. The reason behind this faster moisture removal is the increased residence time of the particles that pass through the cyclone. The residence time of the particles increased due to the swirling flow coming to the cyclone, which provides an additional drying effect. Hence, the cyclone plays an important role in the drying ability of the flash dryer. In other words, the cyclone does not only function as a separator, but also plays a role as a dryer.
110 100
0.20 90 80
Moisture content (Without cyclone) Moisture removal (Without cyclone) Moisture removal (With cyclone) Moisture content (With cyclone)
0.15
0.10
70
Initial absolute humidity : 0.007 kg/kgdry gas Initial gas temperaute : 673 K Initial coal moisture content : 0.274 kg/kg coal
60 50
Moisture Removal (%)
Final Moisture Content (kg water/kg coal)
0.25
0.05 40 0.00 0
200
400
600
800
1000
30 1200
Coal Particle Diameter (m)
Fig. 2. The moisture content and moisture removal efficiency with and without the cyclone Table 2 summarizes the moisture content of the dried Mongolia coal. The initial experiment was conducted with a range of 100–2000 ㎛, which was divided into four groups such as 100-300 ㎛, 300500 ㎛, 500-1000 ㎛, over 1,000 ㎛. As the inlet temperature of the dried coal increased from 400 °C to 600 °C, the moisture content decreased from 6.99 wt% to 6.22 wt%. The moisture content was measured after sieving the dried coals according to their dried particle size range. The moisture content decrease with decreasing dried coal size.
Table 2. M Moisture conteent of dried coal c Inlet gas ttemperature ((°C) 4400 500 Coal pparticle size ((㎛)
600
100–22000
66.99
6.64
6.22
100––300
66.57
5.74
4.92
300––500
77.02
6.19
5.24
500–11000
88.21
6.77
5.61
Over 1000
99.44
8.11
7.53
Fig. 3 shows thee moisture reemoval rate and heatingg value of dried coal aaccording too coal size distributiion. The ressults are obttained after sieving the coals accorrding to theeir particle ssize before conductiing the experriment. The iinlet temperaature of the gas was set att 400 °C. Thee moisture reemoval rate changed according too the coal sizze. When thee particle sizee ranged 1000–300 ㎛, thee moisture reemoval rate h as 91.6 wt%. w The mooisture removval rate decreeased with laarger particle sizes. Withinn the range was as high of 500–11,000 ㎛, the moisture rem moval rate was w 81 wt%. When the paarticle size w was between 1000–2800 ㎛, the moisture rem moval rate decreased d to 57.7 wt%. Meanwhile, the heatingg value show wed similar with a particlle smaller thaan 1000 ㎛, at 5,860–5,872 kcal/kg. H However, in the case of results foor the coals w coals larrger than 10000 ㎛, the heeating value dropped to 55,130 kcal/kgg. This can bbe attributed to the fact that smaller particles have a higheer surface areea by unit weeight of solidd than larger particles, favvouring the rate of hheat and masss transfer beetween solid and gas. Heence, the moisture removval rate and tthe heating value inccreases with decreasing coal size.
Figg. 3. The moiisture removaal and heating value of diifferent particcle sizes
Fig. 4 shows the moisture removal rate aand heating value resultts of the thrree dried coaals. In this experimeent, a raw cooal with a parrticle size rannge 100–2,0000 ㎛ was used. After thhe experimennt, the dried coals weere sieved intto four stagees such as 1000-300 ㎛, 3000-500 ㎛, 500-1000 ㎛, over 1,000 ㎛.. ㎛ For all of the drried coals, thhe moisture ddrying rate inncreased witth a smaller pparticle size.. In the Monngolia coal, the moissture removaal rate increaased from 677 wt% to 788 wt% as thee particle sizze decreased from over 1000 ㎛ to 100–3000 ㎛. In Sam mwha coal annd ECO coaal, the moistuure removal rate of the dried coal showed a similar tenddency, althouugh the moissture removall rate was higgher than thaat of the driedd Mongolia coal as pparticle size ddecreased. Meanwhile, M thhe heating vaalue of the M Mongolia coaal tended to increase as the particcle size decreeased. In parrticular, for thhe coals withh a particle siize larger thaan 1,000 ㎛, tthe heating value drastically deccreased to 5,100 kcal/kg. Hence, usinng a particle size smallerr than 1,000 ㎛ will be more effficient in the heating value of Mongolia coal. In Saamwha coal aand ECO coaal, the heatingg value did not channge much witth the changees in differentt particle sizee. Figs 3 and 4 indicaate that it is aappropriate too conduct thee experimentt after dividinng the particlles in order to obtainn a high moiisture removaal rate. How wever, it is effficient to peerform the exxperiment wiith a broad range off particle sizes in order to increase the overall dryinng rate.
Fig. 44. The moistuure removal rate and heatting value at different parrticle sizes Fig. 5 shows the moisture m rem moval rate andd heating vallue according to inlet tem mperature annd coal size when Moongolia coal was used forr the experim ment. The mooisture removval rate graduually increaseed with the increasinng inlet tempperature. Wheen the inlet teemperature w was 500 °C, tthe moisture removal ratee increased from 73 wt% to 81 w wt% as the drried coal parrticle size deccreased from m larger than 1000㎛ to 1100–300㎛. When thhe inlet tempeerature was 6600 °C, the m moisture rem moval rate increased from m 75 wt% to 83 wt% as the driedd coal particlee size decreaased from larrger than 10000㎛ to 100––300㎛. Since the possibiility of coal pyrolysiss becomes hhigher as inleet temperaturre rises, dryiing at a highh temperaturee is not recoommended. Heating value tended to increase with a higgher inlet tem mperature annd smaller pparticle size. However, heating vvalue does noot change muuch when thee particle sizee is smaller thhan 1,000㎛..
Fig. 5. M Moisture rem moval at diffeerent applied temperature Fig. 6 shows thee moisture removal r ratee according to inlet veloocity in the drying riserr tube and accordinng to the coall size distribuution. The moisture remooval rate tendds to increasee as the inlet velocity of the dryinng riser tube decreases frrom 20 m/secc to 10 m/secc and the parrticle size deccreases. Wheen the inlet velocity is 15 m/sec and 20 m/seec, the moistture removall rate increassed from 70 wt% to 80 w wt% as the coal sizee decreased from over 1000 ㎛ to 100–300 ㎛ ㎛. When the inlet velociity was 10 m/sec, the moisturee removal ratte increased ffrom 83 wt% % to 87 wt% aas the coal siize decreasedd from 500–1000 ㎛ to 100–3000 ㎛. A decreease in inlet vvelocity in thhe drying risser tube impllies that the pparticle stayss inside the drying riiser tube for a longer perriod. Hence, the wet mateerials to be ddried are in ccontact with tthe hot gas for longeer than expeected and thee moisture reemoval rate cconsequentlyy increases foollowing a slower inlet velocity..
Fig. 6.. Heating valuue at differennt applied tem mperatures.
4. Coonclusions In thiss paper, an uppgrading expperiment wass performed oon three typees of LRC usiing a hybrid flash dryer. The expperimental vaariables incluuded inlet teemperature, particle sizee, and inlet vvelocity. Wee drew the followinng conclusionns from the exxperiment results: The reesidence timee of particless increased due d to the sw wirling flow ccoming to thhe cyclone, w which gives providess an additionaal drying effeect. Thereforre, the cyclonne plays an im mportant rolee not only inn separating
the dried coal particles from gas, but also in flash drying because it increases the staying time of the particles. In LRC drying using the hybrid flash dryer, the moisture removal rate increased as the inlet temperature increased and particle size decreased. In a hybrid flash dryer experiment, using a narrow range by sieving the coals in order to obtain a high moisture removal rate is recommended. In order to have an overall stable moisture removal rate, increasing the distribution of the particle size to be dried is necessary.
Acknowledgement This work was supported by the Power Generation & Electricity Delivery of the Korea Institute of Energy Technology Evaluation and Planning(KETEP) grant funded by the Korea government Ministry of Knowledge Economy
Reference [1] WEC, "Survey of Energy Resources, Interim Update 2009" [2] MITHRA Research, White Energy Company, 2008. [3] Satoru Sugita, Tetsuya Deguchi and Takuo Shigehisa, "Demonstration of a UBC process in Indonesia", 神戸製鋼技報, 56(2), 2006, pp.23-26 [4] Anderson, B, "Development of Integrated drying and gasification of brown coal for Power Generation", Institute of Chemical Engineers Conference: Gasification–The Gateway to the Future, Dresden, Germany, September 1998 [5] Myeong-Chi Park, Yeong-Su Gwon, “Design of flash dryer(I)", HWAHAK KONGHAK, 12(6), 1974, pp. 353-360 [6] A.S. Mujumdar, Handbook of Industrial drying, 3th Edition, CRE Press, 2007, pp. 993-1016 [7] P.Bunyawanichakul, M.P.Kirkpatrick, J.E. Sargison and G.J.Walker, "Numerical and Experimental Studies of the flow field in a cyclone dryer", Journal of Fluids Engineering, 128, 2006, pp. 1240-1250
2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010
DRYING KINETICS OF LOW RANK COAL IN MULTI-CHAMBER FLUIDIZED BED Jaehyeon Park, Dr., Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA, T: 82-42-860-3274; F: 82-42-860-3134; E: [email protected] Dowon Shun, Dr., Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA, T: 82-42-860-3672; F: 82-42-860-3134; E: [email protected] Dal-Hee Bae, Dr., Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA, T: 82-42-860-3674; F: 82-42-860-3134; E: [email protected] Sihyun Lee, Dr., Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA, T: 82-42-860-3452; F: 82-42-860-3134; E: [email protected] Jeong Hak Seo, Mr., Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA, T: 82-10-2093-5890; F: 82-42-860-3134; E: [email protected] Jaehyeok Park, Mr., Hanyang University 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA, T: 82-10-3173-5449; F: 82-42-860-3134; E: [email protected]
Abstract: When low rank coals such as subbituminous coals and lignites are combusted with high moisture content, they cause higher flue gas flow rate resulting in lower plant efficiency. A better understanding of the drying process would improve the development of the use of these low-grade coals, and they could be economically treated for further processing such as combustion, gasification and liquefaction after appropriate drying processes. Among various drying technologies, fluidized bed drying has the advantages of temperature control due to uniformity of bed temperature and high drying rate. When hot air is passed upward through a perforated distributor, the coals in the dryer are suspended and the bed has many properties similar to a fluid. It offers a way of drying coal in more economical end environmentally acceptable means. Drying can be done in any state of fluidization, however the optimum condition is highly related to drying gas temperature and gas velocity. The objective of this research work is to develop bench-scale (1ton/day) multi-chamber continuous fluidized bed coal dryer to overcome the disadvantages of low rank coal with high moisture such as low calorific values, costly transportation, high emissions of pollutants, and operation problem. The effects of gas temperature, gas velocity and coal feed rate on drying rate were studied to obtain information relating to optimum operating conditions. Coal characterizations (proximate analysis, ultimate analysis, Thermogravimetric Analysis (TGA), BET, Higher Heating Value (HHV), Lower Heating Value (LHV)) were performed to identify the effect on the change of moisture content. This investigation aims to study the drying process under moderated heating conditions. The coal drying experiments were performed in a 100 mm width, 500 mm length, 1700 mm high fluidized bed dryer consisting of preheater, plenums, bubbling bed, and cyclone. The particles size of crushed coals was minus 3 mm. The gas temperatures and velocities were varied up to 150 °C and 0.9 m/s, respectively. As a result of the experiments it was concluded that the thermal fluidized bed process could be successfully applied to reducing moisture in low-rank coal. Results also indicate that about 80 of total moisture could be reduced, including some of the inherent moisture, yielding high heating value product. The drying rate of coal in a fluidized bed was increased by increasing the temperature and velocity of the drying gas and by decreasing the coal feed rate. However gas temperature had limitations causing from the spontaneous combustion and gas velocity had to be decided considering energy efficiency. Those parameters will be used for the design of pilot scale (10 tons/day) fluidized bed dryer. INTRODUCTION In recent years, drying and upgrading of high moisture coals (25%~65%) obtain attentions worldwide owing to the increased energy consumption. High moisture coals include brown coal and subbituminous coal. It causes serious problems for further processing such as combustion, gasification and liquefaction. For example, if the lignite is combusted with high moisture content, it results in higher fuel flow rate causing lower plant efficiency, higher flue gas rate. Furthermore, the high moisture content and low calorific values make the transportation of the lignite costly, even though lignite resources amount to 1025 billion tons (Gt) and another 207 Gt reserves worldwide
(1). It could be economically treated by appropriate drying processes. And a better understanding of the drying process would improve the development of the use of these low-grade coals. Among various drying technologies, fluidized bed drying has the advantage of temperature control due to uniformity of bed temperature and high drying rate. It offers a way of drying coal in more economical end environmentally acceptable means. Drying can be obtained in any state of fluidization, however the optimum conditions is highly related to drying gas temperature and gas velocity (2, 3). We have performed a batch type laboratory-scale (50kg/day) bubbling fluidized bed drying experiments of Indonesian coal to identify the optimal operation conditions (4). In this work, the coal drying in bench scale (1ton/day) continuous fluidized bed is discussed in terms of gas temperature, gas velocity and coal feeding rate. The pilot scale (10 tons/day) continuous multi-chamber fluidized bed dryer is being built in East-West electric power station in Korea and the optimization of operating parameters will be performed for the stable and efficient operation. EXPERIMENTAL
Figure 1. Photo of bench-scale (1ton/day) multi-chamber continuous fluidized bed coal dryer. Figure 1 shows the photo of bench-scale (1ton/day) multi-chamber fluidized bed coal dryer. It is a continuous type dryer consisting of preheater, plenums, bubbling bed, and cyclone. The width and length of dryer are 100 mm and 500 m, respectively. The height of dryer is 1,700 mm. The drying conditions are changed for each run. For a constant drying gas velocity and feeding rate, the drying gas temperature was varied from 50 °C to 150 °C. The drying gas velocity for constant gas temperature and feeding rate was varied from 0.5 m/s to 0.9 m/s. Coal feeding rate was changed from 5 to 20 kg/hr for constant gas temperature and velocity. The high moisture coal was loaded onto the fluidized bed dryer after the drying gas had reached the drying temperature. During the experiment, dried coal was sampled and the moisture contents were measured using Moisture Analyzer (HR-83P, BK Technology). Thermocouples inserted into the bed were used to measure vertical distribution of bed temperature of multi-chamber fluidized bed coal dryer. RESULTS AND DISCUSSION Figure 2 and 3 shows the temperature and pressure drop profiles along bench-scale continuous fluidized bed dryer. Those are measured at wind box (201), 10 cm(211), 25 cm(221), and 40 cm(231) above distribution plate. The gas temperature was 110 °C, and coal feeding rate was 13 kg/hr. In Figure 2, it is seen that the temperatures at
211, 221, and 231 showed very similar values. It means that the coals in fluidized bed were very well mixed in fluidized bed and the operation was very stable. The temperature of bed was about 40 °C at 211, 221 and 231. The result indicates that most of heats from the drying gas are utilized for the removal of moisture in coal indicating that the fluidized bed drying is useful method in terms of energy efficiency. In Figure 3, pressure drop profiles show steady values also representing stable operation, and used for the control of bed height.
180 160
TE_201
TE_211
TE_221
TE_231
140 120 100 80 60 40 20 0 0:00
1:12
2:24
3:36
4:48
6:00
7:12
Figure 2. Temperature profiles of bench-scale fluidized bed dryer at different vertical locations.
600 PT_201 PT_211 PT_221 PT_231
500
400
300
200
100
0 0:00
1:12
2:24
3:36
4:48
6:00
Figure 3. Pressure profiles of bench-scale fluidized bed dryer at different vertical locations.
7:12
Throughout the experiments, it was confirmed that the drying rate increases as the gas temperature and gas velocity increase. It is evident that the heats to the coal particles for the drying of coal increases as the gas temperature and gas velocity increase and, in turn, the drying rates increase. It is also found that the coal drying rate decreases as coal feeding rate increases. CONCLUSIONS We have performed a bench-scale (1 ton/day) continuous bubbling fluidized bed drying experiments of low-rank coal to identify the optimal operation conditions. It has investigated that the drying of moisture from coal particles in a fluidized bed was increased by increasing the drying gas temperature and velocity, and decreasing the coal feeding rate. During the drying in fluidized bed, the operation was stable and it was confirmed with temperature and pressure drop profiles at various locations. It was also found that fluidized bed drying is an efficient and economical technology for high moisture, low rank coal such as lignites and sub-bituminous coals since most of heats from the drying gas are utilized for the removal of moisture in coal. In our next work, the operation and optimization of pilot scale (10 tons/day) continuous multi-chamber fluidized bed dryer, which is being built in East-West electric power station in Korea, will be discussed. ACKNOWLEDGEMENT This work was supported by the Power Generation & Electricity Delivery of the Korea Institute of Energy Technology Evaluation and Planning(KETEP) grant funded by the Korea government Ministry of Knowledge Economy REFERENCES 1. Thomas Thielemann, Sandro Schmidt, J. Peter Gerling (2007). ―Lignite and hard coal: Energy suppliers for world needs until the year 2100 — An outlook‖, International Journal of Coal Geology 72 (2007) 1–14. 2. Edward K. Levy, Hugo S. Caram, Zheng Yao, Zhang Wei and Nenad Sarunac (2006). "Kinetics of coal drying in bubbling fluidized beds", Proceedings Fifth World Congress on Particle Technology, Orlando, Florida. 3. Nenad Sarunac, Mark Ness and Charles Bullinger (2007), "One year of operating experience with a prototype fluidized bed coal dryer at coal creek generating station", Third International Conference on Clean Coal Technologies for our Future, May 15-17 2007, Cagliari, Sardinia, Italy. 4. Jaehyeon Park, Dowon Shun, Dal-Hee Bae, Sihyun Lee, Jeong Hak Se1, Jae Hyeok Park (2010), " The effect of gas temperature and velocity on coal drying in fluidized bed dryer", Fluidization XIII: New Paradigm in Fluidization Engineering, May 16-21, Hotel Hyundai, Gyeong-ju, Korea.
THE INFLUENCE OF THE TEMPERATURE ON ADSORPTION OF SDS ON COALS Roman Maršálek*, Boleslav Taraba Ostrava University, Dept. of Chemistry, 30.dubna 22, 701 03 Ostrava 1, Czech Republic * E-mail: [email protected] Abstract The influence of temperature on adsorption behavior of anionic surfactant – sodium dodecyl sulfate (SDS) has been studied. Batch mode was used for adsorption of SDS on three different types of coal, namely oxidative altered coal (BO), subbituminous coal (SB) and bituminous coal (BC). The adsorption from aqueous solutions was carried out at temperatures 25, 40, 60 and 80 o C. Linear as well as nonlinear regression analyses of experimental data confirmed that adsorption process can be described using Langmuir adsorption theory of monolayer coverage. Thus, equilibrium constant of adsorption was possible to calculate from Langmuir binding constant b. Further, from the dependence of b constant on temperature, thermodynamic parameters as enthalpy ∆Hads, entropy ∆Sads and Gibbs energy ∆Gads were evaluated. Calculated thermodynamic parameters indicate that the adsorption of SDS on the coals is exothermic (∆Hads < 0) and spontaneous (∆Gads < 0) process. The adsorption capacity of SDS on coal was found to increase with rise in temperature for all studied samples, the highest adsorption capacity being observed for bituminous coal. Also, tight relationship between adsorption capacity and critical micelle concentration (CMC) of SDS has been found. In addition, from zeta potential measurements, correspondence between adsorbed amount of SDS on coal surface and zeta potential value arose.
1. Introduction Surfactants are used commonly in many industrial and commercial products and processes over the world. For example, surfactants are in demand for industrial processes requiring colloid stability, metal treatment, mineral flotation, oil production, emulsion polymerization, pesticides, pharmaceutical formulation. This is positive role of surface – active agents but their applications can produce environmental pollution and increase a series of problems for wastewater treatment plants [1]. Considering the impact of surfactant on the environment, it is important to remove them from waste waters. Adsorption is widely used approach to remove contaminants from aqueous solutions. Study of adsorption of surfactants on coal has at least two aspects – improve yield of coal during froth flotation and purification of waste water. In the both cases, knowledge of surfactant adsorption mechanism is needed. Different adsorption mechanisms and models have been suggested, depending on the adsorbent-adsorbate system. Below the critical micelle concentration when ionic surfactants behave as a single amphiphilic ions there are generally two possible interactions between adsorbent surface and surfactant molecules. Surfactants can be caught on the adsorbent through hydrophobic interactions, by their hydrocarbon coil, or through electrical forces, by their ionic groups. The predominance of one kind of interaction adsorbent - ionic surfactant in the adsorption process is determined by the surface characteristic of the adsorbent. Important role plays for example specific surface area, concentration of polar sites (especially oxygen functional groups), equilibrium moisture content, carbon aromaticity
etc. Various techniques were used for description of surfactant adsorption on the carbonaceous materials, e.g. calorimetry, measurements of contact angle, electrophoresis [2-5]. Hydrophilic adsorption of surfactants was found in three Australian coals [3]. However, measurements of the zeta potential of active coal and carbon blacks in the presence of surfactants led to the idea that the interactions are mainly of hydrophobic character [2, 6]. Such a type of the interactions was also supported by calorimetric measurements of natural and activated coals [7]. In this work, adsorption of SDS on three various coals was studied by means of conventionally measured adsorption isotherms to compare the adsorption properties of the adsorbents. The influence of temperature was studied and thermodynamic parameters, e.g. adsorption enthalpy, entropy and Gibbs free energy, were calculated. Another aim was to use the zeta potential measurements for description of the SDS adsorption from aqueous solution onto coals. The obtained results were used to elucidate possible mechanism of SDS adsorption onto coals. 2. Experimental 2.1. Materials and chemicals 2.1.1. Coal samples Samples of bituminous coal (denoted as BC), bituminous oxidative altered coal (BO) and subbituminous coal were investigated. The elemental compositions and the proximate analysis data of the studied samples are shown in Table 1. Table 1 Proximate and ultimate analyses of sample Sample Ashdry VMdaf Cdaf Hdaf Ndaf (%) (%) (%) (%) (%) BC 6.6 35.3 85.6 5.4 1.1 SB 22.7 55.6 71.9 6.4 1.0 BO 11.5 31.2 76.6 4.1 1.8 VM – volatile matter content
Odaf (%) 6.5 20.5 17.5
Stotaldry (%) 1.2 1.2 2.4
The following techniques were used for the texture and structure characterization of the coal samples: The adsorption isotherms of nitrogen at -196°C and carbon dioxide at 25°C were measured. Surface area of the sample from the adsorption isotherm of nitrogen (SBET) was evaluated according to the BET theory, using a nitrogen molecular area of 0.162 nm2 for the calculation. For the adsorption isotherm of carbon dioxide, the Medek´s equation was applied to the evaluation of the micropores surface area (Smicro). DubininRadushkevich theory was used for calculation of the micropores volume (Vmicro). To estimate a hydrophilic/hydrophobic character of the coal surface, enthalpy of immersion of the dried sample with water, Him, was determined at 30°C using a SETARAM C80 calorimeter provided with a mixing cell.
Equilibrium amount of the coal moisture content (Weq) was measured at the laboratory temperature and relative humidity of air ≈ 60%. The Weq values can be related to the abundance of the hydrophilic sites on coal surface. Diffuse reflectance infrared Fourier transform (DRIFT) spectrum of the sample was measured by Nicolet 7600 FTIR spectrometer (Thermo Nicolet Instruments Co., Madison, USA). The spectrum averaged 512 scans per sample and was recorded with the resolution of 2 cm –1. Two ratios were calculated for characterization of the coal structure Ial (I2920/(I2920+I1610)) and ICO (I1725/(I1725+I1610)). Index Ial reflects an amount of aliphatic carbohydrates in the coal structure. Index was determined as a ratio between the integrated areas of the aliphatic C – H bands (2920 cm-1) and the areas of both aromatic and aliphatic C=C absorption bands (1610 cm-1). Index ICO reflects an amount of the carbonyl function groups in the coal structure. It was determined as a ratio between the integrated areas of the carbonyl groups (1725 cm-1) and areas of the carbonyl groups and aromatic C=C absorption bands (1610 cm-1). 13 C CP/MAS NMR spectra of the sample was measured by Bruker Avance 500 WB/US (Karlsruhe, Germany, 2003) at 125 MHz frequency. The sample was rotated below magic angle. Glycin was used as an external standard. For the quantitative evaluation, carbon aromaticity, fA, was determined as a ratio between the content of carbon in the aromatic part of structure and the total amount of carbon. The results are summarized in Table 2. Table 2 Texture and structure parameters of samples Weq(60%) Sample SBET Smicro Vmicro Him(H2O) m2/g m2/g cm3/g % J/g BC 1.5 133 0.049 2.7+0.3 2.7 SB 49 152 0.055 26.5+1.5 12.5 BO 1.5 235+15 0.084 81+2 14.3
ICO
Ial
0.05 0.26 0.27
0.30 0.36 0.05
Aromaticity fA 0.68 0.50 0.96
2.1.2. Chemicals Sodium dodecyl sulfate (SDS) (or Sodium lauryl sulfate), (Sigma-Aldrich) of analytical reagent grade was used for the adsorption investigations.
Figure 1: Structure of SDS
2.2. Surfactant adsorption measurements The coal samples were diminished and sieved to a particle size smaller than 32 µm. The coal amount of 0.1 g was weighed in the flask and 100 ml of the SDS solution of a known concentration was added. The suspensions were inserted into a thermostatic bath at four different temperatures (25oC, 40oC, 60oC and 80oC) and flasks were permanently
shaken. As it was found in the previous experiments, a 24 hours´ period is needed to reach equilibrium. The zeta potential of the adsorption suspensions was measured. Then, the coal was separated by filtration with paper filter. The amount of SDS adsorbed (a) was determined from the change in the solution concentration before and after equilibrium, according to: (c − ce )V a= 0 (1) m where c0 is the initial concentration of SDS solution, ce is the concentration of SDS solution at the adsorption equilibrium, V is the volume of SDS solution and m is the mass of the coal. The SDS concentration of the filtered solutions was determined by UV/VIS spectrophotometry. Critical micelle concentration of SDS was determined by conductivity measurement. 2.3. Zeta potential measurements Zeta potential measurements was performed by analyzing 0.1 g of coal in 100 ml SDS using the Coulter Delsa 440 SX (Coulter Electronic, USA). Delsa 440 SX uses the scattering effect of Doppler light to determine the electrophoretic mobility. The zeta potential was obtained from the electrophoretic mobility by the Smoluchowski equation: µ.η ζ = (2)
ε
ζ is the zeta potential (V), η represents dynamic viscosity (Pa.s),and εo stands for the dielectric constant. The fixed conditions of measuring were the following ones: temperature (298 K), electric field (15 V), frequency (500 Hz), and the properties of the samples – viscosity (0.0089 kg.m-1.s-1), refraction index (1.333), and dielectric constant (78.36). 3. Results and discussion 3.1. Adsorption capacity of coals and change of zeta potential Figure 2 depicts the typical adsorption isotherms obtained from the experimental data. The shape of isotherms indicates that the adsorption data could be well fitted by the Langmuir adsorption model of monolayer coverage. In a linear form, the Langmuir equation is given by ce ce 1 = + (3) a am am b where a is the amount of SDS adsorbed, ce is an equilibrium concentration of SDS in solution, b represents a monolayer binding constant and am is the monolayer adsorption capacity.
1.2 1
a (mmol/g)
0.8 0.6 SB
0.4
BO BC
0.2 0 0
2
4
6
8
10
ce (m m ol/l)
Figure 2: The adsorption isotherms for SDS adsorption onto three types of coals, temp. 25oC. The adsorption isotherms proved to be consistent with the Langmuir model as deduced from the calculated r-square values close to 1. The most informative parameter in the Langmuir equation is am, providing information on an adsorbed amount at the monolayer surface coverage. The same order in adsorption capacity (BC >BO >SB) was found at all temperature levels. The reason for different sorption capacity of bituminous coal (BC), oxidative altered bituminous coal (BO) and subbituminous coal (SB) has to be searched in coal matrix. Sample BC represents typical bituminous coal. Both surface area and concentration of polar sites of the coal are of the lowest value from the samples to be studied. Sample SB represents low-rank coal and sample BO represents oxidative altered bituminous coal. The oxygen content in both coals is very similar as well as amount of carbonyl groups as deduced from infrared spectroscopy (tab. 1, tab. 2). Values of equilibrium moisture content Weq(60%) are comparable too. These facts confirm hydrophilic character of the coal surfaces, comprising mainly oxygen functional groups. Carbon content of SB and BO samples is similar too, however, there are significant differences in carbon aromaticity (see table 2). Coal BO demonstrates mainly aromatic structure, while the SB sample is approximately of the same proportion between aromatic and aliphatic parts. The essential role (for adsorption capacity) will be played by the content of functional groups, respectively hydrophobicity (hydrophilicity) of coals samples. Also the molecules of anionic surfactant definitely affect the surface charge of adsorbent´s particles and also the values of zeta potential. It can be assumed that adsorption of surfactant on coal surface causes a change of coal surface charge and then a change of zeta potential. As already mentioned, after lapse of contact time the zeta potential of suspensions was measured. The following figures 3 and 4 show a connection between the adsorbed amount and the zeta potential.
1.2
-70
-50
a (mmol/g)
0.8
-40 0.6 -30 adsorption
0.4
-20
zeta potential 0.2
zeta potential (mV )
-60
1
-10
0
0 0
2
4
6
8
10
12
ce (m m ol/l)
Figure 3: SDS adsorption onto bituminous coal vs. the values of zeta potential for each suspension 1.2
35 30
BC
25
am (mmol/g)
0.8
BO 20
0.6 SB
15
0.4
10
0.2
∆zeta potential (mV)
1
5
0
0 0
5
10
15
20
25
oxygen content (%)
Figure 4: Dependence of zeta potential change – absolute value (white circles) and adsorption capacity of SDS (am, black circles) on coals with the equilibrium concentration of the surfactant solution (ce). The following findings can be deduced from experimental data: i. addition of the anionic surfactant (SDS) affects the value of ξ- potential, making the zeta potential values more negative (fig 3); ii. the most significant change of the ξ- potential in the presence of SDS was observed for bituminous coal (BC, fig 4), i.e. for sample of the most hydrophobic surface;
iii. there is relationship between oxygen content of the coal, its adsorption capacity for SDS and change in the zeta potential of (adsorption) suspension (fig 4). Decrease in oxygen content (= increase in hydrophobicity) of the coal leads to the increase in its adsorption capacity for SDS, the adsorption being accompanied by the most significant changes in the zeta potential values of the suspensions. The Langmuir adsorption theory was developed for adsorption of gases on solids and, thus, it should not be apparently valid for adsorption from the solutions. Thus, the experimentally obtained data were also analyzed in terms of non-linear regression. Four adsorption models (Langmuir, Freundlich, Langmuir-Freundlich and Temkin) were used and compared. The non-linear regression was calculated by means of OPstat program using various algorithms – genetic algorithm, simplex, and Levenberg-Marquardt algorithm. Akaic criterion was chosen to evaluate the experimental data correspondence with a given theory. The lowest value of Akaic criterion indicates the best correspondence with evaluated theory. The lowest value of Akaic criterion was found for Langmuir isotherms, in other words also nonlinear regression confirmed Langmuir model applicability. This fact is important for possibility replaces equilibrium constant of adsorption by Langmuir binding constant b. 3.2. Influence of temperature and calculation of thermodynamic parameters. The temperature is a significant factor that affects adsorption process. Next picture shows the influence of temperature on adsorption amount of SDS on bituminous coal (BC). The adsorption isotherms for subbituminous and oxidative altered coals had a similar development. 1.8 1.6 1.4
a (mmol/g)
1.2 1 0.8 298,15 K
0.6
313,15 K
0.4
333,15 K
0.2
353,15 K
0 0
2
4
6
8
10
ce (m m ol/l)
Figure 5: The influence of temperature on SDS adsorption onto bituminous coal
The reason for positive role of temperature on adsorption capacity can be found in the influence of temperature on the critical micelle concentration. 10 9 2
8
6 5
1
4 BO SB
0.5
BC CMC
0 280
300
320
340
cmc (mmol/l)
am (mmol/g)
7 1.5
3 2 1 0 360
Tem perature (K)
Figure 6: The influence of temperature on adsorption capacity (left side, ■▲●) and the influence of temperature on critical micelle concentration of SDS (ride side, ♦). The maximal adsorption capacity increases with a rising temperature for all analysed coals (calculated according to Langmuir theory). Simultaneously, on the minor axis y, it is obvious that CMC of anionic surfactant also increases with temperature. In such a case, more agile monomer molecules SDS are presented in a greater concentration at a higher temperature. Increase in temperature thus increases amount of surfactant monomer molecules which easily adsorb on coal surface in comparison with micelles. The fact that adsorption of SDS on coals was carried out at four different temperatures and the fact that we have “equilibrium” constant (calculated from Langmuir equation), allows us to calculate thermodynamic parameters of the adsorption process. From the dependence of equilibrium constant (b) on temperature (parameters r-square close to one, fig. 7) it is possible to calculate the change of adsorption enthalpy (∆Hads) and the change of adsorption entropy (∆Sads). ln b = −
∆H ads 1 ∆S ads . + R T R
(4)
From the next well known equation we can compute the change of Gibbs free energy (∆Gads). ∆Gads = ∆H ads − T ⋅ ∆S ads (5)
1.3 1.2
y = 554.32x - 0.8343 R2 = 0.9894 y = 604.01x - 0.9563 R2 = 0.9868
1.1
ln b
1 0.9 y = 820.18x - 1.5731 R2 = 0.982
0.8 0.7
SB 0.6 0.0028
BO 0.003
0.0032
0.0034
BC
1/T
Figure 7: The influence of temperature on equilibrium constant of SDS adsorption onto coals. The following table 3 then summarizes experimental and calculated data of SDS adsorption on coals. Table 3. Langmuir constants and thermodynamic parameters of SDS adsorption on coals Temperature am b ∆Hads ∆Sads ∆Gads (K) (mmol/g) (kJ/mol) (J/mol.K) (kJ/mol) 298.15 0.77 2.93 -2.65 BO 313.15 0.92 2.65 -2.53 333.15 1.07 2.30 -2.37 -5.02 -7.95 353.15 1.14 2.16 -2.21 298.15 0.59 2.77 -2.54 SB 313.15 0.73 2.59 -2.44 333.15 0.84 2.26 -2.30 -4.61 -6.94 353.15 0.92 2.10 -2.16 298,15 1.04 3.32 -2.92 BC 313.15 1.22 2.78 -2.72 333.15 1.41 2.38 -2.46 -6.82 -13.08 353.15 1.62 2.16 -2.20 It is obvious (figure 7, table 3) that adsorption of SDS on coals is exothermic (negative values of ∆Hads) and spontaneous (negative values of ∆Gads) process. The decrease in entropy then corresponds with generally accepted idea of decrease in “disorder” of the system during adsorption process.
Conclusions Adsorption of the anionic surfactant (SDS) on coals was confirmed to exhibit the typical Langmuir-type behavior. Adsorption capacity depends on temperature and varied from 0.77 to 1.14 mmol/g (BO), from 0.59 to 0.92 mmol/g (SB) and from 1.04 to 1.62 mmol/g for BC. The adsorption capacity of SDS on coal was found to increase with rise in temperature for all studied samples. This is probably caused by different composition of SDS solutions at different temperatures. Increase in temperature obviously increases amount of surfactant monomer molecules which easily adsorb on coal surface in comparison with micelles. Tight relationship between adsorption capacity, adsorption enthalpy a zeta potential has been found. For example for bituminous coal, the highest adsorption capacity was accompanied by the highest amount of release “heat” (∆Hads) and, simultaneously, the most significant change in the zeta potential values has been found. From the zeta potential measurements as well as from the surface characteristics of coals was deduced that the sorption mechanism of SDS on the coal surface is connected mainly with interactions between the hydrophobic part of surfactant and the coal surface. Acknowledgments Our work was supported by Czech Science Foundation (105/07/P041) and GAAV project IAA 301870801. REFERENCES [1] Wu, H.S., Pendleton, P., J. Colloid Interf Sci 243 (2001) 306. [2] Gallardo-Moreno, A.M., Gonzáles-Garcia, C.M., González-Martín, M.. and Bruque, J.M., Colloid Surface 57 (2004) 249. [3] Crawford R.J., Mainwaring D.E., Fuel 80 (2001) 313. [4] Mishra, S.K., Kanungo, S.B., Rajeev, J Colloid Interf Sci 42 (2003) 267. [5] Singh, B.P., Fuel 78 (1999) 501. [6] Basar, C.A., Appl surf sci 1-4 (2003) 218. [7] Taraba, B., Kula, P., Gúcka, M., International Slovak and Czech calorimetric seminary. pp 49-50 (2001).
For Presentation at the 27thAnnual International Pittsburg Coal Conference, October 11-14, 2010, Istanbul, Turkey
ASSESSMENT OF ELEMENTAL SULPHUR IN BIODESULPHURIZED COALS L. Gonsalvesh*a, S.P. Marinova, M. Stefanovaa, R. Carleerb, J. Ypermanb a b
Institute of Organic Chemistry, Bulgarian Academy of Sciences, bl. 9, Sofia 1113, Bulgaria
Research group of Applied and Analytical Chemistry, CMK, Hasselt University, Agoralaan – gebouw D, B-3590, Diepenbeek, Belgium *correspondence author – [email protected]
Abstract: Recently clean coal use became foreground of coal technology. Desulphurization strategies take considerable part of this trend. A requirement of any research in this field is to acquire an accurate method for determining the forms of sulphur. It is important to evaluate the effect of the treatment on the properly selected coals for sulphur specific desulphurization processes. There are direct standard analytical procedures for sulphur forms determination, i.e. total sulphur (St), pyritic sulphur (Sp) and sulphatic (Ss). The indirect way for organic sulphur assessment (So) by subtracting the sum of Sp and Ss from the St, can create an overestimation of So content in Sel presence. Two Bulgarian high sulphur containing coal samples, sub-butiminious (Pirin) and lignite (Maritza East), and one Turkish lignite (Cayirhan-Beypazari) are used in the experiments. Prior biotreatments the samples are demineralized and depyritized. The white rot fungi “Phanerochaeta Chrysosporium” – ME446 and the thermophilic and acidophilic archae “Sulfolobus Solfataricus” – ATCC 35091 are the microorganisms applied in the biodesulphurization processes. A procedure for elemental sulfur determination is developed in order to specify the changes in the organic and elemental sulfur as a result of the studied coals biotreatments. Its application gives us ground to achieve better sulphur balance. The results of experiments demonstrated that more suitable solvent for the extraction of Sel in the studied coals is chloroform instead of described in the literature c-He. The highest content of elemental sulphur is registered in the preliminary demineralized and depyritized coals. It can be related to their chemical alteration. As a result of the implemented biotreatments the amount of the elemental sulphur is reduced by 55%. It is found that extracted Sel in the samples varied in the range of 0.7% - 4.6% of total sulphur and from 0.8-5.1% of organic sulphur. Keywords: coal, elemental sulphur, biodesulphurization, HPLC
1
1.
Introduction High sulfur content hinders coal exploration, because it provokes environmental and technological
problems. For example during combustion SO2 emissions, slagging, boiler fouling, corrosion and equipment wear are serious obstacles. Recently, the researchers have a keen interest in regards to biological approaches – biodesulphurization, being an effective method for cleaning sulphur in coals. The main advantage is mild experimental conditions with no harmful reaction products where the value of coal is only slightly affected. Development of new tactics for sulphur compounds removal from coal depends, in part, upon a knowledge of their chemical constitution. A fundamental requirement of any research in this field is to acquire an accurate method for determining the forms of sulphur [1]. It is important to evaluate the effect of the treatment on the properly selected coals for sulphur specific desulphurization processes. Sulphur in coal is recognized as existing in three forms: (1) inorganic sulfate (S s), as ferrous sulfate and gypsum; (2) inorganic sulfides (Sp), generally as ferrous sulfide (pyrites), although sulfides of zinc and lead may occur in some coals; and (3) as organic sulphur compounds (So) [2,3]. Forth sulphur form-elemental sulphur is also detected in some weathered or biodesulphurized coal (Sel) [4,5]. There are some consequences when the coal studied contains elemental sulphur [6]. This is related to the fact that there are direct standard analytical procedures for determining total sulphur, pyritic sulphur and sulfatic sulphur. The indirect way for organic sulphur (So) assessment by subtracting the sum of Sp and Ss from the St, can create an overestimation of So content in Sel presence. But although elemental sulphur actively participates in many biogeochemical processes, the studies related to this species are rather poor, mainly due to the limitations of the few available analytical methods for its determination [7]. Recently we put our effort to appreciate the changes that occur with the sulphur, its different species and organic forms in soluble and insoluble products, in the process of effective biodesulphurization of coals from different sources [8]. Therefore the aim of the current study is to try to develop an analytical procedure for Sel determination and to study the effect of biotreatments on the contents of elemental sulphur in coals. Development of a precise procedure for Sel measurement will give us ground to attain a better sulphur balance. 2.
Experimental
2.1. Coal In the experiments two Bulgarian coal samples - Humovitrain Maritza East (M) and Pirin sub-bituminous coal (P), and one Turkish lignite - Cayirhan-Beypazari (B) are used. Coal samples under study are preliminary demineralized [9, 10] and subsequently treated by diluted nitric acid for depyritization [11]. Demineralized and depyritized coal samples are marked as APF (Ash and Pyrite Free). 2.2. Biodesulphurization Phanerochaeta Chrysosporium – ME446 and thermophilic bacteria Sulfolobus solfataricus – ATCC No 35091 are the microorganisms used in the biodesulphurization processes.
2
In Table 1 are included: total sulphur content together with other sulphur forms calculated on a dry, ash free base; proximate analysis determined by TGA method [12]; higher heating value (HHV) calculated by the formula of Chaniwala [13]. In this table, organic sulphur is calculated by the difference from S t and the sum of Sp and Ss. Table 1 Characteristics of the coal samples
Sample
Proximate analysis (%)
P-initial P-APF P-APF-PC P-APF-SS M-initial M-APF M-APF-PC M-APF-SS B-initial B-APF B-APF-PC B-APF-SS
S content (%), daf
Calorific value
Ash
VM
W
Fix C
St
Sp
Ss
So
(MJ kg-1)
10.36 0.01 0.53 1.69 8.47 1.2 0.89 1.43 29.1 0.64 0.96 1.14
30.97 39.66 39.8 38.13 40.76 46.97 46.34 45.65 23.55 40.56 40.87 39.21
6.45 6.16 6.6 5.34 8.06 6.25 7.37 6.99 12.34 6.49 6.19 6.37
52.22 54.17 53.05 54.85 42.71 45.56 45.4 45.93 35.01 52.31 51.98 53.28
5.49 4.10 3.11 3.53 7.06 4.01 3.04 3.87 7.62 3.49 3.13 2.90
0.57 0.23 0.16 0.20 1.64 0.37 0.20 0.38 1.68 0.38 0.11 0.36
0.54 0 0 0 1.54 0 0 0 1.23 0 0 0
4.37 3.87 2.95 3.33 3.88 3.64 2.84 3.49 4.72 3.11 3.02 2.53
22.68 23.65 23.08 26.3 19.29 19.93 20.15 24.24 16.51 22.03 22.58 23.00
2.3. Extraction procedure The following extracts are under consideration: (i) cyclohexane (c-He) extracts; (ii) soluble part (bitumen) of the coals extracted by chloroform and its aliphatic fraction of the neutral oils; (iii) c-He extracts from preliminary extracted by chloroform samples. 2.4. HPLC Analysis HPLC procedure for Sel determination is tested. All analyses are performed on an Agilent high performance liquid chromatography system with 20 µL injection volume and a diode array detector operating at 254 nm with an 8 nm bandwidth. A Varian Chromosphere 5µ C18 reversed phase column (4.6 mm x 250 mm) is used with an eluent comprised of 95:5 methanol (HPLC grade): water at a flow rate of 1 ml/min. 2.5. Elemental Sulphur Standards for HPLC The external standard is prepared using elemental sulphur (Aldrich, 99.998%). 3. Results and Discussion 3.1. Analysis of Standards The quantity of Sel extracted from coal is determined on the base of calibration curve prepared for standard solutions of Sel in perchloroethylene (PCE). Figure 1 represents the overlaid chromatograms of the series standard solutions of Sel in PCE. The peak approximately at 11.5 min is assigned to Sel. It increases with increasing concentration of elemental sulphur standards. The broad peak registered at 4 min as well as the small peak at 3.5 min, are assigned to PCE. The smallest peak appearing at 3 min could be attributed to acetone, a rest from glassware washing. No other peaks are observed in the studied range of elution region up to 15 min.
3
The inset in Figure 1 is the calibration curve for a series of S el standards. The excellent fit to the data is an indication for the highly quantitative nature of HPLC analysis. Figure 1. Overlaid HPLC chromatograms of Sel in PCE in concentration range 0.1 – 240 mg/L. Inset: Calibration of absorbance as a function of Sel concentration. Each data point represents the measured peak area of a corresponing Sel concentration. 7000
linear fit, r =0.99999
6000
5000
Solvent
2500
4000
Peak area
Absorbance (mAU)
Series of standard solutions 2
3000
2000
3000
2000
1000
1500 0 0
1000
50
100
150
200
250
Concentration (mg/L)
Elemental sulphur
500
0 2
4
6
8
10
12
14
Retention time (min)
3.2. Analysis of the extracts In non-polar solvents Sel dissolves at ambient temperature without decomposition [14]. As a rule, the solubility of the sulphur allotropes in organic solvents decreases with their increasing molecular size [15]. Moreover, it is shown that Sel extractability with organic solvents depends on Sel surroundings [16]. According to Gryglewicz et al. [6] for the Sel extraction, c-He is more suitable than PCE. Therefore in the present study, initially exactly c-He is chosen as a proper solvent for the extraction of the Sel. 3.2.1. Extraction with c-He (i). It is found that the content of Sel determined by HPLC UV-VIS, varied in the range 0.001 – 0.032 wt %, calculated on dry, ash free base (daf). The highest amount of Sel is detected in the initial samples. This is probably related to the progress of oxidation processes (weathering), which can also be concluded by the increased sulphate content (Table 1) in the initial samples (Ss > 0.3 wt %). It is known that Sel can be consider as a product of atmospheric oxidation of pyrite by water and oxygen, and in-situ bacteriological action or both together. In our studies it appears that the amount of detected Sel as well as content of Ss correlates also with the content of Sp. The highest content of Sel is detected in M-initial sample, characterized by the highest content of Sp and Ss, then in Binitial sample and lowest content in P-initial sample. In consequence of demineralization and depyritization treatment with mineral acid the content of Sel is reduced in the range of 40 – 91 % related to the oxidation potential of the used mineral acids. 3.2.2. Extraction with chloroform and assessment of the soluble parts (bitumen), (ii). The changes that occur in the soluble coal part (bitumen) and its neutral oils are also under consideration. It is well known that Sel elutes within the aliphatic fraction of the neutral oil. HPLC determined Sel in the aliphatic neutral oil fraction is much higher than the content of Sel extracted with c-He. In some cases the amount of Sel extracted with chloroform is even 50 times higher than the same in c-He extracts.
4
Similar to the extraction with cyclohexane (i), the highest amount Sel in demineralized and depyritized sample is detected
for
B-APF
coal
followed
by
M-APF.
In
all
biotreated with PC and SS APF samples the content of extracted Sel decreases. 3.2.3. Extraction with c-Hexane of the samples, which are preliminary extracted with chloroform (iii). c-He is a reliable organic solvent for the extraction of Sel as it was already noted in previous studies [6]. The initial strategy of our study was: (1) to examine changes that occur with the soluble in organic solvents (in this case chloroform) bitumen, as a result of the applied biotreatments (ii); (2) to examine changes with S el by applying subsequent c-He extraction of preliminary extracted by chloroform samples (iii). The obtained results for the quantity of Sel in (iii), were extremely low. This fact provoked the question whether part of the Sel is already completely extracted by chloroform. Therefore it was necessary to quantify Sel amount in these products ((ii) and (iii)) and in the c-He extracts of none extracted with chloroform samples (i). The results demonstrate that in our experimental and coal set, chloroform is a better solvent than c-He for the extraction of Sel. This fact could be attributed to the differences in the polarity of both solvents or to Sel surroundings. Obviously it appears that accessibility of Sel for various organic solvents is of utmost importance. 3.2. Elemental sulphur balance Formula for calculation of total Sel distribution, based on the results for Sel content in (i), (ii), and (iii) could be proposed. The total elemental sulphur distribution can be calculated as follow: Sel(tot) = Sel(ii) + Sel(iii), where Sel(tot) – total Sel distribution; Sel(ii) – Sel in (ii) extract; Sel(iii) - Sel in (iii) extract; The obtained values are summarized in Table 2 together with Sel content (%) as a part from St and So. The highest presence of Sel is registered for preliminary demineralized and depyritized coal, explained by their chemical alteration. The content of Sel in the APF samples increases in the following sequence: Pirin < Maritza East < Baypazar. Through the implemented biotreatments the amount of Sel is reduced in all samples. More promising biodesulphurization effect toward Sel appears in the treatment with Sulfolobus Solfataricus. 4.
Conclusions A new procedure for Sel determination in coal and its fractions is offered. It includes exhaustive CHCl 3
extraction and subsequent quantitative analysis of the extracts by HPLC using C 18 reversed phase column. The highest presence of Sel is registered in the preliminary demineralized and depyritized coals. As a result of the implemented biotreatments the amount of Sel could be reduced in the range of 16 – 54 %. The maximum reduction is achieved for the most mature coal sample P-APF treated by Sulfolobus Solfataricus. The content of Sel is also assessed as a part of the total sulphur and organic sulphur. The following range of Sel content is measured: 0,01 – 0,16 wt % or 0,4% − 4,5% of total sulfur and 0.4 – 5.0% of organic sulphur. Our results clearly demonstrate the significance of Sel and that the lack of Sel evaluation may cause appreciable errors in the magnitude for organic sulfur.
5
Table 2. Elemental sulphur balance Sel(tot) µg S/g coal
wt %
% St
% So
P-APF
310.5
0.03
0.73
0.78
P-APF-PC
258.4
0.03
0.96
1.02
P-APF-SS
140.6
0.01
0.28
0.30
M-APF
884.3
0.09
2.24
2.47
M-APF-PC
616.9
0.06
1.97
2.11
M-APF-SS
416.9
0.04
1.03
1.15
B-APF
1557.0
0.16
4.58
5.14
B-APF-PC
1021.1
0.10
3.19
3.31
B-APF-SS
1115.6
0.11
3.79
4.35
Sample
St
% - Sel(tot) as a part (%) from the St (on daf basis) % So- Sel(tot) as a part (%) from the So (on daf basis)
Acknowledgements The study was supported within the framework of Cooperation agreement for joint supervision and award of a doctorate between Hasselt University, Belgium and Bulgarian Academy of Sciences, BAS-Bulgarian bilateral project with Hasselt University, FWO-Flanders. References: [1] Newton RJ, Bottrell SH, Louie PKK, Hawthorne SB, Timpe RC. Coal-specific systematic errors in estimation of pyritic and organic forms of sulphur by ASTM and BS methods. Fuel 1996;75:1251-4. [2] Chakrabaritty SK, Iacchelli A. The dilemma of estimating forms of sulfur in low-sulphur low rank coals. Can. J. Chem. 1986;64:861-4. [3] Chou CL. Geochemistry of Sulfur in Coal. In: Wilson LO, Curt MW, editors. Geochemistry of Sulfur in Fossil Fuels, ACS Symposium Series No 429, Washington, DC 1990, p. 30-52. [4] Duran JE, Mahasay SR, Stock LM. The occurrence of elemental sulphur in coals. Fuel 1986; 65:1167-8. [5] Beyer M, Ebner HG, Assenmacher H, Frigge J. Elemental sulphur in microbiologically desulphurized coals. Fuel 1987; 66:551-5. [6] Gryglewicz G, Gryglewicz St. Determination of elemental sulphur in coal by gas chromatography-mass spectrometry. Fresenius J Anal Chem 2001;370: 60-3. [7] Chen YW, Joly HA, Belzile N. Determination of elemental sulphur in environmental samples by gas chromatography-mass spectrometry. Chemical Geology 1997;137:195-200. [8] Gonsalvesh L, Marinov SP, Stefanova M, Yürüm Y, Dumanli AG, Dinler-Doganay G, Kolankaya N, Sam M, Carleer R, Reggers G, Thijssen E, Yperman Y. Biodesulphurized subbituminous coal by different fungi and bacteria studied by reductive pyrolysis. Part 1: Initial coal. Fuel 2008;87:2533-43. [9] Radmacher W, Mohrhauer P. Brennstoff-Chemie 1956;37:353.
6
[10] Jorjani E, Yperman J, Carleer R, Rezai B. Reductive pyrolysis study of sulfur compounds in different Tabas coal samples (Iran). Fuel 2006;85:114–20. [11] Marinov SP, Stefanova M, Stamenova V, Carleer R, Yperman J. Sulphur functionality study of steam pyrolyzed ‘‘Mequinenza” lignite using reductive pyrolysis technique coupled with MS and GC/MS detection systems. Fuel Process. Technol. 2005;86:523-34. [12] Hassel RL, Du Pont Company, № Ta-54, Instr. Products, Scient. and Process Div., Wilmington, DE 19898 "Proximate analysis of coal and coke" [13] Channiwala SA, Parikh PP, Fuel, vol. 81, pp 1051-63 (2002), "A unified correlation for estimating HHV of Solid, liquid and gaseous fuels. [14] Meyer B. Elemental Sulfur, Chemical Reviews, 1976, Vol. 76, No 3, pp 367-88 [15] Steudel S, Eckert B. Solid Sulfur Allotropes. In: Steudel S, editor. Elemental Sulfur and Sulfur-Rich Compounds I, Top Curr Chem (2003) 230:81–116. [16] Toniazzo V, Mustin C, Portal JM, Humbert B, Benoit R, Erre R. Elemental sulphur at the pyrite surfaces: speciation and quantification. Applied Surface Science 1999;143:229-37.
7
For Presentation at the 27thAnnual International Pittsburg Coal Conference, October 11-14, 2010, Istanbul, Turkey STUDY OF BIODESULPHURIZED HIGH SULPHUR COALS FROM BULGARIA S.P.Marinov1a, M. Stefanovaa, L.Gonsalvesha, N. Kazakovaa, V. Groudevab, M.Ilievb, P.Gadjanovc, R.Carleerd, J.Ypermand a
Institute of Organic Chemistry, Bulgarian Academy of Sciences, bl. 9, Sofia 1113, Bulgaria University of Sofia, Faculty of Biology,D.Tzankov 8, Sofia 1164,Bulgaria c Technical Univ. of Sofia, Thermal and Nuclear Power Eng. Dept., Sofia 1756, Bulgaria d Research group of Applied and Analytical Chemistry, CMK, Hasselt University, Agoralaan – gebouw D, B-3590, Diepenbeek, Belgium [email protected] b
Abstract: Biodesulphurization is one of the perspective methods for production of friendly fuels. The aim of the present study is to determinate the changes in two Bulgarian coal samples, their soluble products and residues after treatment with bacteria. Low rank high sulphur Maritza East lignites and Bobov Dol subbituminous coal, are of important energy significance for Bulgaria. The treatment by bacteria Pseudomonas putida B2 attained 44% total desulphurization for lignites while the biotreatment by Acidothiobacillus ferrooxidans F3 of the subbituminous coal achieved 14% desulphurization. The bitumen (chloroform soluble portion) and fractions with different polarity of neutral oils of initial and biotreated coals are under study. Some increase in oxygen containing homologues of polar diterpenoids, i.e. phenols, ketones, quinone and ketophenol derivatives, an indication for oxidative process, is confirmed by GC/MS for subbituminous coal treated by Acidothiobacillus ferrooxidans F3. During the same treatment two new formed oxygen containing compounds, ~25% from total amounts of polar diterpenoids are found. Concerning polar diterpenoids of treated by Pseudomonas putida B2 lignites their relative contents are comparable with initial lignites. Temperature programmed reduction coupled “on-line” with mass spectrometry (AP-TPR/MS) and its “off-line” TD-GC/MS mode to determine the sulphur changes in solid residues after bitumen extraction are applied. TD-GC/MS profiles are quantitatively interpreted by spiking with deuterated sulphur compounds as inner standards. It is concluded that Acidothiobacillus ferrooxidans F3 bacteria generally decrease pyritic sulphur and slightly oxidize coal organic matter, probably due to microbial metabolites secreted into media. Pseudomonas putida B2 bacteria influence lignite organic matter but pyritic sulphur is also affected probably by metabolites’ formation in the experimental conditions.
Keywords: coal, sulphur, biodesulphurization, AP-TPR/MS, bitumen, GC/MS; 1
Author to whom correspondence should be addressed.
1
1. Introduction Clean coal use is actual challenge concerning higher energy consumption and global ecological impact. Sulphur in coal is emitted mainly as sulphur oxides, which contribute to air pollution and acid rains when coal burns during combustion. One way to prevent environment from sulphur oxides is to reduce sulphur content in fuel before combustion. Microbial coal desulphurization offers many advantages over conventional physical and chemical treatments, while energy value of coal is slightly altered. Acidothiobacillus ferrooxidans and Pseudomonas putida are widespread bacteria used in coal biodesulphurization [1, 2]. Acidothiobacillus ferrooxidans is effective only in the pyritic sulphur removal while Pseudomonas putida microorganism is capable to remove organic sulphur as well. There are scanty studies on the changes in extracted organic products (bitumen) and their fractions from coal biodesulphurization with above mentioned bacteria [3]. Investigations on sulphur changes in solid residues after bitumen extraction are totally missing. The aim of the present study is to appreciate sulphur types changes in biotreated coal samples by Acidothiobacillus ferrooxidans and Pseudomonas putida bacteria. The distributions in chloroform soluble bitumen and different fractions of their neutral oils as well as solid residues after bitumen extraction will be under consideration. A set of spectral, chromatographic and pyrolysis techniques will be applied in view of biotreatment specification.
2. Material and Methods 2.1 Coal samples Technological samples of lignites (low rank coal) from Maritza East deposit (Trojanovo-North mine) and subbituminous coal from Bobov Dol deposit, Bulgaria were selected due to their high sulphur content and energy significance for Bulgaria.
2.2 Microorganisms and microbial desulphurization. Acidothiobacillus ferrooxidans (A.ferrooxidans) F3 bacterial strain was obtained from acid drainage waters near coal mines. The bacteria were grown on 9K medium (Silverman) with Fe2+ at pH = 5, To = 28 oC .
2
Pseudomonas putida (Ps.putida) B2 bacterial strain was isolated from contaminated soils with crude oil. These bacteria were grown on Raymond nutrient medium at pH = 6.8 and To = 28 oC . Shake-flask experiments were carried out with both bacterial strains at pulp density 10% (w/v), To = 28 oC for 21 days.
2.3 Bitumen extraction and chromatographic separation Chloroform soluble portions of initial and biotreated lignites (bitumen) were separated into neutral oil and asphaltene after precipitation in n-hexane. Neutral oil as prepared was fractionated on Silicagel column into three fractions eluted by solvents with increasing polarity: n-hexane for the first aliphatic fraction; benzene for the second aromatic fraction and acetone for the third NSO (N, S, O containing elements) polar fraction. First and second ones were GC/MS analyzed.
2.4 GC/MS analysis GC/MS study was performed on HP 6890 GC system plus HP 59763 MS detector operated in electron impact (EI) mode.
2.5 AP-TPR/MS “on-line” and AP-TPR “off-line” TD-GC/MS analyses Solid residue after bitumen extraction was analyzed by temperature programmed reduction approach (AP-TPR). Identification of evolving gases makes it possible to obtain more qualitative and quantitative information about sulphur distribution using variants with mass-spectroscopy (MS) and GC/MS techniques, coupled on-line and off-line, respectively. These bulk analytical techniques are used for identification of the volatile sulphur species released during pyrolysis [4]. For “on-line” AP-TPR/MS analysis the AP-TPR reactor was coupled “on-line” with a mass spectrometer (FUSIONS-VG Thermo lab MS) through a capillary. The same AP-TPR system was used in absorption/desorption mode [5]. The outlet of the AP-TPR reactor was connected to an ice-cooled metal absorption tubes with SilcoSteel Coating, filled with Tenax/Carbopack B/Carbosieve SIII (Marks) as absorbents. After experiment the absorption tubes were desorbed systematically and analyzed by a TD-GC/MS apparatus. Each tube was spiked with d4-thiophene to quantify results for the target sulphur species. NIST library spectra were used for peak identification with special interest to the different sulfur species, liberated or in situ formed during the AP-TPR experiment. 3
3. Results and discussion Proximate analyses, i.e. moisture, ash content, volatile matter and fixed carbon of initial and treated coal samples, are gathered in Table 1. In this Table the results for calorific values determined by calorimetric bomb are included as well. Reduction of ash content with ~5% for both coals and diminishing in calorific value with 0.8% and 4.7%, respectively, for M.East and B.Dol coals, are determined. Data for elemental analysis in Table 2 are combined with the ones for the types of sulphur species. It is demonstrated the highest biodesulphurization effect towards total sulphur (St) achieved for M.East –B2 sample, 44%. In organic sulphur (Sorg) calculation the content of water insoluble sulphate sulphur, produced during microbial treatment is not considered. This fact could be a possible explanation for Sorg increased values after biotreatments of M.East and B.Dol samples (Table 2). Biodepyritization effect after the treatment with A.ferrooxidans F3 and Ps.putida B2 achieve 50% and 69% for “B.Dol” and “M.East” coals. High diminishing in Sp is explained by specific pyrite removal with A.ferrooxidans. Concerning Sp reduction of M.East by Ps.putida bacteria, it is probably caused by forming some metabolites in the experimental conditions. Our investigations were focused on oxidative and sulphur changes in organic soluble matter and solid residues after bitumen extraction of initial and biotreated coals. Yields of bitumen, neutral oils and their fractions with different polarity are presented in Table 3. Contents of bitumen and their neutral oils are reasonable for lignites and subbituminous coal. After biotreatment corresponding bitumen values are slightly affected. Respectively, their contents are higher for lignites with predominance of NSO polar compounds in them. Considerable changes in group composition are revealed for subbituminous coal B.Dol after A.ferrooxidans treatment. The content of NSO compounds increases triple which is an indication for production of oxygen containing compounds. Concerning the fractions of bitumen neutral oils at Ps.putida treatment of lignites the yield of polar fraction increases not so sharply as in the case of A.ferrooxidans treatment of bituminous coal. The first two fractions of column absorption separation, i.e. aliphatic and aromatic components, were amenable for GC/MS study. Generally the same chemical classes are presented in the initial and treated samples and only moderate changes are recognized concerning M.East lignites. For example, content of alkane-2-ones class in aromatic fraction after treatment is double comparing to the content of initial sample. Ferruginol from diterpenoids’ class dominates in both samples. One new formed compound during biotreatment, 7-Keto-totara-5-en-ol, C20H26O2, is registered, 3.6% rel.% . Other compound
4
classes slightly decrease but do not suggest supplementary information for the biotreatment of lignite M.East with Ps.putida-B2 strain. In Table 4 are presented results for the chemical compositions of aliphatic and aromatic fractions from initial and biotreated by bacteria A.ferrooxidans F3 subbituminous coal B.Dol. Aliphatic fraction decreases after biotreatment with preservation compound class pattern of distribution – maximal contents of diterpenoids with Simonellite dominance. Aromatic fraction of subbituminous coal B.Dol demonstrates serious changes: - increase in polar diterpenoids with ~20 %; - rearrangement in distributions of individual compounds when is expressed in rel.%; - appearance of two new oxygen containing compounds with 25.5 rel.%, i. e. 20–Noroxototara-5(10),6-dienol, C19H22O2, and 20-Nor-totara-5(10),6-dienol, C19H24O
strongly
dominating GC distribution. The increase in oxygen containing homologues of polar diterpenoids, i.e. phenols, ketones, quinone and ketophenol derivatives, and emergence of new polar diterpenoids could be considered as an indication for oxidative process ran for subbituminous coal treated by Acidothiobacillus ferrooxidans F3. Presently, the applied GC/MS analysis of bitumen neutral oil fractions of initial and biotreated coal samples did not offer any indications for sulphur containing compounds. Information for them was possible to be obtained by AP-TPR technique “on-line” and “offline” coupled with MS and TD-GC/MS detection modes. Solid residues after bitumen extraction of initial and biotreated coals were analyzed by these techniques. The H2S kinetograms of AP-TPR/MS “on-line” experiments of initial and biotreated samples are presented on Fig.1. As the ions m/z = 34(H2S+) and m/z = 33(HS+) exhibit the same evolution, only the sum of both ion profiles is shown. It is necessary to notice the MS kinetogram interpretation was embarrassed by high pyrite presence in the samples. Additional, this pyrite presence was related to troilite (FeS) formation by reaction of hydrogen with pyrite in pyrolysis process. It maybe concluded from TPR/MS kinetograms that after Ps.putida B2 treatment of M.East lignites there is a clear decrease in sulphides as well as pure aliphatic, aromatic and mixed aliphatic-aromatic sulphides. Within the group of sulphides a decrease in pure aromatic sulphides is more substantial than in pure aliphatic sulphides. Also for the simple and more condensed thiophenic structures there is a clear difference. They seem to be more removed or converted into more complex thiophenic structures. Concerning A.ferrooxidans F3 treatment of B.Dol coal the changes in their ion profiles (Fig.1) are not so well pronounced like in the 5
previous observation. In conclusion for A.ferrooxidans bacterial F3 treatment of B.Dol coal their MS kinetograms of ions m/z 34+/33+ visualize a change or removal of simple thiophenic structures into more complex ones. AP-TPR technique with its extension “off-line” TD-GC/MS was also applied to gain additional information for the sulphur compound classes distributions of solid residues after bitumen extraction. In the chromatograms of the samples are detected typical products of coal pyrolysis, i.e. alkylbenzenes, alkylnaphthalenes, phenols, etc., accompanied by their sulphur containing analogues. The last ones are in dominated quantity owing to specific sorbents use. SO2 and elemental sulphur are also measured. The highest amount is determined for sulphides in M.East sample which after P.putida B2 treatment decreases. Sulphides are not detected in the case of B.Dol samples. Total amounts of sulphur containing compounds decrease after biotreatments by the both strains. Difference in the biodesulphurization mechanism for two bacteria explains elemental sulphur changes in biotreated samples. In the case of A.ferrooxidans F3 treated B.Dol–F3 sample elemental sulphur appearance and its increases are explained by oxidative mechanism [6] while at P.putida B2 treated sample M-East-B2 there is twofold decrease and it is not related to the same mechanism. SO2 quantities in biotreated samples diminish at two bacterial treatments and A.ferrooxidans F3 effect on B.Dol-F3 sample is highly expressed.
4. Conclusions ● Treatment by bacteria P.putida B2 attained 44% total desulphurization for lignites while subbituminous coal treated by A.ferrooxidans F3 strain demonstrated 14% desulphurization. Simultaneously, low diminishing in calorific values for both coals with 0.8% and 4.7%, were measured. ● GC/MS analyses proved some increase in oxygen containing homologues of polar diterpenoids, i.e. phenols, ketones, quinone and ketophenol derivatives for subbituminous coal treated by A.ferrooxidans F3. Two new formed oxygen containing compounds, ~ 25% from total amount of polar diterpenoids of subbitumonous coal are registered. In the case of the treated lignites by bacteria P.putida B2 comparable contents of polar diterpenoids in initial and biotreated samples are obtained. One new formed oxygen containing compound (~ 3.6%) appears as well. ● AP-TPR technique in “on-line” and “off-line” GC/MS mode visualized some organic sulphur species in solid residues after bitumen extraction. AP-TPR “off-line” TD-GC/MS
6
detection gave us opportunity to measure some organic sulphur compound classes in solid bitumen residues. ● A.ferrooxidans F3 bacteria generally decrease pyritic sulphur and slightly oxidize coal organic matter, probably due to microbial metabolites secreted into the media. P.putida B2 bacteria influence lignites organic matter but also affected pyritic sulphur probably by forming some metabolites.
Acknowledgements The study is supported by the National Science Fund of the Ministry of Education and Science, Bulgaria (Project Ref. VU-EEC-304/07) and BAS-Bulgaria bilateral Project with FWO-Flanders, Belgium.
Table1. Characteristics of the coal samples (%). Calorific value, MJ/kg 17.45 17.30 12.91 12.30
Adb Vdaf Fix Cdaf Sample Wa M.East 5.53 34.30 48.33 45.77 M.East-B2 5.44 32.77 48.76 45.61 B.Dol 9.15 53.74 29.45 55.86 B.Dol-F3 6.65 50.77 32.50 57.83 Wa - moisture on analytical basis; Adb - ash on dry basis; Vdaf (volatiles) and Fix Cdaf (fixed carbon) on dry ash free basis; a - as received basis; ∆ - decrease, in %;
∆ Calorific value, % 0.8 4.7
Table 2. Elemental analysis and sulphur distribution in different types, wt. %. Sample M.East M.East-B2 B.Dol B.Dol-F3
Cdaf
Hdaf
Ndaf
61.33 60.58 74.94 70.36
6.25 5.04 6.33 6.02
1.23 2.28 2.01 3.37
Odiffdaf 28.05 28.73 16.10 19.05
Types of sulphur (db) St 5.32 2.97 1.39 1.20
daf - dry ash free; db - dry basis; St - total sulphur; Sp - pyritic sulphur; Ss - sulphate sulphur; Sorg - organic sulphur; Sash - sulphur in ash; Scomb - combustible sulphur; ∆ - decrease, in %;
7
Sp 2.48 0.77 0.92 0.46
Ss 0.84 0 0.22 0.19
Sorg 2.00 2.20 0.25 0.55
∆St 0 44 0 14
Table 3. Yields of chloroform bitumens, their neutral oils and fractions, wt %.
Sample
Bitumen*
M.East 4.30 M.East-B2 3.00 B.Dol 0.30 B.Dol-F3 0.50 * - wt. % of sample; ** - wt. % of chloroform extract; *** - wt. % of neutral oil;
Neutral oil** 35.90 43.30 98.90 89.00
aliphatic 13.10 10.90 74.70 43.80
Contents of fractions*** aromatic NSO polar 38.20 42.00 26.80 57.70 18.60 10.30 25.60 28.40
Table 4. Aliphatic and aromatic fractions of neutral oils of initial and biotreated "B.Dol" coal samples.
Chemical classes
Compounds
Formula
M+
Base peak
Content, rel.% Initial Biotreated
Aliphatic fractions n-Alkanes Sesquiterpenoids Diterpenoids
nC22-nC35(nC29)
8.8 3.5 61.0
i.e. 16α(H)-Phyllocladane
C20H34
274
123
4.5
12.0
Simonellite
C19H24
252
237
12.6
15.5
Retene
C18H18
234
219
4.4
4.7
Des-ATriterpenoids
0.5 Des-A-Lupane
0.9 0.5
Hopanes Benzohopanes
5.7 1.1 74.1
C32-C34 i.e.C32H48
432
0.9
0.5
0.4
0.5
1.0
211
0.3
0.5
Aromatised triterpenoids
0.9
0.1
Aromatic fractions Polar diterpenoids i.e.
36.8
46.3
Abieta-7-en-18-ol
C20H34O
290
165
11.2
3.6
Abieta-8,11,13-trien-7-one
C20H28O
284
269
7.1
6.8
3-Keto-Simonellite
C19H22O
266
251
12.1
6.9
Alkylated phenanthrequinone
C18H16O2 264
221
3.0
2.6
7-Keto-totarol
C20H28O2 300
213
2.1
1.0
(New) 20-Nor-totara-5(10),6-dienol (New) 20-Nor-oxototara-5(10),6dienol
C19H24O
268
267
0
23.0
C19H22O2 282
267
0
2.5
8
9.00E-012
Intensity (mbar)
8.00E-012
AP-TPR/MS (H2) M. East M. East-B2 B. Dol B. Dol-F3
7.00E-012 6.00E-012 5.00E-012 4.00E-012 3.00E-012 2.00E-012 1.00E-012 100
200
300
400
500
600
700
800
900
1000
Temperature (°C)
Fig.1 AP-TPR/MS in H2 atmosphere, m/z (33+34) kinetograms of M.East and B.Dol initial and biotreated coal samples.
References 1. G.Rossi, Biodepyritization of coal: achievement and problems, Fuel, 72, 1581 (1993). 2. J.L.Shennan, Review: Microbial Attack on Sulphur-Containing Hydrocarbons: Implications for Biodesulphurisation of Oils and Coals, J Chem Technol Biotechnol , 67, 109(1996). 3. M.J.Fabianska, L.Lewinska-Preiss, R.Galimska-Stypa: Microbial alteration of organic matter of humic coal during biological desulphurization, Fuel, 82, 165 (2003); Biodegradation of lignites organic matter caused by microbial desulphurization. Proceed. 24-th IMOG, 2009-Bremen, Germany, Abstract, P.383. 4. Yperman J, Maes II, Van den Rul H, Mullens S, Van Aelst J, Franco DV, et al. Sulphur group analysis in solid matrices by atmospheric pressure–temperature programmed reduction, J Anal Chim Acta 1999;395(1–2):143–55. 5. L.Gonsalvesh, S.P.Marinov, M.Stefanova, Y.Yürüm, A.G.Dumanli, G.Dinler-Doganay, N.Kolankaya, M.Sam, R.Carleer, G.Reggers, E.Thijssen, J.Yperman, Biodesulphurized subbituminous coal by different fungi and bacteria studied by reductive pyrolysis, Part 1: Initial coal. Fuel, 87, 2533 (2008). 6. Acharya C., L.B.Sukla and V.N.Misra, Review. Biodepyriritisation of coal, J Chem Technol Biotechnol , 79, 1-12 (2004).
9
Samples in the World Coal Quality Inventory - a USGS Compilation on Global Coal Susan J. Tewalt, Geologist, U.S. Geological Survey MS956 National Center, 12201 Sunrise Valley Drive, Reston, VA USA, 20192; [email protected] Harvey E. Belkin, Geologist, U.S. Geological Survey MS956 National Center, 12201 Sunrise Valley Drive, Reston, VA USA, 20192; [email protected]
Introduction The main objective of the World Coal Quality Inventory (WOCQI) was to collect and analyze a set of samples of mined coal from around the world during a set time period, from about 1995 through 2006. Generally, international collaborators were provided sample collecting guidelines and forwarded their samples to the U.S. Geological Survey (USGS). Samples were subsequently analyzed for major-, minor-, and trace-elements at the USGS Inorganic Geochemistry Laboratory located in Denver, CO and for proximate and ultimate analyses at a commercial laboratory located in PA using ASTM methods (2007). The resultant dataset, in EXCEL 2003 format, includes nearly 1,600 individual samples from 56 countries and is not subject to the inter-laboratory variability present in many coal chemistry compilations. About 70 percent of the WOCQI samples have data from the commercial laboratory, which are presented on an as-received basis. Values for nearly 50 elements from the USGS laboratory were calculated to a consistent dry, whole-coal basis in the dataset. The WOCQI data should be used with care, given two caveats: 1) sampling on a one-time basis is incapable of showing temporal or spatial heterogeneity of a coal deposit and 2) a review of quality assurance in the USGS laboratory in 2008 discovered that standardization (normalization) was beyond accepted limits of 10 percent. Re-analysis of some WOCQI samples by the USGS Laboratory has shown that the original results are generally valid; although they may be plus or minus 30 percent from values obtained more recently under more rigorous quality control procedures (Energy Geochemistry Lab, 2010). Additional coal sample data were added to a second EXCEL file from USGS international studies conducted prior to WOCQI. These are split into two time periods: before 1990 and after 1990. The sample locations for all three datasets are plotted on a map (Figure 1) that depicts surficial coal-bearing areas of the western hemisphere (North and South America) and of Africa, compiled in Geographic Information System (GIS) format by the USGS in 2008.
WOCQI data Only non-U.S. sample data are presented in the WOCQI; geochemical datasets on U.S. coals can be found in the COALQUAL database (Bragg and others, 1997) and the National Coal Quality Inventory (Hatch and others, 2006). The distribution of available WOCQI data include the following countries (organized by continent): North America (Canada, Mexico, Guatemala, Costa Rica); South America (Argentina, Bolivia, Brazil, Chile, Colombia, Peru, Venezuela); Africa (Botswana, Egypt, Mozambique, Nigeria, South Africa, Tanzania, Zambia, Zimbabwe); Asia/Middle East (Russia, Afghanistan, China, India, Indonesia, Iran, Japan, Kazakhstan, Kyrgyzstan, Malaysia, Mongolia, Philippines, Republic of Korea, DPR Korea, Taiwan, Thailand, Vietnam); Europe (Austria, Belgium, Czech Republic, France, Georgia, Greece, Hungary, Ireland, Norway, Portugal, Romania, Serbia, Slovak Republic, Spain, Turkey, Ukraine, United Kingdom); and Oceania (Australia, New Zealand), as well as samples from Antarctica. Figure 1 shows the extent of the GIS representing surficial coal-bearing areas with the WOCQI sample locations as well as the post-1990 and pre1990 international USGS studies samples. In the eastern hemisphere, only Africa has a GIS layer for coal-bearing areas.
The intent of WOCQI was to compile coal chemistry data representing some of the coals currently being mined (Finkelman and Lovern, 2001) by obtaining samples from international collaborators within a set timeframe - from about 1995 through 2006. The most recent world coal production estimates by the World Energy Council (2009), using data from 61 countries in the year 2007, are 6.37 billion tonnes (long or metric) and global proven reserves are estimated to total 826 billion tonnes (distributed within 68 countries). A report by the Massachusetts Institute of Technology (2007) also indicates global reserves of 900 billion tonnes. These estimates indicate that adequate coal exists to serve as a global energy resource for the foreseeable future. Of great importance to utilization is information on the chemical characteristics of the coal, but with the large amount of reserves distributed widely over the planet, providing adequate coverage is an enormous task. Obviously, not all global coal resource occurrences are included in the WOCQI, but 1,580 individual coal samples were eventually collected for analysis. In some cases, single samples represent entire coal basins. A comparison of WOCQI sample count to global coal resources and production shows where the database is inadequate. Table 1 indicates the number of WOCQI samples and the estimated reserves by country as reported by the World Energy Council (2009). Several countries (China, Turkey, United Kingdom) had collaborators able to collect numerous samples over a wide region. In other countries (Afghanistan, Hungary, India, and Kyrgyzstan) sample collection was done collaboratively in the field with USGS personnel. Table 1.Total proved reserves in millions of tonnes (mt; World Energy Council, 2009) by continent-country with number of WOCQI samples indicated (total 1,580). Multiple samples are possible at a single geographic location (bench samples). Note: n.d. means no data. Continent - country Africa Algeria Botswana Central African Republic Congo (Dem. Republic) Egypt Malawi Morocco Mozambique Niger Nigeria South Africa Swaziland Tanzania Zambia Zimbabwe North America Canada Greenland Mexico United States of America South and Central America Argentina Bolivia Brazil Chile Costa Rica Colombia Ecuador Guatemala Peru
Total proved reserves (mt)
WOCQI sample count
59 40 3 88 21 2 n.d. 212 70 190 30408 208 200 10 502
0 17 0 0 23 0 0 1 0 22 40 0 24 14 6
6578 183 1211 238308
12 0 2 0
424 1 7059 155 n.d. 6814 24 n.d. 50
7 1 64 23 2 16 0 1 16
Venezuela Asia and Middle East Afghanistan China India Indonesia Iran Japan Kazakhstan Korea (Dem. People's Rep.) Korea (Republic) Kyrgyzstan Malaysia Mongolia Myanmar (Burma) Nepal Pakistan Philippines Taiwan, China Thailand Turkey Uzbekistan Vietnam Europe and Russia Albania Belgium Bulgaria Czech Republic France Georgia Germany Greece Hungary Ireland Italy Montenegro Norway Poland Portugal Romania Russia Serbia Slovakia Slovenia Spain Ukraine United Kingdom Oceania and Antarctica Australia New Caledonia
479
16
66 114500 58600 4328 1386 355 31300 600 133 812 4 n.d. 2 1 2070 316 1 1354 1814 3000 150
118 328 107 8 57 2 18 50 11 15 2 37 0 0 0 6 4 23 149 0 6
794 n.d. 1996 4501 n.d. n.d. 6708 3900 3302 14 10 n.d. 5 7502 36 422 157010 13885 262 232 530 33873 155
0 10 0 42 3 1 0 2 36 1 0 0 28 0 2 12 12 5 8 0 10 33 84
76200 2
10 0
New Zealand Antarctica
571 n.d.
17 15
Discussion The WOCQI database is an eclectic mix of grab, bench, channel, and run-of-mine samples with varying degrees of assurance on location, sample condition, and even sample collection methods. Table 2 lists average, median, minimum and maximum values for moisture, ash yield, volatile matter, fixed carbon, sulfur and calorific value on an as-received basis, grouped by the collectordesignated geologic time period, which are used to examine general trends in coal chemistry. ‘Older’ coals show a general decrease in moisture and a general increase in fixed carbon. These trends in composition are logical, as coalification generally increases with age. Exceptions to these trends are the Triassic (represented by only 47 samples) for moisture and the Permian for fixed carbon content. However, the statistics in Table 2 are presented to indicate the level of care that should be employed when selecting samples and interpreting results using the WOCQI database. Known parting samples and coal briquettes were excluded prior to the generation of statistics in Table 2, but samples that were labeled ‘coal’ by the collector that had more than 50 percent but less than 70 percent ash yield (as-received), were included in the statistical calculations. Many of the Brazilian coals are extremely high in ash (mean of 48.5 percent, maximum of 83.1) and would not have been included, but Brazilian coals are mined and often washed prior to being utilized; thus they were included in the summary. Additionally, some coal samples in WOCQI were washed, sized, or otherwise ‘prepped’ prior to analysis (e.g., the Triassic-age Vietnam samples are prepared export blends of coal, likely affecting the moisture and ash contents); these were also included in the statistics. Therefore any trends in chemistry (or lack thereof) may be artifacts of mixed sample types or inclusion of non-coals. In addition, statistics may not be meaningful if benched samples from a single location are not aggregated to a single value, prior to combination with complete channel or run-of-mine samples. Table 2 includes all raw values with no prior statistical treatment and is representative of the WOCQI data, but not necessarily a valid representation of the indicated time period. The ultimate analysis parameters of hydrogen, oxygen, and carbon in WOCQI samples were converted to a dry, mineral-matter free (dmmf) basis with mineral matter estimated using the classic Parr formula (as SO3 was not always available to adjust the ash yield value per ASTM, 2007). Atomic hydrogen/carbon (H/C) and oxygen/carbon (O/C) ratios were calculated from those values and reflect only the organic matter within the samples. Three van Krevelen plots (figures 2 to 4) of H/C – O/C ratios, grouped by major geologic era, show a similar general trend of coalification; the direction of arrows indicates increasing coalification. Although some high-ash samples were excluded in these plots, variation in the ratios is apparent and may reflect unusual constituents that warrant further investigation. Table 2. General statistics by collector’s designated geologic age (average, median, minimum and maximum values) for percent moisture, ash yield, volatile matter, fixed carbon, sulfur and calorific value on an as-received basis. All sample types (grab, bench, channel, run-of-mine, sized, washed, etc) are included and not statistically aggregated or separated prior to generating these statistics. Samples with more than 70 percent ash yield, known partings and coal briquettes were excluded. Note: calorific value is in megajoules per kilogram (MJ/kg); number of analyses used may not match Table 1.
Time Period Quaternary average min max median number Tertiary average min max median number
Total moisture
Ash yield
Volatile matter
Fixed carbon
Total sulfur
Calorific value MJ/kg
16.68 12.65 20.71
12.26 2.21 22.30
45.13 30.30 59.95
25.94 25.19 26.69
1.09 0.97 1.21
17.48 15.01 19.95
2
2
2
2
2
2
17.62 0.69 55.6 17.63 227
15.82 0.41 69.89 11.98 227
33.29 6.8 60.47 34.055 226
33.45 1.17 71.67 32.435 226
1.56 0.11 13.89 0.88 227
20.52 1.77 36.05 19.50 227
Cretaceous average min max median number
17.64 2.12 50.76 15.78 61
13.19 2.17 61.57 9.5 61
24.61 1.65 39.05 28.96 61
44.12 12.02 90.41 41.69 61
0.88 0.16 3.79 0.56 61
21.38 9.44 31.54 21.17 61
Jurassic average min max median number
8.98 0.78 33.04 6.69 224
15.90 1.46 65.69 12.37 224
29.90 2.34 56.27 29.39 224
45.22 5.01 81.88 45.17 224
1.24 0.03 12.92 0.69 224
24.39 2.26 32.65 24.91 224
Triassic average min max median number
3.65 0.84 25.26 2.53 47
22.92 2.68 62.80 24.80 47
17.35 4.03 35.58 17.82 47
56.08 21.73 88.47 53.50 47
0.94 0.19 2.58 0.69 47
25.60 9.88 34.37 24.99 47
Permian average min max median number
4.30 0.55 28.64 3.12 365
23.75 3.54 64.20 20.60 365
20.31 2.51 42.68 21.34 354
51.28 15.11 86.22 49.53 353
1.32 0.09 14.60 0.63 366
23.86 0.21 33.68 24.93 353
Carboniferous average min max median number
3.40 0.55 29.84 2.38 208
15.35 0.66 65.48 12.74 208
23.50 2.82 44.94 24.40 208
57.75 15.29 84.31 56.75 208
1.57 0.14 11.20 1.01 208
27.64 8.36 35.94 28.56 208
Figure 2. Plot of H/C versus O/C atomic ratios for Paleozoic samples (all sample types). Plot shows most WOCQI samples of Paleozoic age are furthest along the coalification path, which increases in the direction of arrows. Note: H/C is hydrogen/carbon; O/C is oxygen/carbon; dmmf is dry, mineral-matter-free.
Figure 3. Plot of H/C versus O/C atomic ratios for Mesozoic samples (all sample types). Plot shows concentration of WOCQI samples of Mesozoic age with less coalification than Paleozoic samples. Note: H/C is hydrogen/carbon; O/C is oxygen/carbon; dmmf is dry, mineral-matter-free.
Figure 4. Plot of H/C versus O/C atomic ratios for Cenozoic samples (all sample types). Plot shows the youngest WOCQI samples with least coalification. Note: H/C is hydrogen/carbon; O/C is oxygen/carbon; dmmf is dry, mineral-matter-free.
The major-, minor-, and trace-element data from the USGS Inorganic Geochemistry Laboratory are subject to some data assurance issues. The USGS initiated a laboratory review of quality assurance in 2008, covering quality control and methodology used in inorganic chemical analyses of coal, coal power plant ash, water, and sediment samples. This quality control review found that data generated by the USGS Inorganic Geochemistry Laboratory from 1996 through 2006 did not meet USGS quality practices commonly in use at the time. The most serious shortcomings were (1) the adjustment of raw sample data values based on expected results of reference standards when the instrument values for those standards exceeded acceptable limits or (2) use of an insufficient number of standards to provide adequate quality assurance. Sixteen elements analyzed by inductively coupled plasma-mass spectrometry (Ag, As, Bi, Cd, Cs, Ga, Ge, Mo, Nb, Pb, Rb, Sb, Sn, Te, Tl, and U) are considered to have the largest variance. Some reported USGS results have been confirmed by commercial analyses; therefore these specific USGS data may be used quantitatively. Alternatively, if the data have not been independently confirmed, but the conclusions drawn from the data are qualitative, results should be adequate to
support those conclusions. WOCQI data, however, should not be used to draw conclusions or to base decisions that are critically dependent upon quantitative results. In addition to the elements noted above, changes in instrumentation and sample preparation methods affected mercury and chlorine reporting (detection limit) values (after July 2004 and March 2005, respectively). For these reasons, the WOCQI data from the USGS Inorganic Geochemistry Laboratory should be used with care. For more information, individuals may contact laboratory management at [email protected] with specific questions about particular datasets, analytical attributes, or the possible re-analysis of samples under new, improved quality control measures. The WOCQI database will be released in late 2010 and a link to this publication will be found at: http://energy.er.usgs.gov/coal_quality/wocqi/index.html. Links to information on the petrography of many WOCQI samples can also be found at the WOCQI website.
Conclusions The WOCQI database is a unique compilation of coal chemical data, expressed on a common analytical basis with limited interlaboratory variability. In some cases, the WOCQI data may be the only data available for public use. Care must be used when selecting the sample data for use in statistics or interpretation because the WOCQI includes: single samples with little or no spatial variability; a mix of sample types and collection methods; and the accuracy of the USGS Inorganic Geochemistry Laboratory major-, minor-, and trace-element data is variable. Although the WOCQI does not contain world-wide coverage, nor do its samples represent entire coal deposits, it may prove useful for identifying possible areas for future global coal research. A statistical summary of several proximate parameters, as well as plots of atomic hydrogen-carbon and oxygen-carbon ratios, show the data are robust enough to confirm a global trend of increasing coalification with burial depth and temperature (Valkovic, 2000) using geologic age as a proxy for those variables.
Acknowledgments The following sample submitters, without whom this endeavor would have proved impossible, are thanked for their participation: Oke Adeleke, Guillermo Alfaro, Saifuddin Aminy, Ganbaatar Badgaa, Hassan M. Baioumy, Lucy Barros de Ferreira, Coal Research Limited, Inc., Wu Daishe, P. Desideri, Doria Diaconita, Slobodan Djekic, Natalia Druz, Michiel Dusar, Gbenga Ehinola, Luis Felipe Silvia Oliveira, Deolinda Flores, Luis Gadea, P.K. Gautam, Faribroz Goodarzi, Hadiyanto, Guy Holdgate, Zhu Jianming, Dorani B. Johari, Wolfgang Kalkreuth, Zdenek Klika, Melkzedeck Koddy, Bernard Lengert, Ben S. Mapani, Manuel Martinez, Rosa Martinez Tarazona, Tim Moore, Imasiku A. Nyambe, Kiyofumi Okada, Alv Orheim, Boris Panov, Suk-Whan Park, Mikhail Y. Povarennykh, Benjavun Ratanasthien, Michael Realini, Silvia Rico, R.I. Rifai, Richard Sakurovs, Mosi Shahrabi, Li Shehong, Tatiana Shendrick, Alan Spears, Yo Sumartojo, Calin Tatu, Louis Tsai, Ertem Tuncali, Jordan Tuntufye, Nguyen Hong Van, Lopo Vasconcelos, Neli Verulava, BinBin Wang, Rob Williams, M. Zhai, and Baoshan Zheng. Samples from some countries, such as North Korea, were obtained from collaborators who were not residents of those countries. Iranian samples were obtained through License No. IA-5012 issued by the U.S. Department of Treasury, Office of Foreign Assets Control to John R. SanFilipo, for the purpose of obtaining, analyzing, and including the data in the WOCQI. Funding for commercial analyses of Afghanistan samples was provided by U.S. Agency for International Development. References ASTM, 2007, Annual book of ASTM standards, volume 05.06 Gaseous Fuels; Coal and Coke Bragg, L.J., Oman, J.K., Tewalt, S.J., Oman, C.L., Rega, N.H., Washington, P.M., and Finkelman, R.B., 1997, U.S. Geological Survey coal quality (COALQUAL) database; version 2.0: U.S. Geological Survey Open-file Report 97-34, CD-ROM (available online at http://pubs.usgs.gov/of/1997/of97-134/ - accessed May 2008) Energy Geochemistry Laboratory, 2010, Energy Geochemistry Laboratory procedures and QA/QC manual, v. 1.2, January, 2010: U.S. Geological Survey (available online at http://energy.cr.usgs.gov/gg/geochemlab/ERP%20Lab%20QA_Procedure1.2.pdf _
Finkelman, R.B., and Lovern, V.S., 2001, The World Coal Quality Inventory (WoCQI): U.S. Geological Survey Fact Sheet 2000 15500 (available online at http://pubs.er.usgs.gov/usgspubs/fs/fs15500 - accessed April 2010) Hatch, J.R., Bullock, J.H., Jr., and Finkelman, R.B., 2006, Chemical analyses of coal, coal-associated rocks and coal combustion products collected for the National Coal Quality Inventory: U.S. Geological Survey Open-file Report 2006-1162, no pagination (available online at http://pubs.usgs.gov/of/2006/1162/ - accessed May 2008) Massachusetts Institute of Technology, 2007, The Future of Coal – options for a carbon-constrained world, ISBN 978-0-615-14092-6 Valkovic, Vlado, Trace Elements in Coal, v. 1: Boca Raton, FL, CRC Press, Inc., 210 p. World Energy Council, 2009, Survey of energy resources interim update 2009: http://www.worldenergy.org/publications/survey_of_energy_resources_interim_update_2009/1794.asp - accessed June 2010)