Sustainability in the Process Industry
About the Authors Jiří Klemeš, H DSC, is one of the key personalities of the world-leading Centre of Excellence in Process Integration at the University of Manchester Institute of Science and Technology in the United Kingdom. Ferenc Friedler, DSC, is a leading research figure at the University of Pannonia in Hungary. Igor Bulatov is a researcher at the Centre for Process Integration at the University of Manchester in the United Kingdom. Petar Varbanov is a senior lecturer at the Centre for Process Integration and Intensification (CPI2) at the University of Pannonia in Hungary.
Sustainability in the Process Industry Integration and Optimization Jiří Klemeš Ferenc Friedler Igor Bulatov Petar Varbanov
New York Chicago San Francisco Lisbon London Madrid Mexico City Milan New Delhi San Juan Seoul Singapore Sydney Toronto
Copyright © 2011 by The McGraw-Hill Companies, Inc. All rights reserved. Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, or stored in a database or retrieval system, without the prior written permission of the publisher. ISBN: 978-0-07-160555-7 MHID: 0-07-160555-X The material in this eBook also appears in the print version of this title: ISBN: 978-0-07-160554-0, MHID: 0-07-160554-1. All trademarks are trademarks of their respective owners. Rather than put a trademark symbol after every occurrence of a trademarked name, we use names in an editorial fashion only, and to the benefit of the trademark owner, with no intention of infringement of the trademark. Where such designations appear in this book, they have been printed with initial caps. McGraw-Hill eBooks are available at special quantity discounts to use as premiums and sales promotions, or for use in corporate training programs. To contact a representative please e-mail us at
[email protected]. Information contained in this work has been obtained by The McGraw-Hill Companies, Inc. (“McGraw-Hill”) from sources believed to be reliable. However, neither McGraw-Hill nor its authors guarantee the accuracy or completeness of any information published herein, and neither McGraw-Hill nor its authors shall be responsible for any errors, omissions, or damages arising out of use of this information. This work is published with the understanding that McGraw-Hill and its authors are supplying information but are not attempting to render engineering or other professional services. If such services are required, the assistance of an appropriate professional should be sought. TERMS OF USE This is a copyrighted work and The McGraw-Hill Companies, Inc. (“McGrawHill”) and its licensors reserve all rights in and to the work. Use of this work is subject to these terms. Except as permitted under the Copyright Act of 1976 and the right to store and retrieve one copy of the work, you may not decompile, disassemble, reverse engineer, reproduce, modify, create derivative works based upon, transmit, distribute, disseminate, sell, publish or sublicense the work or any part of it without McGrawHill’s prior consent. You may use the work for your own noncommercial and personal use; any other use of the work is strictly prohibited. Your right to use the work may be terminated if you fail to comply with these terms. THE WORK IS PROVIDED “AS IS.” McGRAW-HILL AND ITS LICENSORS MAKE NO GUARANTEES OR WARRANTIES AS TO THE ACCURACY, ADEQUACY OR COMPLETENESS OF OR RESULTS TO BE OBTAINED FROM USING THE WORK, INCLUDING ANY INFORMATION THAT CAN BE ACCESSED THROUGH THE WORK VIA HYPERLINK OR OTHERWISE, AND EXPRESSLY DISCLAIM ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. McGraw-Hill and its licensors do not warrant or guarantee that the functions contained in the work will meet your requirements or that its operation will be uninterrupted or error free. Neither McGraw-Hill nor its licensors shall be liable to you or anyone else for any inaccuracy, error or omission, regardless of cause, in the work or for any damages resulting therefrom. McGraw-Hill has no responsibility for the content of any information accessed through the work. Under no circumstances shall McGraw-Hill and/or its licensors be liable for any indirect, incidental, special, punitive, consequential or similar damages that result from the use of or inability to use the work, even if any of them has been advised of the possibility of such damages. This limitation of liability shall apply to any claim or cause whatsoever whether such claim or cause arises in contract, tort or otherwise.
Contents
1
2
3
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xv xxi
Introduction and Defi nition of the Field . . . . . . . . 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Energy Efficiency . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Screening and Scoping: Auditing, Benchmarking, and Good Housekeeping . . . 1.4 Balancing and Flowsheeting Simulation as a Basis for Optimization . . . . . . . . . . . . . . . . 1.5 Integrated Approach: Process Integration . . . 1.6 Optimal Process Synthesis and Combinatorial Graphs ................................ 1.7 How to Apply the Process Integration and Optimization Technology . . . . . . . . . . . . . . . . .
1 1 3
Process Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Introduction: The Need for Process Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 What Is Process Integration? . . . . . . . . . . . . . . . 2.3 History and Development of Process Integration 2.4 Pinch Technology and Targeting Heat Recovery: The Thermodynamic Roots . . . . . . 2.5 Supertargeting: Full-Fledged HEN Targeting 2.6 Modifying the Pinch Idea for HEN Retrofit . . 2.7 Mass Exchange and Water Networks . . . . . . . 2.8 Benefits of Process Integration . . . . . . . . . . . . . 2.9 The Role of PI in Making Industry Sustainable 2.10 Examples of Applied Process Integration . . . . 2.11 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
Process Optimization . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Model Building and Optimization: General Framework and Workflow . . . . . . . . . . . . . . . .
5 7 7 8 9
11 12 12 14 15 16 17 18 20 20 22 23 23 24
v
vi
Contents 3.3
3.4 3.5
3.6 3.7
3.8 3.9 3.10
Optimization: Definition and Mathematical Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 What Is Optimization? . . . . . . . . . . . . . 3.3.2 Mathematical Formulation of Optimization Problems . . . . . . . . . . . . Main Classes of Optimization Problems . . . . Conditions for Optimality . . . . . . . . . . . . . . . . 3.5.1 Conditions for Local Optimality . . . . 3.5.2 Conditions for Global Optimality . . . Deterministic Algorithms for Solving Continuous Linear Optimization Problems . . . . . . . . . . . . . Deterministic Algorithms for Solving Continuous Nonlinear Optimization Problems . . . . . . . . . . 3.7.1 Search Algorithms for Nonlinear Unconstrained Problems . . . . . . . . . . . 3.7.2 Algorithms for Solving Constrained Nonlinear Problems . . . . . . . . . . . . . . . Deterministic Methods for Solving Discrete Problems . . . . . . . . . . . . . . . . . . . . . . . . Stochastic Search Methods for Solving Optimization Problems . . . . . . . . . . . . . . . . . . . Creating Models . . . . . . . . . . . . . . . . . . . . . . . . . 3.10.1 Conceptual Modeling . . . . . . . . . . . . . 3.10.2 Mathematical Modeling of Processes: Constructing the Equations . . . . . . . . 3.10.3 Choosing an Objective Function . . . . 3.10.4 Handling Process Complexity . . . . . . 3.10.5 Applying Process Insight . . . . . . . . . . 3.10.6 Handling Model Nonlinearity . . . . . . 3.10.7 Evaluating Model Adequacy and Precision . . . . . . . . . . . . . . . . . . . . . . . . .
4 Process Integration for Improving Energy Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Introduction to Heat Exchange and Heat Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1 Heat Exchange Matches . . . . . . . . . . . . 4.1.2 Implementing Heat Exchange Matches . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Basics of Process Integration . . . . . . . . . . . . . . . 4.2.1 Process Integration and Heat Integration . . . . . . . . . . . . . . . . . . . . . . .
25 25 25 26 28 28 28 29 29 30 31 31 32 33 34 35 37 38 40 41 43 45 45 46 47 47 47
Contents
4.3
4.4
4.5
4.6
5
4.2.2 Hierarchy of Process Design . . . . . . . . 4.2.3 Performance Targets . . . . . . . . . . . . . . . 4.2.4 Heat Recovery Problem Identification Basic Pinch Technology . . . . . . . . . . . . . . . . . . . 4.3.1 Setting Energy Targets . . . . . . . . . . . . . 4.3.2 The Heat Recovery Pinch . . . . . . . . . . 4.3.3 Numerical Targeting: The Problem Table Algorithm . . . . . . . . . . . . . . . . . . 4.3.4 Threshold Problems . . . . . . . . . . . . . . . 4.3.5 Multiple Utilities Targeting . . . . . . . . . Extended Pinch Technology . . . . . . . . . . . . . . . 4.4.1 Heat Transfer Area, Capital Cost, and Total Cost Targeting . . . . . . . . . . . . . . . 4.4.2 Heat Integration of Energy-Intensive Processes . . . . . . . . . . . . . . . . . . . . . . . . 4.4.3 Process Modification . . . . . . . . . . . . . . HEN Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.1 The Pinch Design Method . . . . . . . . . . 4.5.2 Superstructure Approach . . . . . . . . . . 4.5.3 A Hybrid Approach . . . . . . . . . . . . . . . 4.5.4 Key Features of the Resulting Networks . . . . . . . . . . . . . . . . . . . . . . . . Total Site Energy Integration . . . . . . . . . . . . . . 4.6.1 Total Site Data Extraction . . . . . . . . . . 4.6.2 Total Site Profiles . . . . . . . . . . . . . . . . . . 4.6.3 Heat Recovery via the Steam System . . 4.6.4 Power Cogeneration . . . . . . . . . . . . . . . 4.6.5 Advanced Total Site Optimization and Analysis . . . . . . . . . . . . . . . . . . . . . . . . .
Mass Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Water Integration . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Minimizing Water Use and Maximizing Water Reuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1 Legislation . . . . . . . . . . . . . . . . . . . . . . . 5.2.2 Best Available Techniques . . . . . . . . . . 5.2.3 Water Footprint . . . . . . . . . . . . . . . . . . . 5.2.4 Minimizing Water Usage and Wastewater . . . . . . . . . . . . . . . . . . . . . . 5.3 Introduction to Water Pinch Analysis . . . . . . . 5.4 Flow-Rate Targeting with the Material Recovery Pinch Diagram . . . . . . . . . . . . . . . . . .
47 48 48 50 51 54 56 60 61 69 69 71 80 81 81 93 95 96 96 97 97 99 101 102 105 105 106 106 107 108 111 113 116
vii
viii
Contents 5.5 5.6
5.7 6
MRPD Applied to Fruit Juice Case Study . . . . Water Minimization via Mathematical Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.1 Introduction to Mathematical Optimization . . . . . . . . . . . . . . . . . . . . . 5.6.2 Illustrative Example: A Brewery Plant Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
117 118 118 120 122
Further Applications of Process Integration . . . . . 123 6.1 Design and Management of Hydrogen Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 6.2 Oxygen Pinch Analysis . . . . . . . . . . . . . . . . . . . 125 6.3 Combined Analyses, I: Energy-Water, Oxygen-Water, and Pinch-Emergy . . . . . . . . . 126 6.3.1 Simultaneous Minimization of Energy and Water Use . . . . . . . . . . . . . 126 6.3.2 Oxygen-Water Pinch Analysis . . . . . . 128 6.3.3 Emergy-Pinch Analysis . . . . . . . . . . . . 130 6.4 Combined Analysis, II: Budget-Income-Time, Materials Reuse-Recycling, Supply Chains, and CO2 Emissions Targeting . . . . . . . . . . . . . . 131 6.4.1 Budget-Income-Time Pinch Analysis . 131 6.4.2 Materials Reuse-Recycle and Property Pinch Analysis . . . . . . . . . . . . . . . . . . . 133 6.4.3 Pinch Analysis of Supply Chains . . . . 136 6.4.4 Using the Pinch to Target CO2 Emissions . . . . . . . . . . . . . . . . . . . . . . . . 138 6.4.5 Regional Resource Management . . . . 139 6.5 Heat-Integrated Power Systems: Decarbonization and Low-Temperature Energy 142 6.5.1 Decarbonization . . . . . . . . . . . . . . . . . . 142 6.5.2 Low-Temperature Energy . . . . . . . . . . 143 6.6 Integrating Reliability, Availability, and Maintainability into Process Design . . . . . . . . 144 6.6.1 Integration . . . . . . . . . . . . . . . . . . . . . . . 144 6.6.2 Optimization . . . . . . . . . . . . . . . . . . . . . 146 6.7 Pressure Drop and Heat Transfer Enhancement in Process Integration . . . . . . . . . . . . . . . . . . . . 146 6.8 Locally Integrated Energy Sectors and Extended Total Sites . . . . . . . . . . . . . . . . . . . . . . 148 6.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Contents 7
8
9
Process Optimization Frameworks . . . . . . . . . . . . . 7.1 Classic Approach: Mathematical Programming 7.2 Structural Process Optimization: P-Graphs . . 7.2.1 Process Representation via P-Graphs 7.2.2 The P-Graph’s Significance for Structural Optimization . . . . . . . . . . . 7.2.3 The P-Graph’s Mathematical Engine: MSG, SSG, and ABB . . . . . . . . . . . . . . . 7.3 Scheduling of Batch Processes: S-Graphs . . . . 7.3.1 Scheduling Frameworks: Suitability and Limitations . . . . . . . . . . . . . . . . . . . 7.3.2 S-Graph Framework for Scheduling . .
151 151 153 154
Combined Process Integration and Optimization 8.1 The Role of Optimization in Process Synthesis 8.2 Optimization Tools for Efficient Implementation of PI . . . . . . . . . . . . . . . . . . . . . 8.3 Optimal Process Synthesis . . . . . . . . . . . . . . . . 8.3.1 Reaction Network Synthesis . . . . . . . . 8.3.2 Optimal Synthesis of Heterogeneous Flowsheets . . . . . . . . . . . . . . . . . . . . . . . 8.3.3 Synthesis of Green Biorefineries . . . . . 8.3.4 Azeotropic Distillation Systems . . . . . 8.4 Optimal Synthesis of Energy Systems . . . . . . 8.4.1 Simple Heat Integration . . . . . . . . . . . . 8.4.2 Optimal Retrofit Design . . . . . . . . . . . 8.5 Optimal Scheduling for Increased Throughput, Profit, and Security . . . . . . . . . . . . . . . . . . . . . . . 8.5.1 Maximizing Throughput and Revenue 8.5.2 Heat-Integrated Production Schedules 8.6 Minimizing Emissions and Effluents . . . . . . . 8.7 Availability and Reliability . . . . . . . . . . . . . . . . 8.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
165 165
Software Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1 Overview of Available Tools . . . . . . . . . . . . . . . 9.2 Graph-Based Process Optimization Tools . . . 9.2.1 PNS Solutions . . . . . . . . . . . . . . . . . . . . 9.2.2 S-Graph Studio . . . . . . . . . . . . . . . . . . . 9.3 Heat Integration Tools . . . . . . . . . . . . . . . . . . . . 9.3.1 SPRINT . . . . . . . . . . . . . . . . . . . . . . . . . .
191 191 191 191 193 195 195
155 157 159 159 161
166 167 167 169 171 173 176 176 177 179 179 180 183 186 190
ix
x
Contents 9.3.2 HEAT-int . . . . . . . . . . . . . . . . . . . . . . . . 9.3.3 STAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.4 SITE-int . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.5 WORK . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.6 HEXTRAN . . . . . . . . . . . . . . . . . . . . . . . 9.3.7 SuperTarget . . . . . . . . . . . . . . . . . . . . . . 9.3.8 Spreadsheet-Based Tools . . . . . . . . . . . Mass Integration Software: WATER . . . . . . . . Flowsheeting Simulation Packages . . . . . . . . . 9.5.1 ASPEN . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.2 HYSYS and UniSim Design . . . . . . . . . 9.5.3 gPROMS . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.4 CHEMCAD . . . . . . . . . . . . . . . . . . . . . . 9.5.5 PRO/II . . . . . . . . . . . . . . . . . . . . . . . . . . . General-Purpose Optimization Packages . . . . 9.6.1 GAMS . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6.2 MIPSYN . . . . . . . . . . . . . . . . . . . . . . . . . 9.6.3 LINDO . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6.4 Frontline Systems . . . . . . . . . . . . . . . . . 9.6.5 ILOG ODM . . . . . . . . . . . . . . . . . . . . . . Mathematical Modeling Suites . . . . . . . . . . . . . 9.7.1 MATLAB . . . . . . . . . . . . . . . . . . . . . . . . 9.7.2 Alternatives to MATLAB . . . . . . . . . . . Other Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.8.1 Modelica . . . . . . . . . . . . . . . . . . . . . . . . . 9.8.2 Emerging Trends . . . . . . . . . . . . . . . . . . 9.8.3 Balancing and Flowsheeting Simulation for Energy-Saving Analysis . . . . . . . . . 9.8.4 Integrating Renewable Energy into Other Energy Systems . . . . . . . . . . . . .
195 195 198 198 199 200 200 201 202 202 203 204 205 206 206 206 207 208 209 209 210 210 211 211 211 212
Examples and Case Studies . . . . . . . . . . . . . . . . . . . . 10.1 Heat Pinch Technology . . . . . . . . . . . . . . . . . . . 10.1.1 Heat Pinch Technology: First Problem 10.1.2 Heat Pinch Technology: Second Problem . . . . . . . . . . . . . . . . . . . 10.2 Total Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.1 Total Sites: First Problem . . . . . . . . . . . 10.2.2 Total Sites: Second Problem . . . . . . . . . 10.3 Integrated Placement of Processing Units and Data Extraction . . . . . . . . . . . . . . . . . . . . . .
219 219 219
9.4 9.5
9.6
9.7
9.8
10
215 216
224 226 226 231 234
Contents
11
12
10.4 Utility Placement . . . . . . . . . . . . . . . . . . . . . . . . 10.4.1 Utility Placement: First Problem . . . . 10.4.2 Utility Placement: Second Problem . . 10.5 Water Pinch Technology . . . . . . . . . . . . . . . . . . 10.5.1 Water Pinch Technology: First Problem 10.5.2 Water Pinch Technology: Second Problem . . . . . . . . . . . . . . . . . .
238 238 243 247 247
Industrial Applications and Case Studies . . . . . . . 11.1 Energy Recovery from an FCC Unit . . . . . . . . 11.2 De-bottlenecking a Heat-Integrated Crude-Oil Distillation System . . . . . . . . . . . . . . . . . . . . . . . 11.3 Minimizing Water and Wastewater in a Citrus Juice Plant . . . . . . . . . . . . . . . . . . . . . . . . 11.4 Efficient Energy Use in Other Food and Drink Industries . . . . . . . . . . . . . . . . . . . . . . . . . 11.5 Synthesis of Industrial Utility Systems . . . . . . 11.6 Heat and Power Integration in Buildings and Building Complexes . . . . . . . . . . . . . . . . . . . . . . 11.7 Optimal Design of a Supply Chain . . . . . . . . . 11.8 Scheduling a Large-Scale Paint Production System ................................
253 253
Typical Pitfalls and How to Avoid Them . . . . . . . . 12.1 Data Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1.1 When Is a Stream a Stream? . . . . . . . . 12.1.2 How Precise Must the Data Be at Each Step? . . . . . . . . . . . . . . . . . . . . . . . 12.1.3 How Can Considerable Changes in Specific Heat Capacities Be Handled? 12.1.4 What Rules and Guidelines Must Be Followed to Extract Data Properly? . . 12.1.5 How Can the Heat Loads, Heat Capacities, and Temperatures of an Extracted Stream Be Calculated? . . . . 12.1.6 How “Soft” Are the Data in a Plant or Process Flowsheet? . . . . . . . . . . . . . . . . 12.1.7 How Can Capital Costs and Operating Costs Be Estimated? . . . . . . . . . . . . . . . 12.2 Integration of Renewables: Fluctuating Demand and Supply . . . . . . . . . . . . . . . . . . . . . 12.3 Steady-State and Dynamic Performance . . . . .
281 283 284
249
256 262 268 271 275 277 279
285 286 287
289 290 290 292 292
xi
xii
Contents 12.4 Interpreting Results . . . . . . . . . . . . . . . . . . . . . . 12.5 Making It Happen . . . . . . . . . . . . . . . . . . . . . . .
293 293
13
Information Sources and Further Reading . . . . . . 13.1 General Sources of Information . . . . . . . . . . . . 13.1.1 Conferences . . . . . . . . . . . . . . . . . . . . . . 13.1.2 Journals . . . . . . . . . . . . . . . . . . . . . . . . . 13.1.3 Service Providers . . . . . . . . . . . . . . . . . 13.1.4 Projects . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2 Heat Integration . . . . . . . . . . . . . . . . . . . . . . . . . 13.2.1 Conferences . . . . . . . . . . . . . . . . . . . . . . 13.2.2 Journals . . . . . . . . . . . . . . . . . . . . . . . . . 13.2.3 Service Providers . . . . . . . . . . . . . . . . . 13.2.4 Projects . . . . . . . . . . . . . . . . . . . . . . . . . . 13.3 Mass Integration . . . . . . . . . . . . . . . . . . . . . . . . . 13.3.1 Conference . . . . . . . . . . . . . . . . . . . . . . . 13.3.2 Journals . . . . . . . . . . . . . . . . . . . . . . . . . 13.3.3 Service Providers . . . . . . . . . . . . . . . . . 13.3.4 Projects . . . . . . . . . . . . . . . . . . . . . . . . . . 13.4 Combined Analysis . . . . . . . . . . . . . . . . . . . . . . 13.4.1 Conferences . . . . . . . . . . . . . . . . . . . . . . 13.4.2 Journals . . . . . . . . . . . . . . . . . . . . . . . . . 13.4.3 Service Providers . . . . . . . . . . . . . . . . . 13.4.4 Projects . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5 Optimization for Sustainable Industry . . . . . . 13.5.1 Conferences . . . . . . . . . . . . . . . . . . . . . . 13.5.2 Journals . . . . . . . . . . . . . . . . . . . . . . . . . 13.5.3 Service Providers . . . . . . . . . . . . . . . . . 13.5.4 Projects . . . . . . . . . . . . . . . . . . . . . . . . . .
295 295 295 297 297 301 301 301 301 301 303 304 304 304 305 306 306 306 307 307 308 309 309 310 310 311
14
Conclusions and Further Information . . . . . . . . . . 14.1 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . 14.1.1 Books and Key Articles . . . . . . . . . . . . 14.1.2 Lecture Notes and Online Teaching Resources . . . . . . . . . . . . . . . . . . . . . . . . 14.2 Development Trends . . . . . . . . . . . . . . . . . . . . . 14.2.1 Top-Level Analysis . . . . . . . . . . . . . . . . 14.2.2 Maintenance Scheduling, Maintainability, and Reliability . . . . . 14.2.3 Hybrid Energy Conversion Systems . . . 14.2.4 Integration of Renewables and Waste
313 313 313 315 316 316 316 317 317
Contents 14.2.5 Better Utilization of Low-Grade Heat 14.2.6 Energy Planning That Accounts for Carbon Footprint . . . . . . . . . . . . . . . . . 14.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
319 320
Bibliography
.................................
321
.......................................
345
Index
319
xiii
This page intentionally left blank
Preface
T
his book describes and analyzes an amalgamation of two effective ways to considerably improve the efficiency and sustainability of processing industries: Process Integration and optimization. It is the result of collaborative efforts by two groups: researchers at the Centre for Process Integration, the University of Manchester, UK (a renowned center of excellence in this field), and the Faculty of Information Technology at the University of Pannonia, Hungary, as represented by the Centre for Process Integration and Intensification (CPI2) and the Centre for Advanced Process Optimization. The University of Pannonia centers are highly regarded for their achievements in optimization, and the P-graph and S-graph approaches to process optimization originated at this university. The book should provide support for graduate and postgraduate students worldwide as well as for Continuing Professional Development (CPD) courses and for practitioners from various fields of the processing industry. Its chapters analyze a number of problems of practical significance and also suggest various options for solving them to the benefit of modern society. The book provides a wealth of material for postgraduate teaching and further professional training. It is supported by the expertise stemming from the authors’ work as well as by a pool of case studies of varying complexity that have been collected over the years. This wide-ranging material presented here has been selected and refined over years of postgraduate teaching, further CPD courses, and training for the industry. It includes eight industry-based demonstration case studies and nine testing examples with the solutions developed. An unbiased evaluation and overview of the software tools available for learning, teaching, and industrial applications are also included. Text discussions are complemented by many figures to clarify details and enhance understanding. The book contains 14 chapters and a comprehensive bibliography. Chapter 1 is devoted to introducing and defining the field, and it also includes a basic assessment of energy efficiency. It starts with screening and scoping, which include auditing, benchmarking, and recommendations for good housekeeping. Next it describes an
xv
xvi
Preface important tool: balancing and flowsheeting simulation as a basis for optimization. It is on this basis that an integrated approach to optimization, Process Integration, is introduced. This approach is then connected to optimal process synthesis and combinatorial graphs. The important question of how to apply the Process Integration and Optimization Technology arises and is dealt with. This is further tackled and analyzed in the following chapters. Chapter 2 deals with the basic outline and definitions of Process Integration (PI). It begins with a historical and methodological introduction, briefly reviewing refinements and extensions of PI over its years of development. The thermodynamic roots of the Pinch Technology and of targeting heat recovery are introduced, followed by one of the key graphical constructions in the PI methodology: Composite Curves (CCs) for targeting process heat recovery. Supertargeting, or targeting for a full-fledged Heat Exchanger Network (HEN), is the next logical step. These tools are used to assess modifications of the Pinch idea for HEN retrofitting. Although PI was initially based on Heat Integration, an important spin-off was the development of integration for mass exchange and water networks. The chapter concludes with remarks on the role of PI in making industry sustainable. Chapter 3 introduces the other key methodology used for sustainable process design and synthesis: optimization. It presents a general framework and workflow of model building and follow-up optimization, including models that incorporate “black boxes” or “gray boxes.” It is critical to understand both the meaning and the mathematical formulation of optimization. Toward this end, the following questions are answered: What is optimization? What are the main classes of optimization problems? How are optimization problems formulated mathematically? What are the conditions for local or global optimality? This is followed by introducing deterministic algorithms for solving continuous linear and nonlinear, constrained and unconstrained, optimization problems. The most popular optimization methods and algorithms that employ stochastic search are also reviewed. The middle part of Chapter 3 is devoted to model creation. It includes a detailed description of conceptual modeling: extracting data about the operating units, identifying network and topology data, constructing equations to represent the processes, and finally choosing a right objective function. The last part of this chapter discusses how to handle complexity and nonlinearity as well as how to apply process insight when evaluating model adequacy and precision. Chapter 4 covers the core topic from which the development of PI began: improving the energy efficiency of individual processes and the PI extension into Total Sites. The chapter starts by introducing heat exchange, heat recovery, and heat exchange matches so that readers will be prepared for the remaining chapter content. The
Preface strategic view of Process Integration is outlined; this includes an overview of the hierarchy of process design, the meaning of performance targets, and the practical issue of identifying heat recovery problems from process flowsheets. These three topics are closely interrelated. Without properly applying the process design hierarchy, the performance targets cannot find a practical application and use. However, the process design hierarchy would be difficult to apply without employing performance targets and estimating upper bounds on process performance or lower bounds on total cost. Meaningful and practically useful heat-integrated designs cannot be obtained without appropriate identification of the heat recovery problem, a process referred to as data extraction. The chapter proceeds to describe the use of Composite Curves to set heat recovery targets, the Problem Table Algorithm for numerical targeting, and the Heat Recovery Pinch. These tools and concepts form the basis of Pinch Technology, defining the thermodynamic capabilities of the heat recovery problems. The more advanced aspects are discussed next; these include threshold problems, targeting multiple utilities via the Grand Composite Curve, and establishing targets for heat transfer area, capital cost, and total cost. There is a short overview of options for modifying the core process (which defines the heat recovery problem) that highlights the usefulness of Pinch Technology and targeting for improving its energy efficiency. The chapter then focuses on the synthesis of Heat Exchanger Networks; the approach mainly follows the Pinch Design Method, but there is discussion of the superstructure-based and hybrid methods used for HEN synthesis. The next step is Total Site Integration, which provides the necessary knowledge for energy recovery over complete industrial complexes and sites. Chapter 5 deals with an extension of PI known as Mass Integration, the most widely used instance of which is Water Integration (WI). The chapter begins with a description of the methodology and bases for minimizing water use and maximizing water reuse, and the importance of legislatively imposed constraints is discussed. Best available techniques are analyzed and recommended for usage, and the concept of a water footprint is described. At this point, the stage is set for the main task: minimizing freshwater usage and wastewater effluents. For this, the methodology of Water Pinch Analysis is introduced. Also described is a related Mass Integration and targeting technique, the material recovery Pinch diagram. The chapter concludes with water minimization using the mathematical optimization approach. Both the WI and the mathematical approaches to optimization are illustrated with case studies. Chapter 6 addresses further PI opportunities that have arisen as the methodology was developed. These include: Hydrogen Networks Design and Management; Oxygen Pinch Analysis; combined analyses (energy-water, oxygen-water, Pinch-emergy, budget-income-time,
xvii
xviii
Preface materials reuse-recycling); supply chains; CO2 emissions targeting; regional resource management; heat-integrated power systems with decarbonization and low-temperature energy systems; and the integration of reliability, availability, and maintainability with process design. An example is given of a real-life PI problem involving pressure-drop considerations during heat transfer enhancement. Several recent applications are mentioned, including a Locally Integrated Energy Sector and extended Total Sites with multiple energy carriers. Chapter 7 presents an overview of process optimization from the perspectives of Mathematical Programming (MPR) and the P-graph. The main features of these frameworks are analyzed, and it is shown that the P-graph is better suited than MPR for solving combinatorial optimization problems and, in particular, problems involving the synthesis of process networks. Optimization of process scheduling is the next topic. The most popular models and representations of process schedules are analyzed, and an efficient tool for obtaining them is introduced: the S-graph. Chapter 8 presents an integrated view of PI and optimization. It discusses how to efficiently apply them jointly in process synthesis and how to combine them. The chapter presents a number of examples of the P-graph and S-graph frameworks applied to combinations of PI and optimization. These applications are grouped thematically: (1) optimal process synthesis, including examples on reaction networks, green biorefineries, and azeotropic distillation; (2) synthesis of general energy systems involving Heat Integration and optimal retrofit; (3) optimal scheduling for maximizing throughput and revenue; (4) minimizing emissions via optimal synthesis of advanced energy conversion systems using Fuel-Cell Combined Cycles; and (5) availability and reliability features. Chapter 9 reviews the software tools for process modeling, integration, and optimization. The engineering field of sustainable design is complex in terms of scales and relationships, which makes information technology and computer software essential for solving problems—preferably with a user-friendly interface. The chapter reviews a wide spectrum of tools, as follows: (1) graph-based process optimization (process network synthesis solutions implementing P-graphs and the S-Graph Studio software); (2) Energy and Mass Integration tools designed to optimize the implementation of Heat Integration (SPRINT, HEAT-int, HEXTRAN, SuperTarget, spreadsheetbased tools), Total Site Integration (STAR, SITE-int), power generation and combined heat and power (STAR, WORK), and water systems integration (WATER); (3) process flowsheeting and simulation packages developed or supported by the major players (Aspen Plus, HYSYS and UniSim, gPROMS, CHEMCAD, PRO/II); (4) generalpurpose optimization systems (GAMS, MIPSYN); (5) computer algebra systems; and (6) other tools.
Preface Chapter 10 is a collection of examples and case studies that support the material presented in previous chapters. This selection of problems was collected by the authors over decades of teaching and consulting. Most of the problems have step-by-step solutions that feature comments and guidance for mastering the methodology. The examples are organized in several thematic groups: basic Heat Integration, Total Sites, integrated placement of energy-intensive processes, placing utilities, and Water Pinch Technology. Chapter 11 contains advanced examples that are based on industrial case studies, most of which were performed and published by the authors. The case studies presented include: (1) retrofit of a fluid catalytic cracking unit process that featured a large-scale heat recovery network; (2) de-bottlenecking of a heat-integrated crude-oil distillation process; (3) water system optimization of a citrus juice plant; (4) efficient water use in other food industries; (5) synthesis of industrial utility systems; (6) heat and power integration in buildings and building complexes; (6) optimal supply chain design; and (7) scheduling a large-scale paint production system. Chapter 12 distills some important and rare expertise: typical pitfalls and how to avoid them. It addresses possible (and probable) difficulties encountered during problem formulation and data extraction. The chapter raises seemingly trivial yet fundamental questions: When is a stream a stream? How precise must the data be at each step? How can extreme changes in specific heat capacity be handled? What rules and guidelines must be followed to properly extract data? How can the heat loads, heat capacities, and temperatures of an extracted stream be calculated? How “soft” are the data in a plant or process flowsheet? How can capital costs and operating costs be estimated? The provided answers are based on long-term experience and could help readers to solve the right problem—that is, a problem that accurately reflects the reality of the process under consideration. The chapter also addresses the integration of renewables that exhibit fluctuating demand and supply rates as well as steady-state and dynamic performance considerations. The chapter concludes with recommendations for interpreting results and suggestions on the successful advocacy and implementation of sustainable and optimal design. Chapter 13 consists solely of information on sources for further reading and information gathering. The various sources of information about optimization and integration in the process industry are arranged as follows: (1) general sources of information, (2) Heat Integration, (3) Mass Integration, (4) combined analysis, and (5) optimization for sustainable industry. Each topic is subdivided into sections on conferences, journals, service providers, and projects. Chapter 14 presents conclusions and sources of further information. Suggestions for further reading point to books and
xix
xx
Preface lecture notes containing applications and details that could not be adequately described in a book of this size. The book ends with a comprehensive bibliography that provides details of works cited in the text and serves also as a source of information and directions for further study. This book is intended to increase awareness of the principal methodologies that could contribute to improving energy and water efficiency in processing plants while reducing their environmental impact. Many additional illustrative examples could well have been discussed, but space limitations required that the authors select the most important features from a variety of fields and processing industries; even so, hundreds of references are included for seekers of more information. If the authors have managed to raise the interest and awareness of readers, then this book has fulfilled its intentions. Jiří Klemeš Ferenc Friedler Igor Bulatov Petar Varbanov
Acknowledgments
W
e acknowledge the support from the European Community–funded project entitled “Integrated Waste to Energy Management to Prevent Global Warming,” or INEMAGLOW. We would like to thank Prof. Robin Smith and Mr. Simon Perry from the University of Manchester for numerous discussions about the finer points of PI and optimization methodology. We also acknowledge with much gratitude the input of all collaborators who contributed to this book. Their dedication, timely responses, and willingness to accept editorial comments and suggestions are greatly appreciated. We received invaluable help from the staff of the Faculty of Information Technology at the University of Pannonia in Hungary: Dr. Rozália Pigler-Lakner, Dr. István Heckl, Dr. Boton Bertók, Dr. Zoltán Süle, Mr. Máté Hegyháti, and Ms. Adrienn Sas. Substantial contributions were also made by Hon Loong Lam—林汉龙 and Zsófia Fodor, Ph.D. students at CPI2. Special thanks go to colleagues and close collaborators worldwide who shared their latest methodologies and case studies. It would be impossible to list all those involved, but we should like to mention a few: Prof. L. T. Fan, Department of Chemical Engineering, Kansas State University, Manhattan, Kansas, USA; Dr. Dominic C. Y. Foo, Department of Chemical and Environmental Engineering, University of Nottingham, Malaysia Campus; Prof. Zdravko Kravanja, Faculty of Chemistry and Chemical Engineering, University of Maribor, Slovenia; Prof. Valentin Ples¸u, Centre for Technology Transfer for the Process Industries, Department of Chemical Engineering, University POLITEHNICA, Bucharest; and Prof. Petr Stehlík, Brno University of Technology, Institute of Process and Environmental Engineering, UPEI VUT Brno, the Czech Republic. The authors appreciate the editing and production efforts of the following people: Taisuke Soda, Michael Penn, Stephen Smith, Richard Ruzycka, and Michael Mulcahy of McGraw-Hill, and Aloysius Raj and the staff at Newgen. It has been a great pleasure for us to work with all of you.
xxi
This page intentionally left blank
CHAPTER
1
Introduction and Definition of the Field 1.1
Introduction In recent years there has been increased interest in the development of renewable, non-carbon-based energy sources in order to combat the increasing threat of carbon dioxide (CO2) emissions and subsequent climatic change. More recently, the fluctuations and often large increases in the prices of oil and gas have further increased interest in employing alternative, lower-carbon or non-carbon-based energy sources. These cost and environmental concerns have led to increases in the industrial sector’s efficiency of energy use, although the use of renewable energy sources in major industry has been sporadic at best. In contrast, domestic energy supply has moved more positively toward the integration of renewable energy sources; this movement includes solar heating, heat pumps, and wind turbines. However, there have been only limited and ad hoc attempts to design a combined energy system that includes both industrial and residential buildings, and few systematic design techniques have been marshaled toward the end of producing a symbiotic system. This book provides an overview of the Process Integration and optimization methodologies and its application to improving the energy efficiency of not only industrial but also nonindustrial energy users. An additional aim is to evaluate how these methodologies can be adapted to include the integration of waste and renewable energy sources. Industrial production requires a considerable and continuous supply of energy delivered from natural resources—principally in the form of fossil fuels such as coal, oil, and natural gas. The increase in our planet’s human population and its growing nutritional
1
2
Chapter One demands have resulted in annual increases in energy consumption. Furthermore, many nations have accelerated their development in the last 10 years, and countries (such as China and India) with large populations have seen significant increases in energy demands. This growing energy consumption has also resulted in unsteady climatic and environmental conditions in many areas because of increased emissions of CO2, NOx, SOx, dust, black carbon, and combustion process waste (Klemeš et al., 2005a; Klemeš, Bulatov, and Cockeril, 2007). It has become increasingly important to ensure that the production and processing industries take advantage of recent developments in energy efficiency and in the use of nontraditional energy sources (Houdková et al., 2008; Lam, Varbanov, and Klemeš, 2010). The additional cost is related to the amount of emitted CO2 and often takes the form of a centrally imposed tax. A workable solution to this problem would be to reduce emissions and effluents by optimizing energy consumption, increasing the efficiency of materials processing, and also increasing the efficiency of energy conversion and consumption (Klemeš et al., 2005b). Although major industry requires large supplies of energy to meet production, it is not the only sector of the world economy that is increasing its demands for energy. The particular characteristics of the other sectors (e.g., transport, residential) make optimizing for energy efficiency and cost reduction more difficult than in traditional processing industries, such as oil refining, where continuous mass production concentrated in a few locations offers an obvious potential for large energy savings (Al-Riyami, Klemeš, and Perry, 2001). In contrast, for example, agricultural production and food processing are distributed over large areas, and these activities are not continuous but rather structured in seasonal campaigns. Hence, energy demands in this sector are related to specific and limited time periods, so the design of efficient energy systems to meet this demand is more problematic than in traditional, steady-state industries. This chapter proceeds by first outlining the field of energy efficiency, including its scope, actors, and main features. The next step is to describe energy-saving techniques generally and then to specify an integrated approach: Heat Integration. An increasingly prominent issue is assessing and minimizing emissions and the carbon footprint. The carbon footprint (CFP) is defined by the U.K. Parliamentary Office for Science and Technology as the total amount of CO2 and the other greenhouse gases emitted over the full life cycle of a process or product (POST, 2006). There have been numerous studies (see, e.g., Albrecht, 2007; Fiaschi and Carta, 2007) that emphasize the “carbon neutrality” of renewable sources of energy. However, even renewable energy sources make some contribution to the overall carbon footprint, and assessment studies frequently do not account for this. The carbon footprint should also be incorporated
Introduction and Definition of the Field into any product’s life-cycle assessment (LCA); see, for example, Masruroh, Li, and Klemeš (2006).
1.2 Energy Efficiency The task of saving energy—especially at a time of rising energy costs, demand, and carbon emissions—must be taken seriously by all communities and industries. Society is driven by the economics of individual situations, and no section of society worldwide can be expected to save energy at any cost. Thus, energy-saving measures must be considered within the context of such issues as environmental factors, legislatively imposed constraints, and pressure from conscientious consumers. The simplest and most obvious technique involves energy auditing and applying good housekeeping measures. In many cases even these simple measures are not fully understood or completed in sufficient detail. To undertake a worthwhile energy audit, correct measurements are necessary. Also, because in many cases energy demand is not constant and instead fluctuates considerably, the monitoring of energy consumption has to be performed over specific (or extended) periods of time. Recommended monitoring techniques are described by various sources: utility companies, such as SEMPRA ENERGY (2009); governmental agencies, such as the U.S. Department of Housing and Urban Development (2009); and international groups, such as the International Energy Agency (Mandil, 2005). Improvements in energy efficiency must often be achieved by more complex means, such as those associated with improved design and operation. It is of paramount importance that all energy-related processes operate with maximum efficiency and minimum energy input. These systems should also ensure that they are fueled as much as possible by low-value inputs or recycled wastes, such as process outputs—for example, “off-gases” and hotwater waste (AEA Technology, 2000). To ensure that systems are designed to be as efficient as best practice allows, optimization methods are frequently employed for grassroots design, retrofit, control, and intelligent support systems for processes, plants, and buildings. One technology that has a strong reputation for improving energy efficiency through better design is Pinch Technology (Linnhoff and Vredeveld, 1984), which has been in use for more than 20 years. This technology, through feedback from practical applications and industry professionals, has been continuously developed and expanded (Klemeš et al., 1997; Smith, 2005; Klemeš, Smith, and Kim, 2008). Details on the successful applications of Pinch Technology in various industrial sectors are described in Chapter 11. The sustainability of energy systems can also be considerably improved by making use of renewable energy sources (e.g., biofuels,
3
4
Chapter One wind, water power), which significantly reduce the generation of greenhouse gases. Implementing Combined Heat and Power (CHP) systems (AEA Technology, 2000), rather than a separate power system and heat system, can also substantially improve the efficiency of energy supply. In addition, the overall situation can be improved by certain fast-advancing technologies: heat pumps, compact heat exchangers, fuel cells (FCs), and intensified technologies. Some of these approaches are not yet fully commercialized but are gradually becoming available. Some examples discussed in AEA Technology (2000) are as follows: • Advanced gas turbines for both utility and industrial applications, including cogeneration (CHP). • Fuel cells are electrochemical devices that may be fueled by hydrogen, methane, or other organic fuels. High-temperature FCs (MCFC and SOFC) can also use cleaned and conditioned synthesis gas directly. These systems produce high-grade heat (above 500ºC) in addition to electrical power, and they are well suited to cogeneration. It is estimated that FCs typically emit 25 percent less CO2 than a gas turbine. Yet further advances are required before their full application becomes economically practical. One option is to integrate CHP and FCs (Varbanov et al., 2006). • Dividing-wall distillation technology (Triantafyllou and Smith, 1992; Hernández and Jiménez, 1999). This technology involves the separation of three components (or groups of components) in a mixture. In the past this would have required two distillation columns, with heating and cooling provided for each column. The dividing-wall technology combines the separation process into a single vessel to yield energy savings of about 30 percent and capital savings of about 25 percent (MW Kellogg, 1998). • Compact heat exchangers are generally made of thin metallic plates rather than tubes. The plates form complex and small flow passages that result in a large surface area for heat transfer per unit volume. Multistream versions of these exchangers can incorporate 12 or more streams. Compact heat exchangers can yield energy savings and also reduce the costs of capital and installation. In a case study at a U.K. refinery, potential capital savings ranged from 69 to 84 percent (EEO, 1993). Cogeneration is being increasingly applied in most sectors. For example, many oil refineries satisfy a large portion of their power demands by on-site generation, with the balance being supplied by externally purchased electricity. Usually all or almost all heating
Introduction and Definition of the Field needs are met by on-site generation of heat carriers (hot oil, steam, flue gases). CHP generation and even “tri-generation” (simultaneous production of heat, power, and cooling) offer an opportunity to reduce greenhouse gas emissions from the combined power grid– refinery system by utilizing fuel heat content more completely than do most existing power generation technologies. The improved utilization rate is achieved by recovering the heat left in the exhaust streams of the various power generating facilities (gas turbines, FCs).
1.3 Screening and Scoping: Auditing, Benchmarking, and Good Housekeeping Over the years, screening and scoping tools have had a considerable effect on reducing the costs of energy and of treating effluents, thereby improving plant profit margins. For example, energy audits performed on various food and drink processes have resulted in cost savings of 15–30 percent and in attractive returns on investment (NRCan, 2007; U.S. DOE, 2007). Because profit margins are generally small in this sector, efficient management of energy is crucial for increasing profits while simultaneously reducing the production plants’ environmental impacts. The Carbon Trust (2009) has suggested the following steps for reducing energy consumption and thus improving energy efficiency. An analogous approach can be used for optimizing the use of water and wastewater: • Good housekeeping: (1) improving staff attitude and awareness; (2) locating heat leakages; (3) preventive maintenance; (4) insulation; (5) justifying use of heating, cooling, and lighting; (6) prevention or reduction of fouling; and (7) monitoring and control. • Energy audits: (1) examining records of energy cost and consumption; (2) producing an “energy balance sheet”; (3) providing high-quality data on energy consumption and costs; (4) collecting and processing data regularly (recommended for the analysis and review of energy information); and (5) establishing a benchmark of energy consumption based on other organizations or accepted standards. • Energy Efficiency Environmental Management System, ISO 14001: (1) the management system is a network of interrelated elements; (2) these elements include responsibilities, authorities, relationships, functions, processes, procedures, practices, and resources; (3) the management system
5
6
Chapter One establishes policies and objectives and also develops ways of applying the policies to achieve the objectives. • Responsible use of energy: in (1) energy procurement, (2) metering and billing, (3) performance measurement, (4) policy development, (5) assigning energy management responsibility, (6) energy surveying and auditing, (7) training and education, and (8) capital project management. The steps listed here have been found by the U.K. Energy Efficiency Best Practice Programme (EEBPp, 2002) to be central to any resource and waste management program. Energy screening and scoping audits fall into three main categories whose use (either individually or in combination) is based on the required depth of the study, the process to be analyzed, and the plant size: (1) the walk-through audit, which provides a quick snapshot of certain opportunities; (2) the detailed audit, which conducts an in-depth analysis of specific components; and (3) the Process Integration audit, which analyzes the plant as a whole and takes a systematic look at all processing steps and their interconnections. Audits may be performed by plant personnel, by external experts, or by teams with members from both groups. Although there are many possible options and levels of detail, typical activities include the following (NRCan, 2007): 1. Determination of the production “base case” and the reference period. 2. Collection of energy total consumption and cost (information usually available from fuel records, electricity invoices, etc.). 3. Development of a process flow chart that shows materials, energy inputs, and energy outputs for the main processing steps. 4. For the largest consumers, collection of energy data from the plant metering devices, control systems, and process flow diagrams (if current operating conditions are close to design data). In terms of the time required for (and other costs of) identifying opportunities to save energy, one efficient approach is the top-level analysis (Varbanov et al., 2004). This procedure accounts for investment limits or allowances and identifies economically justifiable energysaving opportunities, listing them in the order of their expected economic return. Measurements can be performed using portable instruments (flow rate, temperature, humidity, etc.) to determine the overall plant production of steam, refrigeration, compressed air, and hot water. Interviews with key personnel and operators can also provide valuable information about plant operations.
Introduction and Definition of the Field
1.4 Balancing and Flowsheeting Simulation as a Basis for Optimization Balancing reconciliation and flowsheeting simulation tools are frequently used for sustainability design and savings analysis; in fact, they have become the main tools in a process engineer’s toolbox. These tools help engineers to develop complete material and energy models based on measurements and/or design values and mathematical models. Consequently, these simulation tools play an important role in the technical and economic decision-making activities related to the planning and/or design stage of processes under development and to the operation of existing equipment. A number of computer-based systems have been developed to help process engineers calculate energy and mass balances. However, ongoing development costs have left only a few of these systems on the market: those whose positions have been secured by a substantial number of sales. An early overview of flowsheeting simulation was presented by Klemeš (1977). The balancing, data validation, and reconciliation technology involves a set of procedures incorporated into a software tool. Process data reconciliation has become the main method for monitoring and optimizing industrial processes as well as for component diagnosis, condition-based maintenance, and online calibration of instrumentation. According to Heyen and Kalitventzeff (2007), the main goals of this technology are: (1) to detect and correct deviations and errors of measured data so that they satisfy all balance constraints; (2) to use knowledge about the process system and structure along with measured data to compute unmeasured data wherever possible, especially key performance indicators (KPIs); and (3) to determine the postprocessing accuracy of measured and unmeasured data, including KPIs. More information about available software tools is provided in Chapter 9.
1.5
Integrated Approach: Process Integration Heat Integration is the first part of Process Integration, which provides the design foundation for CHP systems, refrigeration, air conditioning, and heat-pump systems. Process Integration is equally applicable to small, medium, and large industrial sites (e.g., power stations and oil refineries engaged in the production of petrochemicals). The technology answers one of the major challenges in the design of heating and cooling systems—namely, the complexity of energy and power integration—via a mapping strategy based on thermodynamically derived upper bounds on the system’s thermal and power performance. The efficient use of available heating and cooling resources for serving complex systems of various sizes and designations can significantly reduce energy consumption and emissions. This methodology can also be used to integrate renewable
7
8
Chapter One energy sources such as biomass, solar photovoltaic (PV), and solar thermal into the combined heating and cooling cycles. Since 1995, the energy consumption of European Community (EC) member countries has risen by 11 percent to the equivalent of 1637 Mt (megatons) of oil equivalent (Eurostat, 2007). This increase in energy consumption contrasts with the trend of the EC population, which is growing at only about 0.4 percent annually (Eurostat, 2007). The overall share of total energy consumption by industry is declining in most countries. However, domestic energy consumption is rising. In the United Kingdom, for example, residential consumption rose from 35.6 Mt (oil equivalent) in 1971 to 48.5 Mt in 2001—an increase of 36 percent— despite increases in energy efficiency (DTI, 2006). Process Integration Technology (Pinch Technology) has been extensively used in the processing and power generating industry for more than 30 years. It was pioneered by the Department of Process Integration, UMIST (now the Centre for Process Integration, CEAS, the University of Manchester), in the late 1980s and 1990s. Heat Integration is introduced in Chapter 2 and is described in more detail in Chapter 4. Water and mass integration is covered in Chapter 5, and recent developments in the field are reviewed in Chapter 6.
1.6
Optimal Process Synthesis and Combinatorial Graphs Process synthesis is a complex engineering activity that involves process modeling (e.g., chemical engineering) as well as combinatorial challenges. Although the basic process modeling has reached a considerable level of maturity, the combinatorial aspects of the engineering problem still leave significant room for improvement. One innovative approach to process synthesis is to exploit the combinatorial nature of network optimization. This approach is used by the process-graph (P-graph) framework, which explicitly defines sets of process materials and operations and then uses efficient combinatorial algorithms to build a rigorous network superstructure that can be reduced to the optimal network topology. This is different from the Mathematical Programming (MPR) approach, where the combinatorial aspects are modeled by algebraic equations and the structural features are blended with the underlying process models. These approaches are covered in Chapters 3, 7, and 8. The P-graph framework has been successfully applied to and demonstrated on several cases of energy system design. For example, Varbanov and Friedler (2008) explored FC-based systems in a case study that evaluated energy conversion systems to reduce CO2 emissions via Fuel-Cell Combined Cycle (FCCC) subsystems that utilize biomass and/or fossil fuels. The combinatorial complexity of the problem is efficiently handled by using P-graph framework and
Introduction and Definition of the Field algorithms. The authors developed a methodology for synthesizing cost-optimal FCCC configurations that accounts for the system carbon footprint. Their results show that the high energy efficiency of such systems when using renewable fuels makes them economically viable for a wide range of conditions. This study and other case studies (and guided applications) are provided in Chapters 10 and 11.
1.7
How to Apply the Process Integration and Optimization Technology The crucial topic of applications is addressed in Chapter 12, and Chapters 13 and 14 discuss sources of further information for those who would like to learn more or who seek qualified help from leading researchers and providers worldwide. Although every effort has been made to ensure that the provided information is as comprehensive as possible, it cannot be fully exhaustive. The field of Process Integration and optimization is developing rapidly, and every month brings important advancements. For this reason, the reader is strongly encouraged to remain up-to-date by reading about and exploring new developments.
9
This page intentionally left blank
CHAPTER
2
Process Integration 2.1
Introduction: The Need for Process Integration Energy and water saving, global warming, and greenhouse gas emissions have become major technological, societal, and political issues. These issues are of strategic importance because they are all closely related to energy supply. Numerous studies have been performed on the subject of improving energy efficiency while reducing emissions of greenhouse gases, volatile organic compounds, and other pollutants. In response to these industrial and societal requirements, several novel methodologies emerged since 1970. They included process system engineering (Sargent, 1979; Sargent, 1983) and “Process Integration” (Linnhoff et al., 1982; Linnhoff et al., 1994) followed by a number of works from the UMIST Group. Both disciplines were involved in dedicated conferences such as ESCAPE (European Symposium on Computer Aided Process Engineering), which was facilitated by the European Federation of Chemical Engineering Working Party on Computer Aided Process Engineering (CAPE, 2009), and PRES (Conference on Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction; PRES, 2009), which is supported on an annual basis by chemical and chemical engineering societies (e.g., Hungarian Chemical Society, Czech Society of Chemical Engineering, Italian Association of Chemical Engineering, Canadian Society for Chemical Engineering). It has gradually become evident that resource inputs and effluents of industrial processes are often connected to each other. Examples of this connection include the following: 1. Reducing external heating utility is usually accompanied by an equivalent reduction in the cooling utility demand (Linnhoff and Flower, 1978; Linnhoff et al., 1982; Linnhoff et al., 1994); obviously, this also tends to reduce the CO2 emissions from the corresponding sites. 2. Reducing wastewater effluents usually leads to reduced freshwater intake (Wang and Smith, 1994; Bagajewicz, 2000; Thevendiraraj et al., 2003).
11
12
Chapter Two Reducing the consumption of resources is typically achieved by increasing internal recycling and reuse of energy and material streams instead of fresh resources and utilities. Projects for improving process resource efficiencies can offer economic benefits and also improve public perceptions of the company undertaking them. However, motivating, launching, and carrying out such projects requires proper optimization studies that are based on adequate models of the process plants.
2.2
What Is Process Integration? Process Integration (PI) is a family of methodologies for combining several processes to reduce consumption of resources and/or to reduce harmful emissions. It started as mainly Heat Integration (HI), stimulated by the energy crisis of the 1970s (Hohmann, 1971; Linnhoff and Flower, 1978; Linnhoff, Mason, and Wardle, 1979; Linnhoff et al., 1982; Linnhoff and Hindmarsh, 1983; Linnhoff and Vredeveld, 1984). This energy-saving methodology has been used extensively in the processing and power generating industry over the last 30 years. Heat Inetegration examines the potential for improving and optimizing the heat exchange between heat sources and heat sinks in order to reduce the amount of external heating and cooling required, thereby reducing costs and emissions. A systematic, rule-based design procedure has been developed that yields the maximum energy-saving design for a given system. There are several definitions of HI (using Pinch Technology); most refer to the thermal combination of steady-state process streams or batch operations for achieving heat recovery via heat exchange. More broadly, the definition of PI, as adopted by the International Energy Agency (Gundersen, 2000) is as follows: systematic and general methods for designing integrated production systems ranging from individual processes to total sites and with special emphasis on the efficient use of energy and reducing environmental effects.
2.3 History and Development of Process Integration It is remarkable that PI continues to interest researchers even 35 years after its emergence. The HI, which developed as the first part of PI, deals with the integration of heat in Heat Exchanger Networks (HENs). This methodology has been shown to have considerable application potential for complete chemical processing sites, reducing overall energy demand and emissions across the site and thus leading to a more effective and efficient site utility system. The PI and its subset HI method have also been successfully applied to the cogeneration of heat and shaft power. Further details are available elsewhere
Process Integration (Linnhoff et al., 1982; Linnhoff et al., 1994; Shenoy, 1995; Smith, 2005; El-Halwagi, 2006; Kemp 2007; Klemeš, Smith, and Kim, 2008). One of the first works in this field was Hohmann’s (1971) PhD thesis, which introduced a systematic thermodynamics-based reasoning for evaluating the minimum energy requirements for a given HEN synthesis problem. In the late 1970s this work was continued by Linnhoff and Flower, who used Hohmann’s foundation to develop the basis of Pinch Technology—now considered the cornerstone of HI. As is often the case with a pioneering innovation, this work was difficult to publish. Yet the authors’ strong commitment eventually led to the publication of their ideas in Linnhoff and Flower (1978), which has since become the most cited paper in the history of chemical engineering. Similar work (Umeda et al., 1978; Umeda, Harada, and Shiroko, 1979) was independently published in Japan, but it was Linnhoff (supported by teams from UMIST and later Linnhoff March Ltd.) who pushed the new concept through academia and industry. The publication of the first “red” book by Linnhoff et al. (1982) played a key role in the dissemination of HI methodology. This user’s guide to Pinch Analysis detailed the most common process network design problems, including HEN synthesis, heat recovery targeting, and selecting multiple utilities. These methodologies were developed and pioneered by the Department of Process Integration, UMIST (now the Centre for Process Integration, CEAS, the University of Manchester) in the late 1980s and 1990s (Linnhoff et al., 1982; Linnhoff and Vredeveld, 1984; Linnhoff et al., 1994; Klemeš et al., 1997; Smith et al., 2000; Smith, 2005). A second edition of Linnhoff’s user’s guide was published by Kemp (2007). Applications of HI in the food industry were presented in Klemeš and Perry (2007a) and in Klemeš, Smith, and Kim (2008). Tan and Foo (2007) successfully applied the Pinch Analysis approach to carbon-constrained planning for the energy sector, and Foo, Tan, and Ng (2008) applied the cascade analysis technique to carbonfootprint-constrained energy planning. Another important part in process design and optimization is the synthesis phase of process flowsheets. From the earliest stages of PI there have been attempts to combine it with optimization (see, e.g., Giammatei, 1994). Such combining is usually performed after the targeting phase mentioned previously. Ideally, the structure of the entire process—and the configurations of the operating units within it—should be simultaneously designed and optimized, because the performance of each unit influences the others. The main source of complexity in this synthesis is the problem’s dual nature of being both continuous and discrete. There are several known methods for performing the task, including heuristic, evolutionary, and superstructure-based approaches. Two major classes of methods for process synthesis are heuristic and algorithmic (or Mathematical
13
14
Chapter Two Programming) methods. Hybrid methods have also been proposed; these approaches incorporate heuristic rules as well as Mathematical Programming (MPR). Much as with any decision making, process synthesis (or design) is an activity for which no past experience can be ignored, especially when it comes to localized details of the design. The most popular approach is to create a superstructure for the network being designed and then choose the best possible solution network from the superstructure options.
2.4 Pinch Technology and Targeting Heat Recovery: The Thermodynamic Roots Furman and Sahinidis (2002) have compiled a comprehensive review of works tracing the development of the research on HENs in time. Their study shows that there was only mild interest in heat recovery and energy efficiency until the early 1970s, by which time just a few works in the field had appeared. But between the oil crises of 1973– 1974 and 1979 there were significant advances made in HI. Although capital cost remained important, the major focus was on saving energy and reducing related costs. It is exactly this focus that resulted in attention being paid to energy flows and to the energy quality represented by temperature. The result was the development of Pinch Technology, which is firmly based on the first and second laws of thermodynamics (Linnhoff and Flower, 1978). In this way HEN synthesis—one of the most important and common tasks of process design—has become the starting point for the PI revolution in industrial systems design. HENs in industry are used mainly to save on energy cost. For many years the HEN design methods relied mostly on heuristics, as necessitated by the large number of permutations in which the necessary heat exchangers could be arranged. Masso and Rudd (1969) is a pioneering work that defines the problem of HEN synthesis; the paper proposes an evolutionary synthesis procedure that is based on heuristics. An alternative HEN synthesis method is described in Zhelev et al. (1985), one that exploits synergies between heuristics and combinatorics. In this approach, several suboptimal networks are synthesized, and in most cases they are less integrated (consist of more than one subnetwork). A complete timeline and thorough bibliography of HEN design and optimization works is provided in Furman and Sahinidis (2002). The paper covers many more details, including the earliest known HEN-related scientific article: Ten Broeck (1944). The discovery of the Heat Recovery Pinch concept (Linnhoff and Flower, 1978) was a critical step in the development of HEN synthesis. The main idea behind the formulated HEN design procedure was to obtain—prior to the core design steps—guidelines
Process Integration and targets for HEN performance. This procedure is possible thanks to thermodynamics. The hot and cold streams for the process under consideration are combined to yield (1) a Hot Composite Curve representing, collectively, the process heat sources (the hot streams); and (2) a Cold Composite Curve representing the process heat sinks (the cold streams). For a specified minimum allowed temperature difference ΔTmin, the two curves are combined in one plot (see Figure 4.7), providing a clear thermodynamic view of the heat recovery problem. The overlap between the two Composite Curves represents the heat recovery target. The overlap projection on the heat exchange axis represents the maximum amount of process heat being internally recovered. The vertical projection of the overlap indicates the temperature range where the maximum heat recovery should take place. The targets for external (utility) heating and cooling are represented by the nonoverlapping segments of the Cold and Hot Composite Curves, respectively. The methodology is described in more detail in Chapter 4.
2.5 Supertargeting: Full-Fledged HEN Targeting After obtaining targets for utility demands of a HEN, the next logical step is to estimate targets for capital and total costs. Capital costs in HENs are determined by many factors, of which the most significant is the total heat transfer area and its distribution among the heat exchangers. Townsend and Linnhoff (1984) proposed a procedure for estimating HEN capital cost targets by using the Balanced Composite Curves, which are obtained by adding utilities to the Composite Curves obtained previously (see Figure 4.7). The HEN heat transfer area target is computed from the enthalpy intervals in the Balanced Composite Curves by using the heat transfer coefficients given in the HEN problem specification (assuming “vertical heat transfer” and “spaghetti-type topology”). Improvements to this procedure that have been proposed involve one or more of the following factors: 1. Obtaining more accurate surface area targets for HENs that exhibit nonuniform heat transfer coefficients (Colberg and Morari, 1990; Jegede and Polley, 1992; Zhu et al., 1995; Serna-González, Jiménez-Gutiérrez, and Ponce-Ortega, 2007). 2. Accounting for practical implementation factors, such as construction materials, pressure ratings, and different heat exchanger types (Hall, Ahmad, and Smith, 1990). 3. Accounting for additional constraints such as safety and prohibitive distance (Santos and Zemp, 2000).
15
16
Chapter Two Cost estimation usually has a significant impact on a project’s predicted profitability. Taal and colleagues (2003) summarized the common methods used for cost estimation of heat exchange equipment and the sources of energy price projections. This paper showed the importance of choosing the right cost estimation method and a reliable source of energy price forecast when retrofit projects are evaluated for viability. Both retrofit and grassroots design projects require accurate cost estimates in the project’s early stages so that the correct choices are made. There are several methods available, most of which lead to FOB (Free On Board) cost estimates. As for operational costs, the range of possible future energy prices may prove crucial when the margins of a proposed design or retrofit project are analyzed. This is especially true for a large project that has a long payback period and consumes a lot of energy. Such projects are typical of plants in the chemical, petrochemical, refinery, and paper industries. Taal et al. (2003) include a brief review of oil and natural gas price projections reported by centers of excellence in the energy economics field. There are numerous forecasts, sometimes contradictory, but general trends can be observed. Some general insights in oil and gas price and production behavior are mentioned and refer to more detailed sources.
2.6 Modifying the Pinch Idea for HEN Retrofit Bochenek, Jezowski, and Jezowska (1998) compared the approaches of optimization versus simulation for retrofitting flexible HENs. This is an important work that should have generated additional research. Zhu, Zanfir, and Klemeš (2000) proposed a heat transfer enhancement methodology, for HEN retrofit design, from which HI could benefit substantially. This approach is worthy of wider implementation, especially in the context of retrofit studies. Heat Exchanger Network retrofit is a special case of optimization. In retrofit problems, one must accommodate an existing network with existing heat exchangers that are already paid for. This circumstance substantially alters the economics of the problem as compared with a new design. One example of this approach to retrofit was the paper of Tjoe and Linnhoff (1986), which suggested identifying heat exchangers with cross-Pinch heat transfer and, where appropriate, attempting to replace such heat exchangers with others that do not transfer heat across the Pinch, thereby reducing energy consumption. However, an ideal new design would not account for the existing HEN equipment and topological constraints. One method for overcoming these drawbacks is the Network Pinch (Asante, 1996; Asante and Zhu, 1997), which uses continuous nonlinear optimization to identify the bottlenecking heat exchangers within
Process Integration the existing network. The network bottleneck occurs at a heat exchanger, which constrains the load shifting to gain further improvement in heat recovery. At least one of the heat exchanger sides exhibits the minimum allowed temperature difference between the involved streams, and the corresponding point in the network’s temperature-load plot is referred to as the Network Pinch. To overcome this Network Pinch, the network structure must be modified. Possible modifications include the relocation of an existing heat exchanger, the addition of a new exchanger, or a change in the stream splitting arrangement. To identify the most promising modifications, the Network Pinch method uses MPR guided by thermodynamic insights. Nonlinear optimization is employed to evaluate the capital energy trade-offs and to produce a new optimal structure. This approach allows retrofit to be carried out one step at a time, leaving the designer in control to accept or reject suggested modifications at each step. There have been some successful practical applications of these targeting and retrofit methodologies. Pleşu, Klemeš, and Georgescu (1998) demonstrated the wide applicability of PI in the Romanian oil refining and petrochemical industry. In one of the first comprehensive retrofit case studies, Hassan, Klemeš, and Pleşu (1999) presented a PI analysis and retrofit suggestions for a fluid catalytic cracking (FCC) plant. Pinch Technology and its recent extensions offer an effective and practical method for designing the HEN for new and retrofit projects. Al-Riyami, Klemeš, and Perry (2001) demonstrated a HI retrofit analysis for the HEN of an FCC plant. Their study found significant room for improvement in the heat recovery process, and the new network was designed using the Network Pinch method.
2.7 Mass Exchange and Water Networks Water is widely used in the processing industries as an important raw material. It is also frequently used as a utility (e.g., steam or cooling water) and as a mass transfer agent (e.g., for washing or extraction). Large amounts of high-quality water are consumed in many industries that face strict requirements for product quality and the associated manufacturing safety issues. The processing industry is characterized by complex design and operation of storage and distribution systems for water management. Today’s industrial processes and systems that use water are subjected to increasingly stringent environmental regulations on the discharge of effluents. Increases in population and its quality of life have led to increased demand for freshwater. The rapid pace of these changes has made more urgent the need for improved water management. Adopting techniques to minimize both water consumption and wastewater
17
18
Chapter Two discharge can considerably reduce demand for freshwater and also the amount of effluent generated by processing. For these reasons, the success of HI has inspired researchers to apply the Pinch and PI concepts to other areas—in particular, to mass exchange networks (El-Halwagi and Manousiouthakis, 1989). Wang and Smith (1994) developed a method for industrial water networks as a special case of mass exchange networks (see Figure 5.1). Their main objective was to minimize the consumption of freshwater and the disposal of wastewater simultaneously by maximizing the reuse of internal water, again exploiting the idea of recycling and reusing valuable streams and materials in order to save resources and reduce emissions. Wastewater can be further reduced by applying additional techniques for water regeneration that enable further reuse or recycling. For the case of a single contaminant, translating the method of Pinch Analysis to water minimization is straightforward: the water’s Composite Curve is used to construct a plot of contaminant concentration versus contaminant load. Extending the Water Pinch Analysis to multiple-contaminant problems is a complicated and difficult procedure. The principal issue concerns determining which contaminant to use on the Y axis when plotting the Composite Curves. Several approaches have been proposed. One option is to employ MPR, in which case Water Pinch serves as a preliminary visualization tool (Doyle and Smith, 1997). Foo, Manan, and Tan (2005) applied Water Pinch Analysis to synthesize optimal water recovery networks for batch processes. These authors introduced a numerical technique (time-dependent water cascade analysis) that has the advantage of clearly depicting the time-dependent nature of batch water networks. Majozi (2005) also employed mathematical modeling—the mixed integer nonlinear programming (MINLP) approach—to devise an effective technique for wastewater minimization in batch processes. Here, too, wastewater was minimized via application of reuse and recycle.
2.8
Benefits of Process Integration Heat recovery targeting for HEN synthesis problems is based on Composite Curves described in Section 2.4 (Linnhoff, Mason, and Wardle, 1979). The Composite Curves plot is a visual tool that summarizes the important energy-related properties of a process in a single view (see Figure 4.7). It was the resulting recognition of the thermodynamic relationships and limitations in the underlying heat recovery problem that led to development of the Pinch Design Method (Linnhoff and Hindmarsh, 1983), which is capable of producing maximally efficient heat recovery networks. As already discussed in this chapter, PI has been considerably expanded in scope since these initial applications. It is now also used for HEN
Process Integration retrofits, both water and combined water-and-energy minimization, as well as to minimize total site energy consumption—a process that includes Combined Heat and Power, locally integrated energy sectors, integration of renewables and waste-to-energy techniques, and combinatorial tools (P-graph and S-graph; see Chapter 7). In addition, recent applications have extended the PI approach to regional energy and emissions planning, financial planning, batch processes, and the targeting of other constrained resources, such as land, renewable energy, and emissions. In the wake of the initial breakthrough of Pinch Analysis for HEN synthesis, all these new PI applications follow the same simple logic: target setting should precede designing. In the most straightforward cases—such as HEN synthesis for Maximum Energy Recovery (MER) and water network synthesis for maximum water reuse—the targets can be interpreted as indicators of what a rigorous application could actually achieve. However, the applicability and benefits of PI are not limited to these straightforward cases. In fact, the target setting can be applied in various contexts and still yield enormous benefits in terms of reduced computational and project development time. Klemeš, Kimenov, and Nenov (1998) described several applications of Pinch Technology within the food industry, work that was further developed in Klemeš et al. (1999). This research showed that Pinch Technology can provide benefits far beyond oil refining and petrochemicals. The most important property of thermodynamically derived heat recovery targets is that they cannot be improved upon by any real system. Composite Curves play an important role in process design; for HEN synthesis algorithms, they provide strict MER targets. For process synthesis based on MPR, the Composite Curves establish relevant lower bounds on utility requirements and capital cost, thereby narrowing the search space for the following superstructure construction and optimization. The preceding observation highlights an important characteristic of process optimization problems, and specifically those that involve process synthesis and design. By strategically obtaining key data about the system, it is possible to evaluate processes based on limited information—before too much time (or other resources) are spent on the study. This approach follows the logic of oil drilling projects: potential sites are first evaluated in terms of key preliminary indicators, and further studies or drilling commence only if the preliminary evaluations indicate that the revenues could justify further investment. The logic of this approach was systematically formulated by Smith in his books on PI for process synthesis (Smith, 1995; Smith, 2005) and by Daichendt and Grossmann (1997), whose paper integrated hierarchical decomposition and MPR to solve process synthesis problems.
19
20
Chapter Two
2.9 The Role of PI in Making Industry Sustainable Sustainability in the design and operation of industrial processes is becoming a key issue of process systems engineering. There are MPR tools for designing chemical processes that minimize costs (and thus maximize profits). These MPR formulations typically are highly complex; therefore, in practical applications, the original problem is usually decomposed into smaller problems that are solved sequentially. There are several definitions of sustainability. It can be defined as the capacity to sustain the viability of a given system. In this context, sustainable development implies that current actions should not harm future generations and that measurable indicators are needed to ensure compliance. For industrial processes, the relevant indicators are the rates of resource intake and effluent emissions. These factors are usually expressed in terms of “footprints” (De Benedetto and Klemeš, 2009)—for example, the carbon footprint (CFP) and/or water footprint (WFP) of a process. The sustainability of an industrial process depends on minimizing these footprints. For synthesis and design problems, the measured indicators should apply to the proposed system’s complete life cycle as part of a life-cycle assessment (LCA). In short, PI has a direct impact on improving the sustainability of a given industrial process. All PI techniques are geared toward reducing the intake of resources and minimizing the release of harmful effluents, goals that are directly related to the corresponding footprints. Hence, employing PI and approaching the targeted values will help minimize those footprints. Footprint considerations, however, are only part of an industry’s ultimate indicator—that is, its economic performance. This metric takes the form of either cost or profit, depending on the selected system boundaries. Thus, no matter how environmentally attractive a given project may be, it will probably not be implemented unless doing so is economically feasible. There are several ways to merge the objectives of sustainability and profitability. The two most popular are (1) expressing footprints in strictly economic terms and then folding them into the cost or profit objective function; and (2) employing multiobjective optimization, whereby footprint and economic indicators are evaluated in parallel as objectives to be achieved.
2.10
Examples of Applied Process Integration In this section, four examples illustrate the potential of PI to reduce not only resource demand but also operational and capital costs. The problem areas vary widely, yet in each case the application of the PI techniques yields significant benefits. Only the main problem points and outcomes are described here; see Chapter 11 for more details.
Process Integration
Petrochemicals: Fluid Catalytic Cracking Process The FCC unit is a major process element in oil refineries, and improvements in yield and efficiency have been attempted over the years in response to various external drivers. Retrofitting the HEN associated with an FCC usually leads to improved energy recovery and thus to reduced energy use and/or increased throughput. In this example, the HEN of the FCC process includes a main column and a gas concentration section (Al Riyami, Klemeš, and Perry, 2001). The particular FCC plant considered had 23 hot streams and 11 cold streams, and the associated cost and economic data required for the analysis were specified by the refinery owners. The task was to analyze the existing process and then propose a HEN retrofit plan for improving energy recovery. Incremental area efficiency was used for the targeting stage of the retrofit design, which was carried out using the Network Pinch method (Asante and Zhu, 1997) consisting of a diagnosis stage and an optimization stage. In the retrofit, four heat exchangers were added and one existing exchanger was removed. The resultant design produced energy savings of 8.955 MW or 74 percent. This translated into a 27 percent decrease in the plant’s utility bill for an annual savings of $2,388,600. The modified HEN required an investment of $3,758,420, so the payback period was less than 19 months.
Energy Integration of a Hospital Complex Herrera, Islas, and Arriola (2003) studied a hospital complex that included an institute, a general hospital, a regional laundry center, a sports center, and some other public buildings. The diesel fuel used to generate the steam required for the complex amounted to 75 percent of its total energy consumption and 68 percent of its total energy cost ($396,131 in 1999). The Pinch Analysis that was performed estimated the minimum need for external heating at 388.64 kW. Because the actual heating used was consuming 625.28 kW, there existed a potential energy savings of 38 percent.
Sunflower Oil Production As reported by Klemeš, Kimenov, and Nenov (1998), an oil production process operated with a minimal temperature difference of 65°C at the Process Pinch has been analyzed. The external heating required by the system was provided by two types of hot utilities, and the required external cooling was provided by two cold utilities. The study resulted in increasing heat recovery and reducing the minimum temperature difference to 8–14°C. This reduced the hot and cold utility requirements and eliminated the need for water steam and cooling water, which considerably simplified the overall design.
A Whisky Distillery In a study by Smith and Linnhoff (1988), the authors found that steam was being used below the Process Pinch; this resulted in unnecessarily
21
22
Chapter Two high utility usage by the distillery. The steam in question resulted from the use of a heat pump, and so the steam below the Process Pinch was eliminated by reducing the heat pump’s size. Although this approach required that the steam now be used for process heating above the Process Pinch, overall energy costs were still reduced owing to the reduced compressor duty.
2.11
Summary This chapter offered an introduction to and an overview of the field of PI, its HI roots, the expansion of its application areas, and the evolution of process optimization methods toward combining PI techniques and insights with the tools of MPR. The chapter discussed the philosophy of PI as well as its contribution to sustainable development and the economic efficiency of projects. There is a wealth of useful information on these topics, and a number of excellent sources of further information are described in Chapters 13 and 14.
CHAPTER
3
Process Optimization
T
his chapter deals with process optimization: its definition, goals, and application areas within sustainable industrial process design and integration. The aim is to provide information on how to formulate sustainability tasks as optimization problems and on what tools to employ for solving them. The chapter begins with a brief description of the general framework for model building and optimization; this is followed by basics of optimization problems and their classes as well as descriptions of the most common algorithms for solving optimization problems. Finally, the chapter discusses how to build models efficiently, how to handle complexity, and how to ensure model adequacy and sufficient precision. The details of computational implementations of optimization solvers and other software tools are given in Chapter 9.
3.1
Introduction Building and operating industrial processes entail costs and environmental impacts. Emissions and effluents include: gaseous waste streams, which may harbor CO2, SOx and NOx; wastewater and various aqueous streams; and flue gases. When attempting to improve the environmental and economic performance of process systems, it is important to keep in mind that the processing paths, which connect the various system inputs and outputs, usually interact with each other. Therefore, minimizing resource demands and environmental impacts is greatly facilitated by properly modeling the process systems and then deciding which designs and operating policies to pursue—in what priority and to what extent. Maintaining a balance between model accuracy and simplicity is necessary in order to derive meaningful results with minimal computational expense. A system model can be created for different purposes. A lumped steady-state model (i.e., one that neglects variations in time and space) will contain only algebraic equations. To simplify the problem, steady-state models assume that the operating units are “black boxes” or “gray boxes” (a black box is a model that represents an empirical process in terms of its input, output, and transfer parameters but does not describe any internal physics; a gray
23
24
Chapter Three box is a model that incorporates a physical representation, although some of the physics is approximated—see, e.g., Hangos and Cameron, 2001). This approach is appropriate for optimizing complex systems. If time variability has to be accounted for then steady-state modeling can be applied to a set of operating periods, each of which is characterized by its own fixed parameters.
3.2 Model Building and Optimization: General Framework and Workflow A good process model should contain a thorough conceptual description of the involved phenomena, unit operations, actions, events, and so forth. Usually this description involves text, flowsheets, and structural diagrams. Additionally, IT-domain diagrams—for example, UML diagrams—can be used (“UML” is a specification of the Object Management Group; UML, 2010). The UML diagrams include class, object, package, use case, sequence, collaboration, statechart, component, and deployment diagrams. A good process model should also contain a sufficiently precise mathematical description. The mathematical relationships are used to reflect not only physical laws but also technological constraints and company rules. Mathematical models include algebraic equations of some form (i.e., equalities or inequalities) and may be supplemented with dynamic modeling, which uses differential equations to capture variations in time, as well as states and actions to express operational procedures and other dynamic relationships of algorithmic nature. Structural information is also an essential feature of process network models. When translated to Mathematical Programming (MPR) models, such information is expressed by integer (mostly binary) variables. An efficient alternative to representing superstructures with binary variables is the P-graph and its related framework (Friedler et al., 1992b), discussed in Chapter 7. An efficient computational implementation of the mathematical description may take the form of a stand-alone compiled application (e.g., PNS Editor, 2010) or may be modeled within a popular environment for process and mathematical calculations. Examples include MATLAB (MathWorks, 2009), Scilab (2009), simulation and optimization tools tailored for the process industry (AspenTech, 2009c), Modelica (2009a; OpenModelica, 2010), Honeywell UniSim (Honeywell, 2010), and the open-source DWSIM (2010). All model components have to be well synchronized to provide appropriate user interfaces and sufficient visual aids to help understand the process and the optimization results. Models often include only the computational implementation with some mathematical descriptions, but a much better practice is to start with the concepts before deriving the mathematical relationships
Process Optimization and then finally implementing the model computationally. In the course of formulating the mathematical and implementation components, it may be necessary to make some corrections to the conceptual and/or mathematical components.
3.3 Optimization: Definition and Mathematical Formulation 3.3.1 What Is Optimization? Optimization can be applied to different tasks: new system design, the synthesis of a new processing network, and the retrofit design and operational improvements in heat exchanger, reactor, and separation networks. Optimization is employed to find the best available option. An objective function consists of a performance criterion to be maximized or minimized. The system properties that determine this function are of two types: 1. Parameters—a set of characteristics that do not vary with respect to the choice to be made 2. Variables—a set of characteristics that are allowed to vary Some of the variables are specified by the decision maker or are manipulated by the optimization tools; these are referred to as specifications or decision variables. The remaining variables are termed dependent variables, and their values are determined by the specifications and the system’s internal relationships. The objective function can be formulated in terms of a single variable or a combination of dependent and decision variables. The value of the objective function can be changed by manipulating the decision variables.
3.3.2 Mathematical Formulation of Optimization Problems Optimization tasks in industry include increasing heat recovery, maximizing the efficiency of site utility systems, minimizing water use and wastewater discharge, and other tasks. The formulations that are used to solve such optimization problems are known as mixed integer nonlinear programs (MINLPs). However, they are frequently linearized to yield the more tractable mixed integer linear programs (MILPs), and some can be further simplified and solved via linear programming (LPR). In general, optimization problems can be formulated as summarized in Table 3.1. The continuous and discrete domains, together with the constraints, define the feasible region for the optimization. This region contains the set of options from which to choose. The value of the function F depends on the values of the decision variables.
25
26
Chapter Three Minimize (or maximize) F(x, y) where
x ∈ Rn (continuous variables) y ∈ Z n (integer variables)
subject to h(x, y) = 0 g(x, y) ≤ 0 TABLE 3.1
Objective function, performance criterion Continuous domain Discrete domain Equality constraints Inequality constraints
Generic Optimization Problem
Continuous variables are used to model properties (e.g., flow rates and chemical concentrations) that vary gradually within the feasible region. Integer variables are used to model the status (ON versus OFF) of operating devices as well as the selection/exclusion of options for operating units in synthesis problems. Using only the problem formulations, it is possible to generate many combinatorially infeasible sets of the integer variable values, which are later analyzed by the optimization solver. Especially for larger problems, it’s a good idea to eliminate these infeasible combinations from the search space or to build into the optimization solver a mechanism for avoiding them (Friedler et al., 1996). The type of the objective function F dictates which extremum— the minimum or the maximum—to seek. Common performance criteria are to minimize the process cost or to maximize the profit. Because some process subsystems (e.g., water networks) do not usually generate useful product streams, no revenue is directly realized and so minimizing the total annualized cost is used instead as an objective function. For complete production systems and supply chains the objective is usually to maximize the profit. Thus, additional variables (reflecting sales and customer behavior) and their relationships may be added to the formulation. Equality constraints stem not only from material and energy balances but also from constitutive relations that normalize the stream compositions to unity. The balances include those for total flow rates, balances of the chemical components, and energy balances of heat exchangers, boilers, and turbines. Inequality constraints stem from limitations on concentrations, flow rates, temperatures, pressures, throughput, and so forth. One example of a constitutive relation is calculation of the fluid heat capacity flow rate from its mass flow rate and specific heat capacity.
3.4 Main Classes of Optimization Problems This section discusses methods that can be applied to detect optimality and solve optimization tasks. Choosing a particular
Process Optimization method should be based on a clear understanding, and toward this end it is useful to bear in mind three aspects of optimization problems: 1. An optimization problem is convex if it minimizes (or maximizes) a convex objective function and if all the constraints are convex (Williams, 1999). 2. An optimization problem is linear if its objective function and all the constraints are linear; note that all linear optimization problems are also convex (Williams, 1999). 3. A variable that can assume only integer (whole number) values is an integer variable; integer variables that are constrained only to values of 0 or 1 are binary variables. Optimization problems are classified in terms of the following specific features: • Objective type: Minimization or maximization problems. • Presence of constraints: Unconstrained versus constrained. Most practical tasks involve the formulation of constrained optimization problems. • Problem convexity: Convex versus nonconvex. • Linear and nonlinear problems: This aspect depends on the nature of the objective function and/or the constraints. Most process optimization problems are bilinear (i.e., containing products of two optimization variables), for example, they feature componentwise mass balances involving products of mass flow rates and concentration and enthalpy balances involving products of mass flow rates and enthalpies. • Absence of integer variables: In such cases the entire problem is continuous and so linear programming (LPR) or nonlinear programming (NLP) models are employed. • Presence of integer variables: In such cases the problem is referred to as “integer.” Integer problems are further subdivided into pure integer programming (IP) models, which involve only integer variables; mixed integer linear programming (MILP) models, which involve linear relationships with both integer and continuous variables; and mixed integer nonlinear programming (MINLP) models, which involve nonlinear relationships with both integer and continuous variables. Further details on these classifications and their properties can be found in Floudas (1995) and Williams (1999). The most important
27
28
Chapter Three factors are linearity and the existence of integer variables. There are two main reasons for the significance of these factors: 1. Continuous problems are solved by using simpler methods that incorporate gradient or other search mechanisms. The presence of integer variables introduces combinatorial complexity caused by the need to build search trees that branch through the integer variables. 2. Nonlinearity introduces a different type of complexity. Whereas linear problems (LPR and MILP) have been shown to be convex (Floudas, 1995; Williams, 1999), the convexity of nonlinear problems (NLP and MINLP) must be evaluated on a case-by-case basis. Such problems are generally assumed to be nonconvex until proven otherwise.
3.5 Conditions for Optimality One of the most popular methods for handling optimization formulations is the use of MPR solvers. They usually take standardized input in the form of matrices of variables, parameters, and equations, after which they explore the search space, using local search and gradients, to reach an optimal solution.
3.5.1 Conditions for Local Optimality When employing local search (e.g., gradient-based) algorithms, the desired extremum (minimum or maximum) needs to be located and proven. The optimality condition usually employed by solver algorithms is based on the mathematical definition of an extremum. For finding a minimum, the definition requires the existence of a point in the search space such that any small deviation from that point, in any direction within the search space, will result in an increase of the objective function value or in keeping it the same:
F x * , y * d F x , y , x , y vicinity of x * , y *
(3.1)
Further details are given in specialized textbooks on optimization (see, e.g., Edgar and Himmelblau, 1988; Floudas, 1995; Luenberger and Ye, 2008). Rigorous definitions can also be found by searching the Web for “KKT optimality conditions” (aka Karush–Kuhn–Tucker conditions, an extension of the method of Lagrange multipliers).
3.5.2 Conditions for Global Optimality Additional conditions are applied to attain global optimality when using local search algorithms, which in this case require the problem
Process Optimization to be convex. For linear problems (e.g., MILP), the locally optimal solution is guaranteed to be global—in other words, the best possible. Another advantage of the linear model is that linear solvers, unlike with nonlinear ones, do not require initialization with a feasible solution. Nonlinear problems are much more difficult to solve, and attaining feasibility of the solutions is a significant issue. Any result obtained guarantees global optimality only if the optimization problem is shown to be convex (Williams, 1999).
3.6 Deterministic Algorithms for Solving Continuous Linear Optimization Problems The best-known algorithm for LPR problems is the simplex algorithm (Dantzig, 1951; Dantzig, Orden, and Wolfe, 1954). It solves LPR maximization problems by constructing a feasible solution at a vertex of the search polyhedron and then walking along the edges of the polyhedron to other vertices with successively higher objective function values. Although this algorithm is efficient in general, in some cases it may require exponential time to find the optimum. Khachiyan (1979) proposed a method that, in the worst case, finds a solution in polynomial time for the worst case. Today, the LPR problems are usually solved using one of two methods: 1. Revisions of the simplex algorithm: This algorithm has been developed almost continuously over the years since its initial formulation, starting with the revised Simplex Method of Dantzig and Orchard-Hays (1953). A thorough consideration of the simplex algorithm and its computational techniques is given in Maros (2003b). Modern LPR solvers use algorithms that flexibly solve both continuous and integer linear problems. 2. Interior point methods: In contrast to the simplex algorithm, which finds the optimal solution by traversing points on the boundary of the feasible region, interior point methods move through the interior of the feasible region. One such method (see Mehrotra, 1992) is used by MATLAB.
3.7 Deterministic Algorithms for Solving Continuous Nonlinear Optimization Problems Nonlinear programming is used to solve continuous nonlinear problems. A common feature of all deterministic search methods is that, starting from a current solution point, they perform an iterative search for a better feasible solution within the vicinity of the current
29
30
Chapter Three one, repeating this action until no further solution improvement is possible. This strategy means that the methods find only local optima, so there is no guarantee of global optimality. Although modern optimization software packages are capable of solving constrained problems, they are mostly based on the search techniques developed for unconstrained optimization.
3.7.1 Search Algorithms for Nonlinear Unconstrained Problems Many methods have been developed for performing local search for the minimum of an objective function (Luenberger and Ye, 2008). Deterministic search algorithms mainly assume that the objective function is minimized and use the concept of descent: a series of iterations are performed that aim reducing the objective function value at each iteration. A thorough review of the deterministic search algorithms for optimization and their performance, as applied to chemical process models has been provided elsewhere, for example, Klemeš and Vašek (1973). Such methods include pattern search, the Rosenbrock method, and conjugate gradients. All descent-based algorithms rely on a common general strategy. First, an initial solution point is chosen. Then, a rule-based direction of movement and a step size are determined and executed. At the new point, a new pair of direction and step size are determined, and the process is repeated until the desired convergence is achieved. The main difference between the various search algorithms is in the rule applied for determining the iteration steps. When the search domain is a given line, the process of determining the function minimum is known as a line search. Higher-dimensional problems are solved by executing a sequence of successive line searches. The two simplest line search algorithms are the Fibonacci method (e.g., Avriel and Wilde, 1966) and the golden section method (e.g., Kiefer, 1953). Both algorithms search over a fixed interval and assume that the function is unimodal within it. The golden section algorithm has a linear convergence rate of dk+1/dk = 0.618. Another line search algorithm is the Newton method, which is based on the additional assumption of function smoothness; this technique achieves even faster convergence than the golden section one. There are also other line search algorithms that are grouped together as curve-fitting methods. Another important search algorithm—one that is applicable to multidimensional function domains—is the method of steepest descent or gradient method. Its update rule derives the new step based on the gradient information in the current point, identifying the direction of the fastest decrease of the objective function. This is one of the oldest and best-known function minimization methods. Hence it is often used as a benchmark against which the performance
Process Optimization of other methods is measured. See Luenberger and Ye (2008) for additional information.
3.7.2 Algorithms for Solving Constrained Nonlinear Problems Many algorithms for constrained problems proceed in two main steps: (1) The constrained problems are transformed into equivalent unconstrained ones, and then (2) the modified problems are subjected to unconstrained search algorithms. Groups of such algorithms include primal methods, penalty and barrier methods, and primaldual methods. The primal algorithms search for the optimal solution directly through the feasible region. For a problem with n variables and m equality constraints, primal methods work within the (n − m)dimensional feasible space. Each point in the process is feasible, and the value of the objective function constantly decreases. Thus, a feasible solution is guaranteed even if the search is interrupted before it finishes. Most primal methods do not rely on special problem structure (such as convexity), so they are applicable to general NLP problems. The weaknesses of primal algorithms are their requirement that an initial feasible solution be found or specified and their slowness (or even failure) to converge in the case of nonlinear constraints. Penalty and barrier methods utilize unconstrained approximations of the constrained optimization problem. Penalty methods add a penalty term to the objective function, which results in a high cost if constraints are violated. Barrier methods instead add a term that favors points interior to the feasible region over those near the boundary. There are two major issues with applying these methods. First, it is important to ensure the accuracy with which the unconstrained NLP problem being solved approximates the true, constrained one. Second, it is possible that the penalty or barrier term may dominate the objective function, skewing the problem to such a degree that the feasible solutions found are not actually optimal.
3.8 Deterministic Methods for Solving Discrete Problems All integer programming problems (pure IP, MILP, and MINLP) possess combinatorial features. The first solution methods used cutting planes (i.e., added constraints to force integrality), and various algorithms have been devised on this basis. The effective approach has been to divide the problem into a number of smaller problems, a method known as branch and bound. This is a strategy of problem decomposition: it aims to partition the feasible region into more manageable subdivisions and then, if required, into further subdivided partitions. The method was proposed by Land and Doig (1960) for solving LPR problems and has since evolved to a point where it can
31
32
Chapter Three be used to solve IP and MIP problems. The original large MIP problem is divided into a number of subproblems, called nodes, which form an enumeration tree. The algorithm starts from a main node and progresses toward the so-called leaves, adding nodes to the current solution or discarding them as necessary. In this process, an important role is played by the bounding function, which (in the case of objective minimization) provides a lower bound on the remaining part of the problem under the current branch. Thus, if the lower bound on the current subproblem node is higher than the current best solution, then the algorithm can safely discard (prune) the node and all its subnodes. More information on IP solving algorithms can be found in Nemhauser and Wolsey (1999).
3.9 Stochastic Search Methods for Solving Optimization Problems An alternative to deterministic algorithms are those that employ stochastic sampling of the search space and hill climbing: pushing the algorithm to lower objective function values (assuming minimization) by allowing temporary increases in the objective function value. The main advantage of stochastic search methods is that their search is not confined in the neighborhood of a given current solution, which means there is a higher probability of finding the global optimum. Such techniques have proven to be successful in applications—for example, process design and synthesis studies—where the computational requirement for single samplings is modest and the time frame for producing a solution is more relaxed. Their major limitation is that many iterations are required in order to assure some degree of optimality. This is because the algorithms blindly evaluate even infeasible combinations of variable values, especially for the simulated annealing variants. Simulated annealing (SA) is the most prominent stochastic search method (Kirkpatrick, Gelatt, and Vecchi, 1983). An interesting application is the GAPinch toolbox for MATLAB (Prakotpol and Srinophakun, 2004), which implements a genetic algorithm (GA) search to solve MINLP network synthesis problems involving water reuse. Another promising development is the use of the GA framework to optimize schedules and supply chains (Shopova and Vaklieva-Bancheva, 2006). Ant colony optimization (Zecchin et al., 2006) and Tabu search (Cunha and Ribeiro, 2004) have been applied to process design and operation. Other examples of stochastic search methods include the following: 1. McKay, Willis, and Barton (1997) used genetic programming (GP) to identify steady-state models of integrated chemical processes.
Process Optimization 2. Manolas et al. (1997) applied GA techniques to the optimal operation of industrial utility systems. 3. Castell et al. (1998) described a novel GA application to optimize process layout. 4. Genetic algorithms have been proposed for the synthesis of Heat Exchanger Networks (HENs) and other processing systems (Ravagnani et al., 2005; Xiangkun et al., 2007; Fieg, Xing, and Jeżowski, 2009); the paper by Xiangkun and colleagues presents a hybrid GA-SA method. 5. Ahmad et al., (2008) proposed that SA be used for synthesizing Heat Exchanger Networks. This approach employs a completely evolutionary strategy, starting from a trivial network topology that connects all hot streams to coolers and all cold streams to heaters.
3.10
Creating Models The procedure for building a process model is illustrated in Figure 3.1. The modeling begins with accumulating sufficient information about the process—in order to develop an understanding of the elements and the relationships between them—and proceeds to formulating a mathematical description of the process that is implemented on a computational platform. A distinctive characteristic of the procedure is its iterative nature. The mathematical modeling
BEGIN YES Conceptual modeling
YES
Mathematical modeling
Need corrections?
NO Computational implementation NO END
FIGURE 3.1
Model creation procedure.
Need corrections?
33
34
Chapter Three often leads to changes in the conceptual model; the result is an iterative feedback loop, as shown in the figure. A similar correction loop is also present at the output of the implementation block. The discussion that follows addresses only those activities in Figure 3.1 that involve conceptual and mathematical modeling.
3.10.1
Conceptual Modeling
Conceptual modeling involves collecting and organizing essential information about the phenomena in the process under consideration. This step is often referred to as data extraction (Williams, 1999). The process operating units are described along with the relevant features of their behavior. The important constraints for the units’ capacities and other limitations are identified and added to the description, as is topological information about the process network. The main purpose of this description is to serve as an interface between the process operators and the modelers. Therefore, it is important that the description be concise enough that modelers can efficiently grasp the workings of the complete system. That being said, the description must contain sufficient detail to complete the study.
Extracting Data About the Operating Units When a production process is modeled for the purpose of Process Integration, the data extraction involves four main steps: 1. Description of the process operating units and their interconnections; creation of a flowsheet that reflects this description. 2. Identification of the heating and cooling needs of the process through use of the flowsheet and related data about the operating units. 3. Definition of the Heat Integration process streams: identifying for each stream the values for its heat load, as well as the supply and target temperatures. Some process streams may need to be segmented. This is done if the specific heat capacity of a given process stream varies significantly within the interval between its supply and its target temperature. 4. Analysis of the collected data. If the goal is to minimize water use and wastewater discharge, then the various water-using operations are analyzed and their relevant properties are recorded in some standard form. The most popular way to express the water requirements (Wang and Smith, 1994; Kuo and Smith, 1997) of an operation is in terms of the limitations on (1) inlet and outlet concentrations of different contaminants and (2) flow rates of the water to be consumed. This topic is discussed in more detail in Chapter 12.
Process Optimization
Network and Topology Data Identification Obtaining network-related information is needed to account for constraints that are related to the system topology and to the limits imposed by operating units on the suitability of various process streams to serve as inputs or outputs. In water networks, such considerations include the acceptability (or unacceptability) of using the water output from some operations as inputs for other operations. For instance, the final washing of sugar crystals in sugar production would require pure water, and for this the outputs from other waterusing operations would be unacceptable. On the otherhand, used water from blanching might be perfectly acceptable for the initial washing or rinsing of fruits. This type of information is used to formulate additional constraints on the compatibility of different process streams. When supplied to automated process optimization algorithms, these constraints serve to eliminate a number of infeasible combinations of process units. When building pure MPR models (Williams, 1999), network-related information is transformed into explicit mathematical constraints involving expressions with binary selection variables. When using the graph-theoretic approach and/or the P-graph framework (Friedler et al., 1993) to construct a process model, such information is explicitly encoded in the P-graph building blocks (materials and operations) and is then used by algorithms that generate only those topologies that are combinatorially feasible.
3.10.2
Mathematical Modeling of Processes: Constructing the Equations
After the conceptual basis has been established, it is time to begin constructing the explicit mathematical formulations of the problem. The standard procedure in this regard is first to build a superstructure—one that incorporates all possible options and combinations of operating units—and then to reduce the superstructure via optimization techniques. In this context, a superstructure is the union of several feasible flowsheets (see Figure 6.2 for an example of a water reuse network superstructure). When this union includes all possible flowsheets, the superstructure is called the maximal structure (Friedler et al., 1993) or the hyperstructure (Papalexandri and Pistikopoulos, 1996). There are two basic approaches to formulating the superstructure and subjecting it to reduction optimization. 1. Explicit formulation of a superstructure by the design engineer, followed by translation of that structure into an integer programming model: The generated problem is then solved by the corresponding MPR algorithm. Popular codes for solving MILP problems are OSL (GAMS, 2009) and CPLEX (ILOG, 2009); both are included in such commercial optimization software packages as GAMS (2009). If the model does not
35
36
Chapter Three involve choices between structural options and involves decisions on flow rates and capacities only, then continuous optimizers (LPR or NLP) can be used. 2. Automated generation of the maximal superstructure, followed by enumeration of all feasible network topologies using the P-graph framework. The unit operations can be described in terms of (a) the input to the automated procedure, (b) the compatible connections between them, and (c) the corresponding process and cost information. More details are provided in Chapter 7. The construction of a process network model begins with formulation of the mass balances. The following key points should be kept in mind: • Mass balances can be of two types: overall and componentwise. Overall balances are performed over the total contents of the input and output streams of a process unit. For steadystate models of continuous processes, this content is usually expressed as a mass-based flow rate (e.g., in units of kg/s, kg/h, or t/h). Componentwise mass balances reflect the mass conservation principle between the inlet and outlet streams for individual chemical components (or pseudocomponents). • For a given operating unit, the set of all componentwise mass balances is exactly sufficient for completely characterizing the material flows into and out of the unit (adding the overall balance could lead to an “overspecified” system of equations). However, if using the overall material balance is critical to the system model, then the overall balance can be used in place of one of the componentwise balances. • In some cases, tracking all chemical species in various process streams is not necessary. This is true for water networks in which mass balances are written for the water flow rates and for the analyzed contaminants. However, an incomplete list of the material species contained in the water streams will naturally result in incomplete mass balances (Smith, 2005). When complete results are necessary, rigorous simulations of the optimized system must be performed. • The componentwise mass balances of stream mixers and the more complex operating units involve bilinear terms that reflect products of the stream flow rates and the component concentrations. In this case, the result is an NLP or MINLP problem. If either the concentrations or the flow rates are fixed then the model could be linear (an LPR or MILP problem), which would make for an easier computation that might guarantee global optimality. Although such an approach is
Process Optimization often viable, additional analyses are required to ensure that the linearization does not result in an inadequate (inaccurate) model. Similar problems also exist for Heat Exchanger Networks, where nonisothermal mixing leads to equations containing bilinear terms: products of mass flow rates and enthalpies. Process-specific constraints are also included in the optimization problem. The mass balances are supplemented in the model by lower and upper bounds on the stream flow rates and component concentrations. Another source of constraints is the temperature feasibility of interequipment connections. For example, water coming from a blanching operation may be too hot to be used for washing fresh fruits and so may require cooling (perhaps by mixing with a colder stream) before this use; alternatively, this may be rejected as an unacceptable connection, leading to a forbidden match. In MPR, such constraints usually contain integer variables. A frequent problem when synthesizing process networks is obtaining extremely low flow rates for some interconnections. When the model to optimize accounts for the complete capital costs, this problem is less likely to appear. However, if capital costs are underestimated (or disregarded) for some reason, then operating costs may dominate, resulting in “degenerate” solutions with impractically small flow rates. Practical solutions require reasonably accurate estimates of the capital cost, especially if there are fixed costs. A more straightforward option is to stipulate a lower bound on all network flow rates. Another important part of creating a process model is identifying the energy needs of various operating units. It is crucial to account for the heating and cooling needs of the process operations, and these are established by formulating enthalpy balances (Linnhoff et al., 1982). Most process streams need to be transported between the operating units and also moved through them, and such transport involves overcoming certain pressure drops (for fluids) and performing mechanical work (for solids). These operations require mechanical shaft power, which can be supplied by direct-drive machines or electrical motors—elements that define the power requirements of a process. In many cases, additional equations are needed (e.g., constitutive relations as well as calculations of reaction rates and equilibriums).
3.10.3
Choosing an Objective Function
The objective function to choose depends on the goal of the optimization. It is possible to choose from a number of criteria, including (1) maximizing profit; (2) minimizing operating cost; (3) minimizing total annualized cost (TAC); (4) minimizing consumption of certain resources or consumption per unit of product;
37
38
Chapter Three and (5) minimizing the system’s total environmental footprint or the footprint per unit of product. It is frequently necessary to optimize more than one criteria. There are three main approaches to this task (Ehrgott, 2005): 1. Choose one criterion for formulating the objective function; then add the other criteria as constraints to the problem. 2. Combine all the criteria into one objective function by summing them up, where each criterion is weighted with a given coefficient. 3. Perform a multicriteria optimization, accounting explicitly for the conflicts between the chosen objectives (criteria).
3.10.4 Handling Process Complexity Process synthesis and process design tasks—when performed on real-life, industrial-scale problems—tend to involve substantial number of operating units. Examples can be found in many areas: • Synthesizing Heat Exchanger Networks involves a large number of possible combinations of potential heat exchangers. Thermodynamic and process-related constraints usually reduce this number, but even then the complexity remains significant. • Water subsystem design is no exception, and problems with 20 or more water-using operations are common (Bagajewicz, 2000; Thevendiraraj et al., 2003). This number leads to high levels of combinatorial complexity. In a superstructure, each water-using operation (and each intermediate water main) defines at least one mixer. If the number of water-using operations is denoted by Nop, then there can be no fewer than Nop corresponding binary variables in the network superstructure, and the number of combinations of binary variable values to be examined by the corresponding MIP solver would equal 2Nop. Thus, for 20 operations there would be more than a million (106 = 1,000,000) possible combinations. When using MPR superstructure models directly, the number of binary variables is dictated by the number of the candidate operating units. In the worst case, the solution algorithm will have to examine the entire search space, which depends exponentially on the number of the binary variables. One modeling strategy that reduces the search space by several orders of magnitude is to use the Maximal Structure Generation (MSG) and Solution Structures Generation (SSG) algorithms of the P-graph framework (Friedler et al., 1993). These algorithms effectively discard all infeasible combinations of the binary selection variables and retain only the feasible ones.
Process Optimization Another popular technique in process design and software development is modularization or encapsulation. The complexity management efforts in information technology and process modeling led to development of the concepts of object-oriented modeling (see, e.g., Modelica, 2009a; UML, 2010) and object-oriented programming (e.g., C++, C#, Java, Delphi). With these modeling concepts, a number of related objects or operations can be grouped together and represented as a single object or operation. Similarly, flowsheeting and simulation software may offer the option of representing a distillation column as a single operating unit at the level of the entire flowsheet while still allowing simulation of the column at the local level; such a facility is available, for example, in gPROMS (PSE, 2009) and HYSYS (HYSYS, 2010). The key to object-oriented thinking is the principle of information hiding, according to which every object should conceal as many details as possible about its own functionality and provide to other objects only the information relevant to interactions with them. Hence, the interface describes the bundle of information items that are defined as being available to an external object. An example of applying information hiding in water network design is the abstraction of the detailed information about water-using operations into only three relevant pieces of information for each operation and contaminant: the limiting inlet and outlet concentrations and the limiting flow rate (or contaminant load). This bundle of information thus constitutes the operation’s interface to the water minimization problem. Efficiently managing the model and problem complexity is greatly facilitated by the practice of documenting the complete modeling process and optimization results, including their interpretation. It is best to document all stages, starting from the conceptual modeling and ending with the computational implementation and obtained results. All this documentation should be systematic so that the reasoning and results can be clearly understood and traced back to their roots. This style of documentation is extremely useful and makes the work of teams and individual engineers smoother and more efficient in the long run. Another important tool for managing complexity is targeting, which reveals limitations in the underlying design or operation task to which the optimization is being applied. With proper targeting it is possible to obtain an upper bound on system performance and/or a lower bound on system cost. In fact, it is also possible to calculate practically achievable targets: • For water systems, current targeting practices mainly yield the first type of estimate: the maximum possible amount of water reduction. • For HEN synthesis, the Maximum Energy Recovery (MER) targets can be established, and HENs achieving them also
39
40
Chapter Three feature reasonable costs. However, this is seldom the global minimum for the total annualized cost (i.e., the sum of annual operating costs and annualized investment costs). The logic behind evaluating a system in terms of the upper bound on its performance is this: if the best possible system performance is still insufficient to satisfy the specified requirements, then no further time and effort should be spent on designing that system. The next section gives more details on the use of targeting to managing design problem complexity.
3.10.5 Applying Process Insight Mathematical tools are absolutely necessary for optimizing the design and operation of industrial processes. However, meaningful and applicable results are obtained only when process insight is used to guide the model building, the optimizing, and the interpretation of results. Each optimization problem has its own particular features. For example: • The processing of fruits for canning imposes different water requirements and practices from those for the processing of poultry. • When designing HENs, the various underlying process operations that need cooling or heating should be properly examined for data extraction. In some cases, process knowledge may aid in the lumping of different heating/cooling needs, thereby simplifying the flowsheet. Process knowledge is also employed when partitioning properly into segments a process stream whose heat capacity flow rate varies widely. Every specific requirement discovered in the iterative process of model improvement should be thoroughly documented and implemented in the model—for instance, in the form of constraints or simplifying assumptions. A good practice is to perform targeting for the desired application: a heat recovery problem, water management, a separator network, or reactor network design. Targeting provides information on potential performance. Targeting procedures have been developed and well tested for a number of applications and domains (see Chapter 2 for a review). The benefits of using variants of Pinch Analysis or other targeting are twofold. First, the designer can estimate the best possible performance of the system by using simple models and calculations, even before using rigorous design procedures, saving valuable time. The obtained targets can be used in preliminary sensitivity studies to determine which operations and units should be included in the design and which should be left out. This approach can greatly
Process Optimization simplify network design if the number of candidate operations is large and the targets can be used as guides in the network synthesis. Usually the engineer aims either to achieve the targets exactly or to approach them closely with the final design. If the targeting model is too idealized, then the estimates produced will serve as loose performance or cost bounds, not as tight bounds. Yet in many cases this strategy results in a simple design procedure and a nearly optimal outcome. Second, if the targeting model is exact, as in Pinch Analysis for Heat Integration (Linnhoff and Flower, 1978), or if it at least captures all key factors at the corresponding design stage, as with Regional Energy Clustering (Lam, Varbanov, and Klemeš, 2010), then the targeting procedure also provides a convenient partitioning of the original design space. This makes it easier to decompose the problem and to simplify the remaining actions. A good example of partitioning the design space is the division above/below the Pinch in Heat Integration (see Chapter 4).
3.10.6 Handling Model Nonlinearity As discussed in Section 3.5, convex problems are guaranteed to produce globally optimal results when solved with deterministic algorithms that employ local search. In contrast, a nonconvex optimization problem is difficult to solve, and its solution is not guaranteed to be a global optimum. All linear MP models are convex (Williams, 1999). With nonlinear models, however, the problem convexity must be established on a case-by-case basis. Nonlinear models hinder the computation process of the solvers (e.g., by requiring that feasible initial solutions be provided), and they often result in poor numerical convergence. This is why engineers usually seek ways to obtain linear models in some form. Crucial factors in this task are preserving the model’s precision and validity.
Trading Off Precision and Linearization Sometimes it is possible to linearize relationships that are inherently nonlinear. This can be done, for instance, by replacing a single nonlinear relationship with two or more linear ones that, together, approximate the original function over the required range. This technique is known as piecewise linearization. For example, it can be applied when the available cost functions (for piping, distillation columns, or heat exchangers) are too complex. The result of this approach is a small reduction in the overall model precision and an increase in the number of integer variables in the model; thus, combinatorial complexity is increased but computational complexity (due to nonlinearity) is reduced. The principal advantage is the resulting linearity of the model, which almost always makes it easier to solve than the original one. Caution must be exercised in the process of linearization: the loss of precision must be kept as small as
41
42
Chapter Three possible so that the resulting model remains an adequate representation of the underlying process. Another pitfall is a potentially unacceptable increase in combinatorial complexity, which can result if too many linear segments are used to approximate the original relationship.
Discretization of Continuous Process Variables Another approach to avoiding nonlinearity is to define a number of fixed levels for some variables and then to substitute the original nonlinear variables in the model with linear combinations of integer variables and parameters (where the parameters are derived from the original variables). In this way, the bilinear terms in the mass balances of contaminants can be reduced to purely linear expressions. In its consequences and pitfalls, this technique is similar to piecewise linearization.
Other Techniques There are other approaches and techniques for coping with nonlinear models. Two of them are of particular interest to practical process optimization: successive MILP (SMILP), which is used for model decomposition and solving; and model reformulation. Successive MILP can be applied to many process engineering optimization models—as long as the nonlinearities are not too strong. One example is the optimization of utility systems, which consist mainly of a set of steam headers combined with steam turbines, gas turbines, boilers, and letdowns. Most of the nonlinearities in such systems are bound to the enthalpy balances of the steam headers, the steam turbines, and the letdowns. The computational difficulties imposed by the nonlinearities can be overcome by first fixing the values of some system properties during optimization (e.g., enthalpies of steam mains), thereby producing a linear optimization model, and following this with a rigorous simulation after each optimization step. The linear optimization steps are repeated, followed again by simulation, and so on until convergence is achieved; see Figure 3.2. This procedure converges rapidly when applied to the optimization of existing utility systems: usually five iterations at most are required to reach reasonably small error levels. Model reformulation refers to the symbolic transformation of the original nonlinear equations into another set of equations that are equivalent but linear. The resulting set usually contains more equations than the initial one. One such reformulation technique, known as the Glover transformation (Floudas 1995), can transform equations containing the product of a continuous and a binary variable. The essence of the technique is to replace each term that is a product of a continuous variable and a binary variable with additional continuous variables and an additional set of linear inequality constraints.
Process Optimization BEGIN Initialization (+Simulation)
MILP
Convergence?
Simulation
NO
YES END
FIGURE 3.2
3.10.7
SMILP procedure for solving nonlinear optimization problems.
Evaluating Model Adequacy and Precision
Once the model is built, the next step is validation. This process boils down to evaluating how precisely the model predicts real-life phenomena as well as how adequately it represents the modeled system (Steppan, Werner, and Yeater, 1998; Montgomery, 2005). If the model turns out to be imprecise or inadequate, then the reasons for these shortcomings must be discovered and addressed. This iterative process is similar to debugging during software development. It is generally accepted that residuals (and their plots) are sufficient for assessing whether a given model accurately predicts the underlying process. The residual plots can be used to minimize or even eliminate stochastic errors. In addition, parity plots are helpful in exposing any systematic errors in the model. The final check is to analyze the model’s variance (Steppan, Werner, Yeater, 1998; Montgomery, 2005). In essence, this means determining whether the empirically derived coefficients and the model’s predictions have any statistical significance. This is performed by means of a standard procedure for the “Analysis of Variance” (ANOVA).
43
This page intentionally left blank
CHAPTER
4
Process Integration for Improving Energy Efficiency
H
eat recovery is widely applied in industrial processes and has an extensive historical record. However, systematic methods for performing heat recovery are relatively new when compared with the age of modern industry.
4.1 Introduction to Heat Exchange and Heat Recovery In industry, large amounts of thermal energy are used to perform heating. Examples of this can be found in crude oil preheating before distillation, preheating of feed flows to chemical reactors, and heat addition to carry out endothermic chemical reactions. Similarly, some processes—such as condensation, exothermal chemical reactions, and product finalization—require that heat be extracted, which results in process cooling. There are several options for utility heating; these include steam, hot mineral oils, and direct fired heating. Steam is the most prevalent option because of its high specific heating value in the form of latent heat. Utility cooling options include water (used for moderate-temperature cooling when water is available), air (used when water is scarce or not economical to use), and refrigeration (when subambient cooling is needed). Heat recovery can be used to provide either heating or cooling to processes. Heat recovery may take various forms: transferring heat between process streams, generating steam from higher-temperature process waste heat, and preheating (e.g., air for a furnace, air or feed water for a boiler using waste heat). Heat transfer takes place in heat exchangers, which can employ either direct mixing or indirect heat transfer via a wall. Direct heat exchange is also referred to as nonisothermal mixing because the temperatures of the mixed streams are different. Mixing heat exchangers are efficient at transferring heat and usually have low
45
46
Chapter Four capital cost. In most industries, the bulk of the heat exchange must occur without mixing the heat-exchanging streams. In order to exchange only heat while keeping the streams separate, surface heat exchangers are employed. In these devices, heat is exchanged through a dividing wall. Because of its high thermal efficiency, the countercurrent stream arrangement is the most common with surface heat exchangers. To simplify the discussion, counter-current heat exchangers are assumed unless stated otherwise. In terms of construction types, the traditional shell-and-tube heat exchanger is still the most common. However, plate-type and other compact heat exchangers are gaining increased attention. Their compactness, together with significant improvements in their resistance to leaking, have made them preferable in many cases.
4.1.1 Heat Exchange Matches A hot process stream can supply heat to a cold one when paired in one or several physical heat exchangers arranged in parallel or sequence. Each such pairing is referred to as a heat exchange match. The form of the steady-state balance equations for heat exchange matches that is most convenient for Heat Integration calculations is based on modeling a match as consisting of hot and cold sides, as shown in Figure 4.1. The hot and cold part each have a simple, steadystate enthalpy balance that involves just one material stream and one heat transfer flow. The main components of the model are (1) calculations of the heat transfer flows accounted for by the enthalpy balances and (2) estimation of the necessary heat transfer area. For the latter, both the log-mean temperature difference and the overall heat transfer coefficient are employed. The enthalpy balance of the hot and cold parts, and the kinetic equation of the heat transfer, may be written as follows: QHE
mhot hin,hot hout,hot
(4.1)
Hot part hIN,HOT
hOUT,HOT mHOT
mHOT
QHE hIN,COLD
mCOLD
hOUT,COLD mCOLD
Cold part
FIGURE 4.1
Process flow diagram of a heat exchange match.
Process Integration for Improving Energy Efficiency QHE
mcold hout,cold hin,cold
(4.2)
QHE
U A ⌬TLM
(4.3)
where QHE [kW] is the heat flow across the whole heat exchanger, U [kW∙m−2∙°C−1] is the overall heat transfer coefficient, A [m2] is the heat transfer area, and ΔTLM [°C] is the logarithmic-mean temperature difference. More information can be found in Shah and Sekulić (2003), Tovazshnyansky et al. (2004), and Shilling et al. (2008).
4.1.2 Implementing Heat Exchange Matches The heat exchange matches are often viewed as being identical to heat exchangers, but this is not always the case. A given heat exchange match may be implemented by devices of different construction or by a combination of devices—for example, two heat exchangers in sequence may implement a single heat exchange match. The distinction between the concept of a heat exchange match and its implementation via heat exchangers is important because of capital cost considerations.
4.2 Basics of Process Integration 4.2.1 Process Integration and Heat Integration A historical review of the field was given in Chapter 2. Initially, attention was focused on reusing any waste heat generated on different sites. Each surface heat exchanger was described with only a few steady-state equations, and the thermal energy saved by reusing waste heat led to reductions in the expense of utility resources. This approach became popular under the names Heat Integration (HI) and the more general term Process Integration (PI). In this context, HI means integrating different processes to achieve energy savings. Engineers realized that integration could also reduce the consumption of other resources as well as the emission of pollutants. Heat and Process Integration came to be defined more widely in response to similar developments in water reuse and wastewater minimization.
4.2.2 Hierarchy of Process Design Process design has an inherent hierarchy that can be exploited for making design decisions. This hierarchy may be represented by the so-called onion diagram (Linnhoff et al., 1982), as shown in Figure 4.2. The design of an industrial process starts with the reactors or other key operating units (the onion’s core). This is supplemented and served by other parts of the process, such as the separation subsystem (the next layer) and the Heat Exchanger Network (HEN) subsystem.
47
Chapter Four
Ambient Utilities
Reactor
He a
tE
ng xcha er Ne t arato p e r S
rk wo
Feed
Separator
48
Reactor
Feed + Product
Product
Steam Turbine Boiler
FIGURE 4.2
The onion diagram.
The remaining heating and cooling duties, as well as the power demands, are handled by the utility system.
4.2.3 Performance Targets The thermodynamic bounds on heat exchange can be used to estimate the utility usage and heat exchange area for a given heat recovery problem. The resulting estimates of the process performance are a lower bound on the utility demands and a lower bound on the required heat transfer area. These bounds are known as targets for the reason that heat recovery estimates are achievable in practice and usually minimize the total cost of the HEN being designed.
4.2.4 Heat Recovery Problem Identification For efficient heat recovery in industry, the relevant data must be identified and presented systematically. In the field of Heat Integration, this process is referred to as data extraction. The heat recovery problem data are extracted in several steps: 1. Inspect the general process flowsheet, which may contain heat recovery exchangers. 2. Remove the recovery heat exchangers and replace them with equivalent “virtual” heaters and coolers. 3. Lump all consecutive heaters and coolers. 4. The resulting virtual heaters and coolers represent the net heating and cooling demands of the flowsheet streams. 5. The heating and cooling demands of the flowsheet streams are then listed in a tabular format, where each heating
Process Integration for Improving Energy Efficiency demand is referred to as a cold stream and, conversely, each cooling demand as a hot stream. This procedure is best illustrated by an example. Figure 4.3 shows a process flowsheet involving two reactors and a distillation column. The process already incorporates two recovery heat exchangers. The utility heating demand of the process is H = 1760 kW, and the utility cooling demand is C = 920 kW. The necessary thermal data has to be extracted from the initial flowsheet. Figure 4.4 shows the flowsheet after steps 1 through 4. The heating and cooling demands of the streams have been consolidated by removing the existing exchangers, and the reboiler and condenser duties have been left out of the analysis for simplicity (although these duties would be retained in an actual study). It is assumed that any process cooling duty is available to match up with any heating duty. Applying step 5 to the data in Figure 4.4 produces the data set in Table 4.1. By convention, heating duties are positive and cooling ones are negative. (The subscripts S and T denote “supply” and “target” temperatures for the process streams.) The last column of Table 4.1 gives the heat capacity flow rate (CP). For streams that do not change phase (i.e., from liquid to gas or vice versa), CP is defined as the product of the specific heat capacity and the mass flow rate of the corresponding stream: mstream Cp ,stream
CP
(4.4)
Condenser
Column
52°C 2080 kW
78°C
182°C Reactor 2
1760 kW H 138°C
120°C Reactor 1
Reboiler 100°C
120°C 3240 kW
920 kW C
FIGURE 4.3 2005).
30°C 34°C
Data extraction: Example process flowsheet (after CPI, 2004 and
49
50
Chapter Four
Condenser
Column
52°C 78°C
Cooling 2080 kW
Reactor 2 Heating 3840 kW
138°C
182°C
120°C Reactor 1
Reboiler 100°C
Cooling 4160 kW
120°C Heating 3240 kW
34°C
30°C
FIGURE 4.4 Data extraction: Heating and cooling demands (after CPI, 2004 and 2005).
Stream
Type
T S [°C]
T T [°C]
ΔH [kW]
CP [kW/°C]
H1
Hot
182
78
−2080
20
H2
Hot
138
34
−4160
40
C3
Cold
52
100
3840
80
C4
Cold
30
120
3240
36
TABLE 4.1
Data Set for Heat Recovery Analysis
The CP can also be calculated using the following simple equation: CP
⌬H TT TS
⌬H ⌬T
(4.5)
When phase transition occurs, the latent heat is used instead of CP to calculate the stream duties. Some more problems related to data extraction are discussed in Chapter 12.
4.3 Basic Pinch Technology The main strategy of Pinch-based Process Integration is to identify the performance targets before starting the core process design activity. Following this strategy yields important clues and design guidelines. The most common hot utility is steam. Heating with steam is usually approximated as a constant-temperature heating
Process Integration for Improving Energy Efficiency utility. Cooling with water is nonisothermal because the cooling effect results from sensible heat absorption into the water stream and thus leads to increasing the temperature.
4.3.1 Setting Energy Targets Heat Recovery between One Hot and One Cold Stream The Second Law of thermodynamics states that heat flows from higher-temperature to lower-temperature regions. As shown in Eq. (4.3), in a heat exchanger the required heat transfer area is proportional to the temperature difference between the streams. In HEN design, the minimum allowed temperature difference (ΔTmin) is the lower bound on any temperature difference to be encountered in any heat exchanger in the network. The value of ΔTmin is a design parameter determined by exploring the trade-offs between more heat recovery and the larger heat transfer area requirement. Any given pair of hot and cold process streams may exchange as much heat as allowed by their temperatures and the minimum temperature difference. Consider the two-stream example shown in Figure 4.5(a). The amount of heat recovery is 10 MW, which is achieved by allowing ΔTmin = 20°C. If ΔTmin = 10°C, as in Figure 4.5(b), then it is possible to “squeeze out” one more megawatt of heat recovery. To obtain the heat recovery targets for a practical HEN design problem, this principle needs to be extended to handle multiple streams.
Evaluation of Heat Recovery for Multiple Streams: The Composite Curves The analysis starts by combining all hot streams and all cold streams into two Composite Curves or CCs (Linnhoff et al., 1982). For each process there are two curves: one for the hot streams (Hot Composite Curve, HCC) and another for the cold streams (Cold Composite Curve, CCC). Each Composite Curve (CC) consists of a (a) T [°C] 200
(b) T [°C] 200
Steam
150
150
100
100 50
20°C
50
Steam
10°C
CW 0 2
CW 10 10
4
20 ΔH [MW]
Heat recovery at ΔTMIN = 20°C
FIGURE 4.5
0 1
10 11
3
20 ΔH [MW]
Heat recovery at ΔTMIN = 10°C
Thermodynamic limits on heat recovery.
51
Chapter Four temperature-enthalpy (T-H) profile, representing the overall heat availability in the process (the HCC) and the overall heat demands of the process (the CCC). The procedure of HCC construction is illustrated in Figure 4.6 on the data from Table 4.1. All temperature intervals are formed by the starting and target temperatures of the hot process streams. Within each temperature interval, a composite segment is formed consisting of (1) a temperature difference equal to that of the interval and (2) a total cooling requirement equal to the sum of the cooling requirements of all streams within the interval. This is achieved by summing up the heat capacity flow rates of the streams crossing the interval. Next, the composite segments from all temperature intervals are combined to form the HCC. Construction of the CCC is entirely analogous. The Composite Curves are combined in the same graph in order to identify the maximum overlap, which represents the maximum amount of heat that could be recovered. The HCC and CCC for the example from Table 4.1 are shown together in Figure 4.7. Both CCs can be moved horizontally (i.e., along the ΔH axis), but usually the HCC position is fixed and the CCC is shifted. This is equivalent to varying the amount of heat recovery and (simultaneously) the amount of required utility heating and cooling. Where the curves overlap, heat can be recuperated between the hot and cold streams. More overlap means more heat recovery and smaller utility requirements, and vice versa. As the overlap increases, the temperature differences between the overlapping curve segments decrease. Finally, at a certain overlap, the curves reach the minimum allowed temperature difference, ΔTmin. Beyond this point, no further overlap is possible. The closest approach between the curves is termed the Pinch point (or simply the Pinch); it is also known as the heat recovery Pinch. It is important to note that the amount of largest overlap (and thus the maximum heat recovery) would be different if the minimum
(a) T [°C]
(b) T [°C] /°C
CP2 = 40 kW/°C
20
138
=
78
CP
2
=4
78 34
34 2080
4160 6240
ΔH [kW]
The hot streams plotted separately
FIGURE 4.6
= P2 +C C CP 1 kW/° 60
/°C
W 0k
1
138
182
kW
182
CP
52
CP1 = 20 kW/°C
3600 880 ΔH [kW] 6240 The composite hot stream
1760
Constructing the Hot Composite Curve.
Process Integration for Improving Energy Efficiency allowed temperature difference is changed for the same set of hot and cold streams. The larger is the value of ΔTmin, the smaller is the possible maximum heat recovery. Specifying the minimum utility heating, the minimum utility cooling, or the minimum temperature difference fixes the relative position of the two Composite Curves and hence the maximum possible amount of heat recovery. The identified heat recovery targets are not absolute—they are relative to the specified value of ΔTmin. If that value is increased, then the minimum utility requirements also increase and the potential for maximum recovery drops; see Figure 4.8. T [°C] 200
150
Pinch
100
50 ΔTmin = 10°C 0 Qc,min = 328
FIGURE 4.7
Qrec = 5912
Qh,min = 1168
ΔH [kW]
The HCC and CCC at ΔTmin = 10°C.
T [°C] 200
150 Pinch 100
ΔTmin = 10°C
50 ΔTmin = 20°C 0 Qc,min = 328
Qrec = 5912
Qh,min = 1168
Qc,min = 728
Qrec = 5512
Qh,min = 1568
FIGURE 4.8
Variation of heat recovery targets with ΔTmin.
ΔH [kW]
53
54
Chapter Four The appropriate value for ΔTmin is determined by economic tradeoffs. Increasing ΔTmin results in larger minimum utility demands and increased energy costs; choosing a higher value reflects the need to reduce heat transfer area and its corresponding investment cost. Conversely, if ΔTmin is reduced then utility costs go down but investment costs go up. This trade-off is illustrated in Figure 4.9.
4.3.2 The Heat Recovery Pinch The heat recovery Pinch has important implications for the HEN being designed. As illustrated in Figure 4.10, the Pinch sets the absolute limits for heat recovery within the process. COST
Total
Energy
Capital 1
2 ΔTMIN
Optimum
FIGURE 4.9 ΔTmin.
Trade-off between investment and energy costs as a function of
T
QH,min
PINCH
Qc,min ΔH
FIGURE 4.10
Limits for process heat recovery set by the Pinch.
Process Integration for Improving Energy Efficiency The Pinch point divides the heat recovery problem into a net heat sink above the Pinch point and a net heat source below it (Figure 4.11). At the Pinch point, the temperature difference between the hot and cold streams is exactly equal to ΔTmin, which means that at this point the streams are not allowed to exchange heat. As a result, the heat sink above the Pinch is in balance with the minimum hot utility (QH,min) and the heat source below the Pinch is in balance with the minimum cold utility (QC,min), while no heat is transferred across the Pinch via utilities or via process-to-process heat transfer. No heat can be transferred from below to above the Pinch, because this is thermodynamically infeasible. However, it is feasible to transfer heat from hot streams above the Pinch to cold streams below the Pinch. All cold streams—even those below the Pinch— could be heated by a hot utility; likewise, the hot streams (even above the Pinch) could be cooled by a cold utility. Although these arrangements are thermodynamically feasible, applying them would cause utility use to exceed the minimum, as identified by the Pinch Analysis. This is a fundamental relationship in the design of heat recovery systems. What happens if heat is transferred across the Pinch? Recall that it is possible to transfer heat only from above to below the Pinch. If, say, XP units of heat are transferred across the Pinch (Figure 4.12), then QH,min and QC,min will each increase by the same amount in order to maintain the heat balances of the two problem parts. Any extra heat that is added to the system by the hot utility must then be taken away by the cold utility, in addition to the minimum requirement QC,min. Cross-Pinch process-to-process heat transfer is not the only way by which a problem’s thermodynamic Pinch partitioning can be
T Zero crosspinch transfer
ΔTmin
QH,min
Pinch
QC,min ΔH
FIGURE 4.11
Partitioning the heat recovery problem.
55
56
Chapter Four T
QH,min QUC,above
QUH,below
XP
XP
PINCH
XP
FIGURE 4.12
QUH,below
QC,min
QUC,above
ΔH
More in, more out.
violated. This could also happen if the external utilities are placed incorrectly. For example, any utility heating below the Pinch will create a need for additional utility cooling in that part of the system (Figure 4.12). Conversely, any utility cooling above the Pinch will create a need for additional utility heating. The implications of the Pinch for heat recovery problems can be distilled into the following three conditions, which must hold if the minimum energy targets for a process are to be achieved. 1. Heat must not be transferred across the Pinch. 2. There must be no external cooling above the Pinch. 3. There must be no external heating below the Pinch. Violating any of these rules will lead to an increase in energy utility demands. The rules are applied explicitly in the context of HEN synthesis by the Pinch Design Method (Linnhoff and Hindmarsh, 1983) and also before a HEN retrofit analysis to identify causes of excessive utility demands by a process. Other HEN synthesis methods—if they achieve the minimum utility demands—also conform to the Pinch rules (though sometimes only implicitly).
4.3.3 Numerical Targeting: The Problem Table Algorithm The Composite Curves are a useful tool for visualizing heat recovery targets. However, they can be time-consuming to draw for problems that involve many process streams. In addition, targeting that relies solely on such graphical techniques cannot be very precise. The process of identifying numerical targets is therefore usually based on an algorithm known as the Problem Table Algorithm (PTA). Some
Process Integration for Improving Energy Efficiency MPR-oriented authors employ the equivalent “transshipment” model (Cerdá et al., 1990). The steps are as follows: 1. Shift the process stream temperatures. 2. Set up temperature intervals. 3. Calculate interval heat balances. 4. Assuming zero hot utility, cascade the balances as heat flows. 5. Ensure positive heat flows by increasing the hot utility as needed. The algorithm will be illustrated using the sample data in Table 4.2.
Step 1 Because the PTA uses temperature intervals, it is necessary to set up a unified temperature scale for the calculations. If the real stream temperatures are used, then some of the heat content would be left out of the recovery. The problem is avoided by obtaining shifted stream temperatures (T*) for PTA calculations. The hot streams are shifted to be colder by ΔTmin/2 and the cold streams are shifted to be hotter by ΔTmin/2. If the shifted temperatures (T*) of a cold and a hot stream (or their parts) are the same, then their real temperatures are still actually ΔTmin apart, which allows for feasible heat transfer. This operation is equivalent to shifting the Composite Curves toward each other vertically, as illustrated in Figure 4.13. The last two columns in Table 4.2 show the shifted process stream temperatures.
Step 2 Temperature intervals are formed by listing all shifted process stream temperatures in descending order (any duplicate values are Shifter Composite Curves T* T 2
T
Cold
−
ΔTmin
Hot
Temperature
T*
+
ΔTmin 2
ΔH
FIGURE 4.13
Temperature shifting to ensure feasible heat transfer.
57
58
Chapter Four No.
Type
T S [°C]
T T [°C]
1
Cold
2
Hot
250
40
15
245
35
3
Cold
140
230
30
145
235
4
Hot
200
80
25
195
75
20
CP [kW/°C]
180
20
T S*[°C] 25
T T*[°C] 185
TABLE 4.2 Problem Table Algorithm Example: Process Streams Data (ΔTmin = 10°C) Interval Temp. [oC]
Stream Population
∑ CPH −∑ CPC [kW/oC]
∆Hinterval [kW]
10
15
150
Surplus
40
−15
600
Deficit
10
10
−100
Surplus
40
−10
−400
Deficit
70
20
1400
Surplus
40
−5
−200
Deficit
10
20
−200
Deficit
∆Tinterval [oC]
Surplus/ Deficit
2
245 235
4
145 CP = 20
3
75
35 25
CP = 25
CP = 15
185
CP = 30
195
1
TABLE 4.3
Problem Table Algorithm for the Streams in Table 4.2
considered just once). This action creates temperature boundaries (TBs), which form the temperature intervals for the problem. For the example in Table 4.2, the TBs are 245°C, 235°C, 195°C, 185°C, 145°C, 75°C, 35°C, and 25°C.
Step 3 The heat balance is calculated for each temperature interval. First, the stream population of the process segments falling within each temperature interval (the first two columns of Table 4.3) is identified.
Process Integration for Improving Energy Efficiency Next, the sums of the segment CPs (heat capacity flow rates) in each interval are calculated; then that sum is multiplied by the interval temperature difference (i.e., the difference between the TBs that define each interval). This calculation is also illustrated in Table 4.3.
Step 4 The Problem Heat Cascade shown in Figure 4.14 has a box allocated to each temperature interval; each box contains the corresponding interval enthalpy balances. The boxes are connected with heat flow arrows in order of descending temperature. The top heat flow represents the total hot utility provided to the cascade, and the bottom heat flow represents the total cold utility. The hot utility flow is initially assumed to be zero and this value is combined (summed) with the enthalpy balance of the top cascade interval to produce the value for the next lower cascade heat flow. This operation is repeated for the lower temperature intervals and connecting heat flows until the bottom heat flow is calculated, resulting in the cascade shown in Figure 4.14(a).
Step 5 The resulting heat flow values in the cascade are examined, and a feasible heat cascade is obtained; see Figure 4.14(b). From the cascading heat flows, the smallest value is identified; if it is nonnegative (i.e., positive or zero), then the heat cascade is thermodynamically feasible. If a negative value is obtained then a positive utility flow of the same absolute value has to be provided at
HOT UTILITY (b)
(a) 245 °C
0 kW
HOT UTILITY
245 °C ΔH = 150 kW
ΔH = 150 kW
235 °C
150 kW ΔH = −600 kW
235 °C 195 °C
−350 kW −750 kW −650 kW −450 kW
250 kW HOT UTILITY (a) Initial cascade
FIGURE 4.14
* TPINCH 0 kW
75 °C
THOT PINCH = 150°C TCOLD PINCH = 140°C
1400 kW
35 °C
1200 kW
ΔH = −200 kW
ΔH = −200 kW
25 °C
145 °C
ΔH = −200 kW
ΔH = −200 kW
35 °C
400 kW
ΔH = 1400 kW
ΔH = 1400 kW
75 °C
185 °C ΔH = −400 kW
ΔH = −400 kW
145 °C
300 kW ΔH = 100 kW
ΔH = 100 kW
185 °C
900 kW
ΔH = −600 kW
−450 kW
195 °C
750 kW
25 °C
1000 kW HOT UTILITY
(b) Feasible cascade
Heat Cascade for the process data in Table 4.2.
59
60
Chapter Four the topmost heat flow, after which the cascading described in Step 4 is repeated. The resulting heat cascade is guaranteed to be feasible and provides numerical heat recovery targets for the problem. The topmost heat flow represents the minimum hot utility, the bottommost heat flow represents the minimum cold utility, and the TB with zero heat flow represents the location of the (heat recovery) Pinch. It is often possible to obtain more than one zero-flow temperature boundary, each representing a separate Pinch point.
4.3.4
Threshold Problems
Threshold problems feature only one utility type—either hot or cold. They are important mostly because they often result in no utility– capital trade-off below a certain value of ΔTmin, since the minimum utility demand (hot or cold) becomes invariant; see Figure 4.15. (a) T
(b) T
Steam
ΔTmin = 14°C
ΔTmin = 20°C
CW
CW ΔH
More heat recovery, no hot utility
Heat recovery, hot and cold utilities
ΔH
(d) T
(c) T
ΔTmin = 10°C
ΔTmin = 10°C
CW
CW ΔH
No increase in heat recovery
FIGURE 4.15
Steam Generation
CW
Utility substitution
ΔH
Threshold HEN design cases.
Typical examples of threshold Heat Integration problems involve high-temperature fuel cells, which usually have large net cooling demands but no net heating demands (Varbanov et al., 2006; Varbanov and Klemeš, 2008). An essential feature that distinguishes threshold problems is that, as ΔTmin is varied, demands for only one
Process Integration for Improving Energy Efficiency utility type (hot or cold) are identified over the variation range; in contrast, pinched problems require both hot and cold utilities over this range. When synthesizing HENs for threshold problems, one can distinguish between two subtypes (see Figure 4.16): 1. Low-threshold ΔTmin: Problems of this type should be treated exactly as Pinch-type problems. 2. High-threshold ΔTmin: For these problems, it is first necessary to satisfy the required temperature for the no-utility end before proceeding with the remaining design (using the tickoff heuristic; see Figure 4.53).
4.3.5 Multiple Utilities Targeting Utility Placement: Grand Composite Curve In many cases, more than one hot and one cold utility are available for providing the external heating and cooling requirements after energy recovery and it is necessary to find and evaluate the (a)
T
ST
Utilities [MW] CW ST
10°C
ΔTmin = 10°C
ΔTmin
CW ΔH [MW]
Low Threshold ΔTmin T
(b)
ST
Utilities [MW]
ST CW
10°C
ΔTmin High Threshold ΔTmin
FIGURE 4.16
Threshold problems.
ΔTmin = 10°C ΔH [MW]
61
62
Chapter Four cheapest and most effective combination of the available utilities (Figure 4.17). To assist with this choice and to enhance the information derived from the HCC and CCC, another graphical construction has been developed, known as the Grand Composite Curve (GCC) (Townsend and Linnhoff, 1983). The heat cascade and the PTA (Linnhoff and Flower, 1978) offer guidelines for the optimum placement of hot and cold utilities, and this allows one to determine the heat loads associated with each utility. For the previous sections, the assumption has been that only one cold and one hot utility are available—albeit with sufficiently low and high temperatures to satisfy the cooling and/or heating demands of the process. However, most industrial sites feature multiple heating and cooling utilities at several different temperature levels (e.g., steam levels, refrigeration levels, hot oil circuit, furnace flue gas). Each utility has a different unit cost. Usually the higher-temperature hot utilities and the lower-temperature cold utilities cost more than the ones with temperatures closer to the ambient. This fact underscores the need to choose a mix that results in the lowest utility cost. The general objective is to maximize the use of cheaper utilities and to minimize the use of more expensive utilities. For example, it is usually preferable to use low-pressure (LP) instead of high-pressure (HP) steam and to use cooling water (CW) instead of refrigeration. The Composite Curves plot in Figure 4.7 provides a convenient view for evaluating the process driving force and the general heat recovery targets. However, the CCs are not useful for identifying targets when multiple utility levels are available; the GCC is used for this task.
Boiler House and Power Plant Fuel
Fuel
HP
Steam Levels
Power MP
Steam turbine
LP BFW preheat
Heating Heating
Power
Gas turbine
Heating
Heating
Heating (Q+W)
Processes, building complexes
Heat pump
Heating Cooling
Refrigeration
Furnace Fuel
FIGURE 4.17 2005).
Cooling
Air preheat
Heating
Cooling (Q)
Power (W)
Power
Choices of hot and cold utilities (amended after CPI 2004 and
Process Integration for Improving Energy Efficiency
Construction of the Grand Composite Curve The GCC is constructed using the Problem Heat Cascade (Figure 4.14). The heat flows are plotted in the T-ΔH space, where the heat flow at each temperature boundary corresponds to the X coordinate and the temperature to the Y coordinate (Figure 4.18). The GCC can be directly related to the Shifted Composite Curves (SCCs), which are the result of shifting the CCs toward each other by ΔTmin/2 so that the curves touch each other at the Pinch; see Figure 4.19. At each temperature boundary, the heat flow in the Problem Heat Cascade and GCC corresponds to the horizontal distance between the SCCs. The GCC has several fundamental properties that facilitate an understanding of the underlying heat recovery problem. The parts with positive slope (i.e., running uphill from left to right) indicate that cold streams dominate (Figures 4.18 and 4.19). Similarly, the parts with negative slope indicate excess hot streams. The shaded areas in the GCC plot, which signify opportunities for process-toprocess heat recovery, are referred to as heat recovery pockets.
Utility Placement Options The GCC shows the hot and cold utility requirements of the process in terms of both enthalpy and temperature. This allows one to distinguish between utilities at different temperature levels. There are typically
T* (°C)
Hot Utility
ΔH= 150 kW
750 kW 245 °C 900 kW 235 °C
ΔH= 600 kW
ΔH= 100 kW ΔH= 400 kW
300 kW 195 °C 400 kW 185 °C 0 kW
PINCH
145 °C ΔH= 1400 kW 1400 °C 75 °C ΔH= −200 kW 1200 kW ΔH= −200 kW
35 °C 1000 kW 25 °C
Cold Utility 500
1000
1500 Q (kW)
FIGURE 4.18
Constructing the GCC for the streams in Table 4.2.
63
64
Chapter Four T* [°C]
T* [°C]
300
300 750
750 200
900
300
900 200
300 400
400
100
100
1400 1200
1400 1200
1000
1000
0
500
1000
1500
0
2000
4000
6000
8000
ΔH [kW]
Q [kW]
FIGURE 4.19 Relation between the GCC (left) and the SCC (right) for the streams in Table 4.2.
(a) T* [°C]
(b) T* [°C] HP steam
HP steam 200 MP steam
200 PINCH
PINCH
100
100
Cooling Water
Cooling Water 0
500
1000
Single steam level
FIGURE 4.20
Q [kW]
Q [kW] 1500
0
500
1000
1500
Multiple steam levels
Using the GCC to target for single and multiple steam levels.
utilities at several different temperature levels available on a site—for example, there may be a supply of both high-pressure (HP) and medium-pressure (MP) steam. As indicated previously, it is desirable to maximize the use of cheaper utilities and to minimize the use of more expensive ones. Utilities at higher temperature and/or pressure are usually more expensive; see Figure 4.20. Therefore, MP steam is used first: ranging from the Y axis until it touches the GCC (resulting in ΔTmin in this point), which maximizes its usage. Only then is HP steam used. When a utility line or profile touches the GCC, a new Pinch point is created, termed a Utility Pinch (the MP steam line touching the GCC in Figure 4.20). Each additional steam level creates another Utility Pinch and increases the complexity of the utility system.
Process Integration for Improving Energy Efficiency Higher complexity has several negative consequences, including increased capital costs, greater potential for leaks, reduced safety, and more maintenance expenses. Therefore limits are typically placed on the number of steam levels. Higher-temperature heating demands are satisfied by nonisothermal utilities. These include hot oil and hot flue gas, both of which maintain their physical phase (liquid and gaseous) across a wide range of temperatures. The operating costs associated with such utilities are largely dependent on furnace efficiency and on the intensity and efficiency of the pumping or fan blowing. When targeting the placement of a nonisothermal hot utility, its profile is represented by a straight line,1 which runs from the upper right to the lower left in the graph of Figure 4.21. The line’s starting point corresponds to the utility supply temperature and also to the rightmost point for the utility’s heating duty. The utility use endpoint corresponds either to the zero of the ΔH axis—in which case all utility heating is covered by the current nonisothermal utility—or to the rightmost point on the ΔH axis for other, cheaper hot utilities.
T*
FIGURE 4.21 Properties of nonisothermal hot utilities.
* Tsupply
CP1 CP2
CP1 > CP2
T*ambient ΔHloss,2 ΔHloss,1
ΔH
1 This linear representation assumes an approximately constant specific heat capacity of the corresponding stream.
65
Chapter Four As plotted in the figure, the nonisothermal utility’s termination point corresponds to the ambient temperature. The distance from this point to the zero of the ΔH axis represents the thermal losses from using the utility. The heat capacity flow rate of the nonisothermal utility target is determined by making the utility line as steep as possible, thereby minimizing its CP and the corresponding losses (Figure 4.21). Its supply temperature is usually fixed at the maximum allowed by the furnace and the heat carrier composition; the remaining degree of freedom corresponds to the utility’s exact CP. Smaller CP values result in steeper slopes and smaller losses. The placement for a nonisothermal utility (e.g., hot oil) may be constrained by two problem features: the process Pinch point and a “kink” in the GCC at the top end of a heat recovery pocket; see Figure 4.22 for an example. When fuel is burned in a furnace or a boiler, the resulting flue gas becomes available to heat up the corresponding cold-stream medium (for steam generation or direct process duty). Transferring heat to the process causes the flue gas temperature to drop as it moves from the furnace to the stack. The stack temperature has to be above a specified value: the minimum allowed stack temperature, which is determined by limitations due to corrosion. If flue gas is used directly for heating, then the Pinch point, if it is higher, may become more limiting than the minimum allowed stack temperature. If the analyzed process features both high-temperature and moderatetemperature utility heating demands, then flue gas heating may not be appropriate for satisfying all those demands. If steam is cheaper,
ΔTmin= 20
(a) T*
(b) * T 290
290
Minimizing the CP
Tsupply = 300 Minimizing the CP
in
CP
mi
n
The Pinch does not need to be limiting
TReturn, min =130
CP m
66
TReturn, min= 150 Pinch
120
120
Pinch ΔH
ΔH Process Pinch limitation
FIGURE 4.22
Heat recovery pocket limitation
Constraints for placing hot oil utilities, ΔTmin = 20 °C.
Process Integration for Improving Energy Efficiency then combining it with flue gas reduces the latter’s CP and the corresponding stack heat losses. Another option for utility placement is to use part of the cooling demand of a process for generating steam. This is illustrated in Figure 4.23, in which steam generation is placed below the Pinch. The GCC can reveal where utility substitution may improve energy efficiency; see Figure 4.24. The main idea is to exploit heat recovery pockets that span two or more utility temperature levels. The technical feasibility of this approach is determined by both the temperature span and the heat duty within the pocket, which should be large enough to make utility substitution worthwhile when weighed against the required capital costs. Utility cooling below ambient temperatures may be required, a need that is usually met by refrigeration. Refrigerants absorb heat by evaporation, and pure refrigerants evaporate at a constant temperature. Therefore, refrigerants are represented—on the plot of T (or T*) versus ΔH—by horizontal bars, similarly to the steam levels. On the GCC, refrigeration levels are placed similarly to steam levels; see Figure 4.25. When the level of a placed utility is between the temperatures of a heat recovery pocket, the Utility Pinch cannot be located by using
Pinch
T*
160
140 Point of closest approach. Not necessarily at the boiling point
120
100 Preheat Evaporation
Superheat 80
60
40
Cooling Water QC,min
20 0
FIGURE 4.23
Generating steam below the Pinch.
ΔH
67
68
Chapter Four T*
T*
HP Steam Generation
HP Steam Generation
LP Steam use
CW
CW ΔH
FIGURE 4.24
ΔH
Exploiting a GCC pocket for utility substitution.
T*
TCW
−5°C −40°C −70°C ΔH
FIGURE 4.25
Placing refrigeration levels for pure refrigerants.
the GCC. In this case, the Balanced Composite Curves (BCCs) are used. Figure 4.26 shows how the data about the placed utilities can be transferred from the GCC to the BCCs, enabling the correct location of the Utility Pinch associated with LP steam. The BCCs create a combined view in which all heat sources and sinks (including utilities) are in balance and all Pinches are shown.
Process Integration for Improving Energy Efficiency (a) T*
(b) T HP steam
LP steam
HP steam
LP steam
LP steam Pinch
Process Pinch
Process Pinch
LP steam Pinch
CW
CW ΔH Grand Composite Curve
FIGURE 4.26
ΔH Balanced Composite Curve
Locating the LP-steam Utility Pinch.
BCCs are a useful additional tool for evaluating heat recovery, obtaining targets for specific utilities, and planning HEN design regions.
4.4 Extended Pinch Technology 4.4.1 Heat Transfer Area, Capital Cost, and Total Cost Targeting In addition to maximizing heat recovery, it is also possible to estimate the required capital cost. The expressions for obtaining these estimates are derived from the relationship between heat transfer area and the efficiency of a heat exchanger. Methods for targeting capital cost and total cost were initially developed by Townsend and Linnhoff (1984) and further elaborated by others (e.g., Ahmad, Linnhoff, and Smith 1990; Colberg and Morari, 1990; Linnhoff and Ahmad, 1990; Zhu et al., 1995). The HEN capital cost depends on the heat transfer area, the number of the heat exchangers, the number of shell-and-tube passes in each heat exchanger, construction materials, equipment type, and operating pressures. The heat transfer area is the most significant factor, and assuming one-pass shell and tube exchangers it is possible to estimate the overall minimum required heat transfer area; this value helps establish the lower bound on the network’s capital cost. Estimating the minimum heat transfer area is based on the concept of an enthalpy interval. As shown in Figure 4.27, an enthalpy interval is a slice constrained by two vertical lines with fixed values on the ΔH axis. This interval is characterized by its ΔH
69
70
Chapter Four FIGURE 4.27
Enthalpy intervals.
T
Enthalpy interval Amin
=
qstream 1 · ΔTLM streams hstream H
difference, the corresponding temperatures of the CCs at the interval boundaries, its process stream population, and the film heat transfer coefficients of those streams. The minimum heat transfer area target can be obtained by estimating it within each enthalpy interval of the BCCs and then summing up the values over all intervals (Linnhoff and Ahmad, 1990): EI NS ª 1 qs , i º AHEN, min « » « ⌬TLM, i s 1 hs , i ¼» i 1 ¬
¦
¦
Here EI and NS denote the number of enthalpy intervals and the number of streams; i denotes ith enthalpy interval; s, the sth stream; ΔTLM,i, the log-mean temperature difference in interval i (from the CC segments); qs, the enthalpy change of the sth stream; and hs, the heat transfer coefficient of sth stream. The area targets can be supplemented by targets for number of shells (Ahmad and Smith, 1989) and for the number of heat exchanger units, thus providing a basis for estimating the HEN capital cost and the total cost. This approach is known as supertargeting (Ahmad, Linnhoff, and Smith, 1989). With supertargeting it is also possible to optimize the value of ΔTmin prior to designing the HEN. Proposed improvements to the capital cost targeting procedure of Townsend and Linnhoff (1984) mainly involve: (1) obtaining more accurate surface area targets for HENs with nonuniform heat transfer coefficients (Colberg and Morari, 1990; Jegede and Polley, 1992; Zhu et al., 1995; SernaGonzález, Jiménez-Gutiérrez, and Ponce-Ortega, 2007); (2) accounting for construction materials, pressure rating, and different heat exchanger types (Hall, Ahmad, and Smith, 1990); or (3) accounting for safety factors, including “prohibitive distance” (Santos and Zemp, 2000). Further information can be found in the paper by Taal and colleagues (2003), which summarizes the common methods used to estimate the cost of heat exchange equipment and also provides sources of projections for energy prices.
Process Integration for Improving Energy Efficiency
4.4.2 Heat Integration of Energy-Intensive Processes Heat Engines Particularly important processes are the heat engines—steam and gas turbines. They operate by drawing heat from a higher-temperature source, converting part of it to mechanical power; then (after some energy loss) they reject the remaining heat at a lower temperature (see Figure 4.28). For targeting purposes the energy losses are usually neglected. Integrating a heat engine across the Pinch, which is equivalent to a Cross-Pinch process-to-process heat transfer, results in a simultaneous increase of hot and cold utility, which usually leads to excessive capital investment for the utility exchangers. If a heat engine is integrated across the Pinch, then the hot utility requirement is increased by Q and the cold by Q − W (in the notation of Figure 4.28). Heat engines should be integrated in one of two ways. 1. Above the Pinch (Figure 4.29a): This increases the hot utility for the main process by W, but this extra heat is converted into shaftwork. 2. Below the Pinch (Figure 4.29b): This offers a double benefit. It saves on a cold utility, and the process heat below the Pinch supplies Q to the heat engine (instead of rejecting it to a cooling utility). Different heat engines differ in their placement. On the one hand, steam turbines may be placed either below or above the Pinch because they draw and exhaust steam. Figure 4.30 shows steam turbine integration above the Pinch, which has the benefit of cogenerating
T*
T*source
Source Q
W
Q–W T*sink Sink
FIGURE 4.28
Heat engine configuration.
71
72
Chapter Four A+W
T*
T* A–(Q –W)
A
Q HE
Q –W
PINCH
W
PINCH
HE
B–Q
Integration above the pinch (a)
FIGURE 4.29
W
Q–W
100% conversion Heat→Work
B
100% conversion Heat→Work Q
Integration below the pinch (b)
Appropriate placement of heat engines.
HP Steam
T*
W
QHP QLP
QHP
T* QFUEL
LP Steam
Boiler
QLP
Condensate PINCH
CW ΔH
QC,min
FIGURE 4.30
Integrating a steam turbine above the Pinch.
extra power. In contrast, gas turbines—which use fuel as input—are typically used only as a utility heat source for the processes and can be placed only above the Pinch.
Heat Pumps Heat pumps present another opportunity for improving the energy performance of an industrial process. Their operation is the reverse of heat engines. That is, heat pumps take heat from a lowertemperature source, upgrade it by applying mechanical power, and then deliver the combined flow to a higher-temperature heat sink (Figure 4.31). An important characteristic of heat pumps is their coefficient of performance (COP). This metric for device efficiency is defined as the
Process Integration for Improving Energy Efficiency FIGURE 4.31 Heat-pump configuration.
T*
Sink Q + W = 125 kW
T*sink
W = 25 kW Q = 100 kW T*source
Source
ratio between the heat delivered to the heat sink and the consumed shaftwork (mechanical power): Q sink
Q source W
(4.6)
COP
Q sink Q source W = W W
(4.7)
The COP is a nonlinear function of the temperature difference between the heat sink and the heat source (Laue, 2006); this difference is also referred to as temperature lift. Figure 4.32(a) shows the appropriate integration of a heat pump across the Pinch, with the heat source located below the Pinch and the heat sink above it. The GCC facilitates sizing of the heat pump by evaluating the possible temperatures of the heat source and heat sink, and their loads; see Figure 4.32(b). Integrating entirely above the Pinch results in direct conversion of mechanical power to heat. This is a waste of resources because most of the power is generated at the expense of two to three times the amount of fuel energy. The second alternative—placing the heat pump entirely below the Pinch—results in the power flow consumed by the heat pump being added to the cooling utility demand below the Pinch. The procedure for sizing heat pumps to be placed across a (process or utility) Pinch is illustrated in Figure 4.33. First, temperatures are chosen for the heat source and the heat sink. Then the horizontal projections spanning from the temperature axis to the GCC provide the maximum values for the heat source and sink loads. Recall that the GCC shows shifted temperatures. Because real temperatures are used when calculating the heat-pump temperature lift, the GCC values must be modified by subtracting or adding ΔTmin/2 (see Section 4.3.3). The COP value can be derived from the calculated ΔTpump and can be then used to calculate the necessary duties.
73
74
Chapter Four QHmin – (QHP + W)
(a) T*
Steam
(b) T*
QHP + W
QHP +W
PINCH QHP
HEAT PUMP
W
W QHP CW
Qcmin – QHP Appropriate placement
FIGURE 4.32
ΔH Load and temperature lift on the GCC
Heat pump placement against a heat recovery problem.
T*
QHP + W
5
COP
W 6
MAX
QHP + W W QHP
2 3 T*SINK ½ΔTmin ΔTPUMP
T*SOURCE 1 MAX
4
4
3 ΔH
FIGURE 4.33
ΔTPUMP
Procedure for heat-pump sizing.
As a specific example, assume that the GCC in Figure 4.34 reflects an industrial process with ΔTmin = 20°C and that a heat pump is available, described as follows: 0.874 COP 100.18 ⌬Tpump
(4.8)
Focusing on the Pinch “nose” (a sharp nose provides a better integration option) allows choosing a shifted temperature for the heat source, in this example: T*source = 85˚C; see Figure 4.35. Using this value allows extracting a maximum of 6.9 MW from the GCC below the Pinch. Setting T *sink = 100˚C results in an upper bound of 2.634 MW for the sink load. Transforming to real temperatures and taking the difference yields a temperature lift of ΔTpump = 35°C. By Eq. (4.8), the COP is thus equal to 4.4799. Given the upper bounds on the sink
Process Integration for Improving Energy Efficiency ΔH[MW]
T*[°C]
21.90
440
29.40
410
23.82
131
18.00
118
ΔTMIN = 20°C 500 450 400 350 300 T*[°C] 250
1.80
115
0.00
94
150
4.30
91
100
11.50
79
15.00
30
200
50 0 0
10
20
30
40
20
25
ΔH [MW]
FIGURE 4.34
Heat-pump sizing example: Initial data of the GCC.
FIGURE 4.35 Heat-pump sizing example: Attempt 1.
140 120 QSINK T* [°C]
100
Maximum
80 QSOURCE
60 40 20 0
5
10
15
ΔH [MW]
and source heat loads, the smaller one is chosen as a basis. Here the sink bound is smaller, so the sink is sized to its upper bound: Qsink = Qsink,max = 2.634 MW. From this, the required pump power consumption is computed to be 0.588 MW. As a result, the actual source load for the heat pump is 2.046 MW. Comparing this value with the upper bound of 6.9 MW, it is evident that the source heat availability is considerably underutilized. A different selection of source and sink temperatures is needed if the source availability is to be better utilized. As a second attempt, the sink temperature is increased from 100°C to 110°C. The maximum source heat remains 6.9 MW, but the maximum sink capacity increases from 2.634 to 7.024 MW. This results in the desired
75
Chapter Four temperature lift of ΔTpump = 45°C; the COP is now 3.5964, W = 2.657 MW, and the sink load Qsink = 9.557 MW; see Figure 4.36. However, the heat sink is oversized in this case (Figure 4.36), so the heat source temperature has to be shifted upward. An increase of 2.28°C (see Figure 4.37) yields the following values: T *source = 87.28˚C; ΔTpump = 42.72°C; a maximum source load (also taken as the selected source load) of 5.152 MW; a maximum sink load equal to the selected sink load of Qsink = 7.016 MW; COP = 3.7637; and W = 1.864 MW. Better results can be obtained by optimizing the two temperatures simultaneously while using the overall utility cost as a criterion. 140 QSINK
120
T* [°C]
100 The sink is oversized 80 QSOURCE
60 40 20 0
5
10
15
20
25
20
25
ΔH [MW]
FIGURE 4.36
Heat-pump sizing example: Attempt 2.
140 QSINK
120 100 T* [°C]
76
80 QSOURCE 60 40 20 0
5
10
15
ΔH [MW]
FIGURE 4.37
Heat-pump sizing example: Attempt 3.
Process Integration for Improving Energy Efficiency Once utility levels are chosen, the heat pump can be placed also across a Utility Pinch (Figure 4.38). A special case of integration involves the placement of refrigeration levels (Figure 4.39). Refrigeration facilities are heat pumps whose main value is that their cold end absorbs heat. Utilizing their hot ends for process heating can save considerable amounts of hot utility, especially when relatively low-temperature heating is needed.
Distillation Processes and Other Separations The simplest distillation column includes one reboiler and one condenser, and these components account for most of the column’s T*
HP
MP
IP LP
CW ΔH
FIGURE 4.38
Heat-pump placement across the utility Pinch. T*
Rcond W
HEAT PUMP
CW
0 R1 R2 ΔH
FIGURE 4.39
Refrigeration systems.
77
78
Chapter Four energy demands. For purposes of Heat Integration, the column is represented generally by a rectangle: the top side denotes the reboiler as a cold stream, and the bottom side denotes the condenser as a hot stream; see Figure 4.40. There are three options for integrating distillation columns: across the Pinch, as shown in Figure 4.41(a); and entirely below or entirely above the Pinch, as shown in Figure 4.41(b). Integrating across the Pinch only increases overall energy needs, so this option is not useful. The other two options result in net benefits by eliminating the need to use an external utility for supplying the distillation reboiler (above the Pinch) or the condenser (below the Pinch). The GCC is used to identify the appropriate column integration options. Figure 4.42 illustrates two examples of appropriately placed distillation columns.
T* A
Reboiler
A, B
Condenser
B
H
FIGURE 4.40 Distillation column: T–H representation.
(a) T*
A + QDIST
A
(b) T*
QDIST
QDIST
PINCH
PINCH Q
DIST
QDIST B + QDIST Across the pinch
FIGURE 4.41
QDIST
QDIST B Below or above the pinch
Distillation column: Integration options.
Process Integration for Improving Energy Efficiency FIGURE 4.42 Appropriate placement of distillation columns against the GCC.
T*
ΔH
When the operating conditions of a column result in placing it across the Pinch, there are several degrees of freedom that can be utilized to facilitate appropriate placement. It may be possible to change the operating pressure, which could shift the column with respect to the temperature scale until it fits above or below the Pinch. Varying the reflux ratio results in simultaneous changes of the column temperature span and the duties of the reboiler and condenser. Increasing the reflux ratio (the ratio between the reflux flowrate returned to the column and the distillate product flowrate) yields a smaller temperature span and larger duties, whereas reducing the ratio has the opposite effect. It is also possible to split the column into two parts, introducing a “double effect” distillation arrangement. In this approach, one of the effects is placed below and the other above the Pinch; this prevents internal thermal integration of the column effects. Additional options are available, such as interreboiling and intercondensing. When using available degrees of freedom, it is important to keep in mind that the energy–capital trade-offs of the column and the main process are combined and thus become more complicated than the individual trade-offs. Another important issue is controllability of the integrated designs: unnecessary complications should be avoided, and disturbance propagation paths should be discontinued. It is usually enough to integrate the reboiler or the condenser. If inappropriate column placement cannot be avoided, then condenser vapor recompression with a heat pump can be used to heat up the reboiler. Evaporators constitute another class of thermal separators. Because they, too, feature a reboiler and a condenser, their operation is similar to that of distillation columns and so the same integration principles can be applied to them. Absorber-stripper loops and
79
80
Chapter Four dryers are similarly integrated. This subject has been extensively studied by Kemp (2007, sections 6.3 to 6.5).
4.4.3
Process Modifications
The basic Pinch Analysis calculations assume that the core process layers in the onion diagram (Figure 4.2) remain fixed. However, it is possible—and in some cases beneficial—to alter certain properties of the process. Properties that can be exploited as degrees of freedom include: (1) the pressure, temperature, or conversion rate in reactors; (2) the pressure, reflux ratio, or pump-around flow rate in distillation columns; and (3) the pressure of feed streams in evaporators. Such modifications will also alter the heat capacity flow rates and temperatures of the related process streams for Heat Integration; this will cause further changes in the shapes of the CCs and the GCC, thereby modifying the utility targets. The CCs are a valuable tool for suggesting beneficial process modifications (Linnhoff et al., 1982). Figure 4.43 illustrates this application of the CCs in terms of the plusminus principle. The main idea is to alter the CC’s slope in the proper direction in order to reduce the amount of utilities needed. This can be achieved by changing CP (e.g., by mass flow variation). According to Smith (2005), such decreases in utility requirements can be brought about by (1) increases in the total hot stream duty above the Pinch, (2) decreases in the total cold stream duty above the Pinch, (3) decreases in the total hot stream duty below the Pinch, and/or (4) increases in the total cold stream duty below the Pinch. Another guide to modifying processes is the principle of keep hot streams hot (KHSH) and keep cold streams cold (KCSC). As illustrated in Figure 4.44, increasing the temperature differences by process
T
+ − − +
ΔH
FIGURE 4.43
The plus–minus principle.
Process Integration for Improving Energy Efficiency T
T
ΔH
ΔH Hot streams (a)
FIGURE 4.44
Cold streams (b)
(a) Keep hot streams hot; (b) keep cold streams cold.
modification allows for more overlap of the curves and results in improved heat recovery. In particular, energy targets improve if the heating and cooling demands can be shifted across the Pinch. The principle suggests (1) shifting hot streams (cooling demands) from below to above the Pinch and/or (2) shifting cold streams (heating demands) from above to below the Pinch. See Chapter 12 (and Figure 12.7) for more details.
4.5 HEN Synthesis Most industrial-scale methods synthesize heat recovery networks under the assumption of a steady state.
4.5.1 The Pinch Design Method The Pinch Design Method (Linnhoff and Hindmarsh, 1983) became popular owing to its simplicity and efficient management of complexity. The method has evolved into a complete suite of tools for heat recovery and design techniques for energy efficiency, including guidelines for changing and integrating a number of energy-intensive processes.
HEN Representation The representation of HENs by a general process flowsheet, as in Figure 4.45, is not always convenient. The reason is that this representation makes it difficult to answer a number of important questions: “Where is the Pinch?” “What is the degree of heat recovery?” “How much cooling and heating from utilities is needed?” The so-called conventional HEN flowsheet (Figure 4.46) offers a small improvement. It shows only heat transfer operations and is based on a simple convention: cold streams run horizontally and hot
81
82
Chapter Four Feed 2
4
140°C
3 230°C Reactor 2 250°C Product 2
180°C Reactor 1
80°C Off-Gas
Feed 1
40°C
1
20°C 40°C
2
Product 1, 40°C
FIGURE 4.45 Using a general process flowsheet to represent a HEN.
FIGURE 4.46 Conventional HEN flowsheet.
2
250°C
4
12.5 MW
140°C
3
200°C
181.7°C 230°C
205°C
H
7.0 MW
7.5 MW
203.3°C
150°C 8.0 MW
150°C
6.5 MW 1
140°C 52.5°C
106.7°C
20°C C
10.0 MW 40°C
180°C
17.5 MW Above the Pinch 80°C
streams vertically. Although the Pinch location can be marked for simple cases, it is still difficult to see. This representation makes it difficult to express the proper sequencing of heat exchangers and to represent clearly the network temperatures. Furthermore, changing
Process Integration for Improving Energy Efficiency the positions of some matches often results in complicated path representations. The grid diagram, as shown in Figure 4.47, provides a convenient and efficient representation of HENs by eliminating the problems just described. The grid diagram has several advantages: the representation of streams and heat exchangers is clearer, it is more convenient for representating temperatures, and the Pinch location (and its implications) are clearly visible; see Figure 4.48. As in the conventional HEN flowsheet, only heat transfer operations are shown in the grid diagram. Temperature increases from left to right in the grid, which is intuitive and in line with CC diagrams, making (re)sequencing heat exchangers are straightforward.
The Design Procedure The procedure for designing a HEN follows several simple steps: 1. Specification of the heat recovery problem. (a)
HOT 1 HOT 2 COLD 1 COLD 2
H Simplified view
(b)
MP Steam HOT 1 HOT 2
C
COLD 1 COLD 2
H
CW Showing the utility streams explicitly
FIGURE 4.47
H
Hot utility
C
Cold utility
The grid diagram for HENs.
Heat exchange between streams
83
84
Chapter Four 2. Identification of the heat recovery targets and the heat recovery Pinch, as explained in Section 4.3. 3. Synthesis 4. Evolution of the HEN topology The first two steps were discussed in previous sections. The synthesis step begins by dividing the problem at the Pinch and then positioning the process streams as shown in Figure 4.49. The engineering practice suggests starting the network design from the Pinch (the most restricted part of the design owing to temperature differences approaching ΔTmin) and then to place heat exchanger matches while moving away from the Pinch (Figure 4.50). When placing matches, several rules have to be followed in order to obtain a network that minimizes utility use: (1) no exchanger may have a temperature difference smaller than ΔTmin;
No cross-Pinch heat transfer
A cross-Pinch heat transfer match
C C
H H
FIGURE 4.48
The grid diagram and implications of the Pinch.
CP [kW/°C]
PINCH 150°C
40°C 80°C
1
20°C
150°C
140°C 3
140°C
QHmin = 750 kW
FIGURE 4.49
250°C 200°C
2
4
180°C
15 25
20 230°C
30
QCmin = 1000 kW
Dividing at the Pinch for the streams in Table 4.2.
Process Integration for Improving Energy Efficiency T
PINCH
ΔH
FIGURE 4.50
Pinch design principle.
(2) no process-to-process heat transfer may occur across the Pinch; and (3) no inappropriate use of utilities should occur. At the Pinch, the enthalpy balance restrictions entail that certain matches must be made if the design is to achieve minimum utility usage without violating the ΔTmin constraint; these are referred to as essential matches. Above the Pinch, the hot streams should be cooled only by transferring heat to cold process streams, not to utility cooling. Therefore, all hot streams above the Pinch have to be matched up with cold streams. This means that all hot streams entering the Pinch must be given priority when matches are made above the Pinch. Conversely, cold streams entering the Pinch are given priority when matches are made below the Pinch. Now recall the example from Table 4.2. Figure 4.49 shows the scaled grid diagram, indicating the hot and cold Pinch temperatures. The part above the Pinch requires essential matches for streams 2 and 4, since they are entering the Pinch. Consider stream 4. One possibility is to match it against stream 1, as shown in Figure 4.51. Stream 4 is the hot stream, and its CP is greater than the CP for cold stream 1. As shown in the figure, at the Pinch the temperature distance between the two streams is exactly equal to ΔTmin. Moving away from the Pinch results in temperature convergence because the slope of the hot stream line is less steep owing to its larger CP. Since ΔTmin is the lower bound on network temperature differences, the proposed heat exchanger match is infeasible and thus is rejected. Another possibility for handling the cooling demand of stream 4 is to implement a match with stream 3, as shown in Figure 4.52. The CP of stream 3 is larger than the CP of stream 4, resulting in divergent
85
86
Chapter Four CP [kW/°C]
PINCH 150°C 150°C
250°C 200°C
2
3
140°C 140°C
CPHOT > CPCOLD
15 PINCH
4
4
25
HEAT EXCHANGER MATCH 1
T
ΔTmin 1
180°C
20 230°C
ΔH Temperature difference smaller than ΔTmin
30
INFEASIBLE!
FIGURE 4.51
An infeasible heat exchanger match above the Pinch.
PINCH
CP [kW/ºC] 250ºC
150ºC
150ºC
200ºC
T 2
15
PINCH 4
4
25
HEAT EXCHANGER MATCH 1
140ºC 140ºC
3
FIGURE 4.52
3 ΔTmin
180ºC
20 230ºC
ΔH The match is feasible
30
A feasible heat exchanger match above the Pinch.
temperature profiles in the direction away from the Pinch. Thus, the rule above the Pinch may be expressed as follows: CPhot stream d CPcold stream
(4.9)
The logic below the Pinch is symmetric. This part of the design is a net heat source, which means that the heating requirements of the cold streams have to be satisfied by matching up with hot streams. The same type of reasoning as before yields the following requirement: the CP value of a cold stream must not be greater than the CP value
Process Integration for Improving Energy Efficiency of a hot stream if a feasible essential match is to result. Generalizing Eq. (4.9) shows that the CP of the stream entering the Pinch must be less than or equal to the CP of the stream leaving the Pinch: CPentering pinch d CPleaving pinch
(4.10)
The Pinch Design Method incorporates a special tool for handling this stage: CP tables (Linnhoff and Hindmarsh, 1983). Here the streams are represented by their CP values, which are sorted in descending order. This facilitates the identification of promising combinations of streams for candidate essential matches. Sizing the matches follows the so-called tick-off heuristic, which stipulates that the heat exchange match should be as large as possible so that at least one of the involved streams will be completely satisfied and then “ticked off” from the design list; see Figure 4.53.
Completing the Design The HEN design above the Pinch is illustrated in Figure 4.54. The design below the Pinch follows the same basic rules, with the small difference that here it is the cold streams that define the essential matches. Figure 4.55 details the design below the Pinch for the considered example. First the match between streams 4 and 1 is placed and sized to the duty required by stream 4. The other match, between streams 2 and 1, formally violates the Pinch rule for placing essential matches. However, since stream 1 already has another match at the Pinch, the current match (between 4 and 1) is not strictly termed essential. Up
CP [kW/ºC] 150ºC
203.3ºC
250ºC
150ºC
200ºC
140ºC
180ºC
1
QSTREAM [kW]
2
15
1500
4
25
1250
20
800
30
2700
800 kW 140ºC 3
1250 kW
FIGURE 4.53
230ºC
TX = 181.7ºC TX =140 +
The tick-off heuristic.
1250 30
=181.7ºC
87
88
Chapter Four
150°C
203.3°C
150°C
250°C
200°C
2
4
CP [kW/°C]
QSTREAM [kW]
15
1500
25
1250
20
800
30
2700
180°C
140°C 1 800 kW 140°C
TX = 181.7°C
3
1250 kW
FIGURE 4.54
205°C 700 kW
H
230°C
750 kW
The HEN design above the Pinch. CP QSTREAM [kW/°C] [kW]
40°C
C
150°C
106.7°C
2
15
1650
4
25
1750
20
2400
1000 kW 150°C
80°C
1
52.5°C
20°C 650 kW
FIGURE 4.55
140°C 1750 kW
The HEN design below the Pinch.
to the required duty of 650 kW, this match does not violate the ΔTmin constraint, which is the relevant one. The completed HEN topology is shown in Figure 4.56. It is not always possible to follow the basic design rules described previously, so in some cases it is necessary to split the streams so that heat exchange matches can be appropriately placed. Splitting may be required in the following situations: 1. Above the Pinch when the number of hot streams is greater than the number of cold streams (NH > NC); see Figure 4.57. 2. Below the Pinch when the number of cold streams is greater than the number of hot streams (NC > NH); see Figure 4.58. 3. When the CP values do not suggest any feasible essential match; see Figure 4.59.
Process Integration for Improving Energy Efficiency CP [kW/°C]
PINCH 150°C
106.7°C C 1000 kW 80°C
40°C
1
203.3°C
150°C
20°C
200°C
4
140°C
Tx =181.7 °C
205°C
QCmin =1000 kW
QHmin = 750 kW
Completed HEN design.
(a)
CP [kW/°C]
PINCH 100°C 1
12
2
20
3
7
100°C 100°C I?
5
90°C
T > 90°
90°C
T > 90°
15 30 More Hot Streams than Cold Streams
(b) CP [kW/°C]
PINCH 100°C 1
12
2
20
3
7
100°C 100°C 90°C 4
15 90°C
5
30
Split a cold stream
FIGURE 4.57
25 20
230°C 30 H 700 kW 750 kW
1250 kW
4
15
800 kW
650 kW 1750 kW 3
2
180°C
140°C
52.5°C
FIGURE 4.56
250°C
Splitting above the Pinch for NH > NC .
89
90
Chapter Four CP [kW/°C]
PINCH T<100
35
100°C 1
T<100 60
100°C
2
? 24
3
30
4
20
5
90°C 90°C 90°C
(a) More Cold Streams than Hot Streams
PINCH
CP [kW/°C]
100°C
35 100°C
60
24
3
30
4
20
5
1 2
90°C 90°C 90°C
(b) Split a Hot Stream
FIGURE 4.58 Splitting below the Pinch for NC > NH.
The loads of the matches involving stream branches are again determined using the tick-off heuristic. Because each stream splitter presents an additional degree of freedom, it is necessary to decide how to divide the overall stream CP between the branches. One possibility is illustrated in Figure 4.60. The suggested split ratio is 4:3. This approach would completely satisfy the heating needs of the cold stream branches. The arrangement in Figure 4.60 is actually trivial and can be improved upon. Unless the stream combinations impose some severe constraints, there is a large number of possible split ratios. This fact can be exploited to optimize the network. In many cases it is possible to save an extra heat exchanger unit by ticking off two steams with a
Process Integration for Improving Energy Efficiency (a)
Rules: CPH ≥ CPC ; NH ≥ NC
PINCH
30°C
CP [kW/°C]
100°C 1 40°C
5 4
100°C 2 ?
?
15°C
3
90°C
7
CP rules are not satisfied (b) Rules: CPH ≥ CPC ; NH ≥ NC
PINCH
30°C
100°C
40°C
100°C
15°C
CP [kW/°C]
1
5
2
4
90°C
3
4 3
Split the cold Stream
FIGURE 4.59
Splitting to enable CP values for essential matches.
PINCH 30°C
C1
100°C
1
CP [kW/°C]
1
5
2
4
T1 = 100 −300/5 = 40°C T2 = 100 −225/4
40°C
3
C2
T2 = 43.75°C
15°C
2
100°C
90°C
4 3
Q1 = 4 × (90 − 15)= 300 kW
FIGURE 4.60
Splitting and trivial tick-off.
Q2 = 3 × (9 − 15) =225 kW
91
92
Chapter Four single match, as illustrated in Figure 4.61. This satisfies both hot stream 2 and the corresponding cold branch of stream 3, thereby eliminating the second cooler. The complete algorithm for splitting streams above the Pinch is given in Figure 4.62. The procedure for splitting below the Pinch is symmetrical, with the cold and hot streams switching their roles.
Network Evolution Any HEN obtained using the design guidelines described previously is optimal with respect to its energy requirements, but it is usually away from the total cost optimum. Observing the Pinch division PINCH T1 = 100 − 285/5 T1 = 43°C
30°C C1
100°C
1
CP [kW/°C]
1
5
2
4
Step 1: Q2 = 4 (100 − 40) = 240 kW 1
40°C
100°C
2 Step 3: Q1 = 3.8 × (90 − 15) = 285 kW 15°C
3
2
90°C
Q1
3.8
Q2
3.2
Step 2: CPC2 = 240(90 − 15) = 3.2
FIGURE 4.61
Splitting and advanced tick-off.
Stream Data at Pinch Yes
Place Matches
CPH ≤ CPC ?
Yes
NH ≤ CPC ? No
No
Split Cold Stream Split Hot Stream
FIGURE 4.62
Splitting procedure above the Pinch.
Process Integration for Improving Energy Efficiency typically introduces loops into the final topology and leads to larger number of heat exchanger units. The final step in HEN design is evolution of the topology: identifying heat load loops and open heat load paths; then using them to optimize the network in terms of heat loads, heat transfer area, and topology. During this phase, formerly rigorous requirements—for example, that all temperature differences exceed ΔTmin and that cross-Pinch heat transfers be excluded—are usually relaxed. The resulting optimization formulations are typically nonlinear and involve structural decisions, so they are MINLP problems. Different approximations and simplifying assumptions can be introduced to obtain linear and/or continuous formulations. The design evolution step can even be performed manually by breaking the loops and reducing the number of heat exchangers. Eliminating heat exchangers from the topology is done at the expense of shifting heat loads (from the eliminated heat recovery exchangers) to utility exchangers: heaters and coolers. Topology evolution terminates when the resulting energy cost increase exceeds the projected savings in capital costs, which corresponds to a total cost minimum. Network evolution is performed by shifting loads within the network toward the end of eliminating excessive heat exchangers and/or reducing the effective heat transfer area. To shift loads, it is necessary to exploit the degrees of freedom provided by loops and utility paths. In this context, a loop is a circular closed path connecting two or more heat exchangers, and a utility path connects a hot with a cold utility or connects two utilities of the same type. Figure 4.63 shows a HEN loop and a utility path. A network may contain many such loops and paths.
4.5.2 Superstructure Approach As presented so far, the Pinch Design Method is based on a sequential strategy for the conceptual design of HENs. It first develops an
40°
+W
+U
C
1 80°
1
UTILITY PATH
2
−U
−W
4
5
250° 200°
3
20°
4
180° +U LOOP 140° 3
−U H −W
FIGURE 4.63
2
Loop and path in a Heat Exchanger Network.
+W
230°
93
94
Chapter Four understanding of the thermodynamic limitations imposed by the set of process streams, and then it exploits this knowledge to design a highly energy-efficient HEN. However, another approach has also been developed: the superstructure approach to HEN synthesis, which relies on developing a reducible structure (the superstructure) of the network under consideration. An example of the spaghettitype HEN superstructure fragments typically generated by such methods (Yee et al., 1990) is shown in Figure 4.64. Typically, this kind of superstructure is developed in stages (Yee et al., 1990) or blocks (Zhu et al., 1995), each of which is a group of several consecutive enthalpy intervals (discussed previously). Within each block, each hot process stream is split into a number of branches corresponding to the number of the cold streams presented in the block, and all cold streams are split similarly. This is followed by matching each hot branch with each cold branch. Once developed, the superstructure is subjected to optimization. The set of decision variables includes the existence of the different stream split branches and heat exchangers, the heat duties of the exchangers, and the split fractions or flow rates of the split streams. The objective function involves mainly the total annualized cost of the network, although the function may be augmented by some penalty terms for dealing with difficult constraints. Because the optimization procedure makes structural as well as operating decisions about the network being designed, it is called a structure-parameter2 optimization. Depending
Enthalpy Interval
Hot Streams
Cold Streams
FIGURE 4.64
Spaghetti superstructure fragment.
2 This should not be confused with the “parameter” entities from Mathematical Programming.
Process Integration for Improving Energy Efficiency on what assumptions are adopted, it is possible to obtain both MILP and MINLP formulations. Linear formulations are usually derived by assuming isothermal mixing of the split branches and then using piecewise linearization on the heat exchanger capital cost functions. With the superstructure approach it is possible to include other heat exchange options—for example, direct heat transfer units (i.e., mixing) and different heat exchanger types (e.g., double-pipe, platefin, and shell-and-tube). Soršak and Kravanja (2004) presented a method incorporating different heat exchanger types into a superstructure block for each potential heat exchange match. Some other interesting works in this area are by Daichendt and Grossmann (1997), Zamora and Grossman (1998), Björk and Westerlund (2002), and Frausto-Hernández et al. (2003).
4.5.3 A Hybrid Approach It is clear that the superstructure methodology offers some advantages in the synthesis of process systems and, in particular, of HENs. Among these advantages are: (1) the capacity to evaluate a large number of structural and operating alternatives simultaneously; (2) the possibility of automating (to a high degree) the synthesis procedure; and (3) the ability to deal efficiently with many additional issues, such as different heat exchanger types and additional constraints (e.g., forbidden matches). However, these advantages also give rise to certain weaknesses. First, the superstructure approaches, in general, cannot eliminate the inherent nonlinearity of the problem. Hence they resort to linearization and simplifying assumptions, such as allowing only isothermal mixing of split streams. Second, the transparency and visualization of the synthesis procedure are almost completely lost, excluding the engineer from the process. Third, the final network is merely given as an answer to the initial problem, and it is difficult to assess how good a solution it represents or whether a better solution is possible. Fourth, the difficulties of computation and interpreting the result grow dramatically with problem size; this is a consequence of the large number of discrete alternatives to be evaluated. Finally, the resulting networks often contain subnetworks that exhibit a spaghetti structure: a cluster of parallel branches on several hot and cold streams with multiple exchangers between them. Because of how superstructures are constructed, these subnetworks often cannot be eliminated by the solvers. All these considerations highlight the fundamental trade-off between applying techniques that are based on thermodynamic insights, such as the Pinch Design Method, and relying on a superstructure approach. It would therefore be useful to find a combination of both approaches and, if there is, to see whether it offers any advantages. One such middle way is the class of hybrid synthesis methods described next. A method of this class first
95
96
Chapter Four applies Pinch Analysis to obtain a picture of the thermodynamic limitations of the problem; but then, instead of continuing on to direct synthesis, it builds a reduced superstructure. At this point the method follows the route of the classical superstructure approaches, including structure-parameter optimization and topology simplifications. The cycle of optimization and simplification is usually repeated several times before the final optimal network is obtained. All resulting networks feature a high degree of heat recovery, though rarely the maximum possible. A key component of this technique is avoiding the addition of unnecessary features to the superstructure, and this is an area where Pinch Analysis can prove helpful. A good example of a hybrid method for HEN synthesis is the block decomposition method (Zhu, 1997).
4.5.4 Key Features of the Resulting Networks The networks obtained by the different synthesis methods have distinct features, which influence their total cost and their properties of operation and control. Because the Pinch Design Method incorporates the tick-off heuristic rule (Figure 4.53), the networks synthesized by this method tend to have simple topologies with few stream splits and feature minimum number of units (Linnhoff et al., 1982). Both the tick-off rule and the Pinch principle dictate that utility exchangers be placed last, so they are usually located immediately before the target temperatures of the streams. However, the tick-off rule may also result in many process streams not having utility exchangers assigned to them, which may reduce control efficiency. The Pinch Design Method may reduce network flexibility because it relies on the Pinch decomposition of the problem (Figure 4.50) and so, to a large degree, fixes the network behavior. Both the pure superstructure approach and the hybrid approach tend to produce more complex topologies. Their distinctive feature is the greater number of heat exchangers and stream splits, a result of how the initial superstructure is built. Spaghetti-type subnetworks also present a significant challenge to control.
4.6 Total Site Energy Integration The concept of the Total Site was introduced by Dhole and Linnhoff (1993b). Figure 4.65 shows a typical industrial Total Site. Refinery and petrochemical processes usually operate as parts of large sites or factories. These sites have several processes serviced by a centralized utility system involved in steam and power generation. The two major components of Total Site integration are closely related: heat recovery (through the steam or utility system) and power cogeneration.
Process Integration for Improving Energy Efficiency
Emissions Emissions Power Fuel 1 Fuel 2
Condensing
Emissions
Fuel 3
FIGURE 4.65
Plant A
Plant B
HP Steam MP Steam LP Steam
Plant C Cooling
Schematic of an industrial Total Site.
4.6.1 Total Site Data Extraction The heating and cooling requirements of the individual processes are represented by their respective GCCs (see Section 4.3.5). The GCC represents the process-utility interface for a single process. These individual GCCs can be used to identify the potential heat recovery via steam mains. When a site houses several production processes, the GCC of each process may indicate certain steam levels that are suitable for the given process. This suggests that trade-offs need to be made among energy demands of the various processes on a Total Site, since each process usually needs utility heating or cooling at levels different from the other processes.
4.6.2 Total Site Profiles It is possible to set utility targets for sites involving several processes (Dhole and Linnhoff, 1993b; Linnhoff and Dhole, 1993; see also extensions by Klemeš et al., 1997). The procedure is based on thermal profiles for the entire site that are called, naturally enough, Total Site Profiles (TSPs). These profiles are constructed from the GCCs of the individual processes on the site. The first step (see Figure 4.66) is to extract the net heating and cooling demands. Two options are possible: restricting the Heat Integration to each process (so that heat recovery pockets are not considered for Total Site analysis) or to allow extended integration across processes by including the GCC segments forming the pockets. When the scope of site integration is extended, however, some design possibilities
97
Chapter Four Shifed – T** Temperature
98
T** T* T*
PINCH Shifed – T** 1.1. Remove the pockets 1.2. Shift
1. Extract segments
ΔH
T**
0 3. Combine segments
2. Rotate source segments
ΔH
T**
ΔH
0 4. Align profiles
ΔH
FIGURE 4.66 Construction of the Total Site Profiles when the heat recovery pockets are excluded from site integration.
become impractical—for instance, the analysis may try to integrate streams that are distant, or there may be control and start-up problems. Note that the parts of each GCC that are directly satisfied by local utilities (e.g., furnaces within the processes) are also excluded from the analysis; the remaining curve parts are those representing the net heat source and sink demands to be satisfied by the central utility system. As shown in the figure, subseqent steps include the rotation of heat source segments (for purely graphical reasons), the thermal combination of stream segments (much as in the construction of process CCs), and alignment of the resulting Total Site source profile and Total Site sink profile. As shown in Figure 4.66, the source and sink elements extracted from the GCCs are shifted by ΔTmin/2: the temperatures of the heat source segments are reduced while those for the sinks are increased This operation ensures that all temperatures in the picture remain in the scale of the true utility temperatures so that, if a utility profile touches the process-derived profiles, then there will be just enough temperature driving force to effect the heat transfer. The composite of the heat sources elements is the Site Source Profile and that of the sinks is the Site Sink Profile. Site Source Profiles and Site Sink Profiles derive primarily from the process GCCs. Other steam requirements (mostly for process use not directly related to heating) are usually not represented in the GCC; examples include steam for ejectors and reactors as well
Process Integration for Improving Energy Efficiency as unaccountable steam usage. These additional requirements have to be considered when analyzing or designing the site’s utility system. Such steam demands are considered to be sink elements and are added to the Site Sink Profile without any temperature correction.
4.6.3 Heat Recovery via the Steam System The maximum possible heat recovery through a utility system can be targeted by using the Site Sink and Source Profiles in combination with the steam header saturation temperatures. Source CCs for utility generation and usage are constructed that account for feasible heat transfer from the Site Source Profile to the site source CC and from the site sink CC to the Site Sink Profile. The site CCs are analogous to the individual process CCs. The source CC is built starting from the highest feasible steam level. The steam generation at each level is maximized before the next lower levels are analyzed. This ensures maximum utilization of the heat sources’ temperature potential. The remainder of the Site Source Profile (i.e., the part that does not overlap the newly built source CC) is served by CW. Building the sink CC follows a symmetrical procedure only starting from the lowest possible steam level. The use of this level is maximized before moving up to the next higher temperature level, and so forth until steam with the highest possible level is used—including boiler-generated steam. Figure 4.67 illustrates the procedure for building the Total Site CCs; here, the TSPs from Figure 4.66 were reused. There are several From boilers 4
250
3
200 Temperature [°C]
VHP
HP
1 2
MP
2 150
3
1
LP
100
CW 50 −25
4
−20
−15
−10
−5
0
5
10
Enthalpy [MW]
FIGURE 4.67
Constructing Total Site Composite Curves.
15
20
25
99
Chapter Four steam pressure levels, which are represented by their corresponding saturation temperatures. The main distribution levels are highpressure (HP) steam at a saturation temperature of 200°C , mediumpressure (MP) steam at 170°C, and LP steam at 115°C. In addition, very-high-pressure (VHP) steam at 250°C is generated by the steam boilers, and CW is also available. The Site Source Profile in Figure 4.67 can generate at most 2 MW of HP steam, another 2 MW of MP steam, and 15 MW of LP steam. The rest of the heat sources will need to be served by the CW, which completes the Site Source CC. The Site Sink CC, indicates needs for 1 MW of LP steam, 6.5 MW of MP steam, 2.5 MW HP steam, and the remaining 5 MW demand to be satisfied by VHP steam. The two site CCs can be overlapped in the same way as were the process CCs, thereby illustrating the Total Site heat recovery possible through the steam system. Figure 4.68 shows the corresponding site CCs overlapped. The amount of heat recovery for the Total Site is indicated by the amount of overlap between the CCs. Heat recovery is maximized when the two Site Composite Curves touch and cannot be shifted further. The area where the curves touch, which is usually confined between two steam levels (here, the MP and LP levels), is the Total Site Pinch. The steam mains located at the Site Pinch feature opposite net steam loads; in other words, the steam main above the Site Pinch is a net steam user while the one below the Site Pinch is a net steam generator. Just as with the Process Pinch, the site Pinch divides an overall heat recovery problem into a net heat source and a net heat sink.
Temperature [°C]
100
250
VHP
200
HP MP
150
LP
100 Heat recovery via steam
CW 50 −25
−20
−15
−10
−5
0
5
Enthalpy [MW]
FIGURE 4.68
Targeting heat recovery.
10
15
20
25
Process Integration for Improving Energy Efficiency
4.6.4 Power Cogeneration The factors that control the economics of utility systems are fuels and their properties, the ratio of fuel prices to power prices, efficiency of the utility system, and the amount of power to be imported/exported. Depending on these factors and the Total Site power demand, the site may be a net power importer or exporter or be in power balance. Most industrial processes require steam at different pressure levels up to about 30 bar. Central utility boilers usually generate steam at higher pressure (40–100 bar). Back-pressure steam turbines are used to expand the steam from higher- to lower-level steam headers, thus generating power while delivering steam to processes. Another method of cogeneration employs a gas turbine, which is in itself a power generating equipment. A gas turbine produces large amounts of waste heat along with the power—the ratio of heat to power is about 1.5–2. This is high-temperature (450–600°C) waste heat capable of generating even VHP steam. The heat from a gas turbine’s exhaust stream can be utilized by heat recovery steam generators with or without supplementary firing. The generated steam is expanded through steam turbines to produce additional power. The work of Dhole and Linnhoff (1993b) has been further developed by Raissi (1994) and Klemeš et al. (1997). The latter paper describes the development of a tool called the Site Utility GCC (SUGCC). The area enclosed by this curve is proportional to the power cogeneration potential of the site steam system; Klemeš et al. (1997) also defined a simple proportionality coefficient, whose value is usually evaluated for each industrial site separately. This cogeneration targeting model is referred to as “the T-H model” because it is based on heat flows through the steam system. Using SUGCCs allowed Klemeš and colleagues (1997) to set thermodynamic targets for cogeneration of power along with targets for site-scope heat recovery that would minimize the cost of utilities. Satisfying the goal of maximum heat recovery leads to a minimum boiler VHP steam requirement, which in turn can be achieved by maximizing steam recovery. Here the power generation by steam turbines is also minimal, which has the effect of maximizing imported power. This scenario can be represented by the Site CCs that are shifted to a position of maximum overlap (i.e., pinched). This target represents the thermodynamic limitation on system efficiency, but this is not a specification that must be achieved. The case of minimizing the cost of utilities is handled by exploring the trade-off between steam recovery and power cogeneration by steam turbines. If design guidelines are thus based on minimizing cost, then the resulting network design is usually different from that produced when aiming to minimize fuel consumption. Mavromatis and Kokossis (1998) proposed a simple model of back-pressure steam turbine performance. In this model, the performance of a steam turbine is related to its size (in terms of
101
102
Chapter Four maximum shaft power) and to part-load performance; the shaft power is modeled as a function of the steam mass flow that is known as the Willan line. This model was extended to condensing steam turbines by Shang (2000). All these works follow the same model structure and employ the same equations; however, they use different values for the turbine regression coefficients. The intercept of the Willan line was mapped by Mavromatis and Kokossis (1998) and by Shang (2000) as identical to the turbine energy losses, also assuming a fixed loss rate. Varbanov, Doyle, and Smith (2004) introduced improvements to those models by (1) recognizing that the Willan line intercept has no direct physical meaning and is simply the intercept of a linearization, and (2) accounting for both inlet and outlet pressures of the steam turbines. These improved steam turbine models have been incorporated into methodologies for simulating and optimizing steam networks; they have also been used to target heat and power cogeneration by assuming a single large steam turbine for each expansion zone between two consecutive steam headers.
4.6.5 Advanced Total Site Optimization and Analysis A model for optimizing the utility system serves as a tool for reducing site operating costs related to energy and for analyzing the thermodynamic limitations of energy conversions. An advanced approach to these concepts, known as top-level analysis, is one that allows “scoping” which site processes to target for Heat Integration improvement (Varbanov, Doyle, and Smith, 2004). Consider the utility system shown in Figure 4.69 (Smith and Varbanov, 2005), whose operating properties have been optimized for the given steam and power demands. Suppose it were possible to reduce the HP steam demand—for example, by improving the energy efficiency within the processes that use HP steam. What, then, would such a saving in steam actually be worth? Reducing HP steam demand means that less steam needs to be expanded from the VHP level, which could lead in turn to less power cogeneration and increased import of power. As a support tool for deciding how best to utilize the potential steam excess and estimate the value of potential steam savings, Varbanov, Doyle, and Smith (2004) introduced the concept of marginal steam price. This characteristic captures the change in a utility system’s energy cost per unit change in steam demand, and it is specific to a given combination of steam header and operating conditions. By optimizing the utility system at gradual successive reductions of potential steam demand on the headers, it is possible to obtain a curve of the marginal steam price versus the savings that could be obtained. The marginal price curve for the utility system in
Process Integration for Improving Energy Efficiency Natural Gas Coal
Current: 250.0 [50.0; 250.0]
Fuel Oil
Boiler 1
Current: 55.4 [40.0; 180.0]
Boiler 2
VHP (101 bar abs.)
LP Vent
T6
Current: 59.5 [15.0; 65.0]
T4
20.0 0.0 HP (41 bar abs.)
Current: 90.0 [30.0; 9.0] 0.0
Air
Current: 28
Current: 208.4 [115.0; 335.0]
Current: 105.0 [60.0; 165.0] Current: 15.0 [15.0; 75.0]
GT 18 MW
HRSG
Current: 40.9 [30.0; 70.0]
T5
Current: 20.0 [20.0; 20.0]
Current: 108.0 [60.0; 150.0]
T7
CW 0.073 bar abs.
54.5 0.0
MP (15 bar abs.) 55.9
0.0
LP (3 bar abs.) 128.0
Current: 75.0 [40.0; 75.0]
All flowrates are in [t/h]. The lower and upper bounds are shown in the format [min;max].
FIGURE 4.69
Optimized utility system.
FIGURE 4.70 Marginal steam prices for the utility system shown in Figure 4.69.
HP Steam
8
MP Steam
Marginal Price, $/t
7 6 5 4 3 2 1 0 0
10
20
30
40
50
60
70
Steam Savings, t/h
Figure 4.69 is shown in Figure 4.70. The plot indicates that the most potential for improvements is in the HP steam using processes, followed by MP users; the potential for LP steam savings is evidently quite modest.
103
This page intentionally left blank
CHAPTER
5
Mass Integration
T
he two main branches of Process Integration are energy integration and mass integration. Mass integration is a systematic methodology that provides a fundamental understanding of the global flow of mass within the process and then employs this understanding to identify performance targets and to optimize the allocation, separation, and generation of streams and species. In the context of wastewater minimization, a mass integration problem involves transferring mass (contaminant load) from rich process streams to lean process streams in order to achieve their target outlet concentration and simultaneously minimizing waste generation and the consumption of utilities—including freshwater and external mass separating agents (Rašković, 2006).
5.1 Water Integration Water is widely used in various industries as an important raw material. It is also frequently used in the heating and cooling utility systems (e.g., steam production, cooling water) and as a mass separating agent for various mass transfer operations (e.g., washing, extraction). Strict requirements for product quality and associated safety issues in manufacturing contribute to large amounts of high-quality water being consumed by the industry. Stringent environmental regulations coupled with a growing human population that seeks improved quality of life have led to increased demand for quality water. These developments have increased the need for improved water management and wastewater minimization. Adopting techniques to minimize water usage can effectively reduce both the demand for freshwater and the amount of effluents generated by the industry. In addition to this environmental benefit, efficient water management reduces the costs for acquiring freshwater and treating effluents. A number of different methodologies have been applied to minimizing water use and effluents. These include • Minimizing water consumption through efficient management and control of process operations
105
106
Chapter Five • Optimizing the material and energy balances of processes by applying advanced optimization strategies that aim to reduce waste • Integrating optimization and production planning techniques in conjunction with real-time plant measurements to control for product quality and minimize losses • Increasing the use of enhanced intelligent support to operators by applying knowledge-based decision-making procedures to select options that best protect the environment • Employing Process Integration techniques that are based on Pinch Analysis Each processing industry has its own unique and specific features. In all cases, however, it is advisable to progress from the simplest measures—such as good housekeeping based on efficient management, control, and maintenance—to more advanced methodologies. Some processes are continuous and run seven days a week for the whole year; others are intermittent and/or highly dependent on availability of the feed stock. Typical of such “campaign” production are plants that process sugar, fruit juice, and cereal. In contrast, breweries operate on a nearly continuous basis, although the processing is performed in batches. All these factors influence investments in processing plants and the technologies adopted, including those that involve water.
5.2 Minimizing Water Use and Maximizing Water Reuse 5.2.1 Legislation Water use and wastewater discharge are both subject to national and international standards. For the United States, the most significant water-related federal legislation includes: (1) the National Pollutant Discharge Elimination System (NPDES) permit program (1972); (2) the Clean Water Act (Federal Water Pollution Control Amendments of 1972), as amended by the Clean Water Act of 1977; (3) the Safe Drinking Water Act of 1974; (4) the Toxic Substances Control Act of 1976; and (5) the Water Quality Act of 1987. Individual states also legislate regarding water— for example, the comprehensive water legislation passed in California and signed by the governor in 2009. Legislation in member countries of the European Union (EU) follows the directives of the European Commission (EC), which is the executive body of the EU. The most relevant of these directives are published on the official EU web site (EUROPA, 2009). Selected topics and titles include: • The new European water policy: river basin management
Mass Integration • Water Framework Directive (2000/60/EC) • Strategies to prevent chemical pollution of surface water under the Water Framework Directive • “Priority” substances under Article 16 of the Water Framework Directive • A European action program on flood risk management • Discharges of Dangerous Substances Directive (76/464/EEC) • Water pollution stemming from urban wastewater and certain industrial sectors; Urban Waste Water Treatment Directive (91/271/EEC) • Water pollution caused by nitrates from agricultural sources; Nitrates Directive (91/676/EEC) • Bathing water quality of rivers, lakes, and coastal waters; Bathing Water Quality Directive (76/160/EEC) and proposed revisions • Drinking water quality; Drinking Water Directive (98/83/EC) Most of these items directly (or indirectly) concern the water used and wastewater discharged by processing industries.
5.2.2 Best Available Techniques The Integrated Pollution Prevention and Control (IPPC) Directive (96/61/EC) introduced a framework within which EU member states are required to issue operating permits for industrial installations performing certain activities. These permits must prescribe conditions that are based on best available techniques (BATs). Best available techniques are those with the best overall environmental performance that can be introduced at a reasonable cost, and their purpose is to ensure a high level of protection for the environment as a whole. A key aim of the IPPC Directive is to stimulate an intensive exchange of information on BAT between the European member states and affected industries. The European IPPC Bureau (eippcb. jrc.es) organizes this exchange of information and produces BAT reference documents (BREFs), which member states must take into account when establishing permit conditions. The bureau carries out its work through technical working groups (TWGs) consisting of nominated experts from industry, EU member states, countries in the European Free Trade Association, and nongovernmental organizations concerned with the environment. Because the European IPPC Bureau is located in Seville, Spain, activities carried out within the framework of the IPPC Directive are often referred to as “the Seville process.” Several BAT-oriented studies have been made in food processing industries. A good example comes from the Flemish Centre for Best Available Techniques (BAT-CENTRE, 2009). This document contains
107
108
Chapter Five an overview of available information on the fruit and vegetable– processing industry. Using BAT as guidance, the study proposes: • To Flemish authorities: Permit conditions and techniques for which investment support may be offered because they are less detrimental to the environment • To Flemish companies: Guidelines for implementing the concept of BAT As described in the study, the fruit and vegetable–processing industry comprises the sectors of frozen fruits and vegetables, canned fruits and vegetables, processed potatoes, peeled potatoes, and fruit juices. The most important environmental problems are the use of large volumes of ground water and the production of wastewater polluted with organic carbon, nitrogen, and phosphorus. Information on BAT candidates was obtained mostly from expertise present in Belgium and neighboring countries. More than a hundred different techniques were selected and examined in terms of technical and economical feasibility. Best available techniques in wastewater treatment incorporate, for example, primary and aerobic wastewater treatment for small potato-peeling enterprises as well as primary, anaerobic, and aerobic wastewater treatment for large-scale processing sites. The BAT concept was the basis for concluding that Flemish wastewater discharge limits on surface water were technologically and economically feasible, although new limits (25–50 mg/L) on total phosphorus discharge were suggested. Annual wastewater treatment costs for an average enterprise were estimated to be €2.5–3.5 million. For small potato-peeling companies, wastewater discharge into the sewer system was found to be most appropriate. Water-saving measures and the reuse of water may reduce groundwater consumption by as much as 25–30 percent. A good source for BAT practices is ENVIROWISE—a U.K. government program managed by Momenta (a division of AEA Technology Plc) and TTI (a division of Serco Ltd.)—which offers practical environmental advice for business. Their web site (ENVIROWISE, 2009) provides a wide range of information including news and best practice examples.
5.2.3 Water Footprint Hoekstra (2008) defined the water footprint (WFP) as an indicator of direct and indirect water use, which is measured in terms of water volumes consumed, evaporated, and/or polluted. The WFP includes consumptive use of “virtual” green, blue, and gray water. The virtual green water content of a product is the volume of rainwater that evaporated during the production process. For the food industry, this volume is consumed mainly by agricultural products and
Mass Integration represents the total rainwater evaporation from the fields during a crop’s growing period. The virtual blue water content of a product is the volume of surface water or groundwater that evaporates as a result of production—for example, the evaporation of irrigation water from fields, irrigation canals, and storage reservoirs. The virtual grey water content of a product is the volume of water required to dilute pollutants in order to meet water quality standards for reuse or discharge to the environment. A water footprint can be calculated for any product or activity that has a well-defined group of producers and consumers. The water footprint is a geographically and temporally explicit indicator: it reflects not only volumes of water consumption and pollution but also the type of water use as well as where and when the water was used. The idea of water life-cycle assessment has gained more interest since the concept of a water footprint was introduced (Hoekstra and Chapagain, 2007; Hoekstra, 2008). In food supply chains, the actual water content of the final product is usually negligible when compared with the virtual water content, which is the total fresh water used during the various steps of supply and production. Aside from the water that appears as an ingredient in prepared foods, most water use in the food industry consists of the virtual water described in the previous paragraph (Casani, Rouhany, and Knøchel, 2005). The most common water-using operations are as follows: • Heating: Boilers, heat exchangers, etc. • Process water: Cooling towers • Potable uses: Offices, canteens, etc. • Washing: Equipment, bottles, floors, vehicles, etc. • Rinsing: Equipment, bottles, food materials, final products • Firefighting • Transport medium The food industry consumes a large amount of water: its consumption was estimated to be 347.2 Mm3 in Canada (Dupont and Renzetti, 1998) and 455 Mm3 in Germany (Fahnich, Mavrov, and Chmiel, 1998). Other studies (Hoekstra and Chapagain, 2007; Water Footprint Network, 2009) have reported figures for use of virtual water in the production of some common food products; see Table 5.1. Just as in the case of heat recovery, for water recovery it is best to start with simple measures based on efficient management (e.g., good housekeeping) before moving on to more advanced methodologies (e.g., Process Integration techniques). Industrial operations are not always run in continuous mode, since they depend to a great extent on the availability of feed stock and the need to control quality. As mentioned before, sugar, fruit juices, and cereal are typically processed intermittently, and breweries operate continuously. These
109
110
Chapter Five Product
Virtual water (L)
1 glass of beer (250 mL)
75
1 glass of milk (200 mL)
200
1 cup of coffee (125 mL)
140
1 cup of tea (250 mL)
35
1 slice of bread (30 g)
40
1 slice of bread (30 g) with cheese (10 g)
90
1 potato (100 g)
25
1 apple (100 g)
70
1 glass of wine (125 mL)
120
1 glass of apple juice (200 mL)
190
1 glass of orange juice (200 mL)
170
1 bag of potato crisps (200 g)
185
1 egg (40 g) 1 hamburger (150 g)
135 2400
1 tomato (70 g)
13
1 orange (100 g)
50
TABLE 5.1
Virtual Water Consumed while Processing Selected Food Industry Products (after Hoekstra and Chapagain, 2007)
factors will affect the choice of technologies to be adopted, including those that involve water and wastewater. A comprehensive survey of water and wastewater applications for the processing industry can be found in the books of Smith (2005) and Klemeš, Smith, and Kim (2008). Most of these techniques for water minimization fall into one of two groups: 1. Process changes: This category groups fundamental changes in unit operations that consume fresh water. Examples include increasing the number of stages in an extraction process to reduce its consumption of water, changing from wet cooling towers to air coolers, improving energy efficiency to reduce steam demand, increasing the condensate return from steam systems, and improved housekeeping. Good housekeeping practices include analyzing and measuring water use and wastage, reducing water wastage, regular cleaning operations, and equipment maintenance. 2. Reuse, recycling, and regeneration: These options enable the reuse of wastewater between water-consuming operations. Of course, the presence of pollutants in wastewater streams
Mass Integration must be considered so that subsequent water-using processes will not be adversely affected. This aspect is discussed in more detail in Section 5.2.4. Several particular methodologies for water minimization are listed in the chapter by Klemeš and Perry (2007b): • Water Pinch Analysis techniques • Mathematical optimization techniques • Efficient management and control of process operations • Integrating optimization and production planning techniques in conjunction with real-time plant measurements to control for product quality and minimize losses
5.2.4 Minimizing Water Usage and Wastewater Overview of the Measures The task of minimizing water usage and wastewater discharge has received considerable attention during the last few years as water has become more costly and an environmentally strategic concern. Smith (2005) summed up the measures applied to minimizing water usage and wastewater as follows: 1. Process changes: These include all the measures described under item 1 of the listing in Section 5.2.3. Water quality can also be improved by reducing the use of certain processing components, such as hazardous cleaning agents, chemicals, and additives. Additional process changes may be driven through inspection or through process optimization via developed technologies—for instance, Process Integration (Pinch Technology). Reduced consumption and increased efficiency may be achieved either by upgrading equipment or adopting new technologies. 2. Reuse: This is a viable strategy when wastewater from a given operation is used directly in other operations, provided that the pollutants in the reused water do not disturb the processes in the downstream operations. Methods for maximizing water reuse are detailed in Section 5.3, along with a discussion on the use of recycled water in food processing. 3. Regeneration reuse: This is the process of purifying wastewater from one operation and then reusing it in another operation or process. 4. Regeneration recycling: Here the contaminants in the wastewater are only partly eliminated before the water is returned for use in the same process.
111
112
Chapter Five The paper by Blomquist and Brown (2004) offers a useful review of wastewater minimization. The authors examined a large number of preassessment and assessment techniques for respectively identifying waste minimization focus areas (opportunities) and options (solutions) during a waste minimization audit. Blomquist and Brown critically reviewed these techniques and assessed their relative merits. The preassessment techniques were analyzed in terms of their ease and speed of implementation; the assessment techniques were evaluated in terms of their usefulness and applicability.
Wastewater Treatment Methodologies for wastewater handling can be subdivided into different stages of treatment, as follows: • Pretreatment: Mechanical separation of coarse particles (e.g., sticks, plastics). • Primary treatment: Removal of suspended solids by physical or physical-chemical treatment. This process may consist of natural sedimentation or may be assisted via adding coagulants and/or flocculants or via centrifugation. Primary treatment also includes neutralization, stripping (e.g., elimination of ammonia, NH3), and the removal of oils and grease by flotation. • Secondary treatment: The removal of colloids and similar matters from the wastewater. This treatment, which may include chemical and biological processes, minimizes the wastewater’s organic load. Processes commonly used include activated sludge treatment and anaerobic digestion, both of which lead to the critical removal of phosphate, ammonia, and oxygen-depleting contaminants. • Tertiary treatment: This stage comprises physical and chemical processes that eliminate such pollutants as phosphate, ammonia, minerals, heavy metals, and organic compounds. The processes are viewed as a “polishing phase” and are usually more expensive than conventional techniques. The necessity of applying this type of treatment is largely dictated by the following two factors: (i)
Meeting discharge conditions established by environmental quality standards (EQS), which may be stricter than BAT requirements. Subject contaminants include ammonia, so-called List I and List II (BAT-CENTRE, 2009) substances, and suspended solids.
(ii) Recycling the wastewater for further use in the factory as either process water or washing water.
Mass Integration Tertiary treatment is especially important in environmentally sensitive areas where effluent has to have low concentration levels and loads of nitrogen and phosphorus.
5.3 Introduction to Water Pinch Analysis Pinch Analysis was first developed for Heat Exchanger Network synthesis and subsequently extended to yield other energy integration applications (El-Halwagi, 1997; Klemeš et al., 1997; Smith, 2005). The analogous characteristics of heat and mass transfer allowed the application of Pinch Analysis to the synthesis of mass exchange networks and a series of other mass integration problems (El-Halwagi, 1997). Water Pinch Analysis emerged as a special case of mass integration following the seminal work of Wang and Smith (1994). However, that paper’s targeting technique was limited to the fixed load problem, where water-using processes are modeled as mass transfer operations. Later work on Water Pinch Analysis has focused mainly on the fixed flow-rate problem, where flow-rate requirements are viewed as the important constraints for water-using processes (Dhole et al., 1996; Hallale, 2002; El-Halwagi, Gabriel, and Harell, 2003; Manan, Foo, and Tan, 2004; Prakash and Shenoy, 2005). In the context of Water Pinch Analysis, reuse means that the effluent from one unit is used in another unit and does not reenter the unit where it was previously used; in contrast, recycling means that the effluent will reenter the unit where it was previously used, usually after certain purification. In addition, one may also use a regeneration unit (e.g., filter, stripper) to partially purify the water stream prior to reuse or recycling (Wang and Smith, 1994). A typical Pinch Analysis study proceeds in two stages. The first is targeting, whereby minimum freshwater and wastewater flow rates are set; this is followed by network design to achieve the targeted flow rates. It is worth emphasizing that the targeting step is the primary focus in Water Pinch Analysis. The target is needed in order to determine how well a reuse or recycle system can actually perform in terms of thermodynamic constraints. Establishing targets in advance of design provides a clear picture of the mass exchange limitations of the design problem, indicating the smallest achievable freshwater intake and wastewater discharge. Once the targets are established, a water network can be designed using any network design tools. Wang and Smith (1994) described a methodology for determining the amount of water required by a set of operations when water is reused. They showed that significant water savings can be achieved compared with the case when only freshwater is used. The authors employ a simple example that makes use of the limiting Composite Curve (CC) and incorporates four water-using operations. The problem data is presented in Table 5.2.
113
114
Chapter Five Operation number 1
Contaminant mass flow [kg/h]
Cin [ppm]
Cout [ppm]
FL [t/h]
2
0
100
20
2
5
50
100
100
3
30
50
800
40
4
4
400
800
10
TABLE 5.2
Problem Data for Wang and Smith’s (1994) Water-Using Operations
The table lists the maximum inlet and outlet concentrations of a single contaminant for four operations. The last column gives the limiting water flow rate, which is the flow rate required by the operation if the contaminant mass is taken up by the water between the inlet and outlet concentrations. Note, however, that for an operation whose inlet concentration is greater than zero, using uncontaminated freshwater enables a lower flow rate than the limiting water flow rate for that operation. A straightforward analysis of the problem data, assuming that each operation uses freshwater, reveals that the total (uncontaminated) freshwater required by the operations is 112.5 t/h, with the four operations requiring 20, 50, 37.5, and 5 t/h. However, if water reuse is allowed, then an analysis that makes use of the limiting CC produces a target for the minimum water flow rate of 90 t/h. The limiting CC of the four water-using operations is plotted in Figure 5.1. The water supply line—which satisfies the waterusing operations represented by the limiting CC—has its origin at zero concentration and lies below the curve. The slope of the line is such that it touches the CC at one point, which is termed the Water Pinch. Other water supply lines with the same origin could be drawn, but these would not touch the CC and thus would indicate flow rates greater than the (preferred) minimum. If the water supply line were drawn with a steeper slope to indicate a smaller flow rate, then the line would actually cross the limiting CC and so could be part of an infeasible design. Wang and Smith (1994) provided a methodology for calculating the minimum flow rate of water (including reuse) required to remove contaminants from water-using operations. In addition, this paper provided a methodology for designing a water reuse system. Figure 5.2 displays the final system design for the water operations described in Table 5.2. The figure shows that, of the original targeted freshwater amount of 90 t/h, 20 t/h is fed to operation 1 and 50 t/h is fed to operation 2. The remaining 20 t/h is fed to operation 3 along with
Mass Integration C [ppm] 800
MINIMUM WATER SUPPLY FLOWRATE 90 t/h 400
MINIMIZE FLOWRATE
PINCH
100
0
FIGURE 5.1
1
9
21
41
m [kg/h]
Limiting Composite Curve and the Water Pinch.
20 t/h 20 t/h
Operation 1
40 t/h
Operation 3
Feedwater 90 t/h
Wastewater 50 t/h
90 t/h
5.7 t/h Operation 2
Operation 4 44.3 t/h
FIGURE 5.2
Water treatment system designed using Water Pinch methodology.
20 t/h from operation 1. Of the original 50 t/h fed to operation 2, 5.7 t/h is fed to operation 4 and the remaining 44.3 t/h goes directly to wastewater. The authors acknowledge that this design could be evolved further to produce alternative networks. To set the flow-rate targets for water reuse or recycling, various graphical and tabulated targeting techniques may be employed. In addition to the limiting CCs (Wang and Smith, 1994) mentioned previously, the following methods have been used: the water surplus diagram (Hallale, 2002), the material recovery Pinch diagram (El-Halwagi, Gabriel, and Harell, 2003; Prakash and Shenoy, 2005),
115
116
Chapter Five the cascade analysis technique (Manan, Foo, and Tan, 2004), and the source CC (Bandyopadhyay, Ghanekar, and Pillai, 2006). Once the flow-rate targets have been identified, numerous techniques can be used to design a water network that achieves those targets. The works just cited were developed for continuous processes, but there have been several reported efforts to apply Water Pinch Analysis to batch processes; these include the works of Wang and Smith (1995), Liu, Yuan, and Luo (2007), Foo et al. (2006), and Majozi, Brouckaert, and Buckley (2006).
5.4 Flow-Rate Targeting with the Material Recovery Pinch Diagram This section illustrates the targeting technique of Material Recovery Pinch diagram—MRPD (El-Halwagi, Gabriel, and Harell, 2003; Prakash and Shenoy, 2005). Constructing an MRPD requires knowledge about the material flow rates and loads of each process sink and source. Given this information, one may construct an MRPD as follows: 1. Arrange the individual water sources (SRi) and demands (SKj) into two lists, in ascending order of concentration level (C). 2. For each source and demand, calculate the load given by the product of its flow rate and concentration level (i.e., F × C). 3. Plot the cumulative sources and demands, on a diagram of load versus cumulative flow rate, in ascending order of their concentration levels to form the sink and source CCs. In order to render the problem feasible, the cumulative water source CC must lie below the cumulative water demand CC, ensuring that the water purity requirements are satisfied. 4. For pure fresh resources (zero concentration of impurities), the sink and source CCs are separated horizontally until they barely touch—with the source composite lying below and to the right of the sink composite—as shown in Figure 5.3(a). 5. For impure fresh resources, the source CC is shifted along an impure fresh locus until it lies below and to the right of the sink CC; this is shown in Figure 5.3(b). The overlap area of the sink and source CCs represents the maximum recovery among all sink and source streams within the network. The point where the two composites touch is called the Material Recovery Pinch, which is the bottleneck for maximum recovery. The segment where the sink CC extends to the left of the
Mass Integration Load
Load
Sink composite Sink composite
Pinch point
Pinch point
Source composite Source composite
Minimum fresh source
Maximum recovery
Impure fresh locus
Flowrate Minimum waste discharge
Maximum recovery
Minimum fresh source
(a)
FIGURE 5.3
Flowrate Minimum waste discharge
(b)
MRPD for (a) pure fresh resource and (b) impure fresh resource.
source CC represents the minimum feed needed for fresh resources (to be purchased); the region where the source CC extends to the right of the sink CC represents the minimum waste discharge from the network (for final treatment before releasing into the environment). Both the minimum fresh resources needed and the minimum waste generated by the network are network resource targets, and they are determined before the recovery network is designed. In the next section, the MRPD is used to establish flow-rate targets for a case study in the production of fruit juice.
5.5 MRPD Applied to Fruit Juice Case Study Table 5.3 shows the limiting water data for a case study involving the production of fruit juice (Almató, Espuña, and Puigjaner, 1999; Li and Chang, 2006). The water amounts are expressed in cubic meters (m3) per each batch. There is a start time (tS ) and an end time (tT ) for each water sink and source. Prior to water recovery, freshwater SKj
FSKj [m3]
Cj [ppm]
tjS [h]
tjT [h]
SRi
FSKi [m3]
Ci [ppm]
tiS [h]
tiT [h]
SK1
20
0
0.5
2.5
SR1
20
5
2.5
4.5
SK2
20
6
5.0
7.0
SR2
20
14
7.0
9.0
SK3
20
15
9.5
11.5
SR3
20
20
11.5
13.5
SK4
16
5
17.0
19.0
SR4
8
25
17.0
19.0
SK5
20
7
6.0
8.0
SR5
16
10
10.5
14.5
TABLE 5.3
Fruit Juice Production: Limiting Water Data
117
118
Chapter Five and wastewater amounts are calculated to be 96 m3 and 84 m3, respectively (these values are given by the sum of the individual water flows). For this case study, three assumptions are made in order to simplify the analysis: 1. The batch process is operated repeatedly on a yearly basis. Therefore, the process behaves as if it is operated in continuous mode. It has been shown elsewhere (Foo, Manan, and Tan, 2005) that a repeated batch operation achieves the same flow targets as for equivalent continuous operation. 2. An unlimited water storage tank is always available. Thus, water can be stored for later use. 3. Water recovery is always carried out between two consecutive batches. In other words, water sources in an earlier batch will be sent to water storage tanks before being reused or recycled to the water sinks of the next batch operation. A similar assumption was made in the paper by Shoaib and colleagues (2008). Given these assumptions, any established targeting technique for continuous processes can be used to identify rigorous water targets for the case study. For purposes of this example, the MRPD is used. The MRPD is illustrated in Figure 5.4(a), and the network that achieves the MRPD-derived water targets is shown in Figure 5.4(b). As shown in the figure, the minimum freshwater demand (FFW) is determined to be 35 m3; the wastewater flow (FWW), 23 m3. When compared with the total water flow prior to water recovery, this represents a significant reduction of 63.5 and 72.6 percent for freshwater and wastewater, respectively.
5.6 Water Minimization via Mathematical Optimization 5.6.1
Introduction to Mathematical Optimization
Besides Water Pinch Analysis, water minimization problems have also been solved using mathematical optimization techniques. Various mathematical optimization approaches have been developed to complement Water Pinch Analysis in dealing with more complex problems—for example, multicontaminant systems (Takama et al., 1980; Alva-Argáez, Kokossis, and Smith, 1998; Huang et al., 1999); complex operational constraints, which include limiting the number of pipeline connections (Hul et al., 2007); and forbidden/compulsory matches between water-using processes (Bagajewicz and Savelski, 2001; Kim and Smith, 2004; Li and Chang, 2006). New research results
Mass Integration Fww=23m3
(a) [kg] 1200
SR4
1000 800
SR3
600
SK3
400
SR2
SK5
200
SK2 SK4
SK1
0
20 FFW
40
SK1
80
120
MRPD 20
SR1
100 [m3]
=35m3
(b)
SR5 SR1 60
16
ST1 4
10 SK14
20
6
SK1 5.7 0.7
FFw 35m3
SR4
SR1
ST2 8
5
14.3 SK1
15
SR1
Fww 23m3
6 ST3 8.3 SK5
0
SR5
5
10
16
15
20
t [h]
All water amounts are in m3
Network design
FIGURE 5.4
Fruit juice production: (a) MRPD; (b) network design.
and case studies have recently been published by authors from South Africa (Gouws and Majozi, 2008a; Gouws and Majozi, 2008b), Asia (Chen, Chang, and Lee, 2008; Ng et al., 2008; Chen et al., 2009), and Europe (Tokos and Novak Pintarič, 2009).
119
120
Chapter Five
5.6.2 Illustrative Example: A Brewery Plant This section discusses the case study of a brewery plant (Tokos and Novak Pintarič, 2009) as a means to illustrate how mathematical optimization is used to solve water minimization problems. In the brewery studied, the ratio (by volume) of process water consumed to product beer sold was 6.04 : 1; this translated into 653,300 m3/y of water consumed. In terms of the ratio set by BREF (2006), this freshwater consumption exceeded the upper limit by 144,900 m3. In light of these figures, the company undertook to improve its process by retrofitting modifications to its existing water network so that the plant’s usage of freshwater would be minimized. Production at the brewery plant involves a mixture of waterusing batch and semicontinuous processes. Water-using operations in the packaging area are operated mainly in batch mode, with the exception of rinsers for nonreturnable glass bottles and cans. Wastewater stream from semicontinuous processes can be reused in batch processes with a lower purity requirement. Hence, the basic formulation first proposed is designed to enable the efficient integration of semicontinuous and batch water-using processes. The continuous wastewater streams are treated as limited freshwater sources, and the unused wastewater is discharged. In the next step, the model is extended by including options for installing intermediate storage tanks for the collection of unused wastewater streams for reuse over subsequent time intervals. This particular design modification is motivated by differences in the operating schedules of the filling lines. The superstructure representation for water reuse and regeneration reuse (as defined in Section 5.2.4) is depicted in Figure 5.5. Opportunities for regeneration reuse were analyzed in the brewhouse and in the cellar (see Figure 5.6), since these processes were characterized by a high concentration of contaminants. Here the basic model is extended by installing a local (on-site) wastewater treatment unit that can operate in either batch or continuous mode, thereby enabling water regeneration reuse and recycling. The scheduling of batch wastewater treatment units is performed simultaneously so that the treatment schedule will coincide with the fixed schedule of the batch process. The design includes the option to install storage tanks before and after treatment; this enables wastewater and/or purified water to be stored until required by the treatment schedule. As reported in Tokos and Novak Pintarič (2009), the integration of the water network in the packaging area made it possible for wastewater from the can rinser to be reused in the pasteurization processes. In this way, freshwater consumption could be reduced by 23 percent and the common costs of freshwater and wastewater treatment by 22 percent. These improvements do not require the
Mass Integration Local treatment unit
Purified water Freshwater Purified water
Freshwater
Wastewater Storage n
Operation n
Storage ww
Wastewater from continuous operation
Wastewater from central storage tank for continuous operation
FIGURE 5.5 Superstructure for water reuse and regeneration reuse in a brewery plant (Tokos and Novak Pintarič, 2009).
Water for pouring
Filtration
TR
Cleaning in-place system
Cleaning in-place system
Water for pouring
Wort boiling
Fresh water for batch processes Wastewater
Cellar
FIGURE 5.6
Brewhouse
Water reuse opportunities in a brewery plant.
addition of any storage tanks. The net present value of the proposed water network reconstruction is positive (at a 15 percent discount rate), and the payback period is 0.29 years (about 15 weeks). In the brewhouse and the cellar, the continuous water treatment unit (nanofiltration) was selected for wastewater purification in the optimum water network. Purification allows the water from batch material pouring to be reused in the clean-in-place (CIP) system, and wastewater from filtration could be reused directly for pouring the batch material. All told, freshwater consumption could be reduced
121
122
Chapter Five by 28 percent and the joint cost of freshwater and wastewater treatment by 27.9 percent. Investment costs for the modifications, which require a membrane area of 83 m2, amount to €117,205. The net present value of the optimal water network is positive at a 15 percent discount rate, and the payback period is 1.3 years. The price of freshwater has a significant impact on the optimal water network. Increasing freshwater prices result in the identification of additional opportunities for reuse and regeneration reuse. Examples include water reuse between the rinser for nonreturnable bottles and the bottle washer, water reuse between the pasteurizer and the bottle washer, and regeneration reuse between wort boiling and the CIP system. Complete implementation of the proposed design could allow the brewery to reduce its current freshwater demand by about 25 percent and to reduce its costs for freshwater and for wastewater treatment by about 27 percent. Furthermore, the brewery’s ratio of water consumed to beer sold would decrease to 4.53 : 1 (from 6.04 : 1), which is important for cleaner production and sustainable development within the company.
5.7 Summary Water is used in most process industries for a wide range of applications. Today, industrial processes and systems that use water are subject to increasingly stringent environmental regulations concerning the discharge of effluents. Moreover, the demand for fresh water continues to increase. The pace of these trends has increased the need for improved water management and wastewater minimization. Adopting water minimization techniques can effectively reduce overall freshwater demand in water-using processes and also reduce the amount of effluent generated. These reductions bring reductions also in costs incurred to acquire freshwater and treat effluents. The field has been developing rapidly, and every year brings a number of new and more efficient approaches. This chapter has reviewed and demonstrated, through selected case studies, current methodologies that have been applied to minimize water use and wastewater in the processing industry.
CHAPTER
6
Further Applications of Process Integration
P
rocess Integration, also called Pinch Technology, was initially developed for energy and specifically for Heat Integration. Details of the origin and development of Heat Integration were given in Chapters 2 and 4. Its further development resulted in a methodology for integrating mass transfer and water integration in particular; this technology was described in Chapter 5. This chapter (chapter 6) focuses on the additional applications and especially recent developments that have expanded the generic Process Integration ideas in various other directions. Given the rapid development of this methodology, it is not possible to cover all recent achievements. Nonetheless, this chapter explores several interesting directions that have considerable potential for future development.
6.1 Design and Management of Hydrogen Networks The evolution of Pinch Technology has allowed mass integration to be extended to hydrogen management systems. In one of the earliest works in this field, Alves (1999) proposed a Pinch approach to targeting the minimum hydrogen utility. This method was based on an analogy with process heat recovery. Just as the distribution of energy resources in a plant can be analyzed and designed via using Pinch Technology, so can the distribution of hydrogen resources be handled in refineries, which typically have several potential sources (each capable of producing a different amount of hydrogen) and several hydrogen sinks (with varying requirements). However, the designer has more flexibility in determining the hydrogen loads of individual units by varying the throughput of units and operating many processes over a range of conditions. As a result, there is considerable potential for optimizing refinery performance. A liquid hydrocarbon feed stream is mixed with hydrogen-rich gas, heated, and then fed to a reactor. Part of the hydrogen is consumed by reaction with the feed. Light hydrocarbon compounds (methane, ethane, and propane), hydrogen sulfide (H2S), and
123
Chapter Six ammonia are usually formed as part of the reaction products. The effluent from the reactor is then cooled down and sent to a highpressure flash separator. The gas released in the separator is often treated in an amine scrubber that removes H2S. Part of the gas is vented from the process through a high-pressure purge to prevent any buildup of hydrocarbons in the recycle. The remaining hydrogenrich gas is recompressed and then returned to the reactor with a fresh hydrogen makeup stream. The liquid stream removed from the bottom of the high-pressure separator contains some hydrogen, light hydrocarbon gases and H2S in the solution, which is lost from the hydrogen system. This liquid stream is sent to a low-pressure separator, from which off-gases are taken and typically sent to a flare or to the fuel gas system. A two-dimensional plot of total gas flow rate versus purity represents the mass balance of each sink and source in the hydrogen network. A plot that combines the profiles for hydrogen demand (dashed line) and hydrogen supply (solid line) yields the hydrogen Composite Curve (CC) (Figure 6.1). The sink and source profiles start at zero flow rate and proceed to higher flow rates with decreasing purity. The circled “plus” signs in the figure indicate the surplus— where sources provide more hydrogen than is required by sinks. Where the sources do not provide enough hydrogen to the sinks, a circled “minus” sign appears on CCs to indicate a deficit of supply. The area beneath the entire Sink Curve is the flow rate of pure hydrogen that the system should provide to all the sinks. The area beneath the Source Curve is the total amount of pure hydrogen available from the sources. For the hydrogen network to be feasible, there should be no hydrogen deficit anywhere in the network; otherwise, the sources will not be able to provide enough hydrogen to the sinks. The hydrogen utility can be reduced by horizontally moving the curve toward the vertical (purity) axis until the vertical segment between the purities of the sink and the source touches the vertical axis, Source Composite Curve 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
(b)
– Purity [–]
(a)
Purity [–]
124
Sink Composite Curve
0
1
2
3
4
5
6
7
Gas flowrate [106 standard m3/d]
8
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
–
0
0.5
1
Hydrogen surplus [106 standard m3/d]
FIGURE 6.1 Composite Curves and hydrogen surplus diagram (after Alves, 1999).
Further Applications of Process Integration thereby forming the Hydrogen Pinch. Separating the hydrogen source and sink parts then determines the target value for the hydrogen utility minimum flow rate. The procedure for calculating the supply target requires varying the flow rate of gas supplied to the system until a Hydrogen Pinch is found. The sources from hydrogen-consuming processes or from processes generating hydrogen as a secondary product (dehydrogenation plants) have flow rates that are determined by normal process operation; these rates are assumed to be fixed for the purposes of designing a hydrogen network. However, process hydrogen sources with variable flow rates can be regarded as imports from external suppliers and from processes (i.e., steam reformers or partial oxidation units) that generate hydrogen as a main product. Those sources are hydrogen utilities. One approach to minimizing hydrogen utility consumption is to increase the purity of one or more sources. A hydrogen purification system introduces an additional sink (feedstock for purification) and two sources (purified stream and residue stream), resulting in new targets. By employing Hydrogen Pinch Analysis, an engineer can make the best use of hydrogen resources in order to meet new demands and improve profitability.
6.2 Oxygen Pinch Analysis Another extension of Process Integration is Oxygen Pinch Analysis (Zhelev and Ntlhakana, 1999). The idea is to analyze the problem so that targets are derived prior to designing a system for minimizing oxygen consumption of the micro-organisms used for waste degradation. The next step is to design a flowsheet that achieves the targets. In most cases, oxygen is supplied through agitation. Aeration requires energy, so an analysis based on the Oxygen Pinch eventually leads back to the original application of energy conservation. Using the chemical oxygen demand or COD (Monod, 1949) as the baseline range for organic contaminants allows one to set quantitative targets (for oxygen solubility, residence time, and oxidation energy load) as well as additional qualitative targets— namely, the growth rate that is a direct indicator of the age and health of micro-organisms (Zhelev and Bhaw, 2000). Analyzing the information in Figure 6.2 and then matching the oxygen supply line to the CC (so they touch at the Pinch point) yields targeting information on the growth rate of micro-organisms, oxygen solubility, residence time, and oxidation energy load. In the Oxygen Pinch approach, the method recurs to energy but also incorporates extra information concerning environmental issues. An important contribution of this method is its ability to target—in parallel with the concentration of oxygen and the total
125
Chapter Six Dissolved O2 stream
Pinch
COD
COD
Process stream
I/COD
126
Dissolved O2 stream D μm-Specific growth S-Saturation D-Diluation
FIGURE 6.2
Process stream
Substrate supply line Slope = μm/S Intercept = I/S
I/D I/D Slope ~ growth rate, O2 solubility, residence time, oxidation energy load
Oxygen Pinch method (after Zhelev, 2007).
energy required—a quality characteristic: the micro-organisms’ health as assessed by their rate of reproduction.
6.3 Combined Analyses, I: Energy-Water, Oxygen-Water, and Pinch-Emergy 6.3.1 Simultaneous Minimization of Energy and Water Use Water savings can be achieved through the strategic implementation of water reuse between water-using operations. Further minimization of freshwater usage is possible by regenerating water, which is then recycled. The design methodology developed by Smith and colleagues (Wang and Smith, 1994; Kuo and Smith, 1997; Gunaratnam et al., 2005) has proven to be effective in process industries because it provides a systematic means of establishing realistic minimum water requirements for a site as well as conceptual design guidelines for de-bottlenecking water systems. In some process industry sectors (e.g., the food industry), water use is closely linked to energy systems. The diagram shown in Figure 6.3 explains the basic concept and the importance of considering water and energy systems concurrently. In Figure 6.3(a), freshwater is supplied to two water-using operations and then is discharged in a parallel arrangement. The necessary heating or cooling (usually through a heat exchanger) is provided according to process requirements. This conventional practice can be significantly improved by implementing a design for simultaneous water reuse and heat recovery, as shown in Figure 6.3(b). Water reuse between operations reduces water consumption, and the proposed heat recovery between streams will reduce the need for utilities (e.g., steam, cooling water). When the problem is considered jointly, finding the best energy recovery options and water reuse schemes is an extremely complex task because there are strong design interactions between systems
Further Applications of Process Integration (a) Operation 1 Wastewater
Freshwater
Operation 2 Conventional practice: non-integrated (b) Freshwater
Operation 1
Heat recovery
Water reuse
Wastewater
Operation 2 Improved design with water reuse and heat recovery
Legend
Utility Heater
Utility Cooler
FIGURE 6.3 Simultaneous energy and water minimization (after Savulescu, Kim, and Smith, 2005a).
for water and energy. Both the Water Pinch and the Energy Pinch concepts have been accommodated in separate design frameworks. However, the methodological procedure is changed when the interactions between water reuse and energy recovery must be considered; see Savulescu, Kim, and Smith (2005a, 2005b). Further interesting applications have been published (Leewongtanawit and Kim, 2009; Manan, Tea, and Alwi, 2009). The energy-water methodology of Savulescu and Kim (2008) follows a two-step approach: targeting and design. During the targeting phase, theoretical minimum requirements for freshwater and thermal utilities (hot and cold) are obtained via graphical manipulation of streams data (i.e., water flow rate, contaminant levels, and temperature). The purpose of the design phase is to create a water and heat recovery network that can achieve the established target. A useful design tool is the two-dimensional grid diagram (Figure 6.4), which exploits the network arrangement of water streams subject to energy recovery constraints (Savulescu, 1999; Leewontanawit, 2005). An industrial case study conducted recently (Leewongtanawit and Kim, 2008) showed an 18 percent reduction in annualized cost resulting from the integrated approach (when
127
Chapter Six
Freshwater main
Temperature
128
Process water main
Process water users
Wastewater main
Process water users
Overflow
Water quality - Water source - Water demand/sink
FIGURE 6.4 Two-dimensional diagram of a water quality network (after Savulescu and Kim, 2008).
compared with operations when only water minimization is performed). This case study clearly demonstrated the necessity of a holistic approach to designing water and energy systems, since there are significant benefits to simultaneous integration.
6.3.2 Oxygen-Water Pinch Analysis The link between Water Pinch Analysis and Oxygen Pinch Analysis is the use of COD as the concentration variable in Water Pinch Analysis. Here two configurations of wastewater treatment are investigated, centralized and distributed. The analyzed system includes the centralized biological treatment unit and several satellite factories (sites) in the surrounding area that send some portion of their wastes to the centralized treatment unit. The variables of interest when analyzing wastewater treatment are the quantity of wastewater treated and the quantity of oxygen required. The quantity of treated wastewater gives an indication of how much freshwater is used and thus of the water management level. The quantity of required oxygen gives an indication not only of the wastewater quality but also of the energy required by the wastewater treatment process. Both the quantity and the quality of the wastewater treated are related to the cost of the treatment process.
Further Applications of Process Integration The configuration of wastewater treatment serves as an aid in establishing whether or not wastewater treatment costs that are based on quality and quantity differ significantly from costs based on quantity only. The effluent wastewater conditions obtained via the Water Pinch method are used to plot concentration versus COD flow rate on one set of axes. The oxygen required is then calculated using Zhelev’s (1998) method of the limiting oxygen supply line, which is constructed as shown in Figure 6.5. First, the CC must be constructed. This is done by plotting all the site streams on the same set of axes. These streams are then summed within each concentration interval, and the resulting curve is the limiting CC. The limiting oxygen supply line is then constructed as a line drawn between the origin and the Pinch Point. The inverse of the gradient of the limiting oxygen supply line is the flow rate of oxygen required.
Concentration vs Flowrate Curve for Site 1 12000
(b) 10000 Concentration [mg COD/I]
Concentration [mg COD/I]
(a)
8000 4000 0 4000 8000 0 Mass Flowrate [kg COD/d] (c) Concentration [mg COD/I]
10000
Concentration vs Flowrate Curve for Site 2
5000 0 0
5000
10000
15000
Mass Flowrate [kg COD/d]
Concentration vs Flowrate Curve for Site 3
5000
0 0
4000
8000
Mass Flowrate [kg COD/d]
Concentration [mg COD/I]
(d)
Composite Curve and Minimum Oxygen Supply Line 12000 Site 3
8000
Composite Curve
4000 Site 2
Minimum O2 Supply
Site 1
0 0
5000 10000 15000 20000 25000 30000 Mass Flowrate [kg COD/d]
FIGURE 6.5 Construction of the oxygen limiting supply line (after Zhelev and Bhaw, 2000).
129
130
Chapter Six
6.3.3
Emergy-Pinch Analysis
The concept of emergy (embodied energy) was first developed by Odum in the late 1980s (Odum, 1996). Along with other definitions referring to life cycle, it may be defined in terms of solar transformity (Brown and Ulgiati, 2004). Solar emergy is the solar energy directly or indirectly necessary to obtain a flux of energy in a process. The unit of emergy is the solar emergy Joule (seJ), an extensive quantity, which denotes the available energy of a certain type (heat, electrical, etc.) that undergoes transformations. Transformity, an intensive quantity, is defined as the emergy input per unit of exergy (available energy) output: seJ/J. The first step in the practical process of emergy analysis is collecting information for the calculation of solar transformities ST [seJ/unit] of the chain of activities involved in making a resource available to the process. This is the most difficult part of the methodology because transformity databases—although rapidly growing and continuously updated by researchers—are not comprehensive. The second step is the calculation of solar emergy SE [seJ/y] followed by calculation of the solar emergy investment SEI [seJ/g]: SE [seJ/y] = ST [seJ/unit] · Amount [units/y]
(6.1)
SEI [seJ/g] = SE [seJ/y]/Amount [g/y]
(6.2)
The combined Pinch–emergy analysis is used in the preliminary, conceptual design stage. The Emergy Composite Curve (ECC) is analogous to the Pinch CCs. In the ECC, solar transformity is plotted against solar emergy; the CC is then matched up with the total emergy investment (TEI) supply line, which is restricted by the ECC at the Pinch point. Analyses are based on ECC benefit from using both emergy and Pinch features. Each stream in the ECC carries three pieces of information: 1. Transformity: the past emergy investment or “history” of the stream 2. The “market” potential of the stream in terms of usability: the heat (temperature) potential of a thermal stream or the concentration limits of a water stream 3. The stream’s future in terms of further usability (regenerative reuse) In the case of Heat Pinch Analysis, the hot and cold streams will have different signs for this component of final emergy investment. The sign on the required emergy investment to heat the cold streams will be the opposite of the one for available emergy. Hence, at this level of
FIGURE 6.6 Composite Curve in emergy-transformity coordinates (after Zhelev and Ridolfi, 2006).
ST [seJ/J or $]
Further Applications of Process Integration
Pinch
Composite Curve Total Energy supply line
SE [seJ]
analysis, it is possible to relax certain constraints (e.g., ΔTmin) that can lead to minimization of the usage of expensive hot utilities. With Emergy-Pinch analysis, as with classical Pinch Analysis, the processes “overlap” on the vertical axis (temperature range, concentration range) or, as here, the transformities range. The emergy loads (investments) for the different processes are characterized by relative values, which allows their graphical representation to be freely shifted left and right in the ST/SE plot; see Figure 6.6. The TEI is targeted by drawing the line touching the CC and then calculating its slope. The greater the slope of the TEI line, the smaller the rate of TEI. This minimizes the supply of combined resources and their corresponding costs while lifting the emergy supply line to its maximum. This limit is represented by the point where the supply line and the CC meet—that is, the Pinch point. The slope and the Pinch point of the emergy supply line can be used to help compare alternative design or operational options. Transformity is viewed as a “quality” parameter; when plotted against emergy investment, it allows targeting of TEI and determination of the maximum total transformity needed to run a given process.
6.4 Combined Analysis, II: Budget-Income-Time, Materials Reuse-Recycling, Supply Chains, and CO2 Emissions Targeting 6.4.1 Budget-Income-Time Pinch Analysis There are substantial benefits to be derived from applying the process design concept to financial management. The timing, extent, and allocation of Process Integration for minimizing the financial risk is
131
Chapter Six
Time [d]
Time [d]
Income [$] Time [d]
Project budget and income versus time (after Zhelev, 2007).
FIGURE 6.8 Composite Curves for project budget and income (after Zhelev, 2007).
Budget [$/ 5 y]
FIGURE 6.7
Budget [$]
Income [$]
the primary goal of such investigations, which also account for possible uncertainties in model parameters. The concept of combined resources management can lead to more realistic design solutions while helping decision makers account for financial investments. Because time runs in only one direction, the direction of individual vectors and both CCs is to the right; this is illustrated in Figure 6.7 and Figure 6.8. Zhelev (2005a, 2005b, 2005c) reports on two aspects of this broad area: using Pinch principles to choose alternative designs, and amalgamating financial considerations with the management of energy and water. Several different stages can be identified in the processes of investment, design, commissioning, and operation. By applying traditional targeting procedures to the management of financial resources, the following data can be obtained prior to design: maximum investment level, minimum payback period, and maximum benefit. As shown in the upper part of Figure 6.9, this targeting is analogous to other Pinch applications, such as the Water Pinch. First a CC is constructed, after which a capital (investment) supply line is drawn against the CC. The steeper the investment supply line, the shorter the payback period. The steepest slope is constrained by the CC, which meets the supply line at the Pinch point. Lifting the
Budget [$]
132
Time [y]
Further Applications of Process Integration capital supply line up to the maximum allows one to target both the investment level and the expected annual benefit (Figure 6.9).
6.4.2 Materials Reuse-Recycle and Property Pinch Analysis The composition of a substance is only one of several chemical and physical properties that are essential in a chemical process. Other common properties include acidity and alkalinity (as measured by pH), density, viscosity, reflectivity, turbidity, color, and solubility. The process network synthesis associated with these chemical properties cannot be addressed by conventional mass integration techniques, so another generic approach has been developed to deal with this problem (Shelley and El-Halwagi, 2000; El-Halwagi et al., 2004). For systems that are characterized by one key property, Kazantzi and El-Halwagi (2005) introduced a Pinch-based graphical targeting technique that establishes rigorous targets for minimum usage of fresh materials, maximum recycling, and minimum waste discharge. Foo and colleagues (2006) focused on developing an algebraic technique to solve the problem of identifying rigorous targets for property-based recycling and reuse of materials. A key element of these techniques is the concept of material surplus, which generalizes the analogous concept developed for tasks of synthesizing hydrogen and water networks (Alves and Towler, 2002; Hallale, 2002). Foo et al. (2006) developed an algebraic approach called property cascade analysis (PCA) to identify various performance targets for a maximum resource recovery (MRR) network. This paper also introduced network design techniques for the synthesis of an MRR network as
Investment [$]
20 15 10
20 40 Separate stage of design Composite Curve Capital (investment) Supply Line
60
80
1 Grid
2 3 4 F
FIGURE 6.9
Targeting and project management (after Zhelev, 2007).
Savings [$]
133
134
Chapter Six well as a systematic procedure for identifying optimum process modification strategies. The problem of designing a property-based material reuse network is formulated as follows (see Figure 6.10): A process is described as having a number NSK of process sinks (units) and a number NSR of process sources (e.g., process and/or waste streams) that can be considered for possible reuse and/or to replace the use of fresh material. The aim is to design a network of interconnections among the property sinks and sources such that the overall flow rates of fresh resource and waste discharge are minimized without depriving the sinks of adequate quality resources. Each sink j requires a feed with flow rate Fj as well as an inlet property pjin that satisfies the following constraints: max p min d p in j j d pj min
for j 1, 2,! , N SK
(6.3)
max
where pj and pj are the specified lower and upper bounds on admissible properties of streams to unit j. Likewise, each source i has a given flow rate Fi and a given property pi. Also available for service is a fresh (external) resource, with property pF, that can be purchased to supplement the use of process sources in sinks. Each process source may be intercepted via design and/or operating changes in order to modify the flow rate and property of what each sink accepts and discharges. The Pinch diagram shown in Figure 6.11 is a convenient tool, developed by Kazantzi and El-Halwagi (2005), that avoids the drawbacks of traditional iterative procedures (Alves and Towler, 2002; Hallale, 2002): low visualization insight for targeting and
Sources
Processed sources (back to process)
Segregated sources
Sinks i=1
j=1
i=2 j=2
Property Interception Network
j=NSR i=NSK
FIGURE 6.10 Graphical formulation of designing a property-based material reuse network (after Foo et al., 2006).
Further Applications of Process Integration Unused Materials or Discharged Waste
U1 + U2 + U3
Source 2
Load [kg/s]
U1 + U2
U1 Source 1
Flowrate [kg/s] Minimum Fresh Usage
FIGURE 6.11 Property-based material reuse Pinch diagram that combines fresh usages to determine minimum fresh consumption (after Kazantzi and ElHalwagi, 2005).
network design. Another graphical targeting tool that can be used to determine minimum resource targets is the Material Surplus Composite Curve (MSCC). The MSCC was developed by Saw et al. (2009) based on hydrogen and water surplus diagrams (Alves and Towler, 2002; Hallale, 2002), but it eliminates the latter’s iterative steps; see Figure 6.12. The drawbacks of the graphical approach can be resolved by using an equivalent numerical tool, the PCA. This technique is discussed in detail by Foo et al. (2006) for a case study on solvent recycling in metal degreasing (Kazantzi and El-Halwagi, 2005). Applying the PCA makes it possible to use the targeted fresh solvent flow rate to construct a “balanced” material sink and source composite diagram. Foo et al. (2006) also suggested a technique for using the Property Pinch Analysis (graphical or algebraic) to synthesize a property network that achieves previously established resource targets. In addition, their paper discusses applicability of the PCA procedure to process modifications.
135
Chapter Six (b)
Resource Surplus
Flowrate
Purity
(a) Resource Deficit Purity
136
Surplus Composite Curve
Pinch Deficit Composite Curve
Minimum fresh Minimum waste resource discharge Flowrate
FIGURE 6.12 Construction of (a) interval flow rate diagram and (b) the MSCC (after Saw et al., 2009).
6.4.3 Pinch Analysis of Supply Chains The power of Pinch Analysis, which combines quality (e.g., temperature, concentration) with quantity (e.g., heat duty, mass flow), has been successfully applied to analyze supply chains. In this case, (reduced) time is the “quality” and the amount of material (e.g., number of units, mass) is the “quantity.” The objective of the aggregate planning is to satisfy demand in a way that maximizes profit. Demand must be anticipated and forecasted, and production must be planned in advance for that demand. Aggregate planning is particularly beneficial to plants whose products encounter significant fluctuations in demand. Such planning determines the total production level in a plant for a given time period, rather than the quantity of each stockkeeping unit produced. Singhvi and Shenoy (2002) formulated the aggregate planning problem as follows. Given the demand forecast Dt for each period t in a planning horizon that extends over T time periods, maximize the profit over the specified time horizon (t = 1, . . . , T) by determining the optimum levels of the following decision variables: • Production rate Pt = number of units produced in-house in time period t • Overtime Ot = amount of overtime worked in time period t • Subcontracting Ct = number (outsourced) in time period t
of
units
subcontracted
• Workforce Wt = number of workers needed for production in time period t • Machine capacity Mt = number of machines needed for production in time period t
Further Applications of Process Integration • Inventory It = inventory at the end of time period t • Stock out St = number of units stocked out (backlogged) at the end of time period t Figure 6.13 illustrates how material is accumulated at the end of a time period t. The accumulation of material balances can be expressed mathematically as It 1 St 1 Pt Ct
Dt It St
(6.4)
or Previous inventory + Total production = Demand + Current inventory
(6.5)
These equations are reflected in the Supply Chain CCs used for Pinch analysis, as shown in Figure 6.14. Singhvi, Madhavan, and Shenoy (2004) extended the initial methodology to the case of planning for multiple product scenarios. Singhvi (2002) proposed the following algorithm for minimizing inventory cost: 1. List all the products in order of increasing production rates and produce the products in that order. 2. For products that have the same production rate, first produce the one whose inventory holding cost is lower. 3. For products that have the same production rate and the same inventory holding cost, first produce the one for which demand is lower. ACCUMULATOR Current Net Inventory It
−St
Previous Net Inventory Ir−1 IN
−St−1
Time Period t
In-house Pt
Demand Dt
OUT
Subcontract Ct
PRODUCTION
FIGURE 6.13 2002).
Material balance in aggregate planning (after Singhvi and Shenoy,
137
Chapter Six
Ending inventory 6 5 Time [months]
138
Demand Composite Curve
4
Pinch point 3
Ik
dkΔt
Production Composite Curve
2 1 0
It−1 I0
PkΔt + ck 5000
10000
15000
20000
Material quantity [units]
FIGURE 6.14 Supply chain Composite Curves (after Singhvi, Madhavan, and Shenoy, 2004).
6.4.4 Using the Pinch to Target CO2 Emissions Emission targeting via Pinch analysis was investigated in the 1990s by Linnhoff and Dhole (1993), Dhole and Linnhoff (1993b), and Klemeš et al. (1997). The applications, which employ the Total Site concept, address optimization within industrial facilities, not within extended sites as regional or national energy sectors. However, a later work (Perry, Klemeš, and Bulatov, 2008) included the regional dimension in a Total Site Analysis of integrating renewable sources of energy. Tan and Foo (2007) presented a further application of Pinch Analysis to energy-sector planning under carbon emission constraints: the Carbon Emission Pinch Analysis (CEPA). The main problems addressed by the proposed methodology are (1) identifying the minimum quantity of zero-emission energy resources needed to meet the specified energy requirements and emission limits of different sectors or regions in a system and (2) designing an energy allocation scheme that meets the specified emission limits while minimizing use of the energy resources. The sequence of the proposed Pinch Analysis is as follows (Tan and Foo, 2007): • Tabulate the energy source and demand data. The resulting table must contain the quantity of the energy sources (Si) and demands (Dj) and their respective emission factors (Cout,i and Cin,j). • Arrange the energy sources and demands in order of increasing emission factors.
Further Applications of Process Integration • Calculate the emission levels (SiCout,i) and limits (DjCin,j), respectively, of the energy sources and demands. • Plot the Demand CC with the energy quantity (Dj ) as the horizontal axis and the emissions limit (DjCin,j ) as the vertical axis. Hence the slope of the CC at any given point corresponds to the emissions factor (Cin,j ). • Plot the Source Composite Curve in the same manner as with the Demand Composite Curve, but use instead the quantities Si and SiCout,j. In this curve, the slope at any given point corresponds to the emissions factor SiCout,j. • Superimpose the two CCs on the same graph. • Shift the source CC horizontally to the right so that it does not cross the demand CC. In final position, the former should lie diagonally below and to the right of the latter. The two curves must touch each other tangentially without crossing; their point of contact is the Pinch point. • Note the distance from the origin of the graph to the leftmost end of the Source Composite Curve. This distance gives the minimum amount of zero-carbon energy needed to meet the system’s specified emissions limits. • Finding the Pinch point yields valuable insights to decision makers—in particular, it identifies the system bottleneck. The “golden rule” of Pinch Analysis can then be applied to the problem: in order to meet all the specified emission limits for the system, the zero-carbon energy resource is supplied only to those energy demands below the Pinch point. Any allocation of this resource above the Pinch point will either lead to an infeasible solution or require more zero-carbon energy than the minimum quantity established by Pinch Analysis.
6.4.5
Regional Resource Management
Regional Resource Management Composite Curve A novel approach to regional resource management has been developed that tackles simultaneously the two most important issues with biomass supply chains: transportation and land use. The biomass supply chain problem is complex because of the distributed nature of biomass resources and their low energy density, which necessitates large transportation capacities. Growing biomass requires considerable land areas, often leading to competition with food production. To address these problems, a two-level approach to biomass supply chain synthesis—based on a novel Regional Energy Clustering (REC) approach—was proposed by Lam et al. (2009). The
139
140
Chapter Six first level of regional resources management consists of forming clusters of zones for biomass management. The second level involves building the regional energy transfer cascade (RETC) and the Regional Resources Management Composite Curve (RRMCC). The clusters of zones formed in the first level of the methodology are designed to minimize the environmental impact of biomass energy exchanges among the zones within the overall supply chain network (Lam et al., 2009). The carbon footprint (CFP) is used as a criterion for comparing various magnitudes of this impact; cost is another obvious criterion. The main goal of clustering is to partition the area of the considered region into smaller subareas (the clusters) in order to form more coherent entities. Within each cluster, stronger and more efficient interactions and biofuel exchanges result in minimizing the environmental impact of the whole region. Figure 6.15 presents an REC algorithm, whose steps are discussed next: Step 1. Specify energy sources and demands based on available system data. Step 2. Optimize biomass exchange flows between the zones. In this step, linear programming is used to formulate an objective function that minimizes total CFP within the overall region.
FIGURE 6.15 Flowchart of algorithm for Regional Energy Clustering.
Begin
Step 1. Specification of energy sources and demands
Step 2. Optimize biomass exchange flows between zones
Step 3. Display the optimal biomass exchange flows
Step 4. Cluster formation
Output: Clusters and their properties
End
Further Applications of Process Integration Step 3. Display the optimal biomass exchange flows. A visual mapping of interzone biomass exchanges provides critical feedback for the decision maker. The zone “centroids” are plotted in twodimensional Cartesian coordinates. Step 4. Form the clusters. Mixed integer linear programming (MILP) has proven to be a convenient tool for this task.
Regional Energy Surplus-Deficit Curves The formed clusters should be presented visually to help document and explain the proposed solution. For this purpose, the use of Regional Energy Surplus-Deficit Curves (RESDCs) (see Figure 6.16 for an example) is suggested.
Regional Resources Management Composite Curve The RRMCC can be developed based on results obtained from the REC algorithm. In this graphical method, the main idea of Grand Composite Curve has been translated to the problem of regional resource management. Figure 6.17 illustrates two ways of presenting the RRMCC, where panels (a) and (b) employ different directions of cascading. The RRMCC combines information about energy surpluses and deficits as well as land use, allowing one to assess possible trade-offs. The quantity of the energy demand and supply (cumulative energy balance [PJ/y]) is shown on the X axis, and the cumulative zone area [km2] is shown on the Y axis. The RRMCC reveals several options for tackling the problem of resources management in a region in terms of managing land use and energy surpluses and deficits. A demonstration case study on constructing and using the RRMCC is presented in Chapter 11.
Total Imbalance
Cumulative Energy [PJ/y]
10
8
er
3
st
u Cl
6 Cumulative supply curve
ter 2
Clus
4
2
r1
ste
Clu
0
20
Cumulative demand curve
40
60
80
Cumulative Area [km2]
FIGURE 6.16
Regional Energy Supply-Deficit Curves (after Lam et al., 2009).
141
142
Chapter Six Cumulative Area [km2]
(a)
A
H
Zone 5
Cumulative Area [km2]
(b)
B
Zone 1
C
Zone 4 E
Zone 3
G
F
Zone 2 D
Zone 2
D
Zone 3 Zone 4
Zone 1
E
F
G
C B A Cumulative energy balance [PJ/y]
FIGURE 6.17
Zone 5 H Cumulative energy balance [PJ/y]
Regional Resources Management Composite Curve.
6.5 Heat-Integrated Power Systems: Decarbonization and Low-Temperature Energy 6.5.1 Decarbonization Conventional utility systems focus on how to produce and utilize the steam in a steam distribution network (Varbanov, Doyle, and Smith, 2004). Unlike conventional steam-based utility systems, however, power-dominated energy systems exhibit different characteristics because the provision of shaft (driver) power, rather than steam, is of paramount importance. For such power systems (e.g., in natural gas liquefaction), a key issue is selection of the most appropriate drivers to satisfy mechanical shaft demands. The decision factors in driver selection include the optimal number, type, and size of the drivers, helper motors or generators, and power plants—subject to a set of mechanical and electricity demands and relevant economic scenarios. Zheng, Kim, and Smith (2008a) developed a holistic approach to account for design interactions in power systems, given that driver selection entails unique implications for the overall design; these factors include overall cost, fuel consumption, performance, plant availability, carbon emissions, and so forth. Synthesis complexity increases significantly when steam systems are considered together with power-dominated systems. This case arises when a process requires a large amount of heat (steam) or when a steam turbine is preferable (as a direct driver) to a gas turbine or electric motor. In such cases, additional information is required about the on-site power supply and the way drivers interact with generating facilities. Implementation of a CO2 (carbon dioxide) capture process in the plant requires extra compression duty for the
Further Applications of Process Integration CO2 separation as well as a considerable amount of steam for the stripper. The synthesis of power-dominated energy systems is envisaged with the aid of superstructure-based mathematical optimization. The proposed superstructure (Figure 6.18) includes all possible design options, and the optimization involves (1) the systematic screening and evaluation of possible flowsheets and (2) assessing economic trade-offs between capital costs and operating costs. As usual, the optimization objective is to minimize the overall cost (i.e., capital and operating costs) while accounting for the model’s constraints (Zheng, Kim, and Smith, 2008b). This task is typically formulated as an MILP problem in which piecewise linearization is used to capture the capital costs.
6.5.2 Low-Temperature Energy The levels of power required for compression constitute a major component of energy consumption when cryogenic cooling is applied to process streams. Thus, the efficient use of such “cold energy” contributes to the cost-effectiveness of low-temperature processes. Heat Integration—in particular, one of its most powerful tools, the Grand Composite Curve—has a long history of application to saving energy in cryogenic plants (Linnhoff et al., 1982; Linnhoff and Dhole, 1992; Smith, 2005); see Chapter 4 for details. Pure refrigerant systems cannot avoid some degree of thermodynamic inefficiency, since otherwise the heat exchanger(s) would exhibit large temperature differences and this would push the system away from thermodynamic reversibility. However, if mixed refrigerants are used then the refrigeration cycle’s structure is simplified, considerably reducing the duty requirements for compression. The advantage of mixed refrigerants is that they ....
BO-1
.C1 HG/HM
GT-1
HRSG
VHP ST
HG/HM
GT-N
DR/EG
HRSG ST
EM-1
BO-N
HP
DR/EG
PP-1 ST
EM-N
MP
DR/EG
PP-N
LP ST
ExE
DR/EG CON
FIGURE 6.18 Superstructure for energy system used in a low-temperature process (after Zheng, Kim, and Smith, 2008a).
Process Stream
CN
143
144
Chapter Six facilitate a closer match between hot process and cold refrigeration streams. Del Nogal et al. (2008) described a methodology for mixed refrigerant system design based on a superstructure arrangement. The problem is highly nonlinear and features many local optima, which behave as unwanted “traps” for traditional deterministic optimization methods. This paper therefore suggests that a genetic algorithm (GA) be used to solve the optimization problem; see Figure 6.19. The interactions between the GA and the simulator result in a set of the best solutions found over a discretized solution space. The preliminary solutions so obtained then serve as starting points for standard nonlinear programming (NLP) optimization techniques to fine-tune the results and, finally, report the optimal solution. One of the important aspects of this model is that it ensures the feasibility of heat recovery in every exchanger. The design produced during optimization is then simulated, and cold and hot Composite Curves are produced. Finally, the CCs are rigorously checked against the stipulated ΔTmin.
6.6 Integrating Reliability, Availability, and Maintainability into Process Design 6.6.1
Integration
Current practice often views reliability as a mere afterthought to the design process. As a result, systems go through repeated design and redesign in search of greater process reliability, availability, and maintainability (RAM). An alternative approach, as described by Yin and colleagues (2009), is illustrated in Figure 6.20. This new
FIGURE 6.19 Integrated design for low-temperature energy systems (after Del Nogal et al., 2008).
Genetic Algorithm Set of operating condition with structural options Refrigeration simulator Updated power demands Driver selection
Integrated design
Objective function
Further Applications of Process Integration
Start
Plan preventive maintenance interval and tasks to maximize process availability while meeting risk criteria
Process Synthesis Stage
Reliability and Risk Analysis
No
Is the life cycle cost minimized?
Yes A process with the best life cycle cost
FIGURE 6.20
New process design methodology with integrated RAM stage.
methodology incorporates the simultaneous consideration of flexible process design and reliability, and optimal solutions are obtained by Process Integration. To integrate RAM into process conceptual design, the optimal preventive maintenance (PM) interval must be embedded into the superstructure. The superstructure accommodates a number of different designs, each with a different operating mode. Each socalled available design is linked to its associated financial penalty through its asset usage during the design’s life cycle, which is calculated as the ratio of real throughput to ideal throughput. This ratio is equal to the sum of the availability across all operating modes. Thus, Yin et al. (2009) defined the superstructure system availability (SSA) as SSA
Real throughput Ideal throughput
¦¦ A x
i i
j
i
(6.6)
145
146
Chapter Six Here Ai is the system availability in operation mode i for the jth design, and xi is the mode i’s ratio of actual to maximum capacity for the jth design.
6.6.2
Optimization
Within the optimization framework for integrating RAM into process synthesis, the mathematical model aims to minimize the life-cycle cost. In its general form (Yin and Smith, 2008), the optimization problem can be summarized as follows: Minimize: the objective function (expected cost) Subject to: process model constraints, preventive maintenance constraints, process system availability constraints The objective function is usually formulated as Annual cost = Annualized capital cost + Annualized operational cost + Annual lost production penalty + Other costs
6.7
(6.7)
Pressure Drop and Heat Transfer Enhancement in Process Integration Various factors—including flow rate, composition, temperature, and phase—can affect heat capacity Cp. Another factor that should be taken into account is pressure. Polley, Panjeh Shahi, and Jegede (1990) extended the Heat Exchanger Network (HEN) targeting procedure by considering pressure drop. They used the following relationship between the pressure drop ΔP, the heat transfer coefficient h, and the heat transfer area A: ⌬P
KAh m
(6.8)
where K is a pressure-drop relationship constant and m reflects the heat exchanger’s tube- and shell-side–specific coefficients. The allowable pressure drop (rather than the heat transfer coefficient) is specified for each stream. Then the heat transfer coefficients are calculated iteratively to minimize the total area. Thus, when approaching area targets the design is modified based on the fixed pressure drops rather than fixed film coefficients. Ciric and Floudas (1989) suggested a Mathematical Programming– based, two-stage approach to HEN retrofits that includes a match selection stage and an optimization stage. The match selection stage uses an MILP transshipment model to select process stream matches and match assignments. The optimization stage uses an NLP formulation to optimize the match order and flow configuration of matches.
Further Applications of Process Integration Nie and Zhu (1999) developed a strategy for considering pressure drop in HEN retrofits. They assumed that any additional area would involve only a few heat exchange units in order to minimize the piping and civil engineering work. The optimization procedure consists of two stages. The first stage involves selecting a small number of units that require additional area; the second stage considers series or parallel shell arrangements for those units. The topology change options are initially established by applying the Network Pinch method (Asante and Zhu, 1997). Then a two-stage optimization procedure is used to determine area distribution and shell arrangement under pressure-drop constraints. Area distribution and shell arrangement are the design properties that have the greatest effect on pressure drop. Václavek, Novotná, and Dedková (2003) analyzed in more detail the circumstances under which pressure plays a significant role in Heat Integration. The authors formulated some heuristic heat recovery rules for combinations of process streams (tracks), not merely individual streams. Aspelund, Berstad, and Gundersen (2007) described a new methodology, called extended Pinch Analysis and Design (ExPAnD), to account for pressure drops in process synthesis that extends the traditional Pinch Analysis to incorporate exergy calculations. The authors focus on the thermo-mechanical exergy, which is the sum of pressure- and temperature-based exergy. Compared with traditional Pinch Analysis, the problem that Aspelund, Berstad, and Gundersen (2007) consider (a subambient process) is much more complex; there are many alternatives for the manipulation and integration of streams. The authors also provide a number of (general and specific) heuristics that complement the ExPAnD methodology. In a further development, Aspelund and Gundersen (2007) used the concept of an attainable region in proposing a graphical representation of all possible CCs for a pressurized, subambient cold stream along with the cooling effect of the stream expanding to its target pressure. The attainable region is a new tool for process synthesis, extending Pinch Analysis by explicitly accounting for pressure and including exergy calculations. The methodology shows great promise for minimizing total shaft work in subambient processes. For designing a de-bottlenecking retrofit (as distinguished from an energy-saving retrofit), Panjeshahi and Tahouni (2008) suggested a method for optimizing pressure drop. The technique proceeds in two main stages as follows. (1) Simulation of the existing process operating at the desired increased throughput: additional utility is used to maintain required temperatures in the process. (2) Area efficiency specification for the existing network after the area–energy plot is used to increase throughput: a new virtual area, a pseudonetwork, is introduced. Zhu, Zanfir, and Klemeš (2000) suggested a heat transfer enhancement procedure for HEN retrofits. The methodology features a
147
148
Chapter Six targeting stage and a selection stage. The approach uses the Network Pinch Analysis to determine whether and where enhancements should be applied in the conceptual design. One limitation of this technique is that heat transfer enhancement is used only for taking the place of additional area.
6.8 Locally Integrated Energy Sectors and Extended Total Sites Total Site targeting has established a method for analyzing the heat sources and sinks of multiple processes and how heat can be transferred from one process to another via a carrying medium, such as steam (Klemeš et al., 1997). This methodology has also been used to demonstrate the concept of a locally integrated energy sector for distributing heat among small-scale industrial plants and domestic, business, and social premises while integrating renewable energy sources (Perry, Klemeš, and Bulatov, 2008). A conceptual overall design for an energy sector that involves both heat and power is illustrated in Figure 6.21. In this scenario, demands for heating/ cooling and electricity in units (e.g., dwellings, offices, hospitals, schools) can be met locally by renewable energy sources such as wind, solar cells, heat pumps, and/or excess heat and power from the local industry. Locally installed boilers that consume traditional fossil-based fuels, biomass, or waste can also be used to help meet these requirements when demand is high or other sources are unavailable. Heating or cooling and power that is not required by one unit can be fed to a grid system and then passed to another unit Electricity Grid Fossil Fuels
Renewables
Steam turbine
Nuclear
Gas Wind Sun turbine
Heat pump
Fossil fuels Bio-fuels (including waste)
Electricity Steam Hot water Cooling utility
Fossil fuels Bio-fuels (including waste)
FIGURE 6.21
Unit 1
Unit 2
Fossil fuels
Fossil fuels
Unit 3
Unit 4
Unit 5
Unit 6
Unit 7
Locally integrated energy sector with heat and power.
Further Applications of Process Integration that is unable to meet its demands locally. The grid system can distribute power (electricity) and heating in the form of hot water or steam. In geographic locations where air conditioning is required, a cooling distribution main could also be provided. If local sources are unable to provide for the demands of all units in the system, then district renewable sources can be provided. These would include larger-scale wind turbines, solar-cell systems, heat pumps, and combustors fed by waste from the units or by biofuels or fossil fuels. The sources at this level would include power-generating equipment such as turbines driven by steam or gas. Varbanov and Klemeš (2010) presented a further extension of the Total Sites methodology that covers industrial, residential, service, business, and agricultural customers; incorporates renewable energy sources; and accounts for variability on both the supply and demand sides. The challenge of increasing the share of renewables in the energy mix can be met by integrating solar, wind, biomass, and geothermal energy as well as by integrating some types of waste with the fossil fuels. The availability of renewables and the energy demands of the considered sites all vary significantly with the time of day, period of the year, and location. Some of these factors are unpredictable and can change quickly. Total Site Combined Heat and Power energy systems are optimized by minimizing heat waste and carbon footprint while maximizing economic viability. This methodology incorporates state-of-the-art techniques of Total Site Integration (Klemeš et al., 1997), batch Heat Integration (Kemp and Deakin, 1989), HEN sensitivity analysis (Kotjabasakis and Linnhoff, 1986), and time Pinch Analysis (Wang and Smith, 1995); it also applies the concept of Time Slices (see Figure 6.22) to account for the variabilities just described.
6.9 Summary Every attempt has been made to include in this chapter the most recent research results, but the field is developing so rapidly that Time Slice 1: 6–17 h
T [°C] 200
Source SCC
100
T [°C]
MP Steam LP Steam
Sink SCC
CW
100 Solar Storage
0 −1000
Excess heat 0 ΔH [kW]
MP Steam
200 Embed into the source profile
(shifted)
HW
Time Slice 2: 17–20 h
LP Steam
Solar
HW
Retrieved CW 0 1000 −1000
Storage Excess heat 0
1000
ΔH [kW]
FIGURE 6.22 Time Slice and Site targets for solar heat capture and storage (CW = cooling water, HW = hot water, SCC = Site Composite Curve).
149
150
Chapter Six hardly a month passes without several new and relevant research results being published. The most frequent publishers of novel research results are journals covering energy and cleaner technology— for example, Chemical Engineering Transactions, Applied Thermal Engineering, Energy, the Journal of Cleaner Production, Cleaner Technologies and Environmental Policy, and Resources, Conservation and Recycling. New developments in Process Integration are often published by leading chemical engineering journals, including Computers & Chemical Engineering, Chemical Engineering Science, and the AIChE Journal. A conference dedicated exclusively to dealing with related topics is the Conference on Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction (PRES). The Thirteenth PRES is scheduled for autumn of 2010 in Prague; the Fourteenth PRES will be held in Italy during 2011.
CHAPTER
7
Process Optimization Frameworks 7.1 Classic Approach: Mathematical Programming Process system engineering problems, including process synthesis, are typically considered as optimization problems. The solution or solutions of these problems are usually generated by solving the corresponding mathematical models. However, a review of recent publications reveals various failures in modeling process synthesis (Friedler and Fan, 2009). An inappropriate mathematical model may result in a nonoptimal or even an infeasible solution or the model may be unsolvable because of its complexity. A mathematical model should be a valid representation of the process, taking into account all its significant features, and still be solvable. Process optimization problems are formulated as mathematical models, where variables correspond to decisions (e.g., the flow rate of a stream, the amount of heat provided by high-pressure steam) and constraints correspond to the conceptual model of the system (e.g., material balance). Optimization (or Mathematical Programming) aims to find appropriate values for the variables in such a way that (1) constraints involving these variables are satisfied and (2) a specific function—that is, the objective function—of these variables is minimized (or maximized). The constraints define the search space, while the objective function is to determine the most favorable “point” or “points” in this space. Mathematical models are classified according to the types of variables (continuous or integer) and constraints (linear or nonlinear). Therefore, a mathematical model can be linear in constraints and objective function with continuous variables (i.e., linear programming, LPR). Similarly, a mathematical model is viewed as a nonlinear
151
152
Chapter Seven programming (NLP) problem if any of the constraints or the objective function is nonlinear with continuous variables. Models that include both continuous and integer variables are classified as mixed integer programming ones; these include mixed-integer linear programming (MILP) and mixed-integer nonlinear programming (MINLP). Also, linear optimization (LO) problems are usually referred to as linear programming or LP. Similarly, NLO, MILO, and MINLO correspond to NLP, MILP, and MINLP. Linear programming problems appear in a wide range of applications, including transportation, distribution from sources to sinks, and management decisions (Klemeš and Vašek, 1973; Klemeš et al., 1975; Klemeš, 1986; Jeżowski, 1990; Williams, 1999; Jeżowski, Shethna, and Castillo, 2003; El-Halwagi, 2006). LPR problems are easily solved by the simplex method (Dantzig, 1968) and its improvements (see, e.g., Maros, 2003a; Maros, 2003b). In most cases, NLP is difficult to solve, and certain limitations on the constraints and objective function may be necessary for such problems to be practically solvable by specific methods (Seidler, Badach, and Molisz, 1980; Banerjee and Ierapetritou, 2003; Sieniutycz and Jeżowski, 2009). A general technique for solving NLP and mixed-integer programming problems is applied by the branch-and-bound framework (Land and Doig, 1960), where the original complex problem is solved via systematic generation and solution of a set of simpler subproblems. Process synthesis is a creative activity. In fact, it is one of the earliest actions taken by the process designer when creating the structure, network, or flowsheet of a process to satisfy the given requirements in terms of constraints and specifications while attaining the prescribed objectives. The relationships among the mathematical model, the process being modeled, and the solver being deployed are usually complicated, which makes it difficult to establish the most effective and valid model. There is only limited discussion of generating mathematical models in the literature, and the topic is treated in only a few publications (see, e.g., Grossmann, 1990; Kovacs et al., 2000) concerning specific areas. In general, a process synthesis problem is defined by specifying the available raw materials, candidate operating units, and desired products. Each of these is given by an individual mathematical model. The models cannot, by themselves, directly constitute the Mathematical Programming model for the synthesis problem. Construction of the mathematical model from these model elements is not evident with the risk of failure. The major steps of process synthesis are illustrated in Figure 7.1. The main emphasis in this chapter is on an integrated framework for model generation and solution—that is, the P-graph framework. Another class of methods for process synthesis is based on heuristic rules. Implementing heuristic methods is relatively
Process Optimization Frameworks FIGURE 7.1 Major steps of process synthesis.
Cost data and constraints for the operating units. Prices and constraints for the products and raw materials.
Generation of the model
Mathematical Programming model (MILP, MINLP, NLP)
Solution of the Mathematical Programming model
Optimal network (flowsheet)
straightforward, and only moderate computational effort is required. Yet by their nature, heuristics are effective only at the local level. This is because human experiences are almost always localized: they are gained from an often limited number of encounters with (or observations of) specific instances. For this reason, solutions that are globally optimal are seldom obtainable via heuristic methods alone (Feng and Fan, 1996).
7.2
Structural Process Optimization: P-Graphs There are four good reasons to employ graph-theoretic methods: (1) the unambiguous representation of decision alternatives, (2) the algorithmic generation of a mathematical model, (3) the reduced complexity of the solution procedure, and (4) the derivation of multiple alternative solutions. The P-graph or process graph framework, as applied by Friedler and Fan (Friedler et al., 1992a; Friedler et al., 1992b; Friedler, Varga, and Fan, 1995) to process synthesis, involves novel structural representations of complex processes coupled with combinatorial algorithms for generating the superstructure, the mathematical model, and the model’s optimal solution. The P-graph framework is robust, and its algorithms have been validated as mathematically rigorous in that they are based on a set of axioms (Friedler et al., 1992b). These axioms express the necessary
153
154
Chapter Seven structural properties for process networks to be feasible. The algorithms are able to guarantee the resultant mathematical model’s validity, reduce the search space, and generate the optimal solution.
7.2.1
Process Representation via P-Graphs
In a P-graph, one class of nodes is assigned to operating units or activities and the other is assigned to their inputs and outputs. Raw materials, resources (precursors), and preconditions (activating entities) are inputs to the operating units; products, effects (resulting entities), and targets are outputs from the operating units. Table 7.1 shows the P-graph representation of process structure elements. In a process network, functional units that perform operations (e.g., mixing, reacting, separating) are termed operating units. These operating units, which correspond to the blocks in a process flowsheet, alter the physical and/or chemical states of materials being processed or transported. Such transformations are carried out by one or more unit operations, and the overall process converts raw materials into the desired product(s). A process may also generate by-products, which are either to be recovered for further use or to be treated as waste. In process network synthesis, a material is uniquely defined by its components and their concentrations—in other words, by its composition, which is identified by a symbol used to mark the material. Associated with any operating unit are two classes of materials (or material streams): input materials and output materials. For example, operating unit O2 in Figure 7.2 consumes raw materials E and F while producing intermediate material C and by-product B. Note that a material may consist of more than one component. The P-graph provides not only a formal description of the process but also an unambiguous representation of the possibilities for structural decisions. If an operating unit requires multiple inputs, Process element
P-graph representation
Raw material or precursor Final product or final target Intermediate material or entity By-product Operating unit
TABLE 7.1
P-Graph Symbols That Represent Process Elements
Process Optimization Frameworks FIGURE 7.2 P-graphs representing the process structure of three operating units.
E
(a)
F
E
(b)
F
O2
B
O2
O3
C
D
O1
A
O3
B
C
O1
A
each provided by a single operating unit, then structural alternatives cannot be defined. In contrast, if multiple operating units are capable of providing a particular input, then any combination of these units may eventually be used. In Figure 7.2(a), for example, materials C and D are necessary inputs to operating unit O1. Material C can only be produced by operating unit O2, and material D can only be produced by operating unit O3. For unit O1 to operate it is necessary that units O2 and O3 both be included in the process structure. In Figure 7.2(b), however, material C can be produced by unit O1, unit O3, or both. In addition to unambiguous structural representation, the P-graph framework also provides a set of rigorous and effective algorithms for the synthesis and optimization of process networks.
7.2.2 The P-Graph’s Significance for Structural Optimization The extreme complexity of process network synthesis is due mainly to the problem’s combinatorial nature. This complexity grows exponentially with the number n of candidate operating units, because the optimal network must be found among 2n possible combinations of the units (i.e., alternative networks) unless some possibilities can be eliminated (e.g., by heuristics) in advance. The factor 2n is derived by simple induction. First observe that a single additional decision (regarding the inclusion or exclusion of an operating unit) doubles the number of potential design alternatives: 2n × 2 = 2n+1. This means that a designer contemplating a system with a total of 35 operating units is faced with more than 34 billion (235 = 3.436 × 1010) alternative arrangements! Reducing such large numbers of alternatives requires robust decision-making tools that are mathematically rigorous (preferably axiomatic) and effectively implementable on computers. These ends
155
156
Chapter Seven have been met largely by employing the well-established mathematics of graph theory, which can be regarded as a branch of combinatorics. Thus was developed the graph-theoretic, algorithmic method described in this section. The method is based on using P-graphs to extract the universal combinatorial features (properties) inherent in feasible processes. Such properties can be expressed mathematically as a set of axioms that characterize the combinatorial feasibility of processing networks. A given process network is said to be combinatorially feasible (or to be a solution structure) if it satisfies the following five structural axioms: (S1) Every final product and target is represented in the structure. (S2) An entity represented in the structure has no input if and only if it represents a raw material or precursor. (S3) Every operating unit represented in the structure is defined in the problem. (S4) Any operating unit represented in the structure has at least one path leading to a final product or a final target. (S5) An entity belongs to the structure if and only if it is either an input entity to or an output entity from at least one operating unit already represented in the structure. Figure 7.3 illustrates the extreme reduction in the search space that results from this approach. The universe of all possible networks is reduced to a much smaller space containing only those networks that satisfy the axioms—in other words, the combinatorially feasible (CF) networks. Clearly this reduction will drastically reduce the required computational effort. Search-space reductions by a factor of nearly a billion have been reported in some of the real-life process synthesis tasks performed to date using this axiomatic approach. Note that each feasible network, including the optimal network, is an element of the set of combinatorially feasible networks. Figure 7.4 depicts two process structures that are not combinatorially feasible. The P-graph in Figure 7.4(a) shows a process structure in which material F is consumed as an input. Yet because material F is not a raw material and was never produced, the structure is not combinatorially feasible according to Axiom (S2). In the P-graph of Figure 7.4(b), operating unit O3 produces only by-product B. Here O3 does not output any final product or material that is later used to yield a final product, so the process structure violates Axiom (S4). In short, the structural properties expressed by Axioms (S1)–(S5) are necessary conditions for process structures to be feasible. This means that reducing the search space to combinatorially feasible structures does not result in the loss of any practically feasible or optimal processes.
Process Optimization Frameworks
Potential networks (search space)
CF
F
Optimal network
FIGURE 7.3 Reduction in the search space effected by combinatorial axioms (F = feasible networks, CF = combinatorially feasible networks).
FIGURE 7.4 P-graphs representing process structures that violate (a) Axiom (S2) or (b) Axiom (S4).
(a)
B
E
F
O2
O3
C
D
(b)
E
F
O3
O2 C
B
O1
O1
A
A
7.2.3 The P-Graph’s Mathematical Engine: MSG, SSG, and ABB When combined with the structural axioms, P-graph representation makes it possible to implement effective algorithms for structural analysis, synthesis, and optimization of process structures. The maximal structure generation (MSG) algorithm (Friedler et al., 1992a) generates a superstructure that can be rigorously proved to incorporate each combinatorially feasible process structure. Then the solution structures generation (SSG) algorithm (Friedler, Varga, and Fan,
157
158
Chapter Seven 1995) is used to enumerate all the combinatorially feasible process structures that satisfy Axioms (S1)–(S5) or the accelerated branch and bound (ABB) algorithm (Friedler et al., 1996) is used to generate the optimal process structure together with a ranked, finite list of nearoptimal structures. Figure 7.5 illustrates the connections among the three algorithms. Algorithm MSG generates the maximal structure, and it can be followed either by algorithm SSG to generate all combinatorially feasible process structures or by algorithm ABB to generate the optimal and near-optimal processes. Algorithms MSG and SSG require as input the list of candidate operating units, each defined by the set of its input materials (preconditions) and products (effects). Algorithm ABB requires, in addition to these data, quantitative information (e.g., prices of raw materials, costs and capacity constraints of operating units) relevant to assessing network optimality. These algorithms and their description are available online (www.p-graph.com). The methodology has been demonstrated via typical engineering decision problems (see Chapter 8 for details). A considerable advantage of the P-graph framework is its potential for solving large industrial problems, as indicated by its application in solving various process systems engineering problems (see, e.g., Liu et al. 2004; Halasz, Povoden, and Narodoslawsky, 2005; Liu et al., 2006; Fan et al., 2008; Varbanov and Friedler, 2008).
Process Synthesis Problem
Algorithm MSG
Maximal Structure
Algorithm SSG
Algorithm ABB
Combinatorially Feasible Structures (for further evaluation)
Optimal Process (n-best Processes)
FIGURE 7.5
Inputs to and outputs from the three P-graph algorithms.
Process Optimization Frameworks
7.3
Scheduling of Batch Processes: S-Graphs 7.3.1 Scheduling Frameworks: Suitability and Limitations Scheduling problems arise in many various areas, including chemical engineering, supply chain management, operating systems design, or train timetable design. The terminology and details vary among the different types of applications, but the core of scheduling remains the same: assigning tasks to resources at time intervals that satisfy predefined conditions. The goal is usually to find a feasible schedule that performs better—in terms of a particular objective—than any alternative. One of the first contributions to scheduling of chemical processes by Mathematical Programming was the paper by Kondili, Pantelides, and Sargent (1993). These authors developed the state task network (STN) representation in order to formulate production scheduling in multipurpose plants as a MILP problem. Pantelides (1993) also developed the resource task network (RTN) representation, which employs a uniform treatment for all available resources. Originally, the time representation in the formulations are discrete—in other words, a certain number of predefined times on the time horizon can be considered in the schedule. Zhang and Sargent (1996) extended the formulation to accommodate continuous-time representation, where the time points may vary, their number still have to be predefined. Ierapetritou and Floudas (1998a) employed the concept of unit-specific event points. Majozi and Zhu (2001) reformulated the problem to reduce the number of variables in certain applications. Cerdá, Henning, and Grossmann (1997) and Méndez and Cerdá (2003) developed precedence-based models suitable for cases where sequence-dependent changeovers must be considered. The rest of this section addresses some issues associated with conventional approaches to scheduling. These issues motivated the development of a novel, graph-theoretic approach: the S-graph framework (Sanmartí, Friedler, and Puigjaner, 1998; Sanmartí et al., 2002), which is introduced in Section 7.3.2. In most MILP formulations, the time horizon is divided into time intervals by so-called time points, which denote the possible starting and ending times of tasks. The number of time points is one of the model’s parameters, so it must be specified prior to optimization. Even though the quality of the solution depends strongly on this parameter, the minimum number required for the optimal solution is not known in advance. Therefore, an iterative approach is applied to determining the number of time points. First the model is solved using a small number of time points, after which each subsequent iteration increases that number by 1 until the same objective value is obtained for, say, two consecutive steps. However, it is not certain that this initial convergence necessarily yields the optimal solution.
159
Chapter Seven Castro, Barbosa-Póvoa, and Matos (2001) published a case study in which the objective function value increased after this convergence, as illustrated in Figure 7.6. A serious shortcoming of this approach is that it may generate a suboptimal solution. Another modeling issue may arise when one attempts to solve an MILP model of scheduling with no intermediate storage: according to Hegyháti and colleagues (2009), the “optimal” solution that is generated may be infeasible. In particular, the Gantt chart of a schedule—which was reported in two independent journal articles (Kim et al., 2000; Méndez et al., 2003) as the optimal solution for a case study—is not a feasible solution; see Figure 7.7. The figure reveals that, at 30 hours into production, three units (U2, U3, and Storage) attempt to exchange materials. However, this is infeasible because there is no intermediate storage and so the “optimal” schedule cannot be implemented in practice. The true optimal solution of the problem is obtained (Hegyháti et al., 2009) by using the S-graph framework (Holczinger et al., 2002); see Figure 7.8.
Objective
160
Number of time points
FIGURE 7.6 Increase in value of the objective function after initial convergence for a maximum throughput problem.
U1
B
U2
C
B
U3
C
U4
D
D
D A
B
A
B C
Storage 5
C
A
10
15
20
A
D 25
30
35
40
45
50
55
60
FIGURE 7.7 Infeasible solution generated by an MILP approach as optimal (Kim et al., 2000).
Process Optimization Frameworks
U1
B
A B
U2
C
D
U3
B
D A
C
D
U4
C
B
Storage
A
D 5
10
15
20
25
30
35
40
45
50
55
60
65
70
FIGURE 7.8 Optimal solution generated by the S-graph algorithm (Hegyháti et al., 2009).
7.3.2 S-Graph Framework for Scheduling The problems discussed in Section 7.3.1 motivated the development of an alternative methodology, the S-graph or schedule graph framework (Sanmartí, Friedler, and Puigjaner, 1998; Sanmartí et al., 2002), which has been successfully applied to the minimization of time required to complete all tasks (the makespan; see Sanmartí et al., 2002, and Romero et al., 2004) and also to problems of maximizing throughput (Majozi and Friedler, 2006). Basics of the S-graph framework are explained in this chapter; Chapter 9 describes a demonstration program for this framework that is available online (www.s-graph.com). Once all processing tasks have been represented in the “recipe,” the S-graph can be used to generate an optimal schedule. A recipe defines the order of tasks in the process, the material transfers among them, and the set of plausible equipment units for each task. The recipe is represented as an S-graph by assigning a node to each task (task node) and one node to each product (product node). An arc is established between nodes of the consecutive tasks defined by the recipe, and there is an arc also from each productgenerating task node to the corresponding product node. The weight of an arc is given as the processing time of the task that corresponds to the arc’s initial node assuming a single equipment unit is available for the task. If more than one equipment unit can perform this task, then the arc’s weight is given as the shortest processing time of all the feasible units. In the graph representing the recipe, the set of plausible units capable of performing the given task is shown in the task node; see Figure 7.9. Suppose that two batches of product A and one batch of product B are to be produced, where product A is produced in two consecutive steps. Task 1 can be performed by equipment unit E1 and task 2 by either E2 or E3. Product B is produced in three consecutive steps that can be performed by any of the elements in sets {E1, E3}, {E1}, and {E1, E2}, respectively. The recipe is shown in Figure 7.9.
161
162
Chapter Seven The equipment–task assignments and the order of tasks to be performed by an equipment unit defines the solution of a scheduling problem. The schedule is given with an S-graph containing arcs additional to the S-graph representing the recipe. Moreover, a single equipment unit is assigned to each task. As an example, Figure 7.10 depicts an S-graph representing a recipe and Figure 7.11 a corresponding solution.
2
1 {E1}
6
{E2,E3}
6
{E2,E3}
3
9
4
{E1}
5
9
6
{E1,E3}
14
{E1}
7 {E1,E2}
16
8
A
9
A
10
B
8
FIGURE 7.9 S-graph representing the recipe for two batches of product A and one batch of product B.
FIGURE 7.10
Example recipe.
1 {E1}
2 {E3}
9
{E3}
14
{E2}
{E2}
15
{E1}
16
{E3}
7
6
8
7 {E1}
9
5
4 {E2}
3
6
17
9 8
10
11
12
Schedule #1
FIGURE 7.11 Solution for the recipe under NIS policy.
1
2
3
{E1}
{E3}
{E2}
4
5
6
{E2}
{E3}
{E1}
7
8
9
{E1}
{E2}
{E3}
Schedule #1
10
11
12
Process Optimization Frameworks The algorithm generating the optimal schedule depends on the storage policy to be considered: nonintermediate storage (NIS), finite intermediate storage (FIS), or unlimited intermediate storage (UIS). In this chapter we assume a NIS policy, so an equipment unit becomes available only after finishing a task and transferring its intermediate product to the subsequent task in the recipe. On an S-graph representing a schedule under NIS policy, an arc leads from the node subsequent in the recipe to the node of the task to be performed next by the same equipment unit. For example, equipment E1 first performs task 6, then moves to task 1 and finally to task 7, which is represented by arcs drawn from node 11 (subsequent to task 6) to node 1 and from node 2 (subsequent to node 1) to node 7 in Figure 7.11. The advantage of the S-graph framework over conventional Mathematical Programming lies in its ability to exploit the problem’s structure to effect a drastic reduction on computational intensity without requiring unknown information, such as the number of time points. (Visit www.s-graph.com for further information.)
163
This page intentionally left blank
CHAPTER
8
Combined Process Integration and Optimization 8.1 The Role of Optimization in Process Synthesis Process Integration (PI), as defined in Chapter 2, is a family of optimization methodologies for reducing resource or emissions intensity of the analyzed processes and Total Sites. As such, it is tightly related to optimization. In fact, PI and optimization complement each other by their functionality. First, PI sets out the strategy for designing and/or operating industrial processes. This gives engineers some direction regarding how processes can be designed or changed, answering the questions of “where we can go?” and “what is to be done?” in order to achieve the business goals at hand. In addition, PI provides quantitative targets for designers and engineers; it does this by exploiting the physical (in the case of Heat Integration, thermodynamical) background to answer the question “how much is it possible to improve or achieve?” The targets in most cases are upper bounds on the process performance or lower bounds on the extent of resource use or emissions. In many cases the targets are practically achievable, as in the case of designing Heat Exchanger Networks (HENs) or water networks. The most obvious example is heat recovery targets established by the Pinch point when analyzing HEN problems. Since they are based on the Second Law of Thermodynamics, it is proven that better heat recovery cannot be achieved by any feasible system. If a realistic value for ΔTmin is specified for their evaluation, the targets will be also practically achievable. Of course, the additional factor of the capital costs for implementing the heat exchangers generally tends to shift the economic optimum of the designed HEN away from the maximum recovery network.
165
166
Chapter Eight The issues tackled by PI are essentially complex optimization problems. As a result, optimization is used by PI to answer the question of “how should the task be performed?” The general goals and specific targets are usually achieved by employing optimization tools at various stages. For instance, process performance targets are typically evaluated by employing a numerical technique that involves a cascade of some sort—a heat cascade in the case of heat recovery targeting. One way to implement such cascades is by using the transshipment optimization formulations, where the external utility use, resource intake, or emissions rate is set as the objective function to be minimized. Once the PI goals are established, engineers strive to achieve the best possible performance. In the case of grassroots design or network synthesis, the criterion is minimization of the total annualized cost; in the case of a retrofit, the main criterion may be minimizing the investments necessary to achieve a certain performance improvement or minimizing the payback period for a given investment. For operational improvements, the criteria include minimizing operating costs or maximizing marginal financial or performance gains. In all cases, a certain system model—including the appropriate objective function—is formulated. The model is then subjected to optimization toward the end of achieving (or maximally approaching) the PI targets. Another function of targets is to partition complex optimization problems into sets of simpler problems that are easier to solve. This approach exemplifies the problem decomposition principle, applied for decades in the world of software development, also known as the “divide and conquer” strategy.
8.2 Optimization Tools for Efficient Implementation of PI For optimizing process models, a wide variety of linear programming (LPR), nonlinear programming (NLP), and mixed-integer programming (MIP) methods can be used, depending on the nature of the problem being solved. Some of these methods were described in Chapters 3 and 7. Special tools and software (see Chapter 9) incorporating optimization methods have been developed to exploit PI possibilities when performing process synthesis, accounting for the interactions between process operating conditions and the networks for resource recovery (energy and water). There are four main groups of optimization tools applied for PI. First, the Pinch Analysis (Linnhoff et al., 1982) enabled industrial engineers to obtain better results with the simple Pinch Design Method than with Mathematical Programming methods in applications to industrial Heat Integration; see Chapter 4. Second, the graphtheoretic method is based on process graphs (P-graphs), which were originally developed for Process Network Synthesis (PNS) (Friedler et al., 1992b; Friedler, Fan, and Imreh, 1998); see Chapter 7. Third, Papoulias and Grossmann (1983) introduced linear constraints in
Combined Process Integration and Optimization their transshipment model for Heat Integration within a mixedinteger linear programming (MILP) formulation for structural process optimization. This work had been further extensively developed (Duran and Grossmann, 1986; Floudas and Grossmann, 1987a; Floudas and Grossmann, 1987b). Fourth, stochastic optimization has become popular in recent years, applying genetic optimization (Shopova and Vaklieva-Bancheva, 2006) and especially simulated annealing (Kirkpatrick, Gelatt, and Vecchi, 1983; Faber, Jockenhövel, and Tsatsaronis, 2005; Hul et al., 2007; Tan, 2007). Optimization methods can be classified according to the characteristics of the objective function, the decision variables, and the problem constraints (Guinand, 2001). A simplified classification scheme for optimization methods is illustrated in Figure 8.1.
8.3
Optimal Process Synthesis A process network uses a given set of operating units to create desired products from specific raw materials. The objective of PNS is to identify the most favorable (optimal) network for accomplishing the given tasks. The P-graph methodology is a graph-theoretical approach to solve PNS problems.
8.3.1
Reaction Network Synthesis
Every reaction is a material transformation, which corresponds to an operating unit when mapped on a P-graph. Similarly, the maximal
Constraints and Objectives
Decision Variables
LINEAR
NONLINEAR
Linear Programming (LPR)
Nonlinear Programming (NLP)
Integer Linear Programming (ILP)
STOCHASTIC
DETERMINISTIC
Incorporation of Probability Functions
FIGURE 8.1
CONTINUOUS
INTEGER
Integer Programming (IP)
Mixed-Integer Linear Programming (MILP)
Mixed-Integer Nonlinear Programming (MINLP)
DYNAMIC
STATIC
Incorporation of Time Domain
Classification of optimization methods (after Guinand, 2001).
167
168
Chapter Eight reaction network can be generated using the Maximal Structure Generation (MSG) algorithm. From that, all feasible reaction networks can be generated as solution structures. Historically, it has been extremely difficult to construct an exact maximal reaction network. The problem has become solvable only since the arrival of the combinatorial approach based on P-graphs. Reaction network synthesis is completely combinatorial in nature because all chemical species participating in the reactions are defined discretely. The application of PNS to Reaction Network Synthesis (RNS) is illustrated by a chemical process for the manufacturing of vinyl chloride (C2H3Cl). Because of its structural simplicity, the balanced synthesis of vinyl chloride (C2H3Cl) leads to only a single complex route containing one loop. It is a trivial example from the standpoint of constructing the maximal reaction network and the feasible reaction networks, but it will serve to illustrate the RNS methodology. The balanced process aims to produce vinyl chloride (C2H3Cl) and water (H2O), where the former is the desired product and the latter is the by-product, from the three starting reactants: ethylene (C2H4), chlorine (Cl2), and oxygen (O2). This process involves the following three unit reactions: R1, R2, and R3. R1: 2C2H4 + 2Cl2 = 2C2H4Cl2 R2: 2C2H4 + 4HCl + O2 = 2C2H4Cl2 + 2H2O R3: 4C2H4Cl2 = 4C2H3Cl + 4HCl The unit reactions yield the following overall reaction: 4C2H4 + 2Cl2 + O2 = 4C2H3Cl + 2H2O Note that C2H3Cl is the final target and that C2H4, Cl2, and O2 are the starting reactants. From the perspective of materials (M), we have: Starting reactants Final products Intermediates
M { C 2H 4 , Cl 2 , O 2 , C 2H 3 Cl , H O , C H 2 2 4 Cl 2 , HCl } N
Raw materials
Product
By-product
Intermediate materials
One or more of the feasible paths and valid vertices in the input structure may disappear if some of the invalid vertices are eliminated. Thus, the final maximal structure is composed (or reconstructed) from the remaining skeleton of the input structure after the elimination. This is accomplished step by step, linking alternately the vertices of the M-type (for materials) to the vertices of the O-type (for operating units) and vice versa. At each step, the vertices linked are assessed in view of the appropriate axioms (see Section 7.2.2): vertices of the M-type must satisfy axioms (S1), (S2), and (S5); and vertices of the O-type must satisfy axioms (S3) and (S4). The execution is initiated from the structure’s shallowest layer—that is, the final, desired product end. The stepwise procedure for the composition is illustrated in Figure 8.2.
Combined Process Integration and Optimization Cl2
R2
R1 Step 5
O2
C2H4
C2H4Cl2
H2O
R3 C2H3Cl
R1
HCl
R2
R1 H2O
C2H4Cl2
R2
C2H4Cl2
R3
H2O
R3
C2H3Cl
HCl
C2H3Cl
HCl
Step 3
Step 4
C2H4Cl2 R3
R3
C2H3Cl
HCl
C2H3Cl
Step 1
HCl
Step 2
FIGURE 8.2 Steps for the composition of the maximal structure representing the maximal reaction network for the manufacture of vinyl chloride.
The maximal reaction network serves as the input to algorithms SSG and Accelerated Branch-and-Bound (ABB). Algorithm SSG generates all combinatorially feasible reaction networks, and algorithm ABB identifies a set of the most favorable reaction networks.
8.3.2 Optimal Synthesis of Heterogeneous Flowsheets Synthesis of Optimal Workflow Structures Workflow technology has become a general tool for a wide range of business and management applications (Tick, Kovács, and Friedler,
169
170
Chapter Eight 2006), which is mainly due to its ability to increase the efficiency of business and management systems. The mathematical foundation of this technology has been developed mainly based on Petri net theory (Kiepuszewski, ter Hofstede, and van der Aalst, 2003; van der Aalst and ter Hofstede, 2005). Even though this mathematical methodology provides a basis for determining the optimal operation of workflows, it cannot be used to derive an optimal workflow structure. The structural component of a workflow synthesis problem can be identified by the sets of products, resources, and (plausible) activities on materials. The cost of a workflow process that generates a particular quantity of product is given as the sum of (1) the cost of the raw materials and (2) the cost related to the activities appearing in the synthesized workflow process. The cost of an activity is the sum of its running cost and the investments assigned to the period of time examined. Both the running cost and investment cost depend on the “size” of the activity—that is, its output volume. The common objective for synthesizing workflow processes is to minimize the total cost under the assumption of unlimited intermediate storage capacities for any activity. Example 8.1: Workflow Synthesis (after Tick, Kovács, and Friedler, 2006) As an example, a set of activities is given by its inputs and outputs in Table 8.1 and represented by P-graph in Figure 8.3. The P-graph contains the interconnections among the activities. Each feasible activity network corresponds to a subgraph of the P-graph in Figure 8.3. A product document represented by A and B can be generated by an appropriate network of the activities—provided that the problem has at least one feasible solution. It is important to note that a product can usually be generated by different types and numbers of activities. When determining the optimal network for a workflow, all possible networks of each product must be taken into account.
Activities
Input
Output
1
C
A, F
2
D
B
3
E, F
C
4
F, G
C
5
G, H
D
6
H
B
7
J
F
8
K
G
9
K
G
10
L
H
TABLE 8.1 Workflow
Plausible Activities for a
Combined Process Integration and Optimization J
K
7
8
F
E
L
9
G
3
H
4
5
6
D
C
1
A
FIGURE 8.3 activities.
10
2
B
P-graph, where A, . . . , L are the materials and 1, . . . , 10 are the
The number of feasible networks is usually large, so a systematic procedure is needed to determine the optimal network. For this purpose, the P-graph framework is proposed. Algorithm MSG verified that the P-graph in Figure 8.3 corresponds to the maximal structure, and algorithm ABB provided an optimal set of several alternative structures containing the strict optimum and other mathematically suboptimal networks. One of the 50 solution networks for this problem is given in Figure 8.4.
8.3.3 Synthesis of Green Biorefineries P-graph is also employed to examine the contribution that process synthesis can offer to the development of production based on renewable resources (Halasz, Povoden, and Narodoslawsky, 2005). These resources face formidable competition from fossil raw materials. Because their production is usually decentralized, viable processes can result only if the structure of the complete value chain is optimized. In contrast to conventional chemical processes, those using renewable resources have to account for logistic operations affecting the value chain’s structure. Hence, using process synthesis in this application calls for using the highest-quality method for synthesis. The P-graph method, as described in Chapter 7, has proven to be extremely efficient and flexible, so it is well suited for this purpose. Example 8.2: Green Biorefinery Synthesis (Kromus, 2002) The components of this example are summarized by the flowsheet shown in Figure 8.5. The goal of the considered system is the (decentralized) production
171
172
Chapter Eight J
K
7
L
10
9
G
F
H
4
C
5
D
1
A
FIGURE 8.4
2
B
A solution network for the workflow synthesis problem.
of silage on farms. By starting from silage, two crucial steps are combined: storage (thus enabling the downstream processes to operate continuously) and conversion of the carbohydrates in green biomass to lactic acid. In addition, silage production transforms many proteins into amino acids or peptides (Povoden, 2002; Koschuh et al., 2004).
The synthesis method requires a comprehensive list of raw materials, intermediates, and possible products. Note that transport is treated like a processing step: it uses trucks (or tractors), together with the raw materials (or partially processed juice or press cake) and “available time,” in order to derive a realistic logistics pattern. Consequently, there has to be a plant-specific intermediate material flow that leaves this “process step.” The steps listed in Table 8.2 reflect the necessary logistical handling (i.e., the “local” and “central” converters for various materials) involved in the process. Once the cost function (including investment and operating costs) is defined, the synthesis yields the optimal solution to process silage as a part of the maximal structure; see Figure 8.6 (black lines show the optimal solution, while the other options are grayed). One major advantage of process synthesis is that it allows the designer to apply sensitivity analysis not only to the process itself but also to the entire value chain. In this example, sensitivity analysis reveals a remarkable stability of the central biorefinery structure
Combined Process Integration and Optimization Double Pressing 5289 t/y Water added for second pressing Silage Feed Matter 35511 t/y Dry Matter 10000 t/y Dry Matter 28.16 % Lactic Acid 1000 t/y Raw Fibre 2825.7 t/y Org. Dry Matter 8992 t/y
Presscake Feed Matter 13825 t/y Dry Matter 6636 t/y Dry Matter 48 % Lactic Acid 215 t/y Raw Fibre 2824 t/y Org. Dry Matter 6483 t/y Flux-out 1.9 t FM/h
Silage Juice Feed Matter 26976 t/y Dry Matter 3364 t/y Dry Matter 13 % Lactic Acid 785 t/y Org. Dry Matter 2509 t/y Flux-out 3.6 t FM/h Heat: 183442 kWh/y
Drying: 603924 kWh/y
Amino Acids Feed Matter 755 t/y Dry Matter 679 t/y Dry Matter 90 % Lactic Acid 3.08 t/y Crude Prot. 666 t/y Flux-out 0.1 t FM/h Crude Protein 98 % DM
Heat to increase concentration: 167033 kWh/y
Lactic Acid 80% D/L Conc. 80 % Feed Matter Product 884 t/y Dry Matter 707 t/y Org. Dry Matter 706 t/y Pure Lactic Acid 706 t/y Flux-out 0.12 t/h
Energy for Drying: 6451609 kWh/y
Fibres Feed Matter Dry Matter Dry Matter Lactic Acid Raw Fibre Org. Dry Matter Flux-out
7373 t/y 6636 t/y 90 % 215 t/y 2824 t/y 6483 t/y 1 t/h
FIGURE 8.5 Flowsheet of a green biorefinery: Base-case mass flows and concentrations (DM = Dry Matter; FM = Feed Matter).
despite variations in several factors that affect the optimal solution for silage fractionation and biogas plants within the value chain, including prices of key products such as electricity and press cake. Further details on the method and the case study are provided in (Halasz, Povoden, and Narodoslawsky, 2005).
8.3.4
Azeotropic Distillation Systems
Azeotropic distillation is common in chemical and allied industries. Many of the existing distillation processes were designed and developed through extensive trial-and-error efforts. The thermodynamic pinches or boundaries (e.g., azeotropes), distillation boundaries, and boundaries of liquid–liquid equilibrium envelopes are of critical importance for azeotropic distillation. Moreover, the compositions of the feed and product streams must be specified in order to define the synthesis problem for an azeotropic distillation
173
174
Chapter Eight Operating unit
Acronym
Operating unit
Mobile press
MP
Central converter, juice
CCJ
Central press
CP
Local converter, rest of fibers
LCRF
Local fibers production
LF
Central converter, rest of fibers
CCRF
Central fibers production
CF
Central converter, rest of juice
CCRJ
Green biorefinery
GBR
Local transport, silage
LTrS
Local biogas
LBG
Central transport, silage
CTrS
Central biogas
CBG
Local transport, cake
LTrC
Local converter, silage
LCS
Central transport, cake
CTrC
Central converter, silage
CCS
Local transport, juice
LTrJ
Local converter, cake
LCC
Central transport, juice
CTrJ
Central converter, cake
CCC
Central transport, rest of fibers
CTrRF
Local converter, juice
LCJ
TABLE 8.2 Case
Acronym
List of Process Steps Incorporated into the Synthesis of the Base
system. Such information can be represented by a residue curve map (RCM). The RCM of an ethanol-water-toluene system is shown in Figure 8.7. The points E, W, and T represent the pure components ethanol (product), water (by-product), and the entrainer, respectively, while points F and H denote the feed and the ternary azeotrope, respectively. The whole RCM is partitioned into materials corresponding to these points; then the lines L1, . . . , L13 demarcate the areas A1, . . . , A6. A set of operating units for this process can be represented in a P-graph; see Figure 8.8. The maximal structure and each solution structure can be generated by algorithms MSG and SSG, respectively. The Mathematical Programming model of a process network includes constraints on operating units (i.e., mathematical models of those units) as well as constraints on materials (e.g., mass balance constraints). Mathematical models of mixers, separators, and decanters are linear, including their mass balances, because they are based on component flow rates. Thus, the Mathematical Programming model of process networks involving mixers, separators, and decanters gives rise to an LPR problem. Each solution structure satisfying
Combined Process Integration and Optimization S CTrS
MP J
SC
C CP
CTrC LTrS
LTrJ
SL
LTrC CL
CTrJ CC
LF
JC
RFL
JL
GBR
CTrRF CF
F1
RJC RFC LCS
LCJ
LCC
LCRF
CCRF
F2
GODML
CCC
CCRJ
CCS
GODMC CBG
LBG
E2
E1
FIGURE 8.6
LA AA CCJ
Maximal and optimal structure for base-case synthesis.
FIGURE 8.7 Residue curve map (RCM) of the ethanol-watertoluene system.
E L11
A=Area L=Line
L1
A1 L10
L13 L12 A6
A2
A5
F L8
H
L2
L6 L5
L9 W
A4
L7
A3
L4
L3
T
175
176
Chapter Eight FIGURE 8.8 P-graph representing the structure of an azeotropic distillation system.
F
L12
Mixer
L6
Separator
W
L7
L13
Mixer
L1
Separator
E
L11
Decantor
combinatorial constraints has been generated by algorithm SSG and evaluated by linear programming (Feng et al., 2003).
8.4 Optimal Synthesis of Energy Systems 8.4.1 Simple Heat Integration A methodology for combining PNS and HENs for integrated synthesis has been presented by Nagy et al. (2001). It employs the hP-graph, a special graph that incorporates both operating units and heat exchangers. Figure 8.9 shows an example flowsheet and its hP-graph representation.
Combined Process Integration and Optimization M9 M10
7
M8
6
5
M11
M4
2
M7
M9
363 343 M3 1
M6
M2 M1
4 328 6
363
3
M6(6)
M5
M10
M6(4)
M3(3)
4
M11 7
M7 3
353 M5
M6(1)
M8 5
M3(4)
M4
M3(1) 333 1
Heating: Cooling:
2 M1
M2
FIGURE 8.9 Flowsheet (left) and hP-graph (right) representations of a heatexchanging system.
In a hP-graph, the node for a heater is indicated by a bar with a solid lower half and that for a cooler by a bar with a solid upper half. If the content of an operating unit is heated or cooled by latent heat, then its “mode” is extended to the left by an appropriate heat exchanger. For example, the flowsheet in Figure 8.9 shows that operating unit 3 is heated with hot latent heat. Suppose that part of a material stream fed to an operating unit requires temperature modulation but that the remaining part, which feeds another operating unit, does not. In this case, the former is diverted through a heat exchanger and the latter is directed to another operating unit (as with the two streams of M6 in the figure). The ABB algorithm determines whether any of the operating units should be included in the optimal structure. Applications of the PNS algorithms extended to the synthesis of combined process and HENs amply demonstrate the efficacy of the P-graph framework.
8.4.2 Optimal Retrofit Design A novel holistic method based on the P-graph approach has been proposed by Halasz, Povoden, and Narodoslawsky (2002) and Liu et al. (2006) for process retrofitting. Unlike conventional approaches, the proposed methodology resynthesizes the entire process by considering the enhanced performance of the operating units. As such, this approach can account for all possible outcomes, including the network’s inevitable restructuring. The method employs a P-graph implementation originally devised for synthesizing grassroots processes. With the combinatorial feasibility of most operating units largely predetermined, the approach detects any
177
178
Chapter Eight necessary retrofit changes in the network structure. For example, the retrofit of a conventional downstream process for the biochemical production of butanol is accomplished by incorporating newly identified adsorbing units, whose characteristics are summarized in Table 8.3. Note that each of the processing units 24 and 25 are represented by two operating units in the P-graph—that is, 24–1, 24–2 and 25–1, 25–2, respectively. The P-graph of the problem’s maximal structure is presented in Figure 8.10. It includes a gas stripper, an extractor, 27 simple distillation columns, two azeotropic distillation units, a centrifuge, and four adsorption columns. The objective function is to minimize the operating cost (in terms of its present values) for all operating units in the process flowsheet. The cost for the three operating units is estimated heuristically. The optimal flowsheet is illustrated in Figure 8.11. This approach to retrofitting accounts for the fact that any change implemented in downstream processing systems—especially at an upper segment—will tend to propagate throughout the system as a function of each system’s unique sequence of structural features. The approach is also applicable to many other chemical processes that share the same structural features. With this methodology, a set of optimal and near-optimal retrofit flowsheets can be generated and ranked in terms of their costs. One (or more) of the options in this set may be seen as infeasible or unsustainable when the retrofitted system is further assessed by taking into account additional criteria and constraints, including stability and controllability as well as environmental, societal, and regulatory constraints.
Operating units and subunits
Cost [103 $]
Annual cost [103 $/y]
Capital
Annualized capital
Operating
Total
No. of units
No. of subunits
Equipment designation
Type
23
—
C1
Centrifuge
9,240
3,080
1,168
4,248
24
24–1
B1
Adsorption column
25,107
8,369
871
9,248
24
24–2
B2
Adsorption column
25
25–1
B3
Adsorption column
3,806
1,269
132
1,401
25
25–2
B4
Adsorption column
TABLE 8.3
Characteristics, Including Costs of Operating Units, Identified by the P-graph Method (after Liu et al., 2006)
Combined Process Integration and Optimization S00 E1
C1
G1
S05 S51 S11 S16 S03 B4 B2 S1 B3
S52 B1
D2
D1
S01 S53
S07 S06 S13
S54 S08
D5
D9
D11
S15
D7 S17
S31
S32
S33
S35
D6
D8
D10
D12
D16
D18
D15
S34 D17
S36
D13
D14
D21
D19 S37 S38
S46 S43 D25
S43 S44
S39 S40
S45
D20 D22
D27 S48
S49
D28
D29
A2 D3
A1 S02
D26
S09 S19 S20
FIGURE 8.10 P-graph representation of the maximal structure for producing butanol, ethanol, and acetone with the inclusion of adsorption.
S09 [A 7] S08 A7 E2 W 26 A 11 E3 W 35 B 26
D21
S39
S20 [E 2] S00
D22
3(G1)
S40
S19 S05
[B 26] B3
S01
B4 25-1(B3)
25-2(B4)
S08
S54
S05 S00
S01 A 11 E4 W 1773 B 27
FIGURE 8.11
20-1(D21)
S54
G1
E1 W 1738 B1
A4 E1 W 35
S09 Legend: Acetone: Butanol: Ethanol: Water:
S40
S39
20-2(D22) A B E W
kg/h kg/h kg/h kg/h
S20
S19
Optimal flowsheet for the example considering adsorption.
8.5 Optimal Scheduling for Increased Throughput, Profit, and Security 8.5.1 Maximizing Throughput and Revenue Majozi and Friedler (2006) developed an effective search algorithm for determining the globally optimal throughput, revenue, or profit over a predefined time horizon in multipurpose batch plants on the basis of the S-graph framework. To demonstrate its performance, a
179
180
Chapter Eight case study from a real-life multipurpose batch facility is presented next. Example 8.3: Optimal Scheduling The case study is taken from a multinational pharmaceuticals facility that produces lotions, shampoos, conditioners, and various creams. The problem features nonintermediate storage policy (NIS), and the processes involve mixing and packaging. Mixing occurs in four mixing vessels (V1, V2, V3, and V4), and packaging occurs in three packing lines (P1, P2, and P3). Because the stirrers in mixing vessels are of different designs, mixing times vary according to the vessel used. Table 8.4 shows the duration of mixing for each product in each vessel, which have a capacity of about 3 t each. The table also lists the economic contribution made to the company revenue or profit by selling a unit of each product; shampoos have the highest economic contribution. The packing duration for each product is 12 h, regardless of which packing line is employed. The objective in this case study is to maximize the overall economic result for a 24-h period. The S-graph for the recipe for the products manufactured in this facility is given in Figure 8.12, where the sets of candidate equipment units for performing tasks 1, . . . , 15 are defined by sets U1 = {V1, V2, V4}, U2 = {P1, P2, P3}, U3 = {V1, V2, V3}, U4 = {V3}, U5 = {V2, V3}, and U6 = {V1, V2, V4}. The global optimal solution corresponds to two batches of Cream 2 and one batch of Shampoo, which yields revenue of 9.5 cost units. The schedule corresponding to the global optimum is shown in Figure 8.13. Two advantages that this approach has over its Mathematical Programming counterparts are: (1) it guarantees global optimality and (2) no manipulation of the time horizon is required—in particular, it is unnecessary to presuppose “time points” that will discretize the time horizon into equal (or unequal) time intervals. For this reason, the technique qualifies as a true continuous-time methodology.
8.5.2 Heat-Integrated Production Schedules Many algorithmic and heuristic methods have been developed for solving Heat Integration problems in continuous processes: Pinch Technology (Linnhoff et al., 1982), superstructure-based mixed
Product
Economic contribution [cost unit/batch]
Production time in mixing vessel [h]
Cream 1
2
10
5
N/A
5
Cream 2
3
12
10
7
N/A
Conditioner
1
N/A
N/A
12
N/A
V1
V2
V3
V4
Shampoo
3.5
N/A
8
13
N/A
Lotion
1.5
10
6
N/A
9
TABLE 8.4
Scheduling Data for the Case Study of Example 8.3
Combined Process Integration and Optimization FIGURE 8.12 S-graph representing the recipe for the case study of Example 8.3.
P3
1 U1
5
2 U2
1
3 U3
7
4 U2
1
5 U4
1
6 U2
1
7 U5
8
8 U2
1
9 U6
6
10 U2
1
Cream 1
12
Cream 2
13
Conditioner
14
Shampoo
15
Lotion
Task 8 (Batch 1)
P2 Equipment Units
11
Task 4 (Batch 2)
P1
Task 4 (Batch 1)
V4 V3 Task 3 (Batch 1) V2 Task 7 (Batch 1) V1 Task 3 (Batch 2) 7 8
12
19
Time [h]
FIGURE 8.13
Globally optimal schedule for the case study of Example 8.3.
integer programming (Douglas, 1988; Biegler, Grossman, and Westerberg, 1997), and integration with PNS (Nagy et al., 2001). In contrast, scheduling and Heat Integration of batch processes are both complex optimization problems and quite different in nature. The problems of scheduling and Heat Integration could be solved sequentially (in either order). Yet because the solution of one problem will affect the other, the result of this simplistic, sequential approach is usually poor. A better solution may result from an integrated consideration of scheduling and Heat Integration. Since there were not too many methods for solving this integrated problem, the effective design and operation of integrated batch systems required development of a new method (Adonyi et al., 2003). The goal was to operate simultaneously those tasks that involve potential heat
181
182
Chapter Eight exchange without compromising the quality of the scheduling solution. Example 8.4: Heat-Integrated Production Schedules The recipes of products A, B, and C are shown in Table 8.5, and the corresponding S-graph of the recipe is given in Figure 8.14. The parameters of the heat streams are listed in Table 8.6. The available hot and cold utilities are 200°C and 10°C. The heat transfer coefficient between a hot and a cold stream is 1500 W/m2∙°C, and the area of a heat exchanger unit is 3 m2.
Product A
Product B
Product C
Task
Eq.
Time [h]
Eq.
Time [h]
Eq.
Time [h]
1
E1 E2
5 5
E1 E2
5 5
E1 E2
6 6
2
E3 E4
4 4
E5 E6
4 4
E5 E6
3 3
3
E5 E6
4 4
E7 E8
4 4
E7 E8
4 4
4
—
—
E1 E2
5 5
—
—
TABLE 8.5
Recipes for the Products in Example 8.4
1 U1
5
2 U2
4
3 U3
4
4 U4
5
5 U5
4
6 U6
4
7 U7
5
8 U8
4
9 U9
4
10 U10
5
11 U11
4
12 U12
4
13 U13
5
14 U14
4
15 U15
4
16 U16
5
35 U2
B
17 U17
5
18 U18
4
19 U19
4
20 U20
5
36 U2
B
21 U21
5
22 U22
4
23 U23
4
24 U24
5
37 U2
B
25 U25
6
26 U26
3
27 U27
4
28 U28
6
29 U29
3
30 U30
4
FIGURE 8.14
31
A
32
A
33
A
34
A
38
C
39
C
S-graph of the recipe for the products in Example 8.4.
Combined Process Integration and Optimization Name
Type
Initial temp. [°C]
Final temp. [°C]
Heat [MJ]
Product/ Task
c1 h1
Cold
40
120
400
A/2
Hot
140
50
200
B/3
c2 h2
Cold
80
130
100
B/4
Hot
150
40
300
C/2
Equipment Units
TABLE 8.6 Parameter Values for the Heating and Cooling Requirements in Example 8.4
E1
1
4
E2
25
28
13
10
17
2
E3
21
7
5
3 26
14 29
6
18
9
22
12 19
E7 27 10 Product A
FIGURE 8.15
24
11
E5
E8
16 8
E4
E6
20
30
15
20 Time [h] Product B
23 30
36
Product C
Gantt chart of the optimal solution for Example 8.4.
If the heating and cooling duties are satisfied by utilities, then the minimal makespan is 33.1 h with 3100 MJ utility. Extending the upper bound for the makespan to 36 h reduces the required utility to 1100 MJ. Figure 8.15 displays the Gantt chart of the optimal solution.
8.6 Minimizing Emissions and Effluents The task of designing a complete energy system involves significant combinatorial complexity. For this, integer programming procedures are not efficient. The P-graph framework and its associated algorithms are capable of efficiently handling exactly the type of complexity that is inherent to network optimization, and they appear to be some of the best tools for solving this task. The P-graph approach can readily evaluate technologies in their early stages of development, such as fuel-cell combined cycles (FCCCs) based on molten carbonate and solid oxide fuel cells (Varbanov and Friedler, 2008).
183
184
Chapter Eight Example 8.5: FCCC Systems for Reducing Carbon Footprint This example presents a procedure for evaluating energy conversion systems involving FCCC subsystems that use biomass and/or fossil fuels; see Figure 8.16. The procedure provides a tool for evaluating trends in CO2 emission levels and the economics of such systems. The significant combinatorial complexity involved is efficiently handled by P-graph algorithms. Promising system components are evaluated by a methodology for synthesizing cost-optimal FCCC configurations that account for the carbon footprint of the various technology and fuel options using the P-graph framework. The efficiency of an FCCC system varies with the fuel-cell (FC) operating temperature, the type of bottoming cycle, and the degree of cycle integration (see Varbanov et al., 2007). High-temperature fuel cells can be combined with gas turbines, steam turbines, or both; however, combining all three yields only marginal improvements. The main reason is that energy in the FC exhaust can only be shared by the bottoming cycles, and typically this potential for energy generation is most fully utilized by a steam or gas turbine alone. Hence the involvement of more than one bottoming cycle cannot substantially increase overall efficiency, although it can present capital cost trade-offs. Figure 8.17 shows an FCCC system represented by a conventional block-style diagram and a P-graph fragment. The synthesis of a processing network, such as the energy conversion system considered here, requires that the designer choose the best solution from a
Fossil fuels Biomass
FIGURE 8.16
Biofuel
Power
Energy Conversion: FCCCs Boilers ...
Processing: Gasification or Digestion
Heat
FCCC system boundary and processing steps.
F F CO2 FCCC
W
FCCC
Q Block-style flowsheet
W
Q P-graph
Legend F: Fuel; FCCC: Fuel Cell Combined Cycle unit; Q: Heat; W: Power
FIGURE 8.17
Flowsheet and P-graph representations of an FCCC system.
CO2
Combined Process Integration and Optimization number of options. This optimization task may have several different objectives. The most obvious are maximizing the system profit (minimizing its cost) and minimizing the amount of CO2 emissions. Although it is mathematically possible to define a multiobjective criterion to be optimized, using profitability alone seems most coherent with the logic of a market economy because profit drives the behavior of companies. Therefore, in this discussion the system profit is used as the objective (to be maximized); CO2 emissions are then used as an additional criterion during the analysis stage. The materials and streams for the system considered in this example are listed in Table 8.7. The waste products are assigned negative prices, denoting that they generate costs for the system rather than revenue. Other performance
Stream
Type
P-graph classification
Description
Price
BM
Biomass
Raw material
Agricultural residues
Varies
BG
Clean biofuel
Intermediate
Biogas suitable for utilization as a fuel
—
BR
Waste / By-product
Product / Output
Biomass residues (solid remainder from the biomass after gasification)
−10 €/t
CO2
Waste, greenhouse gas
Product / Output
CO2 emissions
Varies
FRT
Useful by-product
Product / Output
Fertilizer obtained as a by-product from the anaerobic digester
50 €/t
NG
Fossil fuel
Raw material
Natural gas
36.8 €/ MWh
PR
Waste / By-product
Product / Output
Particulates left from cleaning the synthesis gas
−10 €/t
Q40
Steam
Intermediate
Steam at P = 40 bar(a) —
Q5
Steam
Product / Output
Steam at P = 5 bar(a) to satisfy user demands
30 €/ MWh
RSG
Intermediate fuel
Intermediate
Raw synthesis gas
—
SG
Clean biofuel
Intermediate
Clean synthesis gas suitable for use as a fuel
—
W
Power product
Product / Output
Electrical power to satisfy user demands
100 €/ MWh
TABLE 8.7
Materials and Streams for Example 8.5
185
186
Chapter Eight and economic data specifications, which provide the basis for appropriate economic evaluation of the designs, are given in Varbanov and Friedler (2008). The P-graph tools can be used to generate different solutions based on the objectives of interest (operating cost, CO2 emissions) and market conditions. The results show that systems of this type that employ renewable fuels are economically viable for a wide range of economic conditions; this finding is due mainly to the high energy efficiency of the FC-based systems. Figure 8.18 shows the P-graph for the base case in which operating cost is minimized.
8.7 Availability and Reliability The significance of waste management systems has increased in recent years because of the growing problems of waste management chains affecting not only the environment, but also the daily lives of millions of people. Several promising approaches have appeared, including the RAMS (reliability, availability, maintenance, and safety) software for modeling waste management systems. This approach was analyzed and evaluated thoroughly by Sikos and Klemeš (2009a, 2009b, 2010). In today’s technological world, most waste management plants depend on the continuous operation of a wide array of complex machinery and equipment to sustain development, safety, human health, and economic welfare. Plant operators expect a wide variety BM 55.2 MW
BGD 4.2 t/h 1.4 t/h CO2 0.17 t/h
BG
32.0 MW
16.9 MW
FRT 15.1 MW BLR_BG 12.8 MW
FCCC_36 (MCFC+ST) 2.2 MW
Q40 15.0 MW
LD_40_5
W
FIGURE 8.18
10.0 MW
Q5
15.0 MW
P-graph solution for the energy system of Example 8.5.
Combined Process Integration and Optimization of appliances to function without unexpected problems or major breakdowns. If these equipment units fail, the consequences can be catastrophic: contamination, smog, acid rain, injury, loss of life, production cutbacks, amassed garbage heaps, energy losses, and so on. Catastrophic failures would also entail substantial added costs. For these reasons, solid waste management is a matter of serious concern, which in some cases (waste collection in Naples, Italy) has even led to a change in government. The models that have been developed to manage waste-producing processes are of two types: optimization models deal with specific aspects of waste-related problems; in contrast, integrated waste management models focus on sustainability. The latter type can be subdivided into three main subcategories: models based on cost–benefit analysis, models based on life-cycle inventory, and multicriteria models (Morrissey and Browne, 2004). However, there is an element of uncertainty or risk associated with most environmental decisions. Multicriteria techniques can be extended to consider reliability issues along the entire waste management chain and need not be limited to comparing the environmental impacts of different waste treatment methods. As the complexity of unit arrangements increases, risk assessment becomes more complicated. Risk is a measure of the plant’s ability to carry out its specific operating mission reliably. The expected return on related investments is a function of the plant equipment’s capacity, which is defined in terms of reliability, availability, durability, and performance. Reliability engineering in waste management addresses all aspects of the waste life cycle, from its collection and treatment processes and through the energy generation lifetime, including maintenance support and availability. The concepts of reliability, maintainability, and availability can be quantified with the aid of reliability engineering principles and life data analysis (Kececioglu, 2002). A significant fraction of any system’s operating cost is due to unplanned system stoppages for unscheduled repair of components or the entire system. One method of mitigating the cost (and impact) of such failures is to improve the system’s reliability and availability. Of course, improvements in reliability that are made by the supplier early in the equipment’s life cycle may well result in additional development cost being passed on to the customer in the form of higher equipment acquisition cost. However, this cost increase can be more than offset by the operational cost reduction associated with improved reliability and increased uptime, which also improve productivity. Note that, in the context of waste management, reliability, availability, and maintenance have specialized meanings. Reliability is the probability that a system will perform satisfactorily for at least a given period of time t when used under stated conditions (Kuo and Zuo, 2003).
187
188
Chapter Eight Availability is the probability of the successful operation of a system in a determined period of time. It can be calculated as the ratio between an equipment’s lifetime and total time between failures (de Castro and Cavalca, 2006): A
Lifetime Total time
Lifetime Lifetime Repair time
MTBF MTBF MTTR
(8.1)
where MTBF is the Mean Time Between Failure (the inverse of the failure rate) and MTTR is the Mean Time To Repair (the inverse of the repair rate). In addition, there are three frequently used terms defined by Ireson, Coombs, and Moss (1996) and elsewhere: inherent, achieved, and operational availability. The expression for inherent availability is Ai
MTBF MTBF MTTR
(8.2)
MTBM MTBM MAMT
(8.3)
for achieved availability, Aa
and for operational availability, Ao
MTBM MTBM MDT
(8.4)
Further specifications also exist—for example, those given by Hosford (1960): • Pointwise availability is the probability that a system will be able to operate within tolerances at a given instant of time. • Interval availability is the expected fraction of a given interval of time that a system will be able to operate within tolerances. • A special type of interval availability, called limiting interval availability, was defined by Barlow and Proschan (1996); it is the expected fraction of time in the long run that a system operates satisfactorily. There are even more specific definitions. For example, the availability of a redundant system represented by a series of parallel systems is formulated by de Castro and Cavalca (2006) as n
As
[1 (1 A ) i
i 1
yi
]
(8.5)
Combined Process Integration and Optimization where Ai is the availability of the components of subsystem i and yi is the number of redundant components in subsystem i. Comparing downtimes is another, intuitive way to express availability. Maintenance covers those activities undertaken after a system is in the field in order to keep it operational or restore it to operational condition after a failure has occurred (Ireson, Coombs, and Moss, 1996). There are several classifications of maintenance, the most important of which are listed as follows: • Breakdown maintenance: An item of the system would be repaired each time it breaks down (Mechefske and Wang, 2003). • Condition-based maintenance (CBM): The critical components are monitored for deterioration, and maintenance is carried out just before the failure occurs (Mechefske and Wang, 2003). • Preventive (scheduled) maintenance: The plant is stopped at intervals, often annually, and is partly stripped and inspected for faults (Mechefske and Wang, 2003). • Reliability-centered maintenance (RCM): A procedure to identify preventive maintenance requirements of complex systems (Cheng et al., 2008). Maintainability is the measure of an item’s ability to be retained in (or restored to) a specified condition when maintenance is performed by personnel having specified skill levels, using prescribed procedures and resources, at each prescribed level of maintenance and repair (Ireson, Coombs, and Moss, 1996). De Castro and Cavalca (2006) defined it as the ability to renew a system or component in a determined period of time, enabling it to continue performing its designed functions. For further information on reliability, availability, and maintenance of waste management systems, see Sikos and Klemeš (2010a). Another difficulty with large systems is that troubleshooting usually requires several problems to be solved, often simultaneously. Data collection is also difficult because of the variety of input data; the characteristics (e.g., type) of waste; production that changes seasonally, weekly, and also randomly or unpredictably (due, e.g., to weather changes, price changes that lead to different consumption priorities, unpredictable natural disasters such as volcanic eruptions); and changes related to a special venue, where a gathering or migration of a mass of people can cause substantial changes as at a football match or rock concert. The associated danger and hazards mean that waste materials have to be appropriately cared for. A wide variety of possible failure causes have to be identified. There are many other issues to consider, too. Waste management systems are quite complex,
189
190
Chapter Eight containing both serial and parallel subsystems. They include several equipment units, each with its own degree of availability and reliability. Scheduled outages need to be differentiated from the downtimes caused by unexpected faults. Maintenance actions— including scheduled outages—should be updated as needed to cope with changes that occur within the system. The recommendations of reliability engineers inform the designer’s task of effectively modeling and optimizing all the specified objectives. Many things need to be improved in waste management. These include policy focus and popular opinion, which should both pay more attention to the importance of this process and its effects. Landfill space is decreasing while solid waste is increasing, so public and private landfills compete for municipal clients to ensure the capital required to extend landfill life while coping with new permits. The situation varies in different countries and geographic regions, so investigation is needed to devise appropriate solutions for specific scenarios.
8.6 Summary This chapter presented several examples of combined PI and optimization. The main focus was on exploiting the advantages of graph-theoretic (P-graph and/or S-graph) frameworks. These methods are well tested and have demonstrated their efficiency across many applications.
CHAPTER
9
Software Tools 9.1 Overview of Available Tools Process Integration, modeling, and optimization problems in chemical engineering are complex in terms of scale and relationships. Solving these problems requires the application of information technology and computer software, which provide fast and as much as possible accurate solutions via a user-friendly interface. Software tools have been widely used for process simulation, integration, and optimization, and this has helped process industry companies achieve their operational goals. There is a large number of efficient tools available, each with its particular advantages. The aim of this chapter is to describe these tools based on comprehensive experience with them and their application to Process Integration, modeling, and optimization. The online encyclopedia Wikipedia (2009) presents a comprehensive list of available software tools for the simulation of material and energy balances of chemical processing plants. Examples include: (1) ASCEND; (2) Aspen HYSYS by Aspen Technology; (3) ASSETT and D-SPICE by Kongsberg Process Simulation; (4) CHEMCAD; (5) COCO simulator; (6) Design II by Winsim; (7) EcosimPro; (8) Environment for Modeling, Simulation and Optimization (EMSO); (9) Dymola; (10) GIBBSim; (11) gPROMS by PSE Ltd; (12) OLGA by SPT Group (Scandpower); (13) Omega by Yokogawa; (14) OpenModelica; (15) Petro-SIM; (16) ProMax; (17) SimSci-Esscor DYNSIM & PRO/II by Invensys; (18) SysCAD; (19) UniSim Design & Shadow Plant by Honeywell; and (20) VMGSim. However, this selection cannot be fully comprehensive and is limited just to process simulation software. It should cover a wider field and be studied in more detail, which is the target of this chapter.
9.2
Graph-Based Process Optimization Tools 9.2.1
Process Network Synthesis Solutions
Process network synthesis (PNS) Solutions is a software package designed to solve problems in PNS. Process synthesis involves
191
192
Chapter Nine finding the optimal structure of a process system, and this includes determining optimal types, configurations, and capacities of the units that perform various operations within the system. The details of the process graph (P-graph) approach were presented in Chapters 7 and 8. Process network synthesis is sometimes called flowsheeting or flowsheet optimization because it involves the creation of a flowsheet for the industrial process under consideration. In order to solve a PNS problem, the designer must examine all feasible structures and select the best among them. The structure’s optimality can be assessed in terms of cost, profit, efficiency, and so on. When designing an optimal process network, both structural information (which processing units are connected, and how) and sizing information (how much is produced from a given material) are needed. The questions addressed by PNS Solutions are as follows: (1) How are the building blocks of a process network best represented? (2) What are the possible solution structures of the problem? (3) What is the maximal structure (which includes all solution structures)? (4) What is the optimal structure? The maximal structure comprises all the combinatorially feasible structures capable of yielding the specified products from the specified raw materials. Certainly, the optimal network or structure is among these feasible structures. The maximal structure generation (MSG) algorithm produces a P-graph (see Figure 9.1) in which each
FIGURE 9.1
Starting state of the MSG algorithm (PNS Solutions).
Sof tware Tools material and operating unit appears exactly once. In the composition phase, the nodes are linked stepwise, layer by layer, starting from the shallowest end (i.e., the final-product end) of the remaining input structure. The algorithm proceeds by assessing whether any of the linked nodes violates any of the axioms described in Chapter 7. The structure generation is performed transparently, and the maximum structure that results is the input for the Solution Structures Generation (SSG) algorithm. If a material has to be produced, then the SSG algorithm generates all possible ways for its production. For example, if M1 can be produced by O2 or by O3 then the possibilities include production by O2 alone or by O3 alone or by using both. Once an operating unit is included, its input materials must be produced as well, and so forth. Materials are selected in a specific order. The parent–child relation between steps ensures that the materials are selected by the process according to this order. The SSG algorithm yields all the combinatorially feasible solution structures of a given problem. Unfortunately, the number of feasible structures at this stage is often too large to be enumerated explicitly. Therefore, the Accelerated Branch-and-Bound (ABB) algorithm is used to determine the optimal structure without generating all the possible solutions. Input to this algorithm includes not only the structural relationships between materials and operating units but also such additional information as the costs of each raw material, the fixed and proportional costs of operating units, and the constraints (if any) on the quantity of materials and the capacity of operating units.
9.2.2 S-Graph Studio S-Graph Studio is a software package that enables the user to design batch processes and to optimize them via various optimization methods (S-Graph, 2009). The program also allows scheduling problems to be defined using graphical tools. It has a modular architecture, so different solvers can be used with the program. S-Graph Studio uses the industry’s standard file format: BatchML (as defined by the ISA-88 standard of the World Batch Forum), which is used to exchange information between industry sites and plants as well as for other purposes. The Excel file format can also be used for both input and output. The software includes a solver that utilizes the S-graph methodology developed at the University of Pannonia (S-Graph, 2009). One of the main goals of batch process optimization is to minimize makespan—that is, finding the shortest time in which a process can be completed using available resources. S-Graph Studio can be used to define batch processes in terms of the tasks to be performed, the available equipment units, and task completion times (as a function of the equipment units used). This information is necessary and sufficient for minimizing the makespan of a process.
193
194
Chapter Nine The results are presented graphically (by Gantt chart or by schedule graph) but can be exported to various formats for further use. Issues addressed by S-Graph Studio include how to represent a scheduling problem and how to generate the optimal schedule structure in the cases of unlimited and no intermediate storage. The schedule associated with the selected solution is displayed as a Gantt chart (Figure 9.2) or a schedule graph, both of which clearly show the unit task assignments and their associated timing.
FIGURE 9.2
The Gantt chart and schedule graph of a solution (S-Graph Studio).
Sof tware Tools
9.3 Heat Integration Tools 9.3.1
SPRINT
SPRINT is the software package used to design energy systems for individual processes on a site (SPRINT, 2009). This software tool provides energy targets and optimizes the choice of utilities for an individual process. It also performs Heat Exchanger Network (HEN) design automatically for the utilities that are selected. Both new design and retrofit can be carried out automatically, though the designer maintains control over network complexity. For example, several retrofit modifications can be generated automatically and then presented to the designer one at a time, so that the number of modifications is minimized, and the final decision on each modification is left to the designer. SPRINT can also be used for HEN operational optimization tasks. The SPRINT and the STAR programs (see Section 9.3.3) are linked by common data structures, which facilitate their interaction (e.g., the same files can be used, and no manual data transfer is required between the programs). SPRINT is used in both academic and industrial settings for the following applications: (1) optimizing the choice and load of utilities for individual processes; (2) automatic design of new HENs; (3) automatic retrofit design of HENs while using a minimum number of modifications; (4) automatic design for multiple utilities (new design and retrofit); (5) interactive network design; (6) simulation of networks via simple models; (7) targeting minimum energy consumption; (8) network optimization; and (9) assessing network operability. User interaction with the network structure is through the graphical editor shown in Figure 9.3.
9.3.2
HEAT-int
HEAT-int (2009) is a product of Process Integration Ltd. This program is used to improve the energy performance of individual processes on a site. It is the next-generation development of SPRINT software (by a related team of developers) to a commercial standard, and it offers more user-friendly interface features. The areas in which HEAT-int is applied are similar to those listed previously for SPRINT.
9.3.3
STAR
STAR is a software package for the design of site utility and cogeneration systems (STAR, 2009); see Figure 9.4 for an example user interface. This program analyzes all the interactions between site processes and the steam system, steam turbines, gas turbines (with auxiliary firing options), the boiler house, local fired heaters, and cooling systems. The analysis is used to reduce energy cost and
195
196
Chapter Nine
FIGURE 9.3
SPRINT software interface (SPRINT, 2009).
FIGURE 9.4
STAR’s graphical user interface (STAR, 2009).
Sof tware Tools to plan infrastructure investment when operational changes are anticipated, changes that may include the replacement of energy equipment. STAR can also be used to investigate flue gas emissions, which should often be reduced to meet tighter environmental regulations. The STAR package incorporates several tools, as described next. Utility system optimization: A given utility system configuration incorporates important degrees of freedom for optimization. Multiple boilers with different efficiencies and different fuels in addition to multiple back-pressure steam turbines, condensing turbines, gas turbine heat recovery steam generators, and letdown valves provide optional heat flow paths that can all be exploited for significant cost reduction. STAR has a utility system optimization facility that allows existing utility systems to be optimized. It can also be used to plan infrastructure de-bottlenecking and investment strategies. Top-level analysis: When studying an existing site, it is important to understand how its infrastructure influences the degrees of freedom to make changes as well as the economic consequences of those changes. These considerations are addressed by STAR’s toplevel analysis, whose results ensure that the designer does not waste time and costs pursuing changes that are not viable (structurally or economically) in the overall site context. Process energy targets: Even though the primary function of STAR is the analysis of utility systems, it includes tools for setting energy targets and selecting utilities for individual processes. Using these tools allows the picture of the Total Site to be built up from the individual processes within STAR. Total Sites: STAR can produce profiles that represent the heating and cooling requirements of the Total Site. This allows targets to be set for fuel consumption in the boilers, cogeneration potential, and energy costs. The Site Profiles can be based either on the full heat recovery data or, more simply, on data for the utility exchangers only. Boiler systems and steam turbine systems: Using STAR allows the designer to establish optimal targets for the amount of steam generated by boilers and gas turbines (with auxiliary firing options). A gas turbine model enables the study of different gas turbine arrangements. Steam turbines are a part of most utility systems, serving to generate power or as allocated drivers for process machines. STAR software incorporates the design of steam turbine networks and the analysis of their operability. Emissions: By relating process energy requirements to the supply of utilities, it is possible to target the amount of fuel required for the utility system. Such targets can be combined with information on the
197
198
Chapter Nine fuel and type of combustion device to generate targets for emissions of CO2, SOx, NOx, and particulates. It is then possible to explore the various options for reducing these emissions.
9.3.4
SITE-int
Like HEAT-int (see Section 9.3.2), SITE-int (2009) is a product of Process Integration Ltd. SITE-int is a state-of-the-art software package for the design, optimization, and integration of site utility systems in process industries. Its main features include methods to: (1) model and optimize site utility systems; (2) minimize operating costs for existing systems without modification; (3) target cogeneration potential; (4) optimize site steam pressures and loads; (5) minimize site energy costs through system modifications; (6) determine the true benefit from saving energy in the individual processes; (7) reduce greenhouse gas emissions from the site; and (8) create partial-load models of utility system components from plant operating data using regression functions and data reconciliation functions.
9.3.5
WORK
WORK is the software package used for the design of low-temperature (subambient) processes (WORK, 2009). Low-temperature processes require heat rejection to refrigeration systems. As a result, the operating costs for such processes are usually dominated by the cost of power to run the refrigeration system. Complex refrigeration systems—including cascade and mixed refrigerant systems—can be analyzed using WORK. For mixed refrigerants, WORK can be used to optimize refrigerant composition. The software enables the user to: (1) understand complex refrigeration systems; (2) target minimum shaft work for low-temperature cooling duties; (3) optimize the number and temperatures of refrigeration levels; (4) target minimum shaft work for cascade refrigeration systems; (5) target minimum shaft work for mixed refrigerant systems; and (6) determine the optimum composition for mixed refrigeration systems. Three of these features are discussed in more detail next. Targeting low-temperature systems: WORK can target minimum shaft work for simple and complex refrigeration cycles. Targets are based on rigorous thermodynamic calculations that are highly accurate even when compared with the results of rigorous simulation. When multiple refrigeration levels are used, trade-offs arise between temperature levels and shaft loads. Adjustments to each temperature level affects not only its own shaftwork requirement but also that of the other levels. Therefore, all levels of refrigeration must be optimized simultaneously. This task is facilitated by WORK’s extremely accurate shaftwork predictions. Simulating refrigeration systems: WORK enables the simulation of simple and complex refrigeration systems (see Figure 9.5), which may involve multiple heat levels and multiple compressors. The refrigerant
Sof tware Tools
FIGURE 9.5 Refrigeration composition options and ideal composition profiles (WORK user interface).
heat loads and temperature levels can be optimized relative to the background process in order to minimize overall shaftwork. Optimizing mixed-refrigerant systems: WORK can optimize the composition of mixed refrigerants to minimize shaftwork requirements. This goal is achieved by matching the composition of the mixed refrigerant to the cooling profile (see Figure 9.5). The software outputs a visual representation of the shaftwork losses in refrigeration cycles, including both mechanical and thermal losses.
9.3.6
HEXTRAN
HEXTRAN is a steady-state simulator that provides a view of heat transfer systems (IPS, 2009a). It is used to design new systems, monitor current systems, optimize existing operations, and solve (or prevent) heat transfer problems. The program simulates integrated processes and allows engineers to monitor the performance of individual exchangers or an entire heat transfer network. It also offers superior postprocessing displays, plotting Grand Composite Curves as well as the results from network targeting and zone analysis. HEXTRAN provides new efficiencies in all types of design and operational analysis work, such as individual exchanger and network designs, Pinch Analysis, exchanger zone analysis, split flows, area payout, and optimal cleaning cycles. HEXTRAN analyzes factors that can make the difference between profits and losses. These factors include: (1) improved process heat transfer, product yield, and quality; (2) increased energy efficiency and significantly reduced operating costs; (3) increased plant flexibility and throughput; (4) optimized cleaning schedule for exchangers; (5) optimal antifouling selection and usage; and (6) improved process designs and revamps. The HEXTRAN simulator for process heat transfer offers features that facilitate straightforward evaluations of complex design, operational, and retrofit situations; in particular, it: (1) enables the design of both simple and complex heat transfer systems that result in cost-effective, flexible processes; (2) allows the
199
200
Chapter Nine designer to retrofit existing equipment and revamp HENs to yield optimum performance; and (3) identifies cleaning incentives and predicts future performance.
9.3.7 SuperTarget SuperTarget is mainly used to improve Heat Integration in new design and retrofit projects by reducing operating costs and optimally targeting capital investment (Linnhoff March, 2009). SuperTarget is also a tool for day-to-day application by novice or occasional users, and it makes Pinch Analysis a routine part of process design. The software features an intuitive user interface that makes the technology accessible to users at all levels of expertise, and advanced tools are available for expert applications. Many of the most time-consuming tasks traditionally associated with Pinch Analysis have been partially or fully automated. SuperTarget takes data directly from most popular process simulation programs through interfaces to Aspen Plus, HYSYS, and PRO/II. Its automatic data extraction system converts raw process data into Pinch data, although the user has the option of overriding the extraction defaults. SuperTarget consists of three program modules: (1) Process is the core program, which is used to optimize energy use within a single process unit; (2) Column performs a thermal analysis of the heat distribution in distillation columns; and (3) Site is used to establish heat and power targets across a Total Site.
9.3.8 Spreadsheet-Based Tools Pinch Analysis provides a comprehensive and systematic approach to maximizing the plant energy efficiency and minimizing the use of utilities. The Pinch technique is amenable to use with commercial spreadsheets, which display a grid of rows and columns made up of multiple cells, each containing alphanumeric text or numeric values. Kemp (2007) developed an Excel spreadsheet for Pinch Analysis that incorporates targeting calculations and plots (see Figure 9.6). The main components of this spreadsheet are: (1) input of stream data; (2) calculation of Composite Curves (CCs), the problem table, energy targets, and the Pinch temperature; (3) plots of CCs and the GCC; (4) plots of the stream population over the temperature range of the problem and the basic grid diagram; and (5) tables and graphs of the variation in energy and Pinch temperature over a range of ΔTmin values. Neither area targeting nor cost targeting is included in the spreadsheet because doing so would add considerable complexity. Suitable data on heat exchanger coefficients is often lacking and most plots of cost against ΔTmin could look fairly flat, which is not always the case when the appropriate cost scale is set. However, topology can still be identified from the graphs of utility use and Pinch temperature against ΔTmin (Kemp, 2007).
Sof tware Tools
FIGURE 9.6 Spreadsheet user interface for Pinch Analysis (Kemp Pinch Spreadsheet, 2006).
9.4 Mass Integration Software: WATER WATER is a software package for the design of water systems in process industries (WATER, 2009). Water is used for a wide variety of operations for mass transfer, washing operations, steam systems, cooling systems, and so forth. WATER targets and designs for minimum water consumption by identifying opportunities for water regeneration, reuse, and recycling. The cost of effluent treatment systems is minimized through design methods that lead to distributed systems. Networks involving water use and effluent treatment are designed automatically, with the designer in full control of network complexity. This software tool is capable of handling multiple contaminants. The six principal issues addressed by WATER are described next. Water minimization: WATER minimizes freshwater use by identifying reuse opportunities. The program works using data on the water quality constraints of each operation of the process. In addition to constraints for multiple contaminants, constraints on maximum and minimum flow rate and on water losses and gains can be specified, as well as forbidden matches. Multiple sources of freshwater: It is often the case that there are a number of different sources of freshwater available, each featuring different qualities and different costs. WATER is able to optimize the use of multiple sources of freshwater.
201
202
Chapter Nine Automatic design of water reuse networks: All constraints relating to maximum and minimum flow rates, forbidden matches, and water losses or gains in individual operations can be accounted for. The designer also has control over network complexity. Regeneration of water: Once water reuse has been maximized, further reduction in water consumption can only result from regenerating wastewater. WATER enables a network designer to examine and compare the effects of regeneration reuse and regeneration recycling. Automatic design of water reuse and effluent treatment networks: WATER can automatically generate not only water reuse networks but also effluent treatment networks at a minimum cost. The design engineer maintains control over network complexity and over the relevant constraints (e.g., forbidden matches, flow-rate ranges, water losses and gains during treatment operations). Pipe work and sewer costs in network design: In addition to the capital cost associated with regeneration and treatment processes, WATER also incorporates the cost of connecting the operations that involve running new pipes and sewers. These factors are included with the freshwater cost and other capital cost when assessing trade-offs between cost and overall performance. It is important to consider pipe work and sewer cost because they have a profound effect on network structure and complexity.
9.5 Flowsheeting Simulation Packages 9.5.1
ASPEN
Aspen Technology, Inc., provides integrated software applications for a variety of industries—oil and gas, petroleum, chemicals, and pharmaceuticals—that manufacture products using chemical processes. Aspen ONE (AspenTech, 2009d) is an application suite that enables process manufacturers to implement best practices for optimizing their engineering, manufacturing, and supply chain operations. This software package addresses inefficiencies throughout the plant, resulting in significant cost savings. Aspen Plus (Aspen-Tech, 2009c) is a core element of the Aspen ONE process engineering suite. It is a process modeling tool used in conceptual design, optimization, and performance monitoring for the chemical, polymer, specialty chemical, metals and minerals, and coal power industries. Aspen Plus includes a large database of pure component and phase equilibrium data for conventional chemicals, electrolytes, solids, and polymers. This information is updated regularly using data from the U.S. National Institute of Standards and Technology. Aspen Plus is well integrated with Aspen’s software for cost analysis (AspenTech, 2009a) and heat exchanger design (AspenTech, 2009b). These software applications enable rigorous sizing and rating of
Sof tware Tools key equipment, such as heat exchangers and distillation columns, within the simulation environment.
9.5.2 HYSYS and UniSim Design The HYSYS software was initially created by Hyprotech for simulating both steady-state and dynamic processes. It includes tools that can be applied to: (1) estimating physical properties, including liquid–vapor phase equilibrium; (2) establishing heat and mass balances; (3) designing and optimizing oil and gas processes; and (4) evaluating and selecting process equipment. HYSYS technology was acquired and modified by Aspen (see Section 9.5.1) and later by Honeywell, where it is known as UniSim Design (Honeywell, 2010). Aspen HYSYS and UniSim Design are similar in terms of application and the working interface. Both include: (1) a library of physical properties of many chemical substances; (2) a set of subroutines for estimating the behavior of several types of plant equipment (heat exchangers, reactors, etc.); and (3) a graphical user interface for inputting case specifications and displaying results. Once the designer describes the process in terms of equipment units interconnected by process streams, the program solves all the equations for mass, energy, and equilibrium while taking into consideration the units’ specified design parameters. The program is built upon proven technologies that for two decades have supplied process simulation tools to the oil and gas industry. Another advantage of Aspen HYSYS and UniSim Design is their interactive and flexible process modeling, which allows engineers to design, monitor, and troubleshoot as well as to make operational improvements and perform asset management. Employing these features leads to decision making that enhances the productivity, reliability, and profitability of a processing plant’s life cycle (Ebenezer, 2005). The HYSYS “fluid package” module requires information on the characteristics of unit components and the physical properties of relevant streams. Also, accurate simulation of processes requires that appropriate thermodynamic models be selected as a framework for the simulation. Note that a process that is otherwise fully optimized in terms of equipment selection, configuration, and operation is of no use whatsoever if the process simulation was based on an incomplete or inaccurate fluid package or on an inappropriate thermodynamic model. HYSYS requires minimal input data from the user; the most important input parameters are the stream’s temperature, pressure, and flow rate (Ebenezer, 2005). The software includes an assortment of utilities that can be attached to process stream and unit operations. These tools interact with the process to provide valuable additional information. For example, the flowsheet used within the HYSYS simulation environment can be manipulated by the designer to estimate desired output.
203
204
Chapter Nine
9.5.3
gPROMS
An advanced modeling environment for process industries is offered by gPROMS (PSE, 2009), which provides advanced custom modeling capabilities within a flowsheeting environment combined with an object-oriented modeling language. The process modeling, process simulation, and optimization capabilities of gPROMS are used to generate accurate process behavior predictions and information for decision support in product and process innovation, design, and operation. Because gPROMS is a flowsheeting environment, the user can optimize complex units within the context of an entire process. The software employs synchronized graphical and text views, which makes it easy to develop, maintain, and assure the quality of the models, and archive them. It is capable of assessing all phases of the process life cycle, from laboratory experimentation to process design and detailed engineering to online operations. For modeling problems and deriving solutions, the environment provided by gPROMS ModelBuilder is employed. ModelBuilder is a flexible environment in which engineering experts can perform custom modeling, process engineers can generate graphical flowsheets, and process operators can run “execution-only” routines. Figure 9.7 shows an example user interface. In summary, gPROMS is an equationoriented modeling system used for building, validating, and executing first-principles models within a flowsheeting framework.
FIGURE 9.7 Flowsheeting environment (ModelBuilder) with dual graphical and text views (PSE, 2009).
Sof tware Tools
9.5.4 CHEMCAD The CHEMCAD software tool for simulating chemical processes includes libraries of chemical components, thermodynamic methods, and unit operations (Chemstations, 2009). Its purpose is to facilitate the simulation of steady-state continuous chemical processes from laboratory-scale experiments to full-scale operations; see Figure 9.8 for the user interface. This software package has recently been upgraded to allow for the dynamic analysis of flowsheets. It offers operability assessment, proportional integral derivative (PID) loop tuning, and operator training as well as online process control and soft-sensor functionality. Models for nonstandard unit operations can simulate the behavior of a process under varying feed rates, product rates, temperatures, pressures, and compositions. The program contains the 1500-component Design Institute for Physical Property (DIPPR) database as well as a separate, userdefined database. Components are selected by ID number, formula, synonym, or class. The program also includes routines for predicting the properties of components not included in the database. Crude oil feeds may be characterized by American Society for Testing and Materials (ASTM) or true boiling point (TBP) curves and then represented in the simulation by a series of pseudocomponents (cuts) with different boiling points. The flowsheet representing the plant layout is input graphically, and data describing each feed stream and unit operation is entered via pull-down menu options. Automatic error checking is used to preclude overspecification or missing data entries. The interactive interface also permits individual unit operations to be run separately
FIGURE 9.8
CHEMCAD software interface.
205
206
Chapter Nine from the complete flowsheet for the purpose of quick “what if” analyses (Chempute Software, 2001). Results output by the program include a full heat and material balance; thermodynamic and physical properties of all streams; component flow rates; stream temperatures, pressures, and vapor fractions; and process equipment parameters. The user may view the results graphically or obtain a printout of either summary or detailed results. CHEMCAD includes additional modules; some of them are (1) CHEMCAD-THERM for the design and rating of shell-and-tube heat exchangers (including air coolers), (2) CHEMCAD-BATCH for simulation of batch distillation processes, and (3) CHEMCAD-REACS for the dynamic simulation of stirred tank reactors. The software package also includes a subroutine, called CONVERT, which translates the process flow diagram generated by the program into a series of drawing exchange format (DXF) files for incorporation into AutoCAD software routines.
9.5.5
PRO/II
The PRO/II computer simulation system is used by process engineers in the chemical, petroleum, natural gas, solids processing, and polymer industries (IPS, 2009b). It includes a large chemical component library and multiple thermodynamic property prediction methods as well as advanced and flexible techniques for evaluating unit operations. The software tool can perform mass and energy balance calculations for modeling steady-state processes. Expert systems, extensive input processing, and error checking are included to help inexperienced users. Typically, PRO/II simulation is applied to the following tasks: (1) process design; (2) evaluating alternative plant configurations; (3) modernizing and revamping existing plants; (4) assessing, documenting, and complying with environmental regulations; (5) troubleshooting and de-bottlenecking plant processes; and (6) monitoring and improving plant yields and profitability.
9.6 General-Purpose Optimization Packages 9.6.1
GAMS
The General Algebraic Modeling System (GAMS, 2009) is a high-level modeling system for Mathematical Programming and optimization. GAMS is designed for modeling linear, nonlinear, and mixed integer optimization problems. The system is well suited to complex, largescale modeling applications, and it allows the user to build and archive large models that can later be adapted to new situations. A particular advantage of GAMS is its ability to handle large, complex, and/or unique designs that may require many revisions before an accurate model is established.
Sof tware Tools The system models problems in a natural and highly compact way. The package includes an integrated development environment and a group of integrated solvers. GAMS was the first algebraic modeling language, and it is formally similar to several common programming languages. Models are described in algebraic statements that are easy for humans and machines to read. The system is capable of handling models of many different types, so switching between model types can be done with a minimum of effort. For instance, the same data, variables, and equations can be reused for a linear and a nonlinear model by simply converting a small number of parameters to variables. GAMS also includes a variety of solvers for different classes of models.
9.6.2
MIPSYN
The MIPSYN (short for Mixed Integer Process SYNthesizer) is a user-friendly computer package for the integrated synthesis of new plants and for the innovative reconstruction of existing plants at different levels of complexity. The tasks to which it can be applied range from simple nonlinear programming (NLP) solutions for plant optimization problems to the mixed integer nonlinear programming (MINLP) optimization of heat-integrated, flexible plants. MIPSYN is the successor to the PROSYN synthesizer (Kravanja and Grossmann, 1990; Kravanja and Grossmann, 1994). As such, it is based on the most advanced modeling and optimization techniques: those rooted in disjunctive MINLP. The MIPSYN software can simultaneously address both discrete optimization (e.g., selection of process units, their operating status, their connectivity, and ranges of operation) and continuous optimization (of temperatures, flows, pressures, etc.). The package integrates the following methods and related components: (1) GAMS with a variety of different NLP and MILP solvers; (2) different versions of the outer approximation (OA) algorithm—including the modified OA/ER algorithm (“ER” denotes “equality relaxation”) and a new logicbased OA/ER algorithm—which are supervised by MIPSYN command files; (3) a simple simulator that serves as an initializer to provide NLP subproblems with feasible (or nearly feasible) starting points; (4) a library of models pertaining to process units, interconnected nodes, and the simultaneous integration of heat and mass; (5) a database of the physical properties of the most common chemical components; and (6) a hybrid modeling environment, with a link to external FORTRAN routines, for solving the implicit part of synthesis models. Execution of the NLP and MILP steps in OA algorithms is performed through the use of GAMS saving and restart capabilities, which enable the user to execute MIPSYN in automated or interactive modes of operation. The synthesizer features many important
207
208
Chapter Nine capabilities: initialization of NLP subproblems; calling different NLP and MILP solvers in a sequence with different option files (text files containing specifications of solver options to be applied); the efficient modeling of different formulations and strategies (e.g., multilevel MINLP); the capacity to solve feasibility problems whose objective functions are augmented by “penalties”; multiobjective optimization; integer-infeasible path optimization; multiperiod optimization; and flexible synthesis for cases where the true parameters are uncertain. Some of these applications were described in Kravanja (2009). MIPSYN can be comprehended and used at different levels of problem abstraction because it includes: (1) an MINLP solver for problems of a general nature; (2) a process synthesizer for generating process flowsheets; and (3) a synthesizer shell for accommodating applications from different engineering domains. A number of case studies have been performed using MIPSYN. In these studies, the synthesis was applied to all basic process systems and subsystems. Examples include: (1) heat-integrated reactor networks in overall process schemes; (2) heat-integrated and flexible separator networks; (3) Heat Exchanger Networks, including retrofits and networks that use more than one exchanger type; (4) mass exchanger networks; (5) heat-integrated overall process schemes based on a sustainable, multiobjective approach; and (6) flexible and heat-integrated flowsheets, together with their HENs, for cases involving as many as 30 uncertain parameters. Note that the MIPSYN synthesizer shell also enables applications in the area of mechanics (Kravanja, Kravanja, and Bedenik, 1998a; Kravanja, Kravanja, and Bedenik, 1998b; Kravanja, Šilih, and Kravanja, 2005). These mechanical applications range from simple NLP optimizations to complex, multilevel MINLP syntheses of structures in which topology, material use, and dimensions are optimized simultaneously.
9.6.3
LINDO
LINDO is a tool for solving linear, integer, and quadratic programming problems (Lindo Systems, 2009). It provides an interactive modeling environment that facilitates the simulation and solution of optimization problems. LINDO has the speed and capacity to solve large-scale linear and integer models. The dynamic link library (DLL) version of LINDO allows users to seamlessly integrate the LINDO solver into Microsoft Windows applications that are written in Visual Basic, C/C++, or any language that supports DLL calls. Workstation users can exploit the linkable object libraries to hook the solver engine to applications written in FORTRAN or C. The latest LINDO version (ODC, 2009) offers a number of enhancements, including: (1) significantly expanded nonlinear capabilities; (2) global optimization tools; (3) improved performance on linear and integer
Sof tware Tools problems; and (4) enhanced interfaces to other systems, such as MATLAB and Java. LINDO API was the first full-featured solver with a callable library to offer general nonlinear and integer nonlinear capabilities. This feature allows developers to incorporate a single, generalpurpose solver into their custom applications. The software’s linear and integer capabilities provide the user with a comprehensive set of routines for formulating, solving, and modifying nonlinear models (although a separate, nonlinear license must be purchased to access these nonlinear capabilities). The global solver combines techniques for range bounding (e.g., interval analysis, convex analysis) and range reduction (e.g., linear programming, constraint propagation) within a branch-and-bound framework to find proven global solutions to nonconvex nonlinear programs or mixed-integer nonlinear programs.
9.6.4
Frontline Systems
Solvers (optimizers) are software tools that help users find the best way to allocate scarce resources. Frontline Systems (Solver.com, 2009) offers products that cover several problem types and that allow model definition via an Excel spreadsheet or via a program written in any of several common programming languages and environments (e.g., Visual Basic, C/C++/C#, VB.NET, Java, MATLAB). Models designed using these solver products include decision variables for quantities of resources as well as calculated results (constraints) that are subject to limits (e.g., budget, capacity, and/or time constraints).
9.6.5
ILOG ODM
The ILOG system is geared toward optimizing decisions and thereby finding the best solutions to complex planning and scheduling problems. It provides an application environment that supports the flexible exploration of all the trade-offs and sensitivities. Thus, the ILOG optimization decision manager (ODM) makes optimization easier (ILOG, 2006); see Figure 9.9. A unique aspect of ODM-based applications is that they provide all the features that business people need to take full advantage of optimization technology. Applications built using ILOG ODM allow users to create, visualize, and compare planning or scheduling scenarios, to adjust any of the model’s inputs or goals, and to comprehend the relevant binding constraints, tradeoffs, sensitivities, and business options. With ODM-based applications, overconstrained problems are automatically relaxed during runtime by ILOG CPLEX, which is programmed to relax the least important (and fewest possible) constraints. This ensures that a solution will always be found, and solutions are presented along with clear information about any
209
210
Chapter Nine
FIGURE 9.9
User interface for an ILOG ODM-based application (ILOG, 2006).
relaxed preferences or constraints. The optimized solution—which includes a recommended plan or schedule and the attendant metrics—can easily be further explored, allowing users to understand the optimization model’s dynamics and perhaps identify better solution scenarios.
9.7 Mathematical Modeling Suites 9.7.1
MATLAB
MATLAB (short for matrix laboratory) is an interpreted language for numerical computation (MathWorks, 2009). It allows performing numerical calculations and then visualizing the results without the need for complicated and time-consuming programming. MATLAB allows users to solve problems accurately, to produce graphics easily, and to generate code efficiently. It also enables matrix manipulation, the plotting of functions and data, implementation of algorithms, creation of user interfaces, and interfacing with programs written in other languages. For technical problem solving, MATLAB has many advantages over conventional computer languages, as described next (see also Chapman, 2009). Ease of use: Programs can be easily written and modified under the built-in integrated development environment, and they can be debugged using the MATLAB debugger.
Sof tware Tools Platform independence: MATLAB is supported on many different computer platforms, which provides a large measure of platform independence. The language is supported on Windows, Linux, several versions of UNIX, and the Macintosh. This means that programs written in MATLAB can migrate to different platforms. Predefined functions: MATLAB comes complete with an extensive library of well-tested, predefined functions that generate solutions to many basic technical tasks. The arithmetic mean, standard deviation, median, and hundreds of other mathematical functions are built in to the MATLAB language, which makes the user’s job much easier. Device-independent plotting: Unlike most other computer languages, MATLAB has many commands for imaging and integral plotting. The images and plots can be displayed on any graphical output device supported by the computer that is hosting MATLAB. Graphical user interface: MATLAB includes tools with which a programmer can interactively construct a graphical user interface for any program. Given this capability, programmers can design sophisticated data-analysis programs that can be operated by relatively inexperienced users. The built-in functions of MATLAB allow users to perform basic minimization and maximization routines. However, compiling and executing a proper optimization program may require the use of add-on packages. One popular add-on is TOMLAB (TOMLAB, 2010), a powerful optimization platform and modeling language for solving applied optimization problems in MATLAB. The TOMLAB environment includes a wide range of features, tools, and services for optimization analyses.
9.7.2
Alternatives to MATLAB
There are two free alternatives to MATLAB software: SCILAB (Scilab, 2009) and OCTAVE (Octave, 2009). Both provide numbercrunching power similar to MATLAB’s but at an advantageous cost/performance ratio (since they are free). In essence, SCILAB and OCTAVE are interpreted, matrix-based programming languages. They have strong similarities to MATLAB: (1) the use of matrices as a fundamental data type; (2) built-in support for complex numbers; (3) powerful built-in math functions and extensive function libraries; and (4) extensibility in the form of user-defined functions and macro languages.
9.8 Other Tools 9.8.1
Modelica
Modelica is an object-oriented, declarative, multidomain language for the component-oriented modeling of complex systems—that is,
211
Chapter Nine systems containing mechanical, electrical, electronic, hydraulic, thermal, control, power, and/or process-oriented subcomponents (Modelica, 2009a; Modelica, 2009b). The free Modelica language is developed by the nonprofit Modelica Association. In this language, models are described by classes, which may contain differential, algebraic, and discrete equations alongside properties and algorithms. The language can be used for “hardware in the loop” simulations and for embedded control systems (Modelica, 2009a). Modelica supports high-level modeling by composing complex models from detailed, component models. Models of standard components are typically available in model libraries. A model can be defined by using a graphical model editor offered by the various language implementations (Modelica, 2010) to draw a composition diagram: positioning icons that represent the model components, drawing connections between the components; and providing parameter values in dialogue boxes (Modelica, 2009b). Constructs for including graphical annotations in Modelica render the icons and composition diagrams portable between different platforms. Typical composition diagrams from various domains are shown in Figure 9.10. In addition to the basic language elements mentioned previously, Modelica also supports arrays (via a MATLAB-like syntax). Array elements may consist of basic data types (e.g., real, integer, boolean, string) or, more generally, of component models. This flexibility allows for convenient descriptions of complex models containing repetitive elements.
9.8.2 Emerging Trends Lead times for the development of new energy technologies—from initial idea to commercial application—can run into years. Reducing
y x
Inertial
b3={0,0,0} 3D mechanics b1={0,0,0}
b4={0,0.73,0} load r5={1,0,0}
r4={0,1,0}
r6={0,1,0} state machines b5={0,0,0}
Rp2 load
Rp1
VolP hydraulics
J=1
q R=200 La
L=L
Rel
power systems Line1
T1 p1
p2
k=k
p3 electrical systems
FIGURE 9.10
T4
C=C power trains
P4
C
Ra
1
R=250
gear
motor
T2
control
time>3
b2={0,0.5,0} control systems
time>5
b0
r3={1,0,0}
r2={1,0,0}
r1={0,1,0}
R=50
212
emf
3rd Order
Line2
G2
Composition diagrams produced using Modelica (2009b).
Sof tware Tools this lead time is a primary objective of the European Commission’s Directorate-General for Transport and Energy (DGTREN), which has funded two related projects, EMINENT (Early Market Introduction of New Energy Technologies) and EMINENT2 (Klemeš et al., 2005b). The principal features of these projects are a software tool and an integrated database of new technologies and sectoral energy supplies and demands. The software tool is for analyzing the potential impact of new and underdeveloped energy technologies emerging in different sectors from different countries. This tool has also been used in case studies that illustrate the new technologies. The aim of the EMINENT software is to assess the market potential of early-stage technologies (ESTs) in various energy supply chains by evaluating their performance in terms of: (1) CO2 emissions, (2) costs of energy supply, (3) use of primary fossil energy, and (4) effects on different subsectors of society. Technology developers and financial supporters are frequently not aware of all the potential applications or the relative market attractiveness of such technology across different countries and sectors of society. Thus, the EMINENT project provides insight into the market potential that can accelerate the development of technologies; this benefits research and development efforts by targeting them more effectively. The EMINENT tool that evaluates ESTs makes use of two databases: (1) national energy infrastructures, which contain information on the number of consumers per sector, type of demand, typical quality of the energy required, and consumption and installed capacity per end user and (2) other ESTs and technologies that are already commercial— including key information on new energy technologies currently under development and proven energy technologies now available and in use. The availability, price, and geographical conditions of primary energy resources differ significantly worldwide, so the impact of ESTs can be evaluated only within the context of a particular (national) energy supply system (Klemeš et al., 2007b). The EMINENT package consists of an integrated resource manager, a demand manager, and an EST manager as well as databases on resources and demand. The methodology, as shown in Figure 9.11, can be briefly summarized as follows: Resource manager selects, enters, and modifies data on country resources (electrical, fuel-based, geothermal, hydro, ocean tidal, wave, and wind energy). Demand manager describes energy demands per subsector in a given country; it selects data for the technology assessment, enters new data, and modifies old data. Technology manager stores key data on existing technologies and ESTs. User input: (1) The sectoral energy demands whose potential supply by EST is being evaluated; (2) any other peripheral technologies
213
214
Chapter Nine Solar
Hydro
Coal
Gas/Oil Hydrogen
Wind
Ground Heat
Biomass
Apply EMINENT tool to assess the new technology potential Energy supply system Intermediate products Biomass/waste Coal Oil
Pretreatmentsystems
Electricity Transportation and storage
Conversion processes
Natural gas Hydro Wind
EST to be evaluated
Distribution and storage
Heat
Transportation fuel
Solar energy Geothermal
FIGURE 9.11 2007b).
Methodology employed in EMINENT toolbox (after Klemeš et al.,
needed to establish full energy supply chains; and (3) resources the EST would require to satisfy the full energy supply chain. Output: (1) aggregate numbers; (2) application potential of ESTs per (sub)sector; (3) annual cost of energy delivery per consumer and per (sub)sector; and (4) annual CO2 emissions. Performance indicators: (1) efficiency of energy supply chain; (2) usage of primary fossil energy; (3) CO2 emissions per MW∙h; and (4) costs of delivered energy [€/MW∙h]. Most of the ESTs analyzed were not yet able to achieve the cost levels of existing technologies. Some of the ESTs—for instance, molten carbonate fuel cells (MCFCs)—could become competitive with relatively small efforts aimed at cost reduction. Various promising future trends located with the help of EMINENT have been supported by policy makers. Some near-term prospects are: (1) diverse energy systems that encompass supplies, management, and control of demand; (2) market-based grids with large power stations, including wind farms of different types; (3) local distributed generation, including biomass, waste, and wind; (4) micro generation,
Sof tware Tools including Combined Heat and Power (CHP), fuel cells, and photovoltaic technology; new homes built with nearly zero carbon emissions; (6) natural gas; (7) coal-fired generation combined with carbon capture and storage; and (8) mixed fuels (e.g., coal mixed with biomass, natural gas mixed with hydrogen). Longer-term prospects include nuclear power stations, fuel cells, hydrogen obtained from nonelectricity sources (e.g., biomass, low-carbon biofuels), and nuclear fusion.
9.8.3 Balancing and Flowsheeting Simulation for Energy-Saving Analysis Tools for balancing, reconciliation, and flowsheeting simulation are frequently used for energy-saving analysis and have become an essential item in the process engineer’s toolbox. These tools help designers develop complete mass and energy models based on actual measurements and/or design values and mathematical models. As a result, these simulation tools play an important role in the technical and economic decision making related to the planning and design stages of processes under development and to the operation of existing equipment. Several computer-based systems have been developed over the years in order to assist process engineers with energy and mass balance calculations. However, ongoing development costs have resulted in a limited number remaining on the market, and these have been secured only by a substantial volume of sales. An early overview of the field was presented by Klemeš (1977). The technology used for balancing and for data validation and reconciliation consists of a set of procedures incorporated into a software tool. Process data reconciliation has become the main method for monitoring and optimizing industrial processes as well as for performing component diagnosis, condition-based maintenance, and online calibration of instrumentation. According to Heyen and Kalitventzeff (2007), reconciliation has three main goals: (1) detect and correct deviations and errors of measurement data so that all balance constraints are satisfied; (2) exploit the structure of and knowledge about the process system by using measured data to estimate unmeasured data whenever possible, in particular as regards key performance indicators (KPIs); and (3) determine the postprocessing accuracy of measured and unmeasured data, including KPIs. A comprehensive system called DEBIL—which included balancing, flowsheeting simulation, and optimization—was created several decades ago (see Klemeš, Luťcha, and Vašek, 1979) and has since been further developed by Belsim into the balancing and reconciliation tool VALI (VALI III User Guide, 2003). This tool has been applied to various energy-efficiency tasks, as reported with respect to nuclear power plants (Langenstein, Jansky, and Laipple, 2004) and regenerative heat exchangers (Minet et al., 2001).
215
216
Chapter Nine
9.8.4 Integrating Renewable Energy into Other Energy Systems There are many different computer tools now available that can be used to analyze the integration of renewable energy. Connolly and colleagues (2010) have reviewed these tools in considerable detail. Their study analyzed 37 computer tools designed for this purpose, as summarized (in alphabetical order) in Table 9.1.
Tool (availability)
Applications
Reference
AEOLIUS (commercial)
Power plant dispatch simulation tool
Universität Karlsruhe (2009)
BALMOREL (free download)
Open-source electricity and district heating tool
Balmorel (2009)
BCHP Screening Tool (free download)
Assesses CHP in buildings
Oak Ridge National Laboratory (2009a)
COMPOSE (free download)
Single-project technoeconomic assessments
Aalborg University (2009a)
E4cast (commercial)
Tool for energy projection, production, and trade
ABARE (2009)
EMCAS (commercial)
Creates techno-economic models of the electricity sector
Argonne National Laboratory (2009a)
EMINENT (registration required)
Early-stage technologies assessment
EMINENT2 (2009)
EMPS (commercial)
Electricity systems with thermal/hydro generators
SINTEF (2009b)
EnergyPLAN (free download)
User-friendly analysis of national energy systems
Aalborg University (2009b)
energyPRO (commercial)
Single-project technoeconomic assessments
EMD International A/S (2009)
ENPEP-BALANCE (free download)
Market-based energysystem tool
Argonne National Laboratory (2009b)
GTMax (commercial)
Simulates electricity generation and flows
Argonne National Laboratory (2009c)
H2RES (internal use only)
Energy-balancing models for island energy systems
Instituto Superior Técnico (2009)
HOMER (free download)
Techno-economic optimization for standalone systems
HOMER (2009)
TABLE 9.1
Software Tools for Analyzing the Integration of Renewable Energy into Other Energy Systems (after Connolly et al., 2010)
Sof tware Tools Tool (availability)
Applications
Reference
HYDROGEMS (free for TRNSYS users)
Renewable and H2 standalone systems
Institute for Energy Technology (2009)
IKARUS (earlier versions are free)
Bottom-up cost optimization tool for national systems
Forschungszentrum Jülich Institute for Energy Research (2009)
INFORSE (free to nongovernmental organizations)
Energy-balancing models for national energy systems
INFORSE-Europe (2009)
Invert (free download)
Simulates promotion schemes for renewable energy
Vienna University of Technology (2009)
LEAP (free for developing countries and for students)
User-friendly analysis for national energy systems
Stockholm Environment Institute (2009)
MARKAL/TIMES (commercial)
Energy–economic tools for national energy systems
International Energy Agency (2009)
MESAP PlaNet (commercial)
Linear network models of national energy systems
Schlenzig (2009)
MESSAGE (free except for the simulators)
Medium- and long-term assessment of national or global energy systems
International Institute for Applied Systems Analysis (2009)
MiniCAM (free download)
Simulates long-term, large-scale global changes
MiniCAM (2009)
NEMS (free except for the simulators)
Simulates the U.S. energy market
Energy Information Administration (2009)
ORCED (free download)
Simulates regional electricity dispatch
Oak Ridge National Laboratory (2009b)
PERSEUS (sold to large European utilities)
Family of energy and material flow tools
Universität Karlsruhe (2009)
PRIMES (projects are undertaken for a fee)
Market equilibrium tool for energy supply and demand
National Technical University of Athens (2009)
ProdRisk (commercial)
Optimizes operation of hydro power
SINTEF (2009a)
RAMSES (projects are undertaken for a fee)
Simulates the electricity and district heating sector
Danish Energy Agency (2009)
RETScreen (free download)
Renewable analysis for electricity and/or heat in systems of any size
National Resources Canada (2009)
TABLE 9.1
Software Tools for Analyzing the Integration of Renewable Energy into Other Energy Systems (Continued)
217
218
Chapter Nine
Tool (availability)
Applications
Reference
SimREN (projects are undertaken for a fee)
Bottom-up supply and demand for national energy systems
Institute for Sustainable Solutions and Innovations (2009)
SIVAEL (free download)
Electricity and district heating sector tool
Energinet.dk (2009)
STREAM (free download)
Overview of national energy systems to create scenarios
Ea Energy Analyses (2009)
TRNSYS16 (commercial)
Modular structured models for community energy systems
TRNSYS (2009)
UniSyD3.0 (contact
[email protected])
National energy systems scenario tool
Unitec New Zealand (2009)
WASP (free to IAEA member states)
Identifies the least-cost expansion of power plants
IAEA (2009)
WILMAR Planning Tool (commercial)
Increasing the use of wind in national energy systems
Risø National Laboratory (2009)
TABLE 9.1
Software Tools for Analyzing the Integration of Renewable Energy into Other Energy Systems (Continued)
CHAPTER
10
Examples and Case Studies
T
his chapter provides a selection of problems that have been collected by the authors over twenty years of teaching and consulting. The problems include step-by-step solutions, which provide guidance for mastering the methodology. Space limitations make it impossible to cover all aspects or to provide fully comprehensive solutions, for which an entire book would be required. The representative problems selected for presentation here are divided into five groups: Heat Pinch Technology, Total Sites, integrated placement of processing units, utility placement, and Water Pinch Technology.
10.1 Heat Pinch Technology The first group of problems involves heat Process Integration, known also as Heat (and, in Australia and South Africa, as Thermal) Pinch Technology.
10.1.1 Heat Pinch Technology: First Problem Problem 1: Task Assignment A chemical process is served by a Heat Exchanger Network (HEN), as displayed in Figure 10.1. The figure shows the supply and the target temperatures at the start and end of each stream. For this problem, the minimum allowed temperature difference ΔTmin = 10°C. The following tasks should be performed: (a) Completion of the network description. (1) Convert the network representation to a grid diagram. (2) Calculate the missing temperatures and heat loads for the recovery heat exchangers, heaters, and coolers. (b) Identification of the Network Pinch. (1) Find and write down a path that connects a heater and a cooler through a recovery heat exchanger. Derive expressions for the heat loads and
219
220
Chapter Ten 140°C
164°C 125°C
170°C
300°C H
H CP [kW/°C] 100 H1
190°C
337°C
C
160°C
160 H2 220°C
50 H3
190 H4
220°C
C
C
100°C
FIGURE 10.1
60°C
45°C
160°C
CP [kW/°C]
40°C
35°C
80°C
60°C
140°C
C1
C2
C3
C4
C5
100
70
175
60
200
Existing Heat Exchanger Network (Problem 1).
temperatures that will vary with shifting amounts of heat X [kW] through the path in order to increase the heat recovery. (2) Shift the maximum amount of load through the path while accounting for the minimum allowed temperature difference. Write down the maximum shifted load, the pinching exchanger, and the stream temperatures for the exchanger. Using the Pinch method yields the Maximum Energy Recovery (MER) targets listed in Table 10.1 and the Grand Composite Curve (GCC) shown in Figure 10.2. (c) Identification of the scope for improvement. (1) Calculate the scope for improvement in heat recovery in terms of the network’s total heating requirement. (2) Find the heat exchangers, implementing cross-Pinch heat transfer, and write them down. Problem 1: Solutions Answer to (a)(1) and (a)(2). The HEN is represented as a grid diagram in Figure 10.3, which also shows the missing HEN parameters (i.e., temperatures and loads). Answer to (b)(1). A path between a cold and a hot utility is called a utility path. Heat duty can be shifted along a utility path, which provides a degree of freedom in the HEN retrofit. Figure 10.4 shows a utility path connecting a heater and a
Examples and Case Studies Interval
T * [°C]
Enthalpy [MW]
1
332
17,780
2
305
20,480
3
215
2,480
4
175
2,880
5
169
2,580
6
155
900
7
145
0 (Pinch)
8
130
1,650
9
105
25
10
85
725
11
65
4,925
12
55
7,625
13
40
10,925
14
35
11,425
TABLE 10.1
Calculated MER Heat Cascade for Problem 1
ΔTmin = 10°C 350 300
T* [°C]
250 200 150 100 50 0 0.0
5.0
10.0 ΔH [MW]
FIGURE 10.2
Grand Composite Curve for Problem 1.
15.0
20.0
221
222
Chapter Ten 99.7°
40.0°
190.0° 337.0° E2 E1
C3 5970 kW
160.0°
220.0°
E4 60.0°
88.0° C6
45.0°
C1 C2 C3 C4 C5
220.0°
E5 118.55° E7 C10
160.0°
H1
100
H2
160
H3
50
H4
190
300.0° H8 14700 kW 5300 kW 164.0° 247.0°
100.0° 35.0°
9030 kW
80.0°
125.0°
7875 kW
60.0°
170.0°
6600 kW
140.0°
300.0°
188.0°
H9
9600 kW
FIGURE 10.3
CP [kW/°C]
100 70 175 60 200
22400 kW
Calculated grid diagram for Problem 1.
Load-shift path CP [kW/°C] 40.0°
C3
E2
TH11
E1
H1
100
220.0°
E4
H2
160
H3
50
88.0° 220.0°
E5
C6
1400 kW 118.55° 45.0° C10
160.0°
E7
13984 kW
100.0° C1
QE1 C2
337.0°
QC3
160.0° 60.0°
TH12
TC11 H8 QH8
80.0°
9030 kW
C3 7875 kW
C4 C5
140.0°
6600 kW
188.0° 9600 kW
FIGURE 10.4
300.0°
TC2,out=164.0°
TC2,in=35.0°°
60.0°
190
H4
H9 22400 kW
The load-shift path identified for Problem 1.
100 70
125.0°
175
170.0°
60
300.0°
200
Examples and Case Studies cooler through a heat recovery exchanger. The temperatures, which will vary with the amount of heat shifted, are also indicated in the figure. The expressions for the variation in the affected heat loads and temperatures in response to a shifted load of X kW are displayed in Eqs. (10.1)–(10.8) (notation as in Figure 10.4): Cooler C3 load: QC3 = 5970 - X
(10.1)
Temperature of stream C1: TC11 = 247 + X /100
(10.2)
Heater H8 load: QH8 = 5300 - X
(10.3)
Exchanger E1 heat load: EE1 = 14700 + X
(10.4)
Temperature 1 of stream H1: TH11 = 190 - X /100
(10.5)
Output temperature of C2: TC2,out = 164
(10.6)
Input temperature of C2: TC2,in = 35
(10.7)
Temperature 1 of stream H1: 99.7 X /100
(10.8)
Answer to (b)(2). The temperature differences should be checked at the affected heat exchangers, which are E1 and E2. As can be seen, the heat capacity flow rate (CP) values for both streams in E1 are equal to 100 kW/°C. Exchanger E2’s hot stream has a higher CP value than that for the cold stream, so the smaller temperature difference will be at its hot end. This allows one to calculate the maximum amount of the load X to shift by solving a few inequalities. For process exchanger E2, the temperature difference should not be less than ΔTmin, and this determines the value of the maximum heat load that can be shifted: TH11 – TC2,out ≥ ∆Tmin
(10.9)
190 – X/100 – 164 ≥ 10
(10.10)
X ≤ 2600 kW
(10.11)
For this value of load shift, the exchanger E2 will be pinched at its hot end as follows: TH11 = 174.0°C and TC2,out = 164.0°C. The temperatures at the cold end will be TH12 = 83.7°C and TC2, in = 35.0°C. Answer to (c)(1). Total heating requirement for the existing network: Σ Q H = Q H8 + Q H9 = 27,700 kW. Total cooling requirement for the network: Σ Q C = Q C3 + Q C6 + Q C10 = 21,345 kW. There is no inappropriate use of utilities. Compared with the targets, both total utility heating and total utility cooling are higher by 9920 kW, which is due to cross-Pinch heat transfer. Answer to (c)(2). The Process Pinch is at 150°C for hot streams and at 140°C for cold streams. Comparing the network temperatures with the Pinch location yields a list of exchangers that violate the Pinch; see Table 10.2.
223
224
Chapter Ten Exchanger
Hot side violation
Cold side violation
E1
No
Yes
E2
Yes
Yes
E4
No
No
E5
Yes
Yes
E7
Yes
No
TABLE 10.2
Exchangers That Violate the Pinch (Problem 1)
Stream
Tin [°C]
Tout [°C]
CP [kW/°C]
1
Cold
38
205
11
2
Cold
66
182
13
3
Cold
93
205
13
4
Hot
249
121
17
5
Hot
305
66
13
TABLE 10.3
Process Stream Data for Problem 2
10.1.2 Heat Pinch Technology: Second Problem Problem 2: Task Assignment The stream data for a process are as listed in Table 10.3. The hot utility is steam at 200°C, the cold utility is water at 38°C, and ΔTmin = 24°C. (a) Plot the Composite Curves (CCs) for this process. (b) Determine Q H,min, QC,min, and the Pinch temperatures. (c) Assuming that the cost of cooling water and steam is 18.13 and 37.78 $ kW−1 y−1, plot the minimum annual cost for the utilities as a function of ΔTmin in the range 51–54°C (in 1° increments). (d) Design a network that features a minimum number of units and maximum energy recovery for ΔTmin = 24°C. Problem 2: Solutions Answer to (a). The Composite Curves for the process stream data are shown in Figure 10.5. Answer to (b). The position of the CCs in Figure 10.5 indicates that, for ΔTmin = 24°C, this is a threshold problem at the cold end: Q H,min = 0 kW. On the cold end there is excess of hot streams, which means that some cold utility is required: Q C,min = 482 kW.
Examples and Case Studies FIGURE 10.5 Composite Curves for Problem 2.
T [°C]
ΔTmin = 24°C
No heating utility
200
100 482 kW 0 0
2000
4000
ΔH [kW]
Utility Cost [$/y] 12000 11000 10000 9000 8000 51
FIGURE 10.6 utility cost.
52
54 ΔTmin [°C]
53
Minimum allowed temperature difference ΔTmin versus annual
ΔTmin [°C] 51
52
53
54
QH [kW]
0
0
27
57
Cost of heating [$/y]
0
0
1,020
2,153
QC [kW]
482
482
509
539
Cost of cooling [$/y]
8,739
8,739
9,228
9,772
Total utility cost [$/y]
8,739
8,739
10,248
11,926
TABLE 10.4
Utility Requirements and Cost for Various ΔTmin Values
Answer to (c). The utility cost plot, which is given in Figure 10.6, is based on the data in Table 10.4. Answer to (d). A network that features the minimum number of units and maximum energy recovery is shown in Figure 10.7.
225
226
Chapter Ten CP [kW/°C]
(a) 121°C
163.4°C
249°C
E4
E1 E2
66°C
103.1°C
C 482 kW
1
305°C
E3
38°C
4
17
CP=9.1 kW/°C
CP=3.9 kW/°C
5
205°C
E2
13
11
1837 kW 2
66°C
121.43°C 182°C E3 E4 720 kW 93°C
3
13
788 kW E1
205°C
13
1456 kW One possible design CP [kW/°C]
(b) E3
E1
121°C
66°C
C
103.1°C
CP=5.22 kW/°C
E2
E4
249°C
CP=11.78 kW/°C
193°C
17 4 305°C
E1
5
13
482 kW 1
38°C
E4
144.3°C
1169 kW 2
66°C
3
205°C
E3 668.16 kW E2
182°C
13
1508 kW
93°C
11
E1
205°C 13
1456 kW A second possible design
FIGURE 10.7
10.2
Optimal Heat Exchanger Network.
Total Sites 10.2.1 Total Sites: First Problem Problem 3: Task Assignment Suppose a Total Site incorporates two processes, A and B. The stream data for process A (assuming ΔTA,min = 10°C) are given in Table 10.5, and the GCC for this process is shown in Figure 10.8. The stream data for process B (assuming ΔTB,min = 10°C) are shown in Table 10.6.
Examples and Case Studies Stream name
Supply temp. [°C]
Target temp. [°C]
ΔH [kW]
A1
50
140
450
A2
100
30
420
A3
100
140
80
TABLE 10.5 Data for Process A (Problem 3)
ΔTmin = 10°C
(330;145)
T* [°C]
150 Sink A1
Sink A2 (59;105)
100
(0;95) Source A1
50 20
Source A2
(40;55)
(220;25) 0
100
200
300
ΔH [kW]
FIGURE 10.8
Grand Composite Curve for process A (Problem 3).
Stream name
Supply temp. [°C]
Target temp. [°C]
CP [kW/°C]
B1
190
120
6
B2
100
240
4
B3
80
60
2
TABLE 10.6
Data for Process B (Problem 3)
High-pressure (HP) steam from a boiler is available at 240°C saturation temperature, and cooling water is available at 20–30°C. It is necessary to analyze these two processes (and hence the Total Site) as follows: (a) Find the Pinch temperatures and minimum utility demands for processes A and B. (b) Construct the GCC for process B.
227
228
Chapter Ten (c) Construct the Total Site Profiles (TSPs). (d) Assuming that the site manager is willing to accept just one more steam main in addition to the HP, identify this steam level on the Site Profiles. What saturation temperature will yield the most energy savings? (e) Determine the load target for this steam level. Problem 3: Solutions Answer to (a). For process A, ΔTmin = 10°C, the minimum hot utility is 330 kW, and the minimum cold utility is 220 kW (see Table 10.7). For process B, ΔTmin = 10oC, the minimum hot utility is 240 kW, and the minimum cold utility is 140 kW (see Table 10.8). Answer to (b). Given the data in Table 10.8, the GCC for process B is drawn as shown in Figure 10.9. Answer to (c). Using the GCC plots from Figures 10.8 and 10.9, the heat source and heat sink segments have been extracted. These segments are listed in Tables 10.9 and 10.10 for process A and process B, respectively. The data reported in Tables 10.9 and 10.10 have been combined into composite heat source and sink profiles. Figures 10.10 and 10.11 illustrate the composition procedure, which is analogous to the one for constructing the process CCs. Figure 10.12 shows the resulting TSPs.
Interval
Temperature [°C]
Enthalpy [kW]
1
145
330
2
105
50
3
95
0 (Pinch)
4
55
40
5
25
220
TABLE 10.7
Interval
Problem Table for Process A (Problem 3(a))
Temperature [°C]
Enthalpy [kW]
1
245
240
2
185
0 (Pinch)
3
115
140
4
105
100
5
75
100
6
55
140
TABLE 10.8
Problem Table for Process B (Problem 3(a))
Examples and Case Studies ΔTmin = 10°C
(240;245)
Sink B1 (0;185) 200 T* [°C]
Source B1 (100;135) (140;115)
(100;105) 100
(100;75)
Source B2 (140;55)
40 0
50
100
150
200
250
ΔH [kW]
FIGURE 10.9
Segment Sink A1
Grand Composite Curve for Process B (Problem 3(b)).
T *start [°C]
T *end [°C]
ΔH [kW]
T **start [°C]
T **end [°C]
280
110
150
105
145
Sink A2
95
105
50
100
110
Source A1
95
55
40
90
50
Source A2
55
25
180
50
20
TABLE 10.9
Heat Source and Sink Segments from the GCC for Process A [Problem 3(c)]
Segment
T *start [°C]
T *end [°C]
ΔH [kW]
T **start [°C]
T **end [°C]
Sink B1
185
245
240
190
250
Source B1
185
135
100
180
130
Source B2
75
55
40
70
50
TABLE 10.10 Heat Source and Sink Segments from the GCC for Process B [Problem 3(c)]
Answers to (d) and( e). By analyzing the TSPs (Figure 10.12), one can see that there is an opportunity to match heat source and heat sink requirements in the steam temperature interval between 100°C and 180°C. The proposed steam saturation temperature is in the interval from 117.1°C to 130°C. Within this interval, the maximum heat recovery through the steam system (equal to 100 kW) is achieved. This results from matching the values for steam generation and steam use (here both are equal to 100 kW).
229
230
Chapter Ten Source Source Source Source A2 B1 B2 A1
TB [°C] 180
ΣCP = 2
ΔH = 100
ΣCP = 0
ΔH = 0
ΣCP = 1
ΔH = 20
ΣCP = 3
ΔH = 60
ΣCP = 6
ΔH = 180
130 90 70 50 20 ΔH [kW]
40
180
100
40
CP [kW/°C]
1
6
2
2
FIGURE 10.10
Combined Site heat sources for Problem 3(c). TB [°C]
FIGURE 10.11 Combined Site heat sinks for Problem 3(c).
Sink A1 Sink A2 Sink B1 250 ΣCP = 4 ΔH = 240 190 ΣCP = 0 ΔH = 0
150
ΣCP = 7 ΔH = 280 110 ΣCP = 5 ΔH = 50 100 ΔH [kW] CP [kW/°C]
280
50
240
7
5
4
Temperature 250 Temperature interval with 100 kW heat recovery via the steam system T
∗∗
T
∗∗
= 130°C
150
= 117.1°C
50
−400
−200
0
200 ΔH [kW]
FIGURE 10.12
Total Site Profiles for Problem 3(c).
400
600
Examples and Case Studies
10.2.2 Total Sites: Second Problem Problem 4: Task Assignment Two processes (A and B) are operating on a site. The stream parameters for process A and process B are given in Tables 10.11 and 10.12, respectively. Medium-pressure (MP) steam from a boiler is available at 220°C saturation temperature, and ΔTmin = 10°C for both processes. The cooling water supply temperature is 18°C and the return temperature is at most 35°C. (a) Use the Problem Table Algorithm (PTA) to identify the minimum heating/cooling duty for each process, and plot the GCCs. (b) Create the TSPs including the GCC pockets. (c) Select the saturation temperature for a second (low-pressure) steam level, and construct the Total Site composites, so as to maximize heat recovery via the steam system. Problem 4: Solutions Answer to (a). Pinch Analysis has been applied to the two processes just described, and the targets thus identified are given in Table 10.13. The GCCs for process A and process B are shown in Figures 10.13 and 10.14, respectively.
Stream
Tsupply [°C]
Ttarget [°C]
CP [MW/°C]
∆H [MW]
160
0.15
2.55
116.01
0.15
6.16
A1
143
A2
75
A3
124
90
0.412
14.0
A4
90
84
0.35
2.1
A5
84
37.5
0.101
4.7
CP [MW/°C]
∆H [MW]
TABLE 10.11
Stream
Stream Data for Process A (Problem 4)
Tsupply [°C]
Ttarget [°C]
158.3
159.9
1.33
B2
70
158.3
0.05
4.42
B3
10
70
0.03
1.8
B4
180
135.6
0.075
3.33
B5
135.6
105
0.105
3.21
B1
TABLE 10.12
Stream Data for Process B (Problem 4)
2.08
231
232
Chapter Ten Q H,min [MW]
Q H,min [MW]
Pinch (hot/cold) [°C]
Process A
2.852
14.955
124 / 114
Process B
1.800
0
None (threshold problem)
TABLE 10.13
Heat Recovery Targets for Problem 4(a)
∗ T [°C]
(2.9; 165.0)
160
(0.3; 148.0)
140
(0.3; 121.0)
120
(8.9; 85.0)
100
(0; 119.0)
(10.3; 79.0)
80
(9.9; 80.0)
60 40 (15.0; 32.5)
20 0 0
FIGURE 10.13
2
4
6
8
10
12
14
ΔH [MW]
GCC for process A [Problem 4(a)].
T* [°C] (1.8; 175.0)
180
(0.5; 163.3)
(2.6; 164.9)
160 140 120
(1.4; 130.6)
100 (3.1; 100.0)
80 (1.8; 75.0)
60 40 20
(0; 15.0)
0 0
FIGURE 10.14
0.5
1
1.5
2
2.5
3
ΔH [MW]
GCC for process B [Problem 4(a)].
Answer to (b). The TSPs for the combination of the two processes are plotted in Figure 10.15. The apparent temperature overlap of the profiles indicates a good opportunity for heat recovery. Answer to (c). The heat recovery on the site can only take place via the steam system. Any steam raised from process cooling has to be utilized to the greatest
Examples and Case Studies extent possible in order to maximize the recovery. Usage may exceed generation; in this case, the difference would be made up by MP steam. The profiles in Figure 10.15 were used to balance steam generation and use by the processes by varying the LP level saturation temperature. This has been achieved for LP generation and use steam loads of 6.9 MW at 109°C saturation temperature, which corresponds to 1.39 bar(a) steam pressure. The Total Site composites are shown in Figure 10.16
T** [°C] 200
100
50
−30
−25
FIGURE 10.15
−20
−15
−10
−5
0 5 ΔH [MW]
10
15
20
10
15
20
Total Site Profiles for Problem 4(b).
T** [°C] 200
6.9 MW 100
50
−30
−25
FIGURE 10.16
−20
−15
−10
−5
0 5 ΔH [MW]
Total Site composites for Problem 4(c).
233
234 10.3
Chapter Ten
Integrated Placement of Processing Units and Data Extraction Problem 5: Task Assignment The process flow sheet for a base-case design is shown in Figure 10.17 with some additional data. The CCs and GCC of the process (with appropriate data) for ΔTmin = 20°C are shown in Figures 10.18 and 10.19, respectively. purge FR
H
120°C
m=0.4
900
C m=1 F1 F2
1300
505
50°C
180°C
CA
130°C
H 50°C
180°C
R1 T=180°C
180°C
m=0.8
220°C
A
H
HM m=1.4
220°C RA
1050
H CB 200°C
P2
200°C
70°C
980
Column A
1000
C
B
m=1.3 Column B
RB
C
Heat exchanged with cold utility
H
Heat exchanged with hot utility
Q
Heat exchanged [kW]
1928
m=0.5 70°C P1
C
FIGURE 10.17
150°C
270°C
H
Process flowsheet for the base-case design (Problem 5).
ΔTmin = 20 °C
T [°C]
m=[kg/s] 270°C
250
2642.6
Pinch
200
Th = 240 °C Tc = 220 °C
150 100 50
10316.6
1958.6 0 0
FIGURE 10.18
2000
4000
6000
8000
10000 ΔH [kW]
Composite Curves for the process in Problem 5.
Examples and Case Studies T (°C)
250 200 150 100 50 0 0
500
FIGURE 10.19
1000
1500
2000
2500
Q (kW)
GCC for the process in Problem 5.
(a) Extract the appropriate stream data for performing a Pinch Analysis. Tabulate the extracted data where appropriate. If a stream involves phase changes (several segments), then give the relevant details (e.g., stream F1 has three segments: liquid phase, vaporization, and vapor phase). (b) In the base-case design, what is the hot and cold utility consumption and how much is the process heat recovery? Compare the utility usage in the base case with the targets shown on the CCs. Why are they different? (c) Explain how the heat recovery system can be improved over the base case without modifying any conditions (e.g., reactor pressure, column pressure) of the main process. (d) Refer to the GCC in Figure 10.19. (1) If a heat pump is proposed, what operating temperature levels will best fit the heating and cooling needs of the process? (2) What qualitative suggestions can be made concerning alternatives for utility saving and appropriate utility placement? (e) If process modifications are allowed, comment on whether they could help to improve the process Heat Integration. Your answer should address the reactors, the distillation columns, feed vaporization, and soft data. For this problem, the following assumptions are made. Utilities generated can be sold, and the unit cost of high-temperature utilities
235
236
Chapter Ten is greater than that for low-temperature utilities. The temperature of flue gas for direct heating ranges from 800°C to 160°C. Steam usage and generation are isothermal utilities (up to four levels between 100°C and 340°C). A heat pump or heat engine can be introduced. Additional data for the problem: • Condensers and reboilers of distillation columns A and B involve only phase changes. • In the condenser of distillation column B, total condensation is assumed (i.e., the outlet is in the liquid phase only). • An endothermic reaction takes place in reactor R1. A heating medium (HM) is necessary to keep the isothermal condition in R1. Some properties of the streams are as follows: F1:
Boiling point = 120°C, ΔHvap = 990 kJ/kg, Cp,vap = 4.5 kJ/kg·°C
F2:
Boiling point = 125°C, ΔHvap = 1250 kJ/kg, Cp,vap = 7.5 kJ/kg·°C, Cp,liq = 15 kJ/kg·°C
Cp,vap = 5.0 kJ/kg·°C HM: Cp,liq = 25 kJ/kg·°C
FR: P1:
Boiling point = 270°C, ΔHvap = 1556 kJ/kg
P2:
Boiling point = 200°C
Problem 5: Solutions Answer to (a). Stream data needed to perform a Pinch Analysis is reported in Table 10.14. Answer to (b). For the base case, Q H = 4828 kW, Q C = 3366.4 kW, and process heat recovery = 3530 kW. The targets are Q H = 2642.6 kW and Q C = 1958.6 kW, and process heat recovery = 5715.8 kW. The base-case design requires much more utility consumption. This is because some heat is transferred across the Pinch and also because there is some unnecessary utility use. For example, some amount of heat is needed in the column B reboiler for heating up P1, and this amount must be taken out again when P1 needs cooling. Answer to (c). • The cold streams below the Pinch are F1, F2, and FR. These streams can be heated up by P1 (parts below the Pinch), P2, and by the condensers of columns A and B. In the base case, stream P1 is used to supply heat to stream F2, resulting in heat transfer across the Pinch. Instead, to heat up stream F2 at its higher temperature range, heat exchange with stream P2 can be employed; for the lower-temperature part of F2, the heat from the condenser of column A can be recovered. • Above the Pinch, P1 can be used to heat up parts of the column A reboiler or of the heating media of reactor R1.
Examples and Case Studies Stream
Type
Ts [°C]
Tt [°C]
m Cp ΔHvap [kg/s] [kJ/kg·°C] [kJ/kg]
CP ΔH [kW/°C] [kW]
F1-liq F1-evap F1-vap
Cold Сold Сold
50 120 120
120 120 180
1 1 1
11 — 4.5
— 990 —
F2-liq F2-evap F2-vap
Cold Cold Cold
50 125 125
125 125 180
0.8 0.8 0.8
15 — 7.5
— 1250 —
12 — 6
900 1000 330
HM
Cold
220
250
1.4
25
—
35
1050
CA
Hot
120
120
—
—
—
—
900
11 — 4.5
770 990 270
RA
Cold
220
220
—
—
—
—
1000
FR
Cold
120
180
0.4
5
—
2
120
CB
Hot
200
200
—
—
—
—
980
RB
Cold
270
270
—
—
—
—
1928
P1-cond P1-liq P2
Hot Hot Hot
270 270 200
270 70 70
0.5 0.5 0.8
TABLE 10.14
— 37.17 12.5
1556 — —
— 18.58 10
778 3716 1300
Stream Data for Pinch Analysis [Problem 5(a)].
Answer to (d)(1). Using the GCC in Figure 10.19, from the viewpoint of a heat pump, two good candidates for its integration are as follows: • Across the Process Pinch. Heat rejection from the heat pump to the process T* = 230°C (real temperature T = 240°C) and heat absorption from the process at T* = 190°C (T = 180°C), would result in temperature lift ΔTlift = 60°C. The main challenge is that the heat available for absorption by the heat pump is relatively smaller than what can be delivered by the heat pump above the Process Pinch. • Across a potential Utility Pinch. Heat rejection level T* = 130°C and absorbtion heat level T* = 110°C. The corresponding real heat pump temperatures are: heat rejection level T = 140°C and heat absorption level, T = 100°C. This results in temperature lift ΔTlift = 40°C. At this level, a large amount of heat is available to be absorbed from the process and an even larger amount can be rejected to the process. Answer to (d)(2). • From the GCC, the heat pump integration can be placed at the levels of 140°C and 100°C with only a small amount of shaft work required. This could save a large part of the cold utility required. • Above the Pinch, a large and quite flat pocket is present in the CCC, providing an opportunity to generate steam at the maximum level of 180°C. • Above the Pinch, steam is preferred over flue gas for a hot utility because the process stream temperatures are not too high.
237
238
Chapter Ten • A heat engine could also be a candidate for the pocket in the GCC below the Pinch. However, since ΔT is small—about 80°C, or [(190 + 10) – (130 – 10)]—only a small amount of work may be produced. Answer to (e). • Reactors. The reaction is endothermic and, to save utility, its heating medium heat demand should be placed below the Pinch. It is now placed above the Pinch. If the pressure and temperature of reactor R1 are reduced, then an alternative heating media with lower temperature levels could be used. As a result of decreased R1 pressure, the boiling points of feeds F1 and F2 would also decrease. This would result in different targets. For instance the GCC could be shifted more to the left and the Pinch point could be changed, resulting in more process heat recovery. • Distillation columns. The reboilers of columns A and B currently evaporate the entire flow from the column bases. The appropriate arrangement would be to move the reboilers to evaporate only the branches of these flows that are intended for the columns. • Feed vaporization. In the base case, the vaporization of F1 and F2 are both placed below the Pinch. However, as mentioned, the current vaporization of F2 results in cross-Pinch heat transfer. Also, the vaporization of F1 employs utility heating below the Pinch, which should be eliminated and substituted for process-to-process heat recovery. • Soft data. The P1 and P2 target temperatures are soft data that can be changed. The temperatures could be increased to reduce cold utility required, but this would not change the CC much.
10.4 Utility Placement 10.4.1
Utility Placement: First Problem
Problem 6: Task Assignment A set of data for a part of a crude distillation process has been extracted, and the stream data are shown in Table 10.15. For this problem, the minimum temperature difference is ΔTmin = 10°C. The available utilities are presented in Table 10.16. A targeting procedure was performed for the initial data summarized in Tables 10.15 and 10.16. Figure 10.20, which exhibits two Pinch points, shows the balanced CCs for the problem. (a) Complete the missing entries in Table 10.15. (b) Identify the Pinches and explain their significance. (c) Draw the design grid for the problem, including the hot utilities. Position the streams relative to the location of the Pinches. (d) Design the MER HEN below the Process Pinch. (e) Design the MER HEN above the Process Pinch. (f) Design the MER HEN above the Utility Pinch.
Examples and Case Studies Stream
Name
Ts [°C]
1
OH_Naphta_1
100
41
27.5113
2
OH_Naphta_2
132
60
23.1850
3
OH_HKD
224
65
4
LCT
268
30
5
Residue
283
45
336.25
1.4128
6
Crude
30
146
2031.50
17.5132
7
Denaphta_1
117
204
977.88
8
Denaphta_2
176
305
1641.45
TABLE 10.15
Tt [°C]
ΔH [kW]
CP [kW/°C]
11.2400 363.12
1.5257
12.7244
Process Streams for Problem 6
Stream
Ts [°C]
Tt [°C]
Cost [£/kWy]
HP steam
320.1
320.0
100
MP steam
260.1
260
60
20.0
40.0
8
Cooling water TABLE 10.16
Available Utilities for Problem 6
HP steam = 654.2 kW 320.0°C MP steam = 874.4 kW
300
320.1°C 305.0°C
260.1°C 260.0°C 250.1°C
T* [°C]
224.0°C 200
204.0°C 176.0°C 132.0°C 100.0°C
100 C .0° 60.0°C
5 41.0°C 4
30.0°C 0
20.0°C
146.0°C 122.0°C
40.0°C Cooling water = 2657 kW 2000
4000
6000
8000
ΔH [kW]
FIGURE 10.20 Problem 6.
Balanced Composite Curves and heat recovery targets for
239
Chapter Ten Problem 6: Solutions Answer to (a). Table 10.17 shows the missing parameters of process streams described in Table 10.15. Answer to (b). There are two Pinches: the Process Pinch (located at 132°C/122°C) and the Utility Pinch (located at 260.1°C/250.1°C; see Figure 10.21. The Utility Pinch is caused by the placement of the MP steam. More significant is the Process Pinch, which divides the design area into two different parts: a net heat sink above the Pinch temperatures and a net heat source below the Pinch temperatures. The Utility Pinch further subdivides the area above the Process Pinch. In the interval between the Process Pinch and the Utility Pinch, the only utility allowed is the MP steam. Above the Utility Pinch, the only utility available remains the HP steam. Answer to (c). See Figure 10.22.
No
Name
Ts [°C]
1
OH_Naphta_1
100
41
1623.17
27.5113
2
OH_Naphta_2
132
60
1669.32
23.1850
Tt [°C]
ΔH [kW]
CP [kW/°C]
3
OH_HKD
224
65
1787.16
11.2400
4
LCT
268
30
363.12
1.5257
5
Residue
283
45
336.25
1.4128
6
Crude
30
146
2031.50
17.5132
7
Denaphta_1
117
204
977.88
11.2400
8
Denaphta_2
176
305
1641.45
12.7244
TABLE 10.17
Process Streams for Problem 6(a) with Missing Data Filled In
HP steam = 654.2 kW MP steam = 874.4 kW
300
260.0°C
320.0°C
200 132.0°C
Utility Pinch
5 41.0°C 4
C 60.0°C
30.0°C
20.0°C
204.0°C 176.0°C
Process Pinch
.0°
0
250.1°C
146.0°C 122.0°C
100.0°C
100
Cooling water = 2657 kW 2000
4000
6000
ΔH [kW]
FIGURE 10.21
320.1°C 305.0°C
260.1°C
224.0°C
T [°C]
240
Pinch identification for Problem 6(b), ΔTmin = 10°C.
8000
Examples and Case Studies 132.0°
260.1°
320.1°
320.0°
MP_steam
11
ΔH = 874.373
1
ΔH = 1623.167
CP = 27.51
OH_Nafta_1
132.0°
60.0° CP = 23.18
OH_Nafta_2
2
ΔH = 1669.32
224.0°
65.0° CP = 11.24 ΔH = 753.08
ΔH = 1034.08
CP = 11.24
OH_HKD
3 268.0°
30.0° CP = 1.63
ΔH = 165.621
ΔH = 195.442
CP = 1.63
CP = 1.53
ΔH = 12.053
ΔH = 122.913
CP = 1.41
ΔH = 180.979
CP = 1.41
LCT
CP = 1.41
ΔH = 32.353
5
Residue
146.0°
30.0°
6
4
283.0°
45.0°
CP = 17.51 ΔH = 1611.214 117.0°
CP = 17.51
ΔH = 420.316 204.0°
CP = 11.24
ΔH = 921.68
7 ΔH = 56.2
CP = 11.24
Crude Denafta_1 305.0°
8
ΔH = 942.878
CP = 12.72
Denafta_2
ΔH = 698.589
CP = 12.72
40.0°
20.0°
9
HP_steam
260.1°
CP = 8743.73
100.0°
41.0°
10
ΔH = 654.163
CP = 6541.63 260.0°
Cooling water
CP = 132.83 ΔH = 2656.687 250.1°
122.0°
ΔTmin = 10°C; All temperatures are in [°C], ΔH are in [kW], CP are in [kW/°C]
FIGURE 10.22
Preliminary grid diagram for Problem 6(c).
Answer to (d). See Figure 10.23. 1623.17 kW
132.0°
100.0°
41.0°
C1
1 57.71 kW
60.0°
132.0°
62.51°
2
2
C3 696.88 kW 65.0° C5
127.3°
132.0° 3
4
155.62 kW 30.0°
132.0°
C6
4
122.91 kW 132.0°
45.0°
5
C7 30.0° 2
6
122.0°
1611.21 kW 7
117.0°
4
122.0°
56.2 kW 20.0° 9
C1
C7
C6
C5
C3 122.0°
FIGURE 10.23
HEN design below the Process Pinch [Problem 6(d)].
241
242
Chapter Ten Answer to (e). See Figure 10.24. 260.1°
132.0°
260.0°
12
214.0° CP = 11.24
8
9
CP = 1.53
224.0°
13
260.1°
11
3 4
10
5
11
CP = 1.41 CP = 8.755
10
CP = 17.51
6
CP = 11.24
7
144.32° 143.49°
142.67° 195.44 kW
11
146.04° 13
180.98 kW
43.90 kW 8
921.68 kW 8
CP = 12.72
176.0°
204.0°
185.32° 12
9
112.4 kW
830.48 kW
122.0°
250.1° All temperatures are in [°C], ΔH is in [kW], CP is in [kW/°C]
FIGURE 10.24
HEN design above the Process Pinch [Problem 6(e)].
Answer to (f). See Figure 10.25. 260.1° 320.0°
16
320.1°
CP = 6541.86 12.05 kW
268.0°
15
4
CP = 1.53 283.0° 14
CP = 1.41
CP = 3.36
8
14
15
251.99°
255.19° 253.59°
CP = 3.36 32.35 kW
5
16
305.0°
CP = 6.72 654.16 kW
250.1°
FIGURE 10.25
MP steam
HEN design above the Utility Pinch [Problem 6(f)].
10
Examples and Case Studies
10.4.2
Utility Placement: Second Problem
Problem 7: Task Assignment A process involves the set of process streams described in Table 10.18. The utilities available to satisfy its heating and cooling requirements are given in Table 10.19. For the HEN to be designed, assume ΔTmin = 30°C. (a) Identify the minimum heating duty, the minimum cooling duty, and the Process Pinch location. (b) Plot the GCC. (c) Perform an appropriate placement of the utilities against the GCC so as to achieve minimum total cost of the utilities. (1) Draft the placement on the GCC. (2) Calculate the total duty for each of the utilities. (3) Calculate the total utility cost for the problem. (d) Consider the following ways to reduce the utility costs further, and comment on their suitability for application to
Stream H1
Ts [°C]
Tt [°C]
CP [MW/°C]
175
75
3.5
H2
100
40
2.4
H3
180
130
1.5
H4
195.1
195
100
C1
50
C2
140
C3
80
140
1.5
C4
20
80
3.0
TABLE 10.18
ΔH [MW]
150
2.0
140.1
150
Process Streams Data for Problem 7
Name
Ts [°C]
Tt [°C]
HP steam
300.1
300
65000
MP steam
200.1
200
50000
Cooling water
15.0
20.0
7000
Chilled water
5.0
10.0
30000
TABLE 10.19
Utilities Available for Problem 7
Cost [$/MWy]
243
244
Chapter Ten this current problem. Wherever appropriate, suggest how an option can be exploited. (1) Flue gas heating (2) Heat pumping (3) Introduction of a new steam level Problem 7: Solutions Answer to (a). The heat recovery targets and the Pinch location can be obtained with the help of the PTA. The heat cascade intervals and the corresponding stream population are shown in Table 10.20, and the computed problem table is given in Table 10.21.
Temperature boundary
Interval number
Stream population
180.1 1
H4
180.0 2 165.0 3
H3, C1
3
H1, H3, C1
4
H1, H3, C1, C2
5
H1, H3, C1, C3
6
H1, C1, C3
7
H1, C1, C4
8
H1, H2, C1, C4
9
H1, H2, C4
160.0 155.1 155.0 115 95.0 85.0 65.0 60.0 10
H2, C4
11
H2
35.0 25.0 TABLE 10.20 Heat Cascade Intervals and Stream Population for Problem 7(a)
Examples and Case Studies Alternatively, the heat recovery targets can be obtained by plotting the CCs, as shown in Figure 10.26. Answer to (b). See Figure 10.27. Answer to (c)(1) and (c)(2). See Figure 10.28. Answer to (c)(3). Total utility cost = 37.5 [MW MP steam] × 50,000 [$/MWy] + 62.5 [MW cooling water] × 7,000 [$/MWy] + 24 [MW chilled water] × 30,000 [$/MWy] = 3,032,500 [$/y]
Interval number
Temperature [°C]
1
180.1
2
180
Enthalpy [MW] 37.5 137.5
3
165
137.5
4
160
135.0
5
155.1
149.7
6 (Pinch)
155
0.0
7
115
60.0
8
95
60.0
9
85
45.0
10
65
63.0
11
60
77.5
12
35
62.5
13
25
86.5
TABLE 10.21
Problem Table for Problem 7(a)
FIGURE 10.26 T [°C] Composite Curves for 250 Problem 7(a), ΔTmin = 30°C.
QH,min = 37.5 MW
200 150 Pinch: 170°C (hot) 140°C (cold)
100 50
QCmin=86.5 MW 0 0
200
400
600
800
ΔH [MW]
245
246
Chapter Ten FIGURE 10.27 Grand Composite Curve for Problem 7(b), ΔTmin = 30°C.
QH,min = 37.5 MW T [°C] 150 100 50 QCmin=86.5 MW 0 0
FIGURE 10.28 Appropriate placement of utilities for Problem 7(c).
50
T [°C]
100
ΔH [MW]
MP steam 37.5 MW
150 100 Cooling Water 62.5 MW
50
Chilled Water 24 MW
0 0
50
100
ΔH [MW]
Answer to (d)(1). The flue gas heating option will probably be more expensive than heating with steam. The temperature levels in the problem do not suggest exploitation of furnace heat. Answer to (d)(2). The appropriate placement of a heat pump is across the Process Pinch, so the pump will transfer heat from below to above the Pinch. In this problem it is technically impractical to install a heat pump with water or steam as a working fluid, since this would require operation at a slight vacuum (the Pinch is at 170°C/140°C). The relatively small duties that could be achieved with this technique indicate that investment in a heat pump here is likely to be economically unattractive. Answer to (d)(3). Introducing a new steam level at a steam temperature of 165°C is a good option that can be exploited by passing the steam (taken at 200°C) through a steam turbine. This has the potential of generating additional power on the site. The steam would be returned to the process at the level of 165°C (the process heating requirement is at 150°C, which makes this arrangement feasible). If the site could use additional power, then this option may reduce the overall utility cost by providing on-site power generation.
Examples and Case Studies
10.5 Water Pinch Technology 10.5.1 Water Pinch Technology: First Problem Problem 8: Task Assignment A certain process contains several water-processing operations, as shown in Table 10.22. (a) Create the limiting CCs by plotting the contaminant concentration, C, against contaminant load, m [kg/h], [ppm]. (b) Calculate the minimum water flow rate for maximum reuse. Problem 8: Solutions Answer to (a). The individual limiting water profiles are graphed in Figure 10.29. Figure 10.30 shows the result when these water profiles are combined.
Operation
Contaminant flow rate [kg/h]
1
Cin [ppm] Cout [ppm]
2
0
2
5
3
30
4
4
TABLE 10.22
Limiting flow rate [t/h]
100
20
50
100
100
50
800
40
400
800
10
Data for Water-Processing Operation of Problem 8
800
C [ppm]
4
400 3
100 1 50 0 2
2 7
37 41 m [kg/h]
FIGURE 10.29
Limiting water profiles for Problem 8(a).
40
247
Chapter Ten Answer to (b). The minimum wastewater is targeted as follows: Minimum flow rate
Mass pickup at pinch Pinch concentration 9 [kg/h] 9 10 6 ª kg º 100 [ppm] 100 «¬ h »¼ 9 10 3 100
ªtº «¬ h »¼
90 [t/h]
See Figure 10.31.
FIGURE 10.30
Combined limiting water profiles for Problem 8(a).
800
C [ppm]
248
400 ine /h) t ly L upp ate 90 S r r e t w Wa m flo imu n i (M
Pinch
100
0
1
9
21 m [kg/h]
FIGURE 10.31
Minimum wastewater targeting for Problem 8(b).
41
Minimize flowrate
Examples and Case Studies
10.5.2 Water Pinch Technology: Second Problem Problem 9: Task Assignment Using the stream data from Problem 8, suggest a design of a network for maximum reuse of water. The limiting CC is shown in Figure 10.32. Problem 9: Solution Cut off the pockets of the limiting CC. The minimum water requirement for each pocket can then be defined for each region (Figure 10.33). This enables one to propose the design strategy shown in Figure 10.34. Following the design strategy, set up a design grid as shown in Figure 10.35. Connect streams to water mains and merge operations that cross boundaries; see Figure 10.36. Remove intermediate water mains where this is appropriate (Figure 10.37). Then connect operations directly, as shown in Figure 10.38. Finally, Figure 10.39 illustrates (in conventional flowsheet form) one of the possible resulting designs of the water system.
C [ppm]
800
400
100
0
1
9
21 m [kg/h]
FIGURE 10.32
Limiting Composite Curve for Problem 9.
41
249
Chapter Ten
C [ppm]
800
45.7 t/h
400
100 50 90 t/h
m [kg/h]
FIGURE 10.33 Water Pinch diagram used to target the minimum water flow rate in Problem 9.
800
C [ppm]
250
45.7 t/h 800 ppm
400 100 50
90 t/h
FIGURE 10.34
90 t/h
45.7 t/h 100 ppm 44.3 t/h100 ppm
m [kg/h]
Design strategy for Problem 9.
Examples and Case Studies Flowrate required for the interval
1
20 t/h
Limiting Flowrates
100 t/h
2 50 ppm
40 t/h
3 50 ppm
3
10 t/h Wastewater
0 t/h 800 ppm
45.7 t/h 100 ppm
90 t/h F.W.
(90-90) 0 t/h
(90-45.7) 44.3 t/h 100 ppm
4 400 ppm
(45.7-0) 45.7 t/h 800 ppm
Water mains
FIGURE 10.35
Design grid for the water system in Problem 9.
90 t/h F.W. 20 t/h
1
45.7 t/h 100 ppm 20 t/h 50 t/h
100 t/h
20 t/h
40 t/h
0 t/h 800 ppm
2 3
40 t/h 20 t/h 5.7 t/h
10 t/h 0 t/h
FIGURE 10.36
44.3 t/h
4 45.7 t/h
Streams are connected with the water mains (Problem 9).
251
252
Chapter Ten 90 t/h F.W. 20 t/h
45.7 t/h 100 ppm
0 t/h 800 ppm
20 t/h
1
50 t/h
100 t/h
20 t/h
40 t/h
2 3
40 t/h 20 t/h 5.7 t/h
10 t/h 0 t/h
4
44.3 t/h
45.7 t/h
FIGURE 10.37 Removing intermediate water mains and then connecting sources and sinks (Problem 9).
90 t/h F.W. 20 t/h
20 t/h
1
50 t/h
100 t/h
20 t/h 40 t/h
2 3 20 t/h
5.7 t/h
10 t/h 0 t/h
FIGURE 10.38
4
44.3 t/h
45.7 t/h
Connecting operations directly (Problem 9).
20 t/h 20 t/h F.W. 90 t/h
50 t/h
Operation 1
Operation 2
40 t/h
5.7 t/h
Operation 3 Wastewater 90 t/h Operation 4 44.3 t/h
FIGURE 10.39
Flowsheet representation of a water system design (Problem 9).
CHAPTER
11
Industrial Applications and Case Studies
T
his chapter provides an overview of selected implementations of the Process Integration methodology for various industrial case studies. The presentation is somewhat condensed because of space limitations; more information is available in the works cited.
11.1
Energy Recovery from an FCC Unit In this case study, the Heat Exchanger Network (HEN) of a Fluid Catalytic Cracking (FCC) unit process consisted of a main column and a gas concentration section (Al Riyami, Klemeš, and Perry, 2001). The stream data was made up of 23 hot streams and 11 cold streams. The associated cost and economic data required for the analysis were specified by the refinery owners. Incremental area efficiency was used for the targeting stage of the retrofit design. This was carried out using the Network Pinch method (Asante, 1996; Asante and Zhu, 1997), which consists of a diagnosis stage and an optimization stage. In the diagnosis stage, a few promising retrofit steps were generated using the UMIST (now the University of Manchester) software package SPRINT (2009). This software was also used to optimize the initial design by trading off capital cost against energy savings. The design options were then compared and evaluated, followed by the final retrofit design proposed for final inspection. The existing ΔTmin of the process was identified as 24°C, and the hot utility consumption of the process was 46.055 MW; the area efficiency of the existing design was 0.805. The potential for energy saving was then derived from the resultant Composite Curves, which are shown in Figure 11.1. As seen in the figure, the Composite Curves are relatively wide apart except in the area around the Pinch. The capital cost was estimated under the assumption that the retrofit distribution of area would be the same as that for the existing network. The resulting optimum minimum temperature approach was found to be about 11.5°C for incremental α and about
253
Chapter Eleven 400 ΔTmin = 11.50°C Temperature [°C]
254
300
200
100
0
1
2
Enthalpy [MW]
FIGURE 11.1 Composite Curves of FCC process with optimum ΔTmin (after Al Riyami, Klemeš, and Perry, 2001).
17.5°C for constant α. The area efficiency α of the existing network was found to be 0.804. This value indicated that the existing design was using the area reasonably efficiently. Even so, there was still room for improvement. Since the constant-α targeting produced a conservative estimate, an incremental α value of 1.0 was used to set the retrofit target, which yielded potential for energy savings of about 12.117 MW. Analysis of the existing design revealed that there were four process-to-process heat exchangers that transferred heat across the Process Pinch (from above to below the Pinch). It was also found that some heaters supplied utility heat to process streams below the Pinch and that some coolers removed heat from process streams above the Pinch. These energy violations of the established Pinch rules generated the scope of the project’s possible energy savings. The retrofit design using the Network Pinch method allowed a limit to be set on the structure’s energy recovery. The next stage consisted of testing a set of modifications that would result in higher levels of energy recovery in the process. The increase in energy recovery would come at the expense of increased heat exchange area. Therefore, any benefit in energy cost reduction had to be weighed against the additional capital cost associated with increasing that area. A number of promising design solutions were generated, which were then optimized for minimum total cost. The four designs identified each involved a payback period of less than two years, yet increases in energy prices rendered the actual payback period closer to one year. The final design chosen for the retrofit situation was the one with the shortest payback period and the least additional area required; it is shown in Figure 11.2. In this design option, four new heat exchangers were added and one existing exchanger (number 1 in Figure 11.2) was removed, since its duty approached zero. In reality the exchanger equipment item
Industrial Applications and Case Studies 4 5
224° C
238°
293° 12
C 38° C 50° C 108°
38° 38° 38° 38° 38° 46°
C C
38°
C C C
38°
109°
C2
210.5°
7
H5 H6 H7
104° 294.25 kW
83.3° 40
12459 kW
97.17° 62.87°
H8 H9
68.11°
44
H10 63.11° H11 52.5° H12 51.5° H13 96.41°
H14
53267 kW 46° 359.6 kW 48.92° C 563 kW 653.3 kW
96.3°
74.3° 11
193°
9
53° 46°
195.83 kW
46.44°
H19
H21 60° H22
206 kW
102°
C
110°
88.2°
60°
40 3802 kW
H23
134.8°
2
10
171.7° H 385° 22841 kW 4282.8 kW 195° 4
3
2381 kW 1803 kW
5
192.6°
3697.5 kW
8
38°
H18
67.24°
7336 kW
38°
H17
H20
32.3 kW
C
100°
H15 H16
53°
C3
10198.9 kW
55° 50°
1413.6 kW
C5 C6
H4
158°
856.4 kW
C
H3
224°
2
58.32°
C4
238°
41
1235 kW
1815.6 kW C1
207.8° 8
C C
38°
1247 kW
74°
14125 kW
38° 60°
10
3324 kW 62.05° C C 1902 kW 2704 kW C 244.5 kW C 73.9 kW
C
38°
54.9°
C
C
46°
481 kW
293° 43
42
38° 38°
H2
3
760.5 kW 123°
H1
366°
150.8°
52°
40°
366°
6
60°
11
8787 kW 193° 12
872.25 kW
184.5°
6 C7 C8 C9 C10 C11
78.56°
9
118.4° 118° 48.23°
7
43
101.7°
41479.7 kW
4258.8 kW 213.7 kW
119.3°
13291 kW 124° 41 H
2402.6 kW
1151.8 kW
49.68°
44 1051.9 kW
51.16° 95.40°
95°
101.78°
54.3° 763.9 kW
42 5669.4 kW
H 102° H
200 kW
FIGURE 11.2 Grid diagram of the chosen retrofit option (after Al Riyami, Klemeš, and Perry, 2001). Exchanger number 1 from the original network is removed and now used repiped as new added area (exchanger number 42); exchanger number 10 is only repiped.
need not be removed. Instead, it could be used in place of one of the additional exchangers called for by the new design by repiping either the hot or cold stream (so in this case, exchanger number 1 becomes exchanger 42 in Figure 11.2), thus further reducing the additional area
255
256
Chapter Eleven required. All the modifications carried out in the diagnosis stage had the effect of moving heat from below to above the Network Pinch. The final retrofit design produced energy savings of 8.955 MW, about 74 percent of the potential of the design. The annual utility cost savings amounted to $2,388,600, a 27 percent decrease in the utility bill. Since the modified HEN required an investment of $3,758,420, the payback period was less than 19 months. The study demonstrated that a combination of targeting and Pinch Technology (process Pinch and Network Pinch) can yield substantial improvements in an existing HEN and thereby reduce the total network cost. The employed method can recognize bottlenecks in an existing system, and then generate a series of potential improvements by searching for modifications capable of shifting heat from below to above the Network Pinch. It was found that targeting for maximum energy savings at each potential modification usually produces a good trade-off between area and energy cost.
11.2 De-bottlenecking a Heat-Integrated Crude-Oil Distillation System A crude-oil preheating system retrofit problem was studied by Seikova, Varbanov, and Ivanova (1999). The objective of the study was to develop a topology modification proposal for the refinery owners that would feature better energy efficiency than the existing network under a flexibility requirement—namely, that the distillation plant be able to work with several alternative feedstocks. The preheating system of the distillation plant was analyzed, and several topology modifications were evaluated. The retrofit aspect of Heat Integration was first considered systematically by Tjoe and Linnhoff (1986), who proposed a framework for retrofit assessment and targeting. Later, Asante and Zhu (1997) developed the Network Pinch methodology. This hybrid approach, which combines heuristics and Mathematical Programming (MPR), was implemented through an iterative procedure whose two main stages are to identify the heat transfer bottleneck of the given HEN and then to implement modifications for overcoming it. Another important Process Integration issue involves accounting for different plant operating conditions stemming from the market availability of various feedstocks. The case study reported in Saboo, Morari, and Woodcock (1985) was based on the HEN retrofit methodologies just described and combined the energy recovery improvement with handling of the flexibility requirements that are common to real-life processing plants. The crude-oil distillation unit (Figure 11.3) consists of a main distillation column (with 6 m3/h capacity) for fractionating crude oil into four products: light distillate, heavy gasoline, diesel fuel, and atmospheric residue. The preheating HEN recovers heat from the
Industrial Applications and Case Studies
Gases
Water Steam
Light Distillate
Steam
Crude Oil
Steam Atmospheric residue
Diesel Oil
Lower pumparound Heavy Distillate
FIGURE 11.3
Flowsheet of the crude-oil distillation unit.
pump-around and from some of the products before the furnace. To bring the final crude oil temperature to the column’s required target, the remaining heat duty is provided by the combustion of noncondensed light gases and of additional natural gas. The process features the following characteristics: (1) temperaturedependent heat capacities; (2) continuous partial phase change of the crude oil; (3) temperature variations, which are small owing to specifications of the distillation process; and (4) large variations of the hot stream flow rates due to changing stockfeed composition. The three utility sources available for HEN operation are light gases (as a furnace fuel for higher temperature levels), steam at 1 atm (1.01325 bar), and cooling water at 18–35°C. The data used to estimate the utility cost are $68.74/kWy for furnace heating, $103.10/kWy for steam, and $30.00/kWy for cooling water. The following area cost law was used: Capital cost [$ = 25,000 + 680 Area0.81 [m2]
(11.1)
The plant processes a range of alternative feedstocks. Three crude-oil types were selected to represent this range. Feed 1 is a light crude oil; it contains a significant amount of the lightest fractions, which require a large amount of cooling in the condensers. However, the relatively small amount of the heavy fraction means that less reboiler heating is required. Feed 2 is medium crude oil and feed 3 is heavy crude oil. These two feedstocks are characterized by a relatively greater amount of the heavy fractions (diesel and atmospheric residue), which means that more heat is available for recovery in the higher temperature range. It was assumed that, over a year’s worth of operation, the plant uses equal amounts of the three feedstocks. Each alternative feedstock is associated with a unique operating point.
257
258
Chapter Eleven At all three operating points, the crude oil is preheated within a large temperature interval—from 20°C to 310°C. The preheating process consists of three phases. In the first phase, the crude oil is heated from the starting temperature to its bubble point. The next phase involves continuous partial evaporation in heat exchangers. The third and hottest phase is heating to the specified column entry temperature of 310°C (this heating is performed in the furnace). The upper pump-around features a large enthalpy change split between condensation and subcooling segments. None of the other process streams has phase transitions. The existing HEN for the preheating train is shown in Figure 11.4. In order to initialize the retrofit procedure, a Pinch Analysis of the stream data for ΔTmin = 10°C was carried out. The results of this analysis are given in Table 11.1. For the first two feeds, the Pinch is located at the temperature boundary of the crude oil bubble point. For feed 3, the Pinch is located at a temperature that is close to the beginning of the phase change. Note that each operating point offers a different potential for heat recovery. After looking at the existing network and thermodynamic targets, one can suggest several retrofit modifications. First, the hot utility usage below the Pinch leads to extremely poor heat recovery as well as to substantially increased utility cooling and thus to using considerable amounts (30–38 m3/h) of cooling water. Taking a look at the coolers in the temperature interval of heater H1 (Figure 11.4), one can see that the sum of the cooler loads at the outlets of streams H3, H4, and H5 offers enough heat supply to satisfy the heating demand in the interval from 20°C to 60°C. This analysis implies that steam heating can be eliminated through repiping of the coolers as recovery matches. It can also be seen that the availability of heat recovery varies in response to the different temperature levels characteristic of each type of crude-oil feed. This effect is most significant for the atmospheric residue stream H5, whose heat capacity flow rate varies from 0.98 to 1.47 kW/°C. To handle this variation in heat availability temperatures, a cascade of recovery
Feed 1
Pinch location [°C]
Minimum hot utility [kW]
Minimum cold utility [kW]
110–120
556.25
560.26
Maximum heat recovery [kW] 610.64
Feed 2
110–120
481.31
428.24
664.33
Feed 3
86.19–96.19
507.34
395.66
696.45
TABLE 11.1 Pinch Analysis Results for Operating Points of the Three Preheating Phases
Industrial Applications and Case Studies CP [kW/°C] CPH11 = 17.68 CPH12 = 17.68 0.38 1.66 0.84 0.98 CPC11 = 2.92 CPC12 = 5.40 CPC13 = 4.23
CPH11
CPH12
28°
65.5°
C2 101.9 kW 70.0° 40.0° C3 11.3 kW 60.0° 25.4 kW C4
C1
94.2° 507.5 kW 111.3°
1
75.3°
186.0°
2
H1 H2 H3
197.0° 129.8° 40.0° H4 3 C5 75.8 kW 296.3° 140.3° 100.0° 4 H5 C6 39.7 kW 119.8° 130.3° 310.0° 60.0° 65.3° 158.7° 20.0° 1 2 C1 4 H2 3 H1 CPC1 CPC1 CPC11 1 2 56.7 kW 153.5 kW 101.7 kW 11.9 kW 183.6 kW 677.3 kW 3 (a) Initial HEN Feed 1
CPH11 = 9.286 CPH12 = 2.086 0.28 1.25 0.87 1.47 CPC11 = 2.87 CPC12 = 4.25 CPC13 = 4.49
101.7°
28.0° 54.7° C2 C1 436.7 kW 55.7 kW 40.0° 70.0° 1 C3 8.6 kW 60.0° 74.2° C4 17.70 kW 122.1° 40.0° C5 21.4 kW 100.0° C6 137.3° 54.8 kW C1
20.0°
H1
H1 111.6° 186.3°
2 3
294.6°
4
60.0°
64.2°
1
2
112.1°
11.9 kW 140.4 kW
114.8 kW
196.7°
127.3°
3
64.7 kW
180.5°
4
H2 H3 H4 H5
310.0°
H2
231.1 kW 582.8 kW
(b) Initial HEN Feed 2 CPH11 = 9.85 CPH12 = 2.27 0.27 1.15 0.88 1.41 CPC11 = 3.43 CPC12 = 4.25 CPC13 = 4.46
28.0° 48.5° C2 46.6 kW 40.0° C3 8.0 kW
70.0°
C1
96.2° 111.2°
1
60.0° 73.2° C4 15.1 kW 109.4° 40.0° C5 61.2 kW 127.6° 100.0° C6 38.7 kW C1
20.0°
H1
60.0°
137.0 kW
H1
469.9 kW
186.0°
2
197.1°
3
294.8°
4 1
63.2°
2
99.4°
3
117.6°
10.9 kW 129.2 kW 77.7 kW
4
172.2°
235.1 kW
H2
H2 H3 H4 H5
310.0°
614.1 kW
(c) Initial HEN Feed 3
FIGURE 11.4
Initial HEN for the crude-oil preheating process.
matches is needed. This requirement suggests the repiping of coolers mentioned previously, which would form three loops for internal heat transfer. Another important observation concerns the amount of recovery from streams H2 to H5. Streams H3 and H4 feature close supply temperatures. For all three operating points, and especially for that of feed 1, H3 has substantially greater heat capacity flow rate than
259
260
Chapter Eleven H4, which results in a relatively slower decrease in its temperature. This fact and the heat exchanger cascade already planned suggest another topology modification: resequence matches 2 and 3 on stream C1. Such resequencing would yield a larger driving force in both matches, although some load would then be shifted to the lower temperature segment of stream C1. The network was modified according to the strategy and changes suggested by the analysis. The result is shown in Figure 11.5. Cooler C3, on the hot stream with the lowest target temperature, is left at its old location. This, together with the introduced three loops, will account for variations in composition of the feedstock. When there is a larger total heat supply from streams H2 to H5, any heat surplus is removed from the system through cooler C3. Comparing the initial and new topologies reveals that a significant improvement in energy efficiency has been achieved. The heat recovery fraction is increased from 67.03 to 91.22 percent for feed 1, from 73.37 to 89.77 percent for feed 2, and from 74.12 to 77.86 percent for feed 3. Although the relative increase for feed 3 is lower, its net increase in heat recovery is the same as that of feed 2. Based on the simulations for the initial and modified networks, the heat transfer area for each match is determined as recorded in Table 11.2 (where H1 is 4 × 6.26 m 2). The area requirement of the recovery matches EC4 (new), 3, EC5, and EC6 increase after the topology change. Note that only the increase of EC4 is significant, which is due to the heat load shifted from match 2. Another effect of the modification is that the four eliminated steam heating
Feed 1 Match
Initial
Feed 2
Retrofit
Initial
Feed 3
Retrofit
Initial
Retrofit
C1
67.69
67.69
61.62
61.62
75.56
75.56
C2
43.42
43.42
30.51
30.51
31.30
31.30
1
10.83
4.74
8.12
3.44
7.39
4.80
3.45
2.00
2.62
1.61
2.42
0.00
98.04
38.43
69.49
27.21
57.75
31.57
C3 2 C4→EC4 3 C5→EC5 4 C6→EC6 H1 TABLE 11.2
4.95
59.41
3.49
39.97
3.02
12.99
10.67
11.83
11.85
12.30
12.97
15.64
7.99
11.43
2.37
11.17
7.47
5.64
14.58
14.44
24.98
23.84
24.20
24.42
1.81
9.68
2.54
12.01
1.89
2.48
21.33
—
20.98
—
25.04
—
Changes in Heat Transfer Area Due to Retrofit, in [m2]
Industrial Applications and Case Studies 28.0°
65.5°
101.85 kW
C2
52.1°
40.0°
507.47 kW
4.54 kW
130.2°
EC4
40.0°
EC5
27.7°
20.0°
C1
EC5
1
3
H4
63.5 kW
22.3 kW
89.4°
296.3°
147.3°
EC6
137.0°
68.9 kW
H5 165.0°
4
2
3
116.6 kW 46.6 kW
4
120.0°
105.3°
EC4
H3 197.0°
EC6 49.4°
H2
186.0°
2
115.3°
100.0°
H1
111.3°
1
C3
60.0°
94.2°
C1
310.0° H2
92.5 kW 146.6 kW
609.9 kW
(a) Modified HEN Feed 1
28.0°
54.7°
55.7 kW
C2
53.2°
40.0°
436.7 kW
111.6°
1
C3 3.79 kW
60.0°
111.7°
C1
EC4
40.0°
134.5°
117.0°
EC5
100.0°
149.8°
16.7 kW
3
EC6
EC4
66.7 kW 93.3 kW
73.2 kW
4
124.0°
107.0°
81.6°
EC5
1
C1
196.7°
3
EC6 49.1°
25.8°
20.0°
186.3°
2
187.0°
138.0° 2
69.0 kW
294.6°
H2
H3 H4 H5 310.0°
H2
4
64.8 kW
H1
212.7 kW 549.2 kW
(b) Modified HEN Feed 2
28.0°
96.2°
C1 436.72 kW
39.8°
40.0° C3 60.0°
48.5°
55.70 kW
C2
111.2°
1
0 kW
99.4° 2
H3 197.1°
65.1°
EC5
3
H4 122.7°
100.0°
EC6
C1
31.9°
25.5°
20.0° 1 18.9 kW
EC5 22.1 kW
EC4
4 89.4°
54.5°
45.2°
45.1 kW
H2
186.0°
EC4
40.0°
H1
EC6 32.0 kW
3
112.7° 2
116.5 kW 99.1 kW
294.8° H5 169.1°
4 241.9 kW
310.0° H2 608.1 kW
(c) Modified HEN Feed 3
FIGURE 11.5
Modified HEN for the crude-oil preheating process.
exchangers can be used to satisfy the increased area needs for matches 3, EC5, and EC6. Thus the investment, which is estimated to be $42,330, is only for the additional area of EC4. Overall, the modified heat recovery system yields a significant reduction in energy cost. The total sum of savings after the retrofit is estimated
261
262
Chapter Eleven to be $17,519 per year, which results in a reasonable payback period of about 2.4 years (or less if energy prices increase). This case study of a combined problem in heat recovery and process flexibility within a distillation unit’s preheating system combined advanced Pinch-based retrofit methodologies with additional heuristic rules. The result was that the flexibility goals were better met by the introduction of loops in the network topology. In short, the new and modified paths enabled a significant improvement in heat recovery. Another benefit was that the more expensive steam heating was partially replaced by slightly increased furnace duty in the case of higher heat demand. This interesting trade-off between two hot utilities yields an economic benefit owing to the lower cost of furnace fuel.
11.3 Minimizing Water and Wastewater in a Citrus Juice Plant This case study describes a water and wastewater minimization project designed for a citrus plant located in Argentina (Thevendiraraj et al., 2003); the study proceeded by applying the Water Pinch technology (Wang and Smith, 1994). Citrus juice–processing plants consume large quantities of freshwater. The principal objective of this study was to reduce both the freshwater consumed and the wastewater produced by the plant. The citrus processing plant housed the following processes: selection and cleaning, juice extraction, juice treatment, emulsion treatment, and peel treatment. Water minimization was achieved by maximizing water reuse and identifying regeneration opportunities. Water-using operations were represented by the maximum inlet and outlet contaminant concentrations, which are functions of equipment corrosion, fouling limitations, minimum mass transfer driving forces, and limiting water flow rate through an operation. Targets determined the minimum freshwater requirement using the Limiting Composite Curve for the water design network. The graphical Pinch methods that are based on single contaminants can be extended to cover multiple contaminants. When dealing with a number of operations, multiple contaminants, and multiple water sources, the problem becomes more complex and so algorithms using the basic Pinch principles have been developed for solving by MPR-based methods (see Smith, 2005). The study began with data extraction: each stream had to be characterized by its contaminant concentration, inlet and outlet concentration levels, and limiting flow rate through each operation. The data were provided in a schematic flow diagram of the citrus plant that incorporated a simplified water distribution network and the mass balance of the plant’s water streams. Eleven freshwaterusing operations were identified. Chemical oxygen demand (COD)
Industrial Applications and Case Studies was chosen as a component to proxy for all contaminants for two reasons: first, COD measures the most significant contaminant load in the majority of water streams; and second, COD exhibits significantly high values. The overall mass balance is closed by using the assumption that evaporation losses from the steam system amount to 1 t/h (otherwise, there is an “inconsistency” of 1 t/h of water that must be accounted for). The extracted data on contaminant concentrations and water flow rate made it possible to establish the amount of water gained or lost by each operation in the process. The total mass load picked up by the freshwater through each operation was then calculated. The eleven water-using operations—together with the water flow rates entering and leaving each operation—can be represented in the form of a simplified water network, as shown in Figure 11.6. The freshwater COD concentration level for the plant is 30 ppm. There was an existing water reuse between processes currently in the plant, and these reuse streams were left unchanged. The simplified water network presented in Figure 11.6 shows the freshwater-using operations with existing water reuse streams “built in” to each identified operation. The current total freshwater consumed and wastewater generated by this citrus plant were, respectively, 240.3 and 246.1 t/h. The existing water network provided a base starting point for the Water Pinch Analysis. The freshwater target was evaluated by using the Composite Curves. The maximum concentration levels were based on the constraints and limitations dictated by process conditions and requirements. This data were represented in the WATER software tool (2005) with identified constraints. The process restrictions on water type permissible for each operation indicated that operations 2, 4, 5, and 10 can use only freshwater as input. Hence, the minimum freshwater required by plant operations was 164.4 t/h. Operation 1 is a batch process for which the analysis assumed freshwater must always be available. This increases the plant’s total minimum freshwater requirement [t/h] to 164.9 = 164.4 + 0.5 for operation 1. The current total freshwater feed to the plant is 240.3 t/h. These figures can be used to calculate the maximum theoretical freshwater reduction (MTFWR) that is achievable: MTFWR
240.3 164.9 100 240.3
31.4 percent
(11.2)
The Water Pinch Analysis was then carried out for the existing water network with maximum concentration levels. The overall freshwater target was calculated using the maximum reuse analysis. Figure 11.7 shows the modified water network represented as a conventional diagram, and the Limiting Composite Curve is plotted
263
264
Chapter Eleven 0.5 t/h
0.5 t/h Packing (1) 1 t/h
Loss to Process 6.265 t/h 7.265 t/h
Loss to Environment 0 t/h
Treatment Plant (Potable) (2)
Boiler
1.000 t/h 26.15 t/h
Blowdown+ Condensate Lossess Neglected 26.15 t/h
Selection / Cleaning (3)
28.093 t/h Gain from Process
Overflow as WW 47.558 t/h 137.558 t/h
APV Condenser & Green Tank (4)
Screen 2&3 & Vincent Press
118.093 t/h
90.0 t/h 240.318 t/h
3.385 t/h
3.385 t/h
246.054 t/h
Vacuum Pump (5) Fresh Water to Plant
1.62 t/h
Screen Taca-Taca (6) Loss to Process
Waste Water from plant
1.62 t/h
4.8 t/h
4.8 t/h Finisher (7) Condensate from distiller stream
1.5 t/h 7.322 t/h Gain from Process Distiller (8) Loss to Process 10.95 t/h 39.6 t/h Screen 1 (9) Loss to Process 15.84 t/h
Condensate from juice vapour 8.822 t/h
28.65 t/h
8.64 t/h Distiller Condenser & Washing Spiral 1 (10) 0.476 t/h
3.6 t/h
7.2 t/h
Gain from Process 4.076 t/h
Centrifugal (11) Total Raw Water Total Waste Water Difference
240.318 t/h 246.054 t/h 5.736 t/h
Total Gain from Process Total Losses to Process Difference
37.391 t/h 31.655 t/h 5.736 t/h
Error % 0.00
FIGURE 11.6
Existing water network (simplified).
in Figure 11.8. The freshwater target was compared with the current freshwater consumption in order to assess the plant’s overall potential for minimizing water and wastewater. The redesigned network yielded a freshwater demand of 169.3 t/h and a wastewater flow rate of 175.1 t/h. Although this amounts to a substantial reduction in both water and wastewater, the design includes reuse of certain streams that will require treatment. As already described, the analysis was based on COD as the pseudocontaminant in reuse streams. However, these streams may well contain other contaminants (e.g., solid waste,
Industrial Applications and Case Studies 0.5 t/h
0.5 t/h
Operation 1
8.96 t/h Loss to Process, 6.265 t/h Recycle, 23.7 t/h 6.265 t/h
2.45 t/h
Operation 2
Loss to Process, 4.8 t/h 4.8 t/h
47.56 t/h Freshwater 169.3 t/h
Operation 3
2.45 t/h
Recycle, 0.83 t/h 0.79 t/h
0.79 t/h Operation 6
Operation 7
Gain from Process, 0.48 t/h 4.01 t/h Operation 11 3.53 t/h Loss to Process, 10.95 t/h 31.83 t/h 39.60 t/h 28.60 t/h Operation 9
Operation 4
Gain from Process, 28.09 t/h 90.00 t/h
Wastewater 175.05 t/h
4.38 t/h 113.72 t/h
Operation 4a
3.39 t/h
3.39 t/h Operation 5 Net gain from Process, 28.09 t/h 8.82 t/h
1.00 t/h Operation 2a/8 Loss to Process 8.64 t/h 15.84 t/h
FIGURE 11.7
Operation 10
7.20 t/h
Water Network after Pinch Analysis as a conventional diagram.
small amounts of chemicals) that require further treatment prior to being used for other processes. Hence additional design options must be developed for dealing with specific process requirements, operating conditions, and suitability standards for water reuse. The necessary design refinements may be achieved by introducing further constraints to potential reuse streams and by utilizing the maximum water reuse analysis to obtain optimal designs that use even less freshwater while meeting all process operating conditions and restrictions. The regeneration reuse analysis can also be used to explore additional design options incorporating the reuse of regenerated water in some operations. Such analysis is based on installation of a treatment unit to regenerate wastewater (by gravity settling, filtration, membranes, activated carbon, biological agents, etc.). This modification would further reduce plant levels of freshwater and wastewater, and it was also evaluated with the WATER software.
265
Chapter Eleven Analysis = Single Component Reference Contm No. 1 [COD] Limiting Composite Curves 30.0 25.0 Concentration [ppb]
266
20.0 15.0 10.0 5.0
0.0
0.2
0.4
0.6
0.8
Mass [t/h] Limiting CC
FIGURE 11.8
Water Supply Line
Limiting Composite Curve generated by WATER software.
Four different design options were generated that considered both the maximum reuse analysis and regeneration reuse analysis (Figure 11.9). Design options A and B were based on the maximum reuse analysis, and both achieve a freshwater consumption of 188 t/h; this is a reduction of about 22 percent from the actual freshwater consumption of 240 t/h. Because of process limitations and restrictions, there was no further scope for water reuse. Yet results from the diagnostic stage indicated that a 31 percent reduction in freshwater and wastewater was possible. This additional extent is achievable by regenerating wastewater and then reusing it in other operations. Design options C and D are based on the regeneration reuse analysis; both result in total freshwater consumption of 169 t/h, which amounts to a 30 percent reduction—that is, nearly all of the theoretical maximum predicted by Eq. (11.2), the expression for MTFWR. The reduction in freshwater use was achieved by rerouting water streams, a modification that requires new pipes. Design option A requires five new pipes; this is fewer than options B or D, which require seven pipes, and option C, which requires nine new pipes. New piping will affect investment cost. The five new pipes identified for design option A are also required in the other options for the same function and at similar flow rates. Therefore, this option requires the least investment and option C the most. Furthermore, design options C and D each require additional investment cost for a
Industrial Applications and Case Studies regeneration unit that is not part of options A and B. These results are summarized in Figure 11.9. Design options A and D are the most attractive ones in terms of maximizing both reuse and regeneration reuse. Option A results in a smaller reduction in freshwater use but at a lower investment cost than option D, which results in a larger reduction in freshwater use but at a much higher investment cost. Wastewater reductions are proportional to freshwater reductions, with corresponding reductions in wastewater treatment costs for each of the options. The cost analysis carried out for design option A indicates attractive financial returns for this low-investment option, whose payback period is only 0.14 years. The outlet water quality of operation 3 required further analysis, so a complete cost evaluation of design option D was not possible. Fully evaluating this option would require additional detailed studies to identify the regeneration process type required and its associated costs. The heat energy of the reuse water streams proposed in the design options was reviewed to ensure that stream temperatures at the inlet of operations remained unchanged. Citrus plant managers reported that nearly all of the process operations occur at ambient temperatures; the only exception was operation 8, which produces wastewater at 90°C. (This particular waste stream is highly contaminated, which imposes some limitations.) All the water reuse streams proposed by the four design options are at appropriate temperatures, so they should not have a thermal effect on operations. The overall hot and cold utility requirements of the plant would not be affected by the changes proposed in the design options. With its existing water network, the plant consumes 240.3 t/h of freshwater and generates 246.1 t/h of wastewater. The proposed design options offer a 30 percent and a 22 percent reduction in freshwater consumption and wastewater generation. For a practical project, the number of modifications is limited. The maximum water
Number of New Pipes Required
FRESHWATER SAVINGS Theoretical Freshwater Reduction Limit, 31% 30
10 5
22
22
5
7
A
B
9
C
Design Options
FIGURE 11.9
Summary of four design options.
30
7 D
% Feed Water Reduction 30 25 20
267
268
Chapter Eleven reuse design requires a minimum of five new pipes, and the regeneration reuse design requires seven new pipes. The reuse analysis predicted a short payback period, but the regeneration analysis was not definitive for the reasons described previously. In sum, a Water Pinch Analysis for this citrus plant demonstrated that the consumption of freshwater could be reduced by as much as 30 percent with low investment and few changes to the existing plant.
11.4 Efficient Energy Use in Other Food and Drink Industries Many studies have employed Pinch Technology (and its associated Heat Integration Analysis) in the food-processing industry. This industry has a far different thermodynamic profile than that of the refining and petrochemical industries. The food-processing industry is characterized by process streams of relatively low temperature (normally 120–140°C), a small number of hot streams, low boilingpoint evaporation of food solutions, considerable deposition of scale in evaporators and heat recovery systems, and seasonal operation. However, a number of studies have also found that the application of Pinch Technology and Heat Integration is hampered by particular aspects of the food-processing systems. These aspects include direct steam heating, difficulties in cleaning heat exchanger surfaces, and high utility temperatures. Despite these drawbacks, the benefits that can be obtained by applying the Pinch Technology (e.g., optimized heat recovery systems and reduced energy consumption) far outweigh the difficulties of performing the studies. There are also other advantages that can be realized by technological improvements: reduced deposition of scale due to reduced utility temperatures, selfregulation of heat processes, and reduced emissions. A case study of the production of refined sunflower oil (Klemeš, Kimenov, and Nenov, 1998) exemplifies the benefits of process analysis based on Pinch Technology and Heat Integration. The process studied operated with a minimal temperature difference of 65°C at the Process Pinch. The external heating required by the system was provided by two types of hot utilities—Dowtherm steam and water steam; the required external cooling was provided by two cold utilities—cooling water and chilled water. The analysis proposed that heat recovery be increased and that the minimum temperature difference be reduced to 8–14°C. The increase in heat recovery (provided by a reduction in the minimum driving force for the process) entailed an increase in the heat transfer area, but this was more than offset by reductions in the hot and cold utility requirements. A further benefit of the analysis was a reduction in the number of utilities needed: eliminating water steam and cooling water considerably simplified the overall design.
Industrial Applications and Case Studies A case study of a whisky distillery by Smith and Linnhoff (1988; see also Caddet, 1994) provides another example of how Pinch Technology and Heat Integration can reduce energy use and increase energy efficiency. In this case it was found that steam was being used below the Process Pinch, resulting in an overall increase in utility usage. The steam was related to use of a heat pump, so the steam used below the Process Pinch was eliminated by reducing the size of that heat pump. Although the steam now had to be used for process heating above the Process Pinch, the overall energy costs were reduced owing to the reduction in compressor duty. Another study involving a whisky distillery was made by Kemp (2007). The hot utility requirement for the process examined was 8 MW, and the Process Pinch was at 95°C. The main hot utility requirements were steam for the distillation system and hot air for the drying system. Kemp showed that the form of the Grand Composite Curve and the temperature of the Pinch could be exploited for heat pumping, and he also suggested that the process would benefit from the introduction of a Combined Heat and Power (CHP) scheme for improved Process Integration and energy efficiency. The site’s power demand was 12 MW, so there were two possibilities for providing both the power and the heat demands from the same utility system. First, a gas turbine that produced 12 MW of power would also supply about 30 MW of high-grade heat from the exhaust. The second option involved the use of backpressure steam turbines, but these would produce nearly 100 MW, much more than what was required. Another advantage of a gas turbine was that its exhaust could be used for drying purposes. Figure 11.10 shows the final configuration of the utility system matched to the GCC. Most of the necessary heat was provided by the gas turbine exhaust and the existing thermo compressors. The existing package boilers were used to provide steam for the thermo compressors. The efficiency of this part of the system was increased by using waste heat from below the Pinch to preheat the boiler feed water. Waste heat boilers driven by the exhaust from the gas turbine provided additional steam. Many processes in the food and drink industry make use of chilling and refrigeration systems. Pinch Technology and Heat Integration have also been used to increase the efficiency of these systems. For example, Fritzson and Berntsson (2006) studied a Swedish slaughtering and meat-processing plant. Their analysis of the plant’s subambient temperature section employed the method proposed by Linnhoff and Dhole (1992) for low-temperature process changes that involve shaftwork targeting. Figure 11.11 shows the plant’s Exergy Grand Composite Curve (EGCC). There is a large gap between the EGCC and the utility curve, which indicates low efficiency in the use of available shaftwork. Improvements in this area could result in a 15 percent reduction in energy demand.
269
Chapter Eleven
xh au st tur
bin
ee
400
300
Ga s
Shifted temperature [°C]
500
200
Waste heat boiler
Site steam Driver steam tu Ga ex rbi s ha ne us t
270
−10
Boiler Feed Water heating Heat pump 0
Cooling water
10
20
30
40
50
60
Net heat flow [MW]
FIGURE 11.10 Final configuration of proposed utility system matched to the process GCC (after Kemp, 2007). 0
500
1000
1500
2000
2500
3000
3500
4000 Q [kW]
0 −0.05
EGCC Utility curve
−0.10 −0.15 −0.20 −0.25 ηc [1]
FIGURE 11.11 Exergy Grand Composite Curve for meat-processing plant (after Fritzson and Berntsson, 2006).
Fritzson and Berntsson (2006) started the energy reduction process by adjusting the loads on the subambient utilities, first maximizing the load on the highest-temperature (−10°C) subambient utility. After this, the load on each lower-temperature utility was
Industrial Applications and Case Studies maximized. The reduction process began at the highest level (the highest temperature), since this utility requirement can be satisfied at a lower cost than refrigeration at lower temperatures. The modified system was then modeled and simulated in HYSYS, which showed a 5 percent reduction in the shaftwork required. However, these results still made for a relatively poor fit between the EGCC and the utility curve. To reduce the gap further, it was suggested that the level of the highest-temperature refrigeration utility be increased from −10°C to −3°C and then the loads be readjusted as before; the result is shown in Figure 11.12. This modified configuration yielded a 10 percent reduction in the shaftwork requirement. Other temperature changes were suggested to reduce further the shaftwork requirements, but these changes were found to be less cost-effective.
11.5 Synthesis of Industrial Utility Systems Varbanov and colleagues (2005) demonstrated the synthesis of a utility system (CHP network) of an industrial Total Site by applying a combination of targeting and Mathematical Programming techniques. Figure 11.13 shows the heating and cooling demands of the chemical site studied. There are two operating scenarios—for winter and summer, with different prices for power and fuel. The basic data for the problem are listed in Tables 11.3 and 11.4. Generic estimates of the corresponding coefficients were specified for the boiler performance (field-erected boilers are slightly more efficient than packaged units), and the capital cost estimates were obtained online (Boiler Cost 2003). These estimates are a function of boiler capacity and steam pressure. The performance and cost of gas
0
500
1000 1500 2000 2500 3000 3500 4000 Q [kW]
0 −0.05
EGCC Utility curve
−0.10 −0.15 −0.20 −0.25 ηc [1]
FIGURE 11.12 Area between the EGCC and the utility curve is further decreased by changing the temperature of the first refrigeration level from −10°C to −3°C (after Fritzson and Berntsson, 2006).
271
272
Chapter Eleven Winter
Summer
T [°C]
T [°C]
300 209.15
209.15 200
−20
130.51
40
20
60
80
−60
−40
198.89 130.51
100
−20
270.68
209.15
200
130.51 LP level candidates
0
HP level candidates
270.68
198.89
130.51 100
−40
300
HP level candidates
0
LP level candidates
20
40
60
80
ΔH [MW]
ΔH [MW]
FIGURE 11.13 Total Site Profiles and candidate steam pressure levels for three steam mains of an industrial chemical site (the Heat Source Profile for summer provides more heat than that for winter).
Winter
Summer
Fraction of year
0.55
0.45
Price of cooling water [$/t]
0.0185
0.0212
Site power demand [MW]
25
31
Maximum export allowed [MW]
10
10
Price of power (import and export) [$/MWh]
20
30
TABLE 11.3
Operating Scenarios for the Total Site
Working hours per year
8600
Interest rate [%]
8
Life of plant [y]
10
Capital installation factor
4.8
ΔTCW [°C]
10
TBFW [°C]
120
P VHP [bar(a)]
90
ΔTVHP,SH [°C]
200
TVP [°C]
50
Natural gas price [$/MWh]
9.369
Distillate oil price [$/MWh]
10.734
Fuel gas price [$/MWh]
4.984
Fuel oil price [$/MWh]
6.226
TABLE 11.4 Site
Configuration Data for the Total
Industrial Applications and Case Studies turbines (ranging from 11.7 to 85.4 MW) were estimated using data from Gas Turbine World (2001). The preliminary Total Site targets estimated minimum utility demand of about 80 t/h for the winter scenario, leading to the choice of a field-erected boiler for stand-alone steam generation. A gas turbine with a Heat Recovery Steam Generator (HRSG) is also selected. For this case, three steam mains (headers) were considered: one for very-high-pressure (VHP) steam generated by the boiler and by the HRSG, and two more intermediate steam mains. The VHP header properties were specified by the problem definition and are given in Table 11.4. Hence, locations for the two other headers still have to be determined. Figure 11.13 illustrates the partitioning of the candidate steam levels, whose ranges are highlighted. For the upper steam main, this graph was used to identify three candidate levels with respective saturation temperatures of 270.68°C, 209.15°C, and 198.89°C; the only candidate for the lower steam main there has a saturation temperature of 130.51°C. The problem superstructure is illustrated in Figure 11.14. Optimizing and reducing the superstructure yields the flowsheet shown in Figure 11.15. It features two fired steam boilers and one steam turbine sized to the steam flow for the winter period. This flowsheet features relatively low on-site power generation and a significant amount of power import. The main reasons behind this approach are that power is cheap and capital is expensive; these factors preclude a better utilization of the steam system’s potential cogeneration. This design also involves a significant amount of CO2 emissions.
Gas turbines gt01
gt01 Process heating demands
Process cooling demands Winter
HRSG
b01
HRSG
b02
VHP: 90 bara, 503.35 °C 12.000 16.885 MW MW 7.700 MW
13.383 MW
1.300 MW
2.877 MW
hdr01: st01
hdr01: st02
hdr01
Canditate pressure levels
tb03 hdr02: st01
15.500 18.136 MW MW
hdr02: st02
hdr02 : tb01
10.632 MW
CT: st01
CT: st02
COND: 0.1235 bara To cooling water
DA
Make-up water
To the stream generators
FIGURE 11.14
3.550 MW
3.175 MW
9.038 MW
5.956 MW
1.635 MW
1.501 MW
50.591 MW
43.777 MW
14.660 MW
10.632 MW
tb01 tb02
8.210 MW
Winter Summer
Summer
Superstructure of the industrial utility system.
Condensate return
273
274
Chapter Eleven Next, potential design measures for reducing greenhouse gas emissions were evaluated. The sensitivity analysis addressed four cases, as summarized in Table 11.5. Case 1 is the base case and thus represents the conditions discussed so far. The other cases gradually increase the price of power and fuel as well as the penalties on emissions. The optimal utility system flowsheet for case 3 is shown in Figure 11.16. Case 4 also replaces fuel oil with a biofuel, which is assumed to have zero CO2 emissions. Analysis of these emission reduction options leads to the following conclusions: (1) Increasing the system efficiency is the cheapest option for CO2 abatement, but it has a relatively limited scope. (2) The next economic option for this particular problem is to close the carbon cycle by using biofuels; in general, however, CO2 capture and sequestration could also be considered.
Winter
b01
Summer
Uses fuel gas Capacity: 51.334 t/h Winter: 51.334 t/h Summer: 45.049 t/h
b02
Uses fuel gas Capacity: 70.000 t/h Winter: 41.990 t/h Summer: 0 t/h
12.000 MW 16.885 MW
Winter
Summer
6.289 t/h
5.625 t/h
VHP: 90 bara, 503.35 °C
3.550 MW 3.175 MW
POWER 7.700 MW 13.383 MW Capacity: 52.409 t/h hdr : st 01 01 Winter: 4.767 MW Winter: 52.409 t/h Summer: 1.941 MW Summer: 26.213 t/h Winter: 15.717 t/h 1.300 MW 2.877 MW Summer: 0.001 t/h 33.139 t/h 25.240 t/h
(tb03)
1.635 MW 1.501 MW 84.620 t/h
23.116 t/h
Winter: 18.769 t/h Summer: 12.493 t/h Winter: 114.025 t/h Summer: 101.182 t/h
To cooling water
To the stream generators
FIGURE 11.15
Case
Condensate return from processes
Winter: 136.671 t/h Winter: 22.646 t/h Summer: 116.369 t/h Summer: 15.187 t/h Make-up water
Winter: 154.502 t/h Summer: 128.237 t/h
17.597 t/h
14.660 MW 10.632 MW
COND: 0.1235 bara DA
77.960 t/h
50.591 MW 43.777 MW
TSAT=130.51 °C (tb01); P=2.74 bara
8.210 MW 10.632 MW
5.956 MW
Winter: 16.645 t/h Summer: 0.558 t/h
29.533 t/h
15.500 MW 18.136 MW
13.210 t/h
9.038 MW
P=15.19 bara
TSAT=198.89 °C;
52.304 t/h
18.907 t/h
Power import [MW]: 21.1 (Winter); 30.0 (Summer) Total annualized cost: 13.124 · 106 $/y
Optimal utility system for Case 1 (base case).
Price of CO2 [$/t]
Price of SOx [$/t]
Price of NOx [$/t]
Power price [$/MWh]
1
0
0
0
20.00
30.00
2
0
0
0
30.00
45.00
3
40
500
1000
40.00
60.00
4
40
500
1000
40.00
60.00
TABLE 11.5
Parameters
Sensitivity Analysis for Reducing Emissions: Basic
Industrial Applications and Case Studies Power Capacity: 33.917 MW Winter: 30.085 MW Summer: 33.916 MW
gt01
Winter
Summer
Capacity: 28.570 t/h Winter: 0.000 t/h Summer: 0.000 t/h
Winter: 95.230 t/h Summer: 45.022 t/h
Winter
b02
HRSG 12.000 MW 16.885 MW VHP: 90 bara, 503.35°C 7.700 MW
1.300 MW
Power Capacity: 52.017 t/h hdr01 : st01 Winter: 7.722 MW Winter: 52.017 t/h Summer: 1.942 MW Summer: 26.023 t/h
13.383 MW
2.877 MW TSAT = 198.89°C;
33.139 t/h 25.240 t/h
(tb03)
Power Capacity: 20.000 t/h hdr02 : st01 Winter: 0.733 MW Winter: 13.997 t/h Summer: 0.000 MW Summer: 0.000 t/h
29.533 t/h
TSAT = 130.51°C; 10.632 MW
P=2.74 bara
Winter: 19.869 t/h Summer: 12.491 t/h
To cooling water DA
(tb01)
5.625 t/h
3.550 MW
3.175 MW
18.907 t/h
13.210 t/h
9.038 MW
5.956 MW
1.635 MW
1.501 MW
Winter: 18.016 t/h Summer: 0.164 t/h
52.304 t/h
15.500 MW 18.136 MW
8.210 MW
P=15.19 bara
Summer
6.289 t/h
Winter: 4.904 t/h Summer: 0.555 t/h
84.271 t/h
77.935 t/h
50.591 MW
43.777 MW
24.272 t/h
17.597 t/h
14.660 MW
10.632 MW
COND: 0.1235 bara Winter: 114.832 t/h Condensate return Summer: 101.157 t/h from processes Winter: 22.758 t/h Summer: 15.119 t/h Make-up Power export [MW]: 10.0 (Winter); 4.4 (Summer) water Total annualized cost: 23.060 · 106 $/y
Winter: 137.590 t/h Summer: 116.343 t/h To the stream generators
FIGURE 11.16
11.6
Winter: 156.466 t/h Summer: 128.209 t/h
Optimal utility system for Case 3.
Heat and Power Integration in Buildings and Building Complexes Herrera, Islas, and Arriola (2003) studied a hospital complex that included an institute, a general hospital, a regional laundry center, a sports center, and some other public buildings. The use of diesel fuel represented 75 percent of its total energy consumption and 68 percent of its total energy cost, which was $396,131 in 1999. In the hospital complex, the heat demand is met by producing steam in boilers fueled by high-price diesel fuel. There is no heat recovery between the existing heat sources and heat sinks. The hot streams were identified as the soiled soapy water from the laundry and the flow of condensed steam not recovered in the condensation network. The stream data are presented in Table 11.6. For this hospital complex, the amount of external heating required (i.e., the hot utility target) is 388.64 kW, which can be seen on the hot and cold Composite Curves in Figure 11.17. This plot was employed to determine what temperature levels of the utilities would satisfy this requirement. The heating utility target of 388.64 kW translates to an annual energy requirement of 12.26 TJ/y. The amount of heating provided is actually 625.28 kW, which represents the heat services that are currently transferred to the complex. This figure represents a potential for energy savings of 38 percent, which is equivalent to an annual reduction in diesel fuel of 246,000 liters (worth about $100,000). To reduce heating energy demands to the targeted value, the Heat Integration Analysis
275
276
Chapter Eleven Stream
Name
T supply [°C]
T target [°C]
DH [kW]
CP [kW/°C]
1 Hot
Soapy water
85
40
23.70
0.53
2 Hot
Condensed steam
80
40
96.32
2.41
3 Cold
Laundry sanitary water
25
55
17.60
0.59
4 Cold
Laundry
55
85
77.27
2.58
5 Cold
Boiler feed water
30
60
7.13
0.24
6 Cold
Sanitary water
25
60
77.12
2.20
7 Cold
Sterilization
30
121
12.50
0.14
8 Cold
Swimming pool water
25
28
151.67
50.56
9 Cold
Cooking
30
100
59.63
0.85
10 Cold
Heating
18
25
100.82
14.40
11 Cold
Bedpan washers
21
121
4.94
0.05
TABLE 11.6
Process Stream Data of Hospital Complex (Herrera, Islas, and Arriola, 2003) T [°C] 120
Hot CC Cold CC
100 80 60 40 PINCH 20 0
FIGURE 11.17
80
160
240
320
400
480
ΔH [MW]
Process Composite Curves for hospital complex (∆Tmin = 20°C).
indicates that four extra heat exchangers should be added to the network. Two are needed in the laundry to cover part of the heat demand, a third is needed in the machinery rooms that help to heat boiler feed water, and a fourth is needed in the condensation tank area that heats the sanitary water. The analysis could be refined
Industrial Applications and Case Studies further by considering several other issues, such as fouling, pressure drop, and variation in the heat demand.
11.7 Optimal Design of a Supply Chain In the global market, optimal management control of a company is necessary for survival. Such control involves a series of effective strategic decisions that are concerned with various aspects of the supply chain—for example, decisions regarding the products themselves as well as the location of production facilities and distribution centers. When the subject is the development or operation of highly complex business processes (e.g., supply chains, value chains), computer-aided decisions are preferred. A conventional approach for systematically evaluating decision alternatives is Mathematical Programming, although a daily management problem is usually not resolved via MPR methods. Even when the MPR can be constructed for the problem, it is still difficult to verify whether the mathematical formulation of relevant decision alternatives is sufficiently accurate and complete to identify the optimal solution. When the mathematical model is not generated systematically, the model’s chances of embodying the optimal solution are very low. This case study involves determining the optimal design of a supply chain. Here, the purpose of the supply chain is to meet a given volume demand for commodity C at location L1. Three options are considered: produce commodity C at location L1; produce commodity C at location L2 and then transport it to location L1; or some combination of these. Production requires the availability of part A and part B at the same location. Part A is available at location L3 and, to a limited extent, at location L4; it can be transported to location L1 or to location L2. Part B is available at location L2 and can be transported to location L1. The list of potential activities is given in Table 11.7. ID
Activity
Location
Precondition
Effect
PL1
Production
L1
A and B at L1
C at L1
PL2
Production
L2
A and B at L2
C at L2
TAL3L1
Transportation
From L3 to L1
A at L3
A at L1
TAL3L2
Transportation
From L3 to L2
A at L3
A at L2
TAL4L1
Transportation
From L4 to L1
A at L4
A at L1
TAL4L2
Transportation
From L4 to L2
A at L4
A at L2
TBL2L1
Transportation
From L2 to L1
B at L2
B at L1
TCL2L1
Transportation
From L2 to L1
C at L2
C at L1
TABLE 11.7
Potential Activities in the Supply Chain Case Study
277
278
Chapter Eleven Given the list of potential activities defined by their preconditions (activating entities) and effects (resulting entities), the Maximal Structure Generation (MSG) algorithm then produces the maximal structure (Friedler et al., 1993). Next, when applied to the maximal structure so derived, the Solution Structures Generation (SSG) algorithm (Friedler et al., 1995) enumerates 15 combinatorially feasible business process structures for the problem. To determine the optimal business practice, the following quantitative information is provided in addition to the case study’s 15 structural alternatives. The required volume of the demand is 20,000 pieces annually. Producing one piece of commodity C requires the availability of one piece of part A and one of part B. At most 5000 pieces of part A are available at location L4 for €230 each. An unlimited number of part A can be purchased at location L3 for €250 each, and an unlimited number of part B can be purchased at location L2 for €310 each. The cost of an activity depends on its volume. To estimate the increase in the cost of an activity as a function of its volume, costs are given for handling 1000 and 2000 pieces per year. If a linear cost function with a fixed charge is adopted to estimate the cost of the activities with different volumes, then this function is determined by the fixed charge and the proportionality constant. Table 11.8 summarizes the cost of processing 1000 versus 2000 pieces per annum as well as the parameters of the fixed-charge linear cost function for each activity. The optimal business process, as determined by the Accelerated Branch-and-Bound (ABB) algorithm (Friedler et al., 1996), corresponds to the activities given in Table 11.9 whose annual cost totals
Activity ID
Cost of processing
Cost function parameters
1000 pcs/y [€/y]
2000 pcs/y [€/y]
Fixed charge [€/y]
Proportionality constant [€/pcs/y]
PL1
8,000
10,000
6,000
2
PL2
9,000
10,000
8,000
1
TAL3L1
4,000
8,000
0
4
TAL3L2
10,000
20,000
0
10
TAL4L1
12,000
24,000
0
12
TAL4L2
2,000
4,000
0
2
TBL2L1
10,000
20,000
0
10
TCL2L1
14,000
28,000
0
14
TABLE 11.8
Cost of Activities in the Supply Chain Case Study
Industrial Applications and Case Studies Activity ID
Optimal
2nd best
3rd best
Volume [pcs/y] PL1
15,000
PL2
5,000
TAL3L1
15,000
20,000 20,000 15,000
TAL3L2
15,000
TAL4L1
5,000
TAL4L2
5,000
TBL2L1
15,000
TCL2L1
5,000
Total cost [€/y] TABLE 11.9 Processes
11,439,000
5,000 20,000 20,000 11,466,000
11,568,000
Activities in the Optimal, Second-Best, and Third-Best Business
€11,439,000. The second-best business process has a total annual cost of €11,466,000, and the third-best business process has a total annual cost of €11,568,000.
11.8 Scheduling a Large-Scale Paint Production System Paint production usually consists of three major operations: grinding and dispersion, mixing and coloring, and discharging and packaging. Paints and coatings are typically produced in batches. They are made in stationary and portable equipment units such as high-speed dispersion mixers, rotary batch mixers, blenders, sand mills, and tanks. The raw materials are solvents, resins, pigments, and additives that include inorganic and organic chemicals. Paint manufacturing does not usually involve chemical reactions between the raw materials, so the finished product consists of a mixture of the different raw materials. Several dozens of products are produced at the manufacturing site, so the corresponding scheduling problem is bound to be highly complex. The S-graph framework of batch scheduling (see Chapter 7) has been extended to solve complex paint production problems (Adonyi et al., 2008). Changeover time is defined for any equipment unit that requires cleaning. Traditionally, minimizing makespan (total time to completion) is the criterion used when assigning equipment units to tasks and scheduling the tasks. Such schedules maximize the production system’s efficiency, but they may lead to unnecessarily high levels of waste generation. Thus, determining which task schedule minimizes cleaning cost will require that the problem’s objective function be modified. Now, rather than minimizing makespan as in the original problem, the reformulation seeks to
279
280
Chapter Eleven minimize the cleaning cost. This change in criterion has only a minor effect on the solution procedure, so an effective solver for the original problem is also useful for the reformulated problem. Twenty-three equipment units, E1 through E23, are available to generate six products, A through F. The changeover time is 70 minutes for equipment units E6, E7, E8, and E9 but 100 minutes for equipment units E1 through E5 and E10 through E20. All other changeover times are presumed to be zero. The number of batches to be produced is given in Table 11.10. Cleaning the equipment units is a costly operation that involves many pollutants. The minimal makespan schedule contains 11 cleaning operations, which are denoted by the dotted changeover arcs on its S-graph; see Figure 11.18. The cleaning cost of the solution with minimal makespan is $14,000. In contrast, the solution based on minimizing the cost involves four (rather than 11) cleaning operations and only $3,500 in cleaning cost; its makespan is 6,910 minutes. If the cleaning cost is limited to reach $5,500, then the corresponding makespan is reduced to 6,700 minutes. Product
A
B
C
D
E
F
Number of batches
3
5
1
3
9
3
TABLE 11.10
E1
60
E6
60
E11
120
0
0 E1
310
Number of Batches Produced of Each Product
E6
310
E11
120
E11
120
E15
98
E22
540
99
120
E22
540
E6
310
E1
60
E7
240
E1
60
E7
240
E1
60
E19
120
E22
E7
240
E11
100 120 E22 540
E1
60
E1
60
E7
E2
60
E8 120
60
E7
0
0 0
0 E7
102
60
E22
540
240
E15
100 120 E22
540
E16
50
E22
540
E18
90
E22
E14
90
E22
540
107
E20
60
E23
720
108
103
0 104
0
E3
60
240
E7
0
240
540
106
0
0
0 60
0 E7
70 E4
A
B 70
E4
240
70
E4
B
E4
B
D
40
E6
120
E10
60
E6
120
E12
90
0
0 40
E21
E6
E6
120
0
0 E10
60
300
E12
E6
300
0
60
0
40
E6
300
90
0
0 40
E6
300
E6
300
0
60
90
E6
40
E7
300
114
0 720
115
0 E21
0
40
720
E21 0
E12
720 113
E21 0
E10
0
0 40
0 E21
0 E12
720 112
E21 0
E10
720 111
E21
90
0
0 40
109
720 110
E21
0
40
720 0
0 E4
720
116
0 E16
90
E21
720
117
100 E5
D
D
0
0 E4
B
C
40
0 E4
100 105
0
E3
A
540 101 B 0
E13
70
E3
E4
0
240 0
0
100
0
0
0
E4
A
0
70
0
0
E22 0
60
97
0 540
0
0 E1
540
E22 0
100
40 0
E7
0 E5
E19
120
E23
0
0 E5
240
40
E7
240 0 240
E20
60
E23
720 0 720
118
119
0 E15
120
E23
720
120
E
E
E
E
E
E
E
E
E
F
F
F
100
FIGURE 11.18 Schedule graph of the solution that minimizes makespan (after Adonyi et al., 2008).
CHAPTER
12
Typical Pitfalls and How to Avoid Them
P
rocess Integration (PI) has proven to be a powerful optimization tool for designing processes that are energy-efficient, environmentally friendly, and sustainable. The methodology provides clear insight into the design process, but this direct simplicity is sometimes misunderstood by potential users. Once proper data are made available, the procedure generates an excellent lead for the design process. But just as with all optimization tools, potential pitfalls when using PI include improper formulation of the problem and incorrect data extraction. Even when the most efficient and well-developed methodology is used to solve an optimization problem with high precision, the results may be suspicious unless we have been solving the right problem. In other words, has the problem been formulated in a way that closely reflects the real process under consideration and (especially) has the correct data been extracted? Negative answers to these questions explain such published statements as: “Pinch technology [or process integration] did not work for this problem.” When these problems are revisited it usually becomes obvious that the PI methodology is not at fault; rather, an inexperienced or overconfident user is typically the cause. Therefore, this chapter is devoted to various mistakes that a designer might unwittingly make. The most basic issue concerns how one starts a PI-based project and how it is run. Kemp (2007) summarized the key steps listed below, which have been further developed based on the authors’ experience. These steps are specifically related to Heat Integration; however, they apply with only small adjustments to mass, water, and other integration as well. 1. Become familiar with the analyzed process. The most efficient way is to closely liaise with the process designer and/or plant manager, especially if the plant is already operating.
281
282
Chapter Twelve 2. Develop a mass and heat balance. This should be based on the designed process flowsheet data and calculations and/ or on measurements taken from the operating plant (if the study is for a retrofit). 3. Select the streams. This is a critical step and, as will be shown in this chapter, not as straightforward as it may seem. 4. Remove all the existing units related to the PI analysis. For Heat Integration, remove all heat-transferring units, for mass water integration remove all water interconnections (the pipes). This step is also critical—without it, the optimized design would not differ from the initial design. Section 4.2.4 provides an example illustrating this activity. 5. Extract the stream data for the PI analysis. Different data are relevant for each PI analysis type. For HI heat loads and temperatures are extracted. 6. Make a qualified initial guess for the ΔTmin value; this value can be adjusted later at various stages of the design optimization. 7. Perform the Pinch analysis: obtain the Pinch temperatures and the utility targets. 8. Design the initial (heat exchanger) network using the criterion of maximizing energy recovery. 9. Check for a cross-Pinch transfer and for inappropriate placement of utilities. 10. Check for proper placement of reactors, separation columns, heat engines, and heat pumps. 11. Investigate the potential for further modifying the process in order to minimize energy consumption and reduce capital costs. Investigate the potential benefits of applying the plus-minus principle (see Chapter 4) and the KHSH and KCSC principles (see Figure 4.44). 12. Investigate the potential for integration with other processes— that is, Total Site Analysis. 13. Consider the implications of pressure drop (trade-offs between heat savings and extra energy for pumping) and the physical layout (capital cost of heat exchangers and/or piping). 14. Make the preselection of heat exchange equipment and perform the preliminary costing. Provision should be made for variations in the future price of energy. 15. Make the first optimization run of the predesign plant or site, and make adjustments to ΔTmin.
Typical Pit falls and How to Avoid Them 16. Based on the optimization, extract adjusted data and return to step 7. Perform an additional loop (or loops) while screening and scoping for potential simplifications. 17. Consider real plant constraints; these include safety, technology limitations, controllability, operability, flexibility, availability, and maintainability. 18. Pay attention to start-up and shutdown of the process; some early designs for highly integrated plants had problems in this area. 19. Run a second optimization for the final tuning accounting for the information added during steps 16 to 18. If necessary, return to any appropriate previous step for adjustment. 20. The design is now ready for detailing. However, optimization is a never-ending procedure, and designs may need to be modified in response to changes in operating conditions (e.g., plant capacity) or the economic environment (e.g., tax policy; prices for energy, materials, and production).
12.1
Data Extraction As emphasized previously, data extraction is a crucial step. Bodo Linnhoff presented one of his last plenary lectures (Linnhoff and Akinradewo, 1998) on the automated interface between simulation and integration. This has been a substantial step toward data extraction for PI software tools. The plenary has been fairly comprehensive and suggested the way forward for this important task. The problem received increased interest following this lecture, and several software packages now offer support for solving it. Nonetheless, more work in this area is needed to satisfy the requirements of routine industrial applications. In their SuperTarget and Pinch Express software packages, Linnhoff March (1998) included procedures for automatic extraction of data for Heat Integration. Even so, thermal data, which involve the stream heating and cooling information and utilities information, are the most critical data required for Pinch Analysis. There are several possibilities for extracting the thermal data from a given heat and material balance. This must be done carefully, as poor data extraction can easily lead to missed opportunities for improved process design. In extreme cases, poor data extraction can falsely present the existing process flow-sheet as optimal in terms of energy efficiency. If the data extraction accepts all the features of the existing flow-sheet then there will be no scope for improvement. If it does not accept any features of the existing flow-sheet then Pinch Analysis
283
284
Chapter Twelve may over-estimate the potential benefits. Appropriate data extraction accepts only the critical sections of the plant which cannot be changed. Data extraction skill develops with increased experience in the application of Pinch Technology (Linnhoff March, 1998). Since the release of these packages, the methodology has developed further and more attempts have been made to extract data automatically. However, experience and following the proper rules remain valuable assets. Basic questions to ask include the following: 1. When is a stream a stream? 2. How precise must the data be at each step? 3. How can considerable changes in specific heat capacities be handled? 4. What rules and guidelines must be followed to extract data properly? 5. How can the heat loads, heat capacities, and temperatures of an extracted stream be calculated? 6. How “soft” are the data in a plant or process flowsheet? 7. How can capital costs and operating costs be estimated?
12.1.1 When Is a Stream a Stream? To those unfamiliar with PI, identifying a stream seems fairly straightforward. In fact, many considerations are involved, and accounting for them properly is key to setting up the problem. First, we need not consider any stream that neither gains nor provides heat; no data needs to be extracted from a stream with identical supply and target temperatures and enthalpies. (Of course, in the absence of perfect insulation, every stream loses or gains some heat; in many cases, however, these small amounts of losses and gains can be neglected.) If we do not extract data from such streams, the problem is considerably simplified. There are also streams that, for one reason or another, should not be included in the PI problem—for example, streams that are remote and streams that should not be altered for safety, product purity, and operational reasons and for other (mostly practical) considerations. Finally, Heat Integration deals with heat flows, which can be carried not only by a pipe line but also by radiation or conduction. Consider the example depicted in Figure 12.1. This example was introduced, along with some data extraction rules, by Linnhoff and colleagues (1982) and was later modified for use in many follow-up books (e.g., Smith, 1995; Smith, 2005; Kemp, 2007) and in many courses based on UMIST (later the University of Manchester) teaching materials. The figure shows part of a flowsheet in which the feed stream is heated to 45°C by recuperated heat in a heat exchanger and then enters a processing unit. After leaving this unit, the stream is
Typical Pit falls and How to Avoid Them
150° 160° ΔH3 190° 120° Reactor
40°
80°
45°
10°
45°
ΔH2 140°
FIGURE 12.1
ΔH1 Unit
Feed
140°
Partial flowsheet for example process.
heated further by two heat exchangers and then enters a reactor. The reactor requires the feed stream to be at 165°C. The question is: How many streams should be extracted? 1. Three streams: 10–45°C, 45–80°C, and 80–165°C 2. Two streams: 10–45°C and 45–165°C 3. One stream: 10–165°C Choosing option 1 yields a design exactly like the original one; there will again be three heat exchangers with the same heat transfer duties as before. This case accounts for claims of “no process improvement” by some PI critics. Option 2 presents more degrees of freedom: the first heat exchanger would be the same, but the other could be modified. Extracting two streams would be appropriate for cases when the processing unit requires a feed temperature of close to 80°C. Option 3 provides the most degrees of freedom and scope for improvement, but with this design the processing unit feed could be at any temperature between the 10°C supply and the reactor target of 165°C. If the processing unit is, say, a filter (as assumed by Smith, 2005), then there would probably be some restriction on the filter supply temperature to ensure proper operation of the filter. If the processing unit is storage (as assumed by Linnhoff et al., 1982), then the supply temperature might be restricted to a different range (depending, e.g., on whether liquid or gas is stored). This simple example demonstrates that choosing the right stream from which to extract data cannot be a fully automatic process. Making this choice requires assessments related to specific processing units and performance requirements for the plant.
12.1.2
How Precise Must the Data Be at Each Step?
There are frequently questions about required data precision, and a common excuse for not applying PI analysis is that the data on a plant is not sufficiently precise. However, the process of applying PI
285
286
Chapter Twelve is itself based on (initially) rough assumptions, which are revised during the course of several design loops. At the beginning there are no specifications for the heat-transferring units, and neither are the feed temperatures (and several other important factors) fully fixed. At this stage, then, extremely precise data is not needed. It is important to recognize that PI and the initial optimization are more about screening and scoping than about detailed design. In seeking potential energy savings, it is the general direction in which optimization should proceed that is the initial concern. If the result of this first step is that a 15 percent savings in energy is possible, then this figure is sufficient to confirm the design approach and it doesn’t much matter if the precise figure is actually 13 or 17 percent. It is in the regions close to the Pinch that the data should be as precise as possible (Linnhoff et al., 1982). Also, it is best to remain inside the Composite Curves in the plot of temperature versus enthalpy. At the start of data extraction, we might have only a vague idea of where (and at what temperature) the Process Pinch will occur; also, the Composite Curves are based solely on data extraction. Therefore, data extraction must necessarily start from rough assessments and then be corrected step by step.
12.1.3 How Can Considerable Changes in Specific Heat Capacities Be Handled? Further analysis of the flowsheet in Figure 12.1 reveals that phase changes are very likely to occur when the temperature increases from 10°C to 165°C. In general, Cp varies with temperature, but latent heat is also a determining factor. Clearly, using a constant value for Cp would not be sufficiently realistic. A segmentation technique has been developed to deal with this problem. This technique is used, for example, in the software tools SPRINT (2009) and STAR (2009). The software tools treat the segments as individual streams that are combined to form input and output streams. Data extraction is affected by how many segments are used and their boundary temperatures (see Figure 12.2). Increasing the number of segments naturally increases the complexity, so this
T
1
T
?
T
? ΔH
FIGURE 12.2
2
3
T
? ΔH
How should we linearize?
4
? ΔH
ΔH
Typical Pit falls and How to Avoid Them number should be minimized for industrial problems involving many streams.
12.1.4 What Rules and Guidelines Must Be Followed to Extract Data Properly? As mentioned in Section 12.1.1, data extraction rules were first introduced by Linnhoff et al. (1982) and frequently used with some modifications thereafter (Smith, 1995; CPI, 2004 and 2005; Smith, 2005; Kemp, 2007). Most of this work is related to Heat Integration, but the principles apply as well to mass (water) integration. These rules are reviewed briefly in this section. When two or more streams of different temperatures are mixed, this nonisothermal mixing constitutes a heat exchange with degradation of the higher temperature. In some cases, such mixing can also cause cross-Pinch transfer problems—as in Figure 12.3(a),
FIGURE 12.3 Nonisothermal stream mixing extracted as (a) three streams and (b) two streams.
(a)
160°
110° 70°
40°
75°
110°
35°
160°
40°
70°
35°
75°
(b)
160°
110° 70°
40°
75°
35°
35°
160°
35°
70°
287
288
Chapter Twelve where three streams are extracted. The correct data extraction for this case involves two streams, as shown in Figure 12.3(b). General guidelines for data extraction may be summarized as follows: 1. Heat losses. In most cases, heat losses can be neglected. However, they should not be neglected when streams (mass and heat flow in pipes) are long or subject to varied temperatures. In such cases, the solution is to introduce hypothetical coolers (or heaters) that represent the heat loss. 2. Extracting utilities. The utilities should never be extracted from the existing plant or flowsheet, for then the solution would likely arrive at the same utility values and perhaps neglect some options that would be more efficient. This rule of thumb applies especially to cases where utilities can be generated on-site and thereby (at least partially) replace costly existing utilities. In this connection it should be remembered that steam is not always a utility—sometimes it is also a process stream (e.g., stripping steam in separation columns). Process streams should remain in place and not be removed. 3. Generating utilities. The Heat Integration analysis may indicate some valuable options for using otherwise wasted heat or cold to generate utilities. The Grand Composite Curves can be used to locate such options. However, when extracting data it must be recognized that steam requires the boiler feedwater (BFW) to be heated, water to be evaporated, and the steam to be superheated; see Figure 12.4. Many mistakes have been caused by designers who simply matched the T Specified Steam Conditions
BFW Conditions ΔH
FIGURE 12.4 Extraction of a cold stream including segments for BFW preheating, evaporation, and superheating.
Typical Pit falls and How to Avoid Them steam generation line without making provisions for preheating and superheating. 4. Extracting at the effective temperature. In some cases a stream cannot be extracted directly because it has to still be used by a related process. For example, a hot stream should be extracted at temperatures at which the heat becomes available. An example is given in Smith (2005, p. 433) for a reactor using a quench liquid. 5. Forced and prohibited matches. In almost any process there exist matches that are necessary for technological reasons—the forced matches—as well as those, such as hot and cold stream matches in a heat exchanger, that should be prohibited (e.g., to prevent contamination of one of the streams). In manual design these constraints have to be observed by the designer, but software tools usually offer this option. If not then the constraints can be secured by an appropriate penalty or bonus (as applies) in the objective function used for the optimization. 6. Keeping streams separate only when necessary. If streams can be merged then it may be possible to eliminate some heatexchanging units. For example, streams that leave the plant to be treated as wastewater often have some heat content that can be utilized.
12.1.5 How Can the Heat Loads, Heat Capacities, and Temperatures of an Extracted Stream Be Calculated? Once a stream has been extracted, the next problem is calculating the heat-related data. There are standard engineering procedures available for the running plants as the measurements with the following data reconciliation (Klemeš, Luťcha, and Vašek, 1979; Minet et al., 2001; VALI III User Guide, 2003). The other option is to develop a flowsheeting simulation model (Klemeš, 1977). An overview of flowsheeting and balancing simulators was given in Chapter 9. If a plant is being designed, some data could be also extracted from the process flow diagram (PFD). But all those options consume time and resources, so in the early design stages (when the process structure is still under development and likely to be changed as a result of PI analysis), it is reasonable and easier to use a simplified approach based on the extracted data. Such an approach is demonstrated in Figure 12.5 for a part of the flowsheet from Figure 12.1. The CPs of the stream segments are assumed constant and calculated from the temperatures and the duties given in the flowsheet. Experience has indicated that the resulting rough preliminary data are sufficient and can later be made more precise by one or more of the procedures listed previously.
289
290
Chapter Twelve T 160°
80° 45° 150° 160°
10°
ΔH3
ΔH1 ΔH2
ΔH
ΔH3
190° 120° Reactor
40°
80°
45°
45°
10°
ΔH2 140°
ΔH1 Unit
Feed
140°
FIGURE 12.5 Obtaining rough data from flowsheet heat loads and temperatures.
12.1.6 How “Soft” Are the Data in a Plant or Process Flowsheet? Distinguishing soft data from hard data is one of the most important aspects of data extraction. Inexperienced persons are usually trying to stick the temperatures shown in the PFD, extract those temperatures, and then perform the PI analysis. However, this approach usually ends up overlooking many opportunities. A better approach is to question every temperature, discuss each one with the process engineer (or plant designer or plant manager), and thereby establish which temperatures are critical (the “hard” data) while the rest (the “soft” data) can be in some way compromised. In practice, most data are at least a little soft, and designers can use this fact to their advantage. Typically, streams that are leaving the plant (see Figure 12.6) are characterized by soft data and thus are suitable for optimization via the plus-minus principle (Figure 12.7). Data softness is closely related to changing conditions and to a design’s flexibility, operability, and resilience.
12.1.7 How Can Capital Costs and Operating Costs Be Estimated? The need to find cost data arises when the appropriate ΔTmin (which should be close to the optimum) is being selected. The optimum ΔTmin depends strongly on economic parameters, and its value is important for both grassroots design and retrofit. Estimating capital costs is usually a time-consuming procedure. However, it is possible to use
Typical Pit falls and How to Avoid Them
110°
70°
Storage 70°
110°
Cooler
FIGURE 12.6
20°
Storage 20°
Cooler
Soft data for streams leaving a plant.
T
T Decrease (–) Hot Stream
Reduced Cold Utility Target
Increase (+) Hot Stream
Reduced Hot Utility Target
ΔH T
ΔH T
Increase (+) Cold Stream
Reduced Cold Utility Target ΔH
Decrease (–) Cold Stream
Reduced Hot Utility Target ΔH
FIGURE 12.7 The plus-minus principle can be used to optimize application targets by using properly extracted soft data.
approximate methods (Taal et al., 2003) at the initial design stage, when little is known about the types of heat transfer units to be used, the ultimate design, the materials required, or the temperature, pressure, and composition of streams. The estimates so derived will usually suffice until the final detailed design cost date, at which time information is obtained from selected manufacturers. Note, however, that equipment cost may vary regionally and may also be related to market conditions (e.g., a large customer can secure discounts, and prices fall during a recession). It is much more difficult to establish operating cost, which is affected by labor, taxation, and so forth but is mainly a function of energy cost. Here the most obvious problem—and greatest potential pitfall—is using the current price of energy. (It is possible to find
291
292
Chapter Twelve many publications and projects where this rule was not followed.) It is better to use the anticipated average energy price for the life span of the plant or, in the case of a retrofit, for the payback period. The problem then becomes one of estimating this future energy price for periods that may be as long as five or ten years. It has been shown (see, e.g., Klemeš and Bulatov, 2001; Donnelly, Klemeš, and Perry, 2005) that even the forecasts of highly qualified experts are frequently inaccurate. One of the obvious potential approaches is to use the scenarios and target the most flexible design that would provide a balanced optimum for various situations.
12.2 Integration of Renewables: Fluctuating Demand and Supply The integration of renewable energy sources was discussed in Chapter 6. Renewable resources are usually available on a smaller scale and are often distributed over a certain region. Their availability (with the exception of biomass) varies significantly with time and location. This variability is due to changing weather and geographic conditions. The energy demands (heating, cooling, and power) of sites also vary significantly with time of the day and period of the year. These variations in the supply and demand of renewables can be predicted in part, and some of the variation is fairly regular—for instance, day versus night in predominantly cloudless areas for solar energy. The availability of wind-generated energy can be less predictable. One approach to dealing with these problems is the advanced PI technique that employs time as an additional problem dimension. A basic methodology along these lines (involving “Time Slices” and Time Average Composite Curves) was developed for the Heat Integration of batch processes (Klemeš et al., 1994; Kemp and Deakin, 1998). This methodology was recently revisited by Foo, Chew, and Lee (2008). This methodology has also been extended to the Heat Integration of renewables. Important steps in this direction were reported by Perry, Klemeš, and Bulatov (2008) and Varbanov and Klemeš (2010). Dealing with variation and fluctuation brought another complexity into data extraction. Data should be collected for all time slices that increase the complexity. Especially, for each case, it is necessary to choose the time horizon for the analysis and number of time slices. This is a fast-developing field, so it is advisable to monitor recently published research papers and conference presentations.
12.3 Steady-State and Dynamic Performance It has been assumed that all analyzed and optimized processes operate in a steady state. Many industrial processes do operate in this
Typical Pit falls and How to Avoid Them fashion, and a common control task is to maintain such processes in a steady state. However, there are also many situations in which the working regime must be changed. This occurs not only at start-up and shutdown but also in response to changes in operating conditions (e.g., outside temperatures) and production conditions (e.g., capacity, volume). Problems involving such variation are generally addressed in one of two ways: (1) accommodating various scenarios via a design objective of minimizing sensitivity to the possible changes or (2) designing to optimize the total system’s overall dynamic performance. The second option is much more complicated and requires more time and resources; hence it is used only when substantial deviations from steady-state operations are anticipated.
12.4
Interpreting Results After data extraction, the most important aspect of PI analysis and optimization is the correct interpretation of results. These results are typically presented in the form of a printout generated by some software tool; in most cases, the software output is in the form of a grid diagram (e.g., STAR, 2009) or PFD supported by printout tables—as with ASPEN (AspenTech, 2009a, 2009b, 2009c) and gPROMS (PSE, 2009). In order to minimize typos and misinterpretation, many software tools feature an interface that facilitates the transfer of data files. This technology is now fairly well developed. The most challenging aspect of data interpretation is assessing the results in terms of possible further development or correction of the important process features. The interpretation will depend on the extent to which the factors mentioned previously (i.e., data uncertainty, data softness, flexibility, operability, controllability, safety, availability, and maintenance) have already been incorporated into the data extraction. If the data extraction did not reflect these considerations, then they must be addressed during the data interpretation. At this stage, the close collaboration of a team that includes the main involved professionals is highly recommended. Designers are strongly advised not to stick with just one solution but rather to explore different potential scenarios associated with various operating conditions and then test their design’s sensitivity to the possible variations. For screening and scoping it is helpful to use all types of Composite Curves as well as information (if provided by software) on which streams are contributing to specific parts of the curves.
12.5 Making It Happen Even when a sustainable and near-optimum design is developed, it must still be put into practice. This involves selling the
293
294
Chapter Twelve proposals—which are often unconventional—to plant management, investors, and contractors. This was a big problem when PI was just beginning, and great strides in this area were made by UMIST, Bodo Linnhoff, and his company Linnhoff March in the 1980s and 1990s. PI has since proven itself and gained in popularity, so decision makers have become more receptive. Much of the situation’s improvement is due to multinational companies that have incorporated PI into their design and operational practice. Because the methodology has become widespread, it is not possible to list all of these companies. However, among the pioneers were members of the UMIST and, after the merger, the University of Manchester Process Integration Research Consortium: Air Products, Aspen, BASF, Bayer, BOC, BP, Canmet, Degussa, EDF, Engineers India, Exxon Mobil, Hydro, IFP, JGC, KBC, Mitsubishi Chemical Corporation, MOL, MW Kellogg, Petrobras, Petroleum Research Centre, Petrom, Petronas, Saudi Aramco, Shell, Sinopec, Technip, Total, UOP, and Vito. These firms were joined in the consortium by several universities, including University POLITEHNICA Bucharest and Petronas Technological University. There have also been strong supporters of PI in the United States, both at universities and in the industry; some of them are listed in Chapter 13. A major goal remains a close collaboration and smooth joint effort among PI specialists and the designers, managers, and owners (or contractors) of processing plants. A project’s best chance of success is when all these stakeholders share the goal of developing an optimized and sustainable process.
CHAPTER
13
Information Sources and Further Reading
T
here are various sources of information on optimization and integration in the process industry. The aim of this chapter is to summarize many of these sources, although their number is growing rapidly. No such list could be fully comprehensive, but every attempt has been made to include the most important sources of information. The chapter is divided into five sections as follows: (1) general sources of information, (2) Heat Integration, (3) Mass Integration, (4) combined analysis, and (5) optimization for sustainable industry. Within each section, listings are further divided into four groups: conferences, journals, service providers, and projects. A listing is repeated when the information is relevant to more than one category; this makes searching more efficient for users.
13.1
General Sources of Information 13.1.1
Conferences
Conferences that address Process Integration (PI) approaches to minimizing the use of energy, water, and other resources can be sorted into three groups. The first group consists of conferences that are directly related to energy and resource minimization. • Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction (PRES), organized annually since 1998, <www.conferencepres.com> • European Symposium on Computer Aided Process Engineering (ESCAPE), organized annually since 1992, <www.cape-wp.eu>
295
296
Chapter Thirteen • AIChE International Congress on Sustainability Science and Engineering (ICOSSE), organized since 2009, <www.aiche. org/IFS/Conferences/index.aspx> • Dubrovnik Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES), organized biannually since 2002, <www.sdewes.fsb.hr> • Italian Conference on Chemical and Process Engineering (ICheaP), organized biannually since 1993, <www.aidic.it/ italiano/pubblicazioni/confse.htm> • Europe Energy Efficiency Conference, <www.esv.or.at/esv/ index.php?id=1484&L=1> The second group includes large conferences organized throughout the world and for which PI topics are part of their scientific programs. • AIChE annual and spring meetings, specialty conferences, and cosponsored conferences, <www.aiche.org/Conferences> • European Congress of Chemical Engineering (ECCE), organized biennially • Chemical Engineering, Chemical Equipment Design and Automation (CHISA), organized biennially since 1962 (odd years), <www.chisa.cz> • World Congress of Chemical Engineering,<www.wcce8.org/ index.html> (the 9th World Congress of Chemical Engineering, incorporating the 14th APCChE Congress, will be held in 2013) • World Bioenergy, <www.elmia.se/worldbioenergy/> • World Sustainable Energy Days, <www.wsed.at> • World Renewable Energy Congress: Innovation in Europe, <www.wrenuk.co.uk> The third group of conferences includes those dealing with specific issues (food processing, pulp and paper, etc.) and at which energy and resource efficiency are addressed as special cases. • Total Food, organized biennially, <www.ifr.ac.uk/totalfood 2009> • TAPPI EPE (Engineering, Pulping and Environmental) Conference, organized annually, <www.tappiepe.org> • CLIMA Congress, official congress of the Federation of European Heating and Air-Conditioning Societies (REHVA), <www.clima2010.org>
Information Sources and Further Reading
13.1.2
Journals
There are many journals that cover topics related to energy, water, and resource minimization from the Process Integration standpoint; just a few of them are listed (alphabetically) here. • Chemical Engineering Transactions, <www.aidic.it/cet> • Journal of Clean Technologies and Environmental Policy, <www. springer.com/environment/sustainable+development/journal/ 10098> • Journal of Cleaner Production, <www.elsevier.com/wps/find/ journaldescription.cws_home/30440/description#description> • Resources, Conservation & Recycling, <www.elsevier.com/ wps/find/journaldescription.cws_home/503358/description #description>
13.1.3
Service Providers
The following alphabetical list includes service providers as well as professional bodies and networks. • ADAS, Inside and Solutions, Woodthorne, Wergs Rd., Wolverhampton, WV6 8TQ, UK, <www.adas.co.uk/contact/ index.html> • Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843, USA, <www.che. tamu.edu>, contact person: Mahmoud M. El-Halwagi • AspenTech, Aspen Technology, Inc., 200 Wheeler Rd., Burlington, MA 01803, USA, <www.aspentech.com> • Dansk Energi Analyse A/S, <www.dea.dk> • Department of Chemical Engineering, Auburn University, 212 Ross Hall, Auburn University, AL 36849-5127, USA, <eng. auburn.edu/programs/chen/about/index.html> • BIS, Department for Business, Innovation and Skills, UK, <www.bis.gov.uk> • Centre for Advanced Process Decision-Making, Carnegie Mellon University, Department of Chemical Engineering, 5000 Forbes Ave., Pittsburgh, PA 15213, USA,
• Center for Engineering and Sustainable Development Research, De La Salle University, Manila, 2401 Taft Avenue, 1004 Manila, Philippines, <www.dlsu.edu.ph/research/ centers/cesdr/strg.asp>, contact person: Raymond Tan • Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of
297
298
Chapter Thirteen Manchester (formerly UMIST), Manchester, M13 9PL, UK, <www.ceas.manchester.ac.uk/research/researchcentres/ centreforprocessintegration>, contact person: Robin Smith • Centre for Process Integration and Intensification (CPI2), European Community Project Marie Curie Chair (EXC) MEXC-CT-2003-042618 INEMAGLOW, Research Institute of Chemical and Process Engineering, Faculty of Information Technology, University of Pannonia, Egyetem u.10, Veszprem, H-8200, Hungary, , contact person: Jiří Klemeš • Centre for Process Systems Engineering, Imperial College London, C507 Roderic Hill Bldg., South Kensington Campus, London, UK, <www3.imperial.ac.uk/centreforprocess systemsengineering>, contact person: Efstratios Pistikopoulos • Centre for Technology Transfer in the Process Industries, University POLITEHNICA Bucharest, 1 Polizu St., Bldg. A, RO-011061, Bucharest, Romania, <www.chim.upb.ro/CTTPI>, contact person: Valentin Pleşu • Charles Parsons Institute, University of Limerick, Limerick, Ireland, <www2.ul.ie/web/WWW/Faculties/Science_%26_ Engineering/Research/Research_Institutes/CPI>, contact person: Toshko Zhelev • Chiyoda Corporation, Energy Frontier Business Development Office, 2-12-1 Tsurumichuo, Tsurumi-ku, Yokohama 230-8601, Japan, <www.chiyoda-corp.com/en>, contact person: Kazuo Matsuda • COWI A/S, Parallelvej 2, DK-2800, Kongens Lyngby, Denmark, tel. +45 45 97 22 11, <www.cowi.com/>, [email protected] • DEFRA, Department for Environment, Food and Rural Affairs, UK, <www.defra.gov.uk> • Department of Chemical and Environmental Engineering, University of Nottingham Malaysia, Broga Rd., 43500 Semenyih, Selangor, Malaysia, <www.nottingham.edu. my/Faculties/Engineering/Research/ENV>, contact person: Dominic Foo • Department of Chemical Engineering, University of Maribor, Smetanova ulica 17, Maribor, Slovenia, , contact person Zdravko Kravanja • Department of Chemical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Ave., Room 66–350, Cambridge, MA 02139, <web.mit.edu/cheme/index.html> • Department of Chemical Engineering, University of Pretoria, Lynwood Rd., Pretoria 0002, South Africa, <web.up.ac.za/
Information Sources and Further Reading default.asp?ipkCategoryID=2063&language=0>, contact person: Thoko Majozi • Department of Computer Science and Systems Technology, Faculty of Information Technology, University of Pannonia, Egyetem u.10, Veszprem, H-8200, Hungary, <www.dcs.vein. hu>, contact person: Ferenc Friedler • Department of Energy and Climate Change, UK, <www. decc.gov.uk> • Department of Energy and Environment, Chalmers University of Technology, SE-412 96, Göteborg, Sweden, <www.chalmers. se/ee/EN>, contact person: Thore Berntsson • Energy Information Administration, USA, <www.eia.doe. gov> • Energy Research Group, School of Science & Engineering, University of Waikato, Hamilton 3240, New Zealand, <sci. waikato.ac.nz/atee>, contact person: Martin Atkins • Energy Science and Engineering, Indian Institute of Technology Bombay, Powai, Mumbai, 400076, India, <www. me.iitb.ac.in>, contact person: Santanu Bandyopadhyay • European Integrated Pollution Prevention and Control Bureau, producing best available techniques (BAT) reference documents (aka BREFs), <www.eippcb.jrc.es/pages/ FActivities.htm> • EVECO Brno, Březinova 42, 616 00 Brno, Czech Republic, <www.evecobrno.cz/index.php/english>, contact person: Jaroslav Oral • Fraunhofer Institute for Environmental, Safety and Energy Technology (UMSICHT), Osterfelder Str., D-46047, Oberhausen, Germany, <www.umsicht.fraunhofer.de> • Global Footprint Network (“Advancing the Science of Sustainability”), 312 Clay St., Suite 300, Oakland, CA 946073510, USA, <www.footprintnetwork.org>, contact person: Susan Burns • Institute for Scientific Research, University of Guanajuato, Lascurain de Retana No 5, 36000, Guanajuato, Gto., México, <www.siia.ugto.mx/cosupera/siac.asp?ID_Des_p=1>, contact person: Martin Picon-Nuñez • Institute of Process and Environmental Engineering, Faculty of Mechanical Engineering, Brno University of Technology (UPEI VUT), Brno, Technická 2896/2, 616 69, Brno, Czech Republic, <www.fme.vutbr.cz/index.html?lang=1>, contact person: Petr Stehlík
299
300
Chapter Thirteen • International Energy Agency, 9 rue de la Fédération, 75739 Paris Cedex 15, France, <www.iea.org>, contact persons: Nobuo Tanaka (Executive Director); Richard H. Jones (Deputy Executive Director) • KBC Energy Services (formerly Linnhoff March), Targeting House, Gadbrook Park, Northwich, Cheshire CW9 7UZ, UK, <www.kbcenergyservices.com> • Laboratory for Industrial Energy Systems, Ecole Polytechnique Federale de Lausanne, Bat. ME A2, Station 9, CH-1015 Lausanne, Switzerland, , contact person: Francois Marechal • Lehrstuhl für Technische Chemie A, Technische Universität Dortmund, Fakultät Bio- und Chemieingenieurwesen, Geschossbau 1, Emil-Figgestr. 66, 44227, Dortmund, Germany, <www.chemietechnik.uni-dortmund.de/tca/> • Mechanical Engineering Faculty, Universität Paderborn, Warburger Str. 100, D-33098, Paderborn, Germany, <www. uni-paderborn.de> • Process Integration Ltd, One Central Park, Northampton Road, Manchester, M40 5WW, UK, <www.proint.org>, contact person: Robin Smith • Scottish Environment Protection Agency (SEPA), <www. sepa.org.uk> • AO Sodrugestvo-T, NTU ‘KhPI’, Frunze Str., 21, 61002, Kharkov, Ukraine, <www.sodr-t.kharkiv.com> • Thermal Energy, Department of Energy and Process Engineering, Norwegian University of Science and Technology, NO-7491, Trondheim, Norway, <www.ntnu.no>, contact person: Truls Gundersen • UK Energy Research Centre, 58 Princes Gate, Exhibition Rd., London SW7 2PG, UK, <www.ukerc.ac.uk> • University College London Energy Institute, Central House, 14 Upper Woburn Place, London, WC1H 0HY, UK, <www. ucl.ac.uk/energy> • U.S. Department of Energy, USA, <www.energy.gov> • U.S. Environmental Protection Agency, National Risk Management Research Lab/USEPA, 26 W. M.L. King Dr., Cincinnati, OH 45268, USA, <www.ep,a.gov>, contact person: Subhas K. Sikdar • Warsaw University of Technology, Płock Campus, Warsaw, Poland, <www.pw.plock.pl>, contact person: Krzysztof Urbaniec
Information Sources and Further Reading
13.1.4
Projects
There have been hundreds of projects related to these general concerns. Several typical projects that are publicly available are listed in subsequent sections dealing with specific applications (13.2.4, 13.3.4, 13.4.4, 13.5.4).
13.2
Heat Integration 13.2.1
Conferences
• Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction (PRES), organized annually since 1998, <www.conferencepres.com> • Dubrovnik Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES), organized biennially since 2002, <www.sdewes.fsb.hr> • Italian Conference on Chemical and Process Engineering (ICheaP), organized biannually since 1993, <www.aidic.it/ italiano/pubblicazioni/confse.htm>
13.2.2
Journals
• Applied Thermal Engineering, <www.elsevier.com/wps/find/ journaldescription.cws_home/630/description#description> • Energy, <www.elsevier.com/wps/find/journaldescription. cws_home/483/description#description> • Heat Transfer Engineering, <www.tandf.co.uk/journals/titles/ 01457632.asp> • Renewable Energy, <www.elsevier.com/wps/find/journal description.cws_home/969/description#description>
13.2.3
Service Providers
• American Process Inc., 817 West Peachtree St., Suite M105, Atlanta, GA 30308, USA [also offices in Athens (Greece), and Cluj-Napoca (Romania)], <www.americanprocess.com>, contact person: Theodora Retsina • Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843, USA, <www.che. tamu.edu>, contact person: Mahmoud M. El-Halwagi • Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester (formerly UMIST), Manchester, Oxford Rd., M13 9PL, UK, <www.ceas.manchester.ac.uk/research/research
301
302
Chapter Thirteen centres/centreforprocessintegration/>, contact person: Robin Smith • Centre for Process Integration and Intensification (CPI2), Research Institute of Chemical and Process Engineering, Faculty of Information Technology, University of Pannonia, Egyetem u.10, Veszpréem, H-8200, Hungary, , contact person: Jiří Klemeš • Charles Parsons Institute, University of Limerick, Limerick, Ireland, <www2.ul.ie/web/WWW/Faculties/Science_%26_ Engineering/Research/Research_Institutes/CPI>, contact person: Toshko Zhelev • Chemical Process Engineering Research Institute (CPERI), Center for Research and Technology-Hellas (CERTH), PO Box 361,Thermi Thessaloniki, Greece, <www.cperi.certh.gr/ en/cperi.htm> • Department of Chemical and Biochemical Engineering, Technical University of Denmark (DTU), Søltofts Plads, Bldg. 229, DK-2800 Kgs. Lyngby, Denmark, <www.kt.dtu.dk/ English/Forskning.aspx> • Department of Chemical and Process Engineering, Rzeszow University of Technology, al. Powstańców Warszawy 6, 35–959 Rzeszów, Poland, <portal.prz.edu.pl/en/>, contact person: Jacek Jeżowski • Department of Chemical Engineering, University of Monash, PO Box 36, Clayton, Victoria 3800, Australia, <www.eng. monash.edu.au/chemical> • Department of Energy and Environment, Chalmers University of Technology, SE-412 96, Göteborg, Sweden, <www.chalmers.se/ee/EN>, contact person: Thore Berntsson • Energy Research Group, School of Science & Engineering, University of Waikato, Hamilton 3240, New Zealand, <sci. waikato.ac.nz/atee>, contact person: Martin Atkins • Energy Science and Engineering, Indian Institute of Technology Bombay, Powai, Mumbai, 400076, India, <www. me.iitb.ac.in>, contact person: Santanu Bandyopadhyay • EVECO Brno, Březinova 42, 616 00 Brno, Czech Republic, <www.evecobrno.cz/index.php/english>, contact person: Jaroslav Oral • Faculty of Technology, Lappeenranta University of Technology, P.O.Box 20, FIN-53851 Lappeenranta, Finland, <www.lut.fi> • Institute of Process and Environmental Engineering, Faculty of Mechanical Engineering, Brno University of Technology
Information Sources and Further Reading (UPEI VUT), Brno, Technická 2896/2, 616 69, Brno, Czech Republic, Contact person: Petr Stehlík, <www.fme.vutbr.cz/ index.html?lang=1>, contact person: Petr Stehlík • KBC Energy Services (formerly Linnhoff March), Targeting House, Gadbrook Park, Northwich, Cheshire CW9 7UZ, UK, <www.kbcenergyservices.com> • Laboratory for Analysis and Synthesis of Chemical Systems, Institut de Chimie-Bâtiment B6, Sart Tilman, 4000 Liège, Belgium, <www.lassc.ulg.ac.be> • Laboratory for Industrial Energy Systems, Ecole Polytechnique Federale de Lausanne, Bat. ME A2, Station 9, CH-1015 Lausanne, Switzerland, , contact person: Francois Marechal • Process Design Center (PDC), B.V. Catharinastraat 21f, NL-4811 XD Breda, The Netherlands, <www.keuken. com/1.0-general.htm> • Process Integration Ltd, One Central Park, Northampton Road, Manchester, M40 5WW, UK, <www.proint.org>, contact person: Robin Smith • AO Sodrugestvo-T, NTU ‘KhPI’, Frunze Str., 21, 61002, Kharkov, Ukraine, <www.sodr-t.kharkiv.com> • Thermal Energy, Department of Energy and Process Engineering, Norwegian University of Science and Technology, NO-7491, Trondheim, Norway, <www.ntnu.no>, contact person: Truls Gundersen • U.S. Environmental Protection Agency, National Risk Management Research Lab/USEPA, 26 W. M.L. King Dr., Cincinnati, OH 45268, USA, <www.epa.gov>, contact person: Subhas K. Sikdar • VTT Technical Research Centre of Finland, PO Box 1000, FI02044 VTT, Finland, <www.vtt.fi> • Warsaw University of Technology, Płock Campus, Warsaw, Poland, <www.pw.plock.pl>, contact person: Krzysztof Urbaniec
13.2.4
Projects
• TOTAL SITE, European Community Non-Nuclear Energy Programme Research & Technology Development Project JOULE II PL920520, “Targeting and Planning for Reduction of Fuel, Power and CO2 on Total Sites” • EMINENT 2, “Early Market Introduction of New Energy Technologies in Liaison with Science and Industry,” <www. eminentproject.com>
303
304
Chapter Thirteen • EUBIONET 3, “Solutions for Biomass Fuel Market Barriers and Raw Material Availability,” <www.eubionet.net> • PEPNET, “Network for the Promotion of RTD Results in the Field of Eco-building Technologies, Small Polygeneration and Renewable Heating and Cooling Technologies for Buildings,” <www.proecopolynet.net> • European Community Project JOULE II PL930287, “Methodology for Optimisation of Operation Site-Scale Energy Used in Production Plants” • European Community Project JOULE II PL 930485, “Process Integration with Combined Heat and Power” • U.S. Environmental Protection Agency, Ernest Orlando Lawrence Berkeley National Laboratory, “Energy Efficiency Improvement and Cost Saving Opportunities for Petroleum Refineries, An ENERGY STAR Guide for Energy and Plant Managers” (by E. Worell & C. Galitsky), LBNL-56183 • U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, “Ultra-Low-Emission, Integrated Process Heater System,” <www1.eere.energy.gov/industry/combustion/ pdfs/processheatersystem.pdf>, accessed 22 December 2009; Bob Gremmer, High-Efficiency Technology Manager, 1000 Independent Avenue SW, Washington, DC 20585, USA • “Development of an Advanced Combined Heat and Power (CHP) System Utilizing Off-Gas from Coke Calcination, Utilization of Process Off-Gas as a Fuel for Improved Energy Efficiency,” <www1.eere.energy.gov/industry/distributed energy/pdfs/coke_calcination.pdf>, accessed 22 December 2009; project partners: Yaroslav Chudnovsky, Gas Technology Institute (GTI), Des Plaines, IL; Mark Zak, Superior Graphite Co. (SGC), Chicago, IL
13.3
Mass Integration 13.3.1
Conference
• Loss Prevention and Safety Promotion in the Process Industries, organized triennially,
13.3.2
Journals
• Journal of Clean Technologies and Environmental Policy, <www. springer.com/environment/sustainable+development/ journal/10098>
Information Sources and Further Reading • Journal of Cleaner Production, <www.elsevier.com/wps/find/ journaldescription.cws_home/30440/description#description> • Resources, Conservation & Recycling, <www.elsevier.com/wps/ find/journaldescription.cws_home/503358/description #description> • Transactions of the IChemE, UK
13.3.3
Service Providers
• American Process Inc., 817 West Peachtree St., Suite M105, Atlanta, GA 30308, USA, [also offices in Athens (Greece) and Cluj-Napoca (Romania)], <www.americanprocess.com>, contact person: Theodora Retsina • Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843, USA, <www.che. tamu.edu>, contact person: Mahmoud M. El-Halwagi • Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester (formerly UMIST), Manchester, Oxford Rd., M13 9PL, UK, <www.ceas.manchester.ac.uk/research/researchcentres/ centreforprocessintegration/>, contact person: Robin Smith • Centre for Process Integration and Intensification (CPI2), European Community Project Marie Curie Chair (EXC) MEXC-CT-2003–042618 INEMAGLOW, Research Institute of Chemical and Process Engineering, Faculty of Information Technology, University of Pannonia, Egyetem u.10, Veszprem, H-8200, Hungary, , contact Person: Jiří Klemeš • Charles Parsons Institute, University of Limerick, Limerick, Ireland, <www2.ul.ie/web/WWW/Faculties/Science_%26_ Engineering/Research/Research_Institutes/CPI>, contact person: Toshko Zhelev • Department of Chemical Engineering, University of Monash, PO Box 36, Clayton, Victoria 3800, Australia, <www.eng. monash.edu.au/chemical> • KBC Energy Services (formerly Linnhoff March), Targeting House, Gadbrook Park, Northwich, Cheshire CW9 7UZ, UK, <www.kbcenergyservices.com> • Process Integration Ltd, One Central Park, Northampton Road, Manchester, M40 5WW, UK, <www.proint.org>, contact person: Robin Smith • School of Chemical and Environmental Engineering, University of Nottingham Malaysia, Broga Rd., 43500
305
306
Chapter Thirteen Semenyih, Selangor, Malaysia, <www.nottingham.edu.my/ Faculties/Engineering/Chemical/Pages/default.aspx>, contact person: Dominic Foo • VTT Technical Research Centre of Finland, PO Box 1000, FI02044 VTT, Finland, <www.vtt.fi> • Water Footprint Network, University of Twente, UNESCOIHE, RIVM, <www.waterfootprint.org/?page=files/Contact>, contact person: A.Y. Hoekstra
13.3.4
Projects
• “Assessing Collaborative and Integrated Water Management in the Maitland River Watershed,” University of Guelph, Ontario, Canada, <sustsci.aaas.org/content.html?listed= 1&contentid=2061> • European Community Programme JOULE III PL950749, “Water Pinch: Simultaneous Energy and Water Use Minimisation” • AWARNET (Agro-food Wastes Minimisation and Reduction Network), project involving various aspects of waste minimization and energy, EC GROWTH, GRD1-CT2000-28033 (de las Fuentes, Sanders, and Klemeš, 2002), <www.cordis.lu/ data/PROJ_FP5/ACTIONeqDndSESSIONeq112242005919nd DOCeq154ndTBLeqEN_PROJ.htm>, accessed 10 December 2005 • EU INCO Project ERB IC18-CT98-0271, “Intelligent Management System for Water and Energy Minimisation in Latin American Food Industries – WATERMAN” • “Waste Minimisation for the Process Industry: New Researchers’ Fund,” University of Nottingham Malaysia, May 2006–April 2009
13.4
Combined Analysis 13.4.1
Conferences
• Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction (PRES), organized annually since 1998, <www.conferencepres.com> • Italian Conference on Chemical and Process Engineering (ICheaP), organized biannually since 1993, <www.aidic.it/ italiano/pubblicazioni/confse.htm>
Information Sources and Further Reading
13.4.2
Journals
• Applied Thermal Engineering, <www.elsevier.com/wps/find/ journaldescription.cws_home/630/description#description> • Chemical Engineering and Processing, <www.elsevier.com/ wps/find/journaldescription.cws_home/504081/description# description> • Chemical Engineering Science, <www.elsevier.com/wps/ find/journaldescription.cws_home/215/description#description> • Chemical Engineering Transactions, <www.aidic.it/cet/> • Computer Aided Chemical Engineering, <www.elsevier.com/ wps/find/bookdescription.cws_home/BS_CCE/description #description> • Ecological Modelling, <www.elsevier.com/wps/find/journal description.cws_home/503306/description#description> • Energy, <www.elsevier.com/wps/find/journaldescription. cws_home/483/description#description> • Journal of Clean Technologies and Environmental Policy, <www. springer.com/environment/sustainable+development/journal/ 10098> • Journal of Cleaner Production, <www.elsevier.com/wps/ find/journaldescription.cws_home/30440/description# description> • Renewable Energy, <www.elsevier.com/wps/find/journal description.cws_home/969/description#description> • Resources, Conservation & Recycling, <www.elsevier.com/ wps/find/journaldescription.cws_home/503358/description #description> • Waste Management, <www.elsevier.com/wps/find/journal description.cws_home/404/description#description>
13.4.3
Service Providers
• AspenTech, Aspen Technology, Inc., 200 Wheeler Rd., Burlington, MA 01803, USA, <www.aspentech.com> • Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester (formerly UMIST), Manchester, Oxford Rd., M13 9PL, UK, <www.ceas.manchester.ac.uk/research/researchcentres/ centreforprocessintegration/>, contact person: Robin Smith • Centre for Process Integration and Intensification (CPI2), European Community Project Marie Curie Chair (EXC)
307
308
Chapter Thirteen MEXC-CT-2003-042618 INEMAGLOW, Research Institute of Chemical and Process Engineering, Faculty of Information Technology, University of Pannonia, Egyetem u.10, Veszprém, H-8200, Hungary, , contact Person: Jiří Klemeš • Charles Parsons Institute, University of Limerick, Limerick, Ireland, <www2.ul.ie/web/WWW/Faculties/Science_%26_ Engineering/Research/Research_Institutes/CPI>, contact person: Toshko Zhelev • Department of Chemical Engineering, University of Massachusetts, Amherst, MA, USA, <www.ecs.umass.edu/ ~chemeng> • Hebling Technik, Hohlstr. 614, CH-8048 Zurich, Switzerland, <www.helbling.ch/htk> • Process Integration Ltd, One Central Park, Northampton Road, Manchester, M40 5WW, UK, <www.proint.org>, contact person: Robin Smith • School of Chemical and Environmental Engineering, University of Nottingham Malaysia, Broga Rd., 43500 Semenyih, Selangor, Malaysia, <www.nottingham.edu.my/ Faculties/Engineering/Chemical/Pages/default.aspx>, contact person: Dominic Foo
13.4.4 Projects • AVICENNE, “Integrated Concept for the Fermentation of Sewage Sludge and Organic Waste as a Source of Renewable Energy and for Use of the Fermented Products as a Hygienic Fertiliser and Soil Improver,” AVI*94005, Universität Stuttgart, Germany, 1997 • ANDI-POWER-CIFRU, “An Anaerobic Digestion Power Plant for Citrus Fruit Residues,” FP5 EESD NNE5/364/ 2000, Envisec S.A., Greece, 2004, <www.cordis.lu/data/ PROJ_EESD/ACTIONeqDndSESSIONeq7826200595nd DOCeq1ndTBLeqEN_PROJ.htm>, accessed 14 March 2006 • Biogas by Bioaugment, “Optimized Biogas Production and Resource Recovery through Bio-Augmentation in a Joint Plant Treating Poultry and Pig Waste,” NNE5/46/1999, Centro para a Conservaçao de Energia, Praceta 1, 2720-537 Amadora, Portugal, <www.cordis.lu/data/PROJ_EESD/ACTIONeq DndSESSIONeq7826200595ndDOCeq4ndTBLeqEN_PROJ. htm>, accessed 14 March 2006 • DEP-PROJECT, “Power Plant Based on Fluidized Bed Fired with Poultry Litter,” FP5 NNE5/75/1999, 2003, Energy Systems BV, De Vest 51, Postbus 218, 5555 XP Valkenswaard,
Information Sources and Further Reading Netherlands, <www.cordis.lu/data/PROJ_EESD/ACTIONeq DndSESSIONeq7826200595ndDOCeq13ndTBLeqEN_PROJ. htm>, accessed 14 March 2006 • “Development of a CHP Process Using Grease and Oil Waste from Plants and Animals,” FP5 EESD ENK5 CT-2000-35008, Sud recuperation Sarl, France, 2001, <www.cordis.lu/data/ PROJ_ EESD/ACT IONe qD ndSES SIONe q2611920 0595 ndDOCeq216ndTBLeqEN_PROJ.htm>, accessed 14 March 2006 • “Development of Property Integration Techniques for Waste Minimisation,” Ministry of Science, Technology and Innovation, Malaysia University of Nottingham Malaysia, September 2007–August 2009 • INEMAGLOW European Community Project MARIE CURIE CHAIR (EXC) MEXC-CT-2003-042618 “Integrated Waste to Energy Management to Prevent Global Warming”, • Project SUCLEAN, “Research on Minimization of Energy and Water Use in Sugar Production by Cooling Crystallization of Concentrated Raw Juice,” IC15960734, , accessed 10 December 2005 • WREED, “Development of an Energy Efficient to Reduce the Cost of Drying Food and Feed Waste,” FP5 EESD ENK6CT2001-30002, Ceramic Drying Systems Ltd, UK, 2004, , accessed 3 May 2010 • “Western Renewable Energy Zones,” a joint initiative of the Western Governors’ Association and U.S. Department of Energy, <www.westgov.org/index.php?option=com_content&view=ar ticle&id=219&Itemid=81>, accessed 29 December 2009 • “Flexible Distributed Energy and Water from Waste for the Food and Beverage Industry,” General Electric, <www1.eere. energy.gov/industry/distributedenergy/pdfs/water_to_ waste.pdf>, accessed 29 December 2009
13.5 Optimization for Sustainable Industry 13.5.1
Conferences
• AIChE meetings, organized twice annually in the United States, <www.aiche.org>
309
310
Chapter Thirteen • European Symposium on Computer Aided Process Engineering (ESCAPE), organized annually since 1992, <www. cape-wp.eu> • AIChE International Congress on Sustainability Science and Engineering (ICOSSE), organized since 2009,
13.5.2
Journals
• AIChE Journal, <www.aiche.org/Publications/AIChEJournal/ index.aspx> • Chemical Engineering and Processing, <www.elsevier.com/ wps/find/journaldescription.cws_home/504081/description #description> • Chemical Engineering Science, <www.elsevier.com/wps/find/ journaldescription.cws_home/215/description> • Chemical Engineering Transactions, <www.aidic.it/cet> • Computers & Chemical Engineering, <www.elsevier.com/wps/ find/journaldescription.cws_home/349/description> • Industrial & Engineering Chemistry Research, • Journal of the Chinese Institute of Chemical Engineers, <www. elsevier.com/wps/find/journaldescription.cws_home/712103/ description> • Journal of Universal Computer Science, <www.jucs.org>
13.5.3
Service Providers
• Advanced Process Combinatorics, Inc., Purdue Technology Center, 3000 Kent Ave., West Lafayette, IN 47906, USA, <www.combination.com> • AspenTech, Aspen Technology, Inc., 200 Wheeler Rd., Burlington, MA 01803, USA, <www.aspentech.com> • Centre for Process Systems Engineering, Imperial College London, C507 Roderic Hill Bldg., South Kensington Campus, London, UK, <www3.imperial.ac.uk/centreforprocesssy stemsengineering>, contact person: Efstratios Pistikopoulos • Computer-Aided Systems Laboratory, Department of Chemical Engineering, Princeton University, Princeton, NJ 08544 USA, , Contact person Christodoulos A. Floudas • Department of Computer Science and Systems Technology, Faculty of Information Technology, University of Pannonia,
Information Sources and Further Reading Egyetem u.10, Veszprém, H-8200, Hungary, <www.dcs.vein. hu>, contact person: Ferenc Friedler • Honeywell, 300-250 York St., London, Ontario, N6A 6K2 Canada, <www.honeywell.com> • Laboratory for Industrial Energy Systems, Ecole Polytechnique Federale de Lausanne, Bat. ME A2, Station 9, CH-1015 Lausanne, Switzerland, , contact person: Francois Marechal • Process Integration Ltd, One Central Park, Northampton Road, Manchester, M40 5WW, UK, <www.proint.org>, contact person: Robin Smith • School of Chemical Engineering, Purdue University, Forney Hall of Chemical Engineering, 480 Stadium Mall Dr., West Lafayette, IN 47907, <engineering.purdue.edu/ChE/index. html> • U.S. Environmental Protection Agency, National Risk Management Research Lab/USEPA,26 W. M.L. King Dr., Cincinnati, OH 45268, USA, <www.epa.gov>, contact person: Heriberto Cabezas
13.5.4
Projects
• ZEROWIN, “Towards Zero Waste in Industrial Networks,” Austrian Society for Systems Engineering and Automation, Gurkgasse 43/2, A-1140 Vienna, Austria, <www.zerowin.eu> • EMINENT 2, “Early Market Introduction of New Energy Technologies in Liaison with Science and Industry,” <www. eminentproject.com> • PRISM, “Knowledge-Based Processing Systems,” FP6 MC Human Resources Project (Prof. E. Pistikopoulos, coordinator), Imperial College of Science Technology and Medicine, London, UK, <www.cpi.umist.ac.uk/prism/>
311
This page intentionally left blank
CHAPTER
14
Conclusions and Further Information
T
he two options when writing this book were either to make it fully comprehensive or to present selected material while directing the reader to sources of further information. The first option would have resulted in a book of more than a thousand pages, so it was decided to keep the book compact and thus accessible to a wider audience. This chapter supplies additional references to provide guidelines for more detailed and comprehensive studies.
14.1
Further Reading Chapter 13 gives an overview of places to find more information. Of course, the most recent information is found in archive research papers and proceedings from related conferences. However, the number of research papers has been increasing, and many of them are dedicated to rather specialized matters. For this reason, it is often useful for a potential user of the methodology to have some material for further reading: guidebooks, handbooks, and teaching texts. The reader should bear in mind that even the most excellent books are only as current as their date of publication. No book can cover methodology developed subsequent to its printing.
14.1.1
Books and Key Articles
Process Integration The most well-known and perhaps most referenced book on Process Integration (PI) is A User Guide on Process Integration for the Efficient Use of Energy (Linnhoff et al., 1982). This “red book,” authored by a team led by Bodo Linnhoff, was updated for an extended second edition in 1994. Although it was published a long time ago, it remains an excellent vehicle for teaching and learning about Heat Integration and several other PI applications, providing the reader with a smooth and understandable learning path.
313
314
Chapter Fourteen Valuable references for engineers (as well as for teachers and students) are the books authored by Robin Smith, Linnhoff’s successor to head of the Department of Process Integration at UMIST and—after the university merged to become the University of Manchester—director of the Centre for Process Integration. Smith published two books: the first in 1995 and a new version in 2005 that included more recent developments. Ian Kemp undertook the task of developing the second (1994) edition of Linnhoff’s red book, and in 2007 he published an updated and extended version. Some of UMIST legacy course materials (CPI, 2004 and 2005) were used in that book (Kemp, 2007), which also includes spreadsheet software available as a web annex. A handy guidebook entitled A Process Integration Primer was developed by Gundersen (2000) with support from the International Energy Agency. Shenoy (1995) described the application of Heat Integration methodology to Heat Exchanger Network synthesis. Furman and Sahinidis (2002) presented a comprehensive review of 461 published works (through the year 2000) on heat exchanger synthesis, although this article is rife with misspelled non-English names (e.g., “Klemeš” was rendered as “Klemebvs”). More recently, Sieniutycz and Jeżowski (2009) authored a book dedicated to PI and energy optimization methods. In addition, a number of books (and chapters of books) have resulted from the collaboration between Centre for Process Integration at the University of Manchester and a Centre for Process Integration and Intensification (CPI2), recently created at the University of Pannonia. Key staff members of CPI2 gained much of their experience at UMIST, and the two centers have been close collaborators. Chapters have been published in handbooks related to energy and to the management of water and waste in food processing, and the material provided includes advanced methodology in addition to a number of case studies. This material can be found in Klemeš and Perry (2007a, 2007b) and in Klemeš, Smith, and Kim (2008). The water footprint’s connection to water integration was examined in a chapter from Klemeš, Varbanov, and Lam (2009), and combining PI with state-of-the-art methods for the recovery of organic materials from process water in the food industry was discussed by Napper, Kim, and Bulatov (2009) in their chapter of that same handbook. Carbon footprint and its relation to energy efficiency are analyzed in a chapter of Klemeš, Bulatov, and Perry (2008). Substantial developments in the field of mass integration have been reported by Mahmud El-Halwagi and colleagues; in particular, the synthesis of mass exchange networks (MENs) was reviewed by El-Halwagi (1997, 1999), El-Halwagi and Spriggs (1998), and Spriggs and El-Halwagi (2000). This information is also available in the book
Conclusions and Further Information form (El-Halwagi, 2006). A fast-developing group at the Malaysian campus of the University of Nottingham has compiled a new book edited by Dominic Foo and his colleagues (Foo, El-Halwagi, and Tan, 2010). An overview of capital cost targeting for MEN synthesis was provided by Fraser and Hallale (2000a, 2000b), and the application of Mathematical Programming to MEN synthesis was reviewed by Galan and Grossmann (1999) and Grossmann, Caballero, and Yeomans (1999).
Optimization and Optimal Design of Industrial Processes Several books covering these topics were published in the late 1980s and the 1990s. Although much time has passed since then, the basics of optimization are still valid. Texts that are worth mentioning include Conceptual Design of Chemical Processes (Douglas, 1988), Nonlinear and Mixed-Integer Optimization: Fundamentals and Applications (Floudas, 1995), and Systematic Methods of Chemical Process Design (Biegler, Grossman, and Westerberg, 1997). The book by Bumble (1999) also provides useful information. A series of books have been published on process system engineering, including titles by Seferlis and Georgiadis (2004), Papageorgiou and Geogiadis (2007a, 2007b), Pistikopoulos, Georgiadis, and Dua (2007a, 2007b), and Georgiadis, Kikkinides, and Pistikopoulos (2008). Along with extensive information on process costing, the wellknown Plant Design and Economics for Chemical Engineers (Peters, Timmerhaus, and West, 2003) contains a wealth of knowledge on optimization methodologies, including P-graph basics.
14.1.2 Lecture Notes and Online Teaching Resources Teaching materials and training courses covering the various aspects of PI have been developed for many years at UMIST and The University of Manchester. These materials are based on more than 25 years research and are continuously updated by the Process Integration Research Consortium. The material has been used as a basis for a number of books in the field; for example, some of it was referenced in CPI (2004 and 2005). It also has been supported by such PI software as SPRINT (2009), STAR (2009), WATER (2009), and WORK (2009). The training courses were closely developed with Linnhoff March (1998) to produce Introduction to Pinch Technology, which later developed independently. Another well-known source of teaching materials is A Process Integration Primer (Gundersen, 2000), which provides a comprehensive overview of the period up to year 2000. Teaching materials are also included in Energy and Process Integration (Georgiadis and Pistikopoulos, 2006). Many other teaching resources have either been developed in-house or been tailor-made for specific fields such as the pulp and paper industry, oil refining industry, and the sugar industry.
315
316 14.2
Chapter Fourteen
Development Trends At the 2009 PRES conference, development trends in PI, optimization, and Mathematical Programming as tools for sustainability in process industries were reviewed by Friedler (2009, 2010). Some of these trends are discussed in this section.
14.2.1
Top-Level Analysis
A typical industrial site comprises different process production units linked to a common utility system. The centralized utility system meets the demands for heat and power, creating indirect links between the processes. Considerable amounts of steam can often be saved by optimizing and retrofitting site processes and their heat exchanger networks (Dhole and Linnhoff, 1993a; Asante and Zhu, 1997). Although such projects might have the potential to save steam at different pressure levels, they may not always save costs or even fuel. Moreover, steam at a given level of pressure can have different prices depending on how much is saved, despite the common practice of attributing a single value to steam at a given level. This issue was first addressed systematically by Makwana and colleagues (Makwana, 1998; Makwana, Smith, and Zhu, 1998), who provided an approach to screening energy projects without the need to collect data from every site process. Instead, their methodology of top-level analysis required the collection and analysis only of basic data relating to the central utility system. This approach was based on comparing current and optional heat flow paths through the utility system. The procedure provided a good conceptual view of the problem, but the required number of manual steps and calculations was not amenable to automation and too complex for application to real sites. However, an important concept derived from this initial version of top-level analysis is that, given marginal steam prices, a small part of the total site can be analyzed to rank the most promising directions for steam-saving projects. Those marginal steam prices are key because they can be computed in a different way—namely, by utility system optimization. This approach led to the development by Varbanov et al. (2004) of a stepwise optimization procedure that can be used to determine the true marginal price of steam at different pressure levels as well as the constraints on steam savings imposed by the utility system. This approach builds on the optimization framework for industrial utility systems presented by Varbanov, Doyle, and Smith (2004), and it enables the analysis of complex utility systems using automated tools for utility system optimization.
14.2.2 Maintenance Scheduling, Maintainability, and Reliability A chief concern of companies and operators is that industrial plants should operate safely and reliably, with minimal costs for maintenance.
Conclusions and Further Information Increasing the efficiency of process systems via PI increases the importance of this issue because PI also increases the interdependence of the various process modules and subsystems. Cheung and Hui (2004) proposed a scheme for Total Site maintenance scheduling for better energy utilization. To minimize the impacts on production and utility systems during routine maintenance, the scheduling must be done carefully and with consideration of sitewide utilities and material balances. The reliability analysis targets maintenance requirements, cost, availability, and maintainability. The methodology described by Yin et al. (2009) considers flexible process design simultaneously with reliability and risk factors, and the best solutions are obtained by optimizing the PI. Sikos and Klemeš (2010a, 2010b) applied a combination of HEN optimization and reliability software packages to develop a methodology with several advantages over the commonly used approach.
14.2.3
Hybrid Energy Conversion Systems
Energy conversion systems for heat and power generation traditionally involve only gas and steam turbines. However, the interest of research and industrial engineers has recently been attracted by a broader range of technologies: the hybrid energy conversion systems involving fuel cells (FCs). Especially interesting are high-temperature fuel cells (HTFCs), which feature electrical efficiency of 40–60 percent (Yamamoto, 2000) compared with 30–35 percent for most gas turbines (Gas Turbine World, 2001). There has been an extensive research aimed at improving the efficiency of FC systems. Karvountzi, Price, and Duby (2004) compared the integration of molten carbonate fuel cells (MCFCs) and solid oxide fuel cells (SOFCs) into hybrid systems. Kurz (2005) focused on the choice of appropriate gas turbines for the given FCs, and Massardo and Bosio (2002) studied the MCFC combinations with gas and steam turbines. One promising option is to integrate the FCs with “bottoming cycles” for dedicated power generation or Combined Heat and Power applications. Varbanov et al. (2006) and Varbanov and Klemeš (2008) studied the benefits of integrating HTFC systems with steam cycles for purposes of industrial cogeneration. The results indicate that HTFCs have great potential in terms of economic viability and a low carbon footprint.
14.2.4 Integration of Renewables and Waste Most energy systems of industrial, residential, service and business, agriculture, and production sites continue to use fossil fuels as their primary energy source. Sites are usually equipped with steam and/ or gas turbines and with steam boilers and water heaters (running on electricity or gas) as energy conversion units. The challenge of increasing the share of renewables in the primary energy mix could be met by integrating solar, wind, biomass, and some types of waste
317
318
Chapter Fourteen with the fossil fuels. Renewable resources typically feature availability distributed over certain areas, varying significantly with time and location. Another source of variability is the energy demands (for heating, cooling, and power) of residential and business consumers, which vary significantly with the time of day and period of the year. (The notable exception is large-scale industries that operate continuous processes; in this case, variations are less pronounced and can often be either neglected or conveniently modeled using multiperiod optimization.) Furthermore, the variations in consumer energy demand are not synchronous but instead are displaced in time. For instance, commercial and office buildings tend to have higher energy demands during normal business hours, whereas residential demands tend to increase after business hours. These factors all make optimizing the design of energy conversion systems using renewable resources more complex than when using fossil fuels only. However, combining the supply and demand streams of individual users may allow such systems to serve industrial plants as well as residential customers and the service sector (hotel complexes, hospitals). The design task is to account for both the demand- and supply-side variability. One approach to solving the task is to employ advanced PI methodology with time as another problem dimension. A basic methodology along these lines has already been developed for Heat Integration of batch processes (Kemp and Deakin, 1989; Klemeš et al., 1994) and was recently revisited by Foo, Chew, and Lee (2008). A further important step in the direction of extending this methodology to Heat Integration of renewables was taken by Perry, Klemeš, and Bulatov (2008), who considered the integration of local energy sectors into extended total sites involving residential and commercial processes and buildings in addition to industrial ones. This work was further developed by Varbanov and Klemeš (2010) to account for the integration of inherently variable renewables. Converting waste via thermal processing is an intriguing option because it simultaneously reduces the demand for fossil fuels and landfilling. Bébar et al. (2001, 2002) demonstrated a method for efficiently utilizing the heat value of incineration products that would partially compensate for the cost of thermal waste treatment. The challenge of utilizing waste as an energy source is not variability, as it is with renewables, since waste is relatively plentiful and generated at significant rates. Rather, the main problems associated with extracting energy from waste are technological. Municipal and other solid wastes are typically incinerated (Bébar et al., 2001; Bébar et al., 2002; Stehlik, 2009), a process that involves three principal issues. First, the incineration must ensure efficient combustion and minimal emissions of pollutants; this is often achieved by co-firing
Conclusions and Further Information the waste with other fuels (Werther, 2007). Second, it is important to prevent the release of toxic pollutants with the flue gas—for example, the emission of dioxins from incinerated solid wastes. These emissions can be minimized by proper waste separation and moisture reduction prior to the incineration (Werther, 2007). Third, the moisture content poses other challenges besides the formation of dioxins. The higher the moisture content, the larger fraction of energy is wasted on evaporating it; thus, increased moisture reduces the specific energy gain from incineration. This means that, for purposes of energy generation, waste with especially high water content is more beneficially treated via other processes. Examples include the anaerobic digestion of wet organic waste from food industry and agriculture (Zhang et al., 2007; Macias-Corral et al., 2008) and the supercritical gasification of black liquor from pulpand-paper plants (Pettersson and Harvey, 2009; Sricharoenchaikul, 2009).
14.2.5 Better Utilization of Low-Grade Heat Desai and Bandyopadhyay (2010) proposed a scheme incorporating organic Rankine cycles (ORCs) in the efficient utilization of lowgrade waste heat for power generation. They provided a graphically assisted procedure for integrating an industrial process with an ORC, thereby reducing the process cooling demand and the need to import external power.
14.2.6
Energy Planning That Accounts for Carbon Footprint
The international community has become increasingly concerned with climate change. Most attention has been focused on the carbon footprint (CFP), which quantifies the impact of greenhouse gas emissions. In parallel with this development, the issue of energy independence has been given more attention in a number of countries that import fossil fuels—especially the United States, European countries (except Norway), and, to a lesser extent, Australia and Japan. An interesting paper by Foo, Tan, and Ng (2008) addresses “Carbon and Footprint-Constrained Energy Planning Using Cascade Analysis Technique.” The authors presented algebraic targeting techniques for energy-sector planning under constraints on CO2 emissions and land availability. This contribution extends the classic Pinch Analysis of Linnhoff and Hindmarsh (1983) to identify the minimum amount of low- or zero-carbon energy sources needed to meet regional or national energy demands while observing limits on CO2 emissions. The concept of Regional Energy Clustering presented by Lam, Varbanov, and Klemeš (2010) involves the synthesis of regional energy targeting and supply chains. The methodology seeks to account for
319
320
Chapter Fourteen the complete spectrum of renewable energy sources and energy carriers—and to evaluate the appropriate combinations of waste-toenergy and fossil fuels—in order to deliver more complete and practical procedures.
14.3
Conclusions In recent years there has been much interest in the development of renewable, non-carbon-based energy sources to combat the threat of increased CO2 emissions and related climatic change. Increases in the price of oil and gas have further boosted interest in such alternative energy sources. These concerns have resulted in increased efficiency of energy and water use in the industrial sector, although the major industry’s use of renewable energy sources has been only sporadic. In contrast, domestic energy supply has moved more positively toward the integration of renewable energy sources, including wind turbines, solar heating, and heat pumps. Yet efforts to design a sustainable combined energy system that includes both industrial and residential buildings have been limited and ad hoc, since there are no systematic design techniques available for producing a symbiotic system on this scale. Increasing the efficiency of energy-using processes is the most effective current method for reducing costs and emissions that affect the stability of the world’s climate, and contributing to sustainable growth. Increasing energy efficiency is the cleanest short-term method to produce green energy—that is, energy that produces the least amount of emissions. The situation is similar with respect to minimizing the consumption of water and the generation of wastewater. Unfortunately, these relatively mundane approaches are often overlooked by those who are seduced by exciting new technologies—such as renewable energy sources—that have generated more publicity. Nonetheless, the primary sustainability issue remains the efficient use of resources: raw materials, energy, and water. Two metrics have recently been developed to assess the impact of energy-efficient methods and energy-reduction proposals. The carbon footprint (CFP) takes account of all carbon emissions over the entire life cycle of a process or product, and the water footprint (WFP) indicates how much water is used during the life cycle of a product or service.
Bibliography Aalborg University, 2009a. EnergyInteractive.NET. <energyinteractive.net> (accessed 11 June 2009). Aalborg University, 2009b. EnergyPLAN: Advanced Energy System Analysis Computer Model. <energy.plan.aau.dk> (accessed 10 January 2009). Aaltola, J., 2002. Simultaneous synthesis of flexible heat exchanger network. Applied Thermal Engineering, 22, 907–918. ABARE, 2009. ABARE Models. <www.abare.gov.au/publications_html/models/ models/models.html> (accessed 22 April 2009). Acevedo, J., Pistikopoulos, E. N., 1997. A multiparametric programming approach for linear process engineering problems under uncertainty. Industrial & Engineering Chemistry Research, 36, 717–728. Adonyi, R., Biros, G., Holczinger, T., Friedler, F., 2008. Effective scheduling of a large-scale paint production system. Journal of Cleaner Production, 16(2), 225–232. Adonyi, R., Romero, J., Puigjaner, L., Friedler, F., 2003. Incorporating heat integration in batch process scheduling. Applied Thermal Engineering, 23, 1743–1762. AEA Technology, 2000. Study on Energy Management and Optimisation in Industry, 2000 (prepared by AEA Technology plc at the request of the Environment Directorate-General of the European Commission). <www. ec.europa.eu/environment/ippc/pdf/energy.pdf> (accessed 23 August 2007). Ahmad, M. I., Chen, L., Jobson, M., Zhang, N., 2008. Synthesis and optimisation of heat exchanger networks for multi-period operation by simulated annealing. 11th Conference on Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction—PRES2008/CHISA2008, Summaries 4. Prague: Czech Society of Chemical Engineering lecture J8.2, Proceedings, pp.1200 Ahmad, S., Linnhoff, B., Smith, R., 1989. Supertargeting: Different process structures for different economics. Journal of Energy Resources Technology, 111(3), 131–136. Ahmad, S., Linnhoff, B., Smith, R., 1990. Cost optimum heat exchanger networks: 2. Targets and design for detailed capital cost models. Computers & Chemical Engineering, 14(7), 751–767. Ahmad, S., Smith, R., 1989. Targets and design for minimum number of shells in heat exchanger networks. Chemical Engineering Research and Design, 67(5), 481–494. Albrecht, J., 2007. The future role of photovoltaics, a learning curve versus portfolio perspective. Energy Policy, 35(4), 2296–2304. Almató, M., Espuña, A., Puigjaner, L., 1999. Optimisation of water use in batch process industries. Computers & Chemical Engineering, 23, 1427–1437. Al-Riyami, B. A., Klemeš, J., Perry, S., 2001. Heat integration retrofit analysis of a heat exchanger network of a fluid catalytic cracking plant. Applied Thermal Engineering, 21, 1449–1487.
321
322
Bibliography Alva-Argáez, A., Kokossis, A. C., Smith, R., 1998. Wastewater minimisation of industrial systems using an integrated approach. Computers & Chemical Engineering, 22, 741–744. Alves, J., 1999. Analysis and design of refinery hydrogen distribution systems. PhD thesis, UMIST, Manchester, U.K. Alves, J. J., Towler, G. P., 2002. Analysis of refinery hydrogen distribution systems. Industry and Engineering Chemistry Research, 41, 5759–5769. Arellano-Garcia, H., Martini, W., Wendt, M., Li, P., Wozny, G., 2003. Chance constrained batch distillation process optimization under uncertainty. Foundations of Computer Aided Process Operations—FOCAPO (Coral Springs, FL, 12–15 January). Arellano-Garcia, H., Wozny, G., 2009. Chance constrained optimization of process systems under uncertainty: I. Strict monotonicity. Computers & Chemical Engineering, 33, 1568–1583. Argonne National Laboratory, 2009a. Electricity Market Complex Adaptive System (EMCAS). <www.dis.anl.gov/projects/emcas.html> (accessed 23 April 2009). Argonne National Laboratory, 2009b. Energy and Power Evaluation Program (ENPEP-BALANCE). <www.dis.anl.gov/projects/Enpepwin.html> (accessed 23 April 2009). Argonne National Laboratory, 2009c. Generation and Transmission Maximization (GTMax) Model. <www.dis.anl.gov/projects/Gtmax.html> (accessed 23 April 2009). Asante, N. D. K., 1996. Automated and interactive retrofit design of practical heat exchanger networks. PhD thesis, UMIST, Manchester, U.K. Asante, N. D. K., Zhu, X. X., 1997. An automated and interactive approach for heat exchanger network retrofit. Chemical Engineering Research and Design, 75(part A), 349–360. Aspelund, A., Berstad, D. A., Gundersen, T., 2007. An extended pinch analysis and design procedure utilizing pressure based exergy for subambient cooling. Applied Thermal Engineering, 27, 2633–2649. Aspelund, A., Gundersen, T., 2007. A new process synthesis methodology utilizing pressure exergy in subambient processes. Computer Aided Chemical Engineering, 24, 1133–1138. AspenTech, 2009a. Aspen Economic Evaluation. <www.aspentech.com/products/ aspen-economic-eval.cfm> (accessed 26 October 2009). AspenTech, 2009b. Aspen Exchanger Design & Rating (EDR). <www.aspentech. com/products/spen-edr.cfm> (accessed 26 October 2009). AspenTech, 2009c. Aspen Plus. <www.aspentech.com/core/aspen-plus.cfm> (accessed 26 October 2009). AspenTech, 2009d. What Is Aspen ONE. <www.aspentech.com/v7/whatis.cfm> (accessed 26 October 2009). Avriel, M., Wilde, D. J., 1966. Optimality proof for the symmetric Fibonacci search technique. Fibonacci Quarterly, 4, 265–269. Bagajewicz, M., 2000. A review of recent design procedures for water networks in refineries and process plants. Computers & Chemical Engineering, 24, 2093–2113. Bagajewicz, M., Savelski, M., 2001. On the use of linear models for the design of water utilization systems in process plants with a single contaminant. Chemical Engineering Research and Design, 79(5), 600–610. Balmorel [Danish Energy Research Program], 2009. <www.balmorel.com/> (accessed 30 December 2009). Bandyopadhyay, S., Ghanekar, M. D., Pillai, H. K., 2006. Process water management. Industrial & Engineering Chemistry Research, 45, 5287–5297. Banerjee, I., Ierapetritou, M. G., 2003. Parametric process synthesis for general non-linear models. Computers & Chemical Engineering, 27, 1499–1512. Barlow, R. E, Proschan, F., 1996. Mathematical theory of reliability. Society for Industrial and Applied Mathematics, New York, USA, ISBN-13: 978-0898713695.
Bibliography BAT-CENTRE, 2009. Vito’s Centre for Best Available Techniques. <www.emis.vito. be/vito%E2%80%99s-centre-best-available-techniques-bat-centre> (accessed 30 April 2010). Bébar, L., Čaněk, J., Kermes, V., Stehlík, P., Oral, J., 2001. Low NOx burners—Recent development, equipment, experience, modelling. International Conference on Incineration and Thermal Treatment Technologies (Philadelphia, 14–18 May). College Park: University of Maryland Press. Bébar, L., Martinák, P., Hájek, J., Stehlík, P., Hajný, Z., Oral, J., 2002. Waste to energy in the field of thermal processing of waste. Applied Thermal Engineering, 22, 897–906. Bertok, B., Friedler, F., Feng, G., Fan, L. T., 2001. Systematic generation of the optimal and alternative flowsheets for azeotropic distillation systems. Computer Aided Chemical Engineering, 9, S351–S356. Biegler, L. T., Grossmann, I. E., Westerberg, A. W., 1997. Systematic methods of chemical process design. Englewood Cliffs, NJ: Prentice-Hall. Björk, K., M., Westerlund, T., 2002. Global optimization of heat exchanger network synthesis problems with and without the isothermal mixing assumption. Computers & Chemical Engineering, 26(11), 1581–1593. Blomquist, P. A., Brown N. J., 2004. A review of the pre-assessment and assessment techniques used in waste minimisation audits. Water SA, 30(2), 131–141. Bochenek, R., Jezowski, J., Jezowska, A., 1998. Retrofitting flexible heat exchanger networks—Optimization vs. sensitivity tables method. CHISA '98/ 1st Conference PRES '98 (Prague), lecture F1. Prague: Czech Society of Chemical Engineering. Boiler Cost (MATCHES), 2003. <www.matche.com/EquipCost/Boiler.htm> (accessed 15 September 2009). BREF, 2006. Reference Document on Best Available Techniques in Food, Drink and Milk Industries (Seville, Spain). (accessed 26 December 2009). Brooke, A., Kendrick, D., Meeraus, A. A., 1992. GAMS—A user’s guide, release 2.25. San Francisco: Scientific Press. Brown, M., Ulgiati, S., 2004. Energy quality, emergy, and transformity: H.T. Odum’s contributions to quantifying and understanding systems. Ecological Modelling, 178(1–2), 201–213. Bumble, S., 1999. Computer generated physical properties. Boca Raton, FL: CRC Press. CADDET [Center for the Analysis and Dissemination of Demonstrated Energy Technologies], 1994. Integrated Heat Recovery in a Malt Whisky Distillery. Project no. UK-94-509. CAPE, 2009. Working Party of Computer Aided Process Engineering, European Federation of Chemical Engineering. <www.cape-wp.eu> (accessed 10 January 2009). Carbon Trust, 2009. <www.carbontrust.com> (accessed 26 December 2009). Casani, S., Rouhany, M., Knøchel, S., 2005. A discussion paper on challenges and limitations to water reuse and hygiene in the food industry. Water Research, 39(6), 1134–1146. Castell, C. M. L., Lakshmanan, R., Skilling, J. M., Bañares-Alcántara, R., 1998. Optimisation of process plant layout using genetic algorithms. Computers & Chemical Engineering, 22(Suppl.), S993–S996. Castro, P., Barbosa-Póvoa, A. P. F. D., Matos, H., 2001. An improved RTN continuoustime formulation for the short-term scheduling of multipurpose batch plants. Industrial & Engineering Chemistry Research, 40, 2059–2068. Cerdá, J., Galli, M. R., 1990. Synthesis of flexible heat exchanger networks—II. Nonconvex networks with large temperature variations. Computers & Chemical Engineering, 14(2), 213–225. Cerdá, J., Galli, M. R., 1991. Synthesis of flexible heat exchanger networks—III. Temperature and flowrate variations. Computers & Chemical Engineering, 15(1), 7–24. Cerdá, J., Galli, M. R., Camussi, N., Isla, M. A., 1990. Synthesis of flexible heat exchanger networks—I. Convex networks. Computers & Chemical Engineering, 14(2), 197–211.
323
324
Bibliography Cerdá, J., Henning, G. P., Grossmann, I. E., 1997. A mixed-integer linear programming model for short-term scheduling of single-stage multiproduct batch plants with parallel lines. Industrial & Engineering Chemistry Research, 36, 1695–1707. Chapman, S. J., 2009. Essentials of MATLAB programming, 2nd ed. Stamford, CT: Cengage. Chempute Software, 2001. CHEMCAD: Process Flowsheet Simulation. <www. chempute.com/chemcad.htm> (accessed 26 October 2009). Chemstations, 2009. CHEMCAD. <www.chemstations.com> (accessed 26 October 2009). Chen, C. L., Chang, C. Y., Lee, J. Y., 2008. Continuous-time formulation for the synthesis of water-using networks in batch plants. Industrial & Engineering Chemistry Research, 47(20), 7818–7832. Chen, C. L., Lee, J. Y., Tan, J. W., Ciou, Y. J., 2009. Synthesis of water-using network with central reusable storage in batch processes. Computers & Chemical Engineering, 33(1), 267–276. Cheng, Z., Jia, X., Gao, P., Wu, S., Wang , J., 2008. A framework for intelligent reliability centered maintenance analysis. Reliability Engineering and System Safety, 93, 784–792. Cheung, K.-Y., Hui, C.-W., 2004. Total-site scheduling for better energy utilization. Journal of Cleaner Production, 12, 171–184. Chopra, S., Meindl, P., 2009. Supply chain management: Strategy, planning and opeation. Englewood Cliffs, NJ: Prentice-Hall. Ciric, A. R., Floudas, C. A., 1989. A retrofit approach for heat exchanger networks. Computers & Chemical Engineering, 13(6), 703–715. Ciric, A. R., Floudas C. A., 1990. A comprehensive optimisation model of the HEN retrofit problem. Heat Recovery Systems and CHP, 10(4), 407–422. Colberg, R. D., Morari, M., 1990. Area and capital cost targets for heat exchanger network synthesis with constrained matches and unequal heat transfer coefficients. Computers & Chemical Engineering, 14(1), 1–22. Connolly, D., Lund, H., Mathiesen, B. V., Leahy, M., 2010. A review of computer tools for analysing the integration of renewable energy into various energy systems. Applied Energy, 87(4), 1059–1082. CPI [Centre for Process Integration], 2004 and 2005. Heat integration and energy systems (MSc course, School of Engineering and Analytical Science, UMIST and The University of Manchester). Manchester, U.K. Cunha, M. C., Ribeiro, L., 2004. Tabu search algorithms for water network optimization. European Journal of Operational Research, 157(3), 746–758. Daichendt, M. M., Grossmann, I. E., 1997. Integration of hierarchical decomposition and mathematical programming for the synthesis of process flowsheets. Computers & Chemical Engineering, 22(1–2), 147–175. Danish Energy Agency, 2009. Ramses. <www.ens.dk/da-DK/Info/TalOgKort/ Fremskrivninger/modeller/ramses/Sider/Forside.aspx> (accessed 1 July 2009). Dantzig, G. B., 1951. Maximization of a linear function of variables subject to linear inequalities. In T. C. Koopmans (Ed.), Activity analysis of production and allocation, pp. 339–347. New York: Wiley. Dantzig G. B., 1968. Linear programming and extensions. Princeton, NJ: Princeton University Press. Dantzig, G., Orchard-Hays, W., 1953. Notes on linear programming, V: Alternative algorithm for the revised simplex method using product form of the inverse. Rand Research Memorandum RM-1268, RAND Corporation, Santa Monica, CA. Dantzig, G. B., Orden, A., Wolfe, P., 1954. The generalized simplex method for minimizing a linear form under linear inequality restraints. Rand Research Memorandum RM-1264, RAND Corporation, Santa Monica, CA. De Benedetto, L., Klemeš, J., 2009. The environmental performance strategy map: An integrated LCA approach to support the strategic decision making process. Journal of Cleaner Production, 17(10), 900–906.
Bibliography de Castro, H. F., Cavalca, K. L., 2006. Maintenance resources optimization applied to a manufacturing system. Reliability Engineering and System Safety, 91, 413–420. de las Fuentes, L., Sanders, B., Klemeš, J., 2002. Awarenet: Agro-food wastes minimisation and reduction. PRES / CHISA 2002 Conference on Process Integration, Modelling, and Optimisation for Energy Saving and Pollution Reduction (Prague), lecture H4.4 [13282002]. Czech Society of Chemical Engineering. Del Nogal, F., Kim, J.- K., Perry, S., Smith, R., 2008. Synthesis of cryogenic energy systems. Computer Aided Chemical Engineering, 25, 67–72. Desai, N. B., Bandyopadhyay, S., 2010. Process integration of organic Rankine cycle. Energy, 34(10), 1674–1686. Dhole, V. R., Linnhoff, B., 1993a. Distillation column targets. Computers & Chemical Engineering, 17(5–6), 549–560. Dhole, V. R., Linnhoff, B., 1993b. Total site targets for fuel, co-generation, emissions and cooling. Computers & Chemical Engineering, 17(Suppl.), 101–109. Dhole, V. R., Ramchandi, N., Tainsh, R. A., Wasilewski, M., 1996. Make your process water pay for itself. Chemical Engineering, 103(1), 100–103. Donnelly, N., Klemeš, J., Perry, S., 2005. Impact of economic criteria and cost uncertainty on HEN network design and retrofit. In J. Klemeš (Ed.), Proceedings of PRES '05, pp. 127–132. Milano, Italy: AIDIC. Douglas, A. P., Hoadley, A. P. A., 2006. A process integration approach to the design of the two- and three-column methanol distillation schemes. Applied Thermal Engineering, 26(4), 338–349. Douglas, J. M., 1988. Conceptual design of chemical processes. New York: McGrawHill. Doyle, S. J., Smith. R., 1997. Targeting water reuse with multiple contaminants. Trans. IChemE, 75(part B), 181–189. DTI, 2006. UK Energy in Brief—July 2006. <www.dti.gov.uk/energy/statistics/ publications/in-brief/page17222.html> (accessed 15 March 2007). Dua, V., Bozinis, N. A., Pistikopoulos, E. N., 2002. A multiparametric programming approach for mixed-integer quadratic engineering problems. Computers & Chemical Engineering, 26(4–5), 715–733. Dupont, D., Renzetti, S., 1998. Water use in the Canadian food processing industry. Canadian Journal of Agricultural Economics, 46(1), 83–92. Duran, M. A., Grossmann, I. E., 1986. An outer-approximation algorithm for a class of mixed-integer nonlinear programs. Mathematical Programming, 36(1), 307–339. DWSIM, 2010. About DWSIM. (accessed 18 January 2010). Ea Energy Analyses, 2010. The STREAM Modelling Tool. <www.streammodel. org> (accessed 5 May 2010). Ebenezer, S. A., 2005. Removal of Carbon Dioxide from Natural Gas for LNG Production (semester project work report, December 2005, Norwegian University of Science and Technology, Norway). <www.ipt.ntnu.no/~jsg/ studenter/prosjekt/Salako2005.pdf> (accessed 29 December 2009). Edgar, T. F., Himmelblau, D. M., 1988. Optimization of chemical processes. New York: McGraw-Hill. EEBPp, 2002. Benchmarking tool for industrial buildings—Heating and internal lighting. London: U.K. Energy Efficiency Best Practice Programme. EEO [Energy Efficiency Office], 1993. Opportunities for compact heat exchangers. Best Practice Programme R&D Profile 36, London. Ehrgott, M., 2005. Multicriteria optimization, 2nd ed. Berlin: Springer. El-Halwagi, M. M., 1997. Pollution prevention through process integration: Systematic design tools. San Diego, CA: Academic Press. El-Halwagi, M. M., 1999. Pollution prevention through mass integration: Systematic design tools. In Proceedings of International Conference on Process Integration, vol. 1, pp. 95–112. Copenhagen: PI99. El-Halwagi, M. M., 2006. Process integration. Amsterdam: Academic Press.
325
326
Bibliography El-Halwagi, M. M., Gabriel, F., Harell, D., 2003 Rigorous graphical targeting for resource conservation via material recycle/reuse networks. Industrial & Engineering Chemistry Research, 42, 4319–4328. El-Halwagi, M. M., Glasgow, I. M., Eden, M. R., Qin, X., 2004. Property integration: Componentless design techniques and visualization tools. AIChE Journal, 50(8), 1854–1869. El-Halwagi, M., Manousiouthakis, V., 1989. Synthesis of mass exchange networks. AIChE Journal, 35(8), 1233–1244. El-Halwagi, M. M., Spriggs, H. D., 1998. Solve design puzzles with mass integration. Chemical Engineering Progress, 94, 25–44. EMD International A/S, 2009. <www.emd.dk/> (accessed 23 April 2009). EMINENT2, 2009. Welcome to EMINENT. <www.eminentproject.com> (accessed 23 April 2009). Energinet.dk, 2009. SIVAEL. <www.energinet.dk/en/menu/Planning/ Analysis+models/Sivael/SIVAEL.htm> (accessed 27 April 2009). Energy Information Administration, 2009. The National Energy Modeling System: An Overview; 2003. <www.eia.doe.gov/oiaf/aeo/overview/index.html> (accessed 26 April 2009). ENVIROWISE, 2009. <envirowise.gov.uk> (accessed on 4 November 2009). EUROPA, 2009. EUROPA—Gateway to the European Union. <europa.eu> (accessed 4 November 2009). European Commission, 2009. Council Directive 96/61/EC of 24 September 1996 (concerning integrated pollution prevention and control). Official Journal L 257, 10 October 1996, pp. 0026–0040. European Commission, DG Environment. 2000. Water Framework Directive (2000/60/EC). <ec.europa.eu/environment/water/water-framework/index_ en.html> (accessed 10 December 2009). Eurostat, 2007. Eurostat Press Office. <ec.europa.eu/eurostat> (accessed 15 March 2007). Faber, R., Jockenhövel, T., Tsatsaronis, G., 2005. Dynamic optimization with simulated annealing. Computers & Chemical Engineering, 29(2), 273–290. Fahnich, A., Mavrov, V., Chmiel, H., 1998. Membrane processes for water reuse in the food industry. Desalination, 119(1–3), 213–216. Fan, L. T., Zhang, T., Liu, J., Seib, P. A., Friedler, F., Bertok, B., 2008. Price-targeting through repeated flowsheet re-synthesis in developing novel processing equipment: Pervaporation. Industrial & Engineering Chemistry Research, 47, 1556–1561. Feng, G., Fan, L. T., 1996. On stream splitting in separation system sequencing. Industrial & Engineering Chemistry Research, 35(6), 1951–1958. Ferrer Nadal, S., Holczinger, T., Méndez, C. A., Friedler, F., Puigjaner, L., 2006. Rigorous scheduling resolution of complex multipurpose batch plants: S-graph vs. MILP. 16th European Symposium on Computer Aided Process Engineering—ESCAPE 16 (Garmisch-Partenkirchen, Germany, 9–13 July). Fiaschi, D., Carta, R., 2007. CO2 abatement by co-firing of natural gas and biomassderived gas in a gas turbine. Energy, 32(4), 549–567. Fieg, G., Xing, L., Jeżowski, J., 2009. A monogenetic algorithm for optimal design of large-scale heat exchanger networks. Chemical Engineering and Processing, 48(11–12), 1506–1516. Floudas, C. A., 1995. Nonlinear and mixed-integer optimization: Fundamentals and applications. Oxford, U.K.: Oxford University Press. Floudas, C. A., Lin, X., 2004. Continuous-time versus discrete-time approaches for scheduling of chemical processes: A review. Computers & Chemical Engineering, 28(11), 2109–2129. Floudas, C. A., Grossmann, I. E., 1987a. Automatic synthesis of multiperiod heat exchanger network configurations. Computers & Chemical Engineering, 11(2), 123–143. Floudas, C. A., Grossmann, I. E., 1987b. Synthesis of flexible heat exchanger networks with uncertain flowrates and temperatures. Computers & Chemical Engineering, 11(4), 319–336.
Bibliography Foo, D. C. Y., Chew, Y. H., Lee, C. T., 2008. Minimum units targeting and network evolution for batch heat exchanger network. Applied Thermal Engineering, 28(16), 2089–2099. Foo, D. C. Y., El-Halwagi, M. M., Tan, R. R. (Eds.), 2010. Recent advances in sustainable process design and optimisation. Singapore: World Scientific. Foo, D., Kazantzi, V., El-Halwagi, M., Mananc, Z. A., 2006. Surplus diagram and cascade analysis technique for targeting property-based material reuse network. Chemical Engineering Science, 61, 2626–2642. Foo, D. C. Y, Manan, Z. A., Tan, Y. L., 2005. Synthesis of maximum water recovery network for batch process systems. Journal of Cleaner Production, 13(15), 1381– 1394. Foo, D. C. Y., Tan, R. R., 2007. Targeting for minimum low and zero-carbon energy resources in carbon-constrained energy sector planning using cascade analysis. Chemical Engineering Transactions, 12, 139–144. Foo, D., Tan, R., Ng, D. K. S., 2008. Carbon and footprint-constrained energy planning using cascade analysis technique. Energy, 33(10), 1480–1488. Forschungszentrum Jülich Institute for Energy Research, 2009. Systems Analysis and Technology Evaluation. <www.fz-juelich.de/ief/ief-ste/index. php?index=3> (accessed 15 June 2009). Fraser, D. M., Hallale, N., 2000a. Capital and total cost targets for mass exchange networks: Part 1. Simple capital cost models. Computers & Chemical Engineering, 23(11–12), 1661–1679. Fraser, D. M., Hallale, N., 2000b. Capital and total cost targets for mass exchange networks: Part 2. Detailed capital cost models. Computers & Chemical Engineering, 23(11–12), 1681–1699. Frausto-Hernández, S., Rico-Ramírez, V., Jiménez-Gutiérrez, A., HernándezCastro, S., 2003. MINLP synthesis of heat exchanger networks considering pressure drop effects. Computers & Chemical Engineering, 27(8–9), 1143–1152. Friedler, F., 2009. Process integration, modelling and optimisation for energy saving and pollution reduction. Chemical Engineering Transactions, 18, 1–26. Friedler, F., 2010. Process integration, modelling and optimisation for energy saving and pollution reduction. Applied Thermal Engineering, doi:10.1016/j. applthermaleng.2010.04.030. Friedler, F., Fan, L. T., 2009. Mathematical modeling in process design and operation: Structural assessment. Seventh International Conference on the Foundations of Computer-Aided Process Design: Design for Energy and the Environment— FOCAPD 2009 (Breckenridge, CO, 7–12 June). Friedler, F., Fan, L. T., Imreh, B., 1998. Process network synthesis: Problem definition. Networks, 28, 119–124. Friedler, F., Tarjan, K., Huang, Y. W., Fan L. T., 1992a. Combinatorial algorithms for process synthesis. Computers & Chemical Engineering, 16, S313–S320. Friedler, F., Tarjan, K., Huang Y. W., Fan, L. T., 1992b. Graph-theoretic approach to process synthesis: Axioms and theorems. Chemical Engineering Science, 47, 1973–1988. Friedler, F., Tarjan, K., Huang, Y. W., Fan, L. T., 1993. Graph-theoretic approach to process synthesis: Polynomial algorithm for maximal structure generation. Computers & Chemical Engineering, 17(9), 929–942. Friedler, F., Varga, J. B., Fan, L. T., 1995. Decision-mapping: A tool for consistent and complete decisions in process synthesis. Chemical Engineering Science, 50(11), 1755–1768. Friedler, F., Varga, J. B., Fehér, E., Fan, L. T., 1996. Combinatorially accelerated branch-and-bound method for solving the MIP model of process network synthesis. In C. A. Floudas and P. M. Pardalos (Eds.), State of the art in global optimization, pp. 609–626. Boston: Kluwers. Fritzson, A., Berntsson, T., 2006. Efficient energy use in a slaughter and meat processing plant—Opportunities for process integration. Journal of Food Engineering, 76(4), 594–604. Furman, K. C., Sahinidis, N. V., 2002. A critical review and annotated bibliography for heat exchanger network synthesis in the 20th century. Industrial & Engineering Chemistry Research, 41, 2335–2370.
327
328
Bibliography Galan, B., Grossmann, I. E., 1999. Optimization strategies for the design and synthesis of distributed wastewater treatment networks. Computers & Chemical Engineering, 23, S161–S164. GAMS, 2009. GAMS Home Page. < www.gams.com> (accessed 26 October 2009). Gas Turbine World, 2001. Handbook 2000–2001, vol. 21. Southport, CT: Pequot. Georgiadis, M. C., Kikkinides, E. S., Pistikopoulos, E. N., 2008. Process systems engineering, vol. V: Energy systems engineering. New York: Wiley. Georgiadis, M. C., Pistikopoulos, E. N., 2006. Energy and process integration. Redding, CT: Begell House. Giammatei, J. A. G., 1994. Simultaneous optimization of a chemical process, its heat exchanger network, and the utility system using a process simulator. PhD thesis, University of Pennsylvania, Philadelphia. (accessed 28 December 2009). Gouws, J. F., Majozi, T., 2008a.A mathematical technique for the design of nearzero-effluent batch processes. Water SA, 34, 291–295. Gouws, J. F., Majozi, T., 2008b. Impact of multiple storage in wastewater minimization for multicontaminant batch plants: Toward zero effluent. Industrial & Engineering Chemistry Research, 47(2), 369–379. Grossmann, I. E., 1985. Mixed-integer programming approach for the synthesis of integrated process flowsheets. Computers & Chemical Engineering, 9(5), 463–482. Grossmann, I. E., 1990. MINLP optimization strategies and algorithms for process synthesis. FOCAPD, CACHE, pp. 105–132. Amsterdam: Elsevier. Grossmann, I. E., Caballero, J. A., Yeomans, H., 1999. Mathematical programming approaches for the synthesis of chemical process systems. Korean Journal of Chemical Engineering, 16, 407–426. Guinand, E. A., 2001. Optimization and network sensitivity analysis for process retrofitting. PhD thesis, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge. Gunaratnam, M., Alva-Argáez, A., Kokossis A., Kim, J., Smith, R., 2005. Automated design of total water systems. Industrial & Engineering Chemistry Research, 44(3), 588–599. Gundersen, T., 2000. A process integration primer—Implementing agreement on process integration. Trondheim, Norway: International Energy Agency, SINTEF Energy Research. H 2RES, 2009. Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb. <www.powerlab.fsb.hr/h2RES/> (accessed 23 April 2009). Halasz, L., Nagy, A. B., Ivicz, T., Friedler, F., Fan, L. T., 2002. Optimal retrofit design and operation of the steam-supply system of a chemical complex. Applied Thermal Engineering, 22, 939–947. Halasz, L., Povoden, G., Narodoslawsky, M., 2005. Sustainable processes synthesis for renewable resources. Resources, Conservation and Recycling, 44, 293–307. Hall, S. G., Ahmad, S., Smith, R., 1990. Capital cost targets for heat exchanger networks comprising mixed materials of construction, pressure ratings and exchanger types. Computers & Chemical Engineering, 14(3), 319–335. Hallale, N., 2002. A new graphical targeting method for water minimization. Advances in Environmental Research, 6(3), 377–390. Hallale, N., Fraser, D. M., 1998. Capital cost targets for mass exchange networks, a special case: Water minimization. Chemical Engineering Science, 53(2), 293–313. Hallale, N., Liu, F., 2001. Refinery hydrogen management for clean fuels production. Advances in Environmental Research, 6(1), 81–98. Hangos, K. M., Cameron, I. T., 2001. Process modelling and model analysis. In Process systems engineering, vol. 4. New York: Academic Press. Hassan, M., Klemeš, J., Pleşu, V., 1999. Process integration analysis and retrofit suggestions for a FCC plant. Integrirovannye Tehnologii i Energosberegenie (Integrated Technologies and Energy Saving), 2, 10–30 (in Russian). HEAT-int, 2009. PIL Web Pages. <www.proint.org/Heat.html> (accessed 14 January 2010).
Bibliography Heckl, I., Friedler, F., Fan, L. T., 2009a. Reduced super-structure for a separation network comprising separators effected by different methods of separation. Computers & Chemical Engineering, 33(3), 687–698. Heckl, I., Friedler, F., Fan, L. T., 2009b. Solution of separation network synthesis problems by the P-graph methodology. Computer Aided Chemical Engineering, 26, S641–S646. Heckl, I., Kovács, Z., Friedler, F., Fan, L. T., Liu, J., 2007. Algorithmic synthesis of an optimal separation network comprising separators of different classes. Chemical Engineering and Processing, 46, 656–665. Hegyháti, M., Adonyi, R., Holczinger, T., Friedler, F., 2006. Demonstration program for S-graph based batch-process scheduling. CHISA 2006 (27–31 August). Prague: Czech Society of Chemical Engineering. Hegyháti, M., Majozi, T., Holczinger, T., Friedler, F., 2009. Practical infeasibility of cross-transfer in batch plants with complex recipes: S-graph vs. MILP methods. Chemical Engineering Science, 64(3), 605–610. Hernández, S., Jiménez, A., 1999. Design of energy-efficient Petlyuk systems. Computers & Chemical Engineering, 23(8), 1005–1010. Herrera, A., Islas, J., Arriola, A., 2003. Pinch technology application in a hospital. Applied Thermal Engineering, 23, 127–139. Heyen, G., Kalitventzeff, B., 2007. Process monitoring and data reconciliation. In L. Puigjaner and G. Heyen (Eds.), Computer Aided Process and Product Engineering, vol. 2, pp. 517–540. New York: Wiley. Himmelblau, D. M., 1988. Applied nonlinear programming. New York: McGrawHill. Hoekstra, A. Y., 2008. The value of water. Research Report Series, 28, UNESCO-IHE, Delft, The Netherlands. Hoekstra, A. Y., Chapagain, A. K., 2007. Water footprints of nations. Water Resources Management, 21(1), 35–48. Hohmann, E. C., 1971. Optimum networks for heat exchange. PhD thesis, University of Southern California, Los Angeles. Holczinger, T., Biros, G., Friedler, F., Sarkozi, N., 2004. Acceleration tools for batch process scheduling. CHISA 2004 (Prague, 22–26 August), Czech Society of Chemical Engineering. Holczinger, T., Majozi, T., Hegyhati, M., Friedler, F., 2007. An automated algorithm for throughput maximization under fixed time horizon in multipurpose batch plants: S-graph approach. Computer Aided Chemical Engineering, 24, 649–654. Holczinger, T., Romero, J., Puigjaner, L., Friedler, F., 2002. Scheduling of multipurpose batch processes with multiple batches of the products. Hungarian Journal of Industrial Chemistry, 30, 305–312. HOMER, 2009. HOMER Energy LLC. <www.homerenergy.com/> (accessed 2 September 2009). Honeywell, 2010. UniSim Design. (accessed 18 January 2010). Hosford, J. E., 1960. Measures of dependability. Operations Research, 8(1), 53–64. Houdková, L., Boráň, J., Ucekaj, V., Elsäßer, T., Stehlík, P., 2008. Thermal processing of sewage sludge—II. Applied Thermal Engineering, 28(16), 2083–2088. Huang, C.-H., Chang, C.-T., Ling, H.-C., Chang, C.-C., 1999. A mathematical programming model for water usage and treatment network design. Industrial & Engineering Chemistry Research, 38(7), 2666–2679. Hul, S., Tan, R. R., Auresenia, J., Fuchino, T., Foo, D. C. Y., 2007. Water network synthesis using mutation-enhanced particle swarm optimization. Process Safety and Environmental Protection, 85(6), 507–514. HYSYS, 2010. Aspen HYSYS. <www.aspentech.com/hysys> (accessed 18 January 2010). IAEA [International Atomic Energy Agency], 2009. Planning and Economic Studies Section (PESS). <www.iaea.org/OurWork/ST/NE/Pess/PESSenergymodels. shtml> (accessed 13 June 2009).
329
330
Bibliography IEA [International Energy Agency], 2009. Energy Technology Systems Analysis Programme. <www.etsap.org/> (accessed 25 April 2009). Ierapetritou, M. G., Floudas, C. A., 1998a. Effective continuous-time formulation for short-term scheduling. Part 1. Multipurpose batch processes. Industrial & Engineering Chemistry Research, 37, 4341–4359. Ierapetritou, M. G., Floudas, C. A., 1998b. Effective continuous-time formulation for short-term scheduling. Part 2. Continuous and semi-continuous processes. Industrial & Engineering Chemistry Research, 37, 4360–4374. ILOG, 2006. ILOG Optimisation Decision Manager. <www.ilog.com/download/ docs/DS-ODM.pdf> (accessed 26 October 2009). ILOG, 2009. ILOG Parallel CPLEX. <www.ilog.com/products/cplex/product/ parallel.cfm> (accessed 26 October 2009). INFORSE-Europe, 2009. Vision 2050: Visions for a Renewable Energy World. <www.inforse.org/europe/Vision2050.htm> (accessed 25 April 2009). Institute for Energy Technology, 2009. HYDROGEMS. <www.hydrogems.no/> (accessed 25 April 2009). Institute for Sustainable Solutions and Innovations, 2009. Energy Rich Japan. <www.energyrichjapan.info> (accessed 4 July 2009). Instituto Superior Técnico, 2010. H2RES (Instituto Superior Técnico & University of Zagreb). <www.powerlab.fsb.hr/h2RES/> (accessed 5 May 2010). International Energy Agency, 2010. Energy Technology Systems Analysis Programme. <www.etsap.org/> (accessed 5 May 2010). International Institute for Applied Systems Analysis, 2009. Energy Modeling Framework: Model for Energy Supply Strategy Alternatives and Their General Environmental Impact (MESSAGE). <www.iiasa.ac.at/Research/ENE/model/ message.html> (accessed 26 April 2009). IPS [Invensys Process Systems], 2009a. HEXTRAN—Heat Transfer. (accessed 26 October 2009). IPS [Invensys Process Systems], 2009b. PRO/II. (accessed 26 October 2009). Ireson W. G., Coombs, C. F., Moss, R. Y., 1996. Handbook of reliability engineering and management. New York: McGraw-Hill. Jansen, P., Koppejan, J., Hetland, J., Klemeš, J., Phuengphaeng, T., Pipio, A., Perry, S., 2004. EMINENT accelerates market introduction of promising early stage technologies for transport and energy. Chemical Engineering Transactions, 5, 187–192. Jegede, F. O., Polley, G. T., 1992. Capital cost targets for networks with non-uniform heat exchanger specifications. Computers & Chemical Engineering, 16(5), 477– 495. Jeżowski, J., 1990. Linear programming based method of heat exchanger network synthesis. Industrial & Engineering Chemistry Process Design, 10, 305–312. Jeżowski, J., Shethna, H. K., Castillo, F. J.L., 2003. Area target for heat exchanger network using linear programming. Industrial & Engineering Chemistry Research, 42(8), 1723–1730. Joint Global Change Research Institute [JGCRI]. 2009. Pacific Northwest National Laboratory and University of Maryland. <www.globalchange.umd.edu/> (accessed 15 June 2009). Karuppiah, R., Grossmann, I. E., 2006. Global optimization for the synthesis of integrated water systems in chemical processes. Computers & Chemical Engineering, 30, 650–673. Karvountzi, G. C., Price, C. M., Duby, P. F., 2004. Comparison of molten carbonate and solid oxide fuel cells for integration in a hybrid system for cogeneration or tri-generation. ASME, Advanced Energy Systems Division AES, 44, 139–150. Kazantzi, V., El-Halwagi, M. M., 2005. Targeting material reuse via property integration. Chemical Engineering Progress, 101(8), 28–37. Kececioglu, D. B., 2002. Reliability engineering handbook, vol 1. Lancaster, PA: DESTech. Kemp Pinch Spreadsheet, 2006. NOV06 Final. (accessed 13 December 2009).
Bibliography Kemp, I. C., 2007. Pinch analysis and process integration. A user guide on process integration for efficient use of energy. Amsterdam: Elsevier. Kemp, I. C., Deakin, A. W., 1989. The cascade analysis for energy and process integration of batch processes. Part 1: Calculation of energy targets. Chemical Engineering Research and Design, 67, 495–509. Khachiyan, L. G., 1979. A polynomial algorithm in linear programming. Doklady Akademii Nauk SSSR, 244, 1093–1096 (in Russian). Kiefer, J., 1953. Sequential minimax search for a maximum. Proceedings of the American Mathematical Society, 4, 502–506. Kiepuszewski, B., ter Hofstede, A. H. M., van der Aalst, W. M. P., 2003. Fundamentals of control flow in workflows. Acta Infromatica, 39(3), 143–209. Kim, S. B., Lee, H. K., Lee, I. B., Lee, E. S., Lee, B., 2000. Scheduling of non-sequential multipurpose batch processes under finite intermediate storage policy. Computers & Chemical Engineering, 24, 1603–1610. Kim, J.-K., Smith, R., 2004. Automated design of discontinuous water systems. Process Safety and Environmental Protection, 82(3), 238–248. Kim, J., Smith, R., 2005. Pinch design and analysis. In Encyclopedia of Chemical Processing, pp. 2165–2180. New York: Marcel Dekker. King, C. J., 1980. Separation processes. New York: McGraw-Hill. Kirkpatrick, S., Gelatt, C. D., Jr., Vecchi, M. P., 1983. Optimization by simulated annealing. Science, 220(4598), 671–680. Klemeš, J., 1977. Chemical process simulation programs—Present state and future development of flowsheeting. Collection of Czechoslovak Chemical Communications, 42(2), 426–445. Klemeš J., 1986. Synthesis of process flowsheets a comprehensive review. 35th National Congress CHISA (Karlovy Vary, Czech Republic), lecture A3.05 (in Czech). Klemeš, J., Bulatov, I., 2001. Integration of technical information with economic and business data in process industries applications. PRES '01 Conference on Process Integration, Modelling, and Optimisation for Energy Saving and Pollution Reduction (Florence, Italy, 20–23 May). Klemeš, J., Bulatov, I., Cockeril, T., 2007. Techno-economic modelling and cost functions of CO2 capture processes. Computers & Chemical Engineering, 31(5–6), 445–455. Klemeš, J., Bulatov, I., Hetland, J., Jansen, P., Koppejan, J., 2005a. Contribution to market introduction of promising early stage hydrogen technologies—An application of EMINENT software tool. Chemical Engineering Transactions, 8, 251–256. Klemeš, J., Bulatov, I., Jansen, P., Koppejan, J., 2005b, Novel energy saving technologies assessment by EMINENT evaluation tool. Chemical Engineering Transactions, 8, 163–168. Klemeš, J., Bulatov, I., Koppejan, J., 2009. Novel energy saving technologies evaluation tool. Computers & Chemical Engineering, 33(3), 751–758. Klemeš, J., Bulatov, I., Perry, S., 2008. Energy efficiency and carbon footprint reduction. In E. Pistikopoulos, M. Georgiadis, and E. S. Kikkinides (Eds.), Process systems engineering, vol. 5, pp. 203–254. Weinheim, Germany: Wiley. Klemeš, J., Cockerill, T., Bulatov, I., Shackley, S., Gough, C., 2006. Engineering feasibility of carbon dioxide capture and storage. In C. Gough and S. Shackley (Eds.), Carbon capture and its storage: An integrated assessment, pp. 43–82. Aldershot, U.K.: Ashgate. Klemeš, J., Dhole, V. R., Raissi, K., Perry, S. J., Puigjaner, L., 1997. Targeting and design methodology for reduction of fuel, power and CO2 on total sites. Applied Thermal Engineering, 7, 993–1003. Klemeš, J., Friedler, F., 2005. Recent novel developments in heat integration—Total site, trigeneration, utility systems and cost-effective decarbonisation. Case studies waste thermal processing, pulp and paper and fuel cells. Applied Thermal Engineering, 25(7), 953–960. Klemeš, J., Kimenov, G., Nenov, N., 1998. Application of pinch-technology in food industry. 1st Conference on Process Integration, Modelling and Optimisation for
331
332
Bibliography Energy Saving and Pollution Reduction—PRES '98 (Prague, August), lecture F6.6. Klemeš, J., Klemsa, J., Luťcha, J., Dohnal, M., Vašek, V., 1975. Local and global extremum finding strategy in optimization of chemical processes. In Computers in the Design and Erection of Chemical Plants, vol. 2, pp. 627–638. Karlovy Vary, Czech Republic: Czechoslovak Chemical Society, Working Group on Computers. Klemeš, J., Linnhoff, B., Kotjabasakis, E., Zhelev, T. K., Gremouti, I., Kaliventzeff, B., Heyen, G., Maréchal, F., Lebon, M., Puigjaner, L., Espuńa, A., Graells, M., Santos, G., Prokopakis, G. J., Ashton, G. J., Murphy, N., de Paor, A. M., Kemp, I.C., 1994. Design and operation of energy efficient batch processes. Brussels: Commission of the European Communities. Klemeš, J., Luťcha, J., Vašek, V., 1979. Resent extension and development of design integrated system DIS. Computers & Chemical Engineering, 3(4), 357–361. Klemeš, J., Nenov, N., Kimenov, G., Mintchev, M., 1999. Heat integration in food industry. Integrirovannye Tehnologii i Energosberegenie (Integrated Technologies and Energy Saving), 4, 9–26 (in Russian). Klemeš, J., Perry, S., 2007a. Process optimisation to minimise energy use in food processing. In K. Waldron (Ed.), Handbook of waste management and co-product recovery in food processing, vol. 1, pp. 59–89. Cambridge, U.K.: Woodhead. Klemeš, J., Perry, S., 2007b. Process optimisation to minimise water use and wastage. In K. Waldron (Ed.), Waste management and co-product recovery in food processing, vol. 1, pp. 90–118. Cambridge, U.K.: Woodhead Klemeš, J., Phuengphaeng, T., Bulatov, I., Jansen, P., Koppejan, J., 2005c. Novel energy saving technologies evaluation tool. In Proceedings of Heat Transfer in Components and Systems for Sustainable Energy Technologies—Heat SET 2005 (Grenoble, France, April), pp. 663–668. Grenoble: Groupement pour la Recherche sur les Echangeurs Thermiques—GRETh. Klemeš, J., Ptáčník, R., 1985. Computer-aided synthesis of heat exchange network. Journal of Heat Recovery Systems, 5(5), 425–435. Klemeš, J., Smith, R., Kim, J.-K. (Eds.), 2008. Handbook of water and energy management in food processing. Cambridge, U.K.: Woodhead. Klemeš, J.,Varbanov, P., 2008. Total sites for combined utilisation of renewables and fossil fuels via improved energy conversion technologies. In Proceedings of CHISA 2008, K4.2, pp. 1242–1243. Prague: Czech Society of Chemical Engineering. Klemeš, J., Varbanov, P., Lam, H. L., 2009. Water footprint, water recycling and food industry supply chain. In K. Waldron (Ed.), Waste management and co-product recovery in food processing, vol. 2, pp. 134–168. Cambridge, U.K.: Woodhead. Klemeš, J., Vašek, V., 1973. Methods for optimising complex chemical processes. In Proceedings of the 2nd Symposium on the Use of Computers in Chemical Engineering (Ústí nad Labem, Czech Republic), pp. 84–102. Prague: Czechoslovak Chemical Society, Working Group on Computers. Klemeš, J., Zhang, N., Bulatov, I., Jansen, P., Koppejan, J., 2005d. Novel energy saving technologies assessment by EMINENT evaluation tool. Chemical Engineering Transactions, 7, 163–167. Kok, A. G., Graves, S. C., 2003. Supply chain management: Design cooperation and operation (Handbooks in Operations Research and Management Science, vol. 11). New York: Elsevier. Kondili, E., Pantelides, C. C., Sargent, R. W. H., 1993. A general algorithm for short-term scheduling of batch operations. I. MILP formulation. Computers & Chemical Engineering, 17(2), 211–227. Koschuh, W., Povoden, G., Hong, T.-V., Kromus, S., Kulbe, H. D., Novalin, S., 2004. Production of leaf protein concentrate from ryegrass (Loliumperenne × multiflorum) and alfalfa (Medicago sativa subsp. sativa). Comparison between heat coagulation/centrifugation and ultrafiltration. Desalination, 163, 253–259. Kotjabasakis, E., Linnhoff, B., 1986. Sensitivity tables for the design of flexible processes (1)―How much contingency in heat exchanger networks is costeffective? Chemical Engineering Research and Design, 64(3), 197–211.
Bibliography Kovacs, Z., Ercsey, Z., Friedler, F., Fan, L. T., 2000. Separation-network synthesis: Global optimum through rigorous super-structure. Computers & Chemical Engineering, 24, 1881–1900. Kovacs, Z., Friedler, F., Fan, L. T., 1993. Recycling in a separation process structure. AIChE Journal, 39, 1087–1089. Kravanja, S., Kravanja, Z., Bedenik, B. S., 1998a. The MINLP optimization approach to structural synthesis. Part I: A general view on simultaneous topology and parameter optimization. International Journal for Numerical Methods in Engineering, 43, 263–292. Kravanja, S., Kravanja, Z., Bedenik, B. S., 1998b. The MINLP optimization approach to structural synthesis. Part II: Simultaneous topology, parameter and standard dimension optimization by the use of the linked two-phase MINLP strategy. International Journal for Numerical Methods in Engineering, 43, 293–328. Kravanja, S., Šilih, S., Kravanja, Z., 2005. The multilevel MINLP optimization approach to structural synthesis: the simultaneous topology, material, standard and rounded dimension optimization. Advances in Engineering Software, 36(9), 568–583. Kravanja, Z., 2009. The development of an advanced systems synthesis environment: Integration of MI(NL)P methods and tools for sustainable applications. Computer Aided Chemical Engineering, 26, 25–35. Kravanja, Z., Grossmann, I. E., 1990. PROSYN—An MINLP process synthesizer. Computers & Chemical Engineering, 14(12), 1363–1378. Kravanja, Z., Grossmann, I. E., 1994. New developments and capabilities in PROSYN—An automated topology and parameter process synthesizer. Computers & Chemical Engineering, 18(11–12), 1097–1114. Kromus, S. G., 2002. Biorefinery Austria—Development of an integrated system for the use of grassland biomass. PhD thesis, Technical University Graz, Austria (in German). Kuhn, H. W., Tucker, A. W., 1951. Nonlinear programming. In J. Neyman (Ed.), Proceedings of the Second Berkeley Symposium on Mathematical Statistics and Probability, pp. 481–493. Berkeley: University of California Press. Kuo, W. J., Smith, R., 1997. Effluent treatment system design. Chemical Engineering Science, 52(23), 4273–4290. Kuo, W., Zuo, M. J., 2003. Optimal reliability modelling—Principles and applications. Hoboken, NJ: Wiley. Kurz, R., 2005. Parameter optimization on combined gas turbine–fuel cell power plants. ASME Journal of Fuel Cell Science and Technology, 2, 268–273. Lam, H. L., Varbanov. P., Klemeš, J., 2009. Optimisation of regional energy supply chains utilising renewables: P-graph approach. Computer Aided Chemical Engineering, 26, 1003–1008. Lam, H. L., Varbanov, P., Klemeš, J., 2010. Minimising carbon footprint of regional biomass supply chains. Resources, Conservation and Recycling, 54(5), 303–309. Land, A. H., Doig, A. G., 1960. An automatic method of solving discrete programming problems. Econometrica, 28(3), 497–520. Langenstein, M., Jansky, J., Laipple, B., 2004. Finding megawatts in nuclear power plants with process ata reconciliation. In Proceedings of the 12th International Conference on Nuclear Engineering—ICONE12 (Arlington, VA). <www.btbjansky. com/pdf/Q_70_ICONE-final-paper.pdf> (accessed 3 October 2007). Laue, H. J., 2006. Heat pumps. In C. Clauser, T. Strobl, and F. Zunic (Eds.), Renewable energy, vol. 3C. Berlin: Springer. Leewongtanawit, B., 2005. Heat-integrated water system design. PhD thesis, UMIST, Manchester, U.K. Leewongtanawit, B., Kim, J., 2008. Synthesis and optimisation of heat-integrated multiple-contaminant water systems. Chemical Engineering and Processing, 47(4), 670–694. Leewongtanawit, B., Kim, J., 2009. Improving energy recovery for water minimisation. Energy, 34(7), 880–893. Li, B. H., Chang, C. T., 2006. A mathematical programming model for discontinuous water-reuse system design. Industrial & Engineering Chemistry Research, 45(14), 5027–5036.
333
334
Bibliography LINDO System, 2009. LINDO System Website. <www.lindo.com> (accessed 26 October 2009). Linnhoff, B., Ahmad, S., 1990. Cost optimum heat exchanger networks—1. Minimum energy and capital using simple models for capital cost. Computers & Chemical Engineering, 14(7), 729–750. Linnhoff, B., Akinradewo, C. G., 1998. Automated interface between dimulation and integration, CHISA 1998 / 1st Conference PRES 1998, Plenary lecture A3.0 [818] Prague: Czech Society of Chemical Engineering. Linnhoff, B., Dhole, V. R., 1992. Shaftwork targets for low temperature process design. Chemical Engineering Science, 47(8), 2081–2091. Linnhoff, B., Dhole, V. R., 1993. Targeting for CO2 emissions for total sites. Chemical Engineering and Technology, 16(4), 252–259. Linnhoff, B., Flower, J. R., 1978. Synthesis of heat exchanger networks: I. Systematic generation of energy optimal networks. AIChE Journal, 24(4), 633–642. Linnhoff, B., Hindmarsh, E., 1983. The pinch design method for heat exchanger networks. Chemical Engineering Science, 38(5), 745–763. Linnhoff, B., Mason, D. R., Wardle, I., 1979. Understanding Heat Exchanger Networks. Computers & Chemical Engineering, 3(1–4), 295–302. Linnhoff, B., Townsend, D. W., Boland, D., Hewitt, G. F., Thomas, B. E. A., Guy, A. R., Marsland, R. H., 1982. A user guide on process integration for the efficient use of energy. Rugby, U.K.: IChemE [revised edition published in 1994]. Linnhoff, B., Vredeveld, D. R., 1984. Pinch technology has come of age. Chemical Engineering Progress, 80(7), 33–40. Linnhoff March, 1998. Introduction to Pinch Technology. <www.linnhoffmarch. com/pdfs/PinchIntro.pdf> (aaccessed 27 October 2009). Linnhoff March, 2009. Introduction to Pinch Technology. <www.kbcat.com/Productsand-Services/Software/Energy-Optimisation-Software-/SuperTarget/> (accessed 11 January 2010). Liu, F., Zhang, N., 2004. Strategy of purifier selection and integration in hydrogen networks. Chemical Engineering Research and Design, 82(A10), 1315–1330. Liu, J., Fan, L. T., Seib, P., Friedler, F., Bertok, B., 2004. Downstream process synthesis for biochemical production of butanol, ethanol, and acetone from grains: Generation of optimal and near-optimal flowsheets with conventional operating units. Biotechnology Progress, 20, 1518–1527. Liu, J., Fan, L. T., Seib, P. A., Friedler, F., Bertok, B., 2006. Holistic approach to process retrofitting: Application to downstream process for biochemical production of organics. Industrial & Engineering Chemistry Research, 45, 4200–4207. Liu, Y., Yuan, X., Luo, Y., 2007. Synthesis of water utilization system using concentration interval analysis method, II. Discontinuous process. Chinese Journal of Chemical Engineering, 15(3), 369–375. Luenberger D. G., Ye, Y., 2008. Linear and nonlinear programming, 3rd ed. New York: Springer. Macias-Corral, M., Samani, Z., Hanson, A., Smith, G., Funk, P., Yu, H., Longworth, J., 2008. Anaerobic digestion of municipal solid waste and agricultural waste and the effect of co-digestion with dairy cow manure. Bioresource Technology, 99(17), 8288–8293. Majozi, T., 2005. An effective technique for wastewater minimisation in batch processes. Journal of Cleaner Production, 13, 1374–1380. Majozi, T., Brouckaert, C. J., Buckley, C. A., 2006. A graphical technique for wastewater minimisation in batch processes. Journal of Environmental Management, 78, 317–329. Majozi, T., Friedler, F. 2006. Maximization of throughput in a multipurpose batch plant under fixed time horizon: S-graph approach. Industrial & Engineering Chemistry Research, 45, 6713–6720. Majozi, T., Zhu, X. X., 2001. A novel continuous time MILP formulation for multipurpose batch plants. 1. Short-term scheduling. Industrial & Engineering Chemistry Research, 40(25), 5935–5949. Makwana, Y., 1998. Energy retrofit and debottlenecking of total sites. PhD thesis, UMIST, Manchester, U.K.
Bibliography Makwana, Y., Smith, R., Zhu, X. X., 1998. A novel approach for retrofit and operations management of existing total sites. Computers & Chemical Engineering, 22(Suppl.), S793–S796. Manan, Z. A., Foo, C. Y., Tan, Y. L., 2004. Targeting the minimum water flow rate using water cascade analysis technique. AIChE Journal, 50(12), 3169– 3183. Manan, Z. A., Tea, S. Y., Alwi, S. R. W., 2009. A new technique for simultaneous water and energy minimisation in process plant. Chemical Engineering Research and Design, 87(11), 1509–1519. Mandil, C., 2005. Saving electricity in a hurry. Paris: IEA Publications. <www.iea. org/textbase/nppdf/free/2005/savingElec.pdf> (accessed 5 August2009). Manolas, D. A., Frangopoulos, C. A., Gialamas, T. P., Tsahalis, D. T., 1997. Operation optimization of an industrial cogeneration system by a genetic algorithm. Energy Conversion and Management, 38(15–17), 1625–1636. Maros, I., 2003a. A piecewise linear dual phase-1 algorithm for the simplex method. Computational Optimization and Applications, 26(1), 63–81. Maros, I., 2003b. Computational techniques of the simplex method. Boston: Kluwer. Masruroh, N. A., Li, B., Klemeš, J., 2006. Life cycle analysis of a solar thermal system with thermochemical storage process. Renewable Energy, 31(4), 537–548. Massardo, A. L., Bosio, B., 2002. Assessment of olten carbonate fuel cell models and integration with gas and steam cycles. Journal of Engineering for Gas Turbines and Power, 124, 103–109. Masso, A. H., Rudd, D. F., 1969. The synthesis of system designs II. Heuristic structuring. AIChE Journal, 15(1), 10–17. MathWorks, 2009. Getting Started with Matlab. <www.mathworks.com/access/ helpdesk/help/techdoc/matlab_product_page.html> (accessed 26 October 2009). Mavromatis, S. P., Kokossis, A. C., 1998. Conceptual optimisation of utility networks for operational variations—II. Network development and optimization. Chemical Engineering Science, 53(8), 1609–1630. McKay, B., Willis, M., Barton, G., 1997. Steady-state modelling of chemical process systems using genetic programming. Computers & Chemical Engineering, 21(9), 981–996. Mechefske, C. K., Wang, Z., 2003. Using fuzzy linguistics to select optimum maintenance and condition monitoring strategies. Mechanical Systems and Signal Processing, 17(2), 305–316. Mehrotra, S., 1992. On the implementation of a primal-dual interior point method. SIAM Journal on Optimization, 2(4), 575–601. Meixell, M. J., Gargeya, V. B., 2005. Global supply chain design: A literature review and critique. Transportation Research Part E, 41, 531–550. Méndez C. A, Cerdá, J., 2003. An MILP continuous-time framework for shortterm scheduling of multipurpose batch process under different operation strategies. Optimization and Engineering, 4, 7–22. Minet, F., Heyen, G., Kalitventzeff, B., Di Puma, J., Malmendier, M., 2001. Dynamic data reconciliation of regenerative heat exchangers coupled to a blast furnace. Computer Aided Chemical Engineering, 9, 1053–1058. MiniCAM, 2009. Pacific Northwest National Laboratory. <www.pnl.gov/gtsp/ research/minicam.stm> (accessed 26 April 2009). Modelica, 2009a. Modelica and Modelica Association. <www.modelica.org> (accessed 26 October 2009). Modelica, 2009b. Modelica—Language, Libraries, Tools, Workshop and EU-Project RealSim. <www.modelica.org/documents/ModelicaOverview14.pdf> (accessed 26 October 2009). Modelica, 2010. Modelica Tools. <www.modelica.org/tools> (accessed 26 April 2010). Monod, J., 1949. The growth of bacterial cultures. Annual Review of Microbiology, 3, 371–394. Montgomery, D. C., 2005. Design and analysis of experiments, 6th ed. New York: Wiley.
335
336
Bibliography Morrissey, A. J., Browne, J., 2004. Waste management models and their application to sustainable waste management. Waste Management, 24, 297–308. MW Kellogg, 1998. Kellogg divided wall column (promotional brochure). Henley, U.K. Nagy, A. B., Adonyi, R., Halasz, L., Friedler, F., Fan, L. T., 2001. Integrated synthesis of process and heat exchanger networks: Algorithmic approach. Applied Thermal Engineering, 21(13–14), 1407–1427. Napper, D., Kim, J.-K., Bulatov, I., 2009. Hygienic and sustainable use and reuse of water and energy in food factories. In K. Waldron (Ed.), Waste management and co-product recovery in food processing, vol. 2, pp. 169–195. Cambridge, U.K.: Woodhead. National Resources Canada, 2009. RETScreen International. <www.retscreen.net> (accessed 26 April 2009). National Technical University of Athens, 2009. Energy–Economics–Environment Modelling Laboratory Research and Policy Analysis. <www.e3mlab.ntua. gr/> (accessed 26 April 2009). Nemhauser, G. L., Wolsey, L. A., 1999. Integer and combinatorial optimization. New York: Wiley. Ng, D. K. S., Foo, D. C. Y., Rabie, A., El-Halwagi, M. M., 2008. Simultaneous synthesis of property-based water reuse/recycle and interception networks for batch processes. AIChE Journal, 54(10), 2624–2632. Nie, X. R., Zhu, X. X., 1999. Heat exchanger network retrofit considering pressure drop and heat-transfer enhancement. AIChE Journal, 45(6), 1239–1254. Nishida, N., Liu, Y. A., Lapidus, L., 1977. Studies in chemical process design and synthesis: III. A Simple and practical approach to the optimal synthesis of heat exchanger networks. AIChE Journal, 23(1), 77–93. NRCan, 2007. Natural Resources Canada Success Stories. (accessed 1 October 2007). Oak Ridge National Laboratory, 2009a. Oak Ridge Competitive Electricity Dispatch (ORCED) Model. <www.ornl.gov/sci/ees/pes/capabilities_ORCED.html> (accessed 26 April 2009). Oak Ridge National Laboratory, 2009b. Whole-Building and Community Integration Program. <www.coolingheatingpower.org/about/bchp-screening-tool.php> (accessed 26 April 2009). Octave, 2009. Octave Website. <www.gnu.org/software/octave> (accessed 26 October 2009). ODC, 2009. LINDO API. <www.odcindia.com/products/optimizationtool/ lindoapi.php> (accessed 26 October 2009). Odum, H. T., 1996. Environmental accounting: Emergy and environmental decision making. New York: Wiley. OpenModelica, 2010. The OpenModelica Project. <www.ida.liu.se/~pelab/ modelica/OpenModelica.html> (accessed 18 January 2010). Panjeshahi, M. H., Tahouni, N., 2008. Pressure drop optimisation in debottlenecking of heat exchanger networks. Energy, 33(6), 942–951. Pantelides, C. C., 1993. Unified frameworks for optimal process planning and scheduling. In D. W. T. Rip-pin, J. C. Hale, and J. Davis (Eds.), Proceedings of the Second International Conference on Foundations of Computer-Aided Process Operations, pp. 253–274. Crested Butte, CO: Computer Aids for Chemical Engineering—CACHE. Papageorgiou, L. G., Georgiadis, M. C., 2008a. Process systems engineering, vol. III: Supply chain optimization, part 1. New York: Wiley. Papageorgiou, L. G., Georgiadis, M. C., 2008b. Processs systems engineering, vol. IV: Supply chain optimization, part 2. New York: Wiley. Papalexandri, K. P., Pistikopoulos, E. N., 1996. Generalized modular representation framework for process synthesis. AIChE Journal, 42(4), 1010–1032. Papoulias, S. A., Grossmann, I. E. 1983. A structural optimization approach in process synthesis. Part I: Utility systems. Computers & Chemical Engineering, 7, 695–706. Pekny, J., Reklaitis, G., 1998. Towards the convergence of theory and practice: A technology guide for scheduling/planning methodology. In J. F. Pekny
Bibliography and G. E. Blau (Eds.), Foundations of computer-aided process operations (AIChE Symposium Series, no. 320), pp. 91–111. New York: AIChE. Perry, S., Bulatov, I., Klemeš, J., 2007. The potential of the EMINENT tool in the screening and evaluation of emerging technologies for CO2 reduction related to buildings. Chemical Engineering Transactions, 12, 709–714. Perry, S., Klemeš, J., Bulatov, I., 2008. Integrating waste and renewable energy to reduce the carbon footprint of locally integrated energy sectors. Energy, 33(10), 1489–1497. Peters, M. S., Timmerhaus, K. D., West, R. E., 2003. Plant design and economics for chemical engineers. New York: McGraw-Hill. Pettersson, K., Harvey, S., 2010. CO2 emission balances for different black liquor gasification biorefinery concepts for production of electricity or secondgeneration liquid biofuels. Energy, 35(2), 1101–1106. Pinto, J. M., Grossmann, I. E., 1994. Optimal cyclic scheduling of multistage continuous multiproduct plants. Computers & Chemical Engineering, 18(9), 797–816. Pinto, J. M., Grossmann, I. E., 1995. A continuous time mixed integer linear programming model for short-term scheduling of multistage batch plants. Industrial & Engineering Chemistry Research, 34, 3037–3051. Pistikopoulos, E. N., Georgiadis, M. C., Dua, V., 2007a. Process systems engineering, vol. I: Multi-parametric programming: Theory, algorithms and applications. New York: Wiley. Pistikopoulos E. N., Georgiadis M. C., Dua ,V., 2007b. Process systems engineering, vol. II: Multi-parametric model-based control: Theory and applications. New York: Wiley. Pleşu, V., Klemeš, J., Georgescu, M., 1998. Applications of process integration in Romanian oil refining and petrochemical industry. CHISA '98 / 1st Conference PRES '98, lecture F6.7. Prague: Czech Society of Chemical Engineering. PNS Editor, 2010. <www.p-graph.com> (accessed 18 January 2010). Polley, G., Panjeh Shahi, M. H., Jegede, F. O., 1990. Pressure drop considerations in the retrofit of heat exchanger networks. Chemical Engineering Research and Design, 68(3), 211–220. Polley, G., Reyes Athie, C. M., Gough, M., 1992. Use of heat transfer enhancement in process integration. Heat Recovery Systems and CHP, 12(3), 191–202. Ponton, J. W., Klemeš, J., 1993. Alternatives to neural networks for inferential measurement. Computers & Chemical Engineering, 17(10), 991–1000. POST [U.K. Parliamentary Office for Science and Technology], 2006. Carbon Footprint of Electricity Generation, no. 268. <www.parliament.uk/documents/ upload/postpn268.pdf> (accessed 29 December 2009). Povoden, G., 2002. Analysis of protein loss in grass juice. MSc dissertation, University of Technology, Graz, Austria (in German). Prakash, R., Shenoy, U. V., 2005. Targeting and design of water networks for fixed flowrate and fixed contaminant load operations. Chemical Engineering Science, 60(1), 255–268. Prakotpol, D., Srinophakun, T., 2004. GAPinch: Genetic algorithm toolbox for water pinch technology. Chemical Engineering and Processing, 43(2), 203–217. Prékopa, A., 1995. Stochastic programming. Dordrecht: Springer. PRES, 2009. Conference on Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction. <www.conferencepres.com> (accessed 1 November 2009). PSE, 2009. gPROMS. <www.psenterprise.com/gproms> (accessed 26 October 2009). Raissi, K., 1994. Total site integration. PhD thesis, UMIST, Manchester, U.K. Rašković, P., 2002. Industrial energy system optimization based on heat exchanger network synthesis. Working paper, Faculty of Mechanical Engineering, University of Nis, Serbia and Montenegro. Rašković, P., 2006. Process integration approach for energy saving and pollution prevention in industrial plants. French-Serbian European Summer University: Renewable Energy Sources and Environment-Multidisciplinary Aspect (Vrnjačka Banja, Serbia, 17–24 October).
337
338
Bibliography Rašković, P., 2010. Water pinch technology for designing wastewater reduction and water conservation systems. Working paper, Faculty of Mechanical Engineering, University of Nis, Serbia and Montenegro. <water-environment. vin.bg.ac.rs/presentations/banja/Raskovic/Raskovic%20paper%20VB%20 2007> (accessed 3 January 2010). Ravagnani, M. A. S. S., Silva, A. P., Arroyo, P. A., Constantino A. A., 2005. Heat exchanger network synthesis and optimisation using genetic algorithm. Applied Thermal Engineering, 25(7), 1003–1017. Risø National Laboratory, 2009. WILMAR: Wind Power Integrated in Liberalised Electricity Markets. <www.wilmar.risoe.dk> (accessed 27 April 2009). Romero, J., Puigjaner, L., Holczinger, T., Friedler, F., 2004. Scheduling intermediate storage multipurpose natch plants using the S-graph. AIChE Journal, 50(2), 403–417. Ryu, J., Pistikopoulos, E. N., 2005. Design and operation of an enterprise-wide process network using operation policies. 1. Design. Industrial & Engineering Chemistry Research, 44, 2174–2182. Saboo, A. K., Morari, M., Woodcock, D. C., 1985. Design of resilient processing plants, VIII. A resilience index for heat exchanger networks. Chemical Engineering Science, 40(8), 1553–1565. Sahdev, M., 2010. Chemical Engineering, Pinch technology: Basics for Beginners. <www.cheresources.com/pinchtech.pdf> (accessed 2 January 2010). Sanmartí, E., Friedler, F., Puigjaner, L., 1998. Combinatorial technique for short term scheduling of multipurpose batch plants based on schedule-graph representation. Computers & Chemical Engineering, 22(Suppl.), S847–S850. Sanmartí, E., Holczinger, T., Puigjaner, L., Friedler, F., 2002. Combinatorial framework for effective scheduling of multipurpose batch plants. AIChE Journal, 48(11), 2557–2570. Santos, L. C., Zemp, R. J., 2000. Energy and capital targets for constrained heat exchanger networks. Brazilian Journal of Chemical Engineering, 17(4–7), 659–670. Sargent, R. W. H., 1979. Flowsheeting. Computers & Chemical Engineering, 3(1–4), 17–20. Sargent, R. W. H., 1983. Computers in chemical engineering—Challenges and constraints. AIChE Symposium Series, 79(235), 57–64. Savulescu, L., 1999. Simultaneous energy and water minimisation. PhD thesis, UMIST, Manchester, U.K. Savulescu, L., Kim, J., 2008. Novel methods for combined energy and water minimisation in the food industry. In J. Klemeš, R. Smith, and J.-K. Kim (Eds.), Handbook of water and energy management in food processing. Cambridge, U.K.: Woodhead. Savulescu, L., Kim, J., Smith, R., 2005a. Studies on simultaneous energy and water minimisation—Part I: Systems with no water re-use. Chemical Engineering Science, 60(12), 3279–3290. Savulescu, L., Kim, J., Smith, R., 2005b. Studies on simultaneous energy and water minimisation—Part II: Systems with maximum re-use of water. Chemical Engineering Science, 60(12), 3291–3308. Saw, S. Y., Lee, L., Lim, M. H., Foo, D. C. Y., Chew, I. M. L., Onyedim, V. F.O., Tan, R. R., Klemeš, J., 2009. A new graphical targeting procedure for water minimisation. 3rd International Conference on Chemical and Bioprocess Engineering—ICCBPE 2009 (Sabah, Malaysia). Schilling, G., Pantelides, C. C., 1996. A simple continuous-time process scheduling formulation and a novel solution algorithm. Computers & Chemical Engineering, 20(Suppl.), S1221–S1226. Schlenzig, C., 2009. Seven2one Modelling: Building Integrated Energy Concepts with Mesap/PlaNet (in German). <www.seven2one.de/de/technologie/ mesap.html> (accessed 30 December 2009). Scilab, 2009. Scilab Website. <www.scilab.org> (accessed 26 October 2009). Seferlis, P., Georgiadis, M. C., 2004. The integration of process design cnd control. Amsterdam: Elsevier.
Bibliography Segurado, R., Pereira, S., Pipio, A., Alves, L., 2009. Comparison between EMINENT and other energy technology assessment tools. Journal of Cleaner Production, 17(10), 907–910. Seidler, I., Badach, A., Molisz, W., 1980. Methods of solving of optimization problems. Warsaw: Wydawnictwa Naukowo Techniczne. Seikova, I., Varbanov, P., Ivanova, E., 1999. Debottlenecking of a heat-integrated crude oil distillation system. In F. Friedler and J. Klemeš (Eds.), Proceedings of PRES '99, pp. 583–588. Budapest: Hungarian Chemical Society. SEMPRA ENERGY, 2009. Energy Saving Tips. <www.sdge.com/business/esc/ promo_tipsresources.shtml> (accessed 5 August 2009). Serna-González, M., Jiménez-Gutiérrez, A., Ponce-Ortega, J.M., 2007. Targets for heat exchanger network synthesis with different heat transfer coefficients and non-uniform exchanger specifications. Chemical Engineering Research and Design, 85(10), 1447–1457. S-Graph, 2009. S-Graph for Scheduling. <www.s-graph.com> (accessed 20 December 2009). Shah, N., Pantelides, C. C., Sargent, R. W. H., 1993. A general algorithm for shortterm scheduling of batch operations. Part II. Computational issues. Computers & Chemical Engineering, 17(2), 229–244. Shah, R. K., Sekulić, D. P., 2003. Fundamentals of heat exchanger design. New York: Wiley. Shang, Z., 2000. Analysis and optimisation of total site utility systems. PhD thesis, UMIST, Manchester, U.K. Sharma, R., Jindal, A., Mandawala, D., Jana, S. K., 1999. Design/retrofit targets of pump-around refluxes for better energy integration of a crude distillation column. Industrial & Engineering Chemistry Research, 38(6), 2411–2417. Shelley, M. D., El-Halwagi, M. M., 2000. Componentless design of recovery and allocation systems: A functionality-based clustering approach. Computers & Chemical Engineering, 24, 2081–2091. Shenoy, U., 1995. Heat exchanger network synthesis: Process optimization by energy and resource analysis. Houston, TX: Gulf Professional. Shilling, R. L., Bell, K. J., Berhhagen, P. M., Flynn, T. M., Goldschmidt, V. M., Hrnjak, P. S., Standiford, F. C., Timmerhaus, K. D., 2008. Heat-transfer equipment. In D. W. Green and R. H. Perry (Eds.), Perry’s chemical engineers handbook, 8th ed, chap. 11. New York: McGraw-Hill. Shoaib, A. M., Said, M. A., Moustafa, E. A., Foo, D. C.Y., El-Halwagi, M. M., 2008. A hierarchical approach for the synthesis of batch water network. Computers & Chemical Engineering, 32(3), 530–539. Shopova, E. G., Vaklieva-Bancheva, N. G., 2006. BASIC—A genetic algorithm for engineering problems solution. Computers & Chemical Engineering, 30(8), 1293–1309. Sieniutycz, S., Jeżowski, J., 2009. Energy optimization in process systems. Amsterdam: Elsevier. Sikos, L., Klemeš, J., 2009. RAMS contribution to efficient waste minimisation and management. Journal of Cleaner Production, 17(10) 932–939. Sikos, L., Klemeš, J., 2010a. Evaluation and assessment of reliability and availability software for securing an uninterrupted energy supply. Clean Technologies and Environmental Policy, 12(2), 137–146. Sikos, L., Klemeš, J., 2010b. Reliability, availability and maintenance optimisation of heat exchanger networks. Applied Thermal Engineering, 30, 63–69. Singhvi, A., 2002. Multi-level planning of supply chains under uncertainty. MTech dissertation, Indian Institute of Technology, Bombay. Singhvi, A., Madhavan, K. P., Shenoy, U. V., 2004. Pinch analysis for aggregate production planning in supply chains. Computers & Chemical Engineering, 28(6–7), 993–999. Singhvi, A., Shenoy, U. V., 2002. Aggregate planning in supply chains by pinch analysis. Chemical Engineering Research and Design, 80(6), 597–605. SINTEF, 2009a. <www.sintef.no/Home> (accessed 12 June 2009).
339
340
Bibliography SINTEF, 2009b. EOPS and EMPS. <www.sintef.no/Home/Petroleum-and-Energy/ SINTEF-Energy-Research/Software/EOPS-and-EMPS/> (accessed 23 April 2009). SITE-int, 2009. PIL Web Pages. <www.proint.org/Site_int.html> (accessed 14 January 2010). Smith, R., 1995. Chemical process design. New York: McGraw-Hill. Smith, R., 2005. Chemical process design and integration, Chichester, U.K.: Wiley. Smith, R., Klemeš, J., Tovazhnyansky, L. L., Kapustenko, P. A., Uliev, L. M., 2000. Foundations of heat processes integration. Kharkiv, Ukraine: NTU KhPI (in Russian). Smith, R., Linnhoff, B., 1988. The design of separators in the context of overall processes. Chemical Engineering Research and Design, 66(3), 195–228. Smith, R.,Varbanov, P., 2005. What’s the price of steam? Chemical Engineering Progress, 101(7), 29–33. Solver.com, 2009. Solver.com Website <www.solver.com/selectqa.htm> (accessed 26 October 2009). Soršak, A., Kravanja, Z., 2004. MINLP retrofit of heat exchanger networks comprising different exchanger types. Computers & Chemical Engineering, 28(1–2), 235–251. Sparrow, R. E., Forder, G. J., Rippin, D. W. T., 1975. The choice of equipment sizes for multiproduct batch plants. Heuristics vs. branch and bound. Industrial and Engineering Chemistry Process Design and Development, 14(3), 197–203. Spriggs, H. D., El-Halwagi, M. M., 2000. Pollution prevention through mass integration. AIChE Symposium Series, 96(323), 367–370. SPRINT, 2009. Process Integration Software (Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester, U.K.). <www.ceas.manchester.ac.uk/research/centres/ centreforprocessintegration/software/packages/sprint> (accessed 26 October 2009). Sricharoenchaikul, V., 2009. Assessment of black liquor gasification in supercritical water. Bioresource Technology, 100(2), 638–643. STAR, 2009. Process Integration Software (Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester, U.K.). <www.ceas.manchester.ac.uk/research/centres/ centreforprocessintegration/software/packages/star> (accessed 26 October 2009). Stehlík, P., 2009. Contribution to advances in waste-to-energy technologies. Journal of Cleaner Production, 17(10), 919–931. Steppan, D. D., Werner, J., Yeater, R. P., 1998. Essential Regression and Experimental Design for Chemists and Engineers. <www.jowerner.homepage.t-online.de/ index.html> (accessed 1 March 2007). Stern, 2006. Stern Review Report on the Economics of Climate Change. <www. hm-treasury.gov.uk/stern_review_report.htm> (accessed 15 January 2010). Stockholm Environment Institute, 2009. Community for Energy, Environment and Development. <www.energycommunity.org/> (accessed 25 April 2009). STREAM, 2009. The STREAM Modelling Tool. <www.streammodel.org/> (accessed 30 June 2009). Taal, M., Bulatov, I., Klemeš, J., Stehlik, P., 2003. Cost estimation and energy price forecast for economic evaluation of retrofit projects. Applied Thermal Engineering, 23, 1819–1835. Takama, N., Kuriyama, T., Shiroko, K., Umeda, T., 1980. Optimal water allocation in a petroleum refinery. Computers & Chemical Engineering, 4(4), 251–258. Tan, R. R., 2007. Hybrid evolutionary computation for the development of pollution prevention and control strategies. Journal of Cleaner Production, 15(10), 902–906. Tan, R. R., 2008. An adaptive swarm-based simulated annealing algorithm for process optimization. In PRES 2008/CHISA 2008, Summaries 4, pp. 1257–1258. Prague: Czech Society of Chemical Engineering.
Bibliography Tan, R. R., Cruz, D. E., 2004. Synthesis of robust water reuse networks for single component retrofit problems using symmetric fuzzy linear programming. Computers & Chemical Engineering, 28, 2547–2551. Tan, R., Foo, D. C. Y., 2007. Pinch analysis approach to carbon-constrained energy sector planning. Energy, 32(8), 1422–1429. Ten Broeck, H., 1944. Economic selection of exchanger sizes. Industrial & Engineering Chemistry Research, 36(1), 64–67. Thevendiraraj, S., Klemeš, J., Paz, D., Aso, G., Cardenas, J., 2003. Water and wastewater minimisation study of a citrus plant. Resources, Conservation and Recycling, 37, 227–250. Tick, J., Kovács, Z., Friedler, F., 2006. Synthesis of optimal workflow structure. Journal of Universal Computer Science, 12, 1385–1392. Timpe, C. H., Kallrath, J., 2000. Optimal planning in large multi-site production networks. European Journal of Operational Research, 126, 422–435. Tjoe, T. N., Linnhoff, B., 1986. Pinch technology retrofit: Setting targets for an existing plant. Chemical Engineering, 93(8), 47–60. Tokos, H., Novak Pintarič, Z., 2009. Synthesis of batch water network for a brewery plant. Journal of Cleaner Production, 17(16), 1465–1479. TOMLAB, 2010. MATLAB Optimization—TOMLAB. (accessed 13 January 2010). Tovazshnyansky, L. L., Kapustenko, P. O., Khavin, G. L., Arsenyeva, O. P., 2004. PHEs in industry. Kharkiv, Ukraine: NTU KhPI (in Russian). Townsend, D. W., Linnhoff, B., 1983. Heat and power networks in process design. Part II: Design procedure for equipment selection and process matching. AIChE Journal, 29(5), 748–771. Townsend, D. W., Linnhoff, B., 1984. Surface area targets for heat exchanger networks. IChemE 11th Annual Research Meeting (Bath, U.K., April). Triantafyllou, C., Smith R., 1992. The design and optimisation of fully thermally coupled distillation columns. Chemical Engineering Research and Design, 70(A2), 118–132. TRNSYS, 2009. A TRaNsient SYtems Simulation Program (University of Wisconsin, Madison). <sel.me.wisc.edu/trnsys> (accessed 27 April 2009). U.K. Government, 2007. Government Response to Kate Barker’s Review of Housing Supply: The Supporting Analysis (Office of the Deputy Prime Minister). <www.odpm.gov.uk> (accessed 15 March 2007). Umeda, T., Harada, T., Shiroko, K. A., 1979. Thermodynamic approach to the synthesis of heat integration systems in chemical processes. Computers & Chemical Engineering, 3(1–4), 273–282. UML, 2010. Object Management Group—UML. <www.uml.org> (accessed 18 January 2010). Unitec New Zealand, 2009. <www.unitec.ac.nz> (accessed 13 June 2009). Universität Karlsruhe, 2009. Institute for Industrial Production. <www.iip.kit. edu/65.php> (accessed 18 June 2009). Uryasev, S. (Ed.), 2000. Probabilistic constrained optimization: Methodology and applications. Dordrecht: Kluwer. U.S. Department of Housing and Urban Development, 2009. How to Identify Primary Opportunities to Save Energy—Energy Saving Techniques. <www. hud.gov/offices/cpd/affordablehousing/training/web/energy/techniques/ techniques.cfm> (accessed 5 August 2009). U.S. DOE, 2007. U.S. Department of Energy, Industrial Assessment Centers, Energy Efficiency and Renewable Energy. <www.iac.rutgers.edu/database> (accessed 30 September 2007). Václavek, V., Novotná, A., Dedková, J., 2003. Pressure as a further parameter of composite curves in energy process integration. Applied Thermal Engineering, 23(14), 1785–1795. VALI III User Guide, 2003. Liege, Belgium: Belsim S.A. van der Aalst, W. M. P., ter Hofstede, A. H. M. 2005. YAWL: Yet another workflow language. Information Systems, 30, 245–275. Varbanov, P., 2004. Optimisation and synthesis of process utility systems. PhD thesis, UMIST, Manchester, U.K.
341
342
Bibliography Varbanov, P., Doyle, S., Smith, R., 2004. Modelling and optimization of utility systems. Chemical Engineering Research and Design, 82(A5), 561–578. Varbanov, P., Friedler, F., 2008. P-graph methodology for cost-effective reduction of carbon emissions involving fuel cell combined cycles. Applied Thermal Engineering, 28(16), 2020–2029. Varbanov, P., Klemeš, J., 2008. Analysis and integration of fuel cell combined cycles for development of low-carbon energy technologies. Energy, 33(10), 1508–1517. Varbanov, P., Klemeš, J., 2010. Total sites integrating renewables with extended heat transfer and recovery. Heat Transfer Engineering, 31(9), 733–741. Varbanov, P., Klemeš, J., Shah, R. K., Shihn, H., 2006. Power cycle integration and efficiency increase of molten carbonate fuel cell systems. Journal of Fuel Cell Science and Technology, 3(4), 375–383. Varbanov, P., Klemes, J., Friedler, F., 2007. Critical Analysis of Fuel Cell Combined Cycles for Development of Low-Carbon Energy Technologies. Chemical Engineering Transactions, 12, 739–744. Varbanov, P., Perry, S., Klemeš, J., Smith, R., 2005. Synthesis of industrial utility systems: Cost-effective decarbonisation. Applied Thermal Engineering, 25(7), 985–1001. Varbanov, P., Perry, S., Makwana, Y., Zhu, X. X., Smith R., 2004. Top-level analysis of site utility systems. Chemical Engineering Research and Design, 82(A6), 784–795. Vienna University of Technology, 2009. Invert. <www.invert.at> (accessed 25 April 2009). Wang, Y. P., Smith, R., 1994. Wastewater minimisation. Chemical Engineering Science, 49(7), 981–1006. Wang, Y. P., Smith, R.,1995. Time pinch analysis. Chemical Engineering Research and Design, 73, 905–914. WATER, 2009. Process Integration Software (Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester, U.K.). <www.ceas.manchester.ac.uk/research/centres/ centreforprocessintegration/software/packages/water> (accessed 26 October 2009). Water Footprint Network, 2009. <www.waterfootprint.org> (accessed 16 August 2009). Werther, J., 2007. Gaseous emissions from waste combustion. Journal of Hazardous Materials, 144(3), 604–613. Wikipedia, 2009. List of Chemical Process Simulators. <en.wikipedia.org/wiki/ List_of_chemical_process_simulators> (accessed 26 October 2009). Williams, H. P., 1999. Model building in mathematical programming, 4th ed. Chichester, U.K: Wiley. WORK, 2009. Process Integration Software (Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester, U.K.). <www.ceas.manchester.ac.uk/research/centres/ centreforprocessintegration/software/packages/work> (accessed 26 October 2009). Xiangkun, M., Pingjing, Y., Xing, L., Wilfried, R., 2007. Synthesis of flexible multistream heat exchanger networks based on stream pseudo-temperature with genetic/simulated annealing algorithms. Journal of the Chinese Institute of Chemical Engineers, 38, 321–331. Yamamoto, O., 2000. Solid oxide fuel cells: Fundamental aspects and prospects. Electrochimica Acta, 45, 2423–2435. Yang, H., Li, Y., Shen, J., Hu, S., 2002. Evaluating waste treatment. Recycle and reuse in industrial system: An application of the emergy approach. Ecological Modelling, 160(12), 13–21. Yin, Q., Smith, R., 2008. Incorporating reliability, availability and maintainability (RAM) into process synthesis. AIChE 2008 Spring National Meeting (New Orleans, LA). Yin, Q., Valls, P., Zheng, X., Smith, R., 2009. Minimizing life cycle cost by incorporating reliability, availability and maintainability (RAM) into process
Bibliography design. In H. Alfadala, G. V. R. Reklaitis, and M. M. El-Halwagi (Eds.), Proceedings of the 1st Annual Gas Processing Symposium. Amsterdam: Elsevier. Zamora, J.M., Grossmann, I.E., 1998. A global MINLP optimization algorithm for the synthesis of heat exchanger networks with no stream splits. Computers & Chemical Engineering, 22(3), 367–384. Zecchin, A. C., Simpson, A. R., Maier, H. R., Leonard, M., Roberts, A. J., Berrisford, M. J., 2006. Application of two ant colony optimisation algorithms to water distribution system optimization. Mathematical and Computer Modelling, 44, 451–468. Zhang, R., El-Mashad, H. M., Hartman, K., Wang, F., Liu, G., Choate, C., Gamble, P., 2007. Characterization of food waste as feedstock for anaerobic digestion. Bioresource Technology, 98(4), 929–935. Zhang, X., Sargent, R. W. H., 1996. The optimal operation of mixed production facilities. Part A. General formulation and some solution approaches for the solution. Computers & Chemical Engineering, 20, 897–904. Zhelev, T. K., 1998. Pinch principles for environment protection. In Proceedings of the 8th International Summer School of Chemical Engineering (Sozopol, Bulgaria), pp. 213–216. Sofia: Institute of Chemical Engineering, Bulgarian Academy of Sciences. Zhelev, T. K., 2005a. On the integrated management of industrial resources incorporating finances. Sustainable systems theory: ecological and other aspects. Journal of Cleaner Production, 13, 469–474. Zhelev, T. K., 2005b. On the optimal synthesis of a micro polymerase chain reactor system for DNA analysis. Computer Aided Chemical Engineering, 20(B), 1579– 1584. Zhelev, T. K., 2005c. Water conservation through energy management. Journal of Cleaner Production, 13, 1461–1470. Zhelev, T, 2007. The conceptual design approach—A process integration approach on the move. Resources, Conservation and Recycling, 50, 143–157. Zhelev, T. K., Bhaw, N., 2000. Combined water–oxygen pinch analysis for better wastewater treatment management. Waste Management, 20(8), 665–670. Zhelev, T., Ntlhakana, L., 1999. Energy-environment closed loop through oxygen pinch. Computers & Chemical Engineering, 23(Suppl.), 79–83. Zhelev, T. K., Ridolfi, R., 2006. Energy recovery and environmental concerns addressed through emergy–pinch analysis, Energy, 31, 2486–2498. Zhelev, T. K., Wagenknecht, M., Hartmann, K., Kauschus, W., 1985. Comparative analysis of synthetic methods for the structuring of heat transfer systems. Wissenschaftliche Zeitschrift THLM, 27, 70–89 (in German). Zheng, X., Kim, J.-K., Smith, R., 2008a. Design of heat-integrated power systems with decarbonisation. Computer Aided Chemical Engineering, 25, 1089–1094. Zheng, X., Kim, J.-K., Smith, R., 2008b. Heat integrated power systems with carbon capture. 11th Conference on Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction—PRES 2008 (Prague, 24–28 August). Zhu, X. X., 1997. Automated design method for heat exchanger networks using block decomposition and heuristic rules. Computers & Chemical Engineering, 21(10), 1095–1104. Zhu, X. X., O’Neill, B. K., Roach, J. R., Wood, R. M., 1995. Area-targeting methods for the direct synthesis of heat exchanger networks with unequal film coefficients. Computers & Chemical Engineering, 19(2), 223–239. Zhu, X. X., Zanfir, M., Klemeš, J., 2000. Heat transfer enhancement for heat exchanger network retrofit. Heat Transfer Engineering, 21(2), 7–18.
343
This page intentionally left blank
Index A ABB algorithm. See Accelerated Branch and Bound algorithm Accelerated Branch and Bound algorithm (ABB), 157–158 ADAS, Inside and Solutions, 297 Advanced Process Combinatorics, Inc., 310 AEA Technology, 4 AIChE International Congress on Sustainability Science and Engineering (ICOSSE), 296, 309, 310 AIChE Journal, 310 algorithms: combinatorial, 8 curve-fitting methods search, 30 descent-based, 30 deterministic search solutions, 29–31 Fibonacci method search, 30 golden section method search, 30 gradient-based (local search), 28–29 HENs, 19 line search, 30 LPR models, 29 MSG, 38 Newton method search, 30 NLP models and simplex, 29 nonlinear unconstrained problems and search, 30 optimality conditions and global/ local search, 28–29 P-graphs and, 38, 158 PNS solutions and starting state of MSG, 192 primal, 31 process synthesis and heuristic, 13– 14 PTA, 56–60 SSG, 38
American Process Inc., 301, 305 Analysis of Variance (ANOVA), 43 ant colony optimization stochastic search method, 32 AO Sodrugestvo-T, 300, 303 Applied Thermal Engineering, 301, 307 aqueous streams, 23 arcs, 161 Artie McFerrin Department of Chemical Engineering, Texas A&M, 297, 301, 305 ASPEN, 202–203 AspenTech, 297, 307, 310 audits of energy, 5, 6 availability: optimization, reliability and, 186–190 P-graph solution for energy system, 186 PI and integrating maintainability, reliability and, 144–146 azeotropic distillation systems: maximal/optimal structure for base-case synthesis, 175 optimal process synthesis and, 173–176 P-graph representing structure of, 176 process steps incorporated into base case, 174 residue of curve map, 175
B Balanced Composite Curves (BCCs), 15, 68–69 heat recovery targets, utility placement and, 239 balancing technology, 7 barrier methods, 31 BAT. See best available techniques
345
346
Index batch processes: increase in value of objective function, 160 infeasible solution generated by MILP approach, 160 number produced of each product, 280 optimal solution generated by S-graph, 161 process optimization frameworks and scheduling of, 159–163 S-graphs and, 160–161 BCCs. See Balanced Composite Curves best available techniques (BAT), water and, 107–108 bilinear optimization problem, 27 binary variable, 27 biofuels, 3 biorefineries, green: flowsheet of, 173 solution network for workflow synthesis problem, 172 synthesis of, 171–173 BIS (Department for Business, Innovation and Skills), 297 black box (steady-state model), 23–24 blue water, virtual, 109 bounding function, role in solving MIP, 32 branch and bound approach, for solving discrete problems, 31 budget-income-time pinch analysis: CCs for project budget and income, 132 PI and, 131–133 project budget and income with time, 132
C Canadian Society for Chemical Engineering, 11 CAPE (Computer Aided Process Engineering), 11 capital costs: data extraction with operating and, 290–292 heat transfer area, total cost targeting and, 69–70 carbon dioxide emissions (CO2), 138– 139 costs related to, 2 FC-based systems and, 9 FCs, gas turbines and, 4 carbon footprint (CFP), 20 conclusions and further information, 319–320
carbon footprint (CFP) (Cont.): defined, 2 PI application, 13 and product life-cycle assessment, 2–3 carbon neutrality studies, 2 Carbon Trust, 5 case studies/examples: de-bottlenecking HI crude-oil distillation system, 256–262 energy efficiency in other food/ drink industries, 268–271 energy recovery from FCC unit, 253–256 heat pinch technology, 219–226 heat/power integration in buildings/building complexes, 275–277 industrial applications and, 253–280 industrial utility systems synthesis, 271–275 minimizing water/wastewater in citrus juice plant, 262–268 optimal design of supply chain, 277–279 scheduling large-scale paint production system, 279–280 total sites, 226–233 utility placement, 238–246 water pinch technology, 247–252 CCs. See Composite Curves Center for Engineering and Sustainable Development Research, De La Salle University, 297 Centre for Advanced Process Decision-Making, Carnegie Mellon University, 297 Centre for Process Integration, University of Manchester, 8, 13, 297, 301, 305, 307 Centre for Process Integration and Intensification, European Community Project, 298, 302, 305, 307 Centre for Process Systems Engineering, Imperial College London, 298, 310 Centre for Technology Transfer in the Process Industries, Bucharest, 298 CFP. See carbon footprint Charles Parsons Institute, University of Limerick, 298, 302, 305, 308 CHEMCAD, 205–206 Chemical Engineering, Chemical Equipment Design and Automation (CHISA), 296
Index Chemical Engineering and Processing, 307, 310 Chemical Engineering Science, 307, 310 Chemical Engineering Transactions, 297, 307, 310 Chemical Process Engineering Research Institute (CPERI), 302 CHISA. See Chemical Engineering, Chemical Equipment Design and Automation Chiyoda Corporation, Energy Frontier Business Development Office, 298 CHP. See Combined Heat and Power (CHP) systems CLIMA Congress, 296 CO2. See carbon dioxide emissions cogeneration, 4–5 in advanced gas turbines/FCs, 4 application of PI/HI method to, 12–13 in oil refineries, 4–5 Cold Composite Curve, 15 combinatorial algorithms, 8 combined analysis: conferences, 306 journals, 307 projects, 308–309 service providers, 307–308 Combined Heat and Power (CHP) systems, 4, 7 compact heat exchangers, 4 Composite Curves (CCs): BCCs, 15, 68–69 cold, 15 data precision and, 286 EGCCs and, 270 FCC process with, 254 GCCs and, 60–64, 79, 221, 227, 229, 232, 235, 246, 270 heat recovery and constructing total site, 99 heat recovery for multiple streams and, 51–54 hot, 15, 52 materials reuse-recycle and supply chain, 138 regional resource management, 142 shifted, 63 utility placement and, 245 water minimization, WATER and limiting, 266 water pinch technology and limiting, 249 Computer Aided Chemical Engineering, 307 Computer Aided Process Engineering (CAPE), 11
Computer-Aided Systems Laboratory, Princeton University, 310 Computers & Chemical Engineering, 310 Conceptual Design of Chemical Processes (Douglas), 315 conceptual modeling: data extraction about operating units, 34 network/topology data identification, 34–35 conclusions: better utilization of low-grade heat, 319 books and key articles, 313–315 development trends, 316–320 energy planning, CFP and, 319–320 hybrid energy conversion systems, 317 integration of renewables and waste, 317–319 lecture notes and online teaching resources, 315 maintenance scheduling, maintainability and reliability, 316–317 optimization and optimal design of industrial processes, 315 PI and, 313–315 top-level analysis, 316 Conference on Process Integration, Modeling and Optimisation for Energy Saving and Pollution Reduction, 11 conferences, 11 combined analysis, 306 general information sources, 295–296 HI, 301 mass integration, 304 optimization for sustainable industry, 309–310 continuous variables, 26 convex optimization problem, 27 cooling: demand and data extraction, 50 HI parameter values for heating and, 183 water, 51 COWI A/S, 298 CPERI. See Chemical Process Engineering Research Institute CPLEX code, 35 crude-oil distillation system: changes in heat transfer area due to retrofit, 260 de-bottlenecking heat-integrated, 256–262
347
348
Index crude-oil distillation system (Cont.): flowsheet of, 257 initial HENs for, 259 modified HENs for, 261 pinch analysis results for operating points of preheating phases, 258 curve-fitting methods search algorithms, 30 cutting planes, for solving discrete problems, 31 Czech Society of Chemical Engineering, 11
D Dansk Energi Analyse A/S, 297 data extraction: avoiding pitfalls, 283–292 calculating heat loads and temperatures of extracted streams, 289–290 changes in specific heat capacities, 286–287 conceptual modeling, 34, 40 estimating capital and operating costs, 290–292 example process flowsheet, 49 extracting at effective temperatures and, 289 extracting utilities and, 288 extraction of cold stream and, 288 from flowsheet heat loads, 290 forced and prohibited matches, 289 generating utilities and, 288–289 heat losses and, 288 heating and cooling demands, 50 HI and, 48 integrated placement of processing units and, 234–238 keeping streams separate when necessary, 289 plus-minus principle with soft, 291 precision, 285–286 rules /guidelines, 287–289 soft data in plant/process flowsheet, 290 streams, 284–285, 289–290 data validation technology, 7 decarbonization, 142–143 decision variables, 25 DEFRA (Department for Environment, Food and Rural Affairs), 298 Department of Chemical and Biochemical Engineering, University of Denmark, 302
Department of Chemical and Environmental Engineering, University of Nottingham Malaysia, 298 Department of Chemical and Process Engineering, Rzeszow University of Technology, 302 Department of Chemical Engineering: Auburn University, 297 MIT, 298 University of Maribor, 298 University of Massachusetts, 308 University of Monash, 302, 305 University of Pretoria, 298 Department of Computer Science and Systems Technology, University of Pannonia, 299, 310 Department of Energy, U.S., 300 Department of Energy and Climate Change, U.K., 299 Department of Energy and Environment, Chalmers University, 299, 302 dependent variables (in process optimization), 25 descent-based algorithms, 30 detailed audit, 6 deterministic search algorithm solutions: constrained nonlinear p roblems and, 31 solving continuous LPR problems and, 29 solving continuous NLP problems and, 29–31 solving discrete problems and, 31–32 unconstrained nonlinear problems, 30 diagrams: CCs, hydrogen networks and hydrogn surplus, 124 FCC grid, 255 heat pinch technology and calculated grid, 222 HENs and grid, 83, 84 IT-domain/UML, 24 materials resue-recycle and interval flow rate, 136 MRPD, 116–117 onion, 48 utility placement and preliminary grid, 241 water pinch technology and minimum water flow rate, 250 distillation processes: appropriate placement against GCCs, 79
Index distillation processes (Cont.): distillation column and, 78 extended pinch technology and, 77–80 heat pump placement across utility pinch, 77 refrigeration systems and, 77 dividing-wall distillation technology, 4 Dubrovnik Conference on Sustainable Development of Energy, Water and Environment Systems, 296, 301
E EC. See European Community ECCE. See European Congress of Chemical Engineering Ecological Modelling, 307 EEBPp/U.K. See Energy Efficiency Best Practice Programme EGCC. See exergy grand composite curve emergy-pinch analysis, 126–129, 130–131 EMINENT 2, 214, 303 emissions/effluents: CO2 emissions, 138–139 FCCC system boundary and processing steps, 184 flowsheet and P-graph of FCCC system, 184 materials, streams and, 185 minimizing, 183–186 types of, 23 energy: audits of, 5, 6 low-temperature, 143–144 recovery case study, 253–256 renewable, 2, 3–4, 216–218, 317–319 superstructure for lowtemperature, 143 targets, 51–54 Energy, 301, 307 energy efficiency: basic pinch technology and, 50–69 Carbon Trust recommendations, 5–6 CFP and, 319–320 defined, 3 EGCC for meat-processing plant and, 270 extended pinch technology and, 69–81 food/drink industries and, 268–271 heat recovery pinch and, 54–56
energy efficiency (Cont.): HENs synthesis and, 81–96 importance of optimization, 2 improvement studies, 11 PI basics and, 47–50 PI for improving, 45–104 Pinch Technology and, 3, 14, 50–69 proposed utility system matched to process GCCs, 270 technology examples, 4 total site energy integration and, 96–104 U.K residential data (1971), 8 Energy Efficiency Best Practice Programme (EEBPp/U.K.), 6 Energy Information Administration, USA, 299 Energy Research Group, University of Waikato, 299, 302 Energy Science and Engineering, Indian Institute of Technology, 299, 302 energy systems: flowsheet and hP-graph, 177 optimal retrofit design and, 177–179 optimization and synthesis of, 176–179 simple HI and, 176–177 Environmental Protection Agency (EPA), U.S., 300, 303, 311 EPA. See Environmental Protection Agency equation construction, 35–37 ESCAPE. See European Symposium on Computer Aided Process Engineering EUBIONET 3, 304 Europe Energy Efficiency Conference, 296 European Community (EC), 8 European Congress of Chemical Engineering (ECCE), 296 European Federation of Chemical Engineering Working Party, 11 European Integrated Pollution Prevention and Control Bureau, 299 European Symposium on Computer Aided Process Engineering, 295, 310 European Symposium on Computer Aided Process Engineering (ESCAPE), 11 evaluation tree, of MILP, 31–32 EVECO Brno, 299, 302 exergy grand composite curve (EGCC), 270
349
350
Index
F Faculty of Technology, Lappeenranta University of Technology, 302 FCC. See Fluid Catalytic Cracking unit FCCC. See Fuel-Cell Combined Cycle (FCCC) subsystems FCs. See fuel cells Fibonacci method search algorithm, 30 flow-rate targeting: MRPD and, 116–117 MRPD for pure/impure fresh resource, 117 flowsheeting simulation packages: ASPEN, 202–203 CHEMCAD, 205–206 gPROMS, 204 HYSYS and UniSim Design, 203 PRO/II, 206 software tools, 202–206 flowsheeting simulation tools, 39, 40 optimization technology and, 7 flue gases, 23 Fluid Catalytic Cracking unit (FCC): CCs and, 254 energy recovery case studies/ examples, 253–256 grid diagram of chosen retrofit option, 255 process example, 21 FOB (Free On Board) cost estimate method, 16 food industry, HI application, 13 footprints: carbon, 2–3, 13, 20, 319–320 water, 108–111, 320 Fraunhofer Institute for Environmental, Safety and Energy Technology (UMSICHT), 299 Frontline Systems, 209 fruit juice case study, 111–112, 117–118, 119 fuel cells (FCs): CO2 emissions and, 4, 9 cogeneration in advanced gas turbines and, 4 Fuel-Cell Combined Cycle (FCCC) subsystems, 8–9 emission/effluents and, 184
G GA technique application (stochastic search method), 32 GAMS, 35, 206–207
gas turbines, advanced, 4 gaseous waster streams, 23 GCCs. See Grand Composite Curves general information sources. See information sources, general genetic programming stochastic search method, 32, 33 Global Footprint Network, 299 global optimality, 28–29 global warming, 11 golden section method search algorithm, 30 good housekeeping approach, water/ wastewater, 5 Gover transformation reformulation technique, 42 gPROMS, 204 gradient-based (local search) algorithms, 28–29 Grand Composite Curves (GCCs): appropriate placement of distillation against, 79 construction and multiple utilities targeting, 63 heat pinch technology and, 221 processing units and, 235 proposed utility system matched to, 270 relation between SCCs and, 64 threshold problems and, 60–61 total sites for process A, 227, 229, 232 total sites for process B, 229, 232 utility placement and, 246 utility placements and, 61–62 graph-theoretical approach, 35 gray box (steady-state model), 23–24 green biorefineries, 171–173 green water, virtual, 108 greenhouse gas emission: CFP and, 2 reduction methods, 3–4, 5, 11 grey water, virtual, 109
H heat: better utilization of low-grade, 319 integration and data extraction, 48 steam, 50–51 heat engines: appropriate placement of, 72 configuration, 71 HI of energy-intensive processes and, 71–72 integrating steam turbine above pinch, 72
Index Heat Exchanger Networks (HENs), 12–13 achievable target calculations, 39 algorithms, 19 CCs and, 15 compact, 4 completed design, 89 completing design, 87–92 conventional flowsheet, 82 cost estimation and, 16 data extraction design considerations, 40 design above pinch, 88 design above process pinch and utility placement, 242 design below pinch, 88 design below process pinch and utility placement, 241 design procedure, 83–87 of the FCC process (example), 21 feasible heat exchanger match above pinch, 86 general process flowsheet to represent, 82 grid diagram and implications of pinch, 84 grid diagrams for, 83 heat recovery and, 45–47 Heat Recovery pinch concept and, 14–15 hot/cold streams with, 89 hybrid approach, 95–96 implementing matches with, 47 infeasible heat exchanger match above pinch, 86 key features of resulting networks, 96 mass balances/problems, 36 matches, 46–47 MER for, 39 multistream, 4 pinch design method, 81–93 pinch design principle and, 85 retrofitting of, 16–17, 21 spaghetti-type topology (of heat transfer), 15 splitting above pinch and, 89 splitting and trivial tick-off, 91 splitting below pinch and, 90 splitting to enable CP values for essential matches and, 91 superstructure approach, 93–95 supertargeting of, 15–16, 17 synthesis, 81–96 synthesis methods, 14, 19, 38–40 tick-off heuristic and, 87
Heat Integration (HI), 7–8 building complexes, power and, 275–277 conferences, 301 crude-oil distillation system and de-bottlenecking with, 256– 262 distillation processes and, 77–80 energy-intensive processes and, 71–80 Gantt chart of optimal solution, 183 globally optimal schedule for case study, 181 heat engines and, 71–72 heat pumps and, 72–77 HENs and, 12–13 IEA definition, 12 journals, 301 methodology dissemination, 113 parameter values for heating/ cooling, 183 power systems and, 142–144 process CCs for hospital complex, 276 process stream data of hospital complex, 276 production schedules, 180–183 projects, 303–304 recipes for products, 182 scheduling data for case study, 180 service providers, 301–303 S-graph representing recipe for case study, 181, 182 Heat Integration tools: HEAT-int, 195 HEXTRAN, 199–200 SITE-int, 198 software and, 195–201 spreadsheet-based, 200–201 SPRINT, 195 STAR, 195–198 SuperTarget, 200 WORK, 198–199 heat pinch technology: calculated grid diagram and, 222 calculated MER heat cascade, 221 case studies, 219–226 exchangers violating pinch, 224 existing HENs and, 220 first problem, 219–224 GCCs, 221 load-shirt path identified, 222 minimum allowed temperature difference, 225 optimal HENs and, 226 process stream data, 224
351
352
Index heat pinch technology (Cont.): second problem, 224–226 utility requirements and cost, 225 heat pumps, 7 configuration, 73 HI of energy-intensive processes and, 72–77 placement against heat recovery problem, 74 procedure for sizing, 74 sizing example, 75–76 heat recovery: constructing hot CCs and, 52 constructing total site CCs and, 99 evaluation of multiple streams, 51–54 heat exchange and, 45–47 heat exchange matches and, 46–47 between hot/cold stream, 51 implementing matches with, 47 limits for process, 54 problem and partitioning, 55–56 problem identification and PI basics, 48–50 via steam systems, 99–100 targeting, 100 thermodynamic limits on, 51 variation of targets with, 53 heat recovery pinch: energy efficiency and, 54–56 HENs and, 14–15 limits for process heat recovery set by, 54 trade-off between investment and energy costs with, 54 heat transfer: area, capital cost and total cost targeting, 69–70 enhancement and pressure drop, 146–148 Heat Transfer Engineering, 301 HEAT-int, 195 Hebling Technik, 308 HENS. See Heat Exchanger Networks HEXTRAN, 199–200 HI. See Heat Integration high-temperature FCs (MCFC/SOFC), 4 hill climbing stochastic search method, 32 Honeywell, Canada, 311 hospital complex example, 21 Hot Composite Curve, 15, 52 hot oil utilities, 66 Housing and Urban Development (HUD) Department, U.S., 3
HUD. See Housing and Urban Development Hungarian Chemical Society, 11 hybrids: approach, 95–96 energy conversion systems, 317 hydrogen networks CCs and hydrogen surplus diagram with, 124 PI and, 123–125 HYSYS/UniSim Design, 203
I ICheaP. See Italian Conference on Chemical and Process Engineering ICOSSE. See AlChE International Congress on Sustainability Science and Engineering IEA. See International Energy Agency ILOG ODM, 209–210 Industrial & Engineering Chemistry Research, 310 industrial applications: case studies and, 253–280 de-bottlenecking heat-integrated crude-oil distillation system, 256–262 energy efficiency in food/drink industries, 268–271 energy recovery from FCC unit, 253–256 heat/power integration buildings and building complexes, 275–277 minimizing water/wastewater in citrus juice plant, 262–268 optimal design of supply chain, 277–279 scheduling large-scale paint production system, 279–280 synthesis of industrial utility systems, 271–275 information hiding, 39 information sources, general: combined analysis, 306–309 conferences, 11, 295–296, 301, 304, 306, 309–310 further reading and, 295–311 general, 295–301 HI, 301–304 journals, 297, 302, 304–305, 307, 310 mass integration, 304–306 optimization for sustainable industry, 309–311
Index information sources, general (Cont.): projects, 301, 303–304, 306, 308–309, 311 service providers, 297–301, 302–303, 305–306, 307–308, 310–311 Institute for Scientific Research, University of Guanajuato, 299 Institute of Process and Environmental Engineering, Brno University, 299, 302 integer programming (IP) model, 25, 27 integer variables (in process optimization), 26, 27 International Energy Agency (IEA), 300 energy monitoring techniques and, 3 Heat Energy definition of, 12 IP. See integer programming Italian Association of Chemical Engineering, 11 Italian Conference on Chemical and Process Engineering (ICheaP), 296, 301, 306
J Journal of Clean Technologies and Environmental Policy, 304, 307 Journal of Cleaner Production, 297, 305, 307 Journal of the Chinese Institute of Chemical Engineers, 310 Journal of Universal Computer Science, 310 journals: combined analysis, 307 general information sources, 297 HI, 301 mass integration, 304–305 optimization for sustainable industry, 310
K KBC Energy Services, 300, 303, 305 key performance indicators (KPIs), 7 KKT (Karush-Kuhn-Tucker) optimality conditions, 28
L Laboratory for Analysis and Synthesis of Chemical Systems, Institut de Chimie-Bâtiment B6, 303
Laboratory for Industrial Energy Systems, Ecole Polytechnique Federale de Lausanne, 300, 303, 311 Lagrange multipliers, 28 laws, second thermodynamic, 51 legislation, water, 106–107 Lehrstuhl für Technische Chemie A, 300 life-cycle assessment, 2–3, 20 LINDO, 208–209 line search algorithm, 30 linear optimization problem, 27 linear programming (LPR) models, 25 absence of integer variables and, 27 branch and bound method solution, 31–32 convex, 28 deterministic algorithm solutions, 29 equation construction, 35, 36 Simplex Method (algorithm) solution, 29 Linnhoff, Bodo, 3, 13, 283, 294, 313–314 local optimality, 28 Loss Prevention and Safety Promotion in the Process Industries, 304 LPR. See linear programming models
M maintainability: maintenance scheduling, reliability and, 316–317 PI, integrating reliability, availability and, 144–146 mass balances, 36–37 mass integration: conferences, 304 design for low-temperature energy systems, 144 flow-rate targeting with MRPD, 116–117 general information sources, 304–306 journals, 304–305 MRPD applied to fruit juice case study, 117–118 new process design methodology, 145 projects, 306 service providers, 305–306 software, 201–202 summary, 122 water, minimizing use and maximizing reuse, 106–113
353
354
Index mass integration (Cont.): water integration and, 105–106 water minimization via mathematical optimization, 118–122 water pinch analysis introduction, 113–116 material recovery pinch diagram (MRPD): flow-rate targeting and, 116–117 fruit juice production, limiting water data and, 117 MRPD for pure/impure fresh resource, 117 materials reuse-recycle: construction of interval flow rate diagram, 136 material balance in aggregate planning, 137 property pinch analysis and, 133–136 supply chain CCs and, 138 mathematical modeling suites: MATLAB, 210–211 MATLAB alternatives, 211 software tools, 210–211 mathematical optimization fruit juice production, MRPD and network design, 119 illustrative example of brewery plant, 120–122 introduction to, 118–119 water minimization via, 118–122 Mathematical Programming (MPR) approach, 8, 14 model building/optimization and, 24 model nonlinearity and, 41–42 Network Pinch and, 17 network-related information transformation, 35 PI and, 14 process optimization frameworks and, 151–153 process synthesis major steps and, 153 superstructure formulation, 35–36, 38 tools for designing chemical processes, 20 MATLAB, 24, 29, 210–211 Maximal Energy Recovery (MER), 39 Maximal Structure Generation (MSG) algorithm, 38, 157–158, 192 MCFCs. See molten-carbonate fuel cells
Mechanical Engineering Faculty, Paderborn University, 300 MER. See Maximal Energy Recovery MILP. See mixed integer linear program MINLP. See mixed integer nonlinear program MIPSYN, 207–208 mistakes. See pitfalls mixed integer linear program (MILP), 25, 29, 160. See also successive mixed integer linear program convexity case-by-case evaluation, 28 described, 27 deterministic methods for solving, 31–32 global optimality and, 29 mass balances/problems, 36 nodes/evaluation tree of, 31–32 mixed integer nonlinear program (MINLP), 25, 27 deterministic methods for solving, 31–32 mass balances/problems, 36 model building and optimization, 24–25 components, 24–25 examples, 24 IT-domain diagrams/UML diagrams, 24 model creation, 33–43 conceptual modeling, 34–35 equation construction, 35–37 evaluating adequacy and precision, 42–43 graph-theoretical approach, 35 handling model nonlinearity, 41 handling process complexity, 38–40 mass balances, 36–37 objective function choice, 37–38 object-oriented modeling, 38–39 P-graph framework, 35 process insight application, 40–41 Modelica modeling language, 24, 211–212 molten-carbonate fuel cells (MCFCs), 4 MPR. See Mathematical Programming (MPR) approach MRPD. See material recovery pinch diagram MSG algorithm. See Maximal Structure Generation algorithm multiple utilities targeting: construction of GCCs and, 63
Index multiple utilities targeting (Cont.): pinch technology and, 61–69 utility placement options, 63–69 utility placements, GCCs and, 61–62 multistream heat exchangers, 4
N network evolution: loops, paths and, 93 splitting and advanced tick-off with, 92 splitting procedure above pinch, 92 network optimization, 8 Network Pinch method, 16–17, 21 network/topology data identification, 34–35 Newton method search algorithm, 30 NLP. See nonlinear programming (NLP) models nodes, of MILP, 31–32 nonisothermal stream mixing, 287 nonisothermal utilities, 65–66 Nonlinear and Mixed-Integer Optimization: Fundamentals and Applications (Floudas), 315 nonlinear programming (NLP) models: absence of integer variables and, 27 convexity case-by-case evaluation, 28 equation construction, 35 mass balances/problems, 36 penalty methods for solving constrained, 31 simplex algorithm for problem-solving, 29 numerical targeting, 56–60
O objective function, 25 object-oriented modeling, 38–39 oil crises (1973–1974, 1979), 14 onion diagram, 48 open-source DWSIM chemical process simulator, 24 operating costs, 290–292 optimal process synthesis: azeotropic distillation systems and, 173–176 classification of optimization methods, 167 heterogeneous flowsheets and, 169– 171 optimization and, 167–176 reaction network synthesis and, 167–169
optimal process synthesis (Cont.): steps for composition of maximal structure, 169 synthesis of green biorefineries and, 171–173 optimal workflow structures: P-graph, 171 plausible activities for, 170 synthesis of heterogeneous flowsheets and, 169–171 optimality conditions: global search algorithms, 28–29 local search algorithms, 28 optimality problems, stochastic search methods for solving, 32–33 optimization: availability, reliability and, 186–190 combined PI and, 165–190 emissions/effluents minimized and, 183–186 methods with optimal process synthesis, 167 network, 8 optimal design of industrial processes and, 315 optimal process synthesis and, 167–176 optimal scheduling for throughput, profit, security and, 179–183 optimal synthesis of energy systems and, 176–179 process synthesis and role of, 165–166 summary, 190 tools for efficient implementation of PI, 166–167 water minimization via mathematical, 118–122 optimization, sustainable industry: conferences, 309–310 journals, 310 projects, 311 service providers, 310–311 optimization packages, general-purpose: Frontline Systems, 209 GAMS, 35, 206–207 ILOG ODM, 209–210 LINDO, 208–209 MIPSYN, 207–208 software tools, 206–210 optimization technology: application, 9 flowsheeting simulation tools, 7 HENs retrofit, 16–17
355
356
Index optimization technology (Cont.): PI integration attempts, 13 sustainable industry and, 309–311 OSL code, 35 oxygen pinch analysis, 125–126 oxygen-water pinch analysis, 128–129
P paint production system, scheduling, 279–280 parity plots, 43 Parliamentary Office for Science and Technology, U.K., 2 penalty methods, for solving constrained NLP, 31 PEPNET, 304 performance targets, 48 P-graphs, 8–9 algorithms and inputs/outputs from, 158 availability and solution for energy system, 186 azeotropic distillation systems structure and, 176 emissions/effluents flowsheet and, 184 mathematical engine, 157–158 model building/optimization and, 24, 35 MSG/SSG algorithms, 38 optimal retrofit design, maximal structure and, 179 optimal workflow structures, 171 process representation via, 154–155 process structure of operating units and, 155 process structure violating axioms and, 157 reduction in search space, combinatorial axioms and, 157 significance for structural process optimization, 155–157 structural process optimization and, 153–158 symbols of process elements and, 154 PI. See Process Integration (PI) Pinch Analysis: benefits of using, 40–41 data extraction and, 283–284 for HI, 41 PI example, 21 pinch design method: completing designs and, 87–92 design procedure and, 83–87
pinch design method (Cont.): HENs representation and, 81–83 network evolution and, 92–93 Pinch Express, 283 Pinch Technology: energy efficiency and basic, 3, 14, 50–69 energy efficiency and extended, 69–81 heat recovery pinch and, 54–56 heat transfer area, capital cost, total cost targeting and, 69–70 HI of energy-intensive processes and, 71–80 multiple utilities targeting and, 61–69 numerical targeting, PTA and, 56–60 process modifications and, 80–81 setting energy targets, 51–54 threshold problems and, 60–61 Pinch Technology (Linnhoff and Vredeveld), 3 pitfalls, 281–282 data extraction, 283–292 integration of renewables, fluctuating demand and supply, 292 interpreting results, 293 making results happen, 293–294 steady-state and dynamic performance, 292–293 PNS solutions: graph-based process optimization tools and, 191–193 starting state of MSG algorithm, 192 power: cogeneration, 101–102 integration, building complexes and HI, 275–277 water, 3 wind, 3 PRES. See Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction pressure drop, 146–148 primal algorithms, 31 Problem Table Algorithm (PTA): heat cascade for process data, 59 numerical targeting and, 56–60 process streams data, 58 temperature shifting to ensure feasible heat transfer and, 57
Index process design: onion diagram and, 48 PI, energy efficiency and, 47–48 Process Design Center, Netherlands, 303 Process Integration, Modelling and Optimisation for Energy Saving and Pollution Reduction (PRES), 295, 301, 306 Process Integration (PI): application of, 7, 9 audit procedure, 6 avoiding pitfalls, 281–294 benefits of, 18–20 data extraction, steps of, 34 defined/described, 12 examples, 20–22 further applications of, 123–150 HI component, 7–8 history and development of, 12–14 improving energy efficiency and, 45–104 MPR approach to, 14 needs for, 11–12 Network Pinch, 16–17 optimization and, 165–190 optimization integration attempts, 13–14 role in industrial sustainability, 20 targeting heat recovery and, 14–15 thermodynamic roots of, 14–15 Process Integration (PI), energy efficiency and: basic pinch technology, 50–69 basics, 47–50 extended pinch technology, 69–81 heat recovery problem identification and, 48–50 HENs synthesis, 81–96 HI and, 47 hierarchy of process design and, 47–48 introduction to heat exchange and recovery, 45–47 performance targets and, 48 total site energy integration, 96–104 Process Integration (PI), further applications of: budget-income-time pinch analysis, 131–133 combined analysis II, 131–142 decarbonization, 142–143 design and management of hydrogen networks, 123–125 energy-water, oxygen-water, pinch-emergy, 126–131
Process Integration (PI), further applications of (Cont.): heat-integrated power systems and, 142–144 integrating reliability, availability and maintainability, 144–146 local integrated energy sectors and extended total sites, 148–149 low-temperature energy, 143–144 materials reuse-recycle and property pinch analysis, 133–136 oxygen pinch analysis, 125–126 pinch analysis of supply chains, 136–138 pinch used to target CO2 emissions, 138–139 pressure drop and heat transfer enhancement in, 146–148 regional resource management, 139–142 summary, 149–150 Process Integration Ltd., 300, 303, 305, 308, 311 process modifications: hot and cold streams with, 81 plus-minus principle and, 80 process network model construction, 36–37 process optimization, 23–43 classes of problems, 26–28 conditions for optimality, 28–29 continuous/integer variables and, 26 creating models, 33–43 decision/dependent variables and, 25 defined, 25 deterministic algorithm solutions, 29–31 deterministic methods for solving discrete problems, 31–32 generic optimization problem, 26t integer/binary variables and, 27 mathematical problem formulation, 25–26 model building and, 24–25 objective function component, 25 stochastic search methods, 32–33 process optimization frameworks: mathematical programming, classical approach and, 151–153 scheduling of batch processes and, 159–163 structural process optimization, P-graphs, 153–158
357
358
Index process optimization tools, graph-based: PNS solutions, 191–193 S-Graph Studio, 193–195 software tools, 191–195 Process Pinch, 21, 22 process synthesis: heuristic/algorithmic methods of, 13–14 major steps and mathematical programming, 153 optimization’s role in, 165–166 P-graph synthesis, 8–9 process system engineering, 11 A Process Integration Primer (Gunderson), 295 processing units: CCs for process in problem 5, 234 data extraction and integrated placement of, 234–238 GCCs for process in problem 5, 235 process flowsheet for base-case design and, 234 stream data for pinch analysis and, 237 PRO/II, 206 projects: combined analysis, 308–309 general information sources, 301 HI, 303–304 mass integration, 306 optimization for sustainable industry, 311 property pinch analysis, materials reuse-recycle and, 133–136 PTA. See Problem Table Algorithm
R recipes, 161 HI, 181–182 S-graph example, 162 S-graph representing case study, 181 reconciliation technology, 7 refrigeration, 7 levels, 67–68 systems and distillation processes, 77 utility systems and, 271 WORK and composition options/ ideal profiles with, 199 reliability: maintenance scheduling, maintainability and, 316–317 optimization, availability and, 186–190
reliability (Cont.): PI and integrating availability, maintainability and, 144–146 renewable energy: carbon neutrality and, 2 energy system sustainability from, 3–4 energy systems and integrating, 216–218 integration of waste and, 317–319 software for analyzing integration of, 216 Renewable Energy, 301, 307 residual plots, 43 resource management, regional: CCs and, 139–141, 142 flowchart of algorithm for, 140 PI and, 139–142 regional energy supply-deficit curves and, 141 surplus-deficit curves, 141 Resources, Conservation & Recycling, 297, 305, 307 retrofit design, optimal: characteristics, operating unit costs and, 178 energy systems and, 177–179 flowsheet for considering absorption, 179 P-graph of maximal structure, 179 revenue, maximizing, 179–180
S SA. See simulated annealing SCCs. See Shifted Composite Curves scheduling: batch processes and process optimization frameworks, 159–163 frameworks, suitability, limitations and S-graphs, 159–161 graph of solution minimizing makespan, 280 large-scale paint production system, 279–280 number of batches produced of each product, 280 optimization, throughput, security and optimal, 179–183 reliability, maintainability and maintenance, 316–317 S-graphs and framework for, 161–163 School of Chemical and Environmental Engineering, University of Nottingham Malaysia, 305, 308
Index School of Chemical Engineering, Purdue University, 311 Scilab computational package, 24 Scottish Environment Protection Agency (SEPA), 300 screening and scoping audits, 5–6 search algorithms, for nonlinear unconstrained problems, 30 second law of thermodynamics, 14–15 segmentation, 286–287 SEMPRA ENERGY, 3 SEPA. See Scottish Environment Protection Agency service providers: combined analysis, 307–308 general information sources, 297–300 HI, 301–303 mass integration, 305–306 optimization for sustainable industry, 311 S-Graph Studio, 193–194 S-graphs: batch processes and, 159–163 example recipe, 162 framework for scheduling, 161–163 HI and recipe for case study with, 181, 182 recipe for two batches, 162 scheduling frameworks, suitability and limitations, 159–161 solution for recipe under NIS policy, 162 Shifted Composite Curves (SCCs), 63 relation between GCCs and, 64 simple heat integration, 176–177 simplex method algorithm for LPR problems, 29 simulated annealing (SA) stochastic search method, 32, 33 simulation tools, 39 for process industry, 24 SITE-int, 198 SMILP. See successive mixed integer linear program SOFCs, 4 software tools: balancing/flowsheeting simulation for energy-saving analysis, 215 emerging trends, 212–215 flowsheeting simulation packages, 202–206 general-purpose optimization packages, 206–210 graph-based process optimization tools, 191–195
software tools (Cont.): HI tools, 195–201 integrating renewable energy into other energy systems, 216–218 mathematical modeling suites, 210–211 Modelica, 24, 211–212 other, 211–218 overview of available, 191 WATER, mass integration software, 201–202, 266 Solution Structures Generation algorithm (SSG), 38, 157–158 specifications, 25 spreadsheet-based tools, 200, 201 SPRINT: HI tools and, 195, 286 software interface, 196 SSG algorithm. See Solution Structures Generation algorithm STAR, 286 graphical user interface, 196 HI tools and, 195–198 steady-state models, 23–24 steam: heating with, 50–51 system and heat recovery, 99–100 stochastic methods: error minimization tools and, 43 optimization problems and, 32–33 streams: aqueous, 23 CCs and heat recovery for multiple, 51–54 data extraction and, 284–285, 289–290 emissions/effluents, materials and, 185 heat exchangers and multi, 4 heat recovery between hot and cold, 51 structural process optimization: mathematical programming and, 153–158 MSG, SSG, ABB and, 157–158 P-graph symbols of process elements and, 154 P-graphs and process structure of operating units, 155 P-graphs and process structures violating axioms, 157 P-graph’s mathematical engine and, 157–158 P-graph’s significance for, 155–157 process representation via P-graphs, 154–155 reduction in search space, combinatorial axioms and, 157
359
360
Index successive mixed integer linear program (SMILP), 42 sunflower oil production, 21 superstructure approach: HENs synthesis and, 93–95 loop and path in heat exchanger network, 93 spaghetti fragment and, 94 SuperTarget, 200, 283 supertargeting, 15–16, 17, 70 supply chain: activities in optimal, second/thirdbest business processes, 279 case study, optimal design of, 277–279 CCs and materials reuse-recycle, 138 cost of activities in case study, 278 potential activities in case study, 277 Systematic Methods of Chemical Process Design (Biegler et al.), 315
T Tabu search, 32 TAPPI EPE (Engineering, Pulping and Environmental Conference), 296 targeting procedures, 39 targets: energy, 51–54 variation of heat recovery, 53 Thermal Energy, Norwegian University of Science and Technology, 300, 303 thermodynamics: limits on heat recovery, 51 roots, 14–15 second law, 51 threshold: HENs design cases, 60 low and high, 61 problems and GCCs, 60–61 throughput: maximizing, 179–180 optimal scheduling for profit, security and, 179–183 tools: complexity management, 39 flowsheeting simulation packages, 202–206 HI, 195–201 mathematical modeling suites, 210–211 MPR and designing chemical processes, 20 optimization and efficient implementation of PI, 166–167
tools (Cont.): optimization packages, 206–210 PI efficient implementation, 166–167 PNS solutions and graph-based process optimization, 191–193 PNS solutions and S-Graph Studio, 193–195 process industry and simulation, 24, 39 software, 191–218 spreadsheet-based, 200, 201 stochastic methods and error minimization, 43 topology data identification, 34–35 Total Food, 296 total sites: case studies, 226–233 combined site heat sources for problem 3(c), 230 data extraction, 107 data for process A, 227 data for process B, 227 energy sector with heat and power, 148 first problem, 226–230 GCCs for process A, 227, 232 GCCs for process B, 229, 232 heat recovery targets for problem 4(a), 232 heat source and sink segments from GCCs for Process A, 229 heat source and sink segments from GCCs for Process B, 229 local integrated energy sectors and extended, 148–149 marginal steam prices for utility system with, 103 optimized utility system and, 103 for problem 4(b), 233 for problem 4(c), 233 problem table for process A, 228 problem table for process B, 228 profiles, 97–99 profiles for problem 3(c), 230 second problem, 231–233 stream data for process A, 231 stream data for process B, 231 total sites energy integration, 96 advanced total site optimization and analysis, 102–104 data extraction, 97 heat recovery pockets, construction of total sites profiles and, 98 heat recovery via steam system, 99–100 power cogeneration, 101–102 profiles, 97–99 schematic of industrial total site, 97
Index trends: conclusions, further information and development, 316–320 software tools and emerging, 212–215 tri-generation, 5
U UK Energy Research Centre, 300 UMSICHT. See Fraunhofer Institute for Environmental, Safety and Energy Technology Uni-Sim Design/HYSYS, 203 UniSim process modeling (Honeywell), 24 United Kingdom: Department of Energy and Climate Change, 299 Energy Efficiency Best Practice Programme, 6 Energy Research Centre, 300 heat exchange case study, 4 Parliamentary Office for Science and Technology, 2 residential energy consumption, 8 universities. See service providers; specific departments University College London Energy Institute, 300 A User Guide on Process Integration for the Efficient Use of Energy (Linnhoff et al.), 313 utility pinch, 64, 77 utility placement: appropriate, 246 available utilities and, 239 BCCs and heat recovery targets, 239 case studies, 238–246 CCs, 245 constraints for placing hot oil utilities and, 66 constructing GCCs for streams and, 63 exploiting GCCs pocket for utility substitution with, 68 first problem, 238–242 GCCs and, 246 generating steam below pinch, 67 heat cascade intervals and stream population, 244 HENs design above process pinch and, 242 HENs design below process pinch and, 241 options and multiple utilities targeting, 63–69
utility placement (Cont.): placing refrigeration levels for pure refrigerants, 68 preliminary grid diagram and, 241 problem table, 245 process streams and, 239, 243 process streams with missing data filled in, 240 properties of nonisothermal hot utilities and, 65 relation between GCCs and SCCs, 64 second problem, 243–246 using GCCs to target for single/ multiple steam levels and, 64 utilities available and, 243 utility systems: case study, synthesis of, 271–275 configuration data for total site, 272 operating scenarios for total site, 272 optimal, 274, 275 sensitivity analysis for reducing emissions, 274 superstructure of industrial, 273 temperature changed with refrigeration levels, 271 total site profiles and candidate steam pressure levels, 272
V variables: continuous/integer, 26 decision/dependent, 25 LPR and absence of integer, 27 vertical heat transfer, 15 Vredeveld, D. R., 3 VTT Technical Research Centre of Finland, 303, 306
W walk-through audit, 6 Warsaw University of Technology, 300, 303 waste, integration of renewables and, 317–319 Waste Management, 307 wastewater, 23 minimizing water usage and, 111–113 optimization of use approach, 5–6 overview of measures, 111–112 treatment, 112–113 WATER, 201–202, 266
361
362
Index water: BAT, 107–108 citrus juice plant, minimizing wastewater and, 262–268 common usage operations, 109 cooling, 51 footprint, 108–111, 320 integration, 105–106 legislation, 106–107 minimizing usage and waste, 111–113 minimizing use and maximizing reuse, 106–113 optimization of use approach, 5–6 pinch analysis introduction, 113–116 power, 3 quality, 128 Water Footprint Network, 306 water footprints, 320 virtual blue water, 109 virtual green water, 108 virtual grey water, 109 virtual water consumed while processing food industry products, 110 water minimization and, 111–112 water minimization: existing water network and, 264 fruit juice production, MRPD, network design and, 119 limiting CCs generated by WATER and, 266 via mathematical optimization, 118–122 summary of four design options, 267 superstructure for water/ regeneration reuse in brewery plant, 121 wastewater in citrus juice plant and, 262–268 water footprints and, 111–112 water network after pinch analysis and, 265 water reuse opportunities in brewery plant, 121
water networks: achievable target calculations, 39 mass exchange and, 17–18 water pinch analysis: limiting CCs and, 115 mass integration and, 113–116 problem data for water-using operations, 114 treatment system design and, 115 water pinch technology: case studies, 247–252 combined limiting water profiles, 248 connecting operations directly, 252 data for water-processing operation, 247 design grid for water system and, 251 design strategy, 250 diagram used to target minimum water flow rate, 250 first problem, 247–248 flowsheet representation of water system design, 252 limiting CCs and, 249 limiting water profiles for, 247 minimum wastewater targeting and, 248 removing water mains, connecting sources and sinks, 252 second problem, 249–252 streams connected with water mains and, 251 whiskey distillery, 22 wind power, 3 WORK: HI tools and, 198–199 refrigeration composition options and ideal profiles, 199 World Bioenergy, 296 World Congress of Chemical Engineering, 296 World Renewable Energy Congress: Innovation in Europe, 296 World Sustainable Energy Days, 296