This page intentionally left blank
Carbon Dioxide Sequestration and Related Technologies
Scrivener Publishing 3 Winter Street, Suite 3 Salem, MA 01970 Scrivener Publishing Collections Editors James E. R. Couper Richard Erdlac Pradip Khaladkar Norman Lieberman W. Kent Muhlbauer S. A. Sherif
Ken Dragoon Rafiq Islam Vitthal Kulkarni Peter Martin Andrew Y. C. Nee James G. Speight
Publishers at Scrivener Martin Scrivener (
[email protected]) Phillip Carmical (
[email protected])
Carbon Dioxide Sequestration and Related Technologies Edited by
Ying (Alice) Wu Sphere Technology Connection
John J. Carroll Gas Liquids Engineering, Ltd. and
Zhimin Du Southwest Petroleum University
Copyright © 2011 by Scrivener Publishing LLC. All rights reserved. Co-published by John Wiley & Sons, Inc. Hoboken, New Jersey, and Scrivener Publishing LLC, Salem, Massachusetts. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., Ill River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. For more information about Scrivener products please visit www.scrivenerpublishing.com. Cover design by Kris Hackerott. Library of Congress Cataloging-in-Publication ISBN 978-0-470-93876-8
Printed in the United States of America 10 9 8 7 6 5 4 3 2 1
Data:
Contents Introduction The Three Sisters - CCS, AGI, and EOR Ying Wu, John J. Carroll and Zhimin Du
xix
S e c t i o n 1: D a t a a n d C o r r e l a t i o n 1.
2.
3.
Prediction of Acid Gas Dew Points in the Presence of Water and Volatile Organic Compounds Ray. A. Tomcej 1.1 Introduction 1.2 Previous Studies 1.3 Thermodynamic Model 1.4 Calculation Results 1.5 Discussion References Phase Behavior of China Reservoir Oil at Different C 0 2 Injected Concentrations Fengguang Li, Xin Yang, Changyu Sun, and Guangjin Chen 2.1 Introduction 2.2 Preparation of Reservoir Fluid 2.3 PVT Phase Behavior for the C 0 2 Injected Crude Oil 2.4 Viscosity of the C 0 2 Injected Crude Oil 2.5 Interfacial Tension for C 0 2 Injected Crude Oil/Strata Water 2.6 Conclusions Literature Cited Viscosity and Density Measurements for Sour Gas Fluids at High Temperatures and Pressures B.R. Giri, P. Biais and R.A. Marriott 3.1 Introduction 3.2 Experimental
3 3 4 5 6 10 11
13
14 14 15 17 19 20 21
23 24 25 v
vi
CONTENTS
3.2.1 Density Measurement 3.2.2 Viscosity Measurement 3.2.3 Charging and Temperature Control 3.3 Results 3.4 Conclusions References 4.
5.
6.
Acid Gas Viscosity Modeling with the Expanded Fluid Viscosity Correlation H. Motahhari, M.A. Satyro, H.W. Yarranton 4.1 Introduction 4.2 Expanded Fluid Viscosity Correlation 4.2.1 Mixing Rules 4.2.2 Modification for Non-Hydrocarbons 4.3 Results and Discussion 4.3.1 Pure Components 4.3.2 Acid Gas Mixtures 4.4 Conclusions 4.5 Acknowledgements References Evaluation and Improvement of Sour Property Packages in Unisim Design Jianyong Yang, Ensheng Zhao, Laurie Wang, and Sanjoy Saha 5.1 Introduction 5.2 Model Description 5.3 Phase Equilibrium Calculation 5.4 Conclusions 5.5 Future Work Reference Compressibility Factor of High C0 2 -Content Natural Gases: Measurement and Correlation Xiaoqiang Bian, Zhimin Du, Yong Tang, and Jianfen Du 6.1 Introduction 6.2 Experiment 6.2.1 Measured Principles 6.2.2 Experimental Apparatus and Procedure 6.2.3 Experimental Results
25 27 30 31 37 37
41 41 42 44 45 47 47 48 52 52 52
55
55 56 58 62 62 63 65
65 67 67 67 68
CONTENTS
6.3
Methods 6.3.1 Existing Methods 6.3.2 Proposed Method 6.5 Comparison of the Proposed Method and Other Methods 6.6 Conclusions 6.7 Acknowledgements 6.8 Nomenclature References
68 68 74 78 83 84 84 85
Section 2: Process Engineering 7.
8.
Analysis of Acid Gas Injection Variables Edward Wiehert and James van der Lee 7.1 Introduction 7.2 Discussion 7.3 Program Design 7.4 Results 7.5 Discussion of Results 7.5.1 General Comments 7.5.2 Overall Heat Transfer Coefficient, U 7.5.3 Viscosity 7.6 Conclusion References Glycol Dehydration as a Mass Transfer Rate Process Nathan A. Hatcher, Jaime L. Nava and Ralph H. Weiland 8.1 Phase Equilibrium 8.2 Process Simulation 8.3 Dehydration Column Performance 8.4 Stahl Columns and Stripping Gas 8.5 Interesting Observations from a Mass Transfer Rate Model 8.6 Factors That Affect Dehydration of Sweet Gases 8.7 Dehydration of Acid Gases 8.8 Conclusions Literature Cited
89 89 90 93 94 96 96 101 104 105 105
107
108 110 111 114 115 118 119 119
CONTENTS
Carbon Capture Using Amine-Based Technology Ben Spooner and David Engel 9.1 Amine Applications 9.2 Amine Technology 9.3 Reaction Chemistry 9.3.1 Nucleophilic Pathway 9.3.2 Acid-Base Pathway (Primary Secondary and Tertiary Amines) 9.4 Types of Amine 9.5 Challenges of Carbon Capture 9.5.1 Prevention 9.5.2 Reclaimers 9.5.3 Purging and Replacing Amine 9.5.4 High Energy Consumption 9.5.5 Size of the Amine Facility 9.5.6 Captured C 0 2 9.6 Conclusion Dehydration-through-Compression (DTC): Is It Adequate? A Tale of Three Gases Wes H. Wright 10.1 Background 10.2 Water Saturation 10.3 Is It Adequate? 10.4 The Gases 10.5 Results 10.6 Discussion References Diaphragm Pumps Improve Efficiency of Compressing Acid Gas and C 0 2 Josef Jarosch, Anke-Dorothee Braun 11.1 Diaphragm Pumps 11.2 Acid Gas Compression 11.3 C 0 2 Compression for Sequestration 11.4 Conclusion Literature
121 121 122 124 124 125 126 128 128 129 129 129 130 130 131
133 133 138 138 141 147 151 152
155 162 164 167 171 172
CONTENTS
ix
Section 3: Reservoir Engineering 12. Acid Gas Injection in the Permian and San Juan Basins: Recent Case Studies from New Mexico David T. Lescinsky; Alberto A. Gutierrez, RG; James C. Hunter, RG; Julie W. Gutierrez; and Russell E. Bentley 12.1 Background 12.2 AGI Project Planning and Implementation 12.2.1 Project Planning and Feasibility Study 12.2.2 Reservoir/Cap Rock Identification and Regulatory Permitting 12.2.3 Well Drilling and Testing 12.2.4 Well Completion and Construction 12.2.5 Reservoir and Seal Evaluation 12.2.6 Documentation, System Start-up and Reporting 12.3 AGI Projects in New Mexico 12.3.1 Permian Basin 12.3.1.1 LinamAGI#l 12.3.1.2 Jal 3 AGI #1 12.3.2 San Juan Basin 12.3.2.1 Pathfinder AGI #1 12.4 AGI and the Potential for Carbon Credits 12.5 Conclusions References 13. C 0 2 and Acid Gas Storage in Geological Formations as Gas Hydrate Farhad Qanbari, Olga Ye Zatsepina, S. Hamed Tabatabaie, Mehran Pooladi-Darvish 13.1 Introduction 13.2 Geological Settings 13.2.1 Depleted Gas Reservoirs 13.2.1.1 Mixed Hydrate Phase Equilibrium 13.2.1.2 Assumptions
175
175 178 178 181 183 186 186 188 190 190 193 196 199 200 204 207 208
209
210 211 211 211 213
x
CONTENTS
13.2.2
Ocean Sediments 13.2.2.1 Negative Buoyancy Zone (NBZ) 13.2.2.2 Hydrate Formation Zone (HFZ) 13.3 Model Parameters 13.3.1 Depleted Gas Reservoir 13.3.2 Ocean Sediment 13.4 Results 13.4.1 Depleted Gas Reservoir 13.4.2 Ocean Sediment 13.5 Discussion 13.6 Conclusions 13.7 Acknowledgment References 14. Complex Flow Mathematical Model of Gas Pool with Sulfur Deposition W. Zhu, Y. Long, Q. Liu, Y. Ju, and X. Huang 14.1 Introduction 14.2 The Mathematical Model of Multiphase Complex Flow 14.2.1 Basic Supposition 14.2.2 The Mathematical Model of Gas-liquid-solid Complex Flow in Porous Media 14.2.2.1 Flow Differential Equations 14.2.2.2 Unstable Differential Equations of Gas-liquid-solid Complex Flow 14.2.2.3 Relationship between Saturation and Pressure of Liquid Phase 14.2.2.4 Auxiliary Equations 14.2.2.5 Definite Conditions 14.3 Mathematical Models of Flow Mechanisms 14.3.1 Mathematical Model of Sulfur Deposition 14.3.2 Thermodynamics Model of Three-phase Equilibrium 14.3.3 State Equations
213 213 214 216 216 217 218 218 221 221 223 224 224
227 227 228 228
229 229
230 231 232 232 232 232 234 236
CONTENTS
14.3.4 14.3.5
Solubility Calculation Model Influence Mathematical Model of Sulfur Deposition Migration to Reservoir Characteristics 14.4 Solution of the Mathematical Model Equations 14.4.1 Definite Output Solutions 14.4.2 Productivity Equation 14.5 Example 14.5.1 Simulation Parameter Selection 14.5.2 Oil-gas Flow Characteristics near Borehole Zones of Gas-well 14.5.3 Productivity Calculation 14.6 Conclusions 14.7 Acknowledgement References
xi
236
237 238 238 239 240 240 240 240 242 242 242
Section 4: Enhanced Oil Recovery (EOR) 15. Enhanced Oil Recovery Project: Dunvegan C Pool Darryl Burns 15.1 Introduction 15.2 Pool Data Collection 15.3 Pool Event Log 15.4 Reservoir Fluid Characterization 15.4.1 Fluid Characterization Program Design Questions 15.4.2 Fluid Characterization Program 15.4.3 Solubility of Acid Gas Mixtures in the Dunvegan C Oil 15.5 Material Balance 15.6 Geological Model 15.7 Geological Uncertainty 15.7.1 Formation Bulk Volume 15.7.2 Porosity 15.7.3 Permeability 15.7.4 Residual (Immobile) Fluid Saturations 15.7.5 Relative Permeability Curve Parameters 15.7.6 Fluid Contacts 15.8 History Match 15.9 Black Oil to Compositional Model Conversion
247 248 249 252 255 255 257 263 263 264 269 269 269 269 270 270 272 272 282
CONTENTS
Recovery Alternatives Economics Economic Uncertainty Discussion and Learning 15.13.1 Reservoir Fluid Characterization 15.13.2 Material Balance 15.13.3 Geological Model 15.13.4 History Match 15.13.5 Black Oil to Compositional Model Conversion 15.13.6 Recovery Alternatives 15.13.7 Economics 15.14 End Note References 15.10 15.11 15.12 15.13
C 0 2 Flooding as an EOR Method for Low Permeability Reservoirs Yongle Hu, Yunpeng Hu, Qin Li, Lei Huang, Mingqiang Hao, and Siyu Yang 16.1 Introduction 16.2 Field Experiment of C 0 2 Flooding in China 16.3 Mechanism of C 0 2 Flooding Displacement 16.4 Perspective 16.5 Conclusion References Pilot Test Research on C 0 2 Drive in Very Low Permeability Oil Field of in Daqing Changyuan Weiyao Zhu, Jiecheng Cheng, Xiaohe Huang, Yunqian Long, and Y. Lou 17.1 Introduction 17.2 Laboratory Test Study on C 0 2 Flooding in Oil Reservoirs with Very Low Permeability 17.2.1 Research on Phase Behavior and Swelling Experiments 17.2.2 Tubule Flow Experiments 17.2.3 Long Core Test Experiments 17.3 Field Testing Research 17.3.1 Geological Characteristics of Pilot 17.3.1.1 Structural Characteristics 17.3.1.2 Characteristics of Reservoir
290 307 312 312 312 315 315 316 317 317 317 317 318 319
319 320 321 324 326 326 329
329 330 330 331 332 333 333 334 334
CONTENTS
Reservoir Properties and Lithology Characteristics 17.3.2 Distribution and Features of Fluid 17.3.3 Designed Testing Scheme 17.3.4 Field Test Results and Analysis 17.3.4.1 Low Gas Injection Pressure and Large Gas Inspiration Capacity 17.3.4.2 Production Rate and Reservoir Pressure Increase after Gas Injection 17.3.4.3 Reservoir Heterogeneity Is the Key to Control Gas Breakthrough 17.3.4.4 C 0 2 Throughput as the Supplementary Means of Fuyu Reservoir's Effective Deployment 17.3.4.5 Numerical Result Shows that Carrying Out Water Flooding after Injecting Certain Amount of C 0 2 Slug is Better 17.4 Conclusion 17.5 Acknowledgement References
Xlll
17.3.1.3
18. Operation Control of C0 2 -Driving in Field Site. Site Test in Wellblock Shu 101, Yushulin Oil Field, Daqing Xinde Wan, Tao Sun, Yingzhi Zhang, Tiejun Yang, and Changhe Mu 18.1 Test Area Description 18.1.1 Characteristics of the Reservoir Bed in the Test Area 18.1.2 Test Scheme Design 18.2 Test Effect and Cognition 18.2.1 Test Results 18.2.2 The Stratum Pressure Status 18.2.3 Air Suction Capability of the Oil Layer 18.2.4 The Different Flow Pressure Control 18.2.5 Oil Well with Poor Response 18.3 Conclusions References
336 339 339 340 340
341
342
343
344 346 349 349 351
352 352 352 353 353 354 356 356 358 359 359
xiv
CONTENTS
19. Application of Heteropolysaccharide in Acid Gas Injection Jie Zhang, Gang Guo and Shugang Li 19.1 Introduction 19.2 Application of Heteropolysaccharide in C 0 2 Reinjection Miscible Phase Recovery 19.2.1 Test of Clay Polar Expansion Rate 19.2.1.1 Test Method 19.2.1.2 Testing results as the Figure 2 and Table 1 shows 19.2.2 Test of Water Absorption of Mud Ball in Heteropolysaccharide Collosol 19.3 Application of Heteropolysaccharide in H2S Reinjection formation 19.3.1 Experiment Process, Method and Instruction 19.3.1.1 Experiment Process 19.3.1.2 Experiment Method 19.3.1.2 Experiment Results 19.4 Conclusions References
361 361 363 364 364 366 367 370 370 370 370 372 373 373
Section 5: Geology and Geochemistry 20. Impact of S 0 2 and NO on Carbonated Rocks Submitted to a Geological Storage of C0 2 : An Experimental Study Stéphane Renard, Jérôme Sterpenich, Jacques Pironon, Aurélien Randi, Pierre Chiquet and Marc Lescanne 20.1 Introduction 20.2 Apparatus and Methods 20.2.1 Solids and Aqueous Solution 20.2.2 Gases 20.3 Results and Discussion 20.3.1 Reactivity of the Blank Experiments 20.3.2 Reactivity with pure S0 2 20.3.3 Reactivity with pure NO 20.4 Conclusion Acknowledgments References
377
377 378 379 380 381 381 384 387 391 392 392
CONTENTS
21. Geochemical Modeling of Huff 'N' Puff Oil Recovery With C 0 2 at the Northwest Mcgregor Oil Field Yevhen I. Holubnyak, Blaise A.F. Mibeck, Jordan M. Bremer, Steven A. Smith, James A. Sorensen, Charles D. Gorecki, Edward N. Steadman, and John A. Harju 21.1 Introduction 21.2 Northwest McGregor Location and Geological Setting 21.3 The Northwest McGregor Field, E. Goetz #1 Well Operational History 21.4 Reservoir Mineralogy 21.5 Preinjection and Postinjection Reservoir Fluid Analysis 21.6 Major Observations and the Analysis of the Reservoir Fluid Sampling 21.7 Laboratory Experimentations 21.8 2-D Reservoir Geochemical Modeling with GEM 21.9 Summary and Conclusions 21.10 Acknowledgments 21.11 Disclaimer References
22. Comparison of C 0 2 and Acid Gas Interactions with reservoir fluid and Rocks at Williston Basin Conditions Yevhen I. Holubnyak, Steven B. Hawthorne, Blaise A. Mibeck, David J. Miller, Jordan M. Bremer, Steven A. Smith, James A. Sorensen, Edward N. Steadman, and John A. Harju 22.1 Introduction 22.2 Rock Unit Selection 22.3 C 0 2 Chamber Experiments 22.4 Mineralogical Analysis 22.5 Numerical Modeling 22.6 Results 22.7 Carbonate Minerals Dissolution 22.8 Mobilization of Fe
XV
393
393 395 395 397 398 400 401 402 403 404 404 405
407
407 409 411 412 413 413 414 416
xvi
CONTENTS
22.9
Summary and Suggestions for Future Developments 22.10 Acknowledgments 22.11 Disclaimer References
418 418 418 419
Section 6: Well Technology 23 Well Cement Aging in Various H 2 S-C0 2 Flui(is at High Pressure and High Temperature: Experiments and Modelling Nicolas Jacquemet, Jacques Pironon, Vincent Lagneau, Jérémie Saint-Marc 23.1 Introduction 23.2 Experimental equipment 23.3 Materials, Experimental Conditions and Analysis 23.3.1 Cement 23.3.2 Casing 23.3.3 Environment 23.3.4 Exposures (Figure 3): 23.3.5 Analyses 23.4 Results and Discussion 23.4.1 Cement 23.4.2 Steel 23.5 Reactive Transport Modelling 23.6 Conclusion Acknowledgments References 24. Casing Selection and Correlation Technology for Ultra-Deep, Ultra- High Pressure, High H2S Gas Wells Yongxing Sun, Yuanhua Lin, Taihe Shi, Zhongsheng Wang, Dajiang Zhu, Liping Chen, Sujun Liu, and Dezhi Zeng 24.1 Introduction 24.2 Material Selection Recommended Practice 24.3 Casing Selection and Correlation Technology
423
424 425 426 426 427 427 427 427 428 428 430 430 432 433 434
437
438 438 441
CONTENTS
Casing Selection and match Technology Below 90°C 24.3.2 Casing Selection and Match Technology Above 90°C 24.4 Field Applications 24.4 Conclusions 24.5 Acknowledgments References
xvii
24.3.1
25. Coupled Mathematical Model of Gas Migration in Cemented Annulus with Mud Column in Acid Gas Well Hongjun Zhu, Yuanhua Lin, Yongxing Sun, Dezhi Zeng, Zhi Zhang, and Taihe Shi 25.1 Introduction 25.2 Coupled Mathematical Model 25.2.1 Gas Migration in Cement 25.2.2 Gas Migration in Stagnant Mud 25.2.3 Gas Unloading and Accumulation at Wellhead 25.2.4 Coupled Gas Flows in Cement and Mud 25.3 Illustration 25.4 Conclusions 25.5 Nomenclature 25.6 Acknowledgment References
442 443 443 445 447 447
449
449 450 451 452 454 456 458 459 460 461 461
S e c t i o n 7: C o r r o s i o n 26. Study on Corrosion Resistance of L245/825 Lined Steel Pipe Welding Gap in H 2 S+C0 2 Environment Dezhi Zeng, Yuanhua Lin, Liming Huang, Daijiang Zhu, Tan Gu, Taihe Shi, and Yongxing Sun 26.1 Introduction 26.2 Welding Process of Lined Steel Pipe 26.3 Corrosion Test Method of Straight and Ring Welding Gaps of L245/825 Lined Steel Pipe 26.4 Corrosion Test Results of Straight and Ring Welding Gaps of 1245/825 Lined Steel Pipe
465
466 466 467 472
xviii
CONTENTS
26.4.1 Atmospheric Corrosion Test Results 26.4.2 Corrosion Test Results at High Pressure 26.4.3 Field Corrosion Test Results 26.5 Conclusions 26.6 Acknowledgments References Index
472 472 474 477 477 477 479
Introduction The Three Sisters - CCS, AGI, and EOR Ying Wu1, John J. Carroll2 and Zhimin Du3 1
Sphere Technology Connection, Calgary, AB, Canada 2 Gas Liquids Engineering, Calgary, AB, Canada 3 Southwest Petroleum University, Chengdu, People's Republic of China
Although there remains some debate about whether or not man is changing the global climate and, if so, whether or not carbon dioxide is the cause of it, there is a significant capital, both political and financial, to reduce carbon emissions. It is not the purpose of this introduction, or this volume for that matter, to enter this debate. The purpose is to review the technology to achieve this and the inter-relations within available technologies. One of the main foci for reducing carbon emission is the so-called process, carbon capture and storage (CCS), removing carbon dioxide from combustion gases and storing them in subsurface formations. The main source of these combustion gases is coal-fired power plants, but other sources are targeted as well. In the petroleum and natural gas business there are two other mature technologies for injecting gas streams. The first of these is acid gas injection (AGI), and the other is injecting carbon dioxide for enhanced oil recovery (EOR). This makes CCS, AGI and EOR three sisters, of sorts. Whereas AGI and EOR are relatively mature processes, CCS is not and there is much those working in the CCS world can learn from both AGI and CCS. Table 1 summarizes the main components for the three technologies. Each of these will be discuss here. Whereas the impetus for acid gas injection is to eliminate sulfurous emissions, and there is little doubt about the effect of these emissions, they also sequester C0 2 . On the other hand, the purpose of injecting C 0 2 for EOR is to produce more oil. Burns [1], in a chapter in this volume discusses, the economics of an EOR xix
xx
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 1. The three sisters: CCS, AGI, and EOR. CCS
AGI
EOR i. virgin ii. recovered
Source of fluid
capture from flue gas
sweetening of natural gas
Pressure
compression
compression
compression
Pipeline
probably a network
commonly a single pipeline
pipeline network
commonly a single well
multiple, injection pattern (5-spot, for example)
disposal
oil recovery
reduced C 0 2 emissions
co2
Well
Purpose By-product
i. probably multiple wells ii. probably deviated wells to achieve high injectivity storage
sequestration
project. Nonetheless, sequestration of C 0 2 is a by-product of these EOR schemes. For CCS the purpose is simply to eliminate carbon emission into the atmosphere. However, C 0 2 captured from flue gas may have value as a source of virgin C 0 2 for EOR projects.
Capture The flue gas stream from a combustion process produces a flue gas that is from 5% to 15% carbon dioxide. The rest of this stream contains mostly nitrogen but also some oxygen and smaller amount of sulfur oxides and nitrogen oxides. The volume of the raw flue gas is too large to make compression and injection feasible. Thus the first step is to "capture" the C 0 2 from the flue gas. In the natural gas business the removal of carbon dioxide (and hydrogen sulfide for that matter) is called sweetening. Much of the technology developed over 75 years in the natural gas business can be transferred to the capture of C0 2 . However there are many
INTRODUCTION
xxi
problems associated with capturing C 0 2 that are not as common in the natural gas business. These include the low pressure of the flue gas stream (near atmospheric pressure versus tens of bars for natural gas) and the contaminants. Oxygen is poison to the common solvents used in the natural gas business. The chapter by Spooner and Engel [2] in this volume discusses the use of amine technology for capturing C 0 2 from flue gas. Among the problems Spooner and Engel address are the high oxygen content of the flue gas and the low pressure. In EOR there must be a source of carbon dioxide when the project begins. This is the so-called "virgin" C0 2 . Once the project starts, some of the C 0 2 will be produced with the oil. This C 0 2 is recovered from the oil and used for re-injection. Initially the recycled C 0 2 will be small but as the project matures this may become as large as 80% or 90% of the carbon dioxide injected.
Compression The next step for each of the three processes is to compress the stream to sufficient pressure such that it can be injected into a subsurface reservoir. In EOR the virgin C 0 2 is usually delivered at such a pressure that little or no compression is required. However the recycled C 0 2 is at low pressure and must be compressed for injection. In AGI the acid gas stream is at low pressure and in comes the sweetening process, where low pressure is used to regenerate the solvent. In acid gas injection and the compression of C 0 2 for EOR it is common to use compression and cooling alone to reduce the water content of an acid gas stream. The water holding capacity of acid gas was discussed in the previous volume in this series by Marriott et al. [3] and also by Satyro and van der Lee [4]. In a chapter in this volume Wright [5] discusses the use of compression and cooling in order to dehydrate an acid gas stream. In particular Wright addresses when dehydration is required and when it is not based on the composition of the gas and its water holding capacity. In some cases, compression alone cannot achieve sufficiently high pressures to inject the stream. In these cases, the stream can be liquefied (using a combination of high pressure and low temperature) and then pumped to higher pressure. Later in this
xxii
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
book Janusch and Braun [6] discuss the pumping of acid gas with diaphragm pumps.
Pipeline For all of the three sisters the compressed gas is transported via pipeline to the injection well(s). In an EOR project the compressed C 0 2 must be distributed through the oil filed such that the optimum oil recovery can be achieved. This requires a network of pipes. For small AGI projects usually only a single injection well is used and thus a single pipeline. However, for very large projects, AGI may require a network of line similar to an EOR project. The volumes injected in a typical CCS project will be very large and thus a single well is probably not an option.
Injection Again in each of the three sisters, the compressed fluid enters a well and travels downward to the target formation. In EOR it is common to have multiple wells arranged in a pattern, some for injecting C 0 2 and some for producing oil. It is also possible to use C 0 2 for huff 'n puff. This involves injecting C 0 2 for a period of time and then allowing the C 0 2 to soak (the "huff"). The same well is the used for producing the oil (the "puff"). Because of the properties of the gas injected and the phase behavior encountered, some unusual behavior can be observed in acid gas injection wells. Mirreault et al. [7] in the previous volume in this series, describe some seeming unusual behaviour in an injection well that have some relatively simple explanation.
Geochemistry The effect of the acid gas, and perhaps more specifically C0 2 , on the reservoir rock is an important consideration in the design of an injection scheme. How does the injected fluid affect the native rock? A case study related to the geochemical interactions is presented in this volume by Holubnyak et al. [8].
INTRODUCTION
xxiii
Summary The three sisters: CCS, AGI, and EOR share many common components. Many lessons can be shared especially between the more mature technologies of AGI and EOR and the newer one, CCS. These commonalities demonstrate that carbon capture and storage is a feasible technology. The remaining chapters in this volume discuss specific aspects of these three sisters and the reader should keep in mind the common aspects of these seemingly different technologies.
References 1. Burns, D. "Enhanced Oil Recovery Project: Dunvegan C Pool", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA. (2011). 2. Spooner, B. and D. Engel, "Carbon Capture Using Amine-Based Technology", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA (2011). 3. Marriott, R.A., E. Fitzpatrick, E Bernard, H. H. Wan, K. L. Lesage, P. M. Davis, and P. D. Clarke, "Equilibrium Water Content Measurements For Acid Gas Mixtures" Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 4. Satyro, M. and J. van der Lee, "The Performance of State of the Art Industrial Thermodynamic Models for the Correlation and Prediction of Acid Gas Solubility in Water", Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 5. Wright, W. "Dehydration-through-Compression: Is it Adequate? A Tale of Three Gases", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 6. Janusch, J. and A.-D. Braun, "Diaphragm Pumps improve Efficiency of Compressing Acid Gas and C0 2 ", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 7. Mireault, R., R. Stocker, D. Dunn, and M. Pooladi-Darvish, "Dynamics of Acid Gas Injection Well Operation", Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 8. Holubnyak, Y.I., S.B. Hawthorne, B.A. Mibeck, D.J. Miller, J.M. Bremer, S.A. Smith, J.A. Sorensen, E.N. Steadman, and J.A. Harju, "Comparison of C 0 2 and Acid Gas Interactions with Reservoir Fluid and Rocks at Williston Basin Conditions ", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 9. Taiman, S.J. and E.H. Perkins, "Concentration Gradients Associated With Acid Gas Injection", Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011).
SECTION 1 DATA AND CORRELATION
1 Prediction of Acid Gas Dew Points in the Presence of Water and Volatile Organic Compounds Ray. A. Tomcej Tomcej Engineering Inc. Edmonton, AB, Canada
Abstract Aromatic hydrocarbons which are present in sour natural gas streams can be absorbed into the amine treating solution at the bottom of the contactor and exit in the rich amine stream. Depending on the process configuration, these dissolved hydrocarbons can end up in the acid gas leaving the amine regenerator. In acid gas injection facilities, trace amounts of heavy hydrocarbons in the acid gas may lead to the formation of a sour hydrocarbon liquid phase in the compressor interstage scrubbers. In this exploratory work, a cubic equation-of-state (EOS) model was used to make predictions of non-aqueous (Lj) dew points in acid gas systems. The objective was to develop a better understanding of the conditions under which this phenomenon can occur, and to reinforce the need for accurate experimental vapor-liquid equilibrium data to support cost effective design and model development.
1.1
Introduction
Benzene, toluene, ethyl benzene and xylene isomers are commonly referred to collectively as BTEX compounds. These compounds are known to be toxic to humans and their containment and disposal are of special interest to the hydrocarbon industry. BTEX environmental contamination is often linked to leakage from underground gasoline storage tanks or accidental spills. Awareness of this toxicity led to regulated clean air emission standards that directly impact Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (3-12) © Scrivener Publishing LLC
3
4
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
the natural gas processing industry as trace amounts of BTEX compounds are associated with produced fluids such as natural gas. Sour gas production generally involves a subsequent processing step in which the hydrogen sulphide (H2S) and carbon dioxide (C0 2 ) are removed to produce an acid gas stream that may be a candidate for acid gas injection. Liquid solvents that are used to remove the H2S and C 0 2 from the gas stream are often aqueous solutions of organic chemicals that have a high affinity for the BTEX compounds. Distribution of the BTEX compounds within the various streams of a natural gas processing plant is a complex phenomenon involving many interrelated process variables such as operating pressures and temperatures, amine composition, amine circulation rates, and others. Of particular interest in acid gas injection, is the amount of BTEX compounds that end up in the acid gas product leaving the amine regenerator. The presence of trace quantities of BTEX compounds in the acid gas, if unaccounted for at the design stage, may lead to the unexpected formation of a sour non-aqueous liquid phase in the compressor train, and considerable operational difficulties. The objective of this work was to develop a better understanding of the conditions under which this phenomenon can occur, and to reinforce the need for accurate experimental vapor-liquid equilibrium data to support cost effective design and model development.
1.2
Previous Studies
In order to estimate the levels of BTEX compounds that will be present in the acid gas, there is a need for accurate vapor-liquid equilibria (VLE) a n d / o r vapor-liquid-liquid equilibria (VLLE) data for BTEX and similar hydrocarbons in amine treating solutions under rich amine conditions. Operating plant data are also useful to verify the predictions of any thermodynamic model. Ng et al. (1999) provided an overview of specific phase equilibria data and physical properties that are required for reliable design of acid gas injection facilities. Hegarty and Hawthorne (1999) presented valuable operating data for a Canadian gas plant using MDEA in which measured BTEX compositions were reported. Mclntyre et al. (2001) and Bullin and Brown (2004) tabulated the experimental data available for hydrocarbon and BTEX solubility in amine treating solutions and demonstrated general trends
PREDICTION OF ACID GAS DEW POINTS
5
in amine plant BTEX absorption using computer simulation. Valtz et al. (2002) presented a comprehensive set of fundamental solubility data for aromatic hydrocarbons in aqueous amine solutions. Miller and Hawthorne (2000) and Jou and Mather (2003) measured the solubility of BTEX compounds in water. Clark et al. (2002) measured bubble and dew points for a nominal 10 mol% H 2 S/90 mol% C 0 2 mixture and regressed an equation of state to match the phase envelope. Satyro and van der Lee (2009) demonstrated that with suitable modification to interaction parameters, a cubic equation of state can provide reliable predictions of phase behavior in sour gas mixtures.
1.3
Thermodynamic Model
A rigorous treatment of the complex phase behavior in the H 2 S-C0 2 water-BTEX system was beyond the scope of this work, which was intended to be exploratory in nature. The Peng-Robinson equationof-state with classical van der Waals mixing rules was used in this study. The interaction parameter for the H 2 S-C0 2 binary was set to 0.1 and all others were set to zero. Table 1 contains the critical properties used for the system components. Table 1. Component critical properties. Critical P, kPa
Critical T, °C
Hydrogen Sulphide
9007.8
100.45
Carbon Dioxide
7386.6
31.05
Benzene
4898.0
Toluene
Component
Acentric Factor
Molecular Weight
0.1
34.076
0.225
44.01
289.0
0.2092
78.112
4105.8
318.7
0.2637
92.138
Ethyl Benzene
3605.9
344.1
0.3026
106.165
o-Xylene
3734.2
357.2
0.3118
106.165
m-Xylene
3536.3
343.9
0.3255
106.165
p-Xylene
3510.8
343.1
0.3211
106.165
6
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The performance of the Peng-Robinson equation of state has been well documented in the literature. The model reproduced the dew point locus of Clark et al. (2002) to within 2.5%.
1.4
Calculation Results
The conditions of the calculations were chosen to encompass those normally found in acid gas injection compression: pressures from 150 kPa to 10 MPa, and temperatures above the hydrate formation curve from 0° to 100°C. Three different nominal acid gas compositions were considered: 20/80, 50/50, and 80 mole% H 2 S/20 mole% C O r Hydrocarbon components studied included: benzene, toluene, ethyl benzene and dimethyl benzenes (xylenes). The model was used to generate the phase envelope for each of the three nominal acid gas compositions. The influence of associated water on the location of the bubble and dew-point loci was not considered in this work. A typical injection profile was generated for each nominal composition using a starting pressure of 150 kPa and constant compression ratio. Temperatures in the compression process were restricted to remain under 150°C. Cooling temperature was set to 50°C. The final pressure was selected to be under 10 MPa but above the mixture critical point. Initial calculations indicated that the phase behavior of the acid gas mixtures in the presence of each of the three xylene isomers was similar. For simplicity only o-xylene was considered in this study. To establish a reasonable range of BTEX compositions, a sensitivity study was undertaken using pure H2S. The model was used to determine the L^ dew point temperature at 4000 kPa using various compositions of benzene and o-xylene ranging from 0 to 5000 ppmv. The results are shown in Figure l. 1 Below concentrations of 100 ppmv, the aromatic compounds increase the dew point temperature by less than 1°C. Hegarty and Hawthorne (1999) reported BTEX content of up to 2500 ppmv in the acid gas of an operating MDEA plant. Using this as a guideline, non-aqueous liquid (L,) dew points were calculated for each of the three nominal acid gas compositions with 500-, 2000- and 5000 ppmv of each of the four aromatic compounds.
1
Figures 1 through 4 appear at the end of this paper.
PREDICTION OF ACID GAS DEW POINTS
7
Figure 1. Effect of BTEX compounds on L, dew point in pure H2S.
Figure 2. Effect of BTEX compounds in 80% H2S - 20% C O r
Clearly this range of calculated points generated a significant amount of data. The results for the 2000 ppmv cases are presented in Figures 2 through 4 and provide an adequate representation of the general trends that were observed. Note that curves labeled as organic compounds represent the dew point loci for the acid gas mixture with 2000 ppmv of only that organic compound.
8
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 3. Effect of BTEX compounds in 50% H2S - 50% C0 2 .
Figure 4. Effect of BTEX compounds in 20% H2S - 80% C0 2 .
Using data from Mclntyre et al. (2001) for BTEX component distribution in the acid gas from an MDEA plant as a guideline, flash calculations were performed at 50°C for the mixture given in Table 2. Identical calculations were performed for a mixture containing 80 mol% H2S and 20 mol% C O r The results are shown in Table 3.
PREDICTION OF ACID GAS DEW POINTS
Table 2. Composition of mixture used for condensation study. Composition, mol %
Component Hydrogen Sulphide
79.82
Carbon Dioxide
19.955
Benzene
1000 ppmv
Toluene
750 ppmv
Ethyl Benzene
250 ppmv
o-Xylene
250 ppmv
Table 3. Condensation study results at 50°C. Pressure, kPa 3268.3
Volume% Lj, BTEX Mixture
Volume %Lj, 80/20 H2S/
co 2
Dew point P
3400
0.009
0
3600
0.031
0
3800
0.074
0
4000
0.169
0
4200
0.420
0
4400
1.29
0 Dew point P
4466.6 4600
3.90
2.46
4800
8.55
7.29
5000
15.6
14.2
5200
26.6
24.8
5400
45.4
42.5
5600
82.8
77.2
5654.1 5674.5
Bubble point P Bubble point P
9
10
1.5
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Discussion
In the absence of experimental data for dew point conditions in acid gases with contaminants, there can be no absolute conclusions drawn on the accuracy of the predictions. This exploratory study clearly emphasizes the importance of experimental research to provide fundamental information for process design and advanced model development. The results in Figures 2 through 4 illustrate that with conservative cooling temperatures and with BTEX contaminant levels in the range of those already measured in an operating MDEA plant, it is possible to enter the three-phase region in the higher pressure interstage coolers and separators in acid gas injection facilities. More aggressive cooling escalates the potential for three-phase conditions. The formation of a second liquid phase in the compression interstage cooling system, in itself is not a problem, provided that the phase behavior phenomenon is understood at design time. The Lj phase is less dense than water, contains u p to 20 mol% BTEX and, if formed, will accumulate in the interstage separators. As pointed out by Hegarty and Hawthorne (1999), it is extremely important to obtain an accurate inlet gas composition, including an extended analysis of the C6+ fraction to determine the aromatic content. Once the BTEX content, if any, is identified it can be accounted for in any process design, modeling, or operational troubleshooting of downstream processes such as acid gas injection. In spite of the purely predictive nature of the calculated results, the following general observations can be made by analyzing Figures 2 through 4. The same behavior is observed in the 500 ppmv and 5000 ppmv calculated results. • At a given pressure, the presence of BTEX compounds in acid gas widens the phase envelope, with this effect being more pronounced in acid gases with higher C 0 2 content. • At a given pressure, the presence of BTEX compounds in acid gas increases the L^ dew point temperature, with this effect being more pronounced in acid gases with higher H2S content. This is, in part, a result of the shift of the acid gas phase envelope to higher temperatures in high H2S mixtures.
PREDICTION OF ACID GAS DEW POINTS
11
• At equal concentration in the acid gas and at equal pressure, BTEX compounds increase the L] dew point temperature in the order: benzene, toluene, ethyl benzene and o-xylene with o-xylene having the most pronounced effect. • In all cases, the possibility of non-aqueous Lj formation is highest in the separator before the final stage of compression. • If compressed acid gas is cooled to lower temperatures (e.g. 30°C) in the compressor facility, this increases the possibility of Lj formation. • If BTEX compounds are present in the acid gas at levels less than 100 ppmv, the acid gas dew point locus is relatively unaffected. The dew point loci shown in Figures 2 through 4 indicate where the first droplet of L^ forms. Table 3 contains an example of the condensation behavior inside the phase envelope at constant temperature. Note that the condensation behavior of the BTEX mixture is similar to the BTEX-free system except for the deep depression of the dew point pressure. Lines of constant liquid volume % are widely spaced in this region of the phase envelope. This behavior is similar to the condensation behavior of rich gas systems. The location of the bubble point is relatively unaffected by the organic compounds.
References Bullin, Jerry A. and William G. Brown, "Hydrocarbons and BTEX Pickup and Control from Amine Systems", Proceedings of the 83rd Gas Processors Association Annual Convention, New Orleans, March 14-17,2004. Clark, M.A., W.Y. Svrcek, W.D. Monnery, A.K.M. Jamaluddin and E. Wiehert, "Acid Gas Water Content and Physical properties: Previously Unavailable Experimental Data for the Design of Cost Effective Acid gas Disposal Facilities, and Emission Free Alternative to Sulfur Recovery Plants", Hycal Energy Research Laboratories, 2002. Hegarty, Mike and Dean Hawthorne, "Application of BTEX/Amine VLE Data at Hanlan Robb Gas Plant", Proceedings of the 78th Gas Processors Association Annual Convention, Nashville, March 1-3,1999. Jou, Fang-Yuan and Alan E. Mather, "Liquid-Liquid Equilibria for Binary Mixtures of Water+Benzene, Water+Toluene and Water+p-Xylene from 273K to 458K", /. Chem. Eng. Data, 48, 750-752(2003)
12
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Mclntyre, G.D., V.N. Hernandez-Valencia and K.M. Lunsford, "Recent GPA Data Improves BTEX Predictions for Amine Sweetening Facilities", Proceedings of the 80th Gas Processors Association Annual Convention, San Antonio, March 12-14,2001. Miller, David J. and Steven B. Hawthorne, "Solubility of Liquid Organics of Environmental Interest in Subcritical (Hot/Liquid) Water from 298K to 473K", /. Chem. Eng. Data, 45, 78-81(2000). Ng, Heng-Joo, John J. Carroll and James Maddocks, "Impact of Thermophysical Properties Research on Acid Gas Injection Process Design", Proceedings of the 78th Gas Processors Annual Convention, Nashville, March 1-3,1999. Satyro, Marco A. and James van der Lee, "The Performance of State of the Art Industrial Thermodynamic Models for the Correlation and Prediction of Acid Gas Solubility in Water", Proceedings of the First International Acid Gas Injection Symposium, Calgary, Alberta, Canada, October 5-6, 2009. Valtz, A., P. Guilbot and D. Richon, "Amine BTEX Solubility", Gas Processors Association Research Report RR-180, 2002.
2
Phase Behavior of China Reservoir Oil at Different COJnjected Concentrations Fengguang Li, Xin Yang, Changyu Sun, and Guangjin Chen State Key Laboratory of Heavy Oil Processing, China University of Petroleum Beijing, People's Republic of China
Abstract The phase behavior of China reservoir oil at different C 0 2 injected concentrations has been studied at the temperature of 339.2 K using a high-pressure PVT unit. Seven groups of reservoir fluids with C 0 2 molar contents of 0, 10.0, 34.1, 44.7, 48.9, 57.8, and 65.0 mol% have been prepared. Saturation pressure of reservoir fluids at seven C 0 2 injected contents were measured. The reservoir oil density and viscosity at different pressures under reservoir temperature were also obtained. The influence of C 0 2 molar contents on the interfacial tension of C 0 2 injected reservoir oil under stratum conditions was determined using a pendant drop method. The experimental data indicated that when C 0 2 content is lower than 45 mol%, the increase of bubble point pressure is slow. After that, the bubble point pressure value increases more sharply with the increase of C 0 2 molar concentrations. The reservoir viscosities decrease sharply with the increase of C 0 2 concentration when the system pressure is above the bubble point for different injection contents. The experimental results of interfacial tension for C 0 2 injected crude oil/stratum water show that it decreases with the increase of C 0 2 injected concentrations. The pressure has a slight effect on the interfacial tension value. These phase behavior data will be helpful for evaluating the effect of C 0 2 injected method to enhance oil recovery.
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (13-22) © Scrivener Publishing LLC
13
14
2.1
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Introduction
The fluid phase behavior study is used as an important basis for miscible-slug process and predominant displacement mechanism, which is of critical importance during the miscible displacement process (1). The conventional fluid phase behavior test is usually conducted using PVT (Pressure-Volume-Temperature) unit. It is of great concern in many high-pressure technologies, such as fluid extraction process, exploration of near-critical gas condensate/ volatile oil reservoir, and gas-injected enhanced oil recovery processes. C 0 2 displacement technology is recognized as a significant and well-established means for oil and gas enhanced recovery both at home and abroad. Miscible gas injection could minimize the trapping effect of capillary forces and is recognized as an economic enhanced oil recovery process. Although some PVT fluid phase behavior data are available in the published papers, they are still insufficient because of the complexity of multi-component reservoir fluid. In this work, the phase behavior of China reservoir fluids collected from Jilin oil field were analyzed at different C 0 2 injected concentrations and pressures using a high-pressure PVT device. The density, bubble point pressure, viscosity, and interfacial tension properties of reservoir fluid at different C 0 2 injected mole percents and pressures under the stratum temperature were systematically measured.
2.2
Preparation of Reservoir Fluid
The reservoir fluid sample was collected from China Jilin oil field at reservoir conditions. The stratum temperature was 339.2K. The reservoir fluid arriving from the well was separated and flashed to standard condition. The molar composition of reservoir fluids was then obtained from analysis of the gas and oil samples. The gas phase was analyzed by HP6890 gas Chromatograph. The liquid phase was analyzed by simulating distillation process using HP5890A. Afterwards, the reservoir fluid composition was obtained by combining the gas and liquid phase compositions using the gasoil ratio (GOR). The measured composition for reservoir fluid was shown in Table 1. Molecular weights of the oil phase were determined by vapor pressure osmometer (VPO) and the determined molecular weight was 420 g/mol.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
15
Table 1. The composition of reservoir fluid. Gas phase (mol%)
Oil phase (mol%)
Reservoir fluid (mol%)
N2
2.491
0.968
co2
0.190
0.074
CH 4
61.921
24.059
C2H6
9.585
3.724
C3H8
11.226
4.361
i"C 4 H 10
1.721
0.669
n-C 4 H 1 0
6.983
2.713
i"C 5 H 12
1.301
0.505
n-C 5 H 1 2
2.721
1.057
C
6H14
1.861
0.723
C
7H16
0.884
0.540
C
8H18
2.998
1.833
C
9H20
2.178
1.332
C
10H22
2.980
1.823
90.960
55.619
C1I+
Seven groups of C 0 2 injected concentration (including 0% C0 2 ) were chosen to study the reservoir fluid behavior under gas injection process. The C 0 2 injected crude oil was prepared using RUSKA PVT device.
2.3
PVT Phase Behavior for the C 0 2 Injected Crude Oil
Phase behavior of China reservoir oil was systematically investigated using a RUSKA high-pressure PVT system which was described in our previous papers (2,3). The PVT data at different
16
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
C 0 2 injected molar components was measured to build the relationship between the volume and pressure of reservoir oil. The bubble point pressure and density of reservoir fluid at different pressures could then be determined according to the measured PVT data, which is useful to calculate the phase behavior properties such as the relatively volume, solubility of injected C 0 2 in oil, and so on. The density of the C 0 2 injected reservoir fluid at different pressure under the strata temperature was plotted in Figure 1. From Figure 1, it can be found that there exists an inflexion for the curve of reservoir fluid density and pressure, showing the process of phase transition. When the C 0 2 concentration achieves 65.0 mol%, there is no significant difference between the gas phase and liquid phase, showing that there may exhibit first contact miscibility condition when the C 0 2 injected contents is 65.0 mol%. The bubble point pressure at seven C 0 2 molar compositions determined from PVT measurement was shown in Figure 2. According to Figure 2, it shows that bubble point pressure increases with the increase of C 0 2 injected concentrations. When C 0 2 content is lower than 45mol%, the increase of bubble point pressure is slow. However, when C 0 2 content is higher than 45mol%, the bubble point pressure value increases more sharply with the increase of C 0 2 molar concentrations. The bubble point pressure data is also used to choose the suitable C 0 2 injected concentration.
Figure 1. Variation of reservoir oil density for C 0 2 injected crude oil at different C 0 2 mole percents and pressures.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
17
Figure 2. Bubble point pressure at different C0 2 injected concentrations for China reservoir crude oil.
2.4 Viscosity of the C0 2 Injected Crude Oil Viscosity is an important transport property in petroleum production and processing. RUSKA falling ball viscometer connected with RUSKA high-pressure PVT device was used in this work to investigate the viscosity of China Jilin oil samples after different C 0 2 content was injected under stratum conditions. The basic principle of falling ball viscometer is based on Stokes law. The fluid viscosity could be exactly calculated by Stokes law according to the time of the ball travels through internal pipe from the top to the bottom. If the falling ball behaves to be laminar flow, the following equation was used: p = kt(pB-pF)
(1)
where pB and pF are the density of the ball and fluid, respectively. t is the travel time. A: is a constant value related to the diameter of the falling ball and the angel of the apparatus. Before the experiment, a falling ball was selected to measure the constant value k in Eqn. (1) using standard silicon oil for the viscometer. Thereafter, the reservoir crude oil viscosities were systematically measured with the same calibrated ball at different C 0 2 injected molar concentrations and pressures. The reservoir fluid viscosity was tested from higher pressure under single phase conditions until close to the
18
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
saturation pressure. After the pressure was lower than the bubble point pressure, a gas exhaust valve was open to slowly reduce to the experimental pressure and the stable time was prolonged to 4-5 h. The measured viscosity for C 0 2 injected crude oil at different C 0 2 mole percents and pressures were plotted in Figure 3. As shown in Figure 3, the viscosity for C 0 2 injected crude oil decreased apparently with increasing of C 0 2 content. When the C 0 2 injected amount changed from 0 to 65.0 mol%, the reservoir oil viscosity value decreased greatly. At about 30 MPa, the viscosity value can decrease from 10.6 cP to 1.1 cP when 65 mol% C 0 2 was injected. It can be found that when the experimental pressure is higher than the saturated value, the reservoir oil viscosity increases with the increase of pressure; When it is lower than the saturated pressure, the reservoir oil viscosity increases with the decrease of pressure. With the decrease of pressure, more C 0 2 was released from the reservoir oil and induced the increase of viscosity of the residual oil. From Figure 3, it can be concluded that C 0 2 injecting is significant in favor of the decrease of viscosity of Jilin reservoir crude oil. However, after C 0 2 content was higher than 44.7 mol%, the reservoir oil viscosity at single phase condition does not decrease significantly with the further increase of C 0 2 injecting concentration. Meanwhile, During the C 0 2 injecting concentration increases from 0 to 44.7 mol%, the bubble point pressure only increases from 11.28 MPa to 14.14 MPa. However, when the C 0 2 injected concentration
Figure 3. Variation of viscosity for C 0 2 injected crude oil at different C 0 2 mole percents and pressures.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
19
increases from 44.7 mol% to 65.0 mol%, the bubble point pressure increases from 14.14 MPa to 25.0 MPa. Therefore, from the view of decrease of viscosity and bubble point pressure, there exists a suitable C 0 2 injecting concentration and high C 0 2 concentration is not needed.
2.5 Interfacial Tension for C0 2 Injected Crude Oil/Strata Water A great amount of reservoir water exists in the stratum after water displacement process of oil field. There is a special need for accurate interfacial tension estimation because the movement of reservoir fluids is influenced to a great extent by capillary forces. The C 0 2 injected concentration also plays an important role on the interfacial phenomena. In this work, the influence of C 0 2 molar contents on the interfacial tension of injected crude oil/water was systematically investigated using the JEFRI pendant drop high-pressure interfacial tension apparatus manufactured by D.B.Robinson (Canada), which the maximum working pressure is 34.5 MPa (5,000 psi) and the operating temperature range is 253-473 K. The experimental apparatus and procedures were detailed described in our previous papers (4,5). The interfacial tension measurement is based on the following principle: If the drop is in equilibrium with its surroundings gas, the interfacial tension (y) values can be calculated directly from an analysis of the stresses in the static, pendant drop, using the following equations developed by Andreas et al. (6): = ApDe2g/H
(2)
l/H = f(ds/de)
(3)
7
where Ap is the density difference between the two phases, De is the unmagnified equatorial diameter of the drop, g is the gravitational constant, ds is the diameter of the drop at a selected horizontal plane at height equal to the maximum diameter de. Andreas et al. have prepared a detailed table of 1/H as a function (djd). The difference in density between reservoir oil and water could be calculated from the measured density data. The interfacial tension of C 0 2 injected crude oil /reservoir water were all measured
20
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
under single-phase conditions at the stratum temperature. The measured interfacial tension data for C 0 2 injected reservoir oil/ water at different C 0 2 injected molar concentrations and pressures are plotted in Figure 4. As shown in Figure 4, the interfacial tension for C 0 2 injected oil/reservoir water decreased apparently with the increase of C 0 2 injected molar concentration when C0 2 content varies from 0 to 65.0 mol%. The dissolvability of C 0 2 in oil has a significant influence on the interfacial tension value. The interfacial tension decreased by about one-third as the C 0 2 injected amount changed from 0 to 65.0 mol%. It also shows that the interfacial tension of the C 0 2 injected crude oil/water increased with increasing pressure. During the experiment process, the experimental pressure was always higher than the bubble point pressure at the corresponding C 0 2 injected condition. Compared with the effect of C 0 2 injected amounts, the pressure has only a slightly effect. When the C 0 2 composition was 65.0 mol%, the C 0 2 injected oil system approached complete irascibility and the interfacial tension data of C 0 2 injected crude oil/ reservoir water changed a little with an increase in pressure.
2.6
Conclusions
The phase behavior of reservoir oil collected from China Jilin oil field was systematically investigated by using a high-pressure RUSKA PVT device at different C 0 2 injected concentrations and
Figure 4. Variation of interfacial tension for C 0 2 injected oil/reservoir water at different C 0 2 mole percents and pressures.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
21
pressures under strata temperature. Seven groups of C 0 2 injected concentrations varying from 0 to 65.0 mol% were prepared. The bubble point pressure increases from 11.28 MPa to 25.0 MPa when C 0 2 content increases from 0 to 65.0 mol%. When the C 0 2 concentration achieves 65.0 mol%, there is no significant difference between the gas phase and liquid phase, showing that there may exhibit first contact miscibility condition under the corresponding C 0 2 injected content. The viscosity for C 0 2 injected crude oil decreased apparently with increasing of C 0 2 content. C 0 2 injecting is significant in favor of the decrease of viscosity of Jilin reservoir crude oil. However, after C 0 2 content was higher than 44.7 mol%, the reservoir oil viscosity under single phase condition does not decrease significantly with the further increase of C 0 2 injecting concentration. The interfacial tension for C 0 2 injected oil/reservoir water decreased apparently with the increase of C 0 2 injected molar concentration when C 0 2 content varies from 0 to 65.0 mol%. When the C 0 2 composition was 65.0 mol%, the C 0 2 injected oil system approached complete miscibility and the interfacial tension data of C 0 2 injected crude oil/reservoir water changed a little with an increase in pressure.
Literature Cited 1. W. Yan, L.K. Wang, L.Y. Yang, T.M. Guo, Fluid Phase Equilibria, Vol. 190, p. 159-178, 2001. 2. M.X. Gu, Q. Li, X.Y Zhou, W.D. Chen, T.M. Guo, Fluid Phase Equilibria, Vol. 82, p. 173-182,1993. 3. H.Q. Pan, T. Yang, T.M. Guo, Fluid Phase Equilibria, Vol. 105, P. 259-271,1995. 4. C.Y. Sun, G.J. Chen, L.Y. Yang, /. Chem. Eng. Data, Vol. 49, p. 1023-1025,2004. 5. C.Y. Sun, G.J. Chen, /. Chem. Eng. Data,, Vol. 50, p. 936-938, 2005. 6. J.H. Andreas, E.A. Hauser, W.B. Tucker, /. phys. Chem., Vol. 42, p. 1001-1019, 1938.
This page intentionally left blank
3
Viscosity and Density Measurements for Sour Gas Fluids at High Temperatures and Pressures B.R. Giri, P. Biais and R.A. Marriott* Alberta Sulphur Research Ltd. Department of Chemistry University of Calgary Calgary, AB, Canada
Abstract Designing an acid gas injection scheme requires an accurate knowledge of the density and viscosity of the injected fluid as these properties are used to optimize compression, monitor transportation and model gas mobility in the reservoir. Fit-for-purpose models are developed based on the available literature data, which in some instances are either inaccurate or studied at industrially irrelevant temperatures and pressures. Moreover, the errors for predicted data at high pressures and temperatures can be as large as 20-50%. An extensive literature search by Schmidt et al. [1] revealed that there are limited data for H2S and its mixtures available in the literature; most of which are limited to gaseous H2S and saturated liquids. The only existing data that extend to higher pressures (p = 10 to 50 MPa) and temperatures (T = 115 to 140°C) are from Monteil et al. [2] which were reported in the late 60's, after which no measurements appeared to have been carried out. Expansion of the literature data to fill the void temperature and pressure regions, especially at relevant conditions for acid gas injection schemes (T = 0 to 150°C and p = 0.1 to 75 MPa) are desired so that the discrepancies of existing data sets can be resolved and reference viscosity models can be further tested and parameterised. It is worthwhile to note that during the recent development of the H2S viscosity model of Schmidt et al., [1] the data set from Monteil et al. [2] was excluded due to inconsistency. This further demonstrates the importance of additional experimental studies for the determination of H2S viscosity and density at elevated pressures and temperatures. Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (23-40) © Scrivener Publishing LLC
23
24
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
We have recently begun an experimental program aimed at measuring the high-pressure densities and viscosities of H2S and other acid gas mixtures using an Anton Parr vibrating tube densimeter and a Cambridge oscillating piston viscometer at p = 1 to 100 MPa and T = 0 to 150°C. This paper discusses how these instruments were commissioned, calibrated and operated. Interim C0 2 , CH 4 and H2S results show the accuracy and reproducibility of the high-pressure measurements.
3.1
Introduction
Design of an acid gas injection (AGI), sour gas injection or C 0 2 injection scheme requires that the density and viscosity properties of the fluid be well known [1,3,4]. From pre-compression to the reservoir, the viscosity is required to assess frictional pressure drops and the density is required to calculate pressure gains due to static head. Expansion of the literature data to fill the applicable temperature and pressure regions, especially at relevant conditions for AGI schemes are desired so that the discrepancies within existing data sets can be resolved and models can be further parameterised. While density and viscosity properties have been well studied for pure C 0 2 and methane, the data for H2S are sparse at industrially relevant conditions, particularly H2S viscosities at higher pressures [1]. A notable exception is Monteil et al. [2] who have reported some H2S viscosities at high pressures (p = 10 to 50 MPa; T = 115 to 140°C). However, it should be noted that, recently Schmidt et ah, [1] have excluded the data set from Monteil et al. [2] due to inconsistency. In order to determine the range of conditions which would be applicable to industry we considered that acid gas streams moving through traditional compression cycles involve a broad range of temperatures from T = 0 to 150°C. For examples of applicable pressures, Mireault et al. [5] have used pressures of 30 MPa for a target aquifer and 3 MPa for a targeted depleted reservoir. If the fluid is being used for reservoir pressure maintenance, one can expect even larger target reservoir pressures (p > 700 MPa). Thus the overall temperature and pressure ranges targeted by our research includes T = 0 to 150°C and p = 0.1 to 100 MPa. Within this range of conditions we intend to measure a variety of pure acid gas components and acid gas mixtures, beginning with C0 2 (calibrant), CH 4
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
25
and H2S. The large temperature and pressure ranges outlined is experimentally challenging in large part because of fluids involved. Pure C 0 2 and H2S are gaseous, liquid and supercritical within these conditions; therefore, measurements must cover a range from p « 10 to 1200 kg m 3 and 77 « 10 to 300 uPa s (0.01 to 0.30 cP). So far we have gained enough data to estimate the accuracy of our new high pressure instruments, vibration tube densimeter (VTD) and oscillating cylinder viscometer (Cambridge). This paper discusses our experimental methods, some preliminary data for C 0 2 (calibrant), CH 4 and H2S; and provides some evaluation of the instruments capabilities.
3.2
Experimental
3.2.1 Density Measurement There are several methods for accurately measuring densities at high pressure. Providing the mass of the fluid can be measured with high confidence, isochoric vessels with good pressure measurement and stable temperature control are simple and have yielded high quality results in all fluid regions [6-8]. With the isochoric method the vessel can be heated to desired temperatures and the resulting pressure measured. A second vessel can be used for controlled isothermal fluid expansion (Burnett Expansion) [9,10]. Another accurate method includes measuring the buoyancy of a sinker, or better yet two sinkers, which are completely immersed in a high-pressure fluid [11,12]. Vibrating Tube Densimeters, VTDs, have the advantage of a small volume, applicability over a wide range of densities, typically p = 1 to 2000 kg m 3 , and they can be used to measure densities for static or flowing fluids. VTDs have long been used by the brewing and distillation industry to quantify alcohol content [13]. The precision of the VTD technique was improved in 1974 by Picker et al. [14] and extended to high pressure in 1984 by Albert and Wood [15]. Provided they are coupled with good temperature and pressure control, they can yield accurate results up to very high pressures. High pressure VTDs have been commercialized by Anton Paar (DMA-HDT and DMA-HPM). In this study densities were measured using an Anton Paar DMA HPM vibrating tube densimeter, VTD. The densimeter
26
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
has Hastelloy C-276 wetted parts, a stated temperature range of T = -10 to 200°C and a pressure limit of p = 140 MPa. A vibrating tube densimeter can be theoretically described using the undamped resonance frequency of a simple harmonic oscillator, co:
(1) where k is the spring constant and m is the mass of the vibrating system consisting of the tube, mt, and the fluid inside the tube, ma or mb. At a specific temperature and pressure, changing the internal fluid from a to b results in a system mass change of ma - my which can be related to the difference in density, pa - pb:
k 2
2
= k'tf-Tt),
(2)
where T.=2jt/COK is the time period of oscillation for the tube containing fluid / and k' is the calibration constant for the instrument. The calibration constant can be determined by measuring the time period for two fluids of well known density. Due to thermal expansion and compressibility of the vibrating tube, the calibration constant, k', of Equation 2 is both temperature and pressure dependent. Because the temperature is reproducible to within 0.01 °C, isothermal calibrations have been determined at T = 0,50,100 and 150°C and from p = 0.09 to 100 MPa. The isothermal expression used for calculating the density was Pr,a = K (P) ■ {rf,a
- 4,b ) + PT,b
O)
For Equation 3, pTa is the density of fluid a, x\A is the oscillation time period of the tube filled with fluid a and x\h is the oscillation time period for the tube filled with air at laboratory pressure. For each temperature, a simple linear expression, kT(p) = c + dp, was fit by least square regression using the time period of oscillation for a = C 0 2 [16] (p = 1, 2, 5,10, 20, 50 and 100 MPa) and air at 0.09 MPa (atmospheric pressure in Calgary).
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
3.2.2
27
Viscosity Measurement
There are many potential high pressure viscosity techniques to choose from, such as a capillary viscometry, falling body viscometers (e.g., Stokes, rolling ball, falling piston, etc.) or oscillating viscometers (oscillating disc, vibrating wire, etc.). Unlike several of the densimeter techniques, most viscometers are built to measure liquid or gaseous viscosity and are rarely designed for a large range of viscosity. Wakeham et al. [17] have recently published a review on the development for some of these techniques. Some common instruments will be briefly discussed here. The high pressure capillary viscometer is similar to the commonly used Ostwald viscometers (u-tube) which are often used for liquids under gravity flow and normally at atmospheric pressure. Rather than gravity flow, most high-pressure capillary viscometers use pistons to drive fluids through a capillary tube either at constant flow (measuring the difference in pressure) or at constant pressure difference (measuring the flow). Through Poiseuille's law for steady state fluid flow, the viscosity can be calculated. Capillary viscometers can be adapted for both liquid and gaseous fluids by changing the size of the capillary line (length a n d / o r internal diameter). A common experimental issue is the low tolerance for small particles which can obstruct flow. Falling body or sinker type viscometers can include falling ball, falling piston and rolling ball viscometers. In general they all involve some object falling through a static fluid under constant gravitational force with an opposing drag. The falling ball and/or falling piston viscometer measurement was originally conceived by Stokes [18] and applied within the work of Flowers [19]. The accuracy of the viscometer depends on the accuracy of the velocity measurement, i.e., the travel time measurement for the object to traverse some known distance. In order to optimize the elapse time the falling object's density (buoyancy) can be changed, the object's shape (drag and tolerance) can be changed, or additional friction can be added by allowing the object to roll/slide on the surface of an inclined tube. Finally several techniques for have been used to better measure elapse time, e.g., optical [20] and electromagnetic [21,22]. Falling body viscometers are well suited for high-pressure applications, because the fluid is static; however, they are commonly used for liquid phase conditions versus the gas phase where the viscosity is very low (77 < 20 uPa-s). Other high pressure instruments adapted
28
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
to low viscosity measurements include oscillating disks, [23-27] vibrating wire [28,29] and torsionally vibrating piezoelectric quartz crystal [30,31]. Finally, a modification of the falling cylinder viscometer has been commercialized by Cambridge Viscosity Inc. This type of viscometer, henceforth referred to as a Cambridge Viscometer, was designed for high viscosity fluids; however, by using a hollow cylinder with less tolerance between the cylinder and vessel wall, it has been possible to reduce minimum measurable viscosity. The primary advantage of this viscometer is the small volume and broad viscosity range, eg., other cylinders can be purchased to accommodate very viscous fluids. The ASRL Cambridge Viscometer is operated with a low-mass silco-coated magnetic piston of diameter 0.312" and an advertised viscosity range of 20 to 200 uPa-s. The piston resides in a cylindrical SS-316 chamber with an internal diameter of 0.314" and operating conditions of T=190°C and p = 140 MPa. For this work the viscometer was held horizontal. The piston is moved a predetermined distance (0.2") at a constant force determined using two magnetic coils outside the SS-316 stainless steel chamber. By alternating the power to the coils, the round trip travel time is measured and translated into absolute viscosity. The measurement is completed for the motion in both directions. The optimal travel time for each viscometer piston is ca. 26 seconds at full scale; therefore, for a 20 - 200 uPa-s viscosity range, a total cycle time of 26 seconds should correspond to a viscosity of 20 uPa-s. To our knowledge, there are some research groups using this instrument; however, no viscosity data from this instrument at these low viscosities have been published in the open literature. Therefore, we have undertaken extensive testing of the instruments performance over a wide range of experimental conditions. Our early testing of the viscometer resulted in the conclusion that the factory calibration settings were inadequate, especially at high-pressures and low-temperatures. We have explored our own calibration procedure using pentane, hexane and pure C0 2 . The Calibration Drive Level (CDL) is the primary parameter which determines the magnitude of current flowing into the magnetic coils to drive the piston at a constant force. To begin our calibration, all other adjustable instrument parameters were initially set equal to zero. After cleaning the internals with isopropyl alcohol, the viscometer was evacuated for several hours, flushed and charged with
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
29
n-pentane (99.5%¿ Fischer Scientific) at a pressure slightly above 1 atm and at T = 25°C. This small pressure ensured that no bubbles were formed inside the chamber. Pentane was chosen to determine CDL because its viscosity is 217.9 uPa-s at p = 1 atm. and T = 25°C, [32] which is close to the high viscosity limit of the piston used. The next step in the calibration procedure was to optimize the high end correction factor (CHC) using hexane as high-viscosity fluid (77 = 296.3 uPa-s at p = 1 atm. and T - 25°C). Through an interative calibration and re-checking the calibration over time, the CDL and CHC were found to be 420 mA and -0.15, respectively. These values are significantly different from the factory calibration settings (CDL = 452 mA and CHC = 1.3). The low-end viscosity correction factor (CLC) was checked by measuring laser grade C 0 2 (PRAXAIR, 99.9995%) at P = 3 bar and T = 25°C. The CLC value was found to be insignificant and was set to zero. The parameters determined above worked well for the several fluids tested during this procedure as long as the measurements were carried out at low pressure. When the pressure or the temperature is changed significantly, the tolerance between the piston and the viscometer chamber also increases thus decreasing the resistance to motion. To compensate for these effects, a corrected viscosity, r¡a, is calculated from raw viscosity, 77^ using an isothermal correction factor which is linear in pressure: Va=^n+dvPynr
(4)
Note that the form of equation is equivalent to those outlined by Cambridge, [33] where (5)
I « = M VA c,=Jfc r =
1 + TPC
(T-25°C)^
(6)
100°C
and j
_ (,/Cp — L)kj
1-
p
PRC 20,O0Opsia
kT
(7)
In this case, cn and dn of Equation 4 were determined at each temperature by least square regression of the raw viscosities for C 0 2 (p = 1, 2, 5, 10, 20, 50 and 100 MPa) and the calculated viscosities
30
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
from Fenghour et al. [34] Again, these correction factors were found to be significantly different from the factory recommended values.
3.2.3
Charging and Temperature Control
A schematic of the experimental charging, control and logging system is shown in Figure 1. H2S was charged at pressure through a custom built SS-316 hydraulic floating piston (ca. 250 cm3). Ethylene glycol hydraulic fluid was delivered using a Waters High Performance Liquid Chromatography, HPLC, pump. Laser grade C 0 2 (PRAXAIR 99.9995%) was charged using a liquid C 0 2 pump (SFT-10, supercritical fluid technology) and methane (PRAXAIR 99.999%) was delivered using an air operated diaphragm gas compressor (pmax = 75 MPa; Supperpressure Inc. 46-14025-1). Pressure was measured via a Hastelloy Honeywell Sensotec TJE pressure transducer with a maximum calibrated pressure of p = 140 MPa. All valves and tubing were SS-316 (pmax = 210 MPa). Extra valves were included for fine adjustment of pressures, i.e., by displacement of the fluid by adjusting the valve stem position. All measurements were completed for static fluids.
Figure 1. A schematic of the vibrating tube densimeter and oscillating cylinder viscometer system. Component details can be found in the text.
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
31
The temperature of the VTD unit was controlled using a NESLAB RTE740 circulating bath which can control to within ± 0.01 from T = -40.00 to 200.00 °C. Temperature was measured at the VTD using an internal platinum resistance thermometer, PRT, and a second PRT (100 Q, 3 wire) inserted into the face plate and between the unit inlet and outlet. This second PRT was previously calibrated using the triple point of pure water and melting point of pure indium (99.9999%) according to ITS-90 (T mn = 0.01 °C; T = 156.5985 °C) [35]. The calibrations for both t,H20
'
m,ln
/ L
J
PRTs were checked by slowly melting distilled water which had been frozen inside the VTD. The inflection in density/time period upon melting was within ± 0.02°C for both PRTs. The temperature of the viscometer was controlled using a Julabo F12 with a range of -20 to 190°C and a stability of ± 0.03°C.
3.3
Results
Figure 2(a) shows the final correlation plot for the C 0 2 VTD calibration data at p = 1, 2, 5, 10, 20, 50 and 100 MPa and all four isothermal temperatures (T = 0, 50, 100 and 150°C). Figure 2(b) shows the differences between the experimental values and the Span and Wagner [16] reference equation used for calibration. Figure 2(b) also show a similar comparison of some literature data. The comparison of the calibrated experimental data shows a pooled standard deviation of 1.2 kg m~3. This accuracy is less than much of the literature data; however, we have found that this densimeter can produce slightly better results if applied to a narrower range of densities. Also the instantaneous time period has been used; whereas, some averaging may improve future results. Previous work with benzene showed an estimated error of 0.4 kg m~3. Figures 3(a) and 3(b) show the similar plots for the C 0 2 viscosity data; experimental viscosity versus those calculated using Fenghour et al. [34] and the relative difference between the experimental and calculated values. With the exception of the three largest pressures at T - 100°C, the pooled standard error based on the correlation plot is ca. 2% which is similar to the stated accuracy of the reference equation. The overall estimated relative error for each measurement has been calculated using 877/77 = [0.0004 + (2 a/77)2]05., where 0.0004 is the square of the calibration confidence (2%) and a is the standard deviation for the averaged measurement. Note
32
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. (a) Correlation plot of the isothermal VTD calibration using C 0 2 at T = 0,50,100 and 150°C (p = 1 to 100 MPa); (b) difference between experimental C 0 2 densities and those calculated using the Span and Wagner EOS;(16) ♦, this work (VTD); o, pooled isochoric densimeter literature;'6"9' A , pooled float/ sinker densimeter literature;" 112 ' x, pooled piezometer data;13738' O , Ihmels and Gmehling (VTD).09»
that each measurement represents an averaged reading of 20 data points. The estimated relative errors for the individual C 0 2 measurements ranged from Ô77/77 = 2 to 6 %. The relative difference plot in Figure 3(b) shows that these estimated relative errors are consistent with the overall differences and the differences shown with the literature data, which is a little sparse at the high pressures and temperatures (larger densities). Those literature values
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
33
Figure 3. (a) Correlation plot of the isothermal viscometer calibration using C 0 2 at T = 0,50,100 and 150°C (p = 1 to 100 MPa); (b) difference plot for the experimental C 0 2 viscosities from this work and the literature, p is calculated from the Span and Wagner EOS;<16) r|(calc), Fenghour et al.;m) o, pooled Kestin et al. data;<24-27'40'4" A, Diller and Ball;<31) x, Iwasaki and Takahashi;(23) O , van der Gulik;<29) +, Michels et al.;ii2) □, Herreman et al.;m; *, Padua et al.fs) ; this work.
which show large relative deviations from each other (and the reference equation) are near the supercritical conditions. A similar but less significant increase in deviation is also observed with the C 0 2 densities differences in Figure 2(b).
34
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
As suggested these instruments are being used to provide acid gas and sour gas viscosities and densities beyond C0 2 . For an additional evaluation of the instruments at low density and viscosity, we have completed measurements for CH 4 over the same range of conditions. Figure 4 shows the difference between the measured CH 4 densities and the densities calculated using the Setzmann and Wagner EOS, [36] which has a stated accuracy of 0.07 % up to 50 MPa. Approximately 75 % of the measurements completed up to p -75 MPa are within the expected accuracy of 1.2 kg m 3 , based on the C 0 2 calibration. The relative difference between measured and calculated [43] viscosities for methane are shown in Figure 5. Like the C 0 2 calibration data, the estimated relative errors for the methane measurements ranged from Ô77/77 = 2 to 6%. Note that only 12 of the 30 CH 4 measurements were within stated viscosity range for the instrument. At the higher viscosities (up to 330 uP s) the estimated relative error is smaller while the larger relative errors, 4-6%, correspond to the small viscosities (77 < 20 uP s). The estimated error is increased at low viscosities because the cylinder travel time is very small, thus increasing the variance within averaged measurements. Density and viscosity measurements for H 2 S are still underway; however, we have shown some preliminary measurements for T = 100 and 150°C. As noted earlier, the literature data available
Figure 4. The difference between the experimental CH 4 densities and those calculated using Setzmann and Wagner. 06 ' Measurements were completed at T = 0,50,100 and 150°C and p = 1,2,5,10, 20,50 and 75 MPa
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
35
for H2S are sparse at high pressures. The difference between experimental densities [39,44-46] and those calculated from the reference equation of Lemmon and Span [47] are shown in Figure 6. For measurements below the supercritical density (pc = 347.28 kg m 3 ) the experimental values for this study are within the expected 1.2 kg m~3, assuming that Lemmon and Span is the most accurate EOS [47]. However, at higher densities (higher pressures) there is more variability for both these measurements and the literature data. In particular the values measured at T = 100.00±0.02°C are larger by as much as 10 kg nr 3 near the supercritical conditions (T = 99.9°C and pc = 9.0 MPa). Also shown in Figure 6, is the difference between the calculated densities of Sakoda and Uematsu [48] and those from Lemmon and Span [47]. The difference between these equations of state, also show a large deviation for T = 150°C. The data measured here seem to agree better with the Sakoda and Uematsu EOS and the experimental data of Liu [46]. The first two isotherms for H2S viscosity also have been completed. The relative differences between these viscosities and those calculated using Schmidt et al, [1] are shown in Figure 7. The estimated relative errors for these measurements are Ô77//7 = 2 to 5%. Again, the larger estimated error is for low viscosities where the
Figure 5. The relative difference between experimental and calculated CH 4 viscosities. Calculated viscosity from Quinones-Cisneros et al.;í43) •, this work (T = 0,50,100 and 150°C; p = 1,2,5,10,20,50 and 75 MPa); A , Stephan and Lucas;(49) □, Gulik et al.;m +, Schley et al.;m) x, Huang et al.f2) O , Ross and Brown;<53> *, Barua et a/.;<54) o, Giddings et al.m
36
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. The difference between experimental and calculated H2S denisities. Densities have been calculated using the Lemmon and Span EOS;<43) •, this work (T = 100°C; p = 1,2,5,10, 20,50 and 75 MPa); ♦, this work (T = 150°C; p = 1,2,5,10,20, 50 and 75 MPa);, Reamer et al.fu\ Imhels and Gmehling;(39) +, Lewis et al.;i45) x, Liu;<46) o, calculated at T = 100°C using Sakota and Uematsu;<48), calculated at T = 150°C using Sakota and Uematsu.' 48 '
Figure 7. The relative difference between some initial viscosity data for H2S and the equation of Schmidt et al.m ;T= 100°C; ♦, T = 150°C.
piston has a fast travel time. Figure 7 shows that data from both isotherms and show larger deviations when compared to Schmidt et al. [1] A discussion regarding these values and additional modeling efforts will be discussed in a subsequent study.
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
3.4
37
Conclusions
An Anton Parr high pressure densimeter and a Cambridge Viscometer have been evaluated by calibration with C 0 2 and measurements of CH 4 and H2S for T = 0 to 150°C and p = 1 to 100 MPa. While these instruments are not as precise or as accurate as some research grade instruments, they are commercially available and have been employed to make measurements over a very broad range of densities and viscosities for gaseous, liquid and supercritical fluids. A calibration procedure has been described. It was found that the factory settings for both these instruments were not accurate. The densimeter was found to be accurate to within 8p = 1.2 kg m~3 and the viscometer was accurate to within 877/77 = 2 to 6%. An expression was given for estimating the relative accuracy for viscosities based on the measurement variance and the accuracy and precision of the calibration. Finally our research is proceeding by measuring more pure H2S viscosities, some acid gas mixtures and sour gas mixtures. The properties can be of aid to the design of high pressure injection schemes.
References 1. K.A.G. Schmidt, S.E. Quinones-Cisneros, J.J. Carroll and B. Kvamme, Energy and Fuels 2008, 22, 3424-3434. 2. J.M. Monteil, F. Lazarre, J. Salvinien and P. Viallet, J. Chim. Phys. Phys.-Chim. Biol. 1969, 66,1673-1675. 3. H-J Ng, J.J. Carroll and J. Maddocks, Impact ofThermophysical Properties Research on Acid Gas Injection Process Design, Gas Processors Association Convention, Nashville, TN, (1999). 4. C E . Stouffer, S.J. Kellerman, K.R. Hall, J.C. Holse, B.E. Gammon and K.N. Marsh,/. Chem. Eng. Data 46,1309-1318 (2001). 5. R. Mireault, R. Stocker, D. Dunn and M. Pooladi-Darvish, Dynamics of Acid Gas Injection Well Operation, AGIS, Calgary, AB (2009)AGIS. 6. J.F. Ely, W M . Haynes and B.C. Bain, /. Chem. Thermodyn. 21, 879-894 (1989). 7. A. Fenghour, W.A. Wakeham and J.T.R. Watson, /. Chem. Thermodyn. 27, 219-223 (1995). 8. W.-W.R. Lau, C.-A. Hwang, J.C. Holste, K.R. Hall, B.E. Gammon and K.N. Marsh, /. Chem. Eng. Data 42,900-902 (1997). 9. J.C. Holste, K.R. Hall, P.T. Eubank, G. Esper, M.Q. Watson, W. Warowny, D.M. Bailey, J.G. Young and M.T. Bellomy, /. Chem. Thermodyn. 19,1233-1250 (1987). 10. C E . Stouffer, S.J. Kellerman, K.R. Hall, J.C. Holse, B.E. Gammon and K.N. Marsh, /. Chem. Eng. Data 46,1309-1318 (2001).
38
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
11. W. Duschek, R. Kleinrahm and W. Wagner, /. Chem. Thermodyn. 22, 827-840 (1990). 12. W. Duschek, R. Kleinrahm and W. Wagner, /. Chem. Thermodyn. 22, 841-864 (1990). 13. D.H. Strunk, J.W. Hamman and B.M. Timmel, /. Assoc. Anal. Chem. 62,653-658 (1979). 14. P. Picker, E. Tremblay and C. Jolicoeur, /. Sol. Chem. 3, 377-384 (1974). 15. H.J. Albert and R.H. Wood, Rev. Sei. Instrum. 55,589-593 (1984). 16. R. Span and W. Wagner, /. Phys. Chem. Ref. Data 25,1509-1596, (1996). 17. W.A. Wakeham, M.A. Assael, J.K. Atakinson, J. Bilek, J.M.N.A. Fareleira, A.D. Fitt, A.R.H. Goodwin and C.M.B.P Oliveira, Int. }. Thermophysics 28, 372-416 (2007). 18. G.G. Stokes, Mathematical and Physical Papers, Cambridge University Press, (1901). 19. A.E. Flowers, Am. Soc. Test. Mat. 14, 565-616 (1914). 20. A. Dandridge and D.A. Jackson, /. Phys. D: Appl. Phys. 14,829-831 (1981). 21. P. Daugé, A. Baylaucq, L. Marlin and C. Boned, J. Chem. Eng. Data 46, 823-830 (2001). 22. J B Irving and A J Barlow, /. Phys. E: Sei. Inst. 4, 232-236 (1971). 23. H. Iwasaki and M. Takahashi, J. Chem. Phys. 74,1930-1942 (1981). 24. J. Kestin and W Leidenfrost, Physica 25,1033-1062 (1959). 25. J. Kestin, J.H. Whitelaw and T.F. Zien, Physica 30,161-181 (1964). 26. J. Kestin, Y. Kobayashi and R.T. Wood, Physica 32,1065-1089 (1966). 27. J. Kestin and J. Yata, /. Chem. Phys. 49,4780-4791 (1968). 28. A. Padua, W.A. Wakeham and J. Wilhelm, Int. } . Thermophys. 15, 767-777 (1994). 29. P.S. van der Gulik, Physica A 238, 81-112 (1997). 30. W Herreman, W. Grevendonk and A. De Bock, /. Chem. Phys. 53, 185-189 (1970). 31. D.E. Diller and M.J. Ball, Int. J. Thermophys. 6, 619-629 (1985). 32. D.G. Friend, NIST Standard Reference Database 14, Version 9.08, Bolder, CO (1992). 33. W.E. Cole, High Pressure and Temperature compensation of the 440 Viscometer, Cambridge Viscosity, (2006). 34. A. Fenghour, W.A. Wakeham, V. Vesovic, /. Phys. Chem. Ref. Data 27, 31-34, (1998). 35. H. Preston-Thomas, Metrología 27, 3-10 (1990). 36. U. Setzmann and W. Wagner,/. Phys. Chem. Ref. Data 20,1061-1151 (1991). 37. A. Michels and C. Michels, Proc. R. Soc. London A 153, 201-214 (1935). 38. A. Michels, C. Michels and H. Wouters, Proc. R. Soc. London A 153, 214-224 (1935). 39. E.C. Ihmels and J. Gmehling, Ind. Eng. Chem. Res. 40,4470-4477 (2001). 40. J.D. Breetveld, R. DiPippo and J. Kestin, /. Chem. Phys. 45,124-126 (1966). 41. R. DiPippo, J. Kestin and K. Oguchi, /. Chem. Phys. 46,4758-4764 (1967). 42. A. Michels, A. Botzen and W. Schuurman, Physica 23, 95-102 (1957). 43. S.E. Quinones-Cisneros, M.L. Huber and U.K. Deiters, /. Phys. Chem. Ref. Data, 2007. 44. H.H. Reamer, B.H. Sage and W.N. Lacey, Ind. Eng. Chem. 42,140-143 (1950)
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
39
45. L.C. Lewis and W.J. Fredericks, /. Chem. Eng. Data 13,482-485 (1968). 46. C.H. Liu, Experimental densities, entropies and energies for pure H2S and equimolar mixtures of H2S/CH4 and H2S/C02 between 300 and 500 K.M.S. Thesis, Texas A&M University, College Station, TX, 1985. 47. E.W. Lemmon and R. Span, /. Chem. Eng. Data 21, 785-850, 2006. 48. N. Sakoda and M. Uematsu, Int.}. Thermophys. 25, 709 (2004). 49. K. Stephan and K. Lucas, Viscosity of Dense Fluids, Plenum Press, NY, (1979). 50. P.S. van der Gulik, R. Mostert and H.R. Van den Berg, Fluid Phase Equilibria 79, 301-311 (1992). 51. P. Schley, M. Jaeschke, C. Küchenmeister, Int. ]. Thermophysics 25, 1623-1651 (2004). 52. E.T.S. Huang, G.W. Swift and F. Kurata, AIChE Journal 12, 932-936 (1966). 53. J.F. Ross and G.M. Brown, Ind. Eng. Chem. 49,2026-2033 (1957). 54. A.K. Barua, M. Afzal, G.P. Flynn and J. Ross, /. Chem. Phys. 41,374-378 (1964). 55. J.G. Giddings, J.T.F. Kao and Riki Kobayashi, /. Chem. Phys. 45,578-586 (1966).
This page intentionally left blank
4
Acid Gas Viscosity Modeling with the Expanded Fluid Viscosity Correlation H. Motahhari, M.A. Satyro, H.W. Yarranton Department of Chemical and Petroleum Engineering, University of Calgary, Calgary, AB, Canada
Abstract The recently developed Expanded Fluid viscosity correlation was adapted to predict the viscosity of acid gases. The correlation was originally developed to predict hydrocarbon viscosities over a broad range of pressures and temperature as a function of fluid densities estimated by the Advanced Peng-Robinson equation of state. The correlation has two adjustable parameters per component: a compressed state density, p°, and an empirical parameter, cr In this study, the correlation was fitted to experimental viscosities for a variety of non-hydrocarbons relevant to acid gas injection activities including water, hydrogen sulfide, and carbon dioxide, with an overall average absolute relative deviation under 4%. The method was used to predict the viscosities of the mixtures, including binaries of C 0 2 with methane or ethane, H2S and methane, and a ternary mixture of H2S, C 0 2 and methane, with an overall average absolute relative deviation under 7% without the use of any adjustable interaction parameters. The proven simplicity and consistency of the method across the critical point and the ability to model mixtures makes the correlation a promising tool for enhancing the quality of process and reservoir simulations.
4.1
Introduction
Sour natural gas mixtures are u s u a l l y processed to separate the acid gases in order to m e e t transport a n d sales specifications. The acid gas stream from the gas treatment process consists of Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (41-54) © Scrivener Publishing LLC
41
42
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
hydrogen disulfide, carbon dioxide as well as small amounts of the light hydrocarbons and water. In many field applications, the acid gases are compressed and re-injected into depleted oil reservoirs or deep saline aquifers for storage and disposal. The accurate design and operation of such processes requires the prediction of the phase behavior of these mixtures as well as their physical and transport properties. This study focuses on viscosity and presents a tool to model the viscosity of acid gases and single phase mixtures of hydrocarbons, water, and acid gases.
4.2
Expanded Fluid Viscosity Correlation
The Expanded Fluid viscosity correlation was developed by Yarranton and Satyro [1], based on the observation that as a fluid expands its fluidity (inverse of the viscosity) increases. A compressed state of the fluid was defined at which the molecules are too closely packed to flow; that is the viscosity becomes infinite. The viscosity was then related to the expansion of the fluid from this state. The framework of the correlation was initially grounded in experimental viscosity and density data and tested for pure hydrocarbons and their mixtures. To eliminate the need of experimental density data for viscosity predictions, the correlation was modified for use with a density model [2]. The Advanced Peng-Robinson (APR) equation of state [3] was selected for the density model because: it can accurately predict the density of non-polar and polar components in gas and liquid states; it is consistent over the critical point and phase transitions; it has been tuned based on an extensive library of interaction parameters used to model vapour-liquid equilibrium mixtures commonly found in the natural gas industry. Note, although any density model could be used for this purpose, the correlation parameters are dependent on the choice of density model. Detailed development of the correlation is explained elsewhere [1,2] and the equation of state version of the correlation is summarized below. The viscosity of the fluid (u in mPa.s) is correlated to the fluid expansion in a form of a viscosity departure function as follows: ju-ßG
= 0.4214 (exp{c2ß}-\)
(1)
ACID GAS VISCOSITY MODELING
43
where ]iG is the dilute gas viscosity of the fluid, c2 is a fitting parameter, and ß is given by:
ß=-
(2)
( ~° \0A872 -1
expxP
j
where p is the density of the fluid predicted by APR equation of state and ps" is the compressed state density. To improve the predictions at higher pressures, a pressure dependency was introduced the compressed state density as follows:
l-C4(l-exp(-c3P))
(3)
where p° is the compressed state density of the fluid in vacuum, P is pressure in kPa, and c3 and c4 are two pressure dependency parameters given by: c, =1.435-lCr 6 MW a 4 2 6 7 [kPa"1] for MW< 97 g /mol:
c4 =0.015 + 0.00042|50-MW|
for M W > 97 g I mol : c4 = 0.035
(4) (5) (6)
The following correlation is used for the dilute gas viscosity: ßG=A
+ BT + CT2
(7)
where T is the temperature in K and A, B and C are fitting parameters and the values were taken from [4]. The remaining parameters, p° and c2, are the characteristic adjustable parameters for each fluid and are determined from the experimental viscosity data of each fluid. Numerical values of these parameters for 39 hydrocarbons were provided by Satyro and Yarranton [2]. As an example, the correlated viscosity of ethane is compared with experimental data from the NIST database [5] in Figure 1. The average absolute relative deviation (AARD) between the correlation and the data are 2.6%. A summary of the correlation
44
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. Comparison between the correlated and experimental viscosity data [5] for ethane: (a) saturated liquid and gas (b) supercritical fluid.
Table 1. The adjustable parameters of the correlation for use with APR equation of state, reproduced from [2]. p\ (kg/m3)
Component
C
2
AARD%
Methane
504.33
0.0221
5.8
Ethane
653.71
0.0321
2.6
Propane
698.73
0.0354
5.0
n-Butane
721.67
0.0351
4.6
performance and its adjustable parameters for light n-alkanes is given in Table 1.
4.2.1
Mixing Rules
The following volumetric mixing rules are used to calculate the adjustable parameters (c2mix and p°mi]) of the mixture:
(8)
~2,mix i=\
k
k
(9) i=l ;=1
ACID GAS VISCOSITY MODELING
45
where k is the number of components in the mixture and
(10)
/w=X4rr-
where ¡uG i and x{ are, respectively, the dilute gas viscosity and mole fraction of component i and:
[
/
-.0.5 /
x 0.25 "I 2
[8(I+MW;/MW,.)J The correlation has two inputs (the density and the pressure) and two parameters for each component (c2 and ps°). The density is calculated with the APR equation of state using the pure component properties including molecular weight, critical properties, acentric factor, and the volume translation factor, which were taken from the VMGSim pure component database [3].
4.2.2
Modification for Non-Hydrocarbons
In order to use the correlation for acid gas injection applications, it must be adapted for typically encountered non-hydrocarbons including water, hydrogen sulfide, and carbon dioxide. Experimental viscosity data for these compounds were used to determine the adjustable parameters of the correlation. Data were extracted from the NIST database [5] and the open literature, Table 2. Preliminary tests showed that the unmodified correlation failed to fit viscosity data for non-hydrocarbons such as water. The original correlation was developed only for hydrocarbons and does not account for the influence of non-dispersion forces such as hydrogen bonding. To overcome this problem, a temperature
46
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
dependency was introduced into the compressed state density as follows: P,° = ^,273.15 (1 + C r ( r - 2 7 3 . 1 5 ) )
(12)
where p°27315 is the compressed state density of the fluid in a vacuum at 273.15 K and T is the temperature in Kelvin. The correlation was fit to experimental viscosity data for each non-hydrocarbon by adjusting the parameters c2 and p°27315 as well as cT if required. The values of the parameters are provided in Table 3 for each compound.
Table 2. Summary of the experimental viscosity data of pure non-hydrocarbon components used for the analysis. Component
No. of Data Pts.
T Range (K)
P Range (kPa)
1795
256 -1139
0.6 - 389000
[5]
18
190 - 483
31 - 4370
[9-15]
520
203 -1100
19 - 453000
[5,16,17]
Water Hydrogen Sulfide Carbon Dioxide
Source
Table 3. Pure component regression summary. Component
C
2
C
4
3
(kg/m )
cT (x 103) (1/K)
AARD %
MARD %
3.8
25
Water
1127.39
0.0981
Eq. (5)
Hydrogen disulfide
1092.76
0.0437
Eq. (5)
0
3.3
13
Carbon Dioxide
1573.93
0.0646
Eq. (5)
0
6.0
16
Carbon Dioxide
1572.84
0.0627
Eq. (5)
3.0
12
Carbon Dioxide
1469.16
0.0519
0.0733
3.5
10
1.781
0.518 0
ACID GAS VISCOSITY MODELING
4.3
47
Results and Discussion
4.3.1 Pure Components As mentioned previously, a temperature dependent compressed state density was required to fit the water viscosity data, Figure 2, with an AARD and MARD of 3.8 and 25%, respectively. The maxim u m deviations occurred near the critical region and for compressed liquid below 0°C. These conditions are not relevant for acid gas injection applications. For carbon dioxide, the unmodified correlation provided a reasonable average fit to the experimental data with AARPD and MARPD values of 6 and 16, respectively. However, since the higher deviations occurred consistently at the high pressure data points suggesting that the correlations of the pressure dependency parameters (c3 and c4) do not apply for carbon dioxide. Carbon dioxide has a quadrupole moment and it is not surprising that it does not conform exactly to a correlation developed for hydrocarbons. To improve the fit, the following two modifications were evaluated: 1. retune parameter c4 to fit the correlation to high pressure data. 2. introduce a non-zero value of cT to fit the correlation to data.
Figure 2. Comparison between correlated and experimental viscosity [5] of water: (a) saturated liquid and atmospheric pressure (b) subcooled liquid conditions.
48
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The numerical values of the parameters for each approach and the errors of the correlation predictions are summarized in Table 3. With both approaches, the AARD was reduced to less than 4% which is within the experimental error of data (average relative uncertainty of 3.5%) [5]. The retuned c4 parameter provided the best fit to the high pressure data but a somewhat poorer fit at lower pressures. It also would require a modified mixing rule. Therefore, we chose to proceed with the temperature dependent compressed state density which required no modification to the mixing rules. Figure 3 compares the correlated viscosities from this approach to the experimental data. For hydrogen sulfide, a zero value of cT was set because the available data were insufficient to test the limits of the unmodified correlation, particularly at high pressure. The resulted average AARD and MARD for H2S were 3.3 and 13%, respectively. Figure 4 compares the correlation with measured viscosities along the saturation curve and at atmospheric pressure. Note, the simulated data points on saturation curve are from a molecular dynamics simulation [15]. The simulated saturated gas data were inconsistent with the experimental dilute gas data and were excluded from the analysis.
4.3.2
Acid Gas Mixtures
To study the accuracy of the correlation predictions for acid gas mixtures, experimental mixture viscosity data for these systems are
Figure 3. Comparison between correlated (with temperature dependent compressed state density) and experimental viscosity [5,16,17] of carbon dioxide: (a) saturated liquid and atmospheric pressure; (b) compressed liquid and supercritical conditions.
ACID GAS VISCOSITY MODELING
49
Figure 4. Comparison between correlated and experimental viscosity [9-15] of hydrogen sulfide (saturated fluid and atmospheric pressure data).
required. Unfortunately viscosity data for acid gas mixtures are rare. An extensive literature survey revealed a few experimental viscosity data points related to this application which are listed in Table 4. The viscosity correlation was applied to this dataset using the mixture density from the APR equation of state and the parameter mixing rules given in Equations 8 to 11. The C7+ fractions of the natural gas mixtures were modeled as n-C12 based on their molecular weight. In all cases, /?. was set to zero and therefore the correlation of mixture viscosities was completely predictive. The AARD and MARD for the correlation are summarized in Table 4. The correlation predicted the viscosity of binary mixtures of carbon dioxide or hydrogen disulfide with light n-alkanes within experimental error. Figure 5 provides an example of the viscosity predictions for a binary mixture of 53.4 mole% carbon dioxide and 46.6 mole% methane over a range of temperatures and pressures. The correlation smoothly follows the transition from the dilute gas to dense fluid phase conditions. Note that all data for the binary and ternary mixtures [19] containing H2S come from molecular dynamics (MD) simulations. The AARD and MARD for these mixtures were calculated by comparing the correlation predictions to the mean of the predicted viscosity values by different intermolecular potential fields and MD simulations. The MD predictions were scattered with an average and maximum deviations of 2.9% and 7.2% around the mean values.
0
17.2 9.4
3.1 8.7 6.1 54.5 63.5
49.4
70
28.2 0 0.3
Sour Natural Gas
Sour Natural Gas
Sour Natural Gas
Sour Natural Gas
Sour Natural Gas
33.3
11.7
28.9
377
310.9
394.3
352.6
322
352.6
35.5
0.5
22.6
290 - 350
350
350
210 - 320
323 - 474
T range (K)
20 -125
Sour Natural Gas
H2S+ C0 2 +CH 4
0.1
101
0.27
4
0.4
2
H,S +CH
44.1
2.1 - 37
25.2 - 75.7 0
3.4 - 70
P range (kPa)
24.5 - 74
C02
0.6
0
0
H2S
Mole Fr. (%)
H2S +CH4
6
206
CO,+C,H,
¿ o
132
C0 2 +CH 4
¿
No. of Data Pts.
Mixture
Table 4. Summary of the experimental viscosity data used in the analysis
-
-
-
-
-
-
5.9
-
-
6.7
1.8
AARD %
39.4
2.5
42.1
5.8
[8]
[8]
[8]
[8]
[8]
[8]
1.1 5.8
[19]
[19]
[19]
[18]
[7]
Source
7.7
3.2
1.6
19.4
8.2
MARD %
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
ACID GAS VISCOSITY MODELING
51
Figure 5. Measured and predicted viscosity of a binary mixture of 53.4 mole% carbon dioxide and 46.6 mole% methane at several temperatures and pressures. Data from [7].
The correlation was tested on data for six sour natural gases from Elsharkawy [8] all of which contained at least 23 mole% acid gas. The predicted viscosities were within 6% of the measured viscosity with two exceptions. The two exceptions were gases with more than 5 mole% C7+ and the correlation underestimated the viscosity by approximately 40%. The large error is probably related to the simple lumping of the plus fraction to the equivalent paraffin by molecular weight. The fit could be improved if aromatic/ naphthenic contributions were accounted for as well. For acid gas applications, the effect of water content on viscosity of gas mixtures can be significant. It is also useful for reservoir engineers to have a predictive tool to model the viscosity of the acid gas injected underground and dissolved in brine. Unfortunately, no data were found for wet acid gases or sour water in the open literature. In the absence of such data, the expanded fluid-based viscosity model along the adjustable parameters of the components determined above can be used to obtain a preliminary estimate of the viscosity of such systems. However, the accuracy of the predictions is not known. Experimental viscosity data of these systems are required to study the validity of the model predictions and further improve its performance.
52
4.4
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Conclusions
The expanded fluid based viscosity model developed originally for hydrocarbons was adapted for non-hydrocarbons including water, carbon dioxide and hydrogen sulfide. A temperature dependency term was introduced into the compressed state density to account for hydrogen bonding. With this modification, the correlation fit the viscosity data for the pure components with an overall AARD of less than 4%. Note that there was little experimental data point for hydrogen sulfide, see Table 2, and it could not be determined whether the correlation required modification for this compound. Therefore, the adjustable parameters for hydrogen sulfide were determined without using the temperature dependency term. The correlation predicted the viscosity of mixtures of light hydrocarbons with the acid gases with an overall AARD of less than 7%. The standard mixing rules for the correlation were used with no adjustment or additional parameter. The expanded fluid-based viscosity model and its mixing rules provide a simple and consistent tool for the prediction of the hydrocarbons and non-hydrocarbons encountered in acid gas injection applications. Its applicability for both and liquid phases as well as its fast execution in computer codes make it a suitable tool for the simulators.
4.5
Acknowledgements
The authors are grateful to Virtual Materials Group Inc. for the use of VMGSim and VMGThermo during the development of this work. We also thank John Carroll for his insight on the availability of experimental viscosity data for hydrogen sulfide.
References 1. Yarranton, H.W.; Satyro, M.A. Ind. Eng. Chem. Res. 2009,48, 3640-3648. 2. Satyro, M.A.; Yarranton, H.W. Fluid Phase Equilibria 2010, doi:10.1016/j. fluid.2010.06.023 3. Virtual Materials Group Inc., VMGSim Version 5.0 User's Manual, Calgary, Alberta (2009)
ACID GAS VISCOSITY MODELING
53
4. Yaws, C.L., "Chemical Properties Handbook: Physical, Thermodynamic, Environmental, Transport, Safety, and Health Related Properties for Organic and Inorganic Chemicals", McGraw-Hill (1999). 5. NIST Standard Reference Database; NIST/TRC Source Database; WinSource, Version 2008 6. Wilke, C.R. /. Chem. Phys. 1950,18,517-519. 7. DeWitt, K.J.; Thodos, G. Can.}. Chem. Eng. 1966,44(3), 148-151. 8. Elsharkawy A.M. Pet. Sei. Tech. 2003, 21(11 and 12), 1759-1787. 9. Rankine, A.O.; Smith, C.J. Philos. Mag. 1921, 42, 615-620. 10. Pal, A.K.; Barua, A.K. Trans. Faraday Soc. 1967, 63, 341-346. 11. Bhattacharyya, P.K.; Ghosh, A.K.; Barua, A.K. /. Phys. B, 1970, 3,526-535. 12. Pal, A.K.; Bhattacharyya, P.K. /. Chem. Phys. 1969, 51, 828-831. 13. Bhattacharyya, P.K., /. Chem. Phys. 1970, 53,893-895. 14. Runovskaya, I.V.; Zorin, A.D.; Devyatykh, G.G. Russ. /. Inorg. Chem. 1970,15, 1338-1339. 15. Nieto-Draghi, C ; Mackie, A.D.; Bonet A.J. /. Chem. Phys. 2005, 123, 014505-1-014505-8. 16. Padua, A.; Wakeham, W.A.; Wilhelm, J. Int. J. Thermophys. 1994,15(2), 767-777. 17. van der Gulik, P.S. Physica A, 1997, 238, 81-112. 18. Diller, D.E.; Van Poolen, L.J.; dos Santos, F.V., J. Chem. Eng. Data, 1988, 33, 460-464. 19. Galliero, G.; Nieto-Draghi, C ; Boned, C ; Avalos, J.B.; Mackie, A.D.; Baylaucq, A.; Montel, F. Ind. Eng. Chem. Res. 2007,46,5238-5244.
This page intentionally left blank
5
Evaluation and Improvement of Sour Property Packages in Unisim Design Jianyong Yang, Ensheng Zhao, Laurie Wang, and Sanjoy Saha Honeywell, Calgary' AB, Canada
Abstract UniSim® Design is a commonly used process simulation tool for the oil and gas processing industry. Typically, the sour Peng-Robinson (PR) and sour Soave-Redlich-Kwong (SRK) models in Unisim Design are chosen by users for handling process systems where H 2 S, C 0 2 or NH 3 are in contact with an aqueous phase. These sour options combine the PR/SRK equation of state and Wilson's API-Sour Model where the equation of state is used to determine the fugacities of the components in vapor and liquid hydrocarbon phases, as well as the enthalpy for all three phases, while the Wilson's API-Sour method is used for the aqueous phase equilibrium calculations, which accounts for the ionization of H2S, C 0 2 and NH 3 in the aqueous phase. The scope of this work includes an evaluation of the efficiency of these two thermodynamic property packages through comparison with industrial and experimental data. The evaluation results will provide a baseline for the enhancements of these property packages for more accurate and reliable modeling of sour systems.
5.1
Introduction
U n i s i m Design is a widely u s e d simulator for oil a n d gas p r o cessing industry. M a n y t h e r m o d y n a m i c m o d e l s are integrated in UniSim Design to h a n d l e a b o a r d r a n g e of process systems. The s o u r Peng-Robinson (Sour PR) a n d sour Soave-Redlich-Kwong (Sour SRK) packages w e r e d e v e l o p e d specially for m o d e l i n g t h e process systems w h i c h involve s o u r gases. These specialty m o d e l s Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (55-64) © Scrivener Publishing LLC
55
56
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
combine the PR/ SRK Equation of State (EOS) and Wilson's API-Sour Model [1] where the equation of state is used to determine the fugacities of the components in vapor and liquid hydrocarbon phases, as well as the enthalpy for all three phases, while the Wilson's API-Sour method is used for the aqueous phase equilibrium calculations, which accounts for the ionization of H 2 S, C 0 2 and NH3 in the aqueous phase. These sour property packages were developed many years ago. This work is to re-evaluate the accuracy of these models and their performance in UniSim Design.
5.2
Model Description
The sour property packages can be selected in UniSim Design, as shown in Figure 1. Once components are attached to the property package, the thermodynamic engine is ready to perform flash and physical property calculations of the defined system. Since this study is focused on the application of the sour property packages in UniSim Design, the scope of our discussion will be dedicated to the special treatments applied in UniSim Design for the implementation of these models. A full description of the Wilson API-Sour Model can be found in the literature [1]. Eqn. (1) from the Steam and Gas Tables by Irvine and Liley [2] is used to calculate the vapor pressure of water instead of the A.S.M.E Steam Tables.
Figure 1. The interface of sour PR in unisim design.
S O U R P R O P E R T Y PACKAGES I N U N I S I M D E S I G N E S
57
( 9
In pw =6.9078 +
no 5>r + T-a 1=0
(1) n
where p*w is the vapor pressure of water, T is the temperature and a. is the coefficient with the values listed in Table 1. In the liquid-liquid or vapor-liquid-liquid equilibrium, algorithm considers the heaviest phase as the aqueous phase only when mole fraction of water, xlw, is greater than 0.5. To ensure the model continuous at x'w - 0.5, a transition range is created where 0.4 <xw< 0.6. If x'w < 0.4, the fugacities of all components are calculated by EOS, and if xlw > 0.6, the API model is used. In this transition range, the fugacities of all the components are corrected by / = has + (fAPI -ÍEOS )• «
-0.4)/(0.6-0.4)
(2)
where / is the fugacity of a compound and subscripts EOS and API represent for the fugacity of the compound calculated by the equation of state and the API model, respectively. The default binary interaction parameters (BIP) in UniSim Design for the EOS part are shown in Table 2. The BIPs between H 2 0 - C 0 2 and H 2 0-H 2 S are the same as those used by Oellrich et al. [3]. The other constants such as Henry constants and chemical equilibrium constants of all components are taken from Wilson's work [1]. Although some of these parameters are user-tunable, the default values are used in this study. Table 1. Parameters in vapor pressure equation of water. i
ai
0
10.4592
6
9.0367E-16
1
-4.0490E-03
7
-1.9969E-18
2
-4.1752E-05
8
7.7929E-22
9
1.9148E-25
3
3.6851 E-07
i
«
,
■
4
-1.0152E-09
10
-3.9681E+03
5
8.6531E-13
11
39.574
58
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 2. Default interaction parameter in unisim design for the EOS part. H20-C02
-0.5572 + 0.001879 x T-1.274 x 10' 6 P
H 2 0-H 2 S
-0.3897 + 0.001565 x T - 1.142 xl0' 6 P
H 2 0-NH 3
-0.2533
NH3-CO3
0.0
NH 3 -H 2 S
0.0
C0 2 -H 2 S
0.1
5.3
Phase Equilibrium Calculation
To evaluate the applicability of the sour property packages, the results from UniSim Design are compared with experimental vapor-liquid equilibrium data from literature for different systems. Since the Sour PR and Sour SRK provide similar results under many conditions, only the results from the Sour PR property package are reported here. Experimental values of partial pressures of solutes in NH 3 water [1], NH 3 -C0 2 -water [4] and NH 3 -H 2 S-water [5] solutions are compared against the values predicted by the Sour PR model from UniSim Design in Figures 2, 3, and 4, respectively. The data points below 5 mmHg in Figure 2 and 5 kPa in Figures 3 and 4 are omitted because of the possible measurement error. It can be see that the prediction from Sour PR model fits the trend of experiment data very well, especially in the dilute region. With the molality of sour gases increasing, as illustrated in Figures 3 and 4, the total vapor pressure decreases first and reaches a minimum. At this minimum point, almost all of the NH 3 is converted into ammonium ions and the ions from the sour gases are saturated. Any excess sour gases injected would go into the vapor phase from this point on, resulting in a sharp increase in total pressure. This behavior of NH 3 solution can be utilized as a controllable operational variable for sour gas treating processes. Figure 5 shows the predicted molality ratios of C 0 2 and NH 3 at the vapor pressure inflection point under different concentrations and temperatures. These ratios are the measure of the captured
SOUR PROPERTY PACKAGES I N U N I S I M DESIGNES
59
Figure 2. Partial pressure of NH 3 in NH 3 -water solution. Experimental data: □, 293.15 K; A, 333.15 K; UniSim Design: —
Figure 3. Partial pressure in NH 3 -C0 2 -water solution at 313 K and mNH3=6 mol/kg. Experimental data: G, C0 2 ; A, NH 3 ; 0, total vapor pressure; UniSim Design: —
C 0 2 per mole of NH 3 . Figure 5 indicates that the amount of C 0 2 captured decreases as the temperature increases. Based on this phenomenon, it is desirable to establish a NH 3 recycle process to absorb C 0 2 at low temperature and to be regenerated at high temperature. This finding has been patented by Eli Gal [6]. Figure 5 also shows that high molality of NH 3 will increase the amount of C 0 2 captured due to higher rate of chemical equilibrium. This behavior has been confirmed by the experiments [4], [5]. At high temperature, impact
60
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 4. Partial pressure in NH 3 -H 2 S-water solution at 313 K and mNH3=3 mol/kg. Experimental data: □, H2S; A, NH 3 ; 0, total vapor pressure; UniSim Design: —
Figure 5. Molality ratio of C 0 2 and NH 3 at the inflection point. The curves from bottom to top: mNH3=3 mol/kg, wNH3=6 m o l / k g and mNH3=12 mol/kg.
of NH 3 concentration on C 0 2 absorption is more prominent and this phenomenon can help an engineer to select suitable process conditions to remove sour gases. Figure 6 shows the relative deviations of partial pressures between the calculated results and experimental data [1], [,4], [5] for ternary and quaternary sour water systems, where wt% is the total weight percentage of the solutes. This figure suggests that
SOUR PROPERTY PACKAGES IN U N I S I M DESIGNES
61
Figure 6. Relative deviation of partial pressure in ternary and quaternary solutions. D, C0 2 ; A, NH 3 ; 0, H2S.
Figure 7. Predicted vapor compositions of C 0 2 and H2S under different temperatures and pressures by Sour PR property package.
the Sour PR property package models dilute sour gas solutions more accurately than that of high concentration solutions. A possible reason for poor prediction at high solute concentrations is that the parameters used in the API model may not have been fitted with experimental data in these regions. Figure 7 shows the predicted compositions of C 0 2 and H2S in vapor phase with the temperature range from 30 to 140°C and
62
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
pressure range from 100 to 260 kPa for a C0 2 -H 2 S-NH 3 -H 2 0 system. The overall compositions in mole fraction are xNH =0.06, x co2 = 0-03 a n c [ xHS= 0.03. It can be seen that C 0 2 has higher solubility than H2S at the low temperature region. As temperature increases, less C 0 2 is absorbed in aqueous phase. Between 50~70°C, the composition of C 0 2 reaches a maximum. Beyond this temperature, NH 3 and water starts to vaporize and therefore absorption of C 0 2 and H2S in aqueous decrease sharply.
5.4
Conclusions
The results obtained by using UniSim Design Sour PR property package in UniSim Design for sour water system are compared with the experimental data. The evaluation indicates: 1. Sour PR property package can predict the phase behavior quite well at low solute concentration, which is more common in Oil and Gas processing where acid gases are removed 2. Although the deviation at the higher solute concentration region is little high, it can be further reduced by tuning the model parameters. 3. The solubility of C 0 2 and H2S in NH 3 solution under different operating conditions is studied. It provides solid rationale for the NH3 recycle process of the sour gas capture.
5.5
Future Work
The natural extension of this study will lead to further enhancements in UniSim Design through 1. Updated model parameters from experimental data, especially in the high concentration region, and 2. Enabled more user-tunable parameters to better represent the real plant conditions.
S O U R P R O P E R T Y PACKAGES I N U N I S I M D E S I G N E S
63
Reference 1. G.M. Wilson, "A New Correlation for NH3, C02, H2S Volatility Data from Aqueous Sour Water System", API Pub. No. 955, American Petroleum Institute, Washington, DC, 1978. 2. T. Irvine, and P. Liley, "Steam and Gas Tables with Computer Equations", New York, Academic Press, p. 21,1984. 3. L, Oellrich, U. PlOcker, J. M. Prausnitz, and H. Knapp, "Equation-of-State Methods for Computing Phase Equilibria and Enthalpies", Internat. Chem. Eng., Vol. 22, P. 1,1981. 4. F. Kurz, B. Rumpf, and G. Maurer, "Vapor-liquid-solid Equilibria in the System NH 3 -C0 2 -H 2 0 from Around 310 to 470 K: New Experimental Data and Modeling", Fluid Phase Equilibria, Vol. 204, P. 261,1995. 5. B. Rumpf, A. Pérez-Salado Kamps, R. Sing, and G. Mauer, "Simultaneous Solubility of Ammonia and Hydrogen Sulfide in Water at Temperatures from 313 K to 393 K", Fluid Phase Equilibria, Vol. 158-160, P. 293,1999. 6. E. Gal, Ultra, "Cleaning combustion gas including the removal of C0 2 ", World Intellectual Property, Patent WO 2006022885,2006.
This page intentionally left blank
6 Compressibility Factor of High C0 2 Content Natural Gases: Measurement and Correlation Xiaoqiang Bian, Zhimin Du, Yong Tang, and Jianfen Du The State Key Laboratory of Oil & Gas Reservoir Geology and Exploitation Engineering, Southwest Petroleum University, Chengdu, People's Republic of China
Abstract The JEFRI-PVT apparatus made in Canada by Schlumberger was used to obtain accurate compressibility factor measurements for high C0 2 content natural gases to study the effect of different C 0 2 content on gas compressibility factors (range covered: temperature, 263.15K to 313.15K; pressure, 3 MPa to 15 MPa). The results showed that gas compressibility factors reduce with increasing CO z content in natural gases and increase with increasing temperature. In addition, a non-integral power polynomial correlation was proposed without an iterative procedure whose coefficients were determined by fitting experimental data. The mixing rules used include: Kay's mixing rule combined with Wiehert-Aziz and Casey correlations (Kay) and Stewart-Burkhardt-Voo mixing rule with Wichert-Aziz and Casey (SBV). Comparison of the DAK-SBV, DAK-Kay, and proposed correlations showed that the presented model yielded the most accurate predictions with the lowest average absolute deviation (0.42%) among them.
6.1
Introduction
In recent years, m o r e a n d m o r e h i g h C 0 2 - c o n t e n t gas reservoirs in the w o r l d h a v e been discovered (Jokhio et al., 2001). Different C 0 2 content leads to different gas compressibility factor w h i c h is
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (65-86) © Scrivener Publishing LLC
65
66
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
involved in calculating gas properties such as formation volume factor, density, compressibility, and viscosity (Shokir, 2008). All these properties are necessary in the natural gas industry for evaluating newly discovered gas reservoirs, calculating initial and gas reserves, and predicting future gas production. Prediction of gas compressibility factor contains four methods: experiment (Adisoemarta et al, 2004), equation of state (Li and Guo, 1991; Elsharkawy, 2002), corresponding state correlations (Riazi, 2005), and empirical formula (Li et al., 2001). Experimental measurement of gas compressibility factor is the most accuracy among all methods, but it is time-consuming and costly. Several correlations fitting the Standing and Katz chart (Standing and Katz, 1942) can be used to calculate the gas compressibility factor (Papay, 1968; Hankinson et al., 1969; Hall and Yarborough, 1973; Yarborough and Hall, 1974; Dranchuk et al., 1974; Brill and Beggs, 1974; Dranchuk and Abou-Kassem, 1975). But DranchukAbou-Kassem (DAK) correlation is the most accurate representation of Standing and Katz chart. In addition, among the mixing rules (Sutton, 1985; Elsharkawy and Elkamel, 2001; Bahadori et al., 2007), Kay's (1936) mixing rule and Stewart-Burkhardt-Voo (1959) are the most widely used. Kay's mixing rule is simple. StewartBurkhardt-Voo (SBV) rule provided the most satisfactory results (Satter and Campbell, 1963). Since the presence of C 0 2 gas, prediction of the compressibility factor is much more difficult than that of sweet gases. Wiehert and Aziz (1972) presented corrections for the presence of H2S a n d / o r C 0 2 for determining compressibility factor of sour gases. Casey (1990) proposed correlations for the presence of nitrogen (N2) and water vapor (H 2 0) to correct the pseudo-critical properties. Fortunately, in this study, due to the absence of C7+ fraction, the critical properties of the C7+ fraction are not calculated from correlations (Keseler and Lee, 1976; Pedersen et al., 1989). This study has two objectives. The first objective is to measure gas compressibility factor for different C0 2 -content natural gases (C0 2 content is about 10%, 30% and 50%, respectively) by applying JEFRI-PVT experimental equipments which were made in Canada by the Schlumberger company. The second objective is to develop a non-iterative empirical correlation to estimate gas compressibility factor based on the experimental results, and to make a comparison among the proposed model, DAK-SBV and DAK-Kay correlation.
MEASUREMENT AND CORRELATION
6.2
67
Experiment
6.2.1 Measured Principles For dry gases, compressibility factors can be calculated using the following equation:
z=PAVT-
(i)
P.-V.-T where P is experimental pressure, P s ambient pressure, T experimental temperature, Ts ambient temperature, AV volume of gas bled from the PVT vessel, and Vs is the volume of gas released at ambient pressure and temperature.
6.2.2 Experimental Apparatus and Procedure In this study, mercury free DBR-PVT vessels made in Canada were used to measure gas compressibility factors. A schematic diagram of the apparatus is shown in Figure 1. The experimental apparatus used consisted of PVT vessel of approximately 135 ml capacity, automatic pump, gas chromatography, dry gasometer, constant temperature air bath, flash separator and ground separator.
Figure 1. Schematic diagram of the experimental apparatus.
68
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The experimental procedure is as follows: 1. Clean PVT vessel and cells, then connect the PVT vessel to the cells and evacuate the cells; 2. Control and maintain the desired temperature using the constant-temperature air bath; 3. Introduce a test gas (about 100 ml) into PVT vessel at the specified temperature and pressures in the oven and keep 5 h, then hold up half an hour and measure the gas volumes of the PVT vessel; 4. Slightly open the valve between the cells and bleed gas from the PVT vessel into the flash separators, at the same time, keep the pressures of the PVT vessel constant using automatic pump; 5. Record the bled gas volumes and remaining gas volumes of PVT vessel; 6. Use Eq.(l) to determine Z-factors; Repeat the procedure (4) ~ (6), and ensure tested Z-factors are about the same at least three times. More detailed descriptions are given by Liu et al. (2002), SY/T 6434-2000, and Varotsis and Pasadakis (1996).
6.2.3 Experimental Results The gas samples used were analyzed using a HP-6890 Gas Chromatograph. The results of analysis were shown in Table 1. The critical pressure and temperature for the pure components normally present in natural gases are also provided in Table 1 (Lu, 1982; Reid et al., 1987). Experimental compressibility factors for different C0 2 -content natural gases were listed in Table 2. It was shown in Table 2 that compressibility factors reduce with increasing C 0 2 content in natural gases and pressures, but increase with increasing temperature.
6.3
Methods
6.3.1 Existing Methods When gas composition is available, pseudo critical properties are calculated using a given mixing rule. In this study, Kay's mixing
MEASUREMENT AND CORRELATION
69
Table 1. Composition of gas mixtures and critical properties of defined component. Component
Sample 1
Sample 2
Sample P/MPa C 3
T /K c
Mw /g-mol 1
Mole fraction C02
0.0984
0.2886
0.5099
7.384
304.21
44.01
N2
0.0205
0.0124
0.0083
3.400
126.20
28.013
Cl
0.8602
0.6773
0.4654
4.595
190.56
16.043
C2
0.0167
0.0159
0.0116
4.871
305.33
30.07
C3
0.0031
0.0042
0.0033
4.247
369.85
44.097
iC4
0.0004
0.0006
0.0005
3.640
407.85
58.123
nC4
0.0005
0.0007
0.0006
3.796
425.16
58.123
iC5
0.0001
0.0003
0.0002
3.381
460.43
72.15
nC5
-
0.0001
0.0001
3.369
469.71
72.15
C6
0.0001
-
-
3.012
507.37
86.177
Table 2. Experimental compressibility factors for high C0 2 natural gases. Pressure (MPa)
Sample 1 313.15K
303.15K
293.15K
283.15K
273.15K
263.15K
3.00
0.9423
0.9307
0.9196
0.9028
0.8959
0.8833
5.00
0.9128
0.8992
0.8803
0.8608
0.8483
0.8337
7.00
0.8901
0.8721
0.8518
0.8252
0.8037
0.7829
9.00
0.8711
0.8563
0.8280
0.7981
0.7719
0.7410
11.00
0.8607
0.8369
0.8079
0.7809
0.7495
0.7054
13.00
0.8500
0.8309
0.8013
0.7714
0.7324
0.6901
15.00
0.8452
0.8286
0.8001
0.7691
0.7303
0.6860
(Continued)
70
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 2. Experimental compressibility factors for high C 0 2 natural gases. (Continued) Pressure (MPa)
Sample 2 313.15K
303.15K
293.15K
283.15K
4.21
-
-
-
0.8892
-
-
4.50
-
-
0.8891
-
0.8505
-
5.00
0.9025
0.8941
0.8718
0.8595
0.8297
0.8066
6.00
0.8745
0.8605
0.8452
0.8344
0.7998
0.7600
7.00
0.8576
0.8358
0.8161
0.8087
0.7710
0.7202
9.00
0.8277
0.8010
0.7809
0.7524
0.7109
0.6589
11.00
0.7957
0.7661
0.7406
0.7075
0.6719
0.6137
13.00
0.7644
0.7429
0.7136
0.6743
0.6395
0.5854
15.00
0.7513
0.7290
0.6882
0.6594
0.6431
0.5877
Pressure (MPa)
273.15K 265.65K
Sample 3 313.15K
303.15K
293.15K
283.15K
275.85K
4.00
-
-
0.8859
0.8574
0.8386
4.50
0.9118
0.8927
-
-
-
5.00
0.9002
0.8710
0.8353
0.8099
0.7674
6.00
0.8627
0.8247
0.7961
0.7710
0.6992
7.00
0.8246
0.7893
0.7683
0.7335
0.6470
9.00
0.7630
0.7423
0.7017
0.6402
0.5576
11.00
0.7158
0.7052
0.6483
0.5653
0.4915
13.00
0.6787
0.6735
0.6117
0.5211
0.4771
15.00
0.6565
0.6507
0.5890
0.5246
0.4937
MEASUREMENT AND CORRELATION
71
rule, Stewart-Burkhardt-Voo (SBV) mixing rule are considered. Kay's (1936) mixing rule, based on molar weighted average critical properties, has the following form:
^ = £yA
(2)
r
(3)
¿=i
P c=Xy.
T
¿=i
d
Stewart-Burkhardt-Voo (1959) (SBV) proposed the following mixing rule for high molecular weight gases. ( *.T
yTc
O ¡=l
\
Í
2
+—
v^ *
«=£ i=i
3
1=1
rzr\
y*
(5)
y/Ve
T*pcVC=K2/J
(6)
Ppc = T pc ' / i PC
(4)
VC '
(7)
>
Eqs. (2) and (3) or (4) through (7) provide critical properties for sweet natural gas systems. For high C 0 2 natural gases, these equations must be corrected for the presence of non-hydrocarbon components. The method proposed by Wiehert and Aziz (1972) (WA) is used to correct the pseudo critical properties of natural gases to H2S a n d / o r C 0 2 components. The WA correlation is given as follows: £ = - [ l 2 0 ( A a 9 - A L 6 ) + 15(Ba5-B4)]
(8)
Where the coefficient A is the sum of the mole fractions of H2S and C 0 2 in the gas mixture and B is the mole fraction of H2S in
72
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
the gas mixture. The corrected pseudo critical properties P' and r/ Tpc
a r e :
T' =TM -E * pc
p/
pc
(9) 7
fVV Tpc+B(l-B)Ç
pc
(10)
Correlations proposed by Casey (1990) are available for correcting pseudo-critical pressure and temperature for the presence of N 2 and H 2 0. The correlations for nitrogen and water vapor are: T p c c 0 r =-136.72y N 2 +222.22y H 2 O
(11)
ppCrC0r=-l.U7yN2+8.756yH2O
(12)
r;c-126.22yN2-647.22yH20 pc
1
1
r
= pc
_
pcœr
yN2
yH,o
P;-3-40yN2-22.06yH2O -• i —
pc,cor
3/N, ~VH7O
where Tl and P' c are the pseudo-critical temperature and pressure correlated for H2S and C 0 2 with the WA correlation. Reduced pressure (P r) and reduced temperature (T r) are calculated from pressure (P) and temperature (T) of interest and critical properties of natural gases ( T"c, P"c ) by the following relationship:
p
''!
p P" pc T PC
(15)
(16)
MEASUREMENT AND CORRELATION
73
The DAK correlation is extensively adopted to calculate the gas compressibility factor (Z) using reduced pressure (Pr) and reduced temperature (T) as follows: A,
Z=l+
A,
A¿
A + A ! A2 tf
Ac Í
T
V
r. r
v
(17)
2 \
-A
A ! A P?+Ao(l + AlPr2) Ä T T L
eX
2
P(-A!Pr2)
r 7
where (18)
Pr
ZTL
The constants Ax through An in Eq. (17) are listed in Table 3: Because the gas compressibility factor appears on both sides of DAK's correlation, Eq. (17), an iteration solution is necessary. Newton-Raphson method is used as follows: :(n)
in) _ yi.n+1) _ y{n)
(19)
dF M where ( . Ar, A-, A, Ac ? Pr + A +— - + ^ r + ^ - + —f- P,* +
r
V
- A A ! A2 V
T
T
r
r
A
'
/ (
pr6 + A 1 0 (l + A n p 2 ) ^
A6+^+A
V 0
\
p; J
3\
exp(-A:ip2)-
0.27P
(20)
Table 3. Constants At throi.ighA M in Eq.(17).
0.05165
A A
0.5475
0.1056
Ao
0.6134
A A
0.3265
Aq
1.07
A A A
0.5339 0.7361 0.721
A A
0.01569 0.1844
74
CO
d¥{n)=\
SEQUESTRATION AND RELATED TECHNOLOGIES
+
I
T 1
r
+
T 4
ir
h
V+^+V % T, A
+
T"J
10
l
\ r
l
T$
xlpr
hr )
X «3öv — ^T.Q
+
2
x6pj?
(21)
Tt
3p^ + A,A?>pr -2A~~py\
exp(-A...pf I
j
(22)
p =0.27-ff-
where Z (n+1) and Z (n) are the new and old values of gas compressibility factors.
6.3.2 P r o p o s e d M e t h o d According to Maclaurin's expansion, the following equation can be obtained:
fix)
f (0)
f{n) (0)
2!
n\
( 23 ^
Eq.(23) is an integral polynomial. Independent variable x is transformed to the power function xa, which can improve the fitting accuracy greatly under the condition of the same number of terms (Sheng et al, 2004). That is, continuous function can be approximate to the following non-integral power polynomial:
g (x) « a0 + a^xb + a2x2h + ... + anxnb
(24)
MEASUREMENT AND CORRELATION
75
where b in Eq.(24) is an undetermined parameter and a0 is equal to g(0). If b = 1, Eq.(24) reduces to Eq.(23). In order to illustrate the rationality of Eq.(24), suppose b>0. Then differentiation of Eq.(24) with respect to x and letting x=l yield:
g'H) = bax + 2ba2 + 3ba3 + ... + nban g"il) = 1Kb - Da, + 2b(2b - \)a2 + ... + nb(nb - \)an (25) g{n\l)
= b(b-l)...[b-(n-l)]al
+
+ ... + nbiyib -l)...[nb-(n-
2b(2b-l)...[2b-(n-l)]a2 Y)]a„
The coefficients a.(i = l,2,---,n) can be determined through Eq.(25). In common, when fitting experimental data, let n=3, which can be sufficiently to satisfy an engineering requirement. Therefore, Eq.(24) can be rewritten as follows:
g(x) = g(0) + ^ 3 - [(6b2 - 5b + l)g'U) - (5b - 3)£"(1) + g'"(l)] xb + -^r-(3& 2 -4b + l)g'(l) + (4b-3)g"(l)-g'"(l)]x2h 2b
(26)
+ -L[(2b 2 -3b + l)g'(l) - 3(b - l)g"(l) + g'"0)] x3
When the conditions of the same number of terms between Eq.(23) and Eq.(24) apply, Eq.(23) becomes: f(x) = /(0) + /'(0)x + £ ^ x
2
+^-x3
(27)
In order to indicate the accuracy of Eq.(26) and Eq.(27), a random function y = (l +x)Z5 is taken for example (see Table 4). The average absolute deviations (AAD) and absolute relative deviations CARD) in the subsequent tables are, respectively, defined as:
0.8
4.3469
4.3600
4.3635
4.3457
4.3305
0.4
2.3191
2.3200
2.3365
2.3194
2.3046
y
f
g(b=0.92)
g(b=0.96)
g(b=\. 00)
X
0
5.6404
5.6556
5.6734
5.6875
5.6569
1
7.1624
7.1777
7.1954
7.2400
7.1789
1.2
10.8854
10.9004
10.9177
11.0800
10.9002
1.6
15.5841
15.5976
15.6121
16.0000
15.5885
2
Table 4. Comparison of the precision of Eq.(26) and Eq.(27) approximating to y=(\+x)2
21.3435
21.3523
21.3585
22.1200
21.3156
2.4
32.1569
32.1466
32.1195
33.8125
32.0000
3
0.256
0.086
0.282
1.718
/
AAD/%
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
MEASUREMENT AND CORRELATION
AAD(%)
=—£ ¡=i
ARD(%) ■■
Cal. - Exp. xlOO Exp.
Cal. - Exp. xlOO Exp.
77
(28)
(29)
It was shown in Table 4 that Eq.(26), whose AAD = 0.086% and b = 0.96, significantly improved the accuracy of the prediction of y = (1+x)25 as compared with Eq.(27) (AAD=1.718%). The proposed correlation of Eq.(26) was extended to apply to bi-variant one as follows: g(x, t) = a0 (t) + a, (t)xb + a2 (t)x2b + a3 (t)x3b a0 \t) = c00 + c10t + C2QI
+ c30t
flj (t) = C0/1 + C u f A + C2lt2A + C3At301
(30)
a2(t) = c02 + cX2tßl + c22t2ß2 + c22t3ßl a3(t) = c03+cl3t
+c23t
+c33t
Finally, a new correlation of gas compressibility factors for high C02-content natural gases can be obtained:
Z(Pr,Tr) = a0(Tr) + a,(Tr)Pra +a2(Tr)Pr2a +a3(Tr)P3a a0(Tr) =
Aa+BaTrß°+CXß°+DaTr3ßo
al(Tr) = Ab +
BhT^+CbT^+DbT3A
a2 (Tr ) = Ac + BcTf2 +CcT2ß2 +
D
Xßl
(31)
a3(Tr) = Ad + BdTrß>+CdT^+DdTr3ß>
where a ß. (i = 0,1,2,3) and A.,B¡,CjrD.{i-a,b,c,d) are undetermined coefficients which can be obtained by fitting experimental data.
78
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
6.5 Comparison of the Proposed Method and Other Methods The tuned coefficients of Eq.(31) can be obtained by fitting experimental data using MATLAB 7.0 (Table 5). But the tuned coefficients of the correlation needed refitting for different samples. Table 6 showed the absolute relative deviations of calculated compressibility factors of sample 2 and 3 by different methods. Figure 2(a-c) showed the predicted gas compressibility factors from the new model compared to the results of the DAK-SBV and DAK-Kay compressibility factor correlations versus the measured gas compressibility of the Sample 3. As can be seen from Table 6 and Figure 2, the proposed method was the most accurate of the three methods tested, giving an overall average absolute deviation of 0.42%. In the order of accuracy Table 5. The tuned coefficients of the proposed model for different samples. A
-8.816390
Ba
24.405726
a
\ Sample 1
A
12.520235
Bb
-31.429384
-5.306294
B
13.338221 c
0.747085
a
0.8
Aa
28.574773
\ Sample 2
A
ß«ßvß, B
65.179604
20.274849
B
-30.885941 C
1.1
A
79.235341
A
Bd
0.813848
c
c* 2.1
C
1.266354
Dd
20.748488
cb c
B
D
-29.974701
Dt
5.006821
14.296895
D
-2.482498 c
-166.38253
C
88.823846
-153.47652 C
28.118034
ßvßvßyßs
cd 2.5
0.4058196
2.3 D
-13.588054
-149.32077
Db
22.973990
83.105283
D
-12.845033
C
c
v*
Dd
a
a
B
-3.328087
a
ßvß,
cb c
0.0965256
2.5
A
2.2
72.278392
1.2
D
/VA 277.43880
a
-9.046675
-2.249317
B>
-13.225356
-2.280668
cd
4.837914
-130.68757
K
Db
c
a
C
21.474317
a
Bb
a
\
-45.647110
-41.125073
a
Sample 3
-1.874226
a
-3.301408
cb c
2.132880 a
c
B<
c
\
D
a
c
\
-16.945150
C
-15.304644
D,
2.374806
1.49
1.84
0.08
0.37
0.22
0.59
0.71
0.31
0.13
0.22
7.00
9.00
11.00
13.00
15.00
3.00
5.00
7.00
2.28
2.95
3.12
283.15K
1.15
0.14
0.40
0.91
0.02
5.00
0.96
313.15K
DAKSBV
0.02
This model
3.00
Pressure /MPa
1.96
2.74
3.01
2.27
1.88
1.48
0.41
0.20
0.78
0.88
DAKKay's
0.05
0.34
0.20
1.09
0.65
0.64
0.13
0.29
0.08
0.02
This model
2.14
2.58
2.90
273.15K
2.75
1.98
0.82
0.48
0.87
1.33
1.59
303.15K
Sample 1
DAKSBV
1.77
2.34
2.77
3.23
2.43
1.20
0.79
0.63
1.18
1.51
DAKKay's
Table 6. The results of average relative deviation by different method (ARD, %).
0.10
0.01
0.10
1.28
2.09
3.15
0.83
1.81
3.01
3.25
2.19
0.72
0.04
1.08
2.02
1.99
DAKKay's
(Continued)
263.15K
2.70
1.67
0.54 0.72
0.27
0.41
1.35
2.20
2.09
293.15K
DAKSBV
0.68
0.09
0.04
0.10
0.16
This model
MEASUREMENT AND CORRELATION 8
0.64
0.05 0.33
1.06
0.98
0.68
1.27
0.30
0.05
0.46
0.05
6.00
7.00
9.00
11.00
0.46
0.26
1.03
0.43
0.41
0.95
1.70
0.58
0.37
1.22
0.03
0.49
0.77
0.92 1.53
0.63
0.39
0.42
0.55
0.10
0.17
This model
0.41
0.82
-
3.24
2.28
1.54
0.16
DAKKay's
0.50
0.99
0.45
0.36
-
0.03
-
0.57
Sample 2
2.49
5.00
0.15
1.56
-
4.50
3.46
0.17
0.89
-
2.82
0.10
15.00
2.47
0.31
0.68
273.15K
Sample 1
DAKSBV
-
1.86
0.01
13.00
1.03
0.12
This model
303.15K
0.50
0.09
11.00
0.65
DAKKay's
313.15K
1.08
283.15K
DAKSBV
0.26
This model
9.00
Pressure /MPa
Table 6. The results of average relative deviation by different method (ARD, %). (Continued)
0.94
0.40
1.45
0.76
0.47
0.10
293.15K
1.96
1.60
0.51
0.02
263.15K
DAKSBV
0.40
0.61
0.75
0.20
0.04
0.47
2.82
2.47
1.32
0.63
DAKKay's
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
1.50
1.72
0.46
-
0.39
0.04
0.19
0.41
0.59
0.82
0.71
13.00
15.00
4.21
4.50
5.00
6.00
7.00
9.00
11.00
13.00
15.00
1.45
1.11
0.32
0.08
0.67
0.41
0.04
-
0.85
283.15K
2.78
2.72
0.59
0.82
1.32
1.31
1.50
1.07
0.54
-
1.24
1.48
1.56
2.48
0.81
1.08
0.17
0.02
0.43
0.79
0.28
-
1.01
0.13
2.64
0.63
1.15
0.19
0.30
0.78
1.14
0.61
273.15K
1.79
1.74
4.92
2.96
3.21
1.38
0.73
0.03
0.54
0.10
0.29
0.40
0.48
0.20
0.32
0.10
0.52
0.42
0.04
0.48
0.12
0.73
1.89
1.61
2.63
3.60
3.11
1.83
265.65K
2.84
1.13
1.86
0.88
0.97
0.63
2.32
2.15
1.13
1.07
0.46
MEASUREMENT AND CORRELATION 81
82
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. Predicted Z-factors from the new model, DAK-SBV, and DAK-Kay vs.the measured Z-factors. (Continued)
MEASUREMENT AND CORRELATION
83
Figure 2. Predicted Z-factors from the new model, DAK-SBV, and DAK-Kay vs.the measured Z-factors.
DAK-Kay (AAD=1.63%) and DAK-SBV (AAD=1.64%) came in the second and third order. In addition, the new method had the additional advantage of non-iterative computation.
6.6
Conclusions
In this work, gas compressibility factors of high C0 2 -content natural gases in the temperatures from 263.15K to 313.15K and pressures from 4MPa to 15MPa were measured using JEFRI-PVT apparatus made in Canada. On the basis of limited experimental evidence, it can be concluded that gas compressibility factors reduce with increasing C 0 2 content in natural gases and increase with increasing temperatures. The proposed method was an explicit correlation and yielded the most accurate prediction with the lowest average absolute deviation (0.42%) among three tested gas compressibility factor correlations.
84
6.7
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Acknowledgements
The authors wish to acknowledge anonymous reviewers for constructive comments and suggestions for improving this paper. The authors also wish to thank the anonymous Associate Editor for his handling of the manuscript and additional suggestions. This work was supported by National Science and Technology Major Project of P.R. China (No.2008ZX05016-001) and a grant from the National Natural Science Foundation of P.R. China (No. 50774062).
6.8
Nomenclature A mole fraction of C0 2 +H 2 S B mole fraction of H2S A, B., C, D. (i=a, b, c, d) tuning coefficient AAD average absolute deviation, % ARD absolute relative deviation, % / SBV parameter, K-MPaA K SBV parameter, K-MPcï05 n the number of data c
P system pressure, MPa Pci critical pressure of the component i, MPa P c pseudo-critical pressure, MPa P c cgr corrected pseudo-critical pressure with Casey correlation, MPa P' c corrected pseudo-critical temperature with Wiehert-Aziz correlation, MPa P" c corrected pseudo-critical pressure with WichertAziz and Casey correlations, MPa Pr reduced pressure P s ambient pressure, MPa T absolute temperature, K Td critical temperature of the component i, K T c pseudo-critical temperature, K T corrected pseudo-critical temperature with Casey correlation, K T c corrected pseudo-critical temperature with Wiehert-Aziz correlation, K T" c corrected pseudo-critical temperature with Wiehert-Aziz and Casey correlations, K
MEASUREMENT AND CORRELATION
85
T reduced temperature T ambient temperature, K Vs volume of gas released at ambient pressure and temperature, ml AV volume of gas bled from the PVT vessel, ml y. mole fraction of component i Z gas compressibility factor a, /?. (¿=0,1, 2,3) tuning exponent pr reduced density t, Wiehert-Aziz pseudo-critical correction Cal. calculated value Exp. experimental data or exact value
References 1. Adisoemarta, P.S., Frailey, S.M., Lawal, A.S., 2004. "Measured Z-factor of C0 2 —dry gas/wet gas/gas condensates for C 0 2 storage in depleted gas reservoirs." Paper SPE 89466 presented at the 2004 SPE/DOE 14th Symposium on Improved Oil Recovery, Tulsa, Oklahoma, 1-11. 2. Bahadori, A., Mokhatab, S., Towler, B.B., 2007. "Rapidly estimating natural gas compressibility factor." /. Nat. Gas Chem. 16, 349-353. 3. Brill, T.P. and Beggs, H.D., 1974. "Two-phase flow in pipes." Univ. of Tulsa, INTERCOMP Course, The Hague. 4. Dranchuk, P.M. and Abou-Kasem, J.H., 1975. "Calculation of Z factors for natural gases using equation of state." /. Can. Petrol. Technol. July-Sept., 34-36. 5. Dranchuk, P.M., Purvis, R.A. and Robinson, D.B., 1974. "Computer calculation of natural gas compressibility factors using the Standing and Katz correlation." Inst. Petrol. Technol. Pap. IP 74-008,1-13. 6. Elsharkawy, A.M., Elkamel, A., 2001. "The accuracy of predicting compressibility factor for sour natural gases." Pet. Sei. Technol. 19(5&6), 711-731. 7. Elsharkawy, A.M., 2002. "Predicting the properties of sour gases and condensates: Equations of state and empirical correlations." Paper presented at the 2002 SPE International Petroleum Conference and Exhibition, Villaherrmosa, Mexico, 1-17. 8. Hall, K.R. and Yaborough, L., 1973. "A new equation of state for Z-factor calculations." Oil Gas J. June 18,82-85, 90,92. 9. Hankinson, R.W., Thomas, L.K. and Philips, K.A., 1969. "Predict natural gas properties." Hydrocarb. Process. April, 106-108. 10. Jokhio, S.A., Tiab, D., and Escobar, EH., 2001. "Quantitative analysis of deliverability, decline curve, and pressure tests in C 0 2 rich reservoirs." Paper SPE 70017 presented at the SPE Permian Basin Oil and Gas Recovery Conference, Midland, Texas, 1-13. 11. Kay, W.B., 1936. "Density of hydrocarbon gases and vapor at high temperature and pressure." Ind. Eng. Chem. 28(9), 1014-1019.
86
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
12. Keseler, M.G. and Lee, B.I., 1976. "Improve prediction of enthalpy of fraction." Hydrocarb. Process. March, 153-158. 13. Lee, J. and Wattenbarger, R.A., 1996. Gas Reservoir Engineering. SPE Text Book Series Vol.5, Texas, pp. 8-9. 14. Li, X.F., Gang, T., Zhuang, X.Q., et al., 2001. An analytic model with high precision for calculating compressibility factor of high-pressure gas. /. Uni. Pet. China. 25(6), 45-46, 51. 15. Li, Q., Guo, T.M., 1991. "A study on the supercompressibility and compressibility factors of natural gas mixtures." /. Pet. Sei. Eng. 6(3), 235-247. 16. Liu, J.Y., Li, S.L., Guo, P., et al., 2002. "Measurement of gas deviation factor." Nat. Gas Ind. 22(2), 63-65. 17. Lu, H.Z., 1982. Petrochemical Industry Basic Handbook. Chemical Industry Press, Beijing, pp. 18-26. 18. Papay, J., 1968. "A termelestechnologiai parameterk valtozasa a gazlelepk muvelese soran." OGIL MUSZ, Tud, KuzL, Budapest, 267-273. 19. Pedersen, K.S., Fredensland, A. and Thomassen, P., 1989. Advances in Thermodynamics 1 137. 20. Riazi, M.R., 2005. "Characterization and properties of petroleum fractions." ASTM Stock Number: MNL50, West Conshohocken, PA, USA, pp. 215-221. 21. Reid, R.C., Prausnitz, J.M. and Poling, B.E., 1987. The Properties of Gases and Liquids. 4th ed., McGraw-Hill, Inc., New York. 22. Satter, A. and Campbell, J.M., 1963. "Non-ideal behavior of gases and their mixtures." SPE /. 3(4), 333-347. 23. Sheng, J.Y, Fang, W.P., Wang, Y.M., et al., 2004. "A modified equation for correlating experimental data—non-integral power polynomial equation." Comput. Appl. Chem, China. 21(5), 725-728. 24. Shokir, E.M. EL-M., 2008. "Novel density and viscosity correlations for gases and gas mixtures containing hydrocarbon and non-hydrocarbon components." /. Can. Petrol. Technol. 47(10), 45-54. 25. Standing, M.B. and Katz, D.L., 1942. "Density of natural gases." Tran. AIME. 146,140-149. 26. Stewart, W.F., Burkhard, S.F. and Voo, D., 1959. "Prediction of pseudo critical parameters for mixtures." Paper presented at the AICHE Meeting, Kansas City, MO. 27. Sutton, R.P, 1985. "Compressibility factors for high molecular weight reservoir gases." Paper SPE 14265 presented at the SPE Annual Technical Meeting and Exhibition, Las Vegas, Sent, 22-25. 28. SY/T 6434-2000. "Analysis for natural gas reservoir fluids physical properties." China National Oil and Gas Industry Standards, 1-23. 29. Varotsis, N. and Pasadakis, N., 1996. "Calibration prerequisite for accurate PVT measurements." Oil Gas }. 94(5), 93-96. 30. Wiehert, E. and Aziz, K., 1972. "Calculate Z's for sour gases." Hydrocarb. Process. 51(5), 119-122. 31. Yarborough, L. and Hall, K.R., 1974. "How to solve equation of state for Z-factors." Oil Gas J. Feb. 18, 86-88.
SECTION 2 PROCESS ENGINEERING
This page intentionally left blank
7
Analysis of Acid Gas Injection Variables Edward Wiehert 1 and James van der Lee2 1
Sogapro Engineering Ltd., Calgary, AB, Canada Virtual Materials Group Inc. Calgary, AB, Canada
2
Abstract The compressor discharge pressure is an important factor in the design of an acid gas compression and injection facility. In this work a methodology is presented to calculate this pressure by accounting for several key factors including the injection zone pressure, hydrostatic pressure and injectivity index. The transport properties needed to calculate these factors were determined using VMGSim. This methodology enables the rapid evaluation of the differing effects expected in an acid gas compression and injections scheme including variations in: acid gas compositions, geothermal temperature gradients, injection flow rate, etc.. A series of sensitivity studies with varying factors that may be difficult to obtain is presented to demonstrate the relative importance of each factor and its impact on the overall acid gas compression and injection facility design.
7.1
Introduction
In the disposal of acid gas b y compression a n d injection into a n u n d e r g r o u n d formation, the compressor discharge pressure h a s to be estimated fairly accurately to design the compression facilities as economically as possible. The discharge pressure can be estimated b y t h e following formula [1]: PD=Pf+*PJ-APH+Ptbfr+Plfr+APH2_m
(1)
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (89-106) © Scrivener Publishing LLC
89
90
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
where P D = compressor discharge pressure, kPaa Pf = formation pressure, kPaa AP = injectivity index pressure, m 3 / d a y / k P a at perforations APH = hydrostatic head pressure in tubing from injected fluid, kPa Ptbfr = friction pressure loss in injection tubing, kPa Plfr = friction pressure loss in line from compressor to well, kPa APH2 H1 = hydrostatic head pressure due to elevation difference between compressor (HJ and well (H2), kPa The formation pressure is provided by the reservoir engineering group, and is usually determined by running a bottom-hole pressure gradient in the well with pressure recorders. At the same time, the temperature at formation depth can also be determined. The injectivity index is a bit more complicated to determine. The best way is to perform an injectivity test with water at two or three different injection rates. If that is not possible or is inconvenient, the injectivity index, J, can be approximated by the following equation, ignoring damage or stimulation effects in the near-wellbore area [2]: J = 0.5356kh /(fiB0 (In {re/rw)~
0.75))
(2)
where J = injectivity into formation through perforations, m 3 / d a y / MPa k = permeability, mD h - formation thickness, m fi = viscosity, cp B = formation volume factor, 1 o
'
re = external boundary radius, m r = wellbore radius, m. This means that some basic data have to be known about the formation into which the acid gas is to be injected, in addition to reservoir pressure and temperature.
7.2
Discussion
A model has been developed to estimate the various pressure components in Eqn. 1. The results are dependent on the composition of
ANALYSIS OF ACID GAS INJECTION VARIABLES
91
the acid gas mixture, and on accurately determining the densities and volumes of the acid gas mixture with changes in pressure and temperature. With this model it is possible to assess the sensitivity of the various parameters that influence the determination of the compressor discharge pressure. The term in Eqn. 1 that has the paramount influence on the compressor discharge pressure is the pressure of the injection zone, P f . The next-most influential term is the hydrostatic head pressure developed by the density of the acid gas in the tubing. This term has a negative sign, and is very influential in reducing the required injection pressure, since at elevated pressures the acid gas will likely be in the liquid or dense phase in the tubing. The third-most influential term is the injectivity index, which is a function of the basic reservoir properties of permeability and thickness, as well as the level of damage or stimulation in the near-wellbore area. The other parameters in Eqn.l are influenced by the properties of the acid gas such as density, viscosity, specific heat capacity, and the dimensions of the injection line and tubing. A model can be used to analyse the influence of the size of the line and tubing with respect to friction losses and temperature changes. At low rates of flow, it can be expected that friction losses would be low, and the fluid temperature in the line or tubing would be greatly influenced by the external ground temperature or the geothermal gradient temperature. Friction losses can be estimated by the general flow equation for gas flow [3]:
(p2-p2)03d25 Q = 0.1034 U l i 2 \ f (GTZLf5
ÍT y
(3)
where Q = gas flow rate in standard cubic metres per day, Sm 3 /day at standard conditions of 101.325 kPa and 288.15 K Pj = upstream pressure, kPaa P 2 = downstream pressure, kPaa d = internal pipe diameter, mm G = gas relative density, air = 1 T = average gas temperature, K Z = gas compressibility factor, at P and T Pavg = 0.6667 (P1 + P2 - (P, x P p / C ^ ?P 2 )) L = line length, m f = Darcy friction factor (large f)
92
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Friction loss in the liquid phase or dense phase can be estimated by the Darcy pressure loss equation [4]: AP = 108.6 fp
Lq2/d5
(4)
where p = fluid density, k g / m 3 q = injection rate in liquid phase, m 3 /day with the other terms as defined for Eqn. 3. The friction factor in both of the above equations can be determined by the Chen equation [5]. The density of the acid gas can be determined by the VMG software program [6]. The temperature loss along the injection line from the compressor to the well can be estimated by the following relationships [7]:
T2-t The rate of heat gained by or transferred to the ground is matched under steady state conditions by the rate of heat lost by the gas, which is determined by: H =qmCr
(Ta - T2)
(6)
The terms in the above equations are defined as follows: Symbol Term SI Units AH Rate of heat transfer to ground W Overall heat transfer coefficient U W/(m 2 °C) Area of heat transfer (pipe surface) A m2 T, Upstream temperature °C Downstream temperature °C T2 t Ground temperature at pipe depth °C Mass flow rate of gas kg/sec qm Specific heat capacity of gas J/(kg °C) c The gas flow rate at standard conditions can be converted to mass rate by the following formula: qm = 14.177 x lO"6 G Q
(7)
ANALYSIS OF ACID GAS INJECTION VARIABLES
93
where G is the gas relative density (air = 1) and Q is the gas flow rate in Sm 3 /d. The above heat loss equations can be combined, as follows and solved for T2:
qm Cp (T, -T2) = ^ - 0 - O W ) ln^—^ T2-t
(8)
Since the rate of gas injection is the main part of the input data, the temperature gradient of the fluid can be estimated, knowing the geothermal gradient temperature with depth of the well. The physical properties of the acid gas mixture can be estimated by software programs such as the VMG suite of programs. The main properties that need to be determined are, first of all, the density of the fluid with changing pressure and temperature, as well as specific heat capacity and viscosity.
7.3
Program Design
The program requires the following basic input data: A. Dimensions: • line size between compressor and well, mm • tubing size, mm • well depth to perforations, m • elevation difference between compressor and injection well, m • absolute roughness of pipes, for friction factor determinations, mm B. Reservoir Properties • pressure, kPaa • temperature, °C • injectivity index C. Acid Gas Composition D. Assumptions • temperature 1 m below ground level, °C • linear temperature gradient between ground level and formation temperature
94
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
In the following calculation results, there are certain variables that are held constant to be able to compare results due to changes in the significant terms, such as injection rate, reservoir pressure and injectivity index. The following parameters were fixed in the calculations: Gas temperature leaving the compressor site at 40 °C Gas line to well, 1000 m horizontal, with ID of 107.9 mm and OD of 114.3 mm Overall heat transfer coefficient of 3 W/(m 2 °C) Ground temperature at 1 m burial depth of 0 °C Absolute roughness of pipe and well tubing of 0.02 mm Geothermal temperature gradient of 3 °C near surface and 60 °C at 2 000 m depth Acid gas liquid viscosity of 0.1 cp, and gas viscosity of 0.015 cp The specific heat capacity, Cp, J/(kg °C), is a function of density and temperature of the fluid as it flows from surface to the formation. An iterative procedure was incorporated in the program to match the specific heat capacity to the density and temperature of the acid gas. After all necessary information is entered into the program, the temperature of the gas at the well is calculated on the basis of the temperature out of the final compressor cooler and heat loss to the ground, followed by the calculation of the temperature gradient of the injection gas along the depth of the well from top down. The pressures are then calculated from bottom up, in steps of one thirtieth of total depth, to the well head and then back to the compressor. This is the required compressor discharge pressure on the basis of the input data.
7.4
Results
The phase diagram and hydrate temperature profile for this acid gas composition are shown in Figure 1. The fluid mass density versus pressure and temperature is shown in Figure 2. During the compression steps of four or five stages, the gas has to be cooled after each stage of compression. The temperature of the gas during cooling between compression stages must not drop below the
ANALYSIS OF ACID GAS INJECTION VARIABLES
95
hydrate line or the dewpoint line of Figure 1. Upon controlling the water content to some necessary low value, it is then possible to cool below the hydrate line, and after the final stage of compression, the acid gas may be safely cooled below the dewpoint line as well. The data in Figure 2 illustrate the changes in density of the acid gas with pressure and temperature. The upper set of curves in the figure, below about 40 °C, indicates that the density of the acid gas reaches about 80 % of the density of water. It is this density in the liquid phase that provides the hydrostatic head benefit in the tubing to reduce the compressor discharge pressure.
Figure 1. Phase envelope and hydrate curve for a 49% H2S, 49% C 0 2 and 2% CH 4 Mixture.
Figure 2. Density vs T and P plot for a 49% H 2 S, 49% C 0 2 and 2% CH 4 mixture.
96
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Given the large number and potential ranges of values of the variables affecting the down-hole pressure calculations, as well as the interaction between the variables, only the results of a select subset of parameters and ranges of their values are analyzed. Six tables are presented of results using one average composition of 49 mol % each C 0 2 and H2S, and 2 mol % CH 4 . Also, the average depth of the perforations in the well has been kept constant at 2 000 m, with nominal 73 mm (2 % inch) tubing. The difference among the first three tables is the input reservoir pressure, showing a low pressure for Table 1, an intermediate pressure for Table 2, and a pressure at hydrostatic gradient pressure for Table 3. To illustrate the effect of the third-most important parameter determining the compressor discharge pressure, namely the injectivity index, Tables 1 to 3 are repeated as Tables 4 to 6, in which case the injectivity index was tripled, from 23.06 to 69.18 m 3 / d / M P a (1 Bbl/d/psi to 3 Bbl/d/psi).
7.5
Discussion of Results
7.5.1 General Comments In the above tables, there are several selected results presented from the output of the computer calculations. The injection rate at Table 1. Calculated results, low reservoir pressure case (7 000 kPaa). injectivity index of 23.06 m 3 /d/MPa (1 Bbl/d/psi). Injection Rate, 10 3 Sm 3 /d
10
25
75
150
Injection rate at perfs, m 3 / d
23.8
52.7
149.9
298.6
Injectivity pressure, kPaa
1031
2 284
6498
12 948
Bottom-hole fluid temperature, °C
39.1
28.3
26.1
53.0
Bottom-hole pressure, kPaa
8 031
9 284
13 498
19 948
8
50
78
300
Wellhead P, kPaa
2 601
2 660
3 545
5 690
Compressor discharge P, kPaa
2 602
2 664
3 567
5 727
Selected Results
Well friction loss, kPa
ANALYSIS OF ACID GAS INJECTION VARIABLES
97
Table 2. Calculated results, intermediate reservoir pressure case (14 000 kPaa). injectivity index of 23.06 m7d/MPa (1 Bbl/d/psi). Injection Rate, 10 3 Sm 3 /d
10
25
75
150
Injection rate at perfs, m 3 / d
21.6
49.8
146.9
287.5
Injectivity pressure, kPaa
936
2161
6 369
12 466
Bottom-hole fluid temperature, °C
41.8
29.7
31.9
35.0
14 936
16161
20 469
26 466
2
10
76
284
Wellhead P, kPaa
2 617
2 924
5 215
10 791
Compressor discharge P, kPaa
2 618
2 928
5 226
10 800
Selected Results
Bottom-hole pressure, kPaa Well friction loss, kPa
Table 3. Calculated results, high reservoir pressure case (19 600 kPaa). injectivity index of 23.06 m 3 /d/MPa (1 Bbl/d/psi). Injection Rate, 10 3 Sm 3 /d
10
25
75
150
Injection rate at perfs, m 3 / d
20.7
48.8
142.9
280.0
Injectivity pressure, kPaa
896
2118
6 196
12143
Bottom-hole fluid temperature, °C
43.8
33.3
32.4
34.2
20 496
21718
25 796
31743
2
10
74
274
Wellhead P, kPaa
4 097
5 321
9 667
15 470
Compressor discharge P, kPaa
4 097
5 322
9 670
15 479
Selected Results
Bottom-hole pressure, kPaa Well friction loss, kPa
the perforations is simply the actual volume of the gas rate at the density of the fluid in the casing at the perforations. The injectivity pressure is the increase in pressure due to the actual volume of the fluid to inject it through the perforations on the basis of the injectivity index. The bottom-hole fluid temperature is the calculated
98
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 4. Calculated results, low reservoir pressure case (7 000 kPaa). injectivity index of 69.18 m 3 /d/MPa (3 Bbl/d/psi). Injection Rate, 10 3 Sm 3 /d
10
25
75
150
Injection rate at perfs, m 3 / d
24.4
53.8
155.3
314.1
Injectivity pressure, kPaa
353
778
2 245
4 541
Bottom-hole fluid temperature, °C
39.3
28.6
25.7
31.2
Bottom-hole pressure, kPaa
7 353
7 778
9 245
11541
8
54
464
1026
Wellhead P, kPaa
2 603
2 649
3 407
4 413
Compressor discharge P, kPaa
2 603
2 653
3 431
4 479
Selected Results
Well friction loss, kPa
Table 5. Calculated results, intermediate reservoir pressure case (14 000 kPaa). injectivity index of 69.18 m 3 /d/MPa (3 Bbl/d/psi). Injection Rate, 10 3 Sm 3 /d
10
25
75
150
Injection rate at perfs, m 3 / d
21.7
50.3
147.9
300.2
Injectivity pressure, kPaa
314
726
2138
4 340
Bottom-hole fluid temperature, °C
41.6
29.2
27.4
33.8
14 314
14 726
16138
18 340
2
10
78
304
Wellhead P, kPaa
2 616
2 799
3 914
5183
Compressor discharge P, kPaa
2 616
2 803
3 934
5 234
Selected Results
Bottom-hole pressure, kPaa Well friction loss, kPa
temperature of the fluid on the basis of the geothermal temperature profile and heat exchange between the fluid and the rock as the fluid moves down the tubing. The bottom-hole pressure is the actual fluid pressure in the casing at the perforations, and is the sum of the reservoir pressure and the injectivity pressure. The well
ANALYSIS OF ACID GAS INJECTION VARIABLES
99
Table 6. Calculated results, high reservoir pressure case (19 600 kPaa). injectivity index of 69.18 m 3 /d/MPa (3 Bbl/d/psi). Injection Rate, 10 3 Sm 3 /d
10
25
75
150
Injection rate at perfs, m 3 / d
20.8
49.0
147.1
292.3
Injectivity pressure, kPaa
300
709
2 126
4 225
Bottom-hole fluid temperature, °C
43.6
32.1
34.2
35.6
19 900
20 309
21726
23 825
2
10
76
290
Wellhead P, kPaa
3 560
3 901
6 303
8 541
Compressor discharge P, kPaa
3 560
3 901
6 306
8 551
Selected Results
Bottom-hole pressure, kPaa Well friction loss, kPa
friction loss is the total friction loss in the tubing. The wellhead pressure is the result of the calculations, starting with the bottom-hole pressure, and the hydrostatic pressure reduction as the calculation proceeds uphole, and adding the friction component per calculation step. The compressor discharge pressure includes friction pressure losses between the wellhead and the compressor location. In this case, there was no assumed elevation difference between the compressor and wellhead elevations. Figure 3 is a copy of the computer input and output data, showing graphically the geothermal gradient temperature profile in blue and the fluid temperature profile in the tubing in red, for the rate of 25 000 Sm 3 /d, in Table 2. The green cells contain the input data, and the numbers in the clear cells are the results of the calculations Upon estimating the required compressor discharge pressure as shown in the above tables, it is necessary to decide how many stages of compression should be included in the design of the compressor. In acid gas compression with reciprocating compressors, the compression ratio range is generally between 2 and about 2.8. At higher compression ratios, the compressor discharge temperature becomes fairly hot. Ideally the discharge temperatures should not exceed about 160 °C. The compression ratio can be estimated by:
c = (pD/psy
(9)
100
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 3. Screenshot of the discharge pressure calculator at an injection rate of 25 000 Sm3/d.
where C = compression ratio per stage P D = compressor discharge pressure after final stage, kPaa P s = suction pressure to first stage, kPaa n = number of compression stages. Assuming an initial suction pressure of 150 kPaa, a final discharge pressure of about 9 000 kPaa can be achieved with a compression ratio of 2.8, not accounting for minor pressure losses between stages of compression. When dealing with depleted reservoirs, i.e., reservoirs that have been on production and consequently their current pressures are lower than their ultimate pressure upon repressurizing, it is necessary to keep in mind the final reservoir pressure when selecting the number of compression stages. In the above tables, and the low injectivity index cases, (Tables 1 to 3) four stages of compression would probably be the selection
ANALYSIS OF ACID GAS INJECTION VARIABLES
101
up to a rate of 75 10 3 Sm 3 /d. It has to be kept in mind that Tables 1 and 2 deal with depleted reservoirs, and the ultimate pressure would reach the discharge pressures of Table 3. At higher rates, ultimately five stages of compression would be required. Another pressure item to keep in mind is that the bottom-hole pressure must not exceed the fracturing pressure of the injection zone. This pressure limit is set by the reservoir engineering section, in accordance with reservoir properties and regulations. The benefit of higher injection capacity is evident when comparing the respective results between Tables 1 to 3 with Tables 4 to 6. The injectivity index can in most injection wells be greatly improved with the application of a massive acid stimulation treatment or a fracture treatment. Such treatments can substantially reduce the required compressor discharge pressure. Concerning the heat transfer estimation between the fluid in the tubing and the surrounding rock, the calculation method uses the relationship of Equation 8. As stated above, the specific heat capacity, C , is calculated for each step for the pressure and temperature throughout the tubing depth. This requires an iterative procedure, as C is a function of pressure and temperature, and temperature is a function of heat transferred to or from the surrounding rock formations. Due to the interdependence of the fluid temperature, pressure, density and C , it takes between 20 to 30 iterations for convergence to be achieved.
7.5.2
Overall Heat Transfer Coefficient, U
The overall heat transfer coefficient, U, was held constant for the results in Tables 1 to 6 at 3 W/(m 2 °C). This is a typical value for non-insulated pipe in gas gathering systems [7]. The overall heat transfer coefficient in the tubing was assumed to be similar to the value used for the heat loss estimation for the acid gas in the buried line between the compressor and the wellhead. To test the effect of variations in the overall heat transfer coefficient, U, in the tubing, some tests were conducted by reducing the value of U by a factor of 3, namely to a value of 1 W/(m 2 °C), as well as tripling the value to 9 W/(m 2 °C), as compared with the value of 3 W/(m 2 °C) for all calculations in Tables 1 to 6. The results are shown in Table 7, for the intermediate reservoir pressure of 14 000 kPaa used in Table 2. The results from Table 2 are repeated in Table 7 in Runs No. 5 to 8 for easy reference.
102
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 7. Effect of variations in overall heat transfer coefficient, U, in tubing. On compressor discharge pressure, at reservoir pressure of 14 000 kPaa. Run No.
Inj. Rate 103 Sm 3 /d
Wellhead U T°C W/(m 2 °C),
Bottom Compr. Discrepancy T°C PkPaa Percent
1
10
1
0.9
24.7
2 443
-6.68
2
25
1
8.7
17.7
2 865
-2.15
3
75
1
30.0
30.6
5 270
0.84
4
150
1
35.0
34.9
10 846
0.43
5
10
3
1.4
41.8
2 618
-
6
25
3
9.1
29.7
2 928
-
7
75
3
29.8
31.9
5 226
-
8
150
3
35.0
35.0
10 800
-
9
10
9
1.9
53.4
2 811
7.37
10
25
9
9.5
45.0
3 003
2.56
11
75
9
29.6
36.5
5180
-0.88
12
150
9
35.0
35.9
10 722
-0.72
The discrepancies are calculated by comparing the compressor discharge pressures at the different values of U and the respective rates with the results obtained with the value of U used in the generation of the results in Tables 1 to 6. As can be seen, The discrepancies are quite small. This is largely due to the fact that in the upper portion of the tubing, the fluid in the tubing is warmer than the surrounding rock. Further down-hole, this temperature relationship changes, so that the rock is warmer than the fluid in the tubing. Due to this reversal in the temperature conditions between upper and lower portion in the well, the effect of heat transfer between fluid in the tubing and the formation is averaged, in that in the upper portion of the tubing the acid gas is cooled, and in the lower portion it is warmed.
ANALYSIS OF ACID GAS INJECTION VARIABLES
103
A pipeline in the ground is surrounded by soil, and the heat loss to the surroundings can be easily verified by recording the temperature change of the fluid between inlet and outlet of the pipe. This could also be done in a well, however the cost of installing bottomhole temperature recorders, even on a temporary basis, is high. Furthermore, the risk of perhaps a wireline failure also detracts operators from performing such temperature measurements. As a result, it is necessary to assume an average value for U for the heat transfer between fluid and surrounding formations. The physical conditions in the tubing and surrounding rock are somewhat different from the conditions of the pipeline. In the latter case, the type of soil that the pipe is buried in is known from digging the pipe trench. Different values of U have been determined experimentally for different soil types [7]. While the formations in a well are logged, the actual relationship between tubing and the formations is not known. When a well is drilled, the hole is larger by several mm than the casing that will be installed. Some portions of the well bore are usually washed out, resulting in thicker concrete sections in those portions. The casing is cemented in the hole, but due to the fact that no well is absolutely vertical, the casing will not necessarily be cemented in the center of the well bore. Centralization of the casing is mainly confined to the formation of interest. Since the well is not necessarily cemented to surface, some length of the upper section outside of the casing would contain drilling mud or water instead of concrete. Furthermore, after the well is completed, the tubing is installed and set on a packer. In the lower portion of the well, the tubing is in compression, which means that it is not located in the middle of the casing, but is in direct contact with the casing as a spiral column. In the upper portion, the tubing is in tension, but it still touches the casing due to the fact that the well is not truly vertical. It is offset somewhat from touching the casing by the fact that the couplings of the tubing joints are somewhat larger in outside diameter than the pipe itself. The space between the tubing and the casing is filled with either water or a hydrocarbon liquid such as diesel fuel or stabilized condensate. The thermal conductivity between these two types of fluids differs substantially. Additionally, there will be convection eddies due to the temperature difference between the tubing wall and the casing wall. For heat transfer to occur between the acid gas in the tubing and the surrounding rock, the heat has to flow through the tubing wall, the liquid in the annulus, the casing,
104
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Table 8. Thermal conductivity of materials associated with acid gas injection well. Material
Thermal Conductivity, W/(m K),
Oil
0.14
Water
0.6
Steel
52
Concrete
1.7
Soil
0.17-3.5
Limestone
1.3
Sandstone
2.4
the concrete sheath around the casing, and then to or from the rock. Different rock formations have different thermal conductivities. To calculate the overall heat transfer coefficient in light of the eccentric location of the tubing in the casing and the casing in the hole, different thicknesses of the concrete sheath, and different thermal conductivities of different rock formations, porous and impermeable, would give uncertain results due to too many unknowns. Thus the estimate of an overall heat transfer coefficient similar to the relationship for pipeline heat transfer is a reasonable approach, especially in light of the fact that large variations in U produce small discrepancies in the compressor discharge pressure, especially at higher rates. The best way to determine the bottom-hole temperature is by instrument, and then adjusting the coefficient to match the measured temperature. Table 8 provides information on the typical thermal conductivity of various materials that play a role in the heat transfer between acid gas in the tubing and the surrounding formations [8,9,10].
7.5.3
Viscosity
The viscosity is a transport property that is used to determine the coefficient of friction in gas flow in pipes, as well as the injectivity index. In this work, the viscosity was kept constant at 0.1 centipoise for acid gas in the liquid state, and 0.015 centipoise in the gas
ANALYSIS OF ACID GAS INJECTION VARIABLES
105
phase. Variations in viscosity had little influence on the results for the compressor discharge pressures.
7.6
Conclusion
A computer model can be used to estimate the required discharge pressure of an acid gas injection compressor. The pressure depends mainly on the reservoir pressure, but is also highly influenced by the injection capacity of the injection zone.
References 1. Wiehert, E., Notes on "Acid Gas Compression and Injection", Chapter 4, Sogapro Engineering Ltd, Calgary, Canada. 2. Craft, B. C. and Hawkins, M. F., Applied Petroleum Reservoir Engineering, Prentice Hall Inc., Englewood Cliffs, N.J. 3. Gas Processors Suppliers Association, Engineering Data Book, SI Version, Section 17,Tulsa,OK. 4. Crane Canada Limited, Technical Paper 410, Metric Edition - SI Units, Montreal, Canada. 5. Chen, N. H., "An Explicit Equation for Friction Factor in Pipes," Ind. Eng. Chem. Fund., 18,296,1979. 6. VMGSim v 5 0b5, August 2009, Virtual Materials Group Inc., Calgary, Canada. 7. Younger, A. H., Notes on "Natural Gas Processing Principles and Technology", Chapter 8, The University of Calgary, Calgary, Canada. 8. Perry, J. H., Chemical Engineers' Handbook, Fourth Edition, McGraw Hill Book Company, New York. 9. Wikipedia, Thermal Conductivity, htt://en.wikipedia.org/wiki/Thermal_ conductivity, retrieved 20/07/2010. 10. Manning, F. S. and Thompson, R. E., Oilfield Processing of Petroleum, Vol. 1: Natural Gas, Pennwell Corporation, Tulsa OK, 1991.
This page intentionally left blank
8 Glycol Dehydration as a Mass Transfer Rate Process Nathan A. Hatcher, Jaime L. Nava and Ralph H. Weiland Optimized Gas Treating, Inc., Sugar Land, TX
Abstract Glycol dehydration is a process that presents some unique challenges from technical and computational standpoints. In the first place, modern designs almost invariably use tower internals consisting of structured packing rather than the more traditional bubble cap trays. Structured packing offers lower pressure drop and considerably higher capacity than trays, and it is well suited to handling the very low L/G ratios common in dehydration. However, until now estimating height of packing used rules of thumb, not science. Mass transfer rate-based modeling, on the other hand, uses science and therefore offers greater reliability of design. The other challenge of dehydration using any glycol is thermodynamic. The dehydration of streams having very high concentrations of acid gases is hard to model reliably because the thermodynamics of vaporliquid phase equilibrium involves water, one of nature's most perverselynonideal chemical species. Interactions between water and the acid gases C 0 2 and H2S, as well as with most hydrocarbons in the gas phase must be taken into account for a thermodynamic model to be reliable. Furthermore, in the liquid phase, aqueous glycol solutions themselves are quite nonideal because both water and glycol are polar molecules. There are other facets of glycol dehydration that are interesting just from an applied science viewpoint. One of them is the heat transfer situation that ensues in a regenerator using both stripping gas and a reboiler (Stahl column). When the hot gas hits the bottom of the packing in the wash section atop the column it finds itself going from an environment in which it is saturated with the water contained in a predominantly TEG stream into an environment where it is grossly under-saturated with respect to the pure water stream in the wash section. This humidification
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (107-120) © Scrivener Publishing LLC
107
108
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
process extracts the necessary heat of vaporization as sensible heat from the liquid water phase and this can drop the wash water temperature by 30°F, 40°F or even more. Optimized Gas Treating, Inc. has recently released a new glycol dehydration model, currently for TEG, and being extended to MEG and DEG. This paper addresses the efficacy of the model in terms of (1) how well it reflects known phase behavior and (2) how closely it predicts known plant performance data using both bubble cap trays and packed columns without recourse to HETP or HTU estimates and other rules of thumb.
8.1
Phase Equilibrium
The concern here is with the accurate calculation of equilibrium water content of high- and low-pressure gases containing very high levels of C 0 2 a n d / o r H2S. The ProTreat™ simulation tool's dehydration model uses the Peng-Robinson equation of state (EOS) for the vapor phase and currently offers a 4-suffix Margules equation activity coefficient model based on the data of Bestani & Shing (1989) for the liquid phase as reported by Clinton et al. (2008). A similar model based on the less conservative data of Parrish et al. (1986) is planned for a future release. There are two important aspects to thermodynamic modeling: water content of the treated gas and the solubility of hydrocarbon, acid gas, and especially the BTEX in the water-laden glycol. Table 1 compares ProTreat model results with GPSA Data Book entries for saturated water content. Generally, ProTreat reproduces measured values of water content to within the accuracy of the data. The Peng-Robinson EOS that performs these saturated water content calculations applies a large number of interaction parameters (k 's) for the interactions between water and the various gases as well as between the gases themselves as outlined, for example, by Carroll and Mather (1995). Other components whose solubility in TEG is pertinent are the acid gases and hydrocarbons, especially the BTEX components. Vapor-liquid equilibrium constants (K-values) for benzene, toluene, ethyl benzene and o-xylene are available in GPA RR-131 and the data there have been used to fit the ProTreat solubility model for these species. The data indicate that at typical contactor conditions approximately 10-30% of the aroma tics in the gas stream may be absorbed in the TEG solution. ProTreat results conform closely
GLYCOL DEHYDRATION AS A M A S S TRANSFER RATE PROCESS
109
Table 1. Saturated water content of gases. GPSA Ref.
Mole Percent (Dry Basis)
Temp
Près
co 2
H2S
(oF)
(psia)
CH 4
H20 Ib/MMscf Meas'd
ProTreat
Ex 20-1
100
0
0
150
1,000
220
216
Ex 20-2
80
20
0
160
2,000
172
188
Fig 20-9
0
100
0
100
500
132
125.3
750
110
102.5
1,000
125
100.7
2,000
215
215.1
3,000
238
247.8
850
88
96.9
1,125
81
99.2
1,500
128
148.6
2,000
139
184.2
5.31
94.69
0
100
Fig 20-16 89
11
0
100
2,000
89
11
0
160
1,000
80
20
0
100
2,000
80
20
0
160
1,000
282
292.5
80
20
0
160
2,000
172
188.5
92
0
8
130
1,500
111
103.5
72.5
0
27.5
160
1,367
247
252.6
83
0
17
160
1,000
292
293.4
30
60
10
100
1,100
81
81.2
9
10
81
100
1,900
442
264.4
5.31
94.69
0
77
1,500
109.2
95
5.31
94.69
0
122
2,000
164.6
234.5
40.6 286 40.6
41.1 283.9 45.1
110
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
to the conclusions of RR-131 (as they should, because ProTreat's solubility model has been regressed to the actual measured BTEX solubilities).
8.2
Process Simulation
The GPSA Data Book contains a nice example of dehydration with TEG (Example 20-11). The gas is water saturated at 600 psia with other details noted in Figure 1. Two cases are detailed, both requiring two theoretical stages. One uses bubble cap trays which at a tray efficiency of 25 to 30%, translates into 6 to 8 actual trays. The other case uses 10-ft of an unspecified structured packing. ProTreat has provision for a separate Stahl column, shown immediately below the stripper in Figure 1 but the stripper can also be simulated without this column if desired. Two condenser outlet streams allow wet stripping gas withdrawal from the system (Stream 19), and removal of a specifiable portion of condensed water (Stream 20), with the remainder returned as reflux. Table 2 shows the effect of the actual tray count on the water content of the dehydrated gas. ProTreat simulation indicates 6 trays are adequate to reduce the water content from 88.7 lb/MMscf to the target level of < 7 lb/MMscf (32°F Dew Point). Tower diameter for
Figure 1. GPSA data book example 20-11.
GLYCOL DEHYDRATION AS A M A S S TRANSFER RATE PROCESS
111
Table 2. Water content vs. tray count. Number of Trays
Water lb/MMscf
5
8.5
6
6.7
7
5.7
8
5.1
70% flood is 3'-0". These values are in line with GPSA data book results which are annotated in Figure 1. In summary, the available data indicate that the model is accurately reflecting literature data on the VLE and general experience as reported by GPSA.
8.3
Dehydration Column Performance
Until now only an equilibrium stage model has been available for calculations involving the performance of structured packing. However, packing size is surely related to the HETP of the particular packing. Packing size can be expressed in terms of specific surface area and crimp size, characteristics that are geometrically related. Under otherwise identical process conditions, one should expect that large crimp packing will require a much deeper bed to give the same performance as a relatively short bed of small crimp packing simply because the surface area of the small crimp material is considerably higher. Figures 2 and 3 simulate how packing size within the Mellapak X-series (higher crimp angle) packings affects dehydration performance. For Sulzer Mellapak structured packings, the packing designation, e.g., M250.X is an approximate indicator of the specific area, in this case 250 m 2 /m 3 . Simulations were all run with 40-ft of packing and the absorber was sized for 70% flood (9 to 11 ft diameter depending on crimp size). The absorber was set up to dehydrate 49,000 lbmol/hr of wet sweet methane (trace C 0 2 and H2S) using 250 gpm of 99.95 wt% TEG at a nominal tower pressure of 200 psig. The gas-phase temperature profiles for the various packing sizes shown in Figure 2 indicate that there is a significant temperature bulge in dehydration columns, caused by the change of phase
112
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. Temperature profiles and the effect of packing size in the mellapak M series.
Figure 3. How water removal depends on packing size and packed bed depth.
GLYCOL DEHYDRATION AS A MASS TRANSFER RATE PROCESS
113
of the water being absorbed. Water absorption generates sensible heating equivalent to its latent heat of condensation. Also, when small-crimp packings are used, the temperature bulge is closer to the bottom of the column because water absorption is much more rapid with the larger surface area. Figure 3 shows the water content of the gas at various positions along the height of the column. It is evident that after traversing the bottom 20 feet of M350.X packing the gas is about as dry as 99.95 wt% TEG at 100°F can get it. With M125.X packing on the other hand, water is still being removed even after the gas has passed through 40 feet of packing. So the bed height needed is very much a function of the packing size. It is not that rules of thumb (ROT) cannot be made to work; rather, it's that the right ROTs, at least for packing, depend on too many parameters (not just packing type and size but also on the gas and solvent fluxes through the column) and this makes them not reliably available. Until now the answer to this dilemma has been to over build the columns; however, in a competitive environment, surely being able to avoid over design puts the knowledgeable contractor and the astute internals vendor in a commercially advantageous position. Mass transfer rate based simulation is the precise tool that allows this to be done. Figure 4 shows that the ROT of 6 to 8 trays for dehydration is a gross over simplification. The number of trays depends at least on the dryness to be achieved, i.e., the dryness of the solvent. If the target dew point is not too stringent (e.g., 10 or 20 lb H 2 0 / MMscf) then 6 or 8 trays seem adequate. But in very low dew point applications such as LNG plants two or three times that number of trays may be required to get to the desired dryness. With 99.97% TEG, for example, it is possible to get to below 1 l b / MMscf water content, but even after 16 trays, water is still being removed. For the particular conditions simulated in this study, it should be mentioned that high TEG viscosity is a consideration and will negatively affect internals performance compared with light hydrocarbons for example. At the solvent moisture levels encountered in these simulations, viscosity is not significantly affected by water content, and at the lowest temperature (feed solvent at 100°F) the viscosity is about 19 cP (for reference, corn syrup is 50-100 cP). This does not necessarily eliminate the possibility of using trays, although it does tend to make structured packing more attractive.
114
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 4. Water removal to low dew points requires deeper beds and more trays.
8.4
Stahl Columns and Stripping Gas
A Stahl column gives an extra stage of regeneration by taking the solvent from the reboiler and contacting it with a flow of dry stripping gas. Stahl columns are essential when the very treated gas must be of very low dew-point. Figure 5 shows the effect of stripping gas rate (SCF per gallon of TEG solvent) on the TEG purity and on the water content of the treated gas for Example 20-11 from GPSA Data Book scaled up by a factor of five and using 6 bubble cap trays in the absorber and a 10-ft bed of FLEXIPAC 1.6Y in the regenerator (includes a 2-ft reflux wash section). In terms of scale, the absorber is nearly 7-ft diameter and is drying gas that is water-saturated at 600 psia and 100°F The regenerator is only 15-in diameter. The simulation results in Figure 5 show that usingeven a modest flow of stripping gas can reduce the water content in the treated gas by more than a factor of two. It removes more water from the solvent and increases the dry TEG from 98.8 wt% to 99.8 wt%. This particular treated gas is dry enough by transmission line standards, but it's a long way from dry enough for an LNG plant, for example. However, this demonstrates the principle that a Stahl column can serve a useful purpose—it would be used when treated gas needs to be drier than usual.
GLYCOL DEHYDRATION AS A M A S S TRANSFER RATE PROCESS
115
Figure 5. Effect of stripping gas flow on solvent dryness and gas water content.
8.5
Interesting Observations from a Mass Transfer Rate Model
The mass transfer rate model uses real trays both in number and mechanical detail, and real packing in terms of actual bed depths of a specific packing with defined geometry including crimp angle, crimp size, surface treatment, and specific surface area. The simulation of a packed column begins with finely segmenting the packed height to approximate the continuous nature of contacting by using a large number of thin cross-sectional slices. As a result, the mass and heat transfer effects can be observed on a fairly detailed scale. Figure 6 shows the vapor and liquid temperature profiles through a 10-ft deep bed of 2-in metal Nutter Rings. The bed starts with a 2-ft deep wash section for TEG recovery followed by an 8-ft deep stripping section for water removal. The stripper was simulated by dividing it into 40 segments, each having a 3-in depth. Finer segmentation is, of course, possible but it adds nothing to the detail and very little to the accuracy of the simulation. We will traverse the regeneration column starting with the condensate (essentially pure water) which enters the column at 180°F and is heated by the gas stream (stripping gas and water) which is at 198°F.
116
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. Effect of stripping gasflowon solvent dryness and gas water content. As the condensate trickles through the bed it continues to be heated by the wet stripping gas, but within the first 18 inches its temperature reaches a peak (194.5°F) and then suddenly plummets to about 160°F. The gas becomes hotter as one descends through the column and eventually experiences a rapid climb to 294°F. The question is why these trends should not be a surprise. The answer for the gas phase is relatively simple: feed, preheated to 300°F enters the column at the 2-ft level and flashes into its vapor and liquid parts. The fact that the vapor temperature changes radically at the feed point is simply the result of the hot feed meeting a cold reflux water stream. But why does the reflux water cool so much? When the vapor leaves the top of the stripping section its water content is very nearly in equilibrium with the liquid in the stripping section. The liquid there is better than 95 wt% TEG and only 5 wt% (17-18 mol%) water. This vapor is suddenly exposed to a pure, hot water stream coming from the wash section so it is seriously under saturated against pure water. The reaction then is for water to evaporate as fast as possible to re-saturate the hot gas. This is a typical humidification operation and has some interesting characteristics. The evaporation process is not mass transfer rate limited! Rather, it is limited by the rate at which the latent heat of vaporization
GLYCOL DEHYDRATION AS A M A S S TRANSFER RATE PROCESS
117
demanded by the humidification process can be drawn from the bulk liquid to the gas-liquid interface. This process is one of heat transfer and, indeed, the rate of humidification right above the feed point is controlled by essentially conductive heat transfer across the liquid film running over the packing. The process is heat transfer rate limited. The mass transfer driving force for humidification is so high that enough heat is drawn from the reflux water to cool it, in this case by approximately 33°F. As the bottom of the stripping section is approached, liquid meets increasingly hotter vapor coming from the reboiler (bubble point is quite sensitive to water content when the water is low). This cools the vapor and heats the liquid, and again some of the heat transfer is a result of water transferring from the vapor back into the liquid around the bottom part of the stripping section. Figures 7(a) and (b) are intended to show how the use of stripping gas affects these temperature profiles. The plots are for an 8-ft bed of FLEXIPAC® 1.6Y structured packing with 2-ft top wash section for TEG recovery. Even when stripping gas is not used, the condensed water leaving the reflux section meets a much hotter gas and a good part of the temperature equilibration takes place by water evaporation from the reflux stream. This is again a humidification process and the reflux water cools as a result of the demand for latent heat for vaporization. In Figure 7(b) the presence of stripping vapor dilutes the gas leaving the stripping section, and this results in slightly more driving force for humidification from the additional water holding capacity of the stripping gas. (How much
Figure 7. Effect of stripping gas flow on temperature profiles.
118
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
the stripping gas adds to driving force and capacity depends on the relative flows of stripping gas versus reboiler vapor.) Again, the reflux water temperature peaks and then drops from 197°F to about 141 °F, for a total cooling of 56°F, the same as for the no-stripping gas case. However, with stripping gas (N2) the temperature bulge in the reflux section is a consequence of the dilution by nitrogen. We note in passing that to maintain water balance the no-stripping-gas case required a 212°F condenser temperature versus 180°F for the case with stripping gas.
8.6
Factors That Affect Dehydration of Sweet Gases
To learn more about the practical limits of glycol dehydration, a sensitivity analysis was conducted starting with the 3 SCF/gallon stripping gas rate at 1.72 MMBtu/hr reboiler duty referenced in the preceding section. First, reboiler duty was increased to 2.5 MMBtu/hr which amounted to an increase from 1124 Btu/gallon to 1634 Btu/gallon. This change allowed the glycol purity to increase from 99.79 wt% to 99.96 wt% TEG. The water content of the dehydrated gas was predicted to drop from 2.7 lb/MMscf to 2.0 lb/MMscf assuming 6 bubble cap trays for contacting. This is not much of an improvement for nearly a 50% increase in reboiler energy. Are there enough absorber trays to take advantage of the greatly improved TEG dryness? The number of glycol contacting trays was next increased from 6 to 8 while maintaining the lean glycol purity of 99.96%wt. This dropped the water content of the gas by an order-of-magnitude, from 2.0 lb/MMSCF to 0.2 lb/MMSCF, so much for ROT. Finally, at 12 trays of dehydrator contacting, some of the reboiler duty was traded for stripping gas. Lowering the reboiler duty from 2.5 to 2.0 MMBtu/hr and doubling the stripping gas rate (from 3 to 6 SCF/gallon) resulted in a predicted water content of 0.13 l b / MMSCF. At this water content, the gas dew point can be expected to be well below the hydrate formation temperature. Simulated glycol purity was 99.975%wt. So it appears that stripping gas is a more effective way to improve dehydration system performance compared to brute force reboiler duty increase. There is much that can be learned by playing with a mass transfer rate model, even for a well-proven process such as glycol dehydration. The rate model allows one to probe the limits of what is in practice possible, for example trading expensive reboiler duty
GLYCOL DEHYDRATION AS A M A S S TRANSFER RATE PROCESS
119
for possibly cheaper stripping gas, or optimizing dehydration unit performance in a demanding LNG setting. Because this kind of tool is so reflective of the real physics, the real chemistry, and the real engineering going on in an amine, DMPEG or glycol plant, using the tool as a virtual laboratory and pilot plant is very attractive.
8.7 Dehydration of Acid Gases Down-well disposal of acid gases (so-called acid gas injection) requires the gas to be compressed to very high pressure. If the gas is wet, compressing it will cause liquid water to drop out and this liquid will be saturated with acid gas at high pressure. Unless one is prepared to build compressors and other equipment from unobtainium, it is paramount that the water be removed from the gas before compression. Therefore, it is of interest to compare dehydration of otherwise pure but water saturated H2S with dehydrating the equivalent sweet gas volume. In order to keep liquid water away from the compression train at pressures up to about 1000 psig, Figures 20-7 through 20-9 in the GPSA Data Books imply that dehydration to below about 100 lb H 2 0/MMscf may be required for some compression paths. A glycol contactor pressure of 10 psig was chosen assuming that acid gas came from a typical amine regenerator operating at 12-15 psig. Assuming an 80:20 mixture of H2S and C 0 2 saturated with water was to be dehydrated; ProTreat predicted that 5 gallons of 99.4%wt TEG would be required for each lb of water removed. Assuming the same glycol was circulated to scrub water-saturated sweet gas at 10 psig, water content in the dehydrated gas was predicted to be 10% lower than in the dehydrated acid gas, or nominally 91 lb/MMScf. So besides being stinkier in general, acid gas is tougher to dehydrate than same amount of sweet gas. Fortunately however for acid gas, this is traded off by a minimum in the water solubility at moderate pressures which may preclude the need for dehydration in some cases.
8.8
Conclusions
Rules of thumb are fraught with danger because often the rules simply do not apply. A blanket tray efficiency of 25% is close to the truth most of the time. But in deep water removal via glycol dehydration, 25% is optimistic and unless one adds several additional
120
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
"safety" trays, failure may be imminent. The situation with packing, be it random or structured, is much worse. Quoting or recommending a single HETP or HTU value can be very dangerous. The right value depends on the packing type as well as on the operating conditions and the gas dryness sought. ROTs had their place when equilibrium stage calculations were the best available tools and reliance had to be placed on experience as expressed (and misexpressed) in ROTs. Today we have available powerful mass transfer rate based simulation tools and reliance on rules of thumb and other approximations and guestimates is no longer warranted. The beauty of ProTreat's mass transfer rate based approach to simulation is that one never has to worry about tray efficiencies, artificial residence times, HETPs, HTUs, and other rules of thumb. ProTreat doesn't use rules of thumb—it uses science and good sound engineering to predict performance. These results were all obtained without any correction factors whatsoever. They are true predictions in every sense of the word.
Literature Cited Bestani, B., Shing, K. S., "Infinite dilution activity coefficients of water in TEG, PEG, glycerol and their mixtures in the temperature range 50 to 140°C," Fluid Phase Equilibria, vol 50,1989 Clinton, P., Hubbard, R. A., Shah, H., "A review of TEG-water equilibrium data and its effect on the design of glycol dehydration units," Laurence Reid Gas Conditioning Conference, Norman, OK, 2008 Parrish, WR, Won, KW, Baltatu, ME, "Phase Behavior of the Triethylene GlycolWater System and Dehydration/Regeneration design for Extremely Low Dew Point Requirements," Proceedings of the 65th annual convention of the GPA, San Antonio,1986 Research Report RR-131 The Solubility of Selected Aromatic Hydrocarbons in Triethylene Glycol - H.-J. Ng, C.-J. Chen, D. B. Robinson, DB Robinson Research Ltd., Edmonton, Alberta. Project 895. December 1991 Carroll, J.J., Mather, A.E., "A generalized correlation for the Peng-Robinson interaction coefficients for paraffin-hydrogen sulfide binary systems," Fluid Phase Equilibria, vol 105,1995 Engineering Data Book, Gas Processors Supplier's Association, 12th Edition, Vol II, Ch.20
9 Carbon Capture Using Amine-Based Technology Ben Spooner and David Engel Amine Experts, Calgary, AB, Canada
Abstract Amine-based solvents have been used in the oil and gas industry for some sixty plus years to remove C0 2 from gas streams. This technology can be used for carbon capture from pre- or post-combustion gasses in power plants as well. Certain challenges will arise however, due to the unique composition of flue gasses as well as the low pressures associated. In this paper we discuss how to utilize as much existing and proven technology as possible for effective carbon capture, as well as the subtle but important differences between historical amine systems and the future. This paper will cover the general purpose and flow scheme of a carbon capture amine plant, some brief comments on operating conditions, as well as an overview of the challenging areas of oxygen reacting with amine, low pressure C0 2 removal, and energy consumption.
9.1
Amine Applications
Amines are currently used to remove C 0 2 from gas streams in several areas of the oil and gas industry, and have been for over sixty years. In refining, amines are primarily used to remove H2S from hydrocarbon gas (and liquid) streams from several various sources. Although C 0 2 may be present, it is not normally a high priority as it is virtually useless to the refinery. The amine system will pick most if not all of the C 0 2 out of the gas, and it eventually goes to incineration. Refineries often have problems related to heat stable salts,
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (121-132) © Scrivener Publishing LLC
121
122
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
which is where stronger acids than H2S or C 0 2 are present and they form a very strong bond with the amine. Amines are also used around the world in natural gas processing plants to remove H2S a n d / o r C0 2 . The level of acid gas removal depends on the "sales gas specification", but on average the treated gas cannot exceed 2 mol% C0 2 . The gas streams being treated in these plants range greatly in pressure and composition, and for this reason many different types of amines are utilized depending on the situation. Gas plants have historically had difficulty in the area of oxygen entering the amine system and causing solution degradation. Tail gas treating is done in both gas plants and refineries, as an option to further reduce the H2S content of the gas exiting a sulphur plant. The primary focus of a tail gas treating plant is to selectively remove H2S from the gas, while leaving C 0 2 in. This is done at extremely low pressures, which makes acid gas removal from the gas very difficult. Utilizing amines in the carbon capture industry is certainly possible, though not without unique difficulties. Although no single one problem is unique, the combination of them is. Carbon capture involves removing C 0 2 from a very low pressure gas stream, which contains high levels of oxygen. The most proven type of amine for C 0 2 removal at low pressure, monoethanolamine, unfortunately will partially degrade when reacted with C0 2 . Carbon capture takes the main problems from each individual application of amine: refining, gas plants, and tail gas; and combines them. Heat stable salt formation, chemical degradation, and low pressure treating are a day-to-day battle in the carbon capture process when using amines.
9.2 Amine Technology Amines, having a pH of approximately 10, are medium strength bases which are used to remove C 0 2 from gas streams. The C0 2 , in the presence of water, is acidic, which then reacts with the amine to form a salt. Amine that has been reacted with C 0 2 is known as "rich amine". The reaction between primary or secondary amines and C 0 2 is almost immediate. The gas and amine contact each other in an absorber tower, which is typically filled with random or structured packing and is several
CARBON CAPTURE USING AMINE-BASED TECHNOLOGY
123
meters in height. Gas enters at the bottom of the tower and amine at the top. The two flow counter-currently, with C 0 2 being steadily transferred from the gas into the liquid. By the time the gas reaches the top of the tower, it lastly contacts the fresh "lean" amine. The amount of C 0 2 in the gas will be in equilibrium with the C 0 2 in the amine; the less C 0 2 in the amine, the more readily C 0 2 will transfer out of the gas. The gas exiting the absorber, known as "treated gas", will be composed of mainly hydrogen and water and can be vented to atmosphere or incinerated. The C0 2 -loaded rich amine is heated in a lean /rich heat exchanger before entering the regenerator tower. The purpose of the regenerator is to reverse the bond between the amine and C0 2 . The reaction is reversed by adding heat to the amine as it travels downward though the tower. Heat is supplied in the form of steam, which is generated at the bottom of the tower in the reboiler. The reboiler is the largest consumer of energy in the amine plant, and therefore a main focus of plant optimization studies. The reboiler is powered by a heat medium, often low pressure steam, but can also be hot oil, glycol, or even direct fired. Inside the reboiler, the water portion of the amine solution boils and produces steam. The steam then rises though the regenerator tower, supplying heat for the endothermic reaction which breaks apart the bond between the amine and C0 2 . The steam generated in the reboiler has three main purposes: sensible heat - to increase the temperature of the amine to the boiling point, reaction heat - to reverse the bond between amine and C 0 2 reflux heat - steam must be exiting the top of the regenerator in order to sweep the now-liberated C 0 2 out of the tower, and also to provide a source of reflux flow for the system. The regenerated amine leaves the reboiler and is cooled first in the lean/rich exchanger, and further cooled in the amine cooler. It is filtered, and ready for re-injection into the absorber. The gas stream leaving the regenerator is almost pure C0 2 . It can be liquefied and sold, or compressed and pumped underground for long-term/permanent storage or sequestration.
124
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
9.3
Reaction Chemistry
The reactions of C 0 2 in the gas phase with as aqueous amine solution commences with the physical dissolution of C 0 2 into the water phase. The C 0 2 molecule has to be transferred from the gas phase to the liquid phase for any meaningful reaction to occur. Some interfacial reactions are possible, but for the most part reactions are in the liquid phase. From that point on, there are two main reaction pathways for C 0 2 reaction with amine molecules. These are: 1. Acid-Base Reactions. The amine acts as a base to react with carbonic acid, a product of the reaction of water and C 0 2 2. Nucleophilic Reactions. The amine reacts directly with dissolved C 0 2 molecules. Subsequent reactions take place but the initial step is an SN2 reaction Acid-base reactions tend to be extremely fast as opposed to nucleophilic reactions that are usually diffusion controlled however in C 0 2 removal the acid base pathway is slow because the fist the slow generating carbonic acid need to be generated. The first event to take place in a C 0 2 reaction with amine is the transfer of C 0 2 from the gas phase to the liquid phase. Then the C 0 2 molecule suffers hydrolysis to produce carbonic acid and bicarbonate. C 0 2 (gas) <—> C 0 2 (physical reaction of C 0 2 dissolving in water) C 0 2 + H 2 0 <—> H 2 C0 3 (chemical reaction of CO, and water to produce carbonic acid) (che H 2,C0 C0 3 <—» <—> HCÛ3HCO^ 11 + H+1 (chemical dissociation of carbonic acid to form bicarbonate) 9.3.1
Nucleophilic Pathway
This reaction mode involves the direct attack of the amine into the central carbon of C0 2 . The intermediate then suffers an internal H+1 transfer to produce a carbamic acid. This further react with the amine to form the ammonium salt. In the case of a tertiary amine this is not possible (no H), hence there is no reaction. It is important to indicate that the reaction is slower and less efficient as the amine goes from a primary to a secondary amine. This is because the R groups physically block the amine N atom from reaching the C0 2 .
CARBON CAPTURE USING AMINE-BASED TECHNOLOGY
125
Reactions with Primary and Secondary Amines (for example MEA and DEA) R ^ N H + C 0 2 <—> R ^ N H - C O O (nucleophilic reaction of amine with C 0 2 forming an internal salt) RjiySÍH-COO <—> R ^ N - C O O H (internal H+1 transfer forming a carbanic acid) RjRjN-COOH + R ^ N H <—» R ^ N - C O O 1 + R ^ N H ^ 1 (ammonium carbamate salt) Reaction with Tertiary Amines (MDEA) R ^ ^ + C 0 2 <—> R J R J N R J - C O O (nucleophilic reaction to generate an internal salt) RjiySfRj-COO <—> RjiySfH-CCXDRg (no reaction, R transfer is not possible)
RJRJNH-COO
Even though a salt is the product of the acid-base reaction, the product is instantaneously dissolved in the water solvent forming a collection of ions (positive ions are called cations and negative ions are called anions). These ions are highly hydrated with a layer of water molecules making them very stable and avoiding reaction reversal under normal conditions. These reactions however, can be reversed if exposed to the correct conditions in terms of pressure and temperature; this is exactly what takes place in the regenerator.
9.3.2
Acid-Base Pathway (Primary, Secondary and Tertiary Amines)
A different pathway is the acid-base reaction. This reaction takes place not with C 0 2 directly but with the product if the reaction of C 0 2 with water (carbonic acid). This reaction is fast and it is basically independent of the alkanol R substituents in an amine. Some amines will have faster reaction rates depending on inductive effect of the substituents. Acid-base reactivity is slightly faster with tertiary amines, followed by secondary and primary amines. R ^ N H + H 2 C0 3 <—» R1R2NH2+1 + HCO; 1 (ammonium bicarbonate)
126
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Both of these reactions mechanisms (acid-base and nucleophilic) represent how C 0 2 is removed from a stream using an amine solutions. One of the most suitable amines for C 0 2 removal from a gas stream is MEA. This particular molecule is the most reactive via a nucleophic pathway (the N atom is highly exposed to react) and also has good acid-base reactivity. As far as process condition and in order to satisfy both reaction mechanisms outline above, a temperature of 43 - 46 °C has been shown to be the most effective.
9.4
Types of Amine
There are several different types of alkanolamines on the market, all of which will remove C 0 2 from gas streams to one degree or another. Given the unique properties of post-combustion gas from a power plant however, only a couple of amines are applicable to this process. Monoethanolamine (MEA), a primary amine, is the most proven and reliable C 0 2 removal agent on the market. It is readily available, and the lowest price of all the different types of amine. MEA has several advantages as well as disadvantages. Advantages of MEA: 1. Strongest amine. MEA is more reactive than any other type of amine and therefore forms the strongest bond with C0 2 . It is very reliable, and will endure higher levels of contamination or improper operating conditions yet still remove C 0 2 compared to other amines. 2. Low molecular weight. MEA is not a complicated molecule, and the molecular weight of 61 is the lowest of any amine. Because of this, even a low strength MEA solution is very potent. Running MEA at low strengths but still being able to remove C 0 2 is advantageous in that it minimizes the risk of corrosion due to degraded MEA and heat stable salts. 3. Price and availability. MEA is the least expensive as well as most readily available of all amines. 4. Online reclaiming. MEA can be reclaimed at atmospheric pressure, something which cannot be said for
CARBON CAPTURE USING AMINE-BASED TECHNOLOGY
127
secondary or tertiary amines. Atmospheric reclaiming is used to remove degradation products and heat stable salts which would otherwise build u p in the solution, increasing the corrosive tendency and reducing the strength of the bond between amine and C0 2 . Disadvantages of MEA: 1. High heat of reaction. Although the strong bond between MEA and CO z is an advantage in the absorber, it is a disadvantage in the regenerator as it requires quite a large amount of energy to break the MEA:C0 2 bond. More so than any other type of amine. 2. Degrades in the presence of C 0 2 and oxygen. Both of these components are present in post combustion gas, so special equipment is needed to keep the degradation products from building to corrosive levels. Another type of amine which would work in post-combustion gas C 0 2 capture is "promoted" MDEA. MDEA is a tertiary amine, which, ironically, has historically been used in situations where C 0 2 removal is NOT desired. MDEA itself does not react directly with the C 0 2 molecule. With the addition of a promoter however (most commonly piperazine), MDEA has the ability to remove C0 2 , but without the two disadvantages of MEA; degradation and high heat of reaction. Advantages of promoted MDEA: 1. Lower heat of reaction compared to MEA, therefore less energy intensive 2. Does not degrade in the presence of C 0 2 Disadvantages of promoted MDEA: 1. Still degrades in the presence of oxygen, and will form corrosive heat stable salts as well 2. Online reclaiming to remove oxygen degradation and corrosion products not possible 3. Very expensive and in lower supply compared to MEA
128
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
9.5
Challenges of Carbon Capture
Although theoretically possible, utilizing amines to remove C 0 2 from the post combustion gas created in coal fired power plants is challenging for several reasons: 1. Low pressure gas increases difficulty of transferring C 0 2 from the gas into amine 2. Oxygen content of the gas can cause amine degradation and acid formation 3. C 0 2 degradation of primary (and secondary) amines 4. High energy consumption 5. Very large facilities 6. Finding a suitable location for the removed C 0 2 Comments on each challenge: Low pressure of absorber: because partial pressure is the "driving force" to transfer C 0 2 from the gas phase into the liquid phase, successfully achieving this under low pressure conditions can be difficult, though it depends somewhat on the target level of C 0 2 in the treated gas. Deeper levels of C 0 2 removal require lower C 0 2 loaded lean amine. This can only be accomplished by increasing the energy duty of the reboiler. Degradation: The oxygen content of the inlet gas will react with the amine and form degradation products and heat stable salts. All amines degrade in the presence of oxygen unfortunately. C 0 2 also causes degradation, but only in primary and secondary amines (MEA and DEA). The degradation products and heat stable salts can increase the viscosity of the amine (requiring more energy to heat and cool) and possibly increase the corrosiveness of the solution. Furthermore, degraded amine is not "amine" anymore and not useful in removing C0 2 . Managing degradation is typically a combination of these techniques: • Prevention • Reclamation • Purge and replace amine
9.5.1
Prevention
The prevention or minimization of degradation (by either 0 2 or C0 2 ) is mainly done by keeping the temperature of the absorber to
CARBON CAPTURE USING AMINE-BASED TECHNOLOGY
129
a minimum. When C 0 2 reacts with amine, it is exothermic, meaning heat is generated. In order to quench the temperature, interstage coolers are used, whereby the amine is taken out of the middle of the absorber, cooled, and re-injected. There can be more than one cooler depending on the height of the tower. 9.5.2
Reclaimers
Thermal reclaimers are used on MEA systems to remove degradation products and heat stable salts. The reclaimer looks much like the reboiler, but runs at a higher temperature. The amine and water are vapourized and returned to the regenerator, but the higher boiling point degradation products remain behind. When the reclaimer is full of degradation products, it is dumped, cleaned, and put back online. Normally 1-3% of the circulation solution is reclaimed. Heat stable salts can also be removed in the reclaimer if caustic is added to the amine solution beforehand. The strong caustic replaces the amine in the HSAS molecule, thus freeing the amine and "neutralizing" the salt. The salt does not boil and is therefore removed in the reclaimer. MDEA cannot be reclaimed at atmospheric pressure, because the boiling point of the amine and the degradation products are too close to the same. 9.5.3
Purging and Replacing A m i n e
This is an "easy" fix to the build up of degradation and heat stable salt products, but should only be done as a last resort. There is cost associated not only with the purchase of fresh amine, but also with the disposal of the current. With proper temperature control of the absorber and operation of the reclaimer, purging amine should not be necessary. 9.5.4
H i g h Energy C o n s u m p t i o n
Energy is consumed in amine systems primarily by the reboiler, because heat is necessary to reverse the reaction between amine and acid gas. Along with reaction heat, the reboiler is also required to produce sensible heat as well as reflux heat (which is the left over steam used to sweep the C 0 2 out of the regenerator). The sensible heat duty [Q=m*c*(T2-Tl)] is the single largest component out of the three. Sensible heat can be kept to a minimum
130
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
by optimizing "m", the mass or circulation rate of amine, and the temperature difference between when the amine enters the regenerator and inside the reboiler. This is the function of the lean/rich exchanger. Amine circulation rate can be minimized by operating the MEA at a high strength of 30% (with a reclaimer online to control degradation products) and a molar rich loading of 0.4 moles of C 0 2 per mole of MEA.
9.5.5
Size of the Amine Facility
Compared to standard gas plants or refineries, carbon capture facilities are much much larger. This is because of the large volume of gas which must be processed as well as the low pressure of the gas. Papers published on the design of carbon capture amine plants list some key design details which can reduce the cost of the amine plant construction: • The inlet gas does not need to be cooled. In typical amine operation, the amine temperature is maintained at approximately IOC warmer than the inlet gas. This is to prevent the condensation of hydrocarbons • Utilizing structured packing has been shown to result in the highest gas capacity for a given area. Therefore if structured packing is specified in the design, the absorber will be a smaller size compared to random packing. • Utilize plate and frame exchangers instead of shell and tube. • To minimize energy to the reboiler, the lean amine can be "flashed" in a separation vessel, whereby the steam and heat being flashed is returned to the regenerator. This may reduce reboiler duty by up to 30%.
9.5.6
Captured C 0 2
The final challenge in this process, is that once the C 0 2 has been removed from the gas, what to do with it? In some areas, there is a market for C 0 2 as a miscible flood in oilfields. The C 0 2 is injected downhole and helps loosen the oil, making it flow easier. This
CARBON CAPTURE USING AMINE-BASED TECHNOLOGY
131
is known as C 0 2 flooding, and there are several successful case studies. If no oilfields are in the vicinity of the power plant however, then the C 0 2 must be sequestered underground. Large underground caverns are used, but the C 0 2 may have to be piped long distances.
9.6
Conclusion
Removing C 0 2 from gas streams using MEA or activated MDEA is a proven, tried-and-true technique. For the most part, existing technology could be used for this application, with perhaps the addition of interstage coolers on the absorber. None of the individual challenges facing carbon capture are unique to the amine industry, however the combination of challenges is. Very large facilities, low pressure treating, combined with oxygen and C 0 2 degradation plus the added problem of what to do with the captured C 0 2 has not been done in any wide-spread way. The technology and know-how exists to handle all of these problems however, and with proper operation to minimize energy consumption amine based technology provides effective, long-term C 0 2 removal from post-combustion gas streams.
This page intentionally left blank
10 Dehydration-through-Compression (DTC): Is It Adequate? A Tale of Three Gases Wes H. Wright John M. Campbell & Company, Weyburn, SK, Canada
Abstract There are many similarities between the disposal of carbon dioxide and hydrogen sulphide mixtures into subsurface reservoirs and the injection of carbon dioxide into oil reservoirs to act as miscible fluid. Experience gained in one activity can often be applied with very little modification to the other. There are however, instances where the unique characteristics of each process require special attention. One such area is the dehydration of the injected fluid. Dehydrationthrough-compression is the practice of removing free water at compressor inter-stage scrubbers to produce an under-saturated acid gas at supercritical or liquid conditions. It has been employed in acid gas injection schemes (AGI's). The same technique is frequently evaluated for use in C 0 2 miscible floods. This paper examines some of the considerations when deciding if a gas is sufficiently dehydrated. Using data from commercial process simulation software (AQUAlibrium© (1) and ProMax® (2)), the effectiveness of Dehydration-through-Compression (DTC) for three gases: two acid gases (high-H 2 S and high-C0 2 contents), and a "typical" recycle gas from a mature C 0 2 miscible flood are evaluated.
10.1 Background C a r b o n dioxide has been injected commercially as a miscible agent to i m p r o v e oil recovery since 1972 (3) (4). A d d i t i o n a l C 0 2 miscible
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (133-154) © Scrivener Publishing LLC
133
134
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
floods have been initiated subsequently at a frequency that could be correlated to the price of oil. As existing oilfields age and concerns over C 0 2 emissions increase, C 0 2 miscible flooding is attracting new interest. Acid Gas Injection (AGI) is relatively more recent, commencing in the late 1980's (5). As environmental regulations tightened and progressively more marginal gas fields were developed against a backdrop of collapsing sulphur prices, alternate methods of disposing of the acid gases removed from natural gas streams were required. Disposal of acid gases into abandoned oil or gas reservoirs or suitable subsurface aquifers has provided an effective alternative to conventional sulphur plants or incineration. Figure 1 shows a simplified schematic for an acid gas injection system. Natural gas containing excessive concentrations of H2S a n d / o r C 0 2 is treated to produce a saleable natural gas stream. The acid gas is compressed, dehydrated (note that wet acid gas injection is not considered here) and transported to an injection or disposal well where it is delivered to the disposal zone. Figure 2 is a similar schematic showing the configuration for a typical C 0 2 miscible flood. C 0 2 is injected into the oil-bearing formation where it mixes with oil. The oil-C0 2 mixture moves through the pore system of the rock and is produced at production wells. The C 0 2 evolves from the oil in the production wells, flowlines and separators, before being compressed, dehydrated and recycled into the reservoir.
Figure 1. Acid gas injection schematic.
DEHYDRATION-THROUGH-COMPRESSION (DTC)
135
Figure 2. Carbon dioxide miscible flood schematic.
In each, the acid gas is separated from the saleable product, compressed, dehydrated and transported to the injection well(s) where it is carried to subsurface formations. Wellhead pressure varies, but in general, wellhead pressures on C 0 2 miscible floods tend to be higher than those in AGI's. This is in large part due to the high reservoir pressures required for miscibility, whereas AGI's often target under-pressured, depleted reservoirs. One of the primary reasons for dehydrating the acid gas in both cases is to reduce corrosion thereby allowing the use of carbon steel components. Other reasons include reducing the risk of hydrate and ice formation within the system, and in some extreme cases, accumulation of condensed liquid water which could result in multiphase flow problems. AGI's and C 0 2 miscible floods differ in two important ways: • the number of injection wells, and • the composition of the injectant. Typically, AGI's have few injection wells. C 0 2 miscible floods often have tens or hundreds of individual injection wells distributed throughout the oilfield. Consequently, the C 0 2 injection systems can incorporate many kilometres of pipeline, booster stations and numerous metering stations. With a single injection well in relatively close proximity to the compression facilities, it is possible to control the operating temperature of the injection system to
136
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
maintain the temperature above the water dew point and hydrate formation temperatures. As the size and complexity of the injection system increases, this option becomes increasingly impractical and reliance on dehydration becomes more critical. In addition, as the size of the injection system increases, the cost of addressing corrosion issues through the use of corrosion resistant alloys becomes prohibitive and the use of inhibitors to avoid hydrates becomes more difficult to administer. Gas composition is another key difference between AGI's and C 0 2 miscible floods. The composition of the injectant in an AGI project tends to remain relatively constant over time. In C 0 2 miscible floods however, unless the recycle gas is processed to remove contaminants, the composition of the re-injected fluid varies continually during the life of the flood. Initial injection is usually with a purchased supply of dry, high-purity C0 2 . As C 0 2 is produced from the reservoir, it contains light hydrocarbons and potentially nitrogen, helium, H 2 S, and other contaminants acquired from the reservoir fluids. Figure 3 illustrates the variation of C 0 2 injection volumes, total Gas-Oil-Ratio (GOR) and the concentration of C 0 2 in the recycle gas over the life of a C 0 2 flood in the Permian Basin, Texas. Notice that the C 0 2 concentration in
Figure 3. Example production profile for a permian basin C 0 2 miscible flood (6).
DEHYDRATION-THROUGH-COMPRESSION (DTC)
137
the recycle gas varies dramatically over time. In the early years, the composition cycles from below 10% up to 20 or 50% several times. This is largely an artefact of the development strategy. C 0 2 floods are usually rolled-out across the field, meaning that the flood initiates in a small area of the reservoir, and then as recycle gas becomes available the flood is expanded into new regions of the field. In the early years, recycle gas volumes are relatively low so small amounts of C 0 2 returning with the produced fluids has a large impact on the gas composition. As the flood moves into new regions of the field, the gas from those regions contains less C 0 2 and the C 0 2 content of the recycle gas drops until C 0 2 begins to be produced from those production wells. Eventually, as the flood matures, the C 0 2 concentration becomes less volatile but tends to increase over time. The continual variation in gas composition has many implications for the design and operation of the flood. Of particular interest is the impact on the water saturation characteristics of the recycle gas. Acid gas can be dehydrated using glycol (and glycerol) absorption, molecular sieve and silica gel adsorption, and by Dehydrationthrough-Compression (DTC). Tri-ethylene Glycol (TEG) absorption is the most common method for existing CÓ 2 miscible floods and has been used in some AGI's. However, TEG dehydrators in acid gas service have a history of operational problems. Acid gases are much more soluble in the glycol than conventional natural gases which can lead to acidification of the glycol, corrosion and high glycol losses. In addition, the glycol has proven to be very effective at stripping iron sulphides, asphaltenes, waxes and other solids form the gas stream causing plugging in filters, heat exchangers and still columns. As a result, acid gas dehydrators generally require more operator attention than their sweet natural gas cousins, and the presence of H2S increases the risk levels for all operator interventions. DTC has numerous advantages over TEG or even molecular sieves. Since the water is condensed out of the acid-gas in interstage scrubbers within the existing compression train, DTC eliminates the contactor, reboiler, filters, pumps, heat exchangers and all associated instrumentation, piping etc. In addition to the obvious cost savings, safety and environmental risks are reduced because the system contains fewer potential leak sites. With fewer pieces of equipment, there are fewer operator interventions, each of which
138
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
has the potential to result in exposure of personnel to toxic H2S (assuming H2S is present in the gas).
10.2
Water Saturation
The water saturation of most gases decreases as pressure increases at constant temperature. The higher the pressure, the less water the gas is capable of carrying. Acid gases behave similarly at lower pressures and typical operating temperatures. However, liquid or dense phase acid gases tend to have significantly higher water saturation. This phenomenon is well documented, and forms the basis for Dehydration-through-Compression. Removing free water from an acid gas stream while in the vapour phase, then pressurizing the stream to become liquid or dense phase can provide an under-saturated (with respect to water) fluid. The degree of under-saturation is dependent on the pressure and temperature at which free water is removed. It is also very strongly influenced by the composition of the acid gas. In general, this phenomenon is most notable in high-H 2 S gases, slightly less in high-C0 2 gases and diminishes with increasing concentrations of hydrocarbons or other contaminants. Figure 4 shows the water saturation of pure C0 2 , pure H 2 S, pure methane (Cl) and a mixture of 50% C 0 2 and 50% Cl all at 25°C for various pressures as predicted by AQUAlibrium©. The "knee" in the saturation curve is obvious for both pure H2S and pure C0 2 . The dotted line indicates the phase change as the acid-gas vapour transitions to liquid. The curve for pure Cl shows the continual decline in saturation with increasing pressure. As the concentration of non-acid gas contaminants in the C 0 2 increases, the reversal of the curve becomes less pronounced and (as shown) when the C 0 2 is diluted with 50% methane, a minimum is almost indiscernible on the graph. This has serious implications for the design of facilities for AGI and C 0 2 miscible floods, but is most significant for the latter where the purity of the recycle gas can vary significantly.
10.3
Is It Adequate?
The operating conditions of the injection system dictate the maximum water dew point temperature a n d / o r hydrate temperature required for the injected fluid to be safely and efficiently handled.
DEHYDRATION-THROUGH-COMPRESSION (DTC)
139
Figure 4. Water saturation at 25°C versus pressure for various gases.
Often facilities design engineers are most familiar with dealing with this in terms of "allowable water content". All major gas transmission lines and sales gas contracts place limits on the amount of water that can be contained in the gas. In Canada, this is commonly stipulated as an allowable water content of 4 pounds of water per million standard cubic feet which translates to a dew point oí approximately -9°C (18°F) at 6.9 MPa (1000 psia). Some contracts limit the maximum water dew-point temperature at a specified pressure (e.g. 15°F at 500 psig, -9.4°C at 3450 kPa). One authoritative source has indicated that the lowest buried pipeline temperature that has been measured in operating conditions is +25°F (-4°C) so these specifications provide a margin of safety (7).
140
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
As illustrated in Figure 4, the actual water dew point is dependent on the composition of the gas. Specifying the maximum allowable water content does not translate directly to a dew-point temperature unless the composition of the gas is also specified. Experience suggests that other factors in the sales gas specification (i.e. limits on H2S and C 0 2 content, heating value, etc.) effectively limit the range of compositions sufficiently that the water content specification has proven adequate. Now, in designing a C 0 2 or Acid Gas Injection system, the engineer could adopt the dew-point temperature criterion used for natural gas (e.g. 15°F at 500 psig, -9.4°C at 3450 kPa). However, Figure 4 indicates that within the range of typical injection pressures for AGI's and C 0 2 floods (>8MPa), high H 2 S/C0 2 content gases are capable of retaining significantly more water than methane. A higher allowable water content is therefore justifiable. Due to the wide range of gas compositions, injection pressures and injection system configurations, a single water-content specification is unrealistic. It is therefore left to the individual owner and designer to define an acceptable limit on a project-by-project basis. This should be based on the expected operating conditions (pressure, temperature and composition) for the system and needs to include a safety margin. A review of what could be expected during upset conditions is also required. One possibility is to identify a range of pressures and temperatures expected in the system. This range could then be overlaid on Pressure-Temperature diagrams (including water dew point and hydrate curves) for the expected fluid compositions. Alternatively, individual, worst-case calculations could be performed to verify that neither solid (hydrate or ice), nor liquid water will pose a problem. The variation of water saturation with acid-gas fluid-phase requires particular care when phase changes are expected within the injection system. If the system normally operates as a liquid, but could operate in the vapour region, more stringent water saturation limits may be required. Alternatively, process controls could be utilized to ensure the system does not operate outside of an allowable range of pressures and temperatures. To demonstrate the potential effectiveness of Dehydrationthrough-Compression (DTC), three gas compositions were selected for evaluation. Two are intended to represent a range of Acid Gas Injection streams (a high H2S case and a high C 0 2 case). The third gas is intended to be a representative Recycle Gas from a mature
DEHYDRATION-THROUGH-COMPRESSION (DTC)
141
C 0 2 miscible flood. The recycle gas contains hydrocarbons that evolve with the C 0 2 from the produced fluids. It should be noted that the graphs and tables DO NOT represent experimental data. ProMax® from Bryan Research and Engineering, and AQUAlibrium© from FlowPhase Inc. were used to generate graphs and data. The operating conditions will be simplified (dramatically) to include a single operating point which has been chosen somewhat arbitrarily, although the intent is to represent conditions at a "typical" injection well during winter operating conditions. The operating point is: ASSUMED OPERATING POINT: 0°C at 14 MPa (32°F at -2000 psi)
10.4 The Gases Table 1 introduces the three gases used in the evaluation. Compositions were designed to provide analogues for two acid gas injection systems and one mature C 0 2 miscible flood recycle gas. ProMax® was used to generate phase diagrams (assuming no water) for each gas and these are presented as Figures 5, 6 and 7. Notice
Table 1. Properties of representative gases. Acid Gas Injection Component
High H 2 S Gas
High C 0 2 Gas
C 0 2 Miscible Flood Recycle Gas
H2S
78
21
0
co2
21
78
82
Cl
0.9
0.9
10.2
C2
0.1
0.1
3.5
C3
1.6
iC4
0.3
nC4
0.7
(Continued)
142
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 1. Properties of representative gases. (Continued) Acid Gas Injection Component
High H2S Gas
High C 0 2 Gas
C 0 2 Miscible Flood Recycle Gas
iC5
0.3
nC5
0.5
C6
0.9
Total Molecular Weight Pseudo-Critical Pressure (MPa) Pseudo-Critical Temperature (°C)
100
100
100
36.0
41.7
41.4
9.2
7.7
7.9
81
37
Figure 5. P-T Phase diagram for high H2S gas (without water).
28
DEHYDRATION-THROUGH-COMPRESSION (DTC)
Figure 6. P-T Phase diagram for high C0 2 Gas (without water).
Figure 7. P-T Phase diagram for recycle gas (without water).
143
144
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
the impact of the heavier hydrocarbons on the shape of the phase diagram for a typical recycle gas (Figure 7). As the concentration of C2+ components increases, the envelope becomes much wider, deviating significantly from the classic "banana-shaped" curve of most Acid Gas Injection systems (as seen in Figures 5 and 6). AQUAlibrium© was used to develop the water saturation curves for these gases and the results are shown in Figures 8, 9 and 10. Each figure shows the water saturation at different temperatures versus pressures ranging from 0.2 MPa up to 16 MPa.
Figure 8. Water saturation for high H2S gas.
DEHYDRATION-THROUGH-COMPRESSION (DTC)
145
Figure 9. Water saturation for high C 0 2 gas.
The effectiveness of DTC is heavily dependent on the pressure and temperature attained at the point where free-water is removed. Since this point exists at a compressor inter-stage pressure, actual operating pressures do not always align with the minimum on the water saturation curve. Additionally, the designer must allow for potential for hydrates and the non-aqueous liquid phase envelope (8) (9) (10) (11) (12). In Figures 8-10, dotted tie-lines lines have been inserted through the region where a non-aqueous liquid phase may be present.
146
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 10. Water saturation for recycle gas.
The water saturation of the vapour may actually decrease further within this region, but the condensed non-aqueous liquid will contain large quantities of acid gas. This non-aqueous component requires special design consideration due to potential for freezing in the liquid drain lines as the non-aqueous liquid flashes across the d u m p valve. Acid gases condensed in the scrubber as non-aqueous liquid will also have to be recycled through the compression train. The
DEHYDRATION-THROUGH-COMPRESSION (DTC)
147
pressure and temperature at which the non-aqueous liquid will appear will vary with relatively small changes in acid gas composition (see Figure 7 for the impact of heavy hydrocarbons on the dew point curve) so caution is recommended when selecting how close the operating point will be to the to the phase envelope. The actual operating-point selected for the removal of the freewater also needs to consider the temperature that can be attained. If aerial coolers are used for inter-stage cooling, there will be a significant range of possible operating temperatures. An after-cooler outlet temperature of 35°C (95°F) seems a reasonable assumption (if not somewhat optimistic) and will be used for this evaluation. Lower temperature isotherms are shown on the graphs for illustrative purposes. If a suitable source of cooling is available, and attempts are made to "take a deeper cut" in order to condense more water, the caveats in the previous paragraph about hydrates and non-aqueous liquids should be reviewed.
10.5 Results Figure 11 provides a simple schematic of the simulations. The gas stream was saturated with water at low pressure and elevated temperature. The water-saturated gas was compressed and then cooled to desired operating conditions before entering a two phase separator. The selected operating points for the separation (which is to perform the role of 'dehydrator") for the three gases are presented in Table 2, along with water saturations, hydrate temperatures. Also included
Figure 11. Simulation schematic.
148
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
in Table 2 are the water dew point and hydrate temperatures for the dehydrated gas streams at 14 MPa (the assumed injection wellhead pressure). Saturated water content for each gas at 14 MPa and 0°C is also shown to illustrate the degree of under-saturation attained. Table 2 shows that Dehydration-through-Compression (DTC) is effective in this example for the High H2S Gas. By separating free
Table 2. S u m m a r y of results.
Separation (Dehydrator) Conditions
High H2S Gas
High C0 2 Gas
Recycle Gas
MPa
2.5
5.5
6.5
°C
35
35
35
2,240
1,460
1,260
lbs/ mmscf
140
90
80
Water Dew Point @ 14 MPa
°C
-19**
_9**
+19
Hydrate Temperature @ 14 MPa
°C
-5
1
+15
mg/ Sm3
4,630
1,900
660
lbs/ mmscf
290
120
40
%
48%
77%
191%
Pressure Temperature Water Content of Vapour
Injection Conditions
Saturated Water Content @ 14 MPa and 0°C Degree of UnderSaturation +
Sm3
Notes: Calculated as the water content of the dehydrated gas divided by the saturated water content of the gas at 14 MPa and 0CC. **Water dew points below hydrate or ice formation temperatures should be used with caution and are provided here only for interest only. ProMax® offers the following warning: "The water dew point temperature is thermodynamically unstable and will not form a free aqueous phase"
DEHYDRATION-THROUGH-COMPRESSION (DTC)
149
water from the gas stream at 35°C and 2.5 MPa, the resulting vapour contains only 48% of the water required to saturate the stream at the design conditions of 0°C and 14 MPa. The results for the High C 0 2 Gas are less encouraging. Although the "dehydrated" gas contains only 78% of the saturated water content at the design conditions, the calculated hydrate temperature is 1°C. Free water should not be present, but potential exists for hydrates to form in the injection system. The Recycle Gas fares poorly in this analysis. The water content of the "dehydrated" gas contains 191% of the water required to saturate the gas stream at 0°C and 14 MPa (i.e. the gas is over-saturated so free-water or hydrates will be present). The simulation predicts that as the gas leaving the compressor station cools, liquid water will begin to form when the temperature falls below 19°C. The results can also be viewed on Pressure-Temperature phase diagrams. Figures 12,13 and 14 show the liquid dew-point curves, bubble-point curves as well as the hydrate and water-ice curves for the three gases as they leave the dehydration separator.
Figure 12. P-T Phase diagram for dehydrated high H2S (including water).
150
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 13. P-T Phase diagram for dehydrated high C 0 2 Gas (including water).
Figure 14. P-T Phase diagram for dehydrated recycle gas (including water).
DEHYDRATION-THROUGH-COMPRESSION (DTC)
151
Once these phase envelopes are generated for the dehydrated gas, operating conditions at various points in the system can quickly be checked to confirm they are not within the hydrate or liquid water regions. An operating window, consisting of minimum and maxim u m pressures and temperatures could be overlaid on the phase diagram and potential problem areas identified.
10.6 Discussion Based on the simulation results, Dehydration-through-Compression can be effective at delivering an under-saturated acid gas at supercritical or liquid conditions. The process is particularly applicable to high-H 2 S content gases, and slightly less so for high-C0 2 content gas. However, the process becomes less effective as the hydrocarbon content of the gas increases. Note that the results in Table 2 do not include a safety margin. Typical recycle gas in a C 0 2 miscible flood contains a considerable amount of hydrocarbons, nitrogen and other contaminants. For the representative C 0 2 recycle gas examined here (82% C 0 2 : 18% hydrocarbon), separating free water at 35°C and 6.5 MPa yielded an injection stream at 14 MPa with a water dew point of approximately +19°C, and a hydrate temperature of +15°C. These results would suggest that given the extent of the injection system commonly seen in C 0 2 miscible floods, it is likely that the injectant would cool to temperatures below the water dew point a n d / o r hydrate temperatures - at least during Canadian winters and in non-flowing dead-legs. The performance of DTC is sensitive to the composition of the acid gas and each acid gas injection scheme may have unique operating conditions. Careful review and analysis is required to assess the adequacy of DTC to produce a sufficiently dehydrated injectant. In schemes where the injectant could see variations in pressure and temperature a n d / o r composition, a range of cases may need to be evaluated. Special care is required when generating the phase envelopes for acid gases that include water. The phase behaviour can be complex, with multiple liquid phases, solids and vapour. Different simulators handle the existence of multiple liquid phases in different ways and it may be necessary to confirm with the software vendors to confirm and to interpret the output (13) (14).
152
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
It should also be noted that experimental data are sparse for many areas of these graphs. Hydrate temperatures for undersaturated, high-pressure acid gases, (with or without hydrocarbon contaminants) is a good example. Simulations are based on proven algorithms, but without actual data to tune the models, there will always be some question as to their validity. Verifying simulation results with some actual experimental data is strongly recommended. Safety margins commensurate with data confidence are advisable.
References 1. AQUAlibrium v3.1 Engineering Soßware. #330,2749-39th Avenue N.E., Calgary AB, CANADA T1Y4T8, www.flowphase.com : FlowPhase Inc. 2. ProMax v3.1 Process Simulation Soßware. P.O. Box 4747, Bryan TX, 77805, USA www.bre.com : Bryan Research & Engineering, Inc. 3. SPE18977 Summary Results ofCOl EOR Field tests, 1972-1987. Brock, W. R. and Bryan, L. A. Denver CO : SPE, 1989. 4. Udvardi, Geza, et al. C 0 2 Dehydration Scheme Aids Hungarian EOR Project. Oil & Gas Journal. Oct. 22,1990. 5. Bachu, Stefan and Gunter, William D. Overview of Acid-Gas Injection Operations in Western Canada. 6. Gas Handling Options & Issues. Wehner, Scott. Midland, TX : s.n., Dec. 5-6,2000. C 0 2 CEED Conference. 7. Younger, A.H. Natural Gas Processing and Principles, s.l. : The University of Calgary, 1989. 8. C02-Rich Mixtures Part 1 C02 Streams Pose New Design Considerations. Case, James L., Ryan, Bernard F. and Johnson, John E. May 6, 1985, s.l. : Oil & Gas Journal, 1985. 9. C02-Rich Mixtures Part 2 Water in High-C02 Stream Complicates Desogn Factors. Case, James L., Ryan, Bernard F. and Johnson, John E. Mayl3,1985, s.l. : Oil & Gas Journal, 1985. 10. Carroll, John J. Natural Gas Hydrates - A Guide for Engineers, 2nd Ed. Oxford : Gulf Professional Publishing, 2009. p. 276. 11. An Analysis of Hydrate Conditions and Property Predictions in Acid Gas Injection Systems. Hendrick, Cory, et al. Austin, TX : GPA, 2010. 12. Acid Gas Water Dew Point, Hydrate and Physical Property Predictions. SnowMcGregor, Kindra and Johnson, Johnny. San Antonio, TX : GPA, 2004. 13. Proper Interpretation of Freezing and Hydrate Prediction Results from Process Simulation. Hlavinka, M. W., Hernandez, V. N. and McCartney, Dan. Grapevine, TX : 85th Annual GPA Convention, 2006. 14. Carroll, John J. Personal communications. August 2010. 15. Salt Creek C02 Enhanced Oil Recovery (EOR) Project. Tatarzyn, Jeff and Brimm, Allen. San Antonio, TX, USA : GPA, 1995. Annual GPA Convention.
DEHYDRATION-THROUGH-COMPRESSION (DTC)
153
16. Sheep Mountain C02 Project Development and Operation . Watson, Kevin D. and Siekkinen, Dan G. New Orleans, Louisiana : s.n., 1986. AIChE 1986 Spring National Meeting. 17. Sloan, Jr., E. Dendy. Clathrate Hydrates of Natural Gases. New York : Marcel Dekker, Inc., 1998. 18. The Water Content of Acid Gas and Sour Gas from 100°F to 220°F and Pressures to 10,000 psia. Carroll, John J. Dallas, TX : GPA, 2002. 81st Annual GPA Convention, p. 36. 19. Campbell, John M. and Hubbard, Robert A. Gas Conditioning and Processing, Volume 1, 8th Edition. Norman, OK : John M. Campbell and Company, 2004. 20. Solid-Liquid-Vapor Phases of Water and Water-Carbon Dioxide Mixtures Using a Simple Analytical Equation of State. Yokozeki, A. Boulder, CO : National Institute of Standards and Technology, 2003. 21. Experimental Pressure-Temperature Data on Three- and Four-Phase Equilibria of Fluid, Hydrate, and Ice Phases in the System Carbon Dioxide-Water. Wendland, Martin, Hasse, Hans and Maurer, Gerd. Kaiserslautern, Germany : J. Chem. Eng. Data, 1999, Vol. 44.
This page intentionally left blank
11 Diaphragm Pumps Improve Efficiency of Compressing Acid Gas and C0 2 Josef Jarosch, Anke-Dorothee Braun LEWA pumps+systems, Leonberg, Germany
Abstract Compression is an important process step in all acid gas injection and carbon capture and storage (CCS) schemes. An efficient method is presented which reduces the significant energy consumption of compression units by replacement of the last compressor stages with a high pressure pump. After liquefaction the pump boosts the pressure in a single step to the necessary pipeline and injection pressure. Since environmental and safety impact of leakages - even small fugitive emissions - are especially critical with highly toxic acid gas and polluted C0 2 , the operators strive for absolutely leak free processes. The diaphragm pump technology has proven in many applications that it is fully qualified to handle even the most difficult, hazardous or toxic fluids. An overview of reciprocating diaphragm pumps is given, with special focus on applications in the acid gas and C 0 2 injection. The methods and technologies used for acid gas and C 0 2 sequestration in geological formations have substantial similarities. The technology and experience developed in acid gas injection operations can be adopted for large-scale C 0 2 sequestration. Increasing attention of the need to protect the environment and more rigorous regulations promote industrial processes with the objective to minimise the emission of harmful substances. Carbon dioxide is the dominating contributor to increased global warming. The effects of climate change are becoming too evident and the Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (155-172) © Scrivener Publishing LLC
155
156
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
level of C 0 2 in the atmosphere is rising as a result of man-caused emissions. Approximately 35% of all C 0 2 emissions originate from fossil fuels for generating electricity. The energy demand is expected to double in coming decades, much of it to be met by fossil fuels. Other industrial processes also generate large amounts of C 0 2 such as steel production, cement industry, fertilizer production and the chemical industry. Additionally the gas industry is increasingly faced with the disposal of C 0 2 received from low quality gas streams, which are sometimes characterised by more than 10 % carbon dioxide. A range of actions will have to be taken in order to bring down the quantity of C 0 2 entering the atmosphere. There is a variety of ways in which C 0 2 emissions can be reduced, for instance by switching to fuels with lower fossil carbon content and increasing the fuel conversion efficiencies of power plants and processes generating C0 2 . Alternatives to fossil fuel such as renewable energies, like solar and wind that produce little or no C 0 2 are growing fast, but cannot supply yet the base load energy demand. The generation of primary energy will continue to be dominated by coal, oil and gas until at least the middle of the century. C 0 2 emissions could be reduced substantially, without major changes in the basic processes by capturing and storing the C 0 2 of fossil fuelled power stations and other large C 0 2 producing single sources. Figure 1 is a simplified schematic of the capture and storage process. The technology promises to be the most effective for reductions in C 0 2 emissions on a large scale. The technology of capture and
Figure 1. Acid gas and C 0 2 capture, compression and storage.
IMPROVE EFFICIENCY OF COMPRESSING ACID GAS AND C 0 2
157
injection is well-known. In the last twenty years oil companies have enhanced production by raising the pressure in ageing oil fields by C 0 2 injection (EOR). The C 0 2 sequestration can build upon the knowledge and experience gained in the oil and gas industry in the last decades. Underground storage has taken place for many years as a consequence of injecting C 0 2 or gas cocktails with large portions of highly toxic H2S [1], [2]. Since C0 2 often represents the largest component in acid gas streams and a separation is costly, large volumes of C 0 2 are injected together with H2S for storage. Over the last 20 years acid gas injection has developed as an effective means to make sure that acid gas is not emitted to the atmosphere. The methods and technologies developed for the acid gas injection can be transferred to the sequestration of pure C0 2 . In all acid gas injection and carbon capture and storage (CCS) schemes the compression of the gases represents an essential subprocess. Today C 0 2 compressors require a large portion of the capital and operating costs of any CCS-system. The compression consumes a significant amount of energy depending on the outlet pressure of the regenerators, pipeline losses, formation pressure and depth and the compression technology chosen. If the reservoir pressure is sufficiently low the C 0 2 and acid gas can be injected as a gas. However, reservoir pressures may be as high as 350 bar, in extreme cases even higher [3]. This requires partly large pressure ratios exceeding 1:200. Low inlet pressures demand large frame compressors and high injection pressures can only be handled with an accordingly high number of stages in the case of a conventional compressor selection. C 0 2 compression is the most energy intensive step of any C 0 2 EOR operation and consumes up to 60-80% of the overall energy requirement. Therefore the definition and design of the most efficient compression technology depends strongly on the upstream separation method used and the geo-mechanics of the storage zone. CCS techniques have energy demands that result in considerable reductions in the net power plant output. The cost penalty for the compression power can be high in the overall efficiency of fossil fuelled power stations. Based on current estimates the C 0 2 compression power for power plants can range from approximately 5 - 12% of the plant rating. A way to reduce this penalty is to consider new compression concepts that raise the pressure to injection levels with the minimal amount of energy required and that can
158
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
be integrated into existing processes. J. Moore and S. Bertolo have explored many different thermodynamic processes and identified optimal C 0 2 compression schemes [4],[5]. There are three equivalent display options to demonstrate multistage compression processes of gases. These are shown in Figure 2. In the following all compression and injection processes from outlet of the regenerators to the storage zone are illustrated in temperature - pressure or phase diagrams. In temperature - pressure charts the phase curves for pure fluids are lines which end in the critical points (A). For binary mixtures the phase curve is given by lining up dew points and bubble points. The gas composition affects the shape of the phase curve. Examples are indicated in (B) for acid gas with various C 0 2 - H2S compositions. In the presence of hydrocarbons the phase envelopes broaden and as a consequence a higher pressure is required to liquefy the acid gas. These phase diagrams are shown in Figure 3. In the acid gas service the knowledge of the characteristic shape of the phase envelope is very important in the design of
Figure 2. Compression of pure gases.
Figure 3. Phase diagrams for pure gases and binary gas mixtures.
IMPROVE EFFICIENCY OF COMPRESSING ACID GAS AND C 0 2
159
compression schemes because the two-phase region must be avoided during the compression by all means. Formation of liquid acid gas represents a high risk of damage to the compressor. Additionally if water is present, C0 2 , H2S and methane can form hydrates, therefore it is necessary to operate above the hydrate forming temperature [6], [7]. In the phase diagrams shown in Figure 4 different compression paths are charted for pure C 0 2 for identical starting conditions at the regenerator exit and identical final conditions of the storage zone. With conditions typical of the exit from amine regenerators the gases leave at low pressures and nearly ambient temperature and are compressed to the necessary pipeline pressure. With a certain pressure drop due to friction losses in the pipeline they reach the injection well. Under the weight of the vertical fluid column the pressure rises to the reservoir condition. a. displays an ideal isothermal compression which would yield the best efficiency. This path could only be approached by an economically inefficient high number of compression steps with intermediate cooling. b. describes a conventional semi - isothermal process with multistage compression and inter-stage cooling
Figure 4. Comparison of different compression paths.
160
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
for reciprocating or turbo compressors. The compression ratio is mostly in a range of 3 - 4. c. shows a conceivable compression solution with a 3 stage shock wave compressor based on supersonic aircraft technology using a high single stage pressure ratio of about 8 - 1 0 [ 8 ]. At the moment a pilot project is supported by the Department of Energy, DOE. The first commercial operation is scheduled for 2014/2015. d. This process utilizes part compression with a few compressor stages followed by a pump. The hybrid approach combines semi isothermal compression, cooling, liquefaction and increasing the pressure in a final single step to the pipeline condition by a pump. Compression in the liquid state consumes less power than in the gaseous state. However the refrigeration power to liquefy and sub cool the gas has to be considered in the energy balance. The power required by each compression option can vary significantly. The energy savings of the improved compression configuration can outweigh the manageable increases in capital cost to incorporate them. By changing the thermodynamics of the compression path in this way the hybrid solution allows significant power savings over a conventional compression in the gaseous state. However there will be no universally valid solution to C 0 2 and acid gas compression duties. Specific storage site conditions such as location, ambient air temperatures, availability of cooling medium etc. will play a decisive role in determining the optimum configuration. In principal different types of pumps can be integrated in the hybrid compression process to accomplish the final step of injection. In Table 1 advantages and disadvantages of different pump types are compiled. Reciprocating pumps offer compression at an unmatched efficiency. Contrary to centrifugal pumps with distinct maximum efficiencies in narrow speed bands reciprocating diaphragm and plunger pumps are characterised by a flat efficiency graph over the whole capacity range. Figure 5 shows the efficiency for various pumps. Although C 0 2 itself is inert and nontoxic, its processing presents some risks: asphyxiation, pressure hazards and danger from the
IMPROVE EFFICIENCY OF COMPRESSING ACID GAS AND C 0 2
161
Table 1. Advantages and disadvantages of different pump types. Advantages Plunger Pumps
• •
high efficienty uncomplicated design
Disadvantages • investment costs • not hermetically tight • flushing required for
co2 •
wear of seal and plunger • no dry running allowed • sensible against particles • flow and pressure pulsations
Diaphragm Pumps
• • •
hermetically tight high efficiency insensible against particles • low maintenance requirements • safe dry operation • low life cycle costs
• investment costs • flow and pressure pulsations
Multi Stage Centrifugal or Single/ two stage high speed Centrifugal Pumps
• •
• •
low investment costs small space requirements • low pulsation
not hermetically tight high maintenance requirements • high life cycle costs • poor efficiency at partial load • extremely sensible against particles
contaminants in the gas stream, mostly highly toxic H 2 S, benzene, toluene, xylene etc. Even small releases can provoke a serious risk to operational staff and the local environment. The only really convincing solution therefore is - as in the case of acid gas - the absolutely leak free handling and transport. To prevent releases to the environment leak free diaphragm pumps have found a wide field of applications in all industries which handle fluids with the potential to be hazardous, dangerous or toxic. Therefore the diaphragm pump technology is fully
162
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 5. Efficiency of different p u m p types.
qualified to convey contaminated C 0 2 and acid gas streams for sequestration or enhanced oil recovery.
11.1 Diaphragm Pumps Diaphragm pumps are mainly used for low to medium flow rates at medium to high pressures. They are predominantly known for low flow - high head applications and for fluids which have to be metered leak free with high accuracy. In the last 25 years however process diaphragm pumps were developed rather unnoticed from the general public to hydraulic powers of several hundred Kilowatts. Today the allowable physical values such as pressure, flow rate and viscosity cover several orders of magnitude. Figure 6 shows the flow rate and pressure ranges for several diaphragm pumps. The typical application limits of plastic diaphragms are pressures u p to 400 bar at temperatures of -50°C to +150°C. Special designs are used in processes up to 800 bar. Pumps with metal diaphragms can reach pressures of 1200 bar at temperatures up to 250°C. Due to the smaller deflection capability of metal diaphragms, the capacities per pump head are limited to approx. 1,5 m 3 /h. Process diaphragm pumps have with few exceptions hydraulically actuated diaphragms.
IMPROVE EFFICIENCY OF COMPRESSING ACID GAS AND C 0 2
163
Figure 7 shows the working cycle for a typical diaphragm pump. The diaphragm is the pumping element. The back and forth deflection is induced by the plunger displacing hydraulic fluid. The hydraulically actuated diaphragm completely separates all susceptible parts of the pump from the process fluid. Leakages in the case
Figure 6. Flow and pressure range of diaphragm pumps.
Figure 7. Working cycle of a diaphragm pump.
164
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
of a damage are avoided by multilayered diaphragms with integral alarm functions. The fluid is still contained in the pump even if one diaphragm is damaged. Other safety features are implemented in the hydraulic part of the pump to render it inherently safe against up-set conditions such as overload or cavitation and even operating errors. The equipment with diagnostic systems and advanced fault detection techniques helps identify possible maintenance requirements and support trouble shooting. In addition to be leak free diaphragm pumps have all the advantages of reciprocating displacement pumps. • pressure firm characteristics (the capacity is nearly unaffected by the backpressure) • linear capacity control by stroke or speed adjustment • high accuracy • high volumetric and mechanical efficiency
11.2 Acid Gas Compression Safety and environmental protection are critical in the design of any high pressure injection, especially for acid gas. The highly toxic nature of acid gas requires rigorous control over possible leaks within the system. Rather unperceived from the public is the employment of diaphragm pumps in the challenging field of acid gas and C 0 2 injection where they are installed amongst others to save operating costs. As far as we know the first diaphragm pump for acid gas compression was commissioned in Alberta in 2001. A 4-stage reciprocating compressor raises the pressure to 55 - 72 bar. After cooling and liquefaction the acid gas is sub cooled by 5° to compensate for the temperature build-up of the pump heads due to heat of compression. In this manner partial cavitation and efficiency loss are avoided. The design flow is 24 m 3 / h at a pipeline delivery pressure of 231 bar. The pressure-temperature diagram for this injection scheme is given in Figure 8. The 3-headed pump is equipped with PTFE multilayered diaphragms. The pressure between the diaphragms is monitored continuously to detect a leak in the process or hydraulic side diaphragm immediately. Pressure transmitters in the hydraulic system of each pump head monitor the function of process check
IMPROVE EFFICIENCY OF COMPRESSING ACID GAS AND C 0 2
165
Triplex diaphragm pump Suction: 55 - 72 barg 231 barg Discharge: Temperature 30 - 43 °C Flow: 22m 3 /ri Turn down: 1 :4 Power consumption: 122 KW
Figure 8. Acid gas compression and injection with reciprocating compressor and triplex diaphragm pump.
valves and the main components of the hydraulic drive system. Since no detailed data about the compressibility and temperature build-up were available the layout of the pump includes an allowance of about 3 m 3 /h. The initial concern that too much C 0 2 and H2S could diffuse through the PTFE sandwich diaphragms proved to be unfounded.
166
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Another application is characterised by extremely varying compositions of the acid gas streams coming from different wells. They are delivered to multiple storage zones with pressures ranging from 300 to 450 bar. To comply for varying gas volumes a turn down ratio of 1 : 10 was realised. Hence, for all components of the compression unit a very high flexibility regarding pressures and flow rates is required. An example of this type of application is shown in Figure 9. Figure 10 presents a general arrangement of a planned compression unit for acid gas.
Suction: 6 9 - 1 1 7 barg Discharge: 224 - 345 barg Temperature 5 - 44 °C Flow: 11,5 m3/h Turn down: 1 : 6,5 Power consumption: 105 KW Figure 9. Acid gas compression with reciprocating compressor and diaphragm pump.
IMPROVE EFFICIENCY OF COMPRESSING ACID GAS AND C 0 2
167
Compressor inlet: 0,81 barg Compressor outlet: 56,59 barg Driving power: 975 KW Pump inlet: Pump outlet: Driving power: Flow rate: Turn down:
55,50 barg 250,00 barg 132 KW 9000 Kg/h 1: 5
Figure 10. 4-stage compressor and diaphragm p u m p for acid gas compression.
The compression train consists of an intercooled 4-stage reciprocating compressor, an adequate cooler for changing the fluid density into the liquid range and a triplex pump with PTFE diaphragms for the final compression step. With this compression unit high outlet pressures u p to 350 bar can be obtained. The proposed combination of reciprocating compressor and reciprocating pump can respond rapidly and precisely to changes in acid gas mass flow.
11.3
C 0 2 Compression for Sequestration
As mentioned above the methods and technologies used for acid gas and C 0 2 injection into geological storage zones bear strong resemblances. The acid gas injection projects provide important practical experience for C 0 2 capture and storage.
168
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
For the compression of C 0 2 the combination of compressor and pump offers similar advantages as in the acid gas case. Weyburn-Midale is an example of a very successful pilot program at commercial scale and the largest carbon storage project which supplies short-term experience with large volume injection of C0 2 . A 3-stage reciprocating compressor boosts the C 0 2 from 13 bar to 75 - 86 bar. The 64-stage horizontal centrifugal pump downstream pressurises the dense phase C 0 2 to the pipe line condition of 160 bar. After the end of the enhanced oil recovery period nearly all injected C 0 2 is permanently stored and monitored. A similar approach with diaphragm pumps was taken in a LNG production facility for the export of liquefied natural gas in Norway. The Sn0hvit field is located in the Barents Sea with the production stations directly installed on the seabed in 300 m depth. The produced raw gas is transported on-shore via a 160 km subsea pipe line for processing and removal of 8 - 10 % C 0 2 contained in the natural gas. The carbon dioxide is compressed, liquefied and pumped back for sequestration into a deep sub-sea formation beneath the gas field for final storage [9]. A simplified schematic of this injection process is given in Figure 11. The initial compression of the C 0 2 leaving the regenerators is accomplished by a 3 - stage integrally geared turbo compressor. After cooling and liquefaction the gas is sent to a 4-headed diaphragm pump at 50 - 60 bar.
Figure 11. Compression and injection of C0 2 .
IMPROVE EFFICIENCY OF COMPRESSING ACID G A S AND C 0 2
169
To avoid partial cavitation the C 0 2 is sub cooled by 5 - 6° before it enters the pump. Two pumps - one duty one stand by - boost the pressure to 215 bar and convey the liquefied gas to the injection zone. The layout of separation and compression units are 110 m 3 / h liquefied C 0 2 (equivalent to ca. 95 tonnes C 0 2 / h ) . The LNG is exported ready for sale by ship. A pressure-temperature diagram for this process is given in Figure 12. The 4-headed pump uses the technique of hydraulically driven PTFE diaphragms mentioned before. This hybrid compression consumes more than 2000 KW less than the conventional compressor process.
Technical data: Compressor inlet: 1 barg Compressor outlet 61 barg 12 000 KW Driving power 60 barg Pump inlet: Pump outlet 215 barg Driving power: 800 KW Mass flow: ca. 100 t/h Turn down: 1 :4 Figure 12. C 0 2 compression with turbo compressor and diaphragm pump.
170
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The actual limits for reciprocating pumps which yield the best efficiencies of all pump types are roughly 120 to C 0 2 / h at 50 bar inlet and 300 bar discharge pressure. With this performance the C 0 2 emissions of many C 0 2 generating processes and fossil fuelled power plants of 200 - 300 MW can be brought to the injection pressure with a single pump. Power plants of 200 - 400 MW would need 2 - 3 reciprocating pumps operated in parallel. The range above u p to 1200 MW - would be the field of turbo pumps, though with significant lower efficiency, especially at part load. An interesting pilot project in Germany explores storage techniques and the efficiency of compression profiles at low temperatures in order to profit from the less energy consuming compression in the liquid phase. Flexibility was built into the project to allow investigation over a large temperature and pressure range. The C 0 2 is delivered by truck. Two medium sized triplex pumps with PTFE diaphragms are installed to pressurize the C 0 2 at temperatures ranging from -20 °C to -48 °C to 30 - 95 bar for storage. This process is shown in Figure 13. Other pilot projects for C 0 2 sequestration are pursued in different countries to help identify promising new concepts. Much research is directed at reducing the costs of CCS. Many current R & D activities are focused on developing and improving the energy and cost intensive C 0 2 sub-processes of separation and compression.
Figure 13. C 0 2 compression with diaphragm pumps at low temperatures.
IMPROVE EFFICIENCY OF COMPRESSING ACID GAS AND C 0 2
11.4
171
Conclusion
C 0 2 and acid gas injection reduces air emissions and has the potential to become a valuable method of disposal once its credibility and acceptability as a safe and reliable long term storage is established. Capture and sequestration of C 0 2 has received increased R & D attention in the last decade as the technology promises to be the most effective for large scale reduction in C 0 2 emissions to stabilize the rising concentration in the atmosphere. Carbon capture and storage technologies are no alternatives to better energy efficiency or increased use of non carbon energy sources. As a transitional solution they offer the possibility to make considerable reductions in C 0 2 emissions without having to abandon the actual fossil fuel based energy infrastructure within a short time. The methods and technologies used for sequestration of pure C 0 2 are analogues to those developed for the acid gas injection. Therefore the CCS schemes can be built upon the knowledge and experience gained in the oil and gas industry in the last 20 years. Operating and electricity costs will increase noticeably if the producers have to restrict venting C 0 2 to the atmosphere. Therefore the main priority for the development and wide-spread application of CCS technology is to minimise the energy requirements and the financial implication of C 0 2 injection. The sub-process of compression is consuming a significant portion of the energy requirements of the overall CCS process. Since significantly less energy is necessary to increase the pressure of a liquid than a gaseous fluid, liquid acid gas and C 0 2 pumping configurations offer the possibility to conceive alternative compression paths with notably less energy use. Reduction of the cost and power requirements for compression will encourage injection of C 0 2 both for existing and future power plants and other large C 0 2 or H2S producing processes. It can be anticipated that substantially more injection schemes of this kind will be planned and implemented in the next years. Diaphragm pumps can contribute in the acid gas and C 0 2 injection to the world wide efforts to keep the emissions of harmful substances from high pressure production plants to a minimum and they offer in combination with upstream compressors a distinct reduction of energy and operating costs [10].
172
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Diaphragm pumps can have a share in the global efforts to develop a technology that could help build a bridge to a low emission energy future.
Literature 1. Stefan Bachu, William D. Gunter: Overview of Acid-Gas Injection Operations in Western Canada; Alberta Energy and Utilities Board, Edmonton, AB, T6B 2X3, Canada; 2. Edward Wiehert, M: Acid Gas Compression and Injection; Course Notes, Frankfurt, October 25, 2002 3. N. Rahimi, RJ. Griffin: Potential for Acid Gas Injection at Kharg Island; Gas Liquids Engineering, Calgary, Alberta, Canada 4. Dr. J. Jeffrey Moore, Ms. Marybeth Nored, Dr. Klaus Brun: Novel Concepts for the Compression of Large Volumes of Carbon Dioxide; Southwest Research Institute, Oil & Gas Journal in June 2007 5. Simone Bertolo: Four Post-Combustion C 0 2 Compression Strategies compared; Carbon Capture Journal Sept. - Oct. 2009 6. Tim B. Boyle, John J. Carroll: Calculation of Acid Gas Density in the Vapor, Liquid, and Dense-Phase Regions; PanCanadian Petroleum Ltd, Gas Liquids Engineering Ltd 7. John J. Carroll, Peter J. Griffin, Saad F Alkafeef: Review and Outlook of Subsurface Acid Gas Disposal; SPE Middle East Oil and Gas Show and Conference 2009, Bahrein 8. Peter Baldwin, Joseph Williams: Capturing C0 2 : Gas Compression vs. Liquefaction; Power Magazine June 1, 2009 9. Matthias Müller: The Goliath of Pumps; TCE Today November 2007 10. Josef Jarosch: Improving Efficiency of Compressing C 0 2 for Re - Injection: Carbon Capture and Storage Making it Happen Institution of Mechanical Engineers London, October 2009
SECTION 3 RESERVOIR ENGINEERING
This page intentionally left blank
12 Acid Gas Injection in the Permian and San Juan Basins: Recent Case Studies from New Mexico David T. Lescinsky 1 ; Alberto A. Gutierrez1, RG; James C. Hunter1, RG; Julie W. Gutierrez1; and Russell E. Bentley 2 1
Geolex, Inc., 500 Marquette Ave. NW, Suite 1350, Albuquerque, NM 87102 (www.geolex.com) 2 Carbon Free Corporation, 11221 Richmond Avenue, Suite 107C, Houston, TX 77082
Abstract Acid gas injection (AGI) is becoming an increasingly popular choice for the disposal of gas processing wastes and C 0 2 sequestration in New Mexico. Four AGIs have been brought on-line and several additional projects have been successfully permitted in the last few years. The first AGI well in northwestern New Mexico has been successfully drilled and is being completed and tested in time for the initiation of operation by year's end. AGI has proven to be a cost-effective and environmentally-beneficial alternative to traditional treatment systems. While New Mexico has an abundance of deep saline aquifers suitable for injection, site selection is complicated by extensive oil and gas production. In this paper we explore reservoir selection and characterization, permitting, design and completion of AGI wells using examples of wells in New Mexico. We also discuss ongoing efforts to register similar projects as permanent C 0 2 sequestration sites and potentially obtain associated carbon credits.
12.1
Background
N e w Mexico h a s a long history of oil a n d gas d e v e l o p m e n t a n d p r o d u c t i o n in t w o p r i m a r y s e d i m e n t a r y basins (Figure 1). The P e r m i a n Basin of west Texas extends w e s t w a r d into southeastern Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (175-208) © Scrivener Publishing LLC
175
176
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. Map of New Mexico showing the location of the San Juan and Permian Basins and the five AGI well locations listed in Table 1.
New Mexico and includes a portion of the Central Basin Platform and the Delaware sub-basin to the west. The San Juan Basin located in the north-central and northwestern portions of the state and is located on the Colorado Plateau. These basins produce oil and both sweet and sour gas. The increased demand for natural gas as a "clean" fossil fuel, combined with discovery of new reserves and increased pipeline capacity to these markets, have resulted in the construction of a large number of natural gas processing plants in these basins over the last 45 years.
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
177
New Mexico has thirteen natural gas processing plants that process sour gas, twelve in the Permian Basin and one in the San Juan Basin. Some of these plants have traditionally used Class II injection wells to dispose of wastewater associated with natural gas processing. The methodology of choice for addressing H2S and C 0 2 in the treated waste acid gas (TAG) from the amine units at natural gas processing plants has been the use of sulfur recovery units (SRUs) to convert this waste stream into native sulfur, C 0 2 and H20 using the Claus process (1). In this process, the waste C 0 2 from the acid gas stream is then released to the atmosphere along with additional C 0 2 emissions from the combustion sources used to keep the SRUs at optimum operating temperature. Many of these units are now 30-40 years old, are expensive to maintain and often have upsets. These upsets require excess flaring of the waste gas stream and result in violations of air quality permits at these plants. In addition, many of these plants are bottle-necked in terms of processing capacity by the capacity of these aging SRUs. These factors, combined with the added pressure on companies to reduce greenhouse gas (GHG) emissions, provide powerful incentives to utilize acid gas injection (AGI) as an alternative to SRUs for treatment and disposal of the waste acid gas from plants which process sour gas. In addition, the GHG reporting rules which go into effect in 2010 (2) contribute to the already existing regulatory and economic drivers which encourage the use of AGI and geologic sequestration of C 0 2 as a technology of choice for addressing both H2S and C 0 2 waste streams at these natural gas processing plants. Geolex, Inc.® (Geolex) is an industry leader in providing comprehensive solutions to the natural gas processing industry in acid gas injection and C 0 2 sequestration. Over the last decade we have developed specialized expertise and significant practical experience in planning and successfully permitting, designing and constructing AGI wells for the natural gas processing industry. These wells include both types of disposal and sequestration systems, TAG-only (dry gas injection) and combined TAG/wastewater (wet gas injection). While many of the same considerations go into the identification and characterization of adequate potential reservoirs for AGI and C 0 2 sequestration, dry and wet gas configurations result in distinctly different equipment and material requirements along with operational considerations and limitations. This paper presents an overview of the critical factors involved in identifying and characterizing potential reservoirs for the geologic sequestration
178
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
of these wastes and the steps involved in successfully planning, permitting, designing and executing both wet and dry gas injection projects. Three case studies have been selected to illustrate the process from initial planning through start-up and operation of both dry and wet gas injection systems. In addition, we provide a synopsis of important considerations and the current potential for registration of carbon credits resulting from the permanent geologic sequestration of C 0 2 associated with both types of AGI projects.
12.2
AGI Project Planning and Implementation
The successful execution of an AGI project requires the implementation of a step-by-step process where the goals of the project and critical constraints are identified and evaluated in the context of specific regulatory requirements and operational limitations unique to each site (Figure 2). In addition to the design, permitting and construction of related surface compression facilities, our process for AGI projects includes the following six steps: • Project Planning and Feasibility Study • Reservoir/Cap Rock Identification and Regulatory Permitting • Well Drilling and Testing • Reservoir and Seal Evaluation • Well Completion and Construction • Documentation, System Start-up and Reporting Each of these steps involves a multi-disciplinary team of professionals working together in close coordination with engineering and operational staff from the facility to ensure that relevant technical and regulatory goals are met on-schedule and within budget.
12.2.1 Project Planning and Feasibility Study The project planning process begins with the definition of project goals and objectives, constraints, regulatory, economic and schedule considerations. This involves a determination of the amount and composition of the acid gas stream to be treated. The amount and composition of the acid gas stream varies with plant capacity and the field gas mix that is being processed at the facility.
Figure 2. Diagram detailing components and general schedule for a typical AGI project.
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
180
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
It is important to consider not only current plant capacity and field gas mix but, to the extent possible, the future potential for capacity expansion and changes in composition of the field gas. Often projects are designed and permitted to handle variable injection volumes over the projected life of a facility. Typical TAG compositions range from 75%-90% C0 2 , 10%-25% H2S and trace-2% hydrocarbons by volume. In addition to volume and composition considerations, it is necessary to determine whether the specific facility requires an integrated solution to the disposal of associated wastewater (in the USA this is referred to as Class II wastewater (3) or whether only acid gas disposal is required. The higher material costs associated with wet vs. dry gas injection must be balanced with the costs associated with handling the two waste streams separately. Wet gas injection requires the use of more expensive materials in most of the well components due to the corrosive nature of the combined acid gas/wastewater stream. Another important consideration when deciding between wet and dry gas injection, is the availability of additional reservoir capacity required to accommodate the added volume of both acid gas and wastewater streams. Once the ranges in volume and composition of the waste stream over the projected life of the facility have been determined, the reservoir requirements can be established. Obviously it is preferable to site the AGI well on or immediately adjacent to the plant; however, sometimes the geologic and regulatory constraints on a technically and economically-feasible sequestration reservoir may require an off-site well location. Most natural gas processing plants are located in established oil and gas provinces where there is a significant amount of subsurface geologic information available to evaluate potential reservoirs; however, it is necessary to consider the degree to which the integrity of the upper and lower seals (cap rock) of the potential reservoir may have been compromised by past or future oil and gas exploration and development. Another important consideration is the proximity to existing mineral, oil and gas reserves and drinking water reservoirs. While adequate sequestration targets are almost exclusively found at depths well below potable groundwater resources, in depleted oil and gas reservoirs or in saline aquifers, care must be taken to assure that these are not impacted by the proposed sequestration of acid gas or acid gas/wastewater. Therefore, the ideal reservoir is a geologic unit which is stratigraphically-located in a manner which it cannot
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
181
be compromised by future exploration/exploitation of oil and gas resources, is located far below potable water resources and has not been significantly penetrated by plugged oil and gas wells whose integrity may be questionable. Also important in the evaluation of the potential reservoir are any natural geologic features such as faulting or fracturing that could create potential paths for the escape of acid gas from the reservoir. As mentioned above, since most gas processing plants are located in developed oil and gas provinces, data required for reservoir/cap rock identification and evaluation are usually available in the form of drilling and geophysical logs from nearby wells, seismic data from previous exploration efforts, records of injection wells, published and unpublished regional geologic studies. These data sources include the US Department of Energy (USDOE), US Geological Survey (USGS), Geological Survey of Canada (GSC), state or provincial oil and gas agencies and well as commercial sources such as Petroleum Information Corporation. In addition to geologic data, production and well location data to evaluate the potential interference with existing oil and gas production is available from provincial or state oil and gas agencies. Domestic and industrial water well data are often available from state or provincial water or environmental agencies and from the Water Resources Division of the USGS and the GSC. The elements of an AGI feasibility study include the gathering and processing of geological and production data to identify and evaluate potential reservoirs within the area of the natural gas processing facility as discussed above. In addition, economic data regarding the costs of well completion and associated above ground facilities and the permitting constraints which apply to the use of the selected reservoir(s) must be developed. Much of the data collected and analyzed for the feasibility study is also used to provide the basis for the permitting application for AGI from the applicable provincial or state agency as described in the following section.
12.2.2 Reservoir/Cap Rock Identification and Regulatory Permitting The identification and preliminary characterization of the most attractive reservoir and associated cap rock is a critical component of the requirements for obtaining the necessary permits for an AGI project and for evaluating the technical and economic risks
182
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
associated with the proposed project. The selected reservoir must have the necessary net porosity, permeability and extent necessary to contain the anticipated volume of acid gas or acid gas and wastewater for the entire life cycle of the project. The reservoir must be contained within a geologic structure or stratigraphie trap that is capable of permanently sequestering the injected volume of wastes. Geologic and production data associated with oil and gas wells in the vicinity of the project are used for the preliminary evaluation of the reservoir and the cap rock until it can be supplemented by site-specific data developed during the drilling and completion of the AGI well. These data are useful in the development of permitting documents required by state or provincial regulatory agencies in order to obtain the appropriate approvals and permits for the drilling, completion and operation of these wells. While permitting procedures and requirements vary considerably with jurisdiction, the fundamental requirement is to demonstrate that the potential reservoir will be adequate to contain the projected volume of injected wastes and the cap and bottom seal rocks will not allow the escape of the acid gas or wastewater into potable groundwater aquifers or contaminate other existing or potential oil and gas producing zones. In addition to the technical demonstrations and regulatory considerations described above, every jurisdiction requires different procedures for the public notice and review of potential AGI projects. For example, for the case studies of projects in New Mexico discussed in this paper it was necessary to submit contingency plans for the potential release of H2S as a result of worst case scenarios developed by the NM Oil Conservation Division (NMOCD) that meet their prescribed requirements. In addition, it is necessary to provide direct written notice by registered mail to all surface property and mineral owners or lessees within a one mile radius of the proposed well and of the required public hearing to evaluate the project permit application once the application has been deemed administratively complete by the regulatory agency. While the nuances of the permitting process are different for each state or province, the overall process can easily take from six months to a year from initial application to final approval by the agency. Approvals are then usually accompanied by periodic testing and reporting requirements as well as maximum flow rate and injection pressure constraints which are a function of the depth and characteristics of the individual reservoir. In addition, pressure
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
183
monitoring requirements to assure well integrity are also typically included as part of the permit approval and operating constraints. The following sections describe the specific design considerations of AGI wells and the detailed characterization of the reservoir and seal rocks which are conducted during the drilling and completion of the wells.
12.2.3
Well Drilling and Testing
A number of methods have been used to drill the AGI wells discussed in this paper, including conventional rotary drilling, use of down hole mud motors and directional devices, and top-drive double-joint rigs. Drilling companies and contractors are selected after a review of the potential companies' safety record, equipment, experience, estimated costs, and availability. Following driller and subcontractor selection, a detailed prognosis is developed describing the schedule and sequence of the drilling program. The document identifies the location, equipment, materials, personnel, and a timeline for progress. Particular attention is paid to the programs for fluids, casing, cementing, coring and logging. A "pre-spud" meeting is held among all of the involved parties to review the plan, schedule, and safety requirements for the project. Although drilling and casing operations are performed by the drilling contractor, specialized subcontractors are usually employed for fluid control, cementing, coring, logging and testing of the well. Drilling fluid ("mud") programs are designed to maximize safe drilling progress, control downhole pressures, protect the target formation from excessive invasion, and to insure a stable borehole during logging and casing operations. Representatives of the mud companies regularly visit the site to test and maintain the fluid characteristics. Mud programs for AGI wells generally do not differ significantly from programs for conventional oil and gas wells in similar geological settings. However; in some circumstances, mud tracer materials are added to the mud in order to allow for the definitive determination that a subsequent formation water sample retrieved after swabbing is free from the influence of mud filtrate which enters the formation during drilling. Although lined, excavated mud pits are still permitted for drilling wells in certain states, closed-loop systems of solids removal and fluid management are currently being used to avoid potential
184
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
contamination of soils or groundwater with drilling fluid and costly close-out requirements for lined mud pits. These systems employ centrifuges to separate the solids from the drilling fluids and have significant advantages in controlling the mud's physical and chemical properties. This reduces water use, the net inventory of mud required during drilling operations, facilitating easy and economical removal and disposal of the solids, allowing some mud to be recycled for subsequent drilling operations. Additional nondrilling advantages of closed-loop drilling include a smaller pad footprint, reduced potential for soil a n d / o r groundwater impacts and streamlined permitting with state or provincial oil and gas agencies (such as the C 144 CLEZ process in New Mexico). The cement contractor installs the cement by pressure methods that force the cement into the annular space between the casing and the borehole. Following a period to allow the cement to set, the well head is capped and tested for prescribed pressures and periods before approval. All cement jobs and pressure tests must be documented and results are generally required to be submitted to the applicable state or provincial regulatory agency. As seen in Figure 3, a typical AGI casing program involves a conductor, surface casing and production casing. The annular space surrounding each casing string is cemented to the surface, and the casing and cement integrity are determined by pressure testing following each cement completion. The conductor is typically set into competent bedrock at depths of 50 to 150 feet. This initial casing element is used to support subsequent casings, and to provide a seal for attaching the BOP used during drilling. Following the installation and testing of the conductor, drilling progresses with a smaller-diameter bit to the target depth for the surface casing. To protect drinking water resources, the surface casing depth is selected to exceed the locally known depth of potable groundwater or other shallow mineral resources such as aggregate or coal, commonly from 500 to 1300 feet. Cementing to the surface and pressure testing are also generally required at this stage. Once the surface casing is installed, cemented and tested, the final bit size is used to advance the well to the target injection zone. As described below, logging while drilling (LWD) and mud logging are employed to determine accurate formation depths in preparation for coring. The coring program is critical for demonstrating the depth, thickness and integrity of the geological seal, and to provide
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
185
Figure 3. Typical completion diagram for an AGI well.
samples to evaluate the physical and chemical properties of the reservoir. After coring and drilling to the selected total depth is accomplished, the well is circulated and conditioned to prepare the borehole for geophysical logging. Following geophysical logging and initial interpretation, depths are selected for the installation of the production casing, preliminary perforation zones, and the location of a Corrosion Resistant Alloy (CRA) joint(s) of the casing. The CRA joint(s), typically 20 to 30 feet in length, is casing segment where the packer is seated. Ideally, the CRA is installed in the production casing section within the seal formation, above the uppermost
186
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
injection zone, and is designed to prevent acid gas from migrating u p the borehole from the injection zone. An additional safety factor is provided by the use of special acid and corrosion-resistant cement in the annulus through the seal formation.
12.2.4
Well Completion and Construction
After coring and logging are completed as described below, the production casing is installed, cemented and tested and temporarily capped, the drilling rig is released and a smaller workover rig is mobilized for the final completion tasks. Completion includes logging of the cement-bond (and remediation cementation if necessary), perforation of the injection zone(s), collection of formation water samples, injection testing (using fresh water or brine compatible with the target formation), installation of the production packer and injection tubing, emplacement of the subsurface safety valve (SSV) (to isolate any TAG in the tubing in the event of a failure upstream), and the installation of the production "Christmas tree" that was designed and constructed using appropriate corrosion-resistant alloys. The annulus of the well is now filled with an inert fluid (dry gas only injection-diesel; combined gas/wastewater injectionbrine) to prevent corrosion and assist in detecting potential tubing leaks. Once all submissions are made to the appropriate regulatory agencies and final approval is obtained for the completed well, it is now ready to be connected to the aboveground facilities and put on line for injection.
12.2.5
Reservoir and Seal Evaluation
Throughout the initial geological feasibility study, detailed prospect evaluation, permitting and well design phases, great care is taken to accurately model the depths, thicknesses and properties of the targeted seal and reservoir system. Although geophysical, driller or mud logs are usually available from nearby reference wells, representative core samples are commonly unavailable, and publically-available geophysical logs may not provide the necessary parameters for reservoir and seal determinations. To provide convincing and assuring data on the properties of the seal and reservoir, it is critical to collect cores that span the interval between these units. Because coring is expensive and timeconsuming, and since a core cannot be collected once a zone has
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
187
been drilled, it is imperative that the core point be selected at the correct stratigraphie interval. Before drilling begins, careful examination of existing reference well logs is used to estimate the anticipated depths and thicknesses of distinct formations a n d / o r marker beds to identify the core point in the location of the proposed AGI well. As drilling progresses, an experienced mud logger and the project geologist continuously monitor the chip returns, drilling rates, and Logging While Drilling (LWD) gamma ray readings to provide an ongoing orientation in the stratigraphie section. For projects where there are no wells within a mile which penetrate the injection zone (such as during drilling of the Pathfinder AGI #1, where the nearest wells penetrating the targeted system were from 3.5 to 5 miles away), LWD tools provide significant advantages in picking core points as observed depths and thicknesses of the relevant formations may depart significantly from the anticipated levels. In the case of the Pathfinder AGI #1 referenced above, correlation of the LWD gamma ray readings with the closest reference wells permitted selection of a core point which resulted in the successful collection of a representative sample of the entire upper seal rock (Wanakah Fm.) and the majority of the reservoir (Entrada Fm.). Immediately after the cores are recovered from the well, the project geologist supervises the visual logging and sampling of the cores. The cores are typically removed from the coring assembly in the aluminum tube lining in the core barrel. This method protects the core from damage and contamination, and preserves the formation fluids in the material. The original cores, thirty feet in length, are cut into three to 6 feet sections while in the tubes, and capped and stored for shipping to the core laboratory. During the cutting process, short (1 to 4 inches) samples are collected for field logging. Selected core segments used for detailed formation fluid studies are refrigerated for preservation. At the laboratory, core analyses include the determination of mineralogy, porosity, permeability, formation fluid versus injection fluid interactions, and general petrology. These tests are selected to determine the initial and long-term performance of the seal and the reservoir during the anticipated life of the injection project. After the total depth is reached, the borehole is cleaned and conditioned for geophysical logging and the drilling string is tripped out of the hole. Then a suite of open hole geophysical logs is run to provide additional information on the in situ properties of the
188
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
seal and the reservoir. The initial logging includes caliper, natural gamma, resistivity, litho density porosity and neutron porosity. The data is provided in both analog and digital electronic formats. Gamma and resistivity logs are used for lithology determination and correlation. The formation density and neutron porosity logs are critical in modeling the porosity volume and reservoir capacity and the litho density photo electric absorption index (Pe) is useful in identifying mineralogy differences. Caliper data is used to generate a borehole volume calculation that is very useful in refining the cement program planning. The zone spanning the interval from below the reservoir to the top of the seal is then logged using a Formation Micro Imaging (FMI) logging tool. This tool uses an array of six circumferential micro-resistance sensing pads to detect very fine-scaled features related to bedding planes and fractures. These logs are used to evaluate the presence and extent of any fractures that might either enhance the permeability of the reservoir, or, conversely, damage the effectiveness of the seal. After the final evaluation and interpretation of the core analyses and the geophysical logging, a comprehensive model of the reservoir - seal system is developed. This model is used to select the final injection zones to be perforated, determine if any well treatment is needed prior to injection and the optimum pressures and flow rates for injection.
12.2.6
Documentation, System Start-up and Reporting
Although documentation and reporting differ in various jurisdictions, the general process is described below giving examples specific to the NM case studies. The permitting process typically begins with the submittal of an application for injection (Form C-108 in NM) to the applicable regulatory agency. Following agency review, a public hearing is usually required and scheduled at which any of the potentially impacted parties identified and notified may attend and raise any relevant questions a n d / o r objections to the proposed project. Following the hearing, the appropriate state or provincial agency can accept, modify or deny the application. The final decision will usually be published as a formal order.
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
189
Following the approved order, the applicants are usually required to follow well drilling permit requirements similar to those for other types of oil and gas or injection wells (Forms C-101 and C-102 in NM)then provide a Form C-101 (Application for Permit to Drill, Re-Enter, Deepen, Plug back or Add a Zone). This application usually summarizes the operators, location, depths, and casing programs for the well. An additional certification usually addresses the operation and removal of drilling fluids. This is the Form C-144/C-144 CLEZ (Closed-Loop System Permit or Closure Plan Application) in NM. This submission usually describes the methods that will be used for either closed-loop drilling (where no subgrade excavated pits will be used), or excavated pits, where a detailed plan for closure, verification and restoration. Following the submittal and approval of the above forms, state or provincial agencies usually require notice prior to "spudding", or initiating drilling operations. Once drilling begins, additional regulatory submissions are usually required following drilling milestones such as casing and cementing, testing, completions, etc. These forms include copies or original documentation on work performed. Internal documents include daily summary reports from the drilling supervisor ("Company Man"), project geologist, mud logger, contractors, and coring and logging specialists. These reports are copied to owner, operator and consultants. Following the completion of the well, the operator is usually required to submit a well completion report (Form C-105 in NM) detailing the final configurations of the well, and identifying the depths of formations encountered during drilling. Copied of geophysical logs and mud logs may also be required in this submittal. Once injection commences, operators are usually required to submit periodic reports (C-115, monthly in NM) to provide the details on the volumes and pressures observed during injection. Regulators also typically require continuous monitoring or periodic pressure testing of the casing between the surface and the packer to assure the integrity of the injection system. Following the final completion, approval and operation of the well, a detailed End of Well Report is prepared. This report documents the well's history from initial design to completion, including
190
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
copies of all regulatory correspondence and submittals, internal reports, logs, contractor reports, and budget documents.
12.3
AGI Projects in New Mexico
There are currently four AGI wells in operation in New Mexico, the oldest of which was installed in 2002. All of these AGI wells are located in the Permian Basin (Figure 1; Table 1). A fifth AGI well is scheduled to begin operations in December 2010, and it will be the first in the San Juan Basin. The AGI wells and associated compression facilities represent a range of designs and injection conditions reflecting: both dry injection (TAG only) and combined TAG/wastewater injection. Injection depths range from approximately 4400 feet to over 11,000 feet. This section provides details from three recent Geolex AGI projects (Linam AGI #1, Jal 3 AGI #1, and Pathfinder AGI#1) in order to illustrate the range of AGI projects and the characteristics of successful projects with varying constraints and in different geologic environments.
12.3.1 Permian Basin The Permian Basin of west Texas and southeastern New Mexico is one of the major oil producing areas of the US; it also contains significant accumulations of natural gas. In 2009, the New Mexico portion of the Permian Basin produced close to 500 MMCF of natural gas (4). This portion of the basin consists primarily of Paleozoic carbonates that were deposited on the basin shelf. The climate at the time of deposition was arid and resulted in limited fluvial runoff. This contributed to the growth of carbonate banks and reefs and the coincident development of dune fields during episodic subaerial exposure of the shallow marine carbonates. The carbonates are capped by a regional evaporite and thick red beds. Oil and gas pools are found throughout almost the entire stratigraphy of the Permian Basin, including: the Tansill; Yates; Seven Rivers; Queen; Grayburg; San Andres; Yeso; Bone Springs; Abo; Lower Bone Springs; Atoka; Devonian; Ellenburger; and others (Figure 4; 5). Production from these formations is localized and depends on the proximity to source rocks, local structural geology, and variations in permeability and porosity. Where hydrocarbons are absent, these zones form extensive reservoirs for saline
TD (feet)
Jal 3 AGI #1
Pathfinder AGI#1
SUGS
WGR
87009100
Linam DCP Midstream AGI#1
6610 Entrada
63556550
5144 San Andres 43755000
9212 Lower Bone Springs
1120711412
24
47
73
14
n/a
1:2.5
Dry
78:20:02
90:10:00
Grayburg
Wanakah/ Todilto
Dry
75:25:00
Dry
75:25:00
Woodford
Abo
1:2500
75:25:00
n/a
2.5
2.5-3.5
0.8
0.3
n/a
1300
1200
1850
1800
3.8
5
5
1
0.5
1985
1600
2644
3240
1980
Injection Injection Reported Max. Rate Permited Injection MMSCFD Max. Rate type Press. (TAG) TAG: MMSCFD Press. (PSI) (TAG) PSI ww
Percentage C0 2 : H2S:
c,-c7
Injection Characteristics
TAG Composition
Woodford
Cap Rock Reservoir Injection Net Depth Porosity Name Feet Feet
Agave 990011600 Devonian Ellenburger 11400 Metropolis AZLSt#l
Well Name
Reservoir Characterization
DCP Duke AGI 11472 Devonian Midstream #1
Marathon
Operator
Well Information
Table 1. Summary of AGI wells in New Mexico.
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
192
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES Age
Permian basin
San Juan basin
Alluvium
Alluvium
I I !
i i í
Ogallala
Fredericksburg trinity Ss.
Morrison Fm. Wanakah Fm. Todilto Ls. Entrada Ss. Santa rosa dewey lake
Chinle Fm.
Dewey lake Ruslter Salado Castile Bel! Canyon Greyburg San Andres Cherry Canyon Bone Springs Wolfcamp/Abo
Cutler Fm.
1 I 1 i !
Kirtiand shale fruitland Fm. pictured cliffs Ss. Lewis shale upper mancos shale Gallup Ss. Greenhorn Ls. Graneros Sh. Dakota Fm.
Cisco Canyon Strawn Atoka
Honaker trail Fm. Paradox Fm. Pinkerton Trail Fm. Mofas Fm.
Woodford
Leadville Ls.
Woodford
Elbert Fm.
Fussleman Montoya Simpson group Ellenburger Ellenburger
Ignacio Qtz.
Undifferentiated Basement rocks
Undifferentiated Basement Rocks
!
1
Figure 4. Stratigraphie columns for the Permian and San Juan Basins of New Mexico, USA.
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
193
Figure 5. Diagrams showing the injection reservoir and caprock geology at Linam AGI#1, Jal 3 AGI#1 and Pathfinder AGI #1.
brines. Tighter, less permeable and less porous units, bound these reservoirs and inhibit vertical migration. 22.3.1.2
Linam AGI #1
During the early through mid 2000s, the Linam Processing Plant in Hobbs (Figure 1), operated by DCP Midstream LP (DCP), experienced numerous problems with its SRU. In order to help reduce the pollution levels and eliminate the need for the SRU, Geolex and
194
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
DCP designed and installed an AGI well during October-December 2007. A two-mile long LP pipeline was constructed to transport TAG from the plant to the AGI well. Following completion of the pipeline in early 2009, the well was reentered and completed. The AGI system has been in operation since September 2009. The feasibility study for the Linam plant found that due to local faulting, there was no reservoir beneath the plant capable of accepting the target injection of 5 MMCFD of TAG. The geology of the surrounding area was examined and a suitable site was found about 1.5 miles away. Design TAG (20% H2S and 80% C02) compressed at the plant to 15-20 PSI, piped to well where the TAG was compressed at the wellhead 1400 PSI prior to injection. Injected through 3 1/2" carbonsteel tubing set in an Inconel packer with an Inconel-clad Christmas tree. The TAG is then injected into the Lower Bone Springs formation. A SSV located at about 260 feet depth prevents the upwards migration of TAG in the case of an emergency. The facility has been designed to inject up to 5 MMCFD over a 30 year lifespan. Two target reservoirs were originally selected for the Linam project, the Brushy Canyon (top found to be at 5023 ft) and the Lower Bone Springs (top found to be at 8696 ft). The Brushy Canyon formation (sandstone) was eliminated as a choice, in part due to its low permeability found by an open hole Drill Stem Test (DST). An analysis of sidewall core samples and open hole logs revealed two promising zones in the Lower Bone Springs, one at 8710-9085 feet and the other at 8445-8538 feet zones (Figure 5). The Lower Bone Springs is composed of a mixed calcite and dolomite packstone. The secondary calcite cement is vuggy, fractured and has dissolution porosities measured up to 15%. Permeabilities of around 100 mD and greater were measured. It was decided to perforate the deeper of the two Lower Bone Springs zones and to save the shallower zone for future injection potential. Based on the porosity logs, the deeper zone has a net porosity of about 73 feet. The Linam AGI #1 was spudded on October 21,2007 and reached a TD of 9212 ft on November 16,2007. Three casing strings (13 3 / 8 " casing to 580 feet, 9 5 / 8 " casing to 4217, and 7" casing to TD) were installed and cemented to the surface (Figure 6). The cement was drilled out to 9137 feet depth. It was decided to perforate and test the deeper portion of the Lower Bone Springs first to ascertain whether it would accept the required 5.0 MMSCFD of TAG at less than the maximum expected surface pressure of 2800 psi. If the lower zone performed adequately, the upper zone would be reserved for
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
195
Figure 6. Diagrams showing AGI well designs for Linam AGI#1, Jal 3 AGI#1 and Pathfinder AGI #1.
future injection potential because perforating the casing to test (and then abandoning it later) would cause unnecessary casing integrity issues. This process was accomplished by installing temporary tubing and a retrievable packer and then using Tubing Conveyed Perforating (TCP) guns. The objective in using the TCP guns was to perforate with a slightly underbalanced hydrostatic pressure and then immediately flow the well through the tubing string into a measuring tank at the surface. Upon firing of the guns, no significant blow or flow at the surface was noticed. It was discovered that only the top 85 ft of guns had shot and that the top 45 (net) ft of perforations had opened up a zone that took fluids on a vacuum; an encouraging sign for injection potential. New guns were run in the
196
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
hole and the missed shots were run. Following swabbing and an extensive acid job on the perforations, several injection tests were performed. An injection test performed at an injection rate of 7.25 MMSCFD (above the maximum anticipated 5.0 MMSCFD injection rate) for 6 hours exhibited only a slight gain in surface pressure from 250 PSI to 300 PSI. As a result a flow choke was considered in order to hold enough back pressure on the flow stream to keep the surface injection pressure above critical pressure at approximately 1,000 PSI. Following completion of the low pressure pipeline and aboveground facilities (the AGI compression facility and pressure regulating devices) in July 2009, the hole was reopened and the permanent packer was installed at 8650 feet. Carbon-steel 3 1/2" production tubing with a subsurface safety valve was inserted downhole and the casing annulus was filled with diesel. The well was put into operation in September 2009 and is currently injecting 2.5-3.5 MMSCFD of dry gas at about 1100-1300 PSI, well below the maxim u m allowed pressure of 2644 PSI. 12.3.1.2
Jal 3 AGI #1
By mid-2007 the SRU at the Jal 3 gas processing plant (Figure 1) had reached its processing limit and the operator, Southern Union Gas Services (SUGS), had to potentially curtail gas production and processing at the plant. Following a favorable feasibility study, it was decided to pursue the installation of an AGI well at the plant to reduce the overall GHG emissions at the plant and with the eventual goal of replacing the SRU entirely by injecting the entire stream which was processed by the former SRU and eliminating associated combustion sources. The Jal 3 AGI #1 was drilled during June 25-July 14, 2008 and completed during December 1-10, 2008. It has been in operation since March 2009, following the completion of the aboveground facilities. The Jal 3 AGI #1 was designed to inject a mixed TAG (78%C0 2 , 20% H 2 S, and 2% Cj-Cp and wastewater stream. Field gas at the plant contains significant water that had previously been disposed of using an SWD located at the plant. At this installation TAG is compressed to 1,600 PSI and then mixed with the Class II plant wastewater. The mixed TAG and wastewater is then choked down from 1,600 PSI to 980 PSI and injected into the AGI well through 3 Vi inch fiberglass-lined tubing set in an Inconel® (a corrosion resistant
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
197
nickel alloy) clad packer, and then, through perforations into the San Andres formation. An automatic subsurface safety valve (SSV) placed on the injection tubing approximately 260 feet below the surface will prevent the injected acid gas from migrating upwards in case of an upset or emergency. The facility has been designed to inject 2,300 to 7,930 barrels per day of mixed wastewater and dissolved TAG (with an approximate ratio of 3:1 wastewater to TAG) over a lifespan of 30 years. The Permian San Andres formation was chosen as the injection reservoir, in part since the majority of oil and gas production within this area is restricted to the shallower Yates-Queen interval. The only other well that penetrated the San Andres formation was the plant SWD that was plugged and abandoned prior to the injection of TAG through the AGI well. The primary local fresh water Tertiary-Quaternary Ogallala formation aquifer (<200 feet depth) has been safeguarded by surface casing extending more than 1000 feet below this zone to include any possible freshwater in the red beds of the Triassic Dockum group. The suitability of the San Andres as the injection reservoir collected sidewall and traditional core samples and open hole well logs (Figure 7). The analysis identified an injection interval of 600 net feet with an approximate average porosity of 7.85%, resulting in a calculated 47 feet of effective porosity at this location. The interval is highly fractured and permeable, as indicated by an FMI analysis and demonstrated by twenty years of successful injection of wastewater in the adjacent salt water disposal well which was plugged and abandoned prior to putting the new combined TAG/wastewater well into service. The overlying Grayburg formation and the upper portion of the San Andres form an impermeable barrier (significant sections with a measured vertical permeability of <0.1 milliDarcy (mD) and no conductive fracturing) above the injection zone, while a salt-rich, low porosity, low permeability layer (1.8 % porosity and 0.01 mD permeability) near the base of the San Andres forms a barrier beneath the injection zone. No faults have been identified through the section indicating that the injection interval is well confined. In addition, the injection well is located within a structural trough that should constrain the injected fluid to an area of <240 acres, forming an ellipse extending <2800 ft from the well in a NE-SW direction. The entire sequence is composed of carbonates that will neutralize the acidity of the injected fluids and lead to long-term sequestration of the C 0 2 and sulfur species.
198
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 7. Diagrams showing the injection reservoir and caprock geology at Jal 3 AGI#1.
The Jal 3 AGI #1 was spudded on June 25, 2008 and TD was reached at 5245 feet on July 11 th . The well was constructed with three casing strings: 16" conductor casing to 51 feet; 9 5 / 8 " surface casing to 1247 feet; and 7" production casing to TD (Figure 6). All three casing strings were cemented to the surface, with the lower portion of the 7" casing cemented using Halliburton's special acid resistant ThermaLock™ cement. Completion of the well began on December 1,2008. Based on the evaluation of the injection reservoir,
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
199
the production casing was perforated from 4,430 feet to 4,970 feet bgs and the packer was set at 4,355 feet bgs. The injection tubing and SSV were inserted downhole and the casing annulus was filled with saltwater. Following connection of the Christmas and testing of the safety devices, a temporary flow line from the plant SWD was installed, connected to the completed well and a test injection performed. That night the new well pressured up and flow was redirected back to the SWD. The perforations were cleaned with an acid job the next day and the injection pressure dropped to 280 PSI at 5 bbl/min. Flow of plant waste water was then returned to the Jal 3 AGI #1 well. Injection of mixed TAG and wastewater into the AGI well commenced on March 26, 2009, shortly after completion of the aboveground facilities (the AGI compression facility, pressure regulating devices, and the new gas /wastewater mixing chamber). The plant is currently injecting an average of 2.5 MMSCFD of TAG mixed with 2500 bbls/day of wastewater with a resulting injection pressure of 1300 PSI (Table 1). 12.3.2
S a n Juan B a s i n
The San Juan Basin accounts for roughly two thirds of the New Mexico's natural gas production (937 MMCF of natural gas was produced in the San Juan Basin of New Mexico in 2009; 4). The Basin, located in northwestern New Mexico and southern Colorado, was filled by numerous cycles of is the result of marine and nonmarine deposition that began in the Pennsylvanian and continued through the Tertiary. The marine deposits are characterized by a range silicic sediments (shale to sandstone) and limestone. The non-marine rocks include sandstones originally deposited as aeolian dunes and or deposited by rivers and streams. During the Late Cretaceous, shoreline migration along the large inland seaway resulted in swamps and coastal plain conditions that produced significant coal deposits. The most notable of the coal deposits is the Fruitland Formation, which is the most prolific coal bed-methane field in the United States (6). Although the Cretaceous coal beds now account for the majority of natural gas produced in the San Juan Basin, numerous deeper oil and gas pools have been produced in the past. These deeper pools are found in marine through subaerial sandstones in the Mesa Verde, Gallup and the Dakota (Figure 4). The Entrada sandstone
200
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
has also seen limited production. These sandy units serve as reservoirs to saline brines and are separated by thick units of shale and mudstones in the Lewis, Mancos, Morrison, and Chinle. 12.3.2.1
Pathfinder AGI #1
Field gas processed at the San Juan River Gas Plant operated by WGR (Figure 1) is sourced primarily from the Barker Dome area and has a high C 0 2 content. The high C 0 2 content in the inlet gas results in a lower than optimal operating temperature and reduced effectiveness of the SRU. In order to eliminate all GHG emissions associated with the SRU and its combustion sources, the stream currently going to the SRU (90% C0 2 , 10% H 2 S, trace C r C 7 ) will be injected using the Pathfinder AGI #1 . This well was drilled in August 2010 to a total depth of 6610 feet. The proposed injection zone for the Pathfinder AGI #1 will be within the Jurassic Entrada Sandstone for all of its thickness of approximately 140 feet in this location (6352-6492 feet). The preliminary core analysis and geophysical logs show that the Entrada in this area consists of friable, well-sorted sandstone with porosities up to 25% (average about 17%) and a net porosity of 23.5 feet (Figure 8). Although reservoir tests have not yet been performed, the friable nature of the sandstone indicates that the Entrada should be very permeable. The Entrada is effectively sealed on top by the overlying Todilto Limestone and Beclabito siltstones of the Jurassic Wanakah Formation and below by the underlying shales and mudstones of the Triassic Chinle Formation. Based on the value of 23.5 feet of net porosity, a thirty-year period of injection at an average of about 2.5 MMSCF per day (1000 barrels of compressed TAG) would occupy an area of approximately 60 acres, covering a radius of approximately 910 feet around the AGI well. At a maximum rate of 5 MMSCF per day (2000 barrels of compressed TAG), the area would be approximately 120 acres, enclosed within a radius of about 1290 feet from the well. There are currently four permitted and operating salt water disposal (SWD) wells completed in the Entrada in the general area of the plant, but the closest well (Salty Dog #5) is approximately 3.7 miles southeast, well outside the one-mile radius of evaluation within the proposed injection zone and the area of review required for the MNOCD C-108 application. According to NMOCD files, these four SWD wells currently accept from 800 to over 2000 barrels
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
201
Figure 8. Diagrams showing the injection reservoir and caprock geology at Pathfinder AGI #1.
of fluids per day, at pressures below their permitted levels. Based on these data, we have concluded that the Entrada provides ample porosity, permeability and volume to serve Anadarko's injection needs.
202
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Since the Pathfinder well has not yet been completed, no samples are available for the injection reservoir. The most representative analysis of fluids from the Entrada was collected in December 2005 from the Salty Dog #5 SWD well, approximately 3.5 miles southeast from the proposed AGI well. These analyses showed that the formation water had a Total Dissolved Solids of 25,624 mg/L. The primary cation was sodium, and the principal anions were chlorides, sulfate, and bicarbonates (Table 2). Three casing strings were installed in the Pathfinder AGI #1: 13 3 / 8 " conductor casing to 134 feet; 8 5 / 8 " surface casing to 1108 feet; and 5" production casing to TD at 6610 feet (Figure 6). The length of the surface casing was chosen to ensure double casing through the Lewis Shale and the Pictured Cliffs, to protect the Fruitland Coal Formation. Both of these formations are considered fresh water aquifers and the Fruitland is an active coal mining and natural gas producing zone. TD reached nearly 150 feet into the Triassic Chinle Formation allowing characterization of the basal cap and ensuring access to full injection zone. Casing was cemented pursuant to applicable requirements. Completion of the well is scheduled for October 2010 and will begin with the cement bond log and the perforation of the production casing. The 2 7 / 8 " 6.5ppf L80 tubing string will be set into an Inconel packer and CRA joint situated in the Wanakah, just above the Entrada injection zone. An SSV (subsurface safety valve) also be constructed of Inconel, will be installed on the production tubing to assure that fluid cannot flow back out of the well in the event of a failure of the injection equipment. In addition, the annular space between the projection tubing and the well bore will be filled with an inert fluid such as diesel fuel. The gates, bonnets and valve stems within the Christmas tree will be nickel coated, while the remainder of the Christmas tree will be made of standard carbon steel components. The Christmas tree will be outfitted with annular pressure gauges that report operating pressure conditions in real time to a gas control center located remotely from the wellhead. In the case of abnormal pressures or any other situation requiring immediate action, the acid gas injection process can be stopped at the compressor and the wellhead shut in using a hydraulically operated wing valve on the Christmas tree. Due to the corrosive nature of the injected fluid, the line that will convey the TAG to the well from the compression facilities will be a 3" stainless steel line.
San Andres
Littmana
Entrada
Approx. 3.5 Miles E of Anadarko Pathfinder
8,809
160
146 N / A
30,900 5,240 2,527 N / A
Approx. 15 Miles NE of Jal 3 AGI #1
535
24,909 1,760
424
1
0
67
N/A
N/A
93
N/A
N/A
0
400
N/A
324
9,000
62,000
40,000
Mg K Fe Sr Zn T-Alk ClCa mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
so4
5,800
2,080
4,145
mg/L
1,708
N/A
395
HC0 3 mg/L
7
pH
25,624
7
93,400 N / A
73,412
TDS mg/L
N/A
N/A
94,800
Conduct uS/cm
Source: From A. Nicholson, Jr., and A. Clebsch, Jr., Geology and Ground-Water Conditions in Southern Lea County, New Mexico, U.S. Geological Survey Ground-Water Report 6, p. 123,1961.
Salty Dog #5
San Juan Basin
L. Bone Spr.
Linam AGI #1
Na mg/L
Linam AGI #1
Formation Location
Permian Basin
Well
Table 2. Formation water compositions for reservoirs of selected AGI wells in New Mexico.
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
204
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
12.4 AGI and the Potential for Carbon Credits Since 80 to 90 percent of the acid gas stream that comes out of amine units is C0 2 , significant amounts of C 0 2 production can be reduced or eliminated through geologic sequestration of GHG using AGI. For example, a 5 MMSCFD acid gas stream with 80% C 0 2 results in 77,000 metric tons of C 0 2 per year, as demonstrated in Figure 9. In sharp contrast to the carbon capture challenge presented by traditional coal-fired power plants, natural gas processing facilities afford a ready opportunity for geologic sequestration of GHG because the waste gas stream from amine units does not contain a significant fraction of non-GHG which would have to be separated or compressed prior to injection/sequestration. In addition, the significantly lower volumes of C 0 2 produced from natural gas processing plants relative to coal-fired power plants make these facilities ideal candidates for practical and economical sequestration projects. Unlike many industrialized nations which have signed on to the Kyoto Protocol, the US has not yet implemented cap and trade legislation or formal rules to regulate the release of GHG. However, reporting requirements for GHG emissions go into effect this year
Figure 9. Diagram showing the relationship between daily injection volume and annual C 0 2 mass injected.
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
205
in the US, and legislation to regulate C 0 2 and other GHG emissions remains a high priority of the Obama administration. In the absence of federal legislation to regulate GHG emissions, a number of states and regions have formed voluntary coalitions to reduce and regulate GHG emissions, including C0 2 . These include such organizations/coalitions as the Western Climate Initiative (WCI), the Regional Greenhouse Gas Initiative (RGGI) and others. The advent of the reporting requirements taking effect this year and the anticipation of potential legislative regulation of GHG emissions are bringing increased visibility to the whole area of GHG emissions from gas processing plants. As environmental legislation is implemented, geologic sequestration of C 0 2 from these facilities may prove to be "low hanging fruit" for gas producers because gas processing plants have a distinct advantage over other GHG emitters such as coal-fired power plants in terms of C 0 2 capture and sequestration as described above because gas plants already separate and capture C 0 2 as a part of their amine process. Therefore, geologic sequestration of C0 2 , from either sour or sweet gas processing is a viable and economical alternative for the gas processing industry now, unlike the coal-fired utility industry. Obtaining carbon credits for C 0 2 reductions is a purely voluntary process in the United States at the present time. One credit can be obtained for each metric ton of C 0 2 reduction. Although there is some market for these credits in the US (approximately $71 million of these credits were traded in US markets in 2006), the real value of these credits in US markets remains to be established when binding GHG legislation comes into effect in this country. In European markets, where that regulation already exists, billions of dollars a year of carbon trading is taking place. Carbon credits can be obtained in a variety of ways, including the purchase of reforestation and planting projects. Clearly, verifying and quantifying the C 0 2 offsets from these sorts of projects is a much less precise process than in the gas processing business where measurement and verification can be done directly at the well head. AGI provides a GHG reduction method for obtaining carbon credits that is both directly measurable and easily verifiable. The actual registration of voluntary carbon credits (with the issuance of Renewable Energy Certificates) is presently fairly well defined by Federal regulation, and there are several groups in the United States that provide certification and verification of these credits—these include the Chicago Climate Exchange (CCX),
206
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
American Carbon Registry (ACR) and others. The process for registration of carbon offsets or credits in these voluntary markets consists of three basic steps. Initially, the calculation of tons per year of CO 2 sequestered must be made; then a formal application must be submitted (project protocols); and subsequently verification of the carbon reduction must be obtained from an independent entity This whole process can cost between US$50 to $100 thousand per project, depending on the size and complexity of the project. The direct economic benefit of registration of offset credits into voluntary programs is not clear at the present time. The motivation for obtaining formal offset reduction credits at the present time lies more in the potential public relations benefit of such actions in terms of enhancing a "green" corporate image and in terms of positioning the companies for a favorable regulatory review. There may also be some potential direct monetary advantage in the early registration of credits if early offset provisions are implemented in forthcoming legislation. However, the definition of the monetary value of voluntary early offset credits in pending federal legislation is very much a moving target at present. In spite of the lack of quantifiable advantages for early offset credits, and the absence of current definitive regulatory legislation, the fact remains that the regulation of C 0 2 and other GHG in the US is inevitable, and natural gas producers should begin thinking about how compliance with these regulations will affect their operations. Geologic sequestration of GHG and H2S through AGI is the best currentlyavailable technology and it has been demonstrated to be a viable and cost-effective methodology for achieving compliance with future mandatory GHG reduction requirements. As geologic sequestration of GHG and related natural gas processing wastes expands in the US, there are numerous legal and regulatory issues which will need to be addressed. One of the most significant of these issues is ownership of pore space in potential reservoirs targeted for geologic sequestration projects. Most states in the US do not define who owns the pore space into which the acid gas will be injected. In contrast, as an example, Wyoming became the only state in the US to legislatively define ownership of pore space. Under Wyoming law, the surface owner is also the owner of underlying pore space, and leases for the anticipated use of pore space must be obtained from those owners in a similar manner that oil and gas leases must now be obtained from mineral owners (7). Similar legislation is anticipated in other oil and gas
ACID GAS INJECTION IN THE PERMIAN AND SAN JUAN BASINS
207
producing states. Another regulatory issue being examined relative to sequestration projects is whether the unitization process currently implemented in many oil and gas producing regions for secondary and tertiary recovery operations might be adapted to AGI/ C 0 2 reservoirs. Also under consideration is whether the federal, provincial or state governments might assume liability associated with these A G I / C 0 2 injection projects. An evolving technical and regulatory issue has to do with the actual monitoring of the injected gas plume and verification that the plume is being contained by the overlying and underlying caprocks and within the boundaries of the leased reservoir. No specific regulatory requirements have yet been developed at the state or national level to standardize methods of acceptable monitoring of these projects and verification of the longevity of carbon credits arising from these projects. The EPA, however, has developed some proposed regulations for Class VI injection wells under its underground injection control program (UIC) program. These regulations seek to standardize the construction and installation of injection wells. Although there is no specific time table for implementation of these regulations, they will be an important factor governing the technical and economic constraints on future AGI projects and should be considered in constructing projects which seek to generate future carbon credits from GHG sequestration as an added benefit to the use of this technology for disposal of acid gas.
12.5
Conclusions
AGI has been demonstrated to be an effective means for disposing of TAG from natural gas processing plants and is well suited to New Mexico and other areas with large reservoirs of saline brine. Building on its experience in AGI, Geolex has developed a structured process for AGI development and identified the following key points in the successful development and implementation of these systems: 1. Identification of a suitable injection reservoir and cap rock is critical. 2. Well design is largely dependent on choice of wet vs. dry injection and characteristics of injection reservoir and cap rocks.
208
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
3. Close communication with regulatory agencies at all stages of the process is important. 4. Proper monitoring of drilling and completion is necessary to ensure that completion of the project is successful, timely and within budget. 5. The accurate characterization of reservoir and caprock (determining capacity and demonstrating isolation) is essential for obtaining regulatory agency approval and for future registration of carbon credits. Since TAG predominantly consists of C0 2 , AGI represents an increasingly popular method of carbon sequestration that can be readily monitored. The combination of AGI and carbon sequestration produce environmental benefits and cost savings for operating companies now and will most certainly produce additional economic benefits as the regulation of GHG emissions becomes more stringent and companies seek to enhance their image as green energy producers.
References 1. J.H. Gary, and G.E. Handwerk, Petroleum Refining Technology and Economics (2nd Edition Ed.). Marcel Dekker, Inc., 1984. 2. U.S. Federal Government, Federal Register, Vol. 74, No. 209, pp. 56250-56519, Friday, October 30, 2009. 3. U.S. Federal Government, 40 CFR Parts 144 and 146, Underground Injection Control Program. 4. EMNRD, New Mexico Energy, Minerals and Natural Resources Department 2009 Annual Report, 2009. 5. E.E. Kinney and F.L. Schatz, The Oil and Gas Fields of Southeastern New Mexico A Symposium, Roswell Geological Society, p. 220,1967. 6. US Energy Information Administration, New Mexico State Energy Profile,
, last updated August 26, 2010. 7. Wyoming State Legislature, Wyoming Statutes WS 34-1-152, Ownership of Pore Space Underlying Surfaces, July 1,2008.
13 C0 2 and Acid Gas Storage in Geological Formations as Gas Hydrate Farhad Qanbari1, Olga Ye Zatsepina1, S. Hamed Tabatabaie1, Mehran Pooladi-Darvish1-2 1
Department of Chemical and Petroleum Engineering, Schulich School of Engineering, University of Calgary, Calgary, AB, Canada 2 Fekete Associates Inc., Calgary, AB, Canada
Abstract With the increasing concern about climate change, the public, industry and government are showing increased interest towards reduction of C 0 2 emissions. Geological storage of C 0 2 is perceived to be one of the most promising methods that could allow significant reduction in C 0 2 emissions over the short and medium term. One major concern against geological storage of C 0 2 is the possibility of its leakage. Carbon dioxide under the pressure and temperature conditions encountered in most geological settings remains more buoyant than water. However, C0 2 -hydrate formation - that leads to trapping of C 0 2 in the solid form - provides the opportunity for secure storage of C0 2 . In this paper, we investigate two different geological settings that are suitable for formation of C 0 2 hydrates. These include storage of C 0 2 in depleted gas pools in Northern Alberta and beneath the ocean floor. Thermodynamic calculations suggest that C 0 2 hydrate is stable at temperatures that occur in a number of formations in northern Alberta, in an area where significant C 0 2 emissions are associated with production of oil sands and bitumen. Simulation results are presented demonstrating that upon C 0 2 injection into such depleted gas reservoirs, pressure would initially rise until conditions are appropriate for formation of hydrates, enabling storage of large volumes of C 0 2 in solid form. Numerical results suggest that because of tight packing of C 0 2 molecules in the solid (hydrate), the C 0 2 storage capacity of these pools, is many times greater
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (209-226) © Scrivener Publishing LLC
209
210
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
than their original gas in situ. This provides a local option for storage of a portion of the C0 2 emissions there. It is shown that in presence of impurities in the injected gas such as H2S, the stability zone of hydrates is significantly expanded allowing an expansion of the candidate reservoirs, by including deeper and warmer reservoirs. In another part of the paper, permanent trapping of C0 2 at a depth of a few hundred meters beneath the ocean floor, where it is of little or no harm to the ocean ecosystem, is studied. Simulation results indicate that injection of C0 2 at a few hundred meters below the ocean floor, would lead to rise of the C0 2 until it arrives at a depth where its density becomes heavier than water. The zone above this depth, where C0 2 becomes heavier than water is called the negative buoyancy zone (NBZ). Beneath the negative buoyancy zone, the C0 2 may become naturally trapped by a gravity barrier. Furthermore, formation of C0 2 hydrate will further reduce formation permeability and introduce a second barrier against C0 2 rise. Sensitivity studies are conducted to determine appropriate ocean depths for large scale C0 2 storage.
13.1
Introduction
Carbon dioxide is a greenhouse gas, so its capture and storage to avoid accumulation in the atmosphere are important components of climate change mitigation. Options for its storage may involve geological settings including sedimentary basins or saline aquifers. However, safety issues are vital in choosing a geological formation to store C0 2 . Sequestration of C 0 2 in the form of hydrate is quite attractive since it provides a low probability of leakage and an effective trapping mechanism as long as it is not heated. Hydrates are solids composed of the framework of water molecules, with gas molecules captured in the lattice vacancies. A solid structure provides a high density for the C 0 2 storage, while kinetics of the hydrate formation is relatively fast [1]. This paper focuses on storing C 0 2 through hydrate formation in depleted reservoirs of natural gas and beneath the ocean floor. The STARS simulator of CMG™ is employed to study the C 0 2 injection in these two geological settings. This simulator has previously been used to predict dissociation/formation of CH 4 -C0 2 hydrate in geological reservoirs [2].
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
211
In the following, first the basics and assumptions used during the course of this study are explained. Since in storing the C 0 2 in the depleted gas reservoir we account for the formation of CH4-CG*2 hydrate, the paper starts with a discussion of the phase diagrams for the pure and mixed gas hydrate. This section is followed by explaining the concepts and principles involved in sub-seabed disposal of C0 2 . Next, the potential of these two geological settings (i.e. depleted gas reservoir and ocean sediment) in storing the C 0 2 is studied through the use of STARS simulator of CMG™. Finally, the results of C 0 2 injection into the hypothetical models representing depleted gas reservoir and ocean sediment are presented.
13.2
Geological Settings
13.2.1 Depleted Gas Reservoirs Natural gas reservoirs represent safe C0 2 -storage facilities since the impermeable rocks have stored the gas for tens or hundreds of millions of years. Unlike saline aquifers, these gas pools are relatively small and well characterized. Alberta is known for its substantial oil and gas production. The northeastern part of the province could be used for sequestration because of the proximity to the source of emission and availability of low temperature depleted gas pools [3]. In this section, we study the storage capacity of a depleted gas reservoir. In particular, we investigate the effect of the in-situ gas in formation of mixed gas hydrate, the effect of rise in temperature as a result of the exothermic reaction of hydrate formation, and suggest a way for avoiding the deleterious formation of hydrates around the wellbore. 13.2.1.1
Mixed Hydrate Phase Equilibrium
When C 0 2 is injected in a reservoir where methane is present, mixed gas hydrates form. The stability conditions of CH 4 -C0 2 hydrates are different from that of either CH 4 or C 0 2 hydrate. Phase diagram for CH 4 -C0 2 hydrate is shown in Figure la [4]. It shows that for the two gases, the hydrate three-phase equilibrium is expanded into a zone between C0 2 - and CH 4 -hydrate boundaries. In this zone, P-T
212
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
conditions of the three-phase equilibrium depend on the composition of the vapor phase (the phase boundaries are shown at 0,20,40, 60, 80, and 100% C 0 2 in the vapor phase). At a fixed temperature, the phase boundary of the mixed hydrate spreads to lower pressures when the fraction of C 0 2 in the vapor phase increases. Consider three-phase equilibrium at a temperature of 6°C. For CH 4 hydrate, the equilibrium pressure at 6°C is 4.61MPa, while for C 0 2 hydrate it is 2.54MPa. When hydrate forms from a gas mixture of, for example, 70% C 0 2 and 30% CH 4 , the equilibrium pressure is 2.80MPa. At a P-T condition denoted by a blue circle (i.e. a shallow gas pool in Northern Alberta at 6°C that may be pressurized to 3MPa), only pure C 0 2 and mixtures of up to 30% methane can form hydrate. Another example of the influence of the composition of the vapor phase on stability of hydrate includes H2S (hydrate of this former is known to exist at 30°C). Similar to CH 4 -C0 2 mixed gas hydrates, H 2 S-C0 2 hydrates form at P-T conditions different from that of pure CÖ 2 or H2S hydrate. Figure l b shows calculated phase diagram of H 2 S-C0 2 hydrate; CSMGeml.10 software [5] was used for predictions. Similar to Figure la, the hydrate three-phase equilibrium is expanded into a zone between C0 2 - and H 2 S-hydrate boundaries. In the zone, P-T conditions of the three-phase equilibrium depend on the composition of the vapor phase (the phase boundaries are shown at 100, 95, 80, and 0% CÖ2 in the vapor phase). At a fixed pressure, the phase boundary of H 2 S-C0 2 mixed hydrate spreads to higher
Figure 1. (a) CH 4 -C0 2 mixed hydrate phase diagram. The blue circle represents a shallow gas pool in Northern Alberta at 6°C that may be pressurized to 3MPa., (b) H2S -C0 2 mixed hydrate phase diagram. Thick lines represent phase boundaries for single component hydrate; while thin lines represent phase boundaries for mixed hydrate (mole fraction of C 0 2 in the vapour phase is written above the lines).
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
213
temperatures when the fraction of H2S in the vapor phase increases. Consider three-phase equilibrium at a pressure of 3MPa. For C 0 2 hydrate, the predicted equilibrium temperature is 7.2°C, while for H2S hydrate it is 29.5°C. When hydrate forms from a gas mixture of, for example, 95% C 0 2 and 5% H2S, the temperature is 13°C. Experimental results confirm an increase in the hydrate stability field even when mole fraction of H2S in the vapor phase is as small as 1% [6]. This allows an expansion of the candidate reservoirs, by including deeper and warmer depleted gas pools. 13.2.1.2
Assumptions
The vapor phase composition affects not only P-T conditions of hydrate stability, but also fractionation of gases in hydrate [7]. We assume that the coefficient of proportionality between compositions of the vapor and hydrate phases is close to 1 [8]. As a result, the vapor and hydrate phases in the modeling have the same compositions. Consequently, the C 0 2 fraction in the hydrate phase would increase with injection time as C 0 2 molecules substitute methane in the structure. In this study, we assume fast intrinsic kinetics, so that the pressure and temperature of hydrate formation follow equilibrium conditions. Heat transfer controls the rate of hydrate formation, and latent heat of the reaction plays an important role. Enthalpy of mixed hydrates is calculated using values for the pure components in the proportion they have in the hydrates
13.2.2 Ocean Sediments Ocean sediments can provide high capacity natural traps for storage of C0 2 . The main trapping mechanisms (hydrate formation a n d / o r gravitational stability of C0 2 ) depend on the thermodynamic conditions in these sediments. 13.2.2.1
Negative Buoyancy Zone
(NBZ)
Figure 2a illustrates the phase behavior of C 0 2 (Red line) [9], the isochors for its pure liquid phase (Dotted lines) [10], and a sample temperature distributions in the ocean in low to middle latitudes [11 ] and its sediments (geothermal gradient of 0.03K/m). This figure shows that at a certain depth (of approximately 2600m in Figure 2a), the density of C 0 2 exceeds the density of
214
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
water. Below the seabed, the density of C 0 2 decreases as temperature increases, such that at a particular depth (of approximately 700m below the seabed in Figure 2a), the density of C 0 2 becomes less than that of ocean water, once again. Therefore, in this situation the density of C 0 2 remains more than that of ocean water between the seabed and this second depth (of 700m below the seabed in Figure 2a). The region of the sediment from the seabed to the neutral buoyancy level in the sediment is called negative buoyancy zone (NBZ) [12]. This zone in the sediment could provide a natural trap for any C 0 2 that may be injected further below. 13.2.2.2
Hydrate Formation Zone (HFZ)
Figure 2b shows the phase behavior of C 0 2 hydrate (using CSMGeml.10 [5]) (the Red line). C 0 2 hydrate can be formed at the conditions on the left side of the phase behavior of the C 0 2 hydrate. In the sediment, temperature increases based on geothermal gradient and increases the pressure requirement for stability of hydrate. At a special point in the sediment the equilibrium pressure exceeds the pressure provided by the water head. Hydrate formation zone (HFZ) is defined as the zone of stability of hydrate in the ocean sediment [12]. Because of the plugging of the pore space with solid hydrates, this zone acts as a cap for the C 0 2 stored underneath. The existence and the thicknesses of these zones (i.e. NBZ and HFZ) depend on the ocean depth and the sediment temperature profile. Four categories of depths are defined with varying degree of suitability to store C 0 2 in the ocean sediments as: (i) shallow, (ii) intermediate, (iii) deep, and (iv) super deep. The names for the categories are adopted from Koide et al. [13], adding the second category. The approach to the classification and consequently, the depth ranges are new. Figure 3 illustrates the thicknesses of these zones as a function of ocean depth. The temperature profile in the ocean is considered to correspond to low to middle latitudes [11] and the geothermal gradient in the ocean sediment is assumed to be 0.03K/m. In shallow sub-sea disposal, HFZ and NBZ do not exist. The parts of the ocean with intermediate depth have merely hydrate formation zone. In contrary to the super deep ocean, in deep parts of the ocean the thickness of HFZ is more than the thickness of NBZ.
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
Figure 2. (a) Schematic illustration of: a. Negative buoyancy zone (NBZ), (b) Hydrate formation zone (HFZ).
215
216
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 3. Schematic illustration of the classification of sub-seabed disposal of C0 2 , HFZ and NBZ.
13.3
Model Parameters
In this section, the properties of the two models developed by the STARS simulator of CMG™ for simulation of the C 0 2 storage as hydrate in depleted gas reservoir and ocean sediments are described.
13.3.1 Depleted Gas Reservoir We consider a depleted gas reservoir with a radius of 300m. It includes a 5-meter thick porous medium with porosity of 30% and permeability of 500mD. The reservoir initial pressure is 500kPa, initial temperature is 6°C, and saturations of gas and water are 75% and 25%, respectively. The initial volume of gas-in-place (methane) is 1.66xl06sm3. There is a layer of shale on the top and bottom of the reservoir (the shale thickness is chosen so that the effect of temperature changes in the reservoir is not felt at the top or the bottom boundaries). A vertical well is fully open, with C 0 2 gas injected at a constant rate of 0.1xl0 6 sm 3 /day at a temperature of 10°C. The injection temperature is chosen so that no hydrate forms in the vicinity of the wellbore: the three-phase equilibrium pressure at 10°C is 4.4MPa and hydrate does not form within a minimum of 10m from the wellbore in the simulations. The reservoir pressure is restricted to no greater than 4MPa in order to avoid appearance of liquid C0 2 . Numerical domain is discretized in radial direction: 300 cells x 1 meter; in vertical direction: 10 cells x 0.5 meter in hydrate reservoir as well as 10 cells x 2.5 meters in base- or cap-rocks. Thermal properties of the reservoir and shale are given in Table 1.
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
217
Table 1. Reservoir parameters for C 0 2 injection into the depleted gas reservoir.
Parameter
Value
Reservoir volumetric heat capacity \y Hydrate heat capacity \( y1/
. „ _1
Shale volumetric heat capacity ( /
Rock thermal conductivity
/
,
203.7
3 °Q\
3.76 x 106
<>,-,
2.468 x 105
Hydrate thermal conductivity
/
Shale total thermal conductivity
Í J/ /
13.3.2
2.6 x 106
3„
,
»p
,
=01
3.396 x 104
1.5 xlO 5
Ocean Sediment
A three-phase (aqueous, liquid C0 2 , and hydrate) three-component (water, C0 2 , and hydrate) model is developed using STARS simulator of CMG™ to simulate the fate of injected C 0 2 into the ocean sediment. Simplified models were built to validate the model for correct representation of the two key mechanisms of C 0 2 hydrate formation (in HFZ) and gravitational stability of liquid C 0 2 (in NBZ). A cylindrical geometry with thickness of 1300 m and radius of 80 km (to represent an extended ocean floor) is considered to represent the storage reservoir. The porosity, horizontal and vertical permeability values are taken to be 0.15 (fraction), 100 mD and 20 mD, respectively. Furthermore, the Carmen-Kozeny constant, rock compressibility, rock heat capacity and thermal conductivity are assumed to be 5 (dimensionless), 5xl0~7 kPa 1 , 2.45xl0 6 J/m 3 .C[14,15], and 2.0736xl0 5 J/day.m.C (= 2.4 W/m.C) [16]. The properties of the pure components are obtained using NIST
218
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Standard Reference Data [10]. The hydration number and density of hydrate is assumed 7.3 (corresponds to the non-ideal stoichiometric reaction) [17,18] and 1100 kg/m 3 [17]. The frequency factor, enthalpy, and activation energy of hydrate formation reaction are 10 20 (gmole/m 3 )-7daykPa, 48.4 kj/gmole, and 81084.2 kj/gmole. The upper boundary of the reservoir is designed to be at constant pressure representing the ocean. The boundary conditions on the periphery and the bottom of the reservoir are considered to be a no flow boundary. Liquid C 0 2 is injected at maximum rate of 3000ton/ day at the temperature of 15°C. In this study, four cases are considered for modeling, each belonging to one of the classes (see Figure 3). The ocean depths are considered to be 600,1500, 2800, and 3500m for cases 1, 2, 3, and 4, respectively. The injection depth in the sediment for all four cases is 800m below the sea floor.
13.4
Results
13.4.1 Depleted Gas Reservoir Figure 4a shows changes in average values of pressure, temperature, and hydrate saturation with time during the C 0 2 injection into depleted gas reservoir. During the first 100 days, pressure increases with injection while no hydrate forms. At an average pressure of 2.54MPa, hydrate saturation is nearly zero. Then, as hydrate forms gas is consumed, so the rate of pressure increase lessens. Formation of hydrate is also signified by an increase in temperature, which continues to rise with increasing hydrate saturation. After 257 days and injection of more than 15 times the original gas-in-place, the pressure and temperature have increased to 3.95MPa and 9.2°C, respectively. Figure 4b provides a measure of storage capacity by showing volumes of C 0 2 and methane stored in the form of hydrate as a function of time. The volume of C 0 2 considerably exceeds that of methane. A comparison between the original gas-in-place and the volume of C 0 2 stored in hydrate suggests that the latter is approximately five times larger than the initial gas-in-place. Two-dimensional distributions of temperature and hydrate saturation at 250 days of C 0 2 injection into depleted gas reservoir are shown in Figures 5a and 5b, respectively. Temperature distribution exhibits the following features. It is higher in the vicinity of the
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
219
Figure 4. (a) Average values of pressure, temperature, and hydrate saturation as functions of time, (b) Volumes of carbon dioxide and methane stored in hydrate form.
wellbore due to injection of warm C0 2 . It is reduced at borders with the cold base- and cap-rocks as well as at the C 0 2 front because of the cold conditions there. Since temperature in the hydrate layer increases with time as hydrate forms, heat diffuses along the temperature gradients so that the lowest temperature is observed on borders with shale. Testing of the results has revealed that vertical temperature profile at the C 0 2 front is associated with gravitational segregation of C 0 2 in the vapor phase.
220
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 5. Two-dimensional distribution of, (a) temperature at 250 days, (b) hydrate saturation at 250 days.
Hydrate saturation in Figure 5b reflects the same features discussed above. First, hydrate is absent in the vicinity of the wellbore because of the high temperature. This allows continued injection without near-wellbore plugging. Second, saturation of hydrate at the borders with shale is higher than in the middle of the reservoir because of the decreased temperature. Hydrate saturation at the C 0 2 front is also increased due to a reduced temperature there. Settling of C 0 2 which promotes formation of hydrate causes radius of the hydrate zone at the bottom of the reservoir to exceed that at the top.
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
221
13.4.2 Ocean Sediment Figure 6 illustrates the C 0 2 volume distribution during C 0 2 injection into ocean sediment for the grids that include the well for Case 1 to 4. In Case 1 (Figure 6a) the C 0 2 front reaches to 50m below the seabed at the end of injection period and to the seabed 5years later. In this case, C 0 2 is highly buoyant and hydrate is not formed. In Case 2 (Figure 6b), where hydrate formation occurs at about 220m below the ocean floor, HFZ cannot prevent upward flow of buoyant C0 2 . C 0 2 enters the ocean 20years after the injection period. Case 3 (Figure 6c) and Case 4 (Figure 6d) exhibit both HFZ (with thickness of 360 and 376m for Case 3 and 4, respectively) and NBZ (with the thicknesses of 225 and 565m for Case 3 and 4, respectively). 450years after the injection period, the fronts of C 0 2 reach to the depths of 215 and 285m below the seabed for Case 3 and 4, respectively. An amount of 550MTonnes of C 0 2 is stored over the injection period in these two cases, without any leakage into the ocean. The figures also show the thicknesses of NBZ and HFZ calculated at the static conditions. The results indicate that the C 0 2 could flow upwards into the what was deemed to be NBZ and HFZ (under static conditions). The reason is that the temperature profile and pressure distribution in the sediment is disturbed because of injection. This suggests that careful flow simulations may be required for design of large-scale C 0 2 storage below ocean floor; however potential for large scale storage is available.
13.5
Discussion
In this study, we examined hydrate formation during a period of eight months of C 0 2 injection into depleted gas reservoir. During this period, the sensible heat of the reservoir between the initial temperature and equilibrium temperature at maximum pressure controls the amount of hydrate formed. However, over longer periods of time (either slower injection or shut-in after the injection period), the increased temperature in the reservoir dissipates and the reservoir cooling would lead to formation of more hydrate a n d / o r pressure reduction. In this case, the rate of hydrate formation would be controlled by conduction of heat. The motivation for this work was the fast kinetics of hydrate formation that could trap C 0 2 in solid cages of C0 2 -hydrate. It should be noted that, kinetics
Figure 6. Distribution of C0 2 volume per unit area for the grids that include well during the C 0 2 injection beneath the ocean floor, (a) Case 1, (b) Case 2, (c) Case 3, (d) Case 4.
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
223
of hydrate dissociation is also fast. Therefore, if pressure and temperature conditions of the reservoir are altered, the hydrate could dissociate and the C 0 2 would be released once again. This may be of practical implications in Northern Alberta, where some of the depleted gas pools are underlain by bitumen reservoirs that may be developed by thermal methods. Work is underway to map those depleted gas pools, where risk of heating by thermal recovery from below is absent.
13.6
Conclusions
We have studied injection of C 0 2 in two geological settings. For the case of a depleted gas reservoir, it was shown that by controlling the temperature of the injected C0 2 , one could avoid hydrate formation in the vicinity of the wellbore, where formation plugging may occur. Formation of hydrate away from the wellbore and at saturations observed in this study is not expected to affect injectivity. Simulation results indicate that more hydrate forms at bottom and top, where heat of hydrate formation dissipates to the baseand cap-rock. The volume of C 0 2 that turned into hydrate was many times the original gas-in-place. A number of other factors, including effect of other hydrocarbons in the in-situ gas or impurities in the injected gas, and reservoir cooling after a period of C 0 2 injection, both of which could increase potential of C 0 2 storage in depleted gas pools are being studied. Sub-ocean floor sequestration of C 0 2 into the deep ocean sediments was also investigated. The idea of permanent storage of C 0 2 in these sediments was numerically modeled and the fate of injected C 0 2 was simulated. It was demonstrated that two important thermodynamic barriers (Hydrate Formation Zone and Negative Buoyancy Zone) are developed in the upper parts of deep and super deep ocean sediments, which restrict upwards flow of C 0 2 towards the seabed. However, the numerical results indicated that the hydrate formation zone and negative buoyancy zone estimations are not static quantities. Pressure and temperature changes as a result of C 0 2 injection and hydrate formation could lead to C 0 2 movement towards the seabed, more than predicted by these static estimates. This study demonstrated an example of use of a numerical simulator that allowed design of C 0 2 injection depth and rate to avoid flow to the seabed. An opportunity for storage of large volumes of C 0 2 was demonstrated.
224
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
13.7 Acknowledgment The assistance of Dr. Dennis Coombe of Computer Modeling (CMG) in use of the STARS simulator for defining mixed-hydrates is gratefully acknowledged. Financial funding for this research was provided by the National Science and Engineering Council of Canada (NSERC).
References 1. O.Y. Zatsepina, and B.A. Buffett, "Nucleation of C0 2 -hydrate in a porous medium/' Fluid Phase Equilibria. 2002; 200 (2): 263-275. 2. M. Uddin, D. Coombe, and F. Wright, "Modeling of C0 2 -hydrate formation in geological reservoirs by injection of C 0 2 gas," Journal of Energy Resources Technology, Transactions of the ASME. 2008; 130 (3): 0325021-03250211. 3. J. Wright, M. Côté, and S. Dallimore, "Overview of regional opportunities for geological sequestration of C 0 2 as gas hydrate in Canada," Proc. 6th Intern. Conf. Gas Hydrates. (2008). Vancouver, Canada. 4. S. Adisasmito, R.J. Frank, and E.D. Sloan, "Hydrates of carbon dioxide and methane mixtures," J. Chem. Eng. Data. 1991; 36: 68-71. 5. E.D. Sloan, and C.A. Koh, Clathrate Hydrates of Natural Gases 3rd Ed. (2007). CRC Press. 6. S.S. Tarn, M.E. Stanton, S. Ghose, G. Deppe, D.F Spencer, R.P. Currier, J.S. Young, G.K. Anderson, L.A. Le, and D.J. Devlin, "A High pressure carbon dioxide separation process for IGCC plants", http://www.netl.doe.gov/ publications/proceedings/01/carbon_seq/lb4.pdf (accessed July 2001). 7. T. Uchida, I.Y. Ikeda, S. Takeya, Y. Kamata, R. Ohmura, J. Nagao, O.Y. Zatsepina, and B.A. Buffett, "Kinetics and Stability of CH4-C02 mixed gas hydrates during formation and long-term storage," Chem. Phys. Chem. 2005; 6 (4): 646-654. 8. B. Buffett, and O.Y. Zatsepina, "Research on mixed gas hydrate equilibrium," Report to Institute of Energy Utilization, AIST, Japan; 2004. 9. R. Span, W. Wagner, "A new equation of state for carbon dioxide covering the fluid region from the triple-point temperature to 1100 K at pressures u p to 800 Mpa," / Phys Chem RefData 1996; 25 (6); 1509-96. 10. National Institute of Standards and Technology (NIST). http://webbook. nist.gov/chemistry/. NIST Chemistry Webbook. NIST Standard Reference Database Number 69. The US Secretary on behalf of the United States of America 2008. 11. UCAR, Windows to the Universe, http://www.windows.ucar.edu/. Regents of the University of Michigan 2008. 12. K.Z. House, D.P. Schräg, CF. Harvey, and K.S. Lackner, "Permanent Carbon Dioxide Storage in Deep-Sea Sediments," Applied Physical Sciences 2006; 103 (33): 12291-5. 13. H. Koide, Y. Shindo, Y. Tazaki, M. Iijima, K. Ito, N. Kimura, and K. Omata "Deep Sub-Seabed Disposal of C 0 2 - The Most Protective Storage," Energy Convers. Mgmt 1997; 30 Suppl: S253-8
C 0 2 AND ACID GAS STORAGE IN GEOLOGICAL FORMATIONS
225
14. R. Wicander, and J.S. Monroe, Essentials of Geology, West Publishing Company. 1995. 15. W.D. Waples, and J.S. Waples, A Review of Specific Heat Capacities of Rocks, Minerals, and Subsurface Fluids. Part 1: Minerals and Nonporous Rocks, International Association for Mathematical Geology. Natural Resources Research 2004; 13 (2). 16. C. Clauser, and E. Huenges, "Thermal Conductivity of Rocks and Minerals," American Geophysical Union 1995. 17. I. Aya, K. Yamane, and H. Nariai, "Solubility of C 0 2 and Density of C 0 2 Hydrate at 30Mpa." Energy 1997; 22:263-71. 18. A. Bozzo, H.S. Chen, J.R. Kass, and A.J. Barduhn,. "The Properties of the Hydrates of Chloride and Carbon Dioxide," Desalination 1975; 16: 303-20.
This page intentionally left blank
14 Complex Flow Mathematical Model of Gas Pool with Sulfur Deposition W. Zhu, Y. Long, Q. Liu, Y. Ju, and X. Huang Civil & Environment Engineering School, University of Science & Technology Beijing, Beijing, People's Republic of China
Abstract The gas-liquid-solid complex flow is a complicated gas pool with sulfur deposition, whose phases may interchange with mass transfer, resulting in changes of gas & liquid saturation in the reservoir. Sulfur deposits on the surface of porous media may contribute to the porous change affecting the penetrating of gas and liquid, which might destroy the reservoir. A multiphase liquid-solid coupling mathematical model was set u p in porous media based on mathematical models of sulfur deposition, capillary force and relative permeability (taking into account capillary & inter facial forces). The numerical simulation results prove that the sulfur deposit decreases the pores and increases the flow resistance in some regions. The prediction without considering sulfur deposition will lead to over estimation of the output. The model also reveals the physical essence of the gas-solid-liquid complex flow.
14.1
Introduction
The g a s pool w i t h sulfur is q u i t e special, s h o w i n g complexity in i n t e r p h a s e m a s s transfer a m o n g sulfur d e p o s i t i o n , liquid p h a s e a n d g a s e o u s p h a s e . Sulfur d e p o s i t i o n is w h a t s e p a r a t e s from h i g h sulfur content gas, k n o w n as sulfur particles, w h i c h s p r e a d w i t h the gas flow a n d g a t h e r / s u b s i d e to block t h e p o r e throats d u e to d e c r e a s e of g a s flow velocity. Besides, t h e h y d r o g e n sulfide
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (227-244) © Scrivener Publishing LLC
227
228
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
hydrolyzes in water, reacts with rocks/minerals, turns into insoluble sulphides and forms the sulfur dirt on the pore surfaces. The phase changes of gas with sulfur are comparatively complicated. Many phases coexist such as gas-solid, gas-liquid-solid and gasliquid with declination of temperature and pressure, and sulfur can exist as gas, liquid and solid. The phase changes may lead to gas & oil saturation changes in the reservoir and sulfur deposition on the surface of the porous media will contribute to pore changes, which affect the gas/liquid penetrating and destroying the reservoir. Gas, liquid and solid distribution in the microspace will affect the gas flow. Additionally, phase changes might result in alteration of the temperature field as well as the interstitial fluid pressure changes, thus influence the characteristics and courses of gas penetration. Meanwhile, due to the high speed of the flow near wells, the fluid movement is in accordance with the generalized Darcy's law. In order to study the gas-liquid-solid multiphase flow with sulfur deposition more accurately, a mathematical model of multiphase complex flow with phase change was set up in this paper.
14.2
The Mathematical Model of Multiphase Complex Flow
14.2.1 Basic Supposition In a gas pool, fluids and rocks meet the following conditions: the water and gas in the reservoir both accord with the non-Darcy flow; the rocks can be compressed a little; there are N fixed hydrocarbon quasi-components in the oil-gas system, which can more precisely reflect the interphase mass transfer among oilgas fluids and can meet the petrochemical engineering and gas pool exploitation requirements; the flow course is regarded as isothermal; solid sulfur is considered to be incompressible; sulfur is absorbed on porous media, but some moves with gas/liquid, and sulfur deposition is determined by phase equilibrium and deposition laws; chemical actions do not occur during the phase change and flow course, so matter characteristics (such as the density, heat capacity and the heat-conduction coefficient) remain constant.
MODEL OF GAS POOL WITH SULFUR DEPOSITION
229
14.2.2 The Mathematical Model of Gas-liquid-solid Complex Flow in Porous Media 14.2.2.1
Flow Differential
Equations
1. Quality conservation equations Gas phase:
d\pLgRsSL+(l-yc-zc)pgs] 0— V dt = -V[pLgRsvL+(l-yc-zc)pgvg]
(1)
Liquid phase: —
^ — — = -^[pLvL+ycpgvg\
(2)
Mixture:
j[pLSL *
+ pLgRssL + Pgsg + psss] 3^
= " v [ A A + PLSKVL
+ Pgvg + psvs ]
(3)
In the formula, 0 stands for the porosity, p , po and p L are density of gas phase, density of gas dissolved in oil phases and density of liquid under normal conditions respectively, Rs is the dissolved GOR, So is the oil saturation, S is the gaseous phase saturation, v is the gaseous phase velocity, vL stands for the liquid phase velocity, vs is the solid phase velocity, yc and zc are the liquid and solid sulfur components in the gaseous phase respectively, while f stands for the time. 2. Momentum equations The speed of gas and liquid flow in porous media, especially near the well casing, is very high. So the fluid movement is in accordance
230
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
with the generalized Darcy's law. The gas and liquid plane radial flows can be expressed as: Gas phase:
(4) Liquid phase: v^KKrL^KKrLlP
(5)
Ik Solid phase: Sulfur flows with gas to the direction in which pressure decreases, whose speed relates to the shear velocity and gas speed when sulfur particles start to move, moreover the sulfur flow course is with the suspended motion and the bed load movement: (6)
vs=as(vg-vgc) V
(7)
ocP
Here, v c is the particle commencement shear velocity, K is the absolute permeability, Kr and XrL are gas and liquid relative permeability respectively, Pc is the capillary pressure, p is the pressure, r is the radial distance, ju and /^are the viscosity of gaseous and liquid phases respectively and as is the equation coefficient. 14.2.2.2
Unstable Differential Equations of Gas-liquid-solid Complex Flow
Substitute Equations (4), (5), (6) for Equation (3), then the following equation can be obtained: ld_ r dr =
d_
~dt
rK
X
* {pL+PLgR5)
pA+PLgRssL BL
+^ r {
P i
+
Psas)
pgs:
- + -B„
+ PSSS
(8)
Here, B and BL are volume coefficients of the gas and liquid phases.
MODEL OF GAS POOL WITH SULFUR DEPOSITION
231
From Equation (8) it is easy to know that all parameters relate to the pressure and the equation is non-linear. A virtual pressure coefficient \ff(p) is introduced to make the equation easy to use: K
y/(p) = [
^(pL
+ PLgRs) +
J^-(px
M,
ßLBL
+ Psas)
•dp
(9)
Based on the average law, Equation (8) can be changed as:
IJLÍrM=JL¿E rdr\
dri
(10)
Dh dt
- f g - (PL + PLgK ) + - J - (Pg + Ptat )
K VLBL (¡> c pL +
D
"
sL
ßxVx pLgR.pg(l-SL-Ss)
-+-
A
B„
(11)
+psss
Here, pb is the saturation pressure, Dh is the defined intermediate variable. 14.2.2.3
Relationship between Saturation and Pressure of Liquid Phase KL
-zc)
Pißi.
R,
rg
KL u BB VL L Since, Vp I
dSn
PogK B„
PLS +
8
Kg
P Vc B uPgB g *
(12)
= 0 > then the saturation equation is:
+s„
dp
f
R,
pg(l-yc-zc)
Po _ PgVc X
B„ v
-R,
B„
B„
Pgtt-Vc-^
B,
PsVc + S„ B„ _ PogK
B„
j
In the equation, Rz is the defined intermediate variable.
(13)
232
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
14.2.2.4
Auxiliary
Equations
Capillary force and relative permeability equations: Kro=KJSg,Nc,oog)
(14)
Krg=Krg(Sg,Nc,aog)
(15)
P^Po
+ Pœg&g'K'Oog)
(16)
In the equations above, Nc is the capillary number, a o is the interfacial tension between gas and oil, po is the oil phase pressure, p is the gaseous phase pressure, and pco is the gas-oil capillary force. Constraint condition: Sg+Sl+Ss=l
(17)
Here, S1 is the liquid phase saturation. 14.2.2.5
Definite
Conditions
1. Gas pool original state „ |f=o __ B o
Po I n + T
PoSi+T '
T|f=o _ T o l I Q+T - L Ci+T°
2. Boundary state The closed outer boundary VO|route). = 0 , the constant pressure outer boundary p\router - Po > the constant bottomhole flow pressure p\nnner = const and the bottomhole constant f[dp/dn~\\rinner = const. Here, Q is the study domain, T is the temperature, V<£> is the potential gradient function, Touter is the outer boundary and Tinner is the inner boundary.
14.3
Mathematical Models of Flow Mechanisms
14.3.1 Mathematical Model of Sulfur Deposition Sulfur deposition mechanisms include molecular diffusion, shear diffusion, Brownian diffusion, gravity subsidence, etc., among
MODEL OF GAS POOL WITH SULFUR DEPOSITION
233
which molecular diffusion and shear diffusion are considered to be major types of sulfur deposition according to some research. So the sulfur deposition model is: dW_=dWL+dWL dt dt dt
(18)
Here, W is the total deposition amount, Wd is the molecular diffusion deposition amount, Ws is the shear diffusion deposition amount. 1. Sulfur molecular diffusion deposition model According to the Fick Diffusion Law, the molecular diffusion deposition speed of sulfur can be expressed as: dWL dt
= c c P d a
A fi
idÇ| l dTl
dT dr
(19)
In the above formula, dW d /df is the mass of dissolved sulfur deposit from molecular diffusion in unit time, Cd is the deposit constant (generally 1500), C : is the liquid phase concentration, A is the surface area for sulfur deposition, ¡x is the liquid viscosity, C is the volume fraction of sulfur to the crude oil, dC/df is the volume fraction gradient of sulfur in the liquid and dT/dr is the radial thermal gradient. 2. Sulfur shear diffusion deposition model Sulfur particles behave two ways of horizontal migration, i.e. Brownian movement and shear diffusion, but the influence of Brownian movement is relatively small. Because of the porous flow speed-gradient field, the sulfur particles suspended in the oil flow will rotate in an angular velocity, contributing to their horizontal movement and shear diffusion. The sulfur shear deposition gradient caused by speed gradient with laminar flows can be expressed as the following: ^ - = Cäk'C'rA dt
(20)
In the equation, dW s /d£ is the mass of dissolved sulfur deposition from shear diffusion in unit time, k* is the shear deposition rate constant, C* is the volume fraction of sulfur particles on the surface and / i s the shear velocity.
234
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
14.3.2 Thermodynamics Model of Three-phase Equilibrium 1. Fugacity equilibrium equations (21)
fi=fi=n
(22)
f!=Mp
(23)
f;=4fr=xtffr
(24)
In the equations above, /,- , f¡ and ff are fugacity of the component i in the gaseous, liquid and solid phases respectively,
X
i
- $i
(25)
X:
Solid-liquid equilibrium:
_ A _ ¡p
(26)
3. Solid phase parameter Fugacity of the solid standard state f°
(1 f? = f?e*P .M RT
T) + h*k
T-
f
R
TÍ b,M ; ■ln-*-)T 2R
wf
+ T-2TJ
(27)
MODEL OF GAS POOL WITH SULFUR DEPOSITION
235
Solid activity coefficient:
RT In the equations above, H¡ stands for the dissolution enthalpy of the component i, T) is the solution temperature of the component i, Mi is the relative molecular weight of the component i, R represents the universal gas constant, bx and b2 are the equation coefficients, ¿%¡ is the solubility parameter of the solid phase mixture, ¿f is the solid solubility parameter of the component i, V* is the solid substance volume of the component i, KJ] and K¡ are gas-liquid and liquid-solid equilibrium constants and r¡ is the solid activity coefficient. 4. Material balance equations of gas-liquid-solid three phases V +L+S =l
(29)
Vxvi+Lx\+Sxsi=Zi
(30)
X*<+X*Í + Z*?=2 Z < = 1
(3D
5. Flash vaporization equations of gas-liquid-solid three phases According to the material conservation principle for gas-liquidsolid phase equilibrium as well as the definition of the equilibrium constant, the three-phase flash vaporization equations are derived as follows: y — j — % _
^V(K?-l)
+ S(K?-l) + l
V
ML
^V(Kf-l)
+ S(Kf-l)
V
MIL
^V(Kf-l)
+ S(Kf-l)
=i
02)
=1
(33)
+l
=i +l
(34)
236
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
In the equations, V, L and S represent the fractions to amount of substance of gaseous, liquid and solid phases in equilibrium, Z is the total amounts of substance of all components. Usually the Newton - Simpson method is adopted to solve equations and to get the fractions to amount of substance V, L and S of gaseous, liquid and solid phases in equilibrium as well as the compositions to amount of substance of these phases xj, x, and x¡ . 14.3.3
State E q u a t i o n s
The PR state equation is selected to express the state equations of the system: _ RT aa(T) P
~V-b
V(V + b) + b(V-b)
R2T> a; =0.45724xP*
(35)
(36)
&..= 0.07780 x ?* (37) Here, R is the universal gas constant, T is the system temperature, p is the system pressure, Tc is the critical temperature and pc is the critical pressure.
14.3.4
Solubility Calculation Model
The main factor influencing sulfur to deposition is the solubility of sulfur in natural gas, and with the increasing of the solubility of sulfur, it becomes difficult for simple substance sulfur to separate and deposition. Therefore, it is essential to set up the solubility prediction model of simple substance sulfur in the natural gas. According to thermodynamics and experiment results, the relationship among solubility of sulfur, pressure and temperature in the acid natural gas is as the following: C = [Mar I(ZRT)]exp(-4666
/ T - 4571)P4
In the formula above: C is the solubility of the natural gas, g/m 3 ; P is the pressure, MPa; T is the temperature, K; M a is 2897, the relative
MODEL OF GAS POOL WITH SULFUR DEPOSITION
237
molecular quality of dry air; R is the density of the natural gas; Z is the deviation factor of the natural gas; R is the gas constant. 14.3.5
I n f l u e n c e M a t h e m a t i c a l M o d e l of S u l f u r D e p o s i t i o n Migration to Reservoir Characteristics
Sulfur deposition migration and porous medium sorption to sulfur particles might influence on porous medium characteristics, such as porosity and permeability. The original porosity of the reservoir is <))0, then the porosity change caused by sulfur deposition is: 0-es/Ps
(38)
In the formula, p s is the solid particle (sulfur) density, es is the particle mass increment in pores of per unit rock volume. During the flow process of solid particles, some deposited solid particles may be released and entered the fluid again because of the surface deposition and the shearing force. Thus deposition speed minus release speed is the net deposition speed. According to the dynamics equations, the deposition speed can be expressed as:
3^ dt
■■R.-R.
(39)
The deposition rate Rr with the particle mass contained in the fluid per unit rock volume and the flow rate is in direct proportion.
dW l + dt I
K=
ar(urujc)
(40)
The re-entering speed Re of particles mainly depends on the hydrodynamic conditions.
When|_^y_^ dx
, it can be expressed as:
dx Jcr
R.
Ms
(41)
238
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
When
dp < dx
-
dp dx R =0
(42)
This means that in order to make the deposited particles re-enter the fluid, the hydrodynamic pressure gradient (-dp/dr) must be greater than the critical pressure gradient (-dp/dr)ci. The permeability reduction is in close relationship with the porosity changes. They are in direct proportion, and the higher the fluid velocity is, the more sulfur deposits and the more permeability drops. Release and migration of solid particles in porous medium flow course will obviously alter the primitive porosity and permeability, aggravate the heterogeneity of porous media and reduce the permeability. Therefore, permeability reducing formula can be expressed as: K_
1—
(43)
v 0o y
In the formula, K is the permeability after sulfur deposition, KQ is the initial penetration, § is the porosity after sulfur deposition, <\>0 is the original porosity, ui is the gas-liquid mixture flow rate, wc is the critical flow rate of gas-liquid mixture when sulfur released and m is the equation exponent.
14.4
Solution of the Mathematical Model Equations
14.4.1 Definite Output Solutions 1 d J 3^1 1 d¥ r dr 1 drl D„ dt
(44)
yr(r,0) = ^
(45) (46)
r dr
2nKh
MODEL OF GAS POOL WITH SULFUR DEPOSITION
^ 1 dr
=0
r=r
239
(47)
<
As to the initial stage of the unstable flow, the gas-liquid flow pressure drop is: , ,. mt 2.25Dht y/i-y/(r,t) = -—t—ln ^ áTüKn r The bottomhole pressure drop is:
,AQ, (48)
¥i-¥(rw,t) = -^-0n^^+2S) AnKh rw
(49)
Here, rw is the shaft radius, mt is the total flow mass of the gas well, h is the effective thickness of the reservoir, \¡f{ is the original reservoir quasi-pressure, r is the radial flow distance and t is the production time. For the quasi-steady stage, the quasi-drop of pressure is: yfi-i/f(r,t) '
= -—'—(—==- +In-2- — + —=-) 2xKh r2 r 4 2r2
(50)
The bottomhole pressure drop is: ^ - V ^ , 0 =^ ( iTtKn
^ re
+
l n ^ - | + SH) rw 4
(51)
Here, SH is the well skin coefficient, yKrw,t) is the bottomhole quasi-pressure at the production time f, y/(r,t) is the quasi-pressure at r away from the well at the production time t. 14.4.2
Productivity Equation
According to the model equations and the solutions, the quasisteady capability equation can be obtained:
mtt=2M y - y y >
r, 3 ln-^--- + SH r,„ 4
In the equation, ty/is the average pressure of the reservoir.
(52)
240
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
14.5
Example
14.5.1
Simulation Parameter Selection
The multiphase complex flow mathematical model above was utilized for a well and the PR state equation was selected in the simulation. The production well section is from 4958.0m to 4963.0m, the gas pool thickness is 5.0 m, the primitive stratum pressure is 55.45 MPa, the stratum temperature is 407.05 K, and these simulation compositions of gas-oil are shown in Table 1. The simulation basic parameters: the single well control radius is 700m, the porosity is 14.5% and the reservoir effective permeability is 4.844 x 10 3 um 2 .
14.5.2
Oil-gas Flow Characteristics near Borehole Zones of Gas-well
The mathematical model was employed to predict the leading edge radius of the multiphase flow region, and the simulation result is as shown in Figure 1. Sulfur deposition might affect the leading edge radius of gas-liquid flow and make the leading edge radius shrink, e.g., the radius of the TH 2 well in this simulation decreased as much as 15 m. Therefore, Sulfur may make pores in some region reduce and the flow resistance increase. It is necessary for low permeability sulfurbearing gas field to control effectively the sulfur deposition during the exploitation for purpose of avoiding harm to the reservoir.
14.5.3
Productivity Calculation
According to the two phase quasi-pressure result and the productivity equation, the IPR curve of the quasi-steady state flow stage can be obtained. From Figure 2 it can be recognized that the two curves Table 1. Fluid compositions in gas pool. Relative Density
0.68
CH 4 Volume Fraction / %
C2H6 Volume Fraction / %
H 2 S Mass Concentration / (g.nr3)
C 0 2 Mass Concentration y / (g.m 3 )
80
0.1
222
69
MODEL OF GAS POOL WITH SULFUR DEPOSITION
241
Figure 1. The influence on sulfur deposition to the leading edge radius in multiphase region.
Figure 2. Impact of sulfur deposition on the IPR curve of Well TH2.
on influence of sulfur deposition to productivity when the initial pressures are 40MPa and 55.5MPa respectively. The output might be over-estimated without considering the sulfur deposition with the impact on the primitive IPR curve, the output reduction can be up to 15% due to sulfur deposition. Therefore, it is very important for reservoir protection and development to control effectively sulfur deposition in the gas field with sulfur. It is especially beneficial to control the pressure of the reservoir in the low permeability gas field and prevent the service life from being shortened because of sulfur deposition.
242
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
14.6
Conclusions
1. Based researched on the gas-liquid-solid flow physical simulation and phase changes, established the mathematical model of multiphase flow with gas-liquid-solid phase changes accompanied by sulfur deposition. The flow changes and the complex flow features caused by sulfur deposition were analyzed, and gas-liquid- solid complex flow characteristics and courses could be described accurately. 2. Studied on the multiphase flow theory of gas-liquid-solid phase changes with sulfur deposition and established the mathematical model of multiphase fluid-solid coupling seepage. The model showed the characteristic of multiphase flow with gas-liquid-solid phase changes while the liquid and sulfur released. This model could serve as theoretical foundation for future research on dynamic exploitation predicting, numerical simulating and field engineering. 3. Sulfur deposition can make pores in some region reduce and make the flow resistance increase. The prediction might be over-estimated without considering the sulfur deposition, to which the output reduction can be attributed. Therefore, it is very important for reservoir protection and development to control effectively sulfur deposition of the gas field with sulfur. 4. The simulation results show that the model could reflect the physical essence of the flow well, and coincide with the pilot results.
14.7
Acknowledgement
This research was supported by the National Natural Science Foundation of China (10772023) and the National Key foundation of China (50934003).
References 1. Bentsen R G. Effect of Momentum Transfer between Fluid Phases on Effective Mobility. / Pet Sei Eng, 1998, 21 (1-2) 27
MODEL OF GAS POOL WITH SULFUR DEPOSITION
243
2. Morrow N R. Interfacial Phenomena in Petroleum Recovery. Monticello, USA, Mercel Dekker Inc, 1991 3. Zhu Weiyao. Liu Xuewei. Luo Kai. "Dynamic Model of Gas-Liquid-Solid Porous Flow with Phase Change of Condensate Reservoirs." Natural Gas Geoscience. 2005,16 (3) 363 4. Zhu Weiyao. "Theoretical Studies on the Gas-Liquid Two-phase Flow." Petroleum Expoloration and Development, 1988,15 (3) 63 5. Zhu Weiyao. "Theoretical Studies on the Gas-Liquid Two-phase Flow (Including a Phase Change Through Porous Media)." Acta Petrolei Sinica, 1990, 9 (1) 15 6. Yu Mingzhou, Lin Jianzhong. "The Dynamics of Nanoparitcle-Laden Multiphase Flow and Its Applications." Mechanics in Engineering, 2010,32 (6) 1 7. Zhang Dihong, Wang Li. Zhang Yi. "Phase Behavior Experiments of Luojiazhai Gas Reservior with High Sulfur Content." Natural Gas Industry, 2005,25 SI
This page intentionally left blank
SECTION 4 ENHANCED OIL RECOVERY (EOR)
This page intentionally left blank
15 Enhanced Oil Recovery Project: Dunvegan C Pool Darryl Burns Green Joule Energy Ltd
Abstract In this project, a template method was developed for using reservoir simulation with computer assisted history match for conventional oil pool characterization and evaluation of enhanced oil recovery alternatives. This method was applied to the Dunvegan C Pool to evaluate recovery and economics of several recovery alternatives including acid gas flood, carbon dioxide flood, additional drilling of horizontal wells, pump off operations, and continued rate limitation (MRL) operations. Only data in the public domain were available for this study. A program was developed to use data available for the Dunvegan C Pool to characterize the reservoir fluid for use in both CMG IMEX black oil, and CMG GEM compositional reservoir simulations. Miscibility of various acid gas mixtures with the reservoir fluid was studied. Computer Modeling Group's CMOST optimization software was used to complete a computer assisted history match using the IMEX black oil reservoir simulator to complete individual simulations. Geological uncertainties in the pool were represented by thirty individual parameters put into CMOST. The history match operation was successful with global objective function error of less than 5% in all optimized cases. The optimized history match cases in IMEX were used to evaluate the horizontal well drilling, pump off, and MRL recovery alternatives. The optimized history match cases were also converted to GEM in order to simulate an acid gas flood on the pool. GEM was also used to evaluate a separate C 0 2 flood case. The p u m p off cases generally realized the highest Net Present Value of the pool, with the minimum capital investment. All acid gas flood economics are marginal. A positive change in oil price, capital cost,
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (247-318) © Scrivener Publishing LLC
247
248
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
royalty framework, or other government policy are required in order for an acid gas flood development to proceed. Comparing royalty received from the base 'Pump Off cases versus potential royalty from the acid gas floods, government has significant negotiation room in the royalty framework.
15.1 Introduction The Dunvegan C Pool in the Grande Prairie Field is located in Section 20-071-04W6M, east of Grande Prairie, Alberta. The geological interpretation of the pool is a deltaic sandstone and shale deposit 600 to 800 m wide and u p to 8 m thick and it is trapped stratigraphically with a down dip water leg (1). Original Oil in Place according to public sources is estimated to be 789 E3M3 of 35.8 API oil at initial reservoir conditions of 8 MPa and 34 °C (2). Since its discovery in 2006, 16.8 E3M3 of oil has been produced from the four wells in the pool, and remaining reserves from primary production estimated to be 62.1 E3M3 (2). Recovery Factor from primary production is expected to be 10%. The pool has exclusively been on primary production, with no waterflood or tertiary recovery schemes considered or implemented. The Energy Resources Conservation Board has approved Good Production Practice (GPP) operation for this pool, indicating that secondary and tertiary recovery methods were not economic at the time of GPP approval. The objectives of this project are; • Provide a template for the process of using reservoir simulation for conventional oil pool characterization, history match, and evaluation of enhanced oil recovery alternatives. • Confirm Original Oil in Place of the pool, identify alternatives for improving recovery from the pool, and quantify incremental recovery, production forecast, injection forecast where applicable, capital costs, operating costs, simplified economic analysis and economic summary of these alternatives. • Technically evaluate the potential for combining acid gas disposal with enhanced recovery operations from this pool.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
249
15.2 Pool Data Collection Data on all oil pools in Alberta is collected by and available through the Energy Resources Conservation Board (ERCB). A variety of tools are available to view and screen this data, including subscriptions to ERCB publications, and database software. The sequence of data acquisition, and data sources for this project were; 1. Database software was used to screen and select a 'typical' small recently discovered conventional oil pool for this project. 2. A search for resource applications related to the pool was then conducted using the ERCB Integrated Application Registry. Several applications were found, and requests to [email protected] were made to retrieve copies of the applications. Geological information including a net pay map and log cross section of the pool was obtained. A core analysis report on 10-20-071-04 W6M should also have been available because core was taken on the well, but the operator had not submitted the report at the time of this study. The net pay map of the pool is given in Figure 1. 3. Well, produced fluids, and pressure data were obtained by request from [email protected]. Fluid data including representative gas, and oil analyses were found. ASTM D86 crude oil distillation data extracted from a .pas file for the 07-20 well is provided in Table 1. Oil viscosity data from the same .pas file for 07-20 is provided in Table 2. Gas analysis from a separate .pas file for the 07-20 well is provided in Table 3. A produced water analysis for the pool was not available from the ERCB. Pool pressure data is presented with pool production data in Figure 2. 4. Drilling fluid and completion treatment data were obtained by requesting daily drilling and completion reports for each of the four wells from the ERCB Core Lab. Core for the 10-20-071-04 W6M well was available for logging and analysis; however the operator had not submitted results of routine core analysis at the time of this report. Independent analysis of core is outside the
250
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
scope of this study, but would normally be part of the reservoir characterization process. 5. Production data was imported from database software. Aggregate pool production and pressure data are provided in Figure 2.
Figure 1. Dunvegan C pool net pay map.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
251
Table 1. ASTM D86 crude oil distillation 07-20-071-04 W6M. ASTM D86 Distillation Data 07-20-071-04 W6M Temperature
Distillate Recovered Vol Fraction
C 84.9
0
108.1
0.05
129.4
0.1
146.6
0.15
172.8
0.2
190
0.25
211.3
0.3
233.5
0.35
255.7
0.4
284.1
0.45
306.3
0.5
331.6
0.55
355.8
0.6
356.8
0.62
IBP
END
Table 2. Crude oil viscosity data 07-20-071-04 W6M. Crude Oil Viscosity Data 07-20-071-04 W6M Temperature °C
Absolute Viscosity mpa-s
Kinematic Viscosity mm2/s
25
8.41
10.14
38
4.37
5.32
50
2.39
2.94
252
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 3. Gas analysis 07-20-071-04 W6M. Gas Analysis 07-20-071-04 W6M Component H2S
0.00%
co 2
0.34%
N2
0.34%
H2
0.00%
He
0.01%
CI
96.23%
C2
1.48%
C3
0.92%
iC4
0.16%
nC4
0.22%
iC5
0.06%
nC5
0.06%
C6
0.04%
C7+
0.14%
Total
15.3
Mol%
100.00%
Pool Event Log
A Pool Event Log was made to correlate events with pool data, assist in the history matching process, and generally better understand the pool. Ideally the event log will capture any operational change to the pool. The event log presented here in Table 4 is abbreviated because daily operation logs for each well were not available for this study. Drilling events were omitted here, but should be added to capture the Gel Chem drilling fluid system used for drilling the wells in the pool.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
253
Figure 2. Dunvegan C pool cumulative production and pressure data.
Table 4. Pool event log.
Dunvegan C Pool
Wells
KBelev.
Event Log
03-20
642.3
Darryl Burns
07-20
643.2
1/21/2010
10-20
642.9
13-20
647.3
Date
Well
Event Description
7-NOV-05
13-20
perforated 960.0-963.0 mKB (312.7-315.7 mSS) 14 spm
26-Nov-05
13-20
fractured
29-Nov-05
13-20
static gradient, IP=B181 kPa
4-NOV-06
07-20
perforated 961.5-967.0 mKb (318.3-323.8 mSS) 17 spm (Continued)
254
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 4. Pool event log. (Continued) Date
Well
Event Description
6-N0V-O6
07-20
20 tonne gelled hydrocarbon fracture stimulation, 58 m3 of fracoil with 1 m3 15% HCI spearhead
lO-Nov-06
07-20
fluid sampled for oil and gas analysis
I6-N0V-O6
07-20
flow and build u p test, P=8840 kPa
26-NOV-06
13-20
15 tonne gelled hydrocarbon fracture stimulation, 44.8 m3 of fracoil with 1 m3 acid spearhead
13-Dec-06
13-20
flow and build u p test, P=8542 kPa
14-Dec-06
13-20
Pump installed Pool Added to ERCB MRL Order list, BWR=8 m 3 / d , MRL=GPP, Base GOR=N/A m 3 /m 3
l-Jan-07
17-Feb-07
10-20
perforated 1080.2-1084.8 mKb (311.6-315. lmSS) 17spm
19-Feb-07
10-20
21 tonne gelled hydrocarbon fracture stimulation, 62 m3 of frac oil with 1 m3 15% HCI spearhead
21-Feb-07
10-20
Kudu pump installed
20-Mar-07
03-20
perforated 1072.0-1076.0 mKb (324.5-327.9 mSS) 17 spm
21-Mar-07
03-20
21 tonne Fracsol oil fracture stimulation, 65 m3 of frac oil with 1 m3 15% HCI spearhead
24-Mar-07
03-20
static gradient, P=7802 kPa
24-Mar-07
03-20
Kudu pump installed ERCB MRL Order Revised, BWR=8 m 3 /d, MRL=8 m 3 / d , Base GOR=100m 3 /m 3
l-May-07 12-Dec-08
03-20
acoustic well sounder, estimated P=6279 kPa
13-Dec-08
07-20
acoustic well sounder, estimated P=6287 kPa
14-Dec-08
10-20
acoustic well sounder, estimated P=7029 kPa
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
255
The Event Log proved valuable for understanding many characteristics of the pool, including qualitative assessment of completion treatment area of influence, timing and quality of pool pressure data measurements, possible causes of water production, and well operating constraints. One important source of information that was missing from the data set for the pool was good quality well test data. Well test data could have been interpreted to determine permeability and skin at each of the wells in addition to indications of directional permeability, faults or other boundaries, and other reservoir properties.
15.4
Reservoir Fluid Characterization
A program was developed to use data available for the Dunvegan C Pool to characterize the reservoir fluid for use in both CMGIMEX black oil and CMG GEM compositional reservoir simulations. Miscibility of various acid gas mixtures with the reservoir fluid was also studied. Acid gas containing 80 mol% H2S and 20 mol% C 0 2 is miscible with the reservoir fluid at 8.2 MPa and reservoir temperature (34 °C). Pure C 0 2 is miscible with the reservoir fluid at 13.3 MPa and reservoir temperature. Minimum Miscibility Pressures are elevated due to the presence of methane in the reservoir fluid. Laboratory confirmation of hydrate formation conditions is recommended since reservoir temperature of 34 °C is approaching hydrate formation temperature of approximately 30 °C with an Acid gas containing 80 mol% H2S and 20 mol% C0 2 . 15.4.1
F l u i d Characterization P r o g r a m D e s i g n Questions
Several questions need to be answered when designing a petroleum reservoir fluid characterization program. These questions, and answers for the Dunvegan C pool, are provided below, followed by a description of the fluid characterization program. 1. What fluid data is already available? Fluid data available for the Dunvegan C Pool included an ASTM D86 distillation curve generated from a crude oil sample from the 07-20 well (Table 1), viscosity data generated from the same
256
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
sample (Table 2), and a separator gas analysis also from the 07-20 well (Table 3). A water analysis was not available, so 20,000 ppm salinity was estimated from data in Alberta Research Council Open File Report 1990-18, Hydrochemistry of the Peace River Arch Area, Alberta and British Columbia. 2. What were the initial fluid saturations in the reservoir? Initial estimated saturations were So=.51 and Sg=.03, based on presimulation material balance calculations assuming Sw=.46, indicating saturated oil in equilibrium with a gas phase. The material balance post simulation was revised with Sw=.56 in response to well log re-interpretation, pool pressure data, and history match results. 3. What are the density, viscosity, and molecular weight of the reservoir fluid(s)? Reported oil gravity from the D86 analysis is 37.6 °API, slightly higher than 35.8 °API reported on the pool card. Oil viscosity data is provided in Table 2. Molecular weight of the oil was not reported, so a procedure was developed to estimate it from the D86 distillation curve. Details of this procedure are provided below in this section of the report. Based on this procedure, the estimated molecular weight of the 'stock tank' oil is 207 kg/kmol. 4. How much industry experience is there with this fluid? Are there correlations in the literature that can be used to estimate properties of this fluid at the anticipated reservoir pressures and temperature? Correlations are available in the literature for crude oils ranging from 10-50 °API gravity, with oil viscosities of up to 100 cp, at reservoir pressures ranging from 3-40 MPa at 15-125 °C. Initial reservoir conditions for the Dunvegan C pool were 8 MPa and 34 °C. The fluid conditions are well within the range of correlations in the literature. Further discussion of this is provided later in this section. 5. How will the results of the fluid characterization be used? For this project, the fluid characterization will be used for 'proof of concept' study purposes. At this stage of project development, it is appropriate to evaluate oil recovery options based on the fluid information on hand. Time and money can be spent on more focused laboratory work once the most attractive enhanced recovery options are identified.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
257
6. What tools will be used to characterize the fluid? (different tools may require different inputs, so it is important to know u p front what you are using) For this project, CMG's Winprop utility was used to characterize the fluid. 7. How will fluid interactions with injected displacing fluids be modeled? CMG's GEM compositional simulator was used to model injection operations on the Dunvegan C pool, so the lack of laboratory fluid and core flood studies was somewhat compensated for by using a compositional reservoir simulator. This approach is only acceptable for 'proof of concept' study purposes. Laboratory studies tailored to the enhanced recovery options of interest would be completed during the detailed enhanced recovery design process. 8. What is the economic value of gathering additional information? The answer to this at the proof of concept stage is qualitative. The value of discarding an economic development scheme is offset by the value of proceeding with a detailed development plan with low quality data. A Value of Information analysis can be done during the detailed enhanced recovery design process.
15.4.2
Fluid Characterization Program
The following procedure was developed to use the available fluid data with CMG Winprop software to characterize the Dunvegan C fluid; 1. Molecular weight and specific gravity of the oil sample are required for oil characterization. The D86 analysis on hand reported specific gravity, but not molecular weight. Molecular weight of the oil sample was estimated from the D86 distillation by the following procedure; a. A linear curve fit of the D86 distillation was generated for the purpose of extrapolating the D86 data to 100% recovery. This curve fit is given in Figure 3. b. The boiling point range of the distillation was cut into pseudo-components at 25 °C (K) boiling point increments for the purpose of estimating
258
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
pseudo-component properties. Increments were reduced near the end points. c. Specific gravity associated with each of the pseudocomponents was estimated using the Watson equation with the constant Watson factor of 12 given in the D86 analysis: Kw = 1.2164
zb
y Kw = Watson factor Tb = normal boiling point y = specific gravity =p
a
(K) /999.024
d. The boiling point and specific gravity of each pseudo-component were entered into Winprop, and an estimated molecular weight for each pseudocomponent returned. e. The molecular weight of the total oil was then estimated based on the molecular weight of each of the pseudo-components. Results of the above calculations are provided in Table 5.
Figure 3. 07-20-071-04 W6M crude oil sample ASTM D-86 distillation curve fit.
op
184.82
198.5
234.5
279.5
324.5
369.5
414.5
459.5
504.5
549.5
594.5
°C
84.9
92.5
112.5
137.5
162.5
187.5
212.5
237.5
262.5
287.5
312.5
Tbi
0.056818
0.056818
0.056818
0.056818
0.056818
0.056818
0.056818
0.056818
0.056818
0.034091
0
fraction
ÄV
0.848089
0.835845
0.823232
0.81022
0.796776
0.782863
0.768436
0.753447
0.737836
0.724854
S.G.
235
217
199
182
166
151
136
122
108
98
MW
0.048187
0.047491
0.046775
0.046035
0.045271
0.044481
0.043661
0.042809
0.041922
0.024711
A S.G. 3
0.20485
0.21864
0.234819
0.252694
0.272453
0.294288
0.320724
0.350555
0.387792
0.251906
kmol/m
Ap/M
0.050476
0.053873
0.05786
0.062264
0.067133
0.072513
0.079027
0.086378
0.095553
0.06207
0
mol frac
X.
Pseudocomponent MW from Winprop using Riazi-Daubert Physical Properties Correlation
0.687148
0.636673
0.582799
0.524939
0.462675
0.395542
0.323028
0.244001
0.157623
0.06207
0
mol frac
ixi
Calculated Cuts at 25 K intervals based on Curve Fit T=440V+85, Watson Characterization Factor = 12
Table 5. 07-20-071-04 W6M Crude oil sample estimated molecular weight.
(Continued)
0.312852
0.363327
0.417201
0.475061
0.537325
0.604458
0.676972
0.755999
0.842377
0.93793
1
mol frac
residue
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
684.5
729.5
774.5
819.5
864.5
909.5
954.5
362.5
387.5
412.5
437.5
462.5
487.5
512.5
0.871568
0.882847
0.893846
0.90458
0.915066
0.925316
0.935344
0.056818
0.056818
0.056818
0.056818
0.056818
0.056818
0.056818
1
0.859988
0.052575
382
206.7233
Bulk MW, kg/kmol
Molar Fraction Balance
4.058299
0.839785
0.035651
0.999976
0.032222
0.144684
0.051992
359
0.037543
0.039688
0.042
0.13077
0.152363
0.051397
337
0.053145
0.16107
0.050787
315
406
0.170452
0.050162
294
0.04449
0.033879
0.180557
0.049521
274
0.047355
mol frac
i
X.
0.137496
0.192186
kmol/m 3
Ap/M
0.048863
A S.G.
254
MW
Molar Balance:
Material Balance (Bulk S G. D86 Test = .837):
Volumetric Balance:
639.5
337.5
S.G.
0.056818
fraction
°C
°F
AV
Tbi
Table 5. 07-20-071-04 W6M Crude oil sample estimated molecular weight. (Continued)
0.999976
0.967754
0.933874
0.898224
0.860681
0.820993
0.778993
0.734504
mol frac
XX,
2.4E-05
0.032246
0.066126
0.101776
0.139319
0.179007
0.221007
0.265496
mol frac
residue
1>
M
o
O O
n M Z
M
O H
M
r
M
> o
% O
H
O c¡ m
M
e/5
n
CTs O
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
261
2. Molecular weight and specific gravity of the crude oil sample were then entered into Winprop. Winprop then generated pseudo-components and the composition of the crude oil based on these pseudo-components. 3. The separator gas analysis from the 07-20 well was then entered into Winprop as a secondary fluid. 4. Using the Winprop Saturation Pressure calculation block, the crude oil and gas were then recombined, and the molar fraction of crude oil (primary fluid) in the mixture adjusted until the saturation pressure matched the initial reservoir pressure. Winprop predicts that with a mix of .68 mol fraction oil and .32 gas, the reservoir liquid will be saturated at a pressure of 8211 kPaa. If there is a target GOR that needs to be achieved, the molar fraction of oil and gas can be calculated by hand (44.5 M3 / m 3 * 1 kmol / 23.7 M3 = 1.88 kmol gas/m3 oil, 840 kg / m 3 oil * 1 kmol / 206 kg = 4.08 kmol oil / m 3 oil). At standard conditions, Winprop reports about 1% light ends remaining in the oil. This is consistent with the D86 reported 1% loss of sample during the distillation. It was likely boiled up, but the lab could not condense and recover it. 5. Since differential liberation data for the reservoir fluid was not available, 'live' reservoir fluid estimates of viscosity, gas-oil ratio, and specific gravity (3) were entered into the Winprop Differential Liberation calculation block. The Winprop Regression utility was then used to adjust properties of the pseudo-components to fit the differential liberation data. 6. Finally, for the IMEX reservoir simulation a Black Oil PVT Data block was added to export the fluid properties to IMEX. For the GEM compositional reservoir simulation, the CMG GEM EOS Model block was added. Properties of the reservoir fluid characterized in Winprop, compared to estimates from Slider, are provided in Figures 4-5. The estimates from Slider are shown as "Experiment" discrete data points. For the purpose of this project, the match on GOR, viscosity, and density is acceptable, especially considering that the match is to estimated properties in the literature. If laboratory differential
262
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
liberation results were available, additional work may be warranted to get a closer match. A complete program for laboratory analysis of reservoir fluids must be conducted as part of the detailed enhanced recovery process design.
Figure 4. Dunvegan C reservoir oil gor and bo vs estimates from slider.
Figure 5. Dunvegan C reservoir oil viscosity vs estimates from slider.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
263
15.4.3 Solubility of Acid Gas Mixtures in the Dunvegan C Oil Solubility of acid gas mixtures with compositions ranging from 100 mol% C 0 2 to 80 mol% H2S, 20 mol% C 0 2 in the live Dunvegan C Oil were studied. The study is not a textbook case for two reasons; 1. The pool is not pressure depleted. As a result, the oil contains significant methane and other light solution gas components. When acid gas comes into contact with the oil, the acid gas components replace methane in the oil, resulting in an acid gas rich oil phase, and a methane gas phase. This acts to increase minimum miscibility pressures. The study also indicates that in this case there is an increase in the oil phase density, rather than the C 0 2 swelling effect normally expected. 2. The average conditions of the pool are right around the critical point for C0 2 . Pool conditions are -7600 kPag at 34 °C. Critical point for C 0 2 is 7374 kPaa @ 31 °C. Critical point for H2S is 8963 kPaa @ 100 °C. In all cases, core flood studies are recommended to confirm solubility and miscibility predictions, and understand their effects on phase mobility and residual saturations. The somewhat unusual conditions under which enhanced oil recovery is being considered for this pool, while presenting opportunity, also further underline the need to conduct laboratory work to confirm predictions.
15.5 Material Balance Material balance calculations were completed prior to reservoir simulation to estimate fluids in place, initial condition of the reservoir oil (under-saturated or saturated), and drive mechanisms for the pool (gas cap, depletion, and water drive). According to information obtained from the ERCB, geological interpretation of the pool and the drive mechanism were among many subjects of discussion during the application process for Good Production Practice (GPP). Understanding the behavior of water in the pool proved to be a challenge throughout the project. Pool pressure data, acquired by acoustic well sounder survey,
264
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
indicate pressure depletion in the pool, leading to the conclusion that water drive is not significant. Well logs indicate high water saturation through the producing interval, but no water-oil contact, leading to the conclusion that the water present is immobile. This conflicts with reported water production, although the water-oil ratio in the down dip well 03-20 is trending downwards, with no clear trends in the other wells. For the purpose of material balance, the pool was assumed to be stratigraphically trapped with a down dip water leg providing insignificant pressure support. Pre-simulation material balance calculations are provided in Table 6. The calculations are in reasonable agreement with pool data, with the exception of pool pressure. Due to the quality of the pool pressure measurements, and the GPP information discussed above, the initial material balance calculations placed more weight on log data and geological interpretation of the pool, and less on the pool pressure data. The pre-simulation material balance indicated that the pool was initially saturated with a gas phase present. The material balance was revisited after the history match. Post simulation material balance calculations are also provided in Table 6. The history match indicated 56%+ water saturation versus 35% reported on the Reserves Summary and 46% average from well logs. The history match also indicated 423.5 E3M3 Original Oil in Place versus 789 E3M3 reported on the Reserves Summary. The post simulation material balance calculations are in better agreement with pool pressure data measurements, and are also in agreement with the history match results for the 56% water saturation case. The conclusion from this work is that the pool contains lower oil reserves than what is on record with the ERCB.
15.6 Geological Model Data available for the geological model consisted of simple geological isopach (thickness) and formation top maps of the pool, and the wireline logs used to generate the geological maps. A summary of the data sources used for this project is given in Table 7. Geological isopach and formation top maps for the Dunvegan C pool were digitized and imported into CMG Builder. Porosity and connate water saturation from logs were entered into CMG Builder and CMOST as initial estimates.
44.5 0.0114 1.11 1
M3 M3 M3/M3 m 3 /M3 m 3 /M3 m 3 /M3
Np
Wp
Rso
Bg
Bo
Bw
Swi
We
cf
oil produced
water produced
gas oil ratio
gas formation volume factor
oil formation volume factor
water formation volume factor
initial water saturation
water influx (reservoir bbl)
formation compressibility
0 5.80E-04
m3 MPa"1
0.46
0
0
0
E3M3
Gp
gas produced
8.2
MPa
Pr
Initial Condition
reservoir pressure
Parameter/ Case
Material Balance Input
0
.35?
-
-
-
-
4,181
15,489
1,722
7.0
Pool Data on Record
Grande Prairie Dunvegan C Material Balance Calculations
Table 6. Material balance calculations.
0
0.46
1
1.1
0.0126
40.7
4,181
15,489
1,722
7.6
Pre Simulation
0
(Continued)
0.56
1
1.09
0.0138
36.9
4,181
15,489
1,722
7.0
Post Simulation
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
30,177
-
-
M3/M3 m3 m 3 /m 3 m3 m3
Bt
Rp
F
m
Eo
Eg
E.
oil + solution gas formation volume factor
cumulative gas oil ratio
net total production (includes (Wp-We))
gas to oil in place ratio
oil expansion
gas expansion
water and formation expansion f,w
0 042
-
M3
N
initial oil in place
m3
m 3 /M3 -
Calculations
7.89E+05
1.3 38E+06 7.89E+05
138E+06
886
3,905
34,966
111
1.148
7.89E+05
m3
Pore Volume and OOIP Estimates
Vp 1.69E+06
Pre Simulation
pore volume
4.79E-04
Pool Data on Record
Cw
MPa"1
Initial Condition
formation water compressibility
Parameter/ Case
Table 6. Material balance calculations. (Continued)
1,052
437
35,459
0.005
36,949
111
1.196
4.13E+05
1.05E+06
Post Simulation
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Sgi
Soi
Swi
Sg
So
Sw
initial oil saturation
initial water saturation
current gas saturation
current oil saturation
currentwater satu ration -
-
fraction fraction
-
fraction
-
-
fraction fraction
-
fraction
Err ->adjust Vp or N until Error=0
initial gas saturation
F- (Eo+Eg+Efw)=0 solver
0.458
0.504
0.039
0.460
0.518
0.022
0
0.556
0.414
0.029
0.560
0.438
0.002
0
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
268
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Omitted from Table 7 are data sources such as correlations from the literature, and generic field or basin data that can be used in the absence of primary or secondary sources. It is strongly recommended that data be collected from primary or secondary sources before conducting a reservoir simulation.
Table 7. Geological model data sources. Geological Parameter
Primary Data Source
Secondary Data Source
Source for this Project
Formation Bulk Volume ( top, thickness, closure)
Well Log Interpretation
Seismic interpretation Well test interpretation Geological interpretation
Geological interpretation
Porosity
Routine Core Analysis
Well Log Interpretation
CMOST optimized with initial estimate from logs
Permeability
Well Test Interpretation
Routine and Special Core Analysis
CMOST optimized
Residual Fluid Saturations (water, oil, liquid, gas)
Special Core Analysis
Well Log Interpretation
CMOST optimized with initial Swc estimate from logs
Relative Permeability curve exponents and end points
Special Core Analysis
Well Skin or near-wellbore permeability
Well Test Interpretation
—
CMOST optimized
Huid Contacts
Well Log Interpretation
Material Balance
CMOST optimized
CMOST optimized
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
15.7
269
Geological Uncertainty
Given the data sources available for this project, there is considerable geological uncertainty associated with the Dunvegan C Pool. This is not unusual for pools of this size and vintage. Computer Modeling Group's CMOST optimization software was used to complete a computer assisted history match using the IMEX black oil reservoir simulator to complete individual simulations. Geological uncertainties in the pool were represented by thirty individual parameters input into CMOST. CMOST cases were completed with connate water saturation at 36%, 46%, and 56%, and the results compared. The history match operation was successful with global objective function error of less than 5% in all optimized cases.
15.7.1 Formation Bulk Volume Uncertainty in formation geometry was not directly accounted for in this project because only one geological interpretation of the pool geometry was available. In many cases, including this one, formation geometry is a significant uncertainty. The history match cases resulted in lower porosity than indicated by well log interpretation. One possible conclusion from this is that the reservoir simulation is responding to a 'pool compressibility', and that the pool is geometrically smaller than indicated by geological mapping.
15.7.2 Porosity Porosity from well log interpretation is 18%. This is based on sandstone matrix density of 2.65 g/cm 3 . A porosity range of 10-24% was used in the history match optimization. The history match returned optimized porosities between 12-14%. Density-neutron well log cross plot interpretation indicates a shaley matrix, with matrix density is between 2.54-2.57 g/cm 3 . Routine and special core analysis is recommended to correct well log data and confirm interpretation.
15.7.3
Permeability
No permeability data was available for this pool. Estimates of permeability in the i, j , and k directions is important data for
270
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
designing optimum enhanced oil recovery programs, and is normally obtained from well test interpretation and routine or special core analysis. For this project, a 0.01-100 md range of permeability values in each of the i, j , and k directions was input into the optimization software. The optimized history match cases returned permeability in the range of 5-50 md in each of the directions, with k (vertical) permeability being unusually high.
15.7.4
Residual (Immobile) Fluid Saturations
Residual fluid saturations are obtained from special core analysis relative permeability results for water-oil and liquid-gas fluid regimes, with attention paid to match the fluid movement history in the core with that anticipated in the field to capture hysteresis effects. For this project, the only residual fluid saturation data available for this project was connate water saturation from well log interpretation and geological interpretation. Pool average connate water saturation in the range of 35-52.5% has been reported by various sources. Connate water saturation calculated from well logs depends on the Archie equation constants used, and therefore is subject to interpretation. Using standard Archie constants of a=l, m=2, and n=2, calculated connate water saturations in the range of 65-68% were calculated. Given the above information, optimized history matches were completed for connate water saturations in the range of 36-56%.
15.7.5
Relative Permeability Curve Parameters
Relative permeability curve exponents and end points are best obtained from special core analysis of water-oil and liquid-gas fluid regimes, with attention paid to match the fluid movement history in the core with that anticipated in the field to capture hysteresis effects. No relative permeability data from core analysis were available for this project. Initial estimates of relative permeability parameters were made by using rule of thumb parameters available in the literature, and production data for each well in the pool. Ranges of relative permeability end points and exponents were input into the optimization software. CMOST optimized end points and exponents are given in Table 8.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
271
Table 8. Fractional flow study inputs versus CMOST results. Dunvegan C Fractional Flow Study Inputs Input Variable
CMOST Results for Comparison
Description
Input
Sw36
Sw46
Sw56
krw*
end point water rel perm
0.5
0.7
0.8
0.7
kro*
end point oil rel perm
1
0.8
0.8
0.6
Swi
irreducible water
0.46
0.36
0.46
0.56
Sor
residual oil
0.2
0.2
0.2
0.2
eo
oil exponent
3.5
4
4
2
ew
water exponent
2
1.5
1.5
1.5
krl*
end point liquid rel perm
1
0.8
0.8
0.6
krg*
end point gas rel perm
1
0.4
0.5
0.4
Sli
irreducible liquid
0.6
0.46
0.56
0.66
Sgr
residual gas
0.04
0.01
0.01
0.01
eg
gas exponent
3
1.5
1
1
el
liquid exponent
2
2
4
4
A study of three phase fractional flow was completed to better understand the expected range of relative permeability end points and exponents, and the fluid saturations required to allow for produced water flow. Stone's method with Aziz and Settari corrections, along with rule of thumb exponent and end point data were used to predict three phase fractional flow (4) (5) (6). Historical fractional flow data from the pool, was obtained by converting production data to fractional flow data at low pressure (near wellbore) and high pressure (reservoir) conditions. The predicted and historical fractional flow data were then compared, and water saturation modified until a match obtained.
272
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
A match at 55% water saturation was obtained for historical fractional water production, with relative permeability exponents in the 2-3 range, end points in the 0.5 to 1 range, Swc (immobile) =0.46, Sor=.2, Srg=.04. Water saturations above 55% were not investigated. Water saturation below 55% resulted in insufficient water flow. The procedure used for this study merits additional review, but the concept of translating relative permeability relationships and actual production into fractional flow data, and comparing the results, is valuable for understanding fluid flow behavior in the reservoir, and predicting the impact of various operations on oil flow and recovery from the pool. Comparison of the study results with the optimized history match results is provided in Table 8. The only value from the study was in knowing the historical fractional flow of each phase at reservoir conditions. Three phase flow through porous media is poorly understood, difficult to study and difficult to predict, but important to consider when evaluating enhanced oil recovery projects. The same can be said for the impact of fluids present on residual saturations.
15.7.6 Fluid Contacts An initial water-oil contact of 335 mSS, from geological interpretation of well logs, was used in the reservoir simulations. Uncertainty in the water-oil contact was not investigated in this project. Geological interpretation of the pool, pressure data, and water production trends from the pool do not indicate an active aquifer. Water production from the pool is not significant in terms of volume of reservoir fluids removed. Although there was no direct evidence of a gas-oil contact from well log interpretation, production data and pool material balance calculations indicated that a small gas cap may be present in the pool. Based on pool geometry and material balance calculations, an initial estimated gas-oil contact of 310 mSS was used, and range between 308-314 mSS allowed in the history match optimization. This translated to initial gas saturations of 7-11 % in the optimized history match cases, at optimized contacts of 310-312 mSS.
15.8
History Match
Computer Modeling Group's CMOST optimization software was used to complete the history match.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
273
Geological uncertainties in the pool were represented by thirty individual parameters input into CMOST. For ease of execution, connate water saturation was held constant in each of the CMOST optimization cases. CMOST cases were completed with connate water saturation at 36%, 46%, and 56%, and the results compared. In each of the cases, CMOST determined the optimum value of 29 parameters that would produce a reservoir simulation that minimized the error between the reservoir simulation and the production history. In CMOST, the production history is represented by a user defined 'global objective function'. For this history match, the global objective function was defined as a weighted average of cumulative gas, oil, and water production for each of the wells, and estimated bottom hole pressures for 03-20, 10-20, and 13-20 wells on artificial lift. Note that IMEX primarily matched oil production in all cases. A summary of the results is provided in Table 9. Charts summarizing pool production history data with the optimized history match cases are provided in Figures 6-8. Observed trends and indications from the history match are; • Global error in the history match decreased with increasing water saturation. This was due to decreasing errors in the cumulative gas and cumulative water production. Match to oil production in all cases was generally excellent. There was no trend amongst the cases in error between the simulation and the constructed bottom hole pressure data. • Porosity is much lower than reported from well log interpretation. • Original Oil in Place in the Sw46 and Sw56 cases is significantly lower than the OOIP on public record for the pool. This is supported by pressure data on file for the pool. • All history match cases indicate the presence of a gas cap, with gas oil contact at 310-312 mSS, and initial gas saturations ranging from 11% in the Sw36 case to 7% in the Sw56 case. • Permeability results in the Sw36 and Sw46 cases were unusual, with k direction (vertical) permeability being higher than i and j direction permeability. Permeability in the Sw56 case met conventional expectations,
2.14E-05 1.20E-05
0.096769
2.14E-05
1.20E-05
21.669
12.618
4.012
20.055
%
%
%
%
%
%
%
W07_20_Oil
W10_20_oil
W13_20_Oil
W03_20_Gas
W07_20_Gas
W10_20_Gas
W13_20_Gas 19.958
7.2683
23.932
8.9173
6.93E-05
1.66E-05
1.66E-05
%
W03_20_Oil
4.09
4.93
1925
Sw_46
%
1280
Sw_36
GlobalObj.
Selected Run
Units Sw_56
4.8637
4.2855
15.444
14.344
1.20E05
2.14E-05
6.93E-05
1.66E-05
3.12
1889
Mar29SW46_PASS01 .cmr
Comparison of optimized base case reservoir simulations
Parameter
Mar29SW36_PASSl 3.cmr
DNVG C Pool EOR Study
2 weighting
4 weighting
2 weighting
4 weighting
9 weighting
18 weighting
22 weighting
27 weighting
Volumetric weighting applied to production
from above .cmr files
Notes
Mar29SW56JPASS02.cmr
CMOST Results Files:
ENCH-699
Table 9. History match results. C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
%
%
%
%
%
fraction
-
W10_20_Water
W13_20_Water
W03_20JPress
W10_20_Press
W13_20_Press
Swc
AQ03RR
GasOilC
AQ13RR
AQ10RR
mSS
-
-
-
%
W07_20_Water
AQ07RR
%
W03^20_Water
14.45 13.455 43.29
14.218
8.0897
46.057
7
7
6 2 310
5
2
312
5
0.45
0.36
5
54.093
37.94
23.913
311
2.5
5
7
7
0.56
30.887
42.636
7.0923
55.289
43.529
4.9649
9.3174
13.409 9.7239
2.7389
3.4641
16.151
gas-oil contact.
(Continued)
aquifer strength indicator for well 13 (located mid-updip edge fmn)
aquifer strength indicator for well 10 (located mid-updip center fmn)
aquifer strength indicator for well 07 (located mid fmn)
aquifer strength indicator for well 03 (located low in fmn)
Reported Sw=.46from logs
1 weighting
1 weighting
1 weighting
2 weighting
4 weighting
5 weighting
7 weighting ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POO
md
fraction
fraction
fraction
permk
porosity
Sgcon
Slcon
Soirw
fraction
-
md
permj
SLT
md
-
-
-
Units
permi
krwiro
krocw
krgcl
Parameter
0.1
0.05
g=l, o=4
g=1.5,o=2
0.02
0.05
0.25
g=L o=4
0.1
0.12
0.12
0.05
10
50
0.1
0.02
0.14
50
10
50
5
10
5 10
end point relative water permeability in w-0 system
0.7
0.8
0.7
residual oil saturation in w-o system
relative permeability curve sets for 1-g system (exponents)
residual oil saturation in 1-g system [total Slr=Swc+Slcon)
residual gas saturation in 1-g system
reported porosity =0.18 from logs, .1-.24 range in CMOST
reservoir permeability in k direction
reservoir permeability in J direction
reservoir permeability in direction
end point relative oil permeability in w-o system
0.6
0.8
0.4
end point relative gas permeability in 1-g system
Notes
0.8
Sw_56 0.4
Sw_46 0.5
Sw_36
Table 9. History match results. (Continued) C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
md
md
md
md
md
md
md
md
w03fracpermk
w03fracpermj
w07fracpermi
w07fracpermk
w07fracpermj
wlOfracpermi
wlOfracpermk
-
w03fracpermi
SWT
0.01
100
1.00E+05
1.00E-02
1E5
1E5
1.00E+02
1.00E-01
w=l .5,o=4
0.01
1
(Continued)
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
1E2
1E6
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
relative permeability curve sets for o-w system (exponents)
simulating near wellbore stimulation or damage
0.001
1000
1E5
1E4
1E-2
w= 1.5,o=2
1E3
1000
0.01
1E6
1E3
1.00E+05
1E0
w=1.5,o=4
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POO
1E-6
1E9
700.5
45.9
529.2
md
md
E3M3
E6M3
E3M3
wl3fracpermk
wl3fracpermj
OOIP
OGIP
OWIP
E3m3
E3m3
Initial Oil
Solrn + Free Gas
523.26
777.555
0.1
md
wl3fracpermi
Reservoir Volume Calculations
100
Sw_36
md
Units
wlOfracpermj
Parameter
Table 9. History match results. (Continued)
705.5
579.6
379.62
302.1
470.085
26.5
33.3
560.217
423.5
504.7
Calculated at Reservoir conditions, Boi=l.ll, Bg=.0114, Bw=l
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
0.001
IE-6 1.00E+09
simulating near wellbore stimulation or damage
1
0.01
1.00E+09
simulating near wellbore stimulation or damage
Notes
1E1
Sw_56
1EO
Sw_46
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
OGIP-solution gas. Rsoi=44.5 Calculated, OIP/(OIP+WIP+Gas Cap Gas) Calculated, WIP/(OIP+WIP+Gas Cap Gas) Calculated, GAS Cap Gas/ (OIP+WIP+Gas Cap Gas)
87.3 0.37 0.56 0.07
123.6 0.44 0.46 0.10
167.9
0.53
0.36
0.11
So
Sw
Sg
E3m3
Gas Cap Gas
705.5
579.6
529.2
E3m3
Initial Water
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POO
280
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. History match pool cumulative oil.
Figure 7. History match pool cumulative gas.
with vertical permeability being 1/10 of the maximum horizontal permeability. Horizontal directional permeability was predicted in the Sw36 and Sw56 cases, important for designing enhanced oil recovery programs.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
281
Figure 8. History match pool cumulative water.
• Near wellbore permeability parameters were included in the history match to simulate stimulation or damage from drilling and completion operations. There was extreme variation in the order of magnitude of near wellbore permeability in the optimized cases. This method of modeling near wellbore reservoir properties appears to be flawed. A better approach may be to use skin factor from well test interpretation if available, or estimate skin based on the completion program. • The strength of the aquifers added to each well in order to provide free water production generally was highest with wells located down dip in the pool, with strength decreasing in the u p dip wells, and thinner pay wells. This method for accounting for water production is also flawed. Water production is likely from water mobilized during completion operations. A manual history match exercise was completed prior to using CMOST to complete the optimized history match. Observations from this exercise were; • Gas production estimates from IMEX were relatively easily tuned to the field history by adjusting gas oil
282
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
contact, gas relative permeability exponent, oil relative permeability end point, and i direction permeability in perforated grid blocks. • Matching the water production (WOR=0.2 to 0.3 for up dip and down dip wells) was particularly difficult. Well logs and geological interpretation do not indicate a water oil contact in the pool, nor do they indicate any high permeability connection to a water source. No mobile water bearing zone adjacent to D N V G C . Matrix expansion is insufficient to mobilize the produced water volume. Cumulative Water-Oil ratio is declining for each well. No indication of water bank arrival at up dip wells. Attaching a Carter-Tracy limited extent aquifer to each well provides the closest match to field history. ( Fetkovitch, Carter Tracy Infinite, and "Old" aquifer models were also investigated) Strength of the aquifer for all wells is set at 10, with the exception of 13-20 set at 1.5. Note that pay height at 13-20 is 1 m compared to the other wells at 5-9 m. Note Completions are generally 20 tonne, 45-65 m 3 gelled oil fracs, water production on all wells is present immediately, and 13-20 (one of the up dip wells) was re-completed and had highest water production immediately after recompletion. No production evidence of a 'flood front' moving into the wells. If a lower layer is being swept out (waterflooded), it would be indicated by a distinctive flood front arrival.
15.9 Black Oil to Compositional Model Conversion The optimized history match cases were converted from IMEX to GEM in order to simulate an acid gas flood on the pool. Comparison of field history with history match cases in IMEX, and the cases converted to GEM, is provided in Figures 9-17. A comparison of IMEX and GEM parameters for the various cases is provided in Table 10. The GEM simulations produced an unacceptable match to IMEX and field historical gas and water production. Predicted well bottom hole pressures in the GEM simulations were also in poor agreement with the IMEX simulations. Although the conversion process
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 9. Sw36 IMEX-GEM comparison pool cumulative oil.
Figure 10. Sw36 IMEX-GEM comparison pool cumulative gas.
283
284
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 11. Sw36 IMEX-GEM comparison pool cumulative water.
Figure 12. Sw46 IMEX-GEM comparison pool cumulative oil.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 13. Sw46 IMEX-GEM comparison pool cumulative gas.
Figure 14. Sw46 IMEX-GEM comparison pool cumulative water.
285
286
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 15. Sw56 IMEX-GEM comparison pool cumulative oil.
Figure 16. Sw56 IMEX-GEM comparison pool cumulative gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
287
Figure 17. Sw56 IMEX-GEM comparison pool cumulative water.
was unsuccessful here, the GEM models were used to simulate various acid gas floods for proof of concept purposes. It is also worth noting that the acid gas floods were simulated with horizontal injection and production wells, and all existing wells shut in to control flood the flood front location and defer breakthrough of injected acid gas to the producing wells. The following factors likely contributed to the unsuccessful conversion; 1. There was some difficulty in getting the GEM simulations to run successfully. Through experimentation with the simulations it was found that the simulations would run if variation of permeability in the reservoir was restricted. Near wellbore permeability for each of the wells was restricted to 1 to 100 md. Refer to Table 10. 2. Use of a large number of components in the GEM fluid model likely contributed to the instability of the simulations. Some components, such as iso-butane and normal-butane, have very similar properties and would be difficult for the simulator to resolve. One important lesson from this exercise is that lumping of such components will give better results. The
-
md
md
md
fraction
fraction
krwiro
permi
permj
permk
porosity
Sgcon
mss
GasOilC
-
-
AQ13RR
krocw
-
AQ10RR
-
-
AQ07RR
krgcl
-
fraction
Swc
AQ03RR
Units
Parameter 0.46 7
0.46 7
0.36 7
0.8
0.8 0.8 10
0.4 0.8 0.7 5
0.01
0.14
50
10
5
0.7
0.8
0.02
0.14
1
10
0.5
0.5
312
312
0.4
310
310
2
2
0.05
0.12
0.12 0.01
1
10
10
0.8
2
50
10
2
6
5
5 6
5 5
Sw_46 GEM
Sw_46 IMEX
Sw_36 GEM
5
5
7
0.36
Sw_36 IMEX
Table 10. Comparison of IMEX and GEM parameters.
0.01
0.12
10
50
5
0.7
0.6
0.4
311
2.5
5
7
7
0.56
Sw56 IMEX
0.02
0.12
1
50
5
0.7
0.6
0.4
311
2.5
5
7
7
0.56
Sw_56 GEM
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
fraction
-
fraction
-
md
md
md
md
md
md
md
md
md
md
md
md
SI con
SLT
Soirw
SWT
w03fracpermi
w03fracpermk
w03fracpermj
w07fracpermi
w07fracpermk
w07fracpermj
wlOfracpermi
wlOfracpermk
wlOfracpermj
wBfracpermi
wl3fracpermk
wl3fracpermj
1E9
IE-6
0.1
100
0.01
100
1.00E+05
1.00E-02
1E5
1E5
1.00E+02
l.OOE-01
w=1.5, o=4
0.2
g=1.5,o=2
0.46
10
1
100
10
1
100
1.00E+09
IE-6
0.01
1E0
1
1E6
1000
0.01
1 10
IE6
1E3
1.00E+05
1E0
10
1
100
10
1
100
10
1
100
10
1
100
w=1.5,o=4
0.05
0.2 w=1.5, o=4
g=l, o=4
0.51
g=l,o=4
0.56
100
10
1
100
w=1.5, o=4
0.1
g=1.5,o=2
0.46
1.00E+09
0.001
1
1E1
0.01
1E2
1E3
0.001
1000
50
1
100
50
1
100
50
1
100
50
1
1E4 1E5
100
w=1.5, o=2
0.25
g=l, o=4
0.66
1E-2
w=1.5,o=2
0.2
g=L o=4
0.66
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
290
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
geological model arrived at through IMEX simulations had to be compromised due to the onerous fluid system that the user imposed on the GEM simulations. Well deliverability, and near wellbore reservoir properties need to be more accurately represented in the simulations. An approach of modeling wellbore deliverability by skin, rather than permeability variations, may have resulted in simple stable GEM models. Extreme values and order of magnitude variations in near wellbore permeability likely contributed to GEM model instability.
15.10
Recovery Alternatives
A number of recovery alternatives were studied in this project, but the alternative of interest is a miscible acid gas flood. This alternative is conceptually attractive for a number of reasons; 1. Theoretical residual oil saturation in a miscible flood is 0, resulting in complete oil displacement and recovery from the reservoir volume swept by the injected fluid. The challenge is achieving acceptable sweep efficiency. 2. Environmental benefits of sequestering acid gas, including pure C0 2 , can be coupled with additional oil recovery. This means that the cost of acid gas or C 0 2 capture and storage would partially or fully paid for by the additional revenue from recovering additional oil from the reservoir. 3. There is significant potential for additional recovery from conventional oil pools. In Alberta, an estimated 7.7 Billion m 3 of oil remains unrecovered in conventional pools, 74% of the Original Oil in Place (7). A generic program was designed, consisting of one horizontal injection well placed up dip in the pool, and two horizontal production wells placed down dip in the pool. The program was then operated under the following recovery alternatives; 1. Acid gas flood, miscible at injection pressure down to 8 MPa, producing wells at 4 MPa. Identical operating
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
2.
3.
4.
5.
291
conditions were used and applied to the 36%, 46%, and 56% water saturation cases. While not comprehensive, this illustrates how the impact geological uncertainty on the project could be evaluated by optimizing a program based on one geological interpretation, then applying that program to other possible geological realities to determine what the technical and economic impact of implementing a non ideal program would be. These discrete cases, with probabilities attached to them, can then be used in a Decision Analysis process to determine the Expected Monetary Value of the project. Production from the horizontal wells only, with no flood implemented. This scenario was run to understand the incremental difference between simply investing in additional horizontal wells versus implementing a flood using the wells. Identical operating conditions were used and applied to the 36%, 46%, and 56% water saturation cases. Production from existing wells only, under 'pump off operation with wells at 1000 kPa bottom hole pressure. This scenario was run to understand the incremental difference between simply lowering the pressure in the existing wells, and producing without Maximum Rate Limitation (MRL) versus investing in additional horizontal wells, also producing without the MRL. Identical operating conditions were used and applied to the 36%, 46%, and 56% water saturation cases. Production from existing wells only, under the initial ERCB Maximum Rate Limitation (MRL) of 8 m 3 / d and corresponding Gas Oil Ratio of 100 m 3 /m 3 . C 0 2 flood, miscibility developed by increasing pool pressure above 13 MPa, producing wells at 4 MPa. A stand alone single case, using 46% water saturation and 18% porosity, was run to demonstrate the potential of coupling C 0 2 disposal with Enhanced Oil Recovery.
Event logs, describing pool operation under each of the above recovery alternatives, are provided in Tables 11-15. Screen shots of the pool are provided in Figures 18-20. Production and injection forecasts for the pool, and individual wells, are provided in Figures 21-42.
292
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 11. Acid gas flood event log. Program: Acid Gas Flood Event Description
Well
Date l-Oct-10
Hz Injector-1
begin drill, complete, tie in operations
l-Oct-10
Hz Producer-2
begin drill, complete, tie in operations
1-Jan-ll
Hz Producer-2
Produce with Constraint Min BHP=4000 kPa
1-Jan-ll
13-20
Shut In
1-Jan-ll
10-20
Shut In
1-Jan-ll
07-20
Shut In
1-Jan-ll
03-20
Shut In
1-Jan-ll
Hz Injector-1
Inject with Constraints Max Rate=30 E3M3/D, Max BHP=15 MPa
l-Oct-22
Hz Producer-3
begin drill, complete, tie in operations
l-Oct-22
Hz Producer-2
Prepare to convert the well form a Producer to an Injector
l-Jan-23
Hz Producer-3
Produce with Constraint Min BHP=4000 kPa
l-Jan-23
Hz Injector-2
Inject with Constraints Max Rate=30 E3M3/D, Max BHP=15 MPa
Table 12. Horizontal wells event log. Program: Horizontal Producers Date
Well
Event Description
l-Sep-09
13-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
10-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
07-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
03-20
Produce with Constraint Min BHP=1000 kPa
l-Oct-10
Hz Producer-2
begin drill, complete, tie in operations
1-Jan-ll
Hz Producer-2
Produce with Constraint Min BHP=4000 kPa, Max oil rate=40 m 3 /d
l-Oct-22
Hz Producer-3
begin drill, complete, tie in operations
l-Jan-23
Hz Producer-3
Produce with Constraint Min BHP=4000 kPa, Max oil rate=40 m 3 /d
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
293
Table 13. Pump off event log. Program : Pump Off Date
Well
Event Description
l-Sep-09
13-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
10-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
07-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
03-20
Produce with Constraint Min BHP=1000 kPa
Table 14. Maximum Rate Limitation (MRL) event log. Program : Status Quo Date
Well
Event Description
l-Sep-09
13-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3
l-Sep-09
10-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3
l-Sep-09
07-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3 , Min BHP=7000 kPa (no p u m p on this well)
l-Sep-09
03-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3
Table 15. C 0 2 Flood event log. Program : C 0 2 Flood Date
Well
Event Description
l-Oct-10
Hz Injector-1
begin drill, complete, tie in operations
l-Oct-10
Hz Producer-2
begin drill, complete, tie in operations
1-Jan-ll
Hz Producer-2
Produce with Constraint Min BHP=5000 kPa (Continued)
294
C 0 2 SEQUESTRATION A N D RELATED TECHNOLOGIES
Table 15. C 0 2 Flood event log. (Continued) Date
Well
Event Description
1-Jan-ll
13-20
Shut In
1-Jan-ll
10-20
Shut In
1-Jan-ll
07-20
Shut In
1-Jan-ll
03-20
Shut In
1-Jan-ll
Hz Injector-1
Inject with Constraints Max Rate=40 E3M3/D, Max BHP=15 MPa
l-Oct-19
Hz Producer-3
begin drill, complete, tie in operations
l-Jan-20
Hz Producer-2
Shut In
l-Jan-20
Hz Producer-3
Produce with Constraint Min BHP=5000 kPa
Figure 18. Screen shot of the dunvegan C pool in CMG results.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
295
Figure 19. North-South cross section of the pool under C 0 2 flood.
Figure 20. East-West cross section indicating trajectory of horizontal producer 002.
296
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 21. Enhanced oil recovery pool cumulative oil.
Figure 22. Enhanced oil recovery pool cumulative gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 23. Enhanced oil recovery pool cumulative water.
Figure 24. Enhanced oil recovery cumulative gross tonnes acid gas injected.
297
298
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 25. Enhanced oil recovery pool net tonnes acid gas sequestered.
Figure 26. Enhanced oil recovery Mol% C 0 2 in producer 2 gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 27. Enhanced oil recovery Mol% H2S in producer 2 gas.
Figure 28. Enhanced oil recovery k g / d C 0 2 in producer 2 oil.
299
300
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 29. Enhanced oil recovery k g / d H2S in producer 2 oil.
Figure 30. Enhanced oil recovery Mol% C 0 2 in producer 3 gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 31. Enhanced oil recovery Mol% H2S in producer 3 gas.
Figure 32. Enhanced oil recovery kg/d C0 2 in producer 3 oil.
301
302
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 33. Enhanced oil recovery k g / d H2S in producer 3 oil.
Figure 34. Sw36 Primary production pool cumulative oil production.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 35. Sw36 Primary production pool cumulative gas production.
Figure 36. Sw36 Primary production pool cumulative water production.
303
304
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 37. Sw46 Primary production pool cumulative oil production.
Figure 38. Sw46 Primary production pool cumulative gas production.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 39. Sw46 Primary production pool cumulative water production.
Figure 40 Sw56 Primary production pool cumulative oil production.
305
306
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 41. Sw56 Primary production pool cumulative water gas production.
Figure 42. Sw56 Primary production pool cumulative water production.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
307
Recovery alternatives that may be attractive, but were not investigated in this project, include surfactant flood and gas flood. In the case of surfactant flood, incremental oil recovery would be expected due to driving the residual oil saturation down (reduction in interfacial tension) along with high sweep efficiency. Oil viscosity is very attractive here, resulting in high sweep efficiency. In the case of gas flood, incremental oil recovery would be expected due to pressure maintenance in combination with low residual oil saturation in a gas-liquid system. A summary of the technical performance of recovery alternatives is provided in Table 16. The highest oil volumes recovered (248-305 E3M3), and recovery factors (35-74%), were realized by the Acid Gas and C 0 2 Flood programs, followed by horizontal wells (79-160E3M3, 19-23%), pump off without horizontal wells (58-102E3M3, 14-15%), and continued rate restriction operation (58-63E3M3, 9-14%). Refer to the Economics section of this report for economic performance of the recovery alternatives. Original Oil in Place for the C 0 2 flood is larger than the other Sw46 cases because it was run as a separate stand alone case using porosity and water saturation from well log interpretation. In the flood programs, Recovered oil volumes generally do not 'plateau' in the 2011-2030 time periods, indicating that further optimization could be done to accelerate recovery.
15.11
Economics
The economic summary of development cases is provided in Table 17. Cumulative cash flows are shown in Figures 43-45. Calculations are on a full project, go forward basis. The Pump Off cases represents the base case for each geological realization, against which the various development cases can be compared. Results are not incremental to the base case. The pump off cases generally realized the highest Net Present Value ($ 2.2-5.3 MM) of the pool, with the minimum capital investment. The exception is the 36% water saturation (Sw36) case, where drilling horizontal wells resulted in the highest NPV. However, the probability that the Sw36 case is reality is low given this case had the highest error in history match, and is not supported by the current geological evidence.
248
35%
285
E3M3
%
E3 Tonnes
Oil Recovered at 2030
Recovery Factor
Acid Gas Sequestered at 2030
700
E3M3
Sw36
336
74%
374
505
Sw46
285
72%
0
23%
160
700
424
305
Sw36
0
20%
102
505
Sw46
0
19%
0
15%
102
700
424
79
Sw36
Sw56
Horizontal Wells
Sw56
Acid Gas Flood
Original Oil In Place
Case
Table 16. Technical performance of recovery alternatives.
0
0
14%
58
71
14%
424
505
0
9%
63
700
0
12%
60
505
Sw46
Sw36
Sw46 Sw56
MRL
Pump Off
0
14%
58
424
Sw56
co 2
231
36%
255
716
Sw46
Flood
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
>40%
1
1.83
N/A
0
N/A
N/A
10,432
Rate of Return
Payout, years
Profit Investment Ratio (0%)
Profit Investment Ratio (10%)
Royalty Revenue to Alberta, M$ CAD
19,684
1.90
7,528
5,331
Net Present Value (10%), M$CAD
Sw36 Hz Wells
Sw36 Pump Off
Profitability Indicators
35,402
-0.06
0.68
12
8.5%
-1,094
Sw36 AG Flood
5,263
N/A
N/A
0
N/A
3,019
Sw46 Pump Off
Table 17. Economic summary of recovery alternatives.
10,225
0.76
0.60
1 year
>40%
3,011
Sw46 Hz Wells
55,971
0.04
1.22
3,766
N/A
N/A
0
11
7,003
0.24
0.18
2
>40%
N/A
10.6%
Sw56 Hz Wells 945
Sw56 Pump Off 2,173
735
Sw46 AG Flood
44,259
-0.05
0.98
11
9.1%
35,192
0.09
0.70
6
12.6%
1,519
-944
co 2 Flood
Sw56 AG Flood
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
310
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 43. Cumulative cash flow after tax, acid gas and C 0 2 floods.
Figure 44. Cumulative cash flow after tax, horizontal wells.
All acid gas flood economics are marginal. A positive change in oil price, capital cost, royalty framework, or other government policy are required in order for an acid gas flood development to proceed. Comparing royalty received from the base 'Pump Off cases versus potential royalty from the acid gas floods, government has significant negotiation room in the royalty framework as shown by Figure 46.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 45. Cumulative cash flow after tax, pump off.
Figure 46. Pool net present value and royalty revenue comparison.
Calculations were completed using the following parameters; • 2011 project start. • 10% Discount Rate. • No escalation.
311
312
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
• • • • •
$377/M3 ($60/STB) Oil. 40% Royalty. $63/M3 ($10/STB) operating cost for oil production. $ 3/tonne operating cost for acid gas disposal (8). $30/tonne capital cost for acid gas disposal, spend in 2011 (8). This capital cost classified as Class 41 with 25% declining balance deduction with Vi year rule. • $0 revenue / $0 cost for the acid gas or C 0 2 itself (operating and capital costs carried per above). • $3 MM per well drill, complete, and tie in capital cost. This capital cost classified as CDE with 30% declining balance deduction. • 28% Income Tax.
15.12
Economic Uncertainty
The impact of +/-50% changes to oil price, capital cost (Capex), royalty, and operating cost (Opex) on the acid gas flood economics were calculated, and the results reported in Table 18. The price of oil had the largest impact on project NPV, closely followed by capital cost and royalties respectively. Operating cost had the lowest impact on project economics, although this cost is significant to the project.
15.13
Discussion and Learning
15.13.1 Reservoir Fluid Characterization Regarding the oil characterization a case was initially completed where mole fraction, MW, S.G., and Boiling point of the D86 cuts were entered directly into plus fraction splitting, but the result appeared to be a sample with the light and heavy components missing from the distribution. This may have had more to do with how the cuts were generated rather than Winprop. A constant K assumption was used because it was quick and easy, and the D86 analysis did give a characterization factor = 12. This warrants additional study. It would be good to go back and re-visit the characterization method, and do a comparison on a volume basis. A comparison
2.59
0.62
1.22
0.04
55,971
Profit Investment Ratio (0%)
Profit Investment Ratio (10%)
Royalty Revenue to Alberta, M$ CAD
83,957
7
11
Payout, years
19.5%
10.6%
12,519
735
Net Present Value (10%), M$CAD
Rate of Return
$90 Oil
Sw46 AG Flood
Profitability Indicators
0.42
-0.54
27,986
2.13
-0.15
27,986
8
20
16.6%
8,591
-11,049
-2.0%
20% Royalty
$30 Oil
Table 18. Economic sensitivity of acid gas flood.
83,957
-0.35
0.30
17
3.3%
-7,121
60% Royalty
55,971
0.21
1.62
9
13.3%
4,201
$5/STB $1.5/T Opex
55,971
-0.13
0.81
13
7.6%
-2,731
$15/STB $4.5/T Opex
55,971
0.86
3.15
7
23.0%
8,685
$11 MM Capex
55,971
-0.24
0.57
7
5.7%
-7,216
$33 MM Capex
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
314
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 47. Economic sensitivity of acid gas flood.
Figure 48. Mol fraction oil distilled versus temperature.
of experimental vs. characterized data on a molar basis was done, and it appeared that the very light ends and very heavy ends were truncated.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
315
Complete component lumping before any simulation to avoid inconsistencies. Reservoir temperature of 34 °C is approaching hydrate formation temperature of approximately 30 °C for 80% H2S:20% C 0 2 acid gas mixture. Reservoir temperature should be confirmed, and hydrate formation conditions verified by laboratory experiment.
15.13.2
Material Balance
It was very useful in this case to use material balance to understand the relative influence of the various drive mechanisms for this pool. In future studies it would be useful to expand on this and complete calculations for various geological interpretations covering the range of geological uncertainty.
15.13.3
Geological Model
It is strongly recommended that data be collected from primary or secondary sources before conducting a reservoir simulation. It is important to discuss results of calculations with the geologist and geophysicist on the team, acknowledge that there can be more than one depositional interpretation, and get a number of possible geological pay maps to include in the history match exercise. The reason for water production from the wells is not well understood, and attempting to match it by adding aquifers may be leading CMOST astray. Generally, non equilibrium conditions as a result of well completion operations are not well understood or simulated. Consider using skin next time instead of manipulating near wellbore perms. Wellbore performance in general needs to be better understood. Regarding use of relative permeability tables in IMEX, the end points that are specified below the relative permeability tables in the .dat file generally over-ride the tables. The tables are re-scaled to match the end points. In the case of SLCON, IMEX seems to have ignored the SLCON value specified and run with the values in the table. To be sure of what end points IMEX is using, run one time step and have IMEX output the end points that it is using. Note that
316
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
the tables in the .out file may only be régurgitation of input, rather than what IMEX has actually used.
15.13.4
History Match
Next time adhere to a 'formula' approach on objective term weightings for the history match. Stick with production weighting based upon cumulative volume of reservoir fluid withdrawn at initial reservoir conditions until a more logical method is revealed. Adjusting weightings on terms, and including 'artificial' terms without data to get a perceived 'better match' only serve to put the basis of the simulation in question. Recommend against using any artificial objective function terms. In this project, bottom hole pressure data for the three wells on artificial lift was constructed in an effort to produce simulations with lower bottom hole pressures. This likely was not helpful. The reservoir simulation is working with monthly production volume data. Given that the pool is in initial production stages, and was until September 2009 on rate restriction, there were daily variations in well operating conditions that the reservoir simulation cannot capture. This may be an explanation for gas production history match discrepancies. Including well cumulative water production terms in the history match objective function was likely not beneficial in this case because the method for accounting for water production did not reflect actual reservoir conditions. The reservoir simulator needs to be improved to reflect 'non equilibrium' conditions present due to well completion operations. One possible explanation for the water production in this pool is that the gelled oil completions mobilized connate water and created a water bank that was produced back once the wells were put on production. Complete a sensitivity analysis in CMOST prior to the history match. Eliminating variables that are found to have minor influence allows CMOST to produce results more quickly, and produce optimized simulations that may be more stable in GEM. Rather than completing a manual history match exercise to 'get a feel' for simulation behavior, similar knowledge can be gained quickly from the sensitivity analysis. The method of modeling near wellbore reservoir properties appears to be flawed. A better approach may be to use skin factor from well test interpretation if available, or estimate skin based on the completion program.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
317
15.13.5 Black Oil to Compositional Model Conversion One important lesson from this exercise is that lumping of such components will give better results. The geological model arrived at through IMEX simulations had to be compromised due to the onerous fluid system that the user imposed on the GEM simulations. Well deliverability, and near wellbore reservoir properties need to be more accurately represented in the simulations. An approach of modeling wellbore deliverability by skin, rather than permeability variations, may have resulted in simple stable GEM models. Extreme values and order of magnitude variations in near wellbore permeability likely contributed to GEM model instability. 15.13.6
Recovery Alternatives
Additional optimization can be done, this is conceptual study. The evaluation method used here was to optimize an enhanced recovery program for one particular geological realization, and then apply that identical program to the other geological realizations to understand the impact of making the wrong decision. This was done to demonstrate the core of what would be done in a full Decision Analysis type study. The method of looking at incremental capital investment steps (pump off->horizontal wells->acid gas flood) was also done to demonstrate the core of what would be done in a full Decision Analysis type study. 15.13.7
Economics
Economic uncertainty was evaluated for one particular geological realization for demonstration purposes. In a full Decision Analysis study, this uncertainty would be evaluated for all geological realizations.
15.14
End Note
This report has been condensed International Acid Gas Injection 2010, Calgary, AB. The author and not give any warranty express or
for presentation at the Second Symposium, September 28-29, his employer and affiliations do implied, and shall not be liable
318
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
for any loss, claims, costs, damages or any other action caused by direct or indirect use of this material. Application of the information contained in this material is entirely at the risk of the user
References 1. Fekete Associates Inc. Application for Good Production Practice - Primary Depletion Pool Dunvegan C Pool, Grande Prairie Field. Calgary : s.n., 2009. 2. IHS. Oil Reserves Summary, Dunvegan C Pool, Grande Prairie Field. Calgary : s.n., 2010. 3. Slider, H.C. Worldwide Practical Petroleum Reservoir Engineering Methods. Tulsa, Oklahoma : PennWell Publishing Company, 1983. 4. Aziz, K., Settari, T. Petroleum Reservoir Simulation. London : Applied Science Publishers, 1979. 5. Evaluation of Normalized Stone's Methods for Estimating Three Phase Relative Permeabilities. Fayers, F.J., Matthews, J.D. 1984, Society of Petroleum Engineers Journal, pp. 224-232. 6. An improved model for estimating three phase oil-water-gas relative permeabilities from two phase oil-water and oil-gas data. Maini, B.B., Kokal, S.L.,. 1990, The Journal of Canadian Petroleum Technology, pp. 105-114. 7. Energy Resources Conservation Board. ST98: Alberta's Energy Reserves and Supply/Demand Outlook. Calgary : s.n., 2006. 8. Study shows 'huge' C 0 2 storage potential in Alberta. Carbon Capture Journal. March/April, 2010,14. 9. Alberta Research Council. Hydrochemistry of the Peace River Arch Area, Alberta and British Columbia, Open File Report 1990-18.1990. 10. Energy Resources Conservation Board. Directive 65: Resources Applications. Calgary : s.n., 2009. 11. —. Directive 007-1: Allowables Handbook-Guidelines for Calculation of Monthly Production Allowables. Calgary : s.n., 2007. 12. A New Method for Petroleum Fractions and Crude Oil Characterization. Castells, F., Hernandez, J., Miquel, J. 1992, SPE Reservoir Engineering, pp. 265-270. 13. Alberta Geological Survey, Energy Resources Conservation Board. ERCB/ AGS Special Report 094: Stress Regime at Acid Gas Injection Operations in Western Canada. Edmonton : s.n., 2008.
16 C0 2 Flooding as an EOR Method for Low Permeability Reservoirs Yongle Hu1, Yunpeng Hu2, Qin Li2, Lei Huang1, Mingqiang Hao1, and Siyu Yang1 ^hina Petroleum Exploration and Development Research Institute Beijing, People's Republic of China 2 China University of Geosciences Beijing, People's Republic of China
Abstract
Carbon dioxide flooding is an efficient enhanced oil recovery (EOR) method for low permeability reservoirs. C0 2 swelling oil, reducing oil viscosity significantly, and obtaining miscibility at specified temperature and pressure can decrease the surface tension considerably. Simultaneously, injecting C0 2 into reservoirs is an important way for C0 2 sequestration. The C0 2 flooding technique has not been widely implemented in China. Technology suitable for low permeability reservoirs in China should be developed further.
16.1
Introduction
Carbon dioxide injection can effectively make up the voidage of low permeability reservoirs. Because of the injection difficulties and poor pressure transmission of low permeability reservoirs, it is difficult for water flooding to build up an effective displacing system to maintain reservoir pressure. In practical, the pressure level of low permeability reservoir developed by water only maintains at 70% of the primary pressure, which seriously affects the oilfield development effects. The viscosity of C 0 2 is far less than that of water, so C 0 2 can be injected into low permeability formation more easily and pressure can be recovered efficiently.
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (319-328) © Scrivener Publishing LLC
319
320
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Carbon dioxide injection can effectively decline the lower limit of the reservoir put to use to improve oil use rate. C 0 2 injection can effectively decrease lower limit of used extent of formations, and increase the oil producing degree. C 0 2 can flow into tiny pores that water cannot, to mobile the reserves and improve the injection profile. For water sensitive low permeability formation, C 0 2 is a favorable alternative. Carbon dioxide flooding can increase displacement efficiency and oil recovery factor. The mobile oil saturation of low permeability reservoir is low, the efficiency of water flooding and water flooding recovery are low, too. For those main oil fields in China, in terms of development stage under high water content, it has been proved that C 0 2 flooding can further improve the recovery of old fields with high water content. Flooding experiments in long cores showed that: compared with water flooding, C 0 2 flooding can dramatically increase oil displacement efficiency. While C 0 2 is implemented in a field with water content 95%, oil recovery can increase 10% [1].
16.2
Field Experiment of C 0 2 Flooding in China
In China C 0 2 flooding was focus in the early 1960s and some experiments and pioneering tests were carried out. In 1963, Daqing Oilfield first researched on C 0 2 flooding and designed some pilotscale experiments in the field. These experiments showed C 0 2 flooding technology can improve the recovery by about 10%. From 1990 to 1995, experiments of water alternating C 0 2 gas were implemented in the Well Zone 45, 3-3 C, located in eastern transition zone of Sanan area, and the water content in field was up to 98%. Oil recovery was enhanced by 6% and C 0 2 utilization efficiency was 0.23 t / t C0 2 . In 1999, C 0 2 flooding experiments are tried in Xinli Oilfield, Jilin Province. There was 5200 ton more oil extracted while 1500 ton C 0 2 was injected into subsurface. Also, experiment of water alternating C 0 2 gas to form miscible displacement was taken in block 14, Jiangsu Oilfield in 1998 with water content in field above 95%. The oil recovery was increased by 4% while the C 0 2 utilization efficiency is 0.4t/tCO 2 . Since 2006, experiments of Water and gas synchronizing injection have commenced in Caoshe Oilfield. Although the experiment is still in progress, it has obtained miscible displacement and promoted well production
C 0 2 FLOODING AS AN EOR METHOD
321
from the data available. In 2008, C 0 2 flooding was used as primary oil recovery in the Tree 101 block and Songfangtun block, Daqing Oil field, and preliminary effect of enhancing oil recovery has been realized. In 2008, C 0 2 flooding experiments started in the Black 59 block, and the effect of gas injection is remarkable. Well production was substantially promoted compared with its initial production. At present, oil fields in China such as Daqing, Jilin, Shengli, Liaohe, Jiangsu have fulfilled some significant work of C 0 2 flooding on research and implementation in field. However, C 0 2 flooding in China is still immature, since the related research has just advanced in a short period and the fields in which tests were implemented are small. More attention should be paid to the future research, such as C 0 2 flooding tests, integrated technique of C 0 2 flooding and resolution of the key technology of C 0 2 flooding.
16.3 Mechanism of C0 2 Flooding Displacement C 0 2 is a kind of gas with high solubility in both water and oil. A large amount of C 0 2 dissolving in crude oil can result in crude oil's volume inflation, viscosity decrease and interfacial tension decline. In addition, the Carbonic acid generated after the C 0 2 dissolves in the water could play a role of acidification. If the compositions of the crude oil are favorable, C 0 2 could be mixed with oil at certain pressure and the recovery efficiency would be significantly increased. It has been proved that C 0 2 is an efficient medium for enhancing oil recovery through a large number of laboratory and field experiments. C 0 2 flooding falls into three categories: miscible phase displacement (semi-miscible phase included), non- miscible phase displacement and carbonated water displacement. The high efficiency of C 0 2 displacing oil in porous medium mainly attribute to following merits: 1. Distention C 0 2 can significantly dissolve into crude oil. The full dissolution can give rise the crude oil to a high volume expansion which is commonly about 10% - 40% [2]. The volume expansion can play an important role in oil displacement. Firstly, the residual oil in-situ after water flooding is reciprocal to the expansion coefficient, i.e., the higher the expansion coefficient, the less the residual oil in-situ.
322
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Secondly, the dissolved oil drops could extrude the water out of the porous space to form a water wet system with water drainage rather than water suction. This can lead to higher oil relative permeability curve of oil drainage than that of oil suction. Therefore, a beneficial oil flowing environment is imposed in any given saturation conditions. Thirdly, the oil volume expansion, on one hand, can increase stratum's elastic energy significantly; on the other hand, the residual oil after expansion could completely or partly escape from the bound water and become mobile oil [3]. 2. Viscosity reduction effect When the crude oil is saturated with C 0 2 gas, its viscosity of crude oil can be greatly reduced. Under the subsurface condition, the higher the pressure is, the more the C 0 2 dissolves in the crude oil, and the reduction of the crude oil viscosity will be more significant [4]. The crude oil viscosity may reduce by 1.5-2.5 times after C 0 2 dissolve in it. In general, the viscosity reduction ratio is proportional to the viscosity of crude oil, i.e., the viscosity reduction ratio in heavy crude oil is much greater than that in light crude oil after C 0 2 dissolution. Therefore, it is suggested that C 0 2 should be used to develop the heavy crude oil, since the viscosity of heavy crude oil with saturated C 0 2 can decline remarkably. The mobility ratio improves and oil relative permeability will be correspondingly promoted, too. 3. Improvement on the mobility ratio and reduction on the interfacial tension When C 0 2 dissolves in water, the water's viscosity can increase 20%-30% and its mobility increase by 2 to 3 times. In the meanwhile, with the decreasing of oil mobility, oil /water mobility ratio and their interfacial tension will be further reduced, so that the oil could flow more easily. 4. Improvement on injection capacity and acidification [5] C0 2 -water mixture is slightly acidic and it can react with the formation matrix as follows: C02+H2O^H2C03 H2C03 + CaCOs -> Ca(HC03 ) 2
(1) (2)
C 0 2 FLOODING AS AN EOR METHOD
H2C03 + MgC03
-> Mg(HC03
)2
323
(3)
The generated bicarbonate can easily dissolve in water and increase the permeability of reservoir, particularly those formations whose vicinity around bore hole a great amount of water and C 0 2 pass by. In addition, due to acidification, C0 2 -water mixture can relieve inorganic scale blocking, dredge the oil flowing pathway and recover single well production to a certain extent. 5. The role of dissolved gas drive [6] The solubility of C 0 2 in crude oil is very high. With gas injecting, part of the C 0 2 will dissolve in crude oil and the amount of C 0 2 dissolution will increase with increasing injection pressure. After C 0 2 injection into reservoir, the reservoir pressure will reduce with oil extraction. As a result, the C 0 2 dissolving into the crude oil will be separated from the crude oil, which can play a role as gas drive similar to the natural type of solution gas drive. 6. Extraction and vaporization of the light components of crude oil There is a good miscibility between light hydrocarbons and C0 2 . When pressure exceeds a certain value which depends on the oil properties and temperature), C 0 2 can make the light components extraction and vaporization, which is more prominent for the light crude oil. Extraction and vaporization of light hydrocarbons in crude oil is one of the main mechanisms of enhancing oil recovery through C 0 2 injection. 7. Miscibility [7] Under the reservoir temperature, the pressure at which C 0 2 and oil reach miscible phase is called minimum miscibility pressure (MMP), which depends on the pureness of C0 2 , oil component and reservoir temperature. When the reservoir temperature goes up, the MMP increases; in addition, it also increases while the molecular weight of component above C5 in crude oil increased. MMP can be influenced by pureness (impurity) of C0 2 . MMP will decrease while the critical temperature of impurity is lower than that of C0 2 , and vice versa. The mixture of C 0 2 and primary oil can not only extract and vaporize light hydrocarbon, but also can realize an oil zone mixed with C 0 2 and light hydrocarbon, which is the most effective oil displacement process when the oil zone is mobile.
324
16.4
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Perspective
Different from marine sediment reservoir, most of the oil fields found in China belong to continental sedimentary reservoir feathered with a complex tectonic geological features, serious heterogeneity, high content of heavy component in crude oil and larger viscosity. In order to implement C 0 2 flooding successfully, we need to resolve the following issues: 1. Determine the screening criteria to implement C 0 2 flooding and evaluation methods of enhancing oil recovery of C 0 2 flooding based on the geological characteristics in China. 2. Currently, there are a number of oversea screening criteria with respect to C 0 2 enhancing oil recovery; however, it is still uncertain that those standards are suitable for continental sedimentary reservoirs in China, especially for the low permeability reservoir. Therefore, the domestic geological features should be considered while we determine the range of application of C 0 2 miscible flooding, immiscible flooding and throughput, formulate the reservoir screening criteria of C 0 2 enhancing oil recovery and C 0 2 sequestration and the evaluation methods of EOR under low permeability reservoirs conditions. All standards should base on the geological features in china: under continental sedimentary geology, synthesizing the complex tectonic geological features (including basin characteristics, geological structure, sedimentary faces characteristics, fault characteristics, cap sealing characteristics, etc.), reservoir characteristics factors (including factors affecting the ability of the reservoir injection, such as the reservoir permeability boundaries, heterogeneity parameters, etc.), and the quality of crude oil factors (including the influence of crude oil components, and C 0 2 purity, etc.). 3. Develop phase evaluation and characterization technology for the Conformation fluid mixing system Components exchange will take place between C 0 2 and crude oil during the process of C 0 2 flooding, which may give rise to a complex process of phase change. Therefore, the principle to compile scientific
C 0 2 FLOODING AS AN EOR METHOD
325
scheme for C 0 2 flooding should base on phase evaluation of Conformation fluid mixing system under formation conditions. The current Conformation fluid mixing system phase evaluation is mainly accounted for through PVT experiments, including sampling, mixing with samples, testing and data calculations and analysis. Inaccuracy in each step could lead to incorrect results. How to make an experiment more closely reproduce C0 2 -crude oil system phase under formation conditions should be focused. With respect to the phase characterization, in order to build a basic principle for the following C 0 2 flooding simulation, integrated methods should be proposed on the evaluation of thermal stability of well flow properties, the division and combination of pseudocomponents, the solid precipitation characterization, the adjustment of phase equation, the calculation methods of the minimum miscibility pressure (MMP), and so on. 4. Develop the applicable C 0 2 flooding fine reservoir characterization technology and the numerical simulation technology Most oil fields in China belong to continental deposit featured with strong heterogeneity, small sand body distribution, and more interbeds. Investigations on the distribution characteristics of the sand body, the development characteristics of intercalation and the connectivity between injection wells and production wells should be based on the reservoir characteristics with thin bed and narrow channel, so that a reliable geological recognition should be provided for implementing C 0 2 flooding. To the numerical simulation technology, we should further develop the multiphase and multi -component simulation considering advection and diffusion effect among different phases, since there are multi-liquid flow and solid-Phase Precipitation in the process of C 0 2 flooding. Therefore, applicable C 0 2 simulation methods can be built as a good technical storage for the implementation of C 0 2 to the large quantities of complex type of reservoir in China. 5. Develop dynamic monitoring technology for displacement front of C 0 2 gas drive It's very important to monitor the displacement front of C 0 2 gas drive in the process of C 0 2 flooding. Development of the monitoring
326
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
technology for the displacement front of C 0 2 gas drive based on the seismic data, well logging data and production performance data is appealing very urgently. 6. Improve the drilling and the surface engineering for the C 0 2 flooding As research on the drilling and the surface engineering for the C 0 2 flooding in China starts late, the application effect of process equipment related to C 0 2 flooding needs to be tested and the C 0 2 anticorrosion technology as well as C 0 2 separation technology needs to be further developed, too. 7. Develop new technology of C 0 2 flooding for enhancing the oil recovery In view of the shortcomings of the conventional technology, the new generation of C 0 2 flooding technology for enhancing the oil recovery has the following improvements: a. using horizontal wells to adjust the well pattern and the displacement methods, and improve the sweep extent to the remaining oil and the displacement efficiency, b. increasing the mobility ratio and controlling the viscous fingering of C 0 2 to expand swept volume, c. reducing the minimum miscibility pressure (MMP) by adding miscible agents, d. paying attention to integrating all technologies.
16.5
Conclusion
In conclusion, the implementation of the C 0 2 enhancing oil recovery in China is still in initial stage. We need to further our research urgently and try our best to provide the technical support for the large-scale industrial implementation of C 0 2 flooding in the low permeability oil field.
References 1. A.T.F.S. Gaspar, S.B. Suslick, D.F. Ferreira, and G.A.C. Lima, "Economic Evaluation of Oil Production Project with EOR: C 0 2 Sequestration in Depleted Oil Field," SPE94922.
CO
FLOODING AS AN EOR METHOD
327
2
2.
3.
4.
5.
6. 7. 8.
Cai Xiulin, "The Mechanism and Application of Single Well C 0 2 cyclic injection technology for production improvement," Petroleum Drill Technology, vol. 24(24), pp. 45-46,2002 (in Chinese). Wang Shouling, et a l , "Research on the mechanism of production improvement and application of C 0 2 cyclic injection technology," Drill Technology, vol. 1, pp. 91-94,2004 (in Chinese). Yu Yunxia, "The Application of Single Well C 0 2 cyclic injection technology for production improvement in oil field," Drill Technology, Vol. 27, pp. 89-90, 2004 (in Chinese). Liang Fuyuan, "The Application of C 0 2 cyclic injection technology in Fault Block Hydrocarbon Reservoir," Producing Test Technology, Vol. 22(3), pp. 31-33, 2001 (in Chinese). Chen Tielong, "The Tertiary Oil Recovery Introduce," Petroleum Industry Press, 2000 (in Chinese). F. Stalkup, "Field Developing by Miscible Displacement," Petroleum Industry Press, Beijing, 1989. J.H. Goodrich, "Target reservoir for C 0 2 Miscible Flooding," Report DOE/ MC/08341-17, U.S.DOE, Washington,DC, 1984.
This page intentionally left blank
17 Pilot Test Research on C 0 2 Drive in Very Low Permeability Oil Field of in Daqing Changyuan Weiyao Zhu1, Jiecheng Cheng2, Xiaohe Huang1, Yunqian Long1, and Y. Lou1 1
Civil and Environmental Engineering School, University of Science Technology Beijing, People's Republic of China 2 Daqing Petroleum Administration Bur of Petroleum Daqing People's Republic of China
Abstract
The oil reserves about 3.7xl08tonne do not obtain economic development by water flooding in Daqing Changyuan. For obtaining an availability development method to fit a very low permeability oil field, according to the research results of some experiments and reservoir engineering, some testing schemes are designed and numerical simulations are investigated. Based on the testing results of C0 2 injection, some injection gas feasibility and immiscible displacement condition for C0 2 drives are presented. The technology ambit and product change curve is given. The appropriate technology and C0 2 injection condition is gained. Thus, in a very low permeability oil field the C0 2 drive has succeeded in enhancing well production and oil recovery.
17.1 Introduction According to the statistical results of more than 70 foreign oil field, for the very low-permeability oil field, gas injection (especially C 0 2 flooding) is the main technical measure to improve development effectiveness and to establish effective deployment system [1-8]. Utility of gas-water alternating injection and miscible injection could enhance oil recovery by 7 to 15 percent. Proportion of C 0 2 Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (329-350) © Scrivener Publishing LLC
329
330
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
flooding in low permeability oil reservoirs is high, but decreases as permeability increases. Proportion of nitrogen injection is close to that of hydrocarbon gas injection, and decreases as permeability increases. C 0 2 flooding projects usually have the following common conditions: depth less than 2000m; crude oil density between 0.8 and 0.9 g/cm 3 . But some nitrogen and hydrocarbon gas injection projects are used when crude oil density is smaller or larger. The number of C 0 2 flooding projects increases as viscosity increases and the fact shows that CÖ 2 flooding is capable of exploiting highviscosity oil. Gas injection in domestic low-permeability oil fields was blocked because of gas supply shortage in the last few years. Now more and more field tests for gas injection projects are carried out in low-permeability reservoirs. Laboratory test results show that gas injection in very low permeability reservoirs differs significantly from that in common permeability and low permeability reservoirs: gas flow has significant non-Darcy flow characteristics; oil and water have obvious threshold pressures. The characteristics above are also revealed in field tests. Formation conditions and fluid characteristics of lowpermeability oil reservoirs in the periphery of Changyuan Daqing satisfy the selection criteria for C 0 2 flooding, which has better adaptability than hydrocarbon gas flooding. Consequently it is necessary to carry out C 0 2 flooding theory research and field tests in order to summarize experience and lay a solid foundation for further development.
17.2
Laboratory Test Study on C 0 2 Flooding in Oil Reservoirs with Very Low Permeability
Study of phase behavior and experiments of swell, tubule flow and long core flow were carried out on the natural rocks of the Fuyu oil layer and the oil/gas samples collected from the object regions.
17.2.1 Research on Phase Behavior and Swelling Experiments Experiments of simple degassing, P-V relationship, multi-stage degassing and swelling were performed on the object samples.
PILOT TEST RESEARCH ON C 0 2 DRIVE
331
The simple degassing experiment aims to obtain the main parameters such as gas/oil ratio, volume factor, density of initial oil in place and so on; the P-V experiment aims to measure the parameters such as the saturated pressure of the fluid, the fluid density and volume factor under the saturated pressure and so on; the multistage degassing experiment aims to measure the dissolved gas/ oil ratio, volume factor, density, viscosity and the change of liquid volume under the conditions of multi-stage degassing. Simple degassing experiments: Under the formation temperature 86 °C and the formation pressure 22.64 MPa, a simple degassing experiment produced a gas/oil ratio 18 m 3 /m 3 , a crude oil volume factor 1.088 m 3 /m 3 , a crude oil shrinkage factor 8.06%, an average solubility of gas 3.51 m 3 /(m 3 .MPa), and a crude oil viscosity 3.314 mPa.s. These data indicate that the Fuyu oil layer is a reservoir with high oil density and low volume factor, swell, shrinkage and solubility. P-V experiments: According to the measured data, the crude oil volume factor changed a little (1.0566-1.0673) with varied pressures, which suggests a small amount of energy for the volume swelling. Multi-stage degassing experiments: The measured parameters for crude oil under varied pressures show that the viscosity and density of the crude oil increase with decreased pressures while the gas/oil ratio and the volume factor decrease with decreased pressures. That is, the crude oil possesses such characteristics as intermediate density, high viscosity, small expansibility and shrinkage, and low density for the displaced gas. Swelling experiments: For Shengqi Well 1-4 and Fangshen Well 6, the experiments under C 0 2 flooding presented swell factors 1.10, 1.15 and 1.26, respectively. Generally, the parameters for the raw oil hardly changed with varied pressures.
17.2.2
Tubule Flow Experiments
The tubule flow experiments were designed to determine whether the injected gas is miscible with the crude oil. According to such experiments, the lowest miscible pressure at the formation temperature (86 °C) was 47 MPa, with an oil-displacement efficiency 92% (see the pressure dependent oil-displacement efficiencies plotted in Figure 1). Therefore, the oil-displacement experiment under C 0 2 flooding in the Fuyu oil layer was immiscible displacement.
332
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. The variation of oil-displacement and pressure.
17.2.3
Long Core Test Experiments
The conditions in long rock test experiments are usually much more close to the actual situations in the formation. The used rocks that were 28.85 cm in the length with an average air-permeability 2.694xl0~3 urn2. Five series of oil-displacement experiments were carried out on the long rocks, which used gas flooding for Fangshen Well 6, gas flooding, gas/water alternate flooding, pure water flooding and pure C 0 2 flooding for Shengqi Well 1-4, respectively. The following conclusions can be made based on the experiments: 1. the displacement under gas flooding is more facile than that under water flooding, which shows that the threshold pressure difference under gas flooding is smaller than that under water flooding (2.06-2.19 MPa vs 5.45-5.77 MPa). 2. The displacement efficiency under the gas/water alternate flooding is not ideal. For Shengqi Well 1-4 under the water/gas alternate flooding with a threshold pressure difference of 5.77 MPa, the injection pressure kept increasing until that is close to the formation fracturing pressure, which did not lead to both water and gas to break through, and the final recovery factor was only 25.96%. 3. Before the breakthrough point, the recovery factor under gas flooding is higher than that under water flooding (27.4-29.081% vs 23.28% at the breakthrough point). 4. The recovery efficiencies under C 0 2 flooding increase with increased injection pressure. For example, with the injection pressure going up from 6.0 MPa to 35 MPa, the recovery factor at the breakthrough point increased from 32.61% to 44.76% and the final efficiency increased from 39.06% to 56.27%.
PILOT TEST RESEARCH ON CO. DRIVE
333
Table 1. Data comparison of different injected medium. Projects
Recovery Factor at the Breakthrough Point (%)
Starting Pressure (MPa)
Stating Pressure Gadient (MPa/m)
Fangshen 6 C 0 2 flooding
2.19
7.59
29.08
34.32
Shengqi 1-4 C 0 2 flooding
2.06
7.14
27.41
32.20
Shengqi 1-4 gas/water alternate flooding
5.77
20.00
Not breakthrough
25.96
Water flooding
5.45
18.89
23.28
Final Recovery Factor (%)
/
Table 2. Data comparison of different injection pressure. Starting Pressure (MPa)
Stating Pressure Gradient (MPa/m)
Recovery Factor at the Breakthrough Point (%)
Final Recovery Factor (%)
6.0
2.43
8.42
32.61
39.06
22.64
2.29
7.94
41.80
48.15
35.0
2.26
7.83
44.76
56.27
Injection Pressure (MPa)
5. C 0 2 flooding can strongly improve the oil recovery factor. For example, the value at the breakthrough point under C 0 2 flooding was 41.08%, 12.72% higher than that for Fangshen Well 6; the final value was 48.15%, 13.83% higher than that for Fangshen Well. 6. All experiment results are listed in table 1 and table 2.
17.3
Field Testing Research
17.3.1 Geological Characteristics of Pilot Fang 48 fault block is located on the southeast of Songfangtun oilfield, and on the Zhaozhou nose structure, which is in the Sanzhao
334
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
depression of central northern Songliao basin. The three exploration well, Fang 48, Zhaoshen 4 and Zhao 401, which drilled through Fuyu oil layer, were successfully finished drilling from 1989 to 1990 in Fang 48 fault block. 3D high-resolution seismic exploration with a bin of 20x20m, was finished to explore deep gas in 1996, and the developmental condition of the structure and fault was identified in Fang 48. In 1998, for Fang 48 fault block, with other six wells, such as Zhou 7, have been submitted and proved reserve was 812xl0 4 t, oil bearing area was 23km 2 (unit coefficient is 5.5xl0 4 t/km 2 .m). In 1999, for designing Putaohua oil layer development wells, giving consideration to Fuyu oil layer, 5 development control wells were drilled, and oil test was conducted in the 5 wells. At present, there are 8 exploration and development control wells, of which 5 wells test oil yield is more than 1.5t/d. 17.3.1.1
Structural
Characteristics
Fang 48 fault block is located on nose structure of east Songfangtun oilfield. Songfangtun nose structure uplifted extent is small, actually a moderate slope, so Fang 48 fault block is flat air streamed structure. The nose structure has the maximum uplifted extent at -1600m contour line in the structure map of Fuyu oil layer. Fractures are developed around Fang 48 fault block. There are two near northsouth faults, MF13 and MF16, which consist of Fang 48 horst block, in the test area from Tl-1 and T2 reflection layer structure map. The extension of the two faults is about 5 km, and the fault throw is the big end up mold. There are up to 34 minor faults in Fang 48 horst block. The fault throw is about 50m of the top surface fault in Fuyu oil layer, and the horst block scale is larger than that of PI group. Though the faults around test area developed well, that of gas injection test area didn't develop well. 17.3.12
Characteristics of Reservoir
Fuyu oil layer in Sanzhao area is cretaceous for Quan 4 and upper Quan 3 segments, and the distribution is relatively stable. Fuyu oil layer in Fang 48 fault block is in the sandstone enrichment zone, which is effected by northern Songfangtun and southern Zhaoyuan water systems. Fuyu oil layer is river-lake flood plain faciès deposition, which is formed in the HST ancient lake, and the lithology is dark purple, purple mixed green and gray mudstone and gray green, green, gray muddy siltstone, siltstone and gray-brown,
PILOT TEST RESEARCH ON C 0 2 DRIVE
335
oily brown powder, and fine sandstone. The formation thickness of Quan 4 segment (Fu 1 group) is about 100m, and the Quan 4 segment is divided into 7 small layers. The formation thickness of upper Quan 3 segment (Fu 2 and 3 groups) is about 100m, and the segment is divided into 5 small layers. The top surface depth of Fuyu oil layer of Fang 48 fault block is about 1742m. Only Fu 1 group developed oil layer. 10 wells were drilled, and the average drilled sandstone thickness is 13.9m and the effective thickness is 9.3m. Vertical evolution sequences: fluvial facies-Lake flood plain facies-delta facies. The main type of oil layer sand body is channel sand, and the shape of sand body is short strip and intermittent banding. The drilling ratio is 90% of the main oil layers, F14 and F17, and the drilled thickness is 24.5% and 64.4% of total effective thickness respectively. The layer FI4 belongs to the branched channel sand of lake flood facies, the micro-gradient curve is bell-shaped, the lithology is positive cycle, and from bottom to top is calcareous siltstone - sandstone oil powder - muddy siltstone, the bedding is parallel bedding, small oblique bedding and wavy bedding, the sandstone thickness is l~5.8m, the effective thickness is 0.6~2.6m, the average drilled sandstone thickness is 3.5m and effective thickness is 2.1m, and the width of sand body is about 500m based on the well drilling. The layer F17 belongs to the branched channel sand of fluvial facies, the micro-gradient curve is box-shaped, the lithology is dual structure, the bedding is parallel and small oblique bedding, the sandstone thickness is 5.8~10.0m and the effective thickness is 4.8~10.0m, the average drilled sandstone thickness is 7.3m and the effective thickness is 6.0m, the width of sand body is 600m based on well drilling. The development wells (well pattern:300x300m) in the Zhou 16 Pu-Fu commingled test area, which is 9km far away from Fang 48 fault block, is proved that the width of Fu 1 group is about 600m too (See Figure 2). The buried depth of F17 sand body is -1696- -1703m, and the sand body tilted from east to west. The effective sandstone thickness is the largest in the vicinity of Fang 190-138, and is 10.0m. The minimum thickness is in the vicinity of 187-138 (6.0m and 5.6m respectively). The sand body is getting thicker from north to south. Based on the distribution of porosity and permeability of layer F17, in general, the reservoir properties of layer F17 in Fang 48 block is not good, and it belongs to low porosity, ultra-low permeability reservoir. The porosity ranges from 5.8%~17.4%, and the average
336
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. Test well location map of gas injection. is 14.5%; the range of permeability is 0.02~3.66xl0-3um2, and the average is 1.4xl0-3um2. Based on the surface distribution, there is little change in porosity and permeability from north to south (See Figure 3 and Figure 4). 27.3.2.3
Reservoir Properties and Lithology
Characteristics
The typical lithological features of Fuyu oil formation from the Fang 48 well is argillaceous siltstone and fine sandstone containing mud with secondary quartz developing well in the pore space. The pores are mainly narrowing intergranular pores, most of which are not connected, and the rock core analysis indicated that the porosity is 9.0-17.6% with an average value of 14.5% while the average air permeability is 1.4xl0"3um2 with the maximum value is 5.22xl0"3|i,m2 and the minimum is 0.1x10 3 |im 2 (the permeability ratio is 279.5 and mutation coefficient is 5.7). Besides, the sandstone is mainly composed of quartz (21-26.7%), feldspar (29.2-36.2%) and the average rock debris is 33.8% while median grain diameter is 0.068-0.111 and sorting coefficient is 5.7-10.08. Slice analysis suggested that the dominant cementation types of the reservoir are shale cementation and calcareous mixing cementation with average 9.7% shale content. And calcic cementation accumulated locally while shale cementation is mainly recrystallization and distributed in clusters and films. The quartz and feldspar have
PILOT TEST RESEARCH ON C 0 2 DRIVE
337
Figure 3. Distribution map of FI7 porosity.
Figure 4. Distribution map of FI7 permeability.
characters of secondary enlargement and regenerated cementation, which are mainly types of pore-film, regeneration-pore-film. The clay minerals of the Fuyu oil formation are mainly illite (31%) and chlorite (39%), 70% of which is enriched in iron and mixing layers of montmorillonite/illite and montmorillonite /chlorite also composed the clay minerals. The primary attitudes are listed in table 3.
3.5
7.3
FI4
FI7
14.5
14.7
31 39
0.9 1.4
30.8 36.0 41.6 36.1
0.865
0.874
0.872
0.869
Fang 190-138
Zhao 401
Fang 48
Average
125
157
148
80
33
30
35
35
32
124
38.1
0.870
Fang 184-136
35
115
33.8
0.866
Freezing Point (°C)
48
0
IBP (°C)
Oil Viscosity (mPa.s)
Fang 190-140
^~~~-~-^Pro j ects Oil Density Well N o / " ~ \ ^ ^ (t/m3)
Clay Mineral Components (%)
17.0
18.1
19.2
14.0
17.8
16.0
Gel Content (%)
4
4
25.1
28.6
26.2
30.8
20.7
19.4
Wax Content (%)
7
65
Porosity Permeability Illite Chlorite Montmorillonite/ Montmorillonite/ Illite Chlorite 3 2 (10 iim ) (%)
Physical Properties
Table 4. Character table of oil properties from fault block Fang 48.
6
2.1
Sandstone
Thickness (m)
Effectiveness
Level No.
Table 3. Reservoir character of FI4 and FI7 from fault block Fang 48.
M
h-1
n
O1 r O
n
a
25
M
z > a *>
O
I—I
H
CD
M
O ci
M
n
p
oo
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
PILOT TEST RESEARCH ON C 0 2 DRIVE
339
Capillary pressure curve of Fuyu oil formation in Sanzhao didn't show obvious platform and the pore distributed in single peak shape with complicated pore throat structure and well developed micro pore, which indicated that the original pore and secondary pore are all well developed with the maximum pore throat radius is 2.14um and the average is 0.257um. Core observation suggested that fractures didn't develop well and sixteen years' of water flooding experiences from Shengnan testing field proved that the fractures didn't cause bad influences.
17.3.2 Distribution and Features of Fluid Oil reservoir in Fuyu area was mainly formed by controls of lithology and distribution of oil was also controlled by fault block. Generally, oil column are higher in horst block and no united oil-water interface exists. The types of the reservoirs were horstlithology reservoirs. Oil-water distribution in Fang 48 fault block are characterized by pure oil layer in Fuyi formation and dry layer or water layer in F2 and F3 formation. Statistical analysis of crude oil property from five wells in Fuyu oil layer shows that averaging density of crude oil, crude oil viscosity; freezing point, glue content and wax content are 0.869t/m 3 , 36.1mPa.s, 33.0 °C, 17.0% and 25.1% respectively. Analysis of high pressure property of the samples from well Fang 48 and well Zhou 7 shows that averaging density of crude oil, crude oil viscosity, saturation pressure; volume factor and original gas oil ratio are 0.815t/ m3, 6.6mPa.s, 5.3MPa, 1.089 and 17.5 m 3 /t respectively. See table 4. Averaging CLcontent in formation water in Fuyu oil layer is 3067.6mg/L. Total salinity is 7158.0mg/L. Water type is NaHC0 3 . Original strata pressure is between 17.06 and 24.19MPa (average 20.4MPa); pressure gradient ranges from 0.9426 to 1.3151MPa /100m, with an averaging value of 1.1212 MPa/100m. Reservoir temperature ranges from 81.1 to 87.8°C, with the mean value of 85.9°C. Geothermal gradient are 4.51-4.85°C/100m (average 4.72°C/100m), which belongs to normal geo thermal field.
17.3.3 Designed Testing Scheme According to experimental results, recovery factor increases significantly when the injected carbon dioxide slug is lower than 0.3PV and recovery factor increases little when it is more than 0.3PV.
340
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
So, it is determined to inject carbon dioxide slug before 0.3PV and then transform it to water drive after that. Specific schemes are as follows: 1. Injected medium: liquid carbon dioxide (1-6 year); water (after 6 year); 2. Injection-production ratio: 1.5 (early stage); 1.0 (stable stage); 3. Daily oil production designed in early stage: 2.0-2.5t; 4. Daily gas injection of a single well: liquid C 0 2 of 8m 3 /d in early stage; 6m 3 /d in stable stage; About 1.5xl04m3 of liquid carbon dioxide will be injected in above process in six years, then, it is transformed to water drive. Gas injection rate will be investigated and adjusted according to dynamic variations of production wells in implementation process of designed schemes.
17.3.4 Field Test Results and Analysis In 2003, a pilot area for C 0 2 flooding was pioneered in the Fuyu reservoir in the southern Songfangtun oil field. The oil-bearing area was 0.43 km 2 , average air permeability was 1.4xl0~3um2, and effective porosity was 14.5%. Currently the pilot area has one gas injection well and five production wells. The average sandstone thickness of the target layer (FI7) is 8.2 m and the effective thickness is 6.6 m. Fong 188-138 gas injection well started testing in March 2003, only penetrating FI7 layer. The sandstone thickness is 10.3 m; effective thickness is 6.0m; air permeability is 0.79~1.35xl0"3um2. Gas was injected without fracturing. Injection pressure currently is 12.5-13.0MPa, cumulative volume of injected liquid C 0 2 is 16500m3 (0.33PV) and the cumulative injection production ratio is 2.5 (See Figure 5). 173.4.1
Low Gas Injection Pressure and Large Gas Inspiration Capacity
From July to November in 2004, the average bottom-hole pressure was 30.2 MPa at the average daily liquid C 0 2 injection rate of 68m3 in Fong 188-138 gas-injection-well. From August to December 2005
PILOT TEST RESEARCH ON C 0 2 DRIVE
341
Figure 5. The variation of recovery percent and injection pore volume.
the bottom-hole pressure was 30 MPa and the average apparent gas inspiration index was 0.57m 3 / (d.MPa.m) at the average daily liquid C 0 2 injection rate of 50 m3. In Zhou 2 pilot area, the geologic characteristics are similar to Fong 188-138, it has two water injection wells with well spacing 212 m. The wells started fracture injection in December 1999. At the beginning the average oil pressure per well was 13.3 MPa, average water injection rate per day was 16 m3, and the apparent water-intake index per effective thickness calculated according to bottom hole pressure was 0.05 m 3 / (d.MPa.m). After two years the apparent water-intake index per effective thickness was 0.079m 3 / (d.MPa.m). Compared with the two water injection wells in Zhou 2 pilot area, the apparent gas inspiration index per effective thickness of gas injection wells without fracturing was 7.2 times more than that of water injection wells fractured. It shows that gas injection pressure is lower and the gas inspiration capacity is larger in Fuyu layer. 17.3.4.2
Production Rate and Reservoir Pressure Increase after Gas Injection
At the beginning, of the five oil wells in the pilot area, average production rate per day was 2.8t, intensity of oil recovery was 0.28t/d.m. Currently average production rate per day was 1.5t, and the intensity of oil recovery was 0.15t/d.m. Cumulative oil recovery was 7751t, recovery percent to OOIP was 3.37%, oil recovery rate was 0.92%, and total water-cut was 5.2%.
342
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The production change of the well group shows that, from the commissioning date to July 2004 production followed the elastic recovery law. Production per day of well group increased steadily after gas injection test quickening in July 2004; decreased slightly after gas breakthrough of Fong 190-136, 190-140 in March 2005; and stayed above 7t because of response of Fong 188-137,190-138, and increase of production of Fong 190-136, 190-140. It was found that reservoir pressure increased from 8.6 MPa to 12.2 MPa after response through monitoring Fong 187-138 well. By analysis of production change since commissioning date in Zhou 2 pilot area, gas injection took effect in Fong 48 well group in August 2004 and cumulative incremental oil production was 1524 tons. Currently daily incremental oil production of the test well group was about 4 tons. 173.4.3
Reservoir Heterogeneity Is the Key to Control Gas Breakthrough
Breakthrough of wells Fong 190-136 and 190-140 occurred in March 2005. The present quantity of C 0 2 contained in casing pipe were 90.3% and 91.8%. The earlier breakthrough of the two wells was mainly due to the reservoir heterogeneity. From the horizontal distribution, the porosity and permeability of layer FI7 increased gradually from north to south. The permeability of well Fong 190-140 was the highest (about 2.6xl0-3um 2 ), and that of the other five wells were about 1.6-1.8x10 3 um 2 . In view of the vertical rhythmic profile of layer FI7, thickness with the relatively high permeability of well Fong 190-136 and 190-140 were apparently greater than other two wells. Due to reservoir heterogeneity, C 0 2 breakthrough of wells Fong 190-136,190-140 occurred earlier. Production well after gas breakthrough has following characteristics: 1. Production rate increased steadily. Seen from the curve of Fong 190-136,190-140 well's production change, oil production per day increased steadily from November 2004. The production slightly decreased early after gas breakthrough in March 2005, but increased steadily afterwards.
PILOT TEST RESEARCH ON C 0 2 DRIVE
343
2. C 0 2 content in the gas production increased gradually as well with gas-oil ratio and casing head pressure. Current gas-oil ratio of Fong 190-136 andl90-140 calculated according to Molar composition of gas production are 186 and 218m 3 /m 3 respectively; daily liquid C 0 2 production is 0.48~0.75m3. Besides, casing head pressure after gas breakthrough in production wells is gradually increased. At present the casing head pressures of the two gas breakthrough wells are between 3.7 and 4.6MPa, and that of wells without breakthrough of gas are below 0.5MPa. 3. Reservoir pressure is relatively high. Integral well test was carried out in the injection well group from May 2005 to June 2005. The reservoir pressure of Fong 190-136 and 190-140 well were 13.4 and 14.8MPa respectively, which were significantly higher than the other 3 wells (between 3.6 and 10.6MPa). 173.4.4
C02 Throughput as the Supplementary Means Reservoir's Effective Deployment
ofFuyu
Test well Fong 188-137 was put into production with 80m well spacing in August 2004. The complete well was only perforated in layer FI17. Sandstone thickness is 8.4m. Effective thickness is 5.7m. Besides, the well was put into production without fracturing. Early daily oil production was only 0.02t on in the test, and the daily oil production was between 0.2 and 0.3t from January to May 2005. C 0 2 throughput test was carried out in well Fong 188-137, and overall of liquid C 0 2 injected was 120m3. Early after throughput daily oil production was 2.3t, and oil recovery rate was 0.4t/d.m. Afterwards daily oil production was between 0.6t and 1.3t. According to the well's production change, daily oil production gradually increased from late August, which shows that C 0 2 throughput plays an early role (See Figure 6). Besides, Fong 190-138 well, which had a low oil recovery rate since it had been put into production, carried out C 0 2 throughput. Early incremental production was relatively high. However, due to the influence of the pump operating duty and project, the validity only lasted 50 days, and cumulative incremental oil production was 61ton (See Figure 7).
344
C0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. Fong 188-137 well daily oil production curve and water cut curve.
Figure 7. Fong 190-138 well daily oil production curve and water cut curve.
17.3.4.5
Numerical Result Shows that Carrying Out Water Flooding after Injecting Certain Amount of C02 Slug Is Better
In order to simulate the technical measures improving effect of gas injection, nine numerical schemes of four types were designed (See Table 5).
PILOT TEST RESEARCH ON C 0 2 DRIVE
345
Table 5. Numerical simulation program. No.
Description
Basic
1
Close gas injection well and keep production well producing
Gas injection
2
Maintain injecting Inject liquid C 0 2 at rate of gas in injection well 4, 9,14,18,27m3/d and producing oil in production well
3
Impulse gas injection
Carry out impulse gas injection (close wells with breakthrough of gas. After completing gas injection, open all the wells) and keep continuous gas injection after completion of impulse gas injection. Simulate the effect of impulse gas injection with different cycles
4
Inject a water slug first and then carry out gas flooding
Study the effect of various water slug sizes and velocities of follow-up gas injection on displacement efficiency
5
Carry out water flooding directly
Three water injection velocities: 10,15,20m3/d
6
Continue injecting certain amount of gas and then carry out water flooding
Continue injecting 4000, 6000, 8000,10000,15000, 25000, 30000m3 liquid C02 and then carry out water flooding. Study the effect of gas injection with various injection velocities
7
Continue impulse gas injection and then carry out water flooding
Carry out impulse gas injection at first and then change to water flooding after gas injection is completed
Category
Continue gas injection and change to water flooding afterwards
Details
(Continued)
346
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 5. Numerical simulation program. (Continued) Category Gas-water alternating flooding
No.
Description
Details
8
Inject gas and water alternatively and then change to water flooding
After injecting water and gas alternatively, carry out water flooding. Simulate the effect of gas injection adopting different proportion of gas alternating water
9
Carry out tapered gas-water alternating injection and then change to water flooding
Change gas-water ratio. Increase water and decrease gas, or contrarily
In summary, due to low permeability and high underground crude oil viscosity (6mPa.s), if basic scheme is adopted, productivity and the ratio of total oil produced to OOIP would be few. Concerning the three gas injection schemes; gas oil ratio increased so fast that after 6 to 8 years it would be greater than 1000 m 3 /m 3 and the well had to be closed. The ratio of total oil produced to OOIP was low. In regard to the water alternating gas displacement, persistent increase in the injecting pressure made the scheme hard to carry out. Consequently, the preferred scheme is to carry out water flooding after injecting certain amount of C 0 2 slug. However, numerical simulation did not take non-Darcy flow into account, and its conclusion needs to be further studied (See Table 6).
17.4
Conclusion
1. Both laboratory research and field test results proved that gas injection could reduce interfacial tension and enhance oil recovery, having unique advantages of developing very low-permeability oil reservoirs similar to Fuyu oil layer. 2. The pressure of miscible phase was 47 MPa in the laboratory research on gas injection in Fuyu layer. However, field tests could only adopt immiscible
Gas-water alternating flooding
Continue gas injection and change to water flooding afterwards
Gas injection
Basic
Category Prediction 10 Years Later
42000
37000
9470
15000
14500
29800
25500
4
5
6
7
8
9
39000
2
3
9470
19765
26440
29545
27982
29372
0
0
0
0
23495
30527
23855
24577
20653
21963
18700
18741
14861
8634
15665
8994
9716
5792
7982
5183
5707
0
(Continued)
9.08
11.79
9.22
9.49
7.98
8.48
7.23
7.24
5.74
Cumulative Liquid Cumulative Water Cumulative Oil Cumulative Oil Ratio of total C0 2 (m 3 ) Injection (m3) Production (t) Production (t) Oil Produced to OOIP (%)
1
No.
Table 6. Results of numerical simulation.
PILOT TEST RESEARCH ON C 0 2 DRIV
Gas-water alternating flooding
Continue gas injection and change to water flooding afterwards
Gas injection
Basic
Category
0
18107
53139 53914 58047 58800 36551
9470
15470
14500
29800
25500
5
6
7
8
9
28262
37516
31422
31924
27458
10155
19409
13315
13817
9352
10.92
14.50
12.14
12.33
10.60
Continue injecting gas for eight years and injecting 5700 m 3 water and then close oil wells.
7.0
4
0
Ratio of total Oil Produced to OOIP (%)
Continue injecting gas for seven years and then close oil wells.
Continue injecting gas for six years and then close oil wells.
9470
Cumulative Liquid Cumulative Water Cumulative Oil Cumulative C0 2 (m 3 ) Injection (m3) Production (t) Incremental Oil production (t)
Prediction 20 Years Later
3
2
1
No.
Table 6. Results of numerical simulation. (Continued)
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
PILOT TEST RESEARCH ON C 0 2 DRIVE
349
flooding. In long-core test, threshold pressure of gas injection was lower than that of water injection by 11 MPa/m, and breakthrough recovery of gas injection was higher than that of water injection by 4-6 percentages. Consequently it is feasible to test C 0 2 flooding in Fuyu layer. 3. Compared with hydrocarbon gas injection, oil recovery of C 0 2 flooding increased by 15 percentage points, and therefore achieved better results. 4. C 0 2 flooding test in Fuyang reservoir shows that injection pressure is lower and gas inspiration capacity is larger. The advantage proves that C 0 2 flooding could build u p effective deployment system in very low-permeability Fuyu reservoir without grown fractures. 5. Gas injection could form breakthrough hard to control and adjust, which could cause imbalance of effect in the horizontal after breakthrough of gas in some wells. Balance of gas drive in the horizontal is the key to improve sweep efficiency.
17.5 Acknowledgement This research was supported by the National Natural Science Foundation of China (10772023) and the National Key foundation of China (50934003).
References 1. Bentsen R G. "Effect of Momentum Transfer Between Fluid Phases on Effective Mobility." / Pet Sei Eng, 1998, 21 (1-2), 27 2. Morrow N R. Interfacial Phenomena in Petroleum Recovery. Monticello: USA, Mercel Dekker Inc, 1991 3. Zhu Weiyao. Liu Xuewei. Luo Kai. "Dynamic Model of Gas-Liquid-Solid Porous Flow with Phase Change of Condensate Reservoirs." Natural Gas Geoscience. 2005,16 (3): 363 4. Zhu Weiyao. "Theoretical Studies on the Gas-Liquid Two-phase Flow." Petroleum Expoloration and Development, 1988,15 (3): 63 5. Zhu Weiyao. "Theoretical Studies on the Gas-Liquid Two-phase Flow (Including a Phase Change)Through Porous Media." Acta Petrolei Sínica, 1990, 9(1): 15
350
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
6. Yu Mingzhou, Lin Jianzhong. The Dynamics of Nanoparitcle-Laden Multiphase Flow and Its Applications. 7. T.P Fishlock,C.J Probert. "Waterflooding of gas-condensate reservoirs." SPERE, 1996,11(3) 8. Prieditis J,Brugman R J."Effects of Recent relative Permeability data on C 0 2 flood modeling (A)." In: the 68 Annual Technical Conferences and Exhibition of SPE (C). SPE26650, Huston, Texas, 1993, 467-481.
18 Operation Control of C02-Driving in Field Site. Site Test in Wellblock Shu 101, Yushulin Oil Field, Daqing Xinde Wan, Tao Sun, Yingzhi Zhang, Tie j un Yang, and Changhe Mu CNPC Daqing Branch, Daqing, People's Republic of China
Abstract Based on C0 2 -driving test of extra-low permeable Fuyang oil layer in Well block ShulOl, Yushulin Oil Field, Daqing, the relationship of pore volume and injection mode, intensity of gas injection and injection rate was found in stratum that has an air permeability of around 1 millidarcy. Methods of adjusting injection profiles and production profiles were initially formed, according to the injection situation and dynamical property. Liquid state C 0 2 can be injected into oil layer as required in order to complement producing energy and oil production. Injecting gas in advance based on the stratum pressure and then bringing in oil wells can guarantee that oil wells take affect earlier and achieve economic yield without taking other measures. The strong ability of absorbing gas in stratum keeps the reservoir pressure high, for a longer time, and creates conditions for miscible displacement. Taking the different flow pressure control and systematic management can control one-way gas onrush and postpone gas channeling in line with oil yield, production intension and beneficial situation. For oil wells that are not responding, we can improve the benefit rate by C 0 2 throughput lead to well connected oil wells. Yushulin Oil Field is regarded as the typical large-scale oil deposit with extra low permeability, low fluidity, and low yield; with extra low permeability Fuyang oil layer as the interest bed under the main development, with the average air permeability of 2.71 xl0~3 urn2 and porosity of 10.8%. In addition, it has a bad water drive development effect, which is featured by low daily oil yield per well (0.7 t / d ) , low oil production speed (0.56% currently), low recovery degree, and low geologic reserve recovery degree (8.5%); thus, it is difficult to adopt the water drive for the oil layer with the Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (351-360) © Scrivener Publishing LLC
351
352
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
permeability lower than 1.5xl0"3um2. Accordingly, by the end of 2007, the site test of C02-driving was done in the Well block Shu 101 of Yushulin Oil Field, which was intended to explore the effective development approach of the reservoir bed with the permeability of 1.0xl0"3um2 and difficult exploration.
18.1 Test Area Description 18.1.1 Characteristics of the Reservoir Bed in the Test Area The test area is with the oil-bearing area of 2.36 km 2 and geologic reserve of 217.8xl0 4 t, which is mainly used to explore Fuyang oil layer. Besides, Fuyu oil layer is with the average porosity of 10.0% and air permeability of 1.16x10" 3 um 2 ; Yangdachengzi oil layer is with the porosity of 10.8% and air permeability of 0.96xl0~3um2. Additionally, the oil deposit is with the buried depth of 1806-2283 m; and the average original saturated pressure is 4.94 MPa, average initial gas-oil ratio is 22.8 m 3 /t, crude oil viscosity of the stratum is 3.6 MPa-s, original stratum pressure is 22.05 MPa, and the stratum temperature is 108°C. Through the slime-tube test, we find out that the min. miscible-phase pressure is 32.2 MPa, and the test area is the C0 2 - immiscible driving.
18.1.2 Test Scheme Design The test area is applied with the well pattern of 300x250 m rectangle five-spot area, which is featured by 23 wells in 7 rows, 7 injection wells and 16 exploratory wells, well array direction of NE77.5 0 and consistent with the max main stress direction. Firstly, three main oil layers including YI 6, YII 41, and YII 42 will be perforated, with the reserve of 118.7xl0 4 t, accounting for 54.5% of the total reserve; in the later stage, the upper part of Fuyu oil layer will be perforated, which will cost the total reserve of 148.5xl0 4 t. The scheme predicts that, the recovery ratio will be 20.1%. Based on the advance gas injection of six months and normal injection of three months and rest period of one month, it is designed that the well head injection pressure < 25.5 MPa and oil well production flow pressure a 5 MPa.
OPERATION CONTROL OF C 0 2 - D R I V I N G IN FIELD SITE
18.2
353
Test Effect and Cognition
Based on the advance gas injection principle, injection wells were successfully put into operation in December 2007, and all oil wells were put into operation till April 2009. By the end of June 2010, the cumulatively-injected liquid C 0 2 w a s 5.86xl0 4 t, cumulative oil production was 2.12xl0 4 t, recovery degree was 1.79%, oil production speed was 1.18%, cumulatively-injected HCPV amount was 0.063, annual injection-production ratio was 2.08, cumulative injection-production ratio was 3.38, annual oil replacement ratio was 0.65 t / m 3 , and cumulative oil replacement ratio was 0.39 t/m 3 .
18.2.1 Test Results The test shows that the liquid C 0 2 can be injected into the oil layer as specified to timely supply the oil layer energy and keep the stable production of the oil well. With the same gas injection amount, the gas-injection wells in the Well block Shu 101 are with the initial gas injection pressure of 18.5 MPa, and the current gas injection pressure is 17.8 MPa. With the reduction of the injection allocation amount and extension of the shut-in period of partial wells, the gas injection pressure is stable with a slight decline. In addition, the water injection pressure of Yushulin Oil Field is with typical increase, and the water injection pressure and daily water injection amount of the adjacent Well block Shu 8 increases by 4.4 MPa and reduces by 50% in the same period (see Figure 1). The first 2 wells are with the initial air suction index of 115.2 t / d . MPa and air suction pressure of 17.4 MPa. Currently, seven gasinjection wells are with the air suction index of about 42.0t/d.MPa and air suction pressure of about 17.2 MPa (see table 1). Viewing from the yield variation, it is always kept stable. On average, the daily oil yield per well keeps at 2.7 t; however, the index of the water drive yield is in a diminishing law, which is of larger reduction amplitude. By the end of the next year, the daily oil yield per well is only 33.3% of that of the initial period and the yield reduces by 66.7%, which is largely different from the law of the gas drive yield (see Figure 2).
354
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. Variation of injection pressure of well block Shu 101 and well block Shu 8.
Table 1. Index curve test result of well block ShulOl. Test Date
Air Suction Pressure (MPa)
Air Suction Index (t/d. MPa)
December 2007
17.4
115.2
April 2008
18.0
79.0
November 2008
17.4
42.3
March 2009
17.5
41.5
March 2010
17.2
42.4
18.2.2 The Stratum Pressure Status Based on the stratum pressure status, if the advance gas injection is properly done, the oil well will become effective earlier and have higher natural productivity. The gas-injection well is injected with the liquid C 0 2 of 2531 tons based on the advance gas injection of 6 months, which is with the averagely-injected HCPV times of 0.021 per well. In case the oil
Figure 2. Comparison between production status of well block Shu 101 and well block Shu 8.
I
H M W Oí
^ *-(
o
I-
w1
l—i
»Tí
2
O
<
I—I
o
IsJ
n o
TI
r O
w o1
H
n o
% O
M W
o
OPERATION CONTROL OF C 0 2 - D R I V I N G IN FIELD SIT
356
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
well without fracture treatment is put into operation, the average daily oil yield per well in the initial period will be 2.71 and with the oil production strength of 0.28 t/d.m; whereas, in case the adjacent water drive well block with fracture treatment is put into operation, the average daily oil yield per well in the initial period will be 3.3 t with the oil production strength of 0.22 t/d.m. In a word, the initial oil production strength of the gas drive well block is 1.27 times of that of the adjacent water drive well block.
18.2.3
Air Suction Capability of the Oil Layer
With the stronger air suction capability of the oil layer, the oil deposit pressure can keep higher and give conditions for the miscible driving. Viewing from the results of stratum pressure test of six oil wells with fixed-point monitoring, we know that, the average stratum pressure of 2008 (before the oil well is put into operation) was 22.0 MPa, the average stratum pressure of 2009 was 29.0MPa, and the currently average stratum pressure is 30.2 MPa, which is 8.1 MPa higher than the original stratum pressure. In addition, it is 1.2 MPa higher than the 2009' value and 2 MPa lower than the miscible pressure (see Table 2). Through the numerical simulation study, we find out that, the area around the injection well can form the miscible phase. Currently, the farthest miscible radius can reach 169 m. And, the average flow pressure of the middle part of the oil layer of the gas-injection well is 37.0 MPa, which is 4.8 MPa higher than the miscible phase pressure. The average stratum pressure of the gas-injection well is 34.7 MPa, which is 2.5 MPa higher than the miscible phase pressure (see Table 3).
18.2.4
The Different Flow Pressure Control
The different flow pressure control based on the oil well yield, oil production strength, response status, and classified management can effectively control the unidirectional gas onrush and delay the gas channeling. Class I well: It is with better response, daily oil yield per well above 4 t, and oil production strength above 0.35 t/d.m. In addition, the yield of such kind of well always keeps in a higher level or rising trend, which is applied with the high flow
OPERATION CONTROL OF C 0 2 - D R I V I N G IN FIELD SITE
357
Table 2. Comparison of the stratum pressure before and after the oil well is put into operation. Middepth of the Oil Layer(m)
2008.4-5 Stratum Pressure (MPa)
2009.3-7 Stratum Pressure (MPa)
2010.2-6 Stratum Pressure (MPa)
Shu 91-Carbon inclined 18
2201.0
20.8
26.6
24.7
Shu 93-Carbon 16
2177.4
29.2
33.0
31.8
Shu 95-Carbon 13
2207.6
21.0
27.5
30.2
Shu 95-Carbon 14
2188.6
20.7
23.2
27.6
Shu 96-Carbon 12
2206.6
20.5
27.7
32.7
Shu 96-Carbon 16
2198.0
19.7
36.1
34.4
Average
2196.5
22.0
29.0
30.2
Well No.
pressure (10~15 MPa) for production restriction. Currently, there are 6 Class I wells, which account for 37.5% of total wells. Besides, it is with the daily oil yield per well of 5.2t and oil production strength of 0.51 t/d.m. Class II well: It is with moderate response, daily oil yield per well of 1~4 t, and oil production strength about 0.15-0.35 t/d.m. In addition, the yield of such kind of well are always kept stable, with the flow pressure of 7 to 10 MPa. Currently, there are 5 Class II wells, which are with the daily oil yield per well of 1.4 t and oil production strength of 0.15 t/d.m. Class III well: It is with poor response, Daily oil yield per well is less than It and oil production strength is less than 0.15 t/d.m. In addition, the flow pressure of such kind of well always keeps at 5 to 7 MPa. Currently, there are 5 Class III wells, which are with the daily oil yield per well of 0.7 t and oil production strength of 0.09 t/d.m.
358
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 3. Fitting result of miscible radius of well block Shu 101. Miscible Radius (m)
Well No. YI6
YII4 1
YII4 2
Shu 92-Carbon 17
76
69
Shu 94-Carbon 14
148
151
Shu 94-Carbon 15
124
126
160
Shu 94-Carbon 16
152
101
169
Shu 96-Carbon 13
135
135
Shu 96-Carbon 14 Shu 96-Carbon 15
100
169
135
135
152
In addition, in order to control the rising speed of the gas-oil ratio of the well with gas breakthrough, the periodical oil production mode will be applied. Based on C 0 2 content in the output gas, the different startup and shut-in periods will be applied. With regard to the well with gas breakthrough with the C 0 2 content larger than 50%, the exploration will be controlled by the high flow pressure, flowing oil production, and periodical oil production mode with 15 d startup period and 15 d shut-in period. With regard to the well with gas breakthrough with the C 0 2 content less than 50%, the periodical oil production mode with 20 d startup period and 10 d shut-in period will be applied.
18.2.5
Oil Well with Poor Response
The oil well with poor response can be connected with the superior oil well for the C 0 2 breakthrough reconstruction to improve the response degree. In order to improve the response degree of the oil well with poor response, the Shu 93-Carbon 15 wells with good connection shall be selected for the C 0 2 breakthrough reconstruction. Currently, the oil production of this well before breakthrough is only 0.31, which will become 4.01 after breakthrough. Currently, the daily oil yield keeps at 2.4 t, with the cumulative oil increment of 733 t.
OPERATION CONTROL OF C 0 2 - D R I V I N G IN FIELD SITE
359
18.3 Conclusions 1. The reservoir bed with the air permeability about 1 millidarcy and difficult exploration can timely supply energy to the oil layer with the C0 2 -driving to keep a stable yield of the oil well; 2. Based on the stratum pressure status, if the advance gas injection is properly done, the oil well will become effective earlier have higher yield without reconstruction; 3. Taking advantage of the proper gas injection strength, the oil deposit pressure can keep at a higher level and give conditions for the local miscible phase; 4. The different flow pressure control based on the oil well yield, oil production strength, response status, and classified management can effectively control the unidirectional gas onrush and delay the gas channeling; 5. The oil well with poor response can be connected with the superior oil well for the C 0 2 breakthrough reconstruction to improve the response degree.
References 1. Guo Wankui and others, Recovery Ratio Improvement Technology by the Gas Injection. Beijing: Petroleum Industry Press, 2003 2. Zhang Chuanru and others, Co2 Gas Well Test and Evaluation Method. Beijing: Petroleum Industry Press, 1999 3. Shen Pingping and Liao Weixin, Co2 Geological Storage and Petroleum Recovery Ratio Improvement Technology. Beijing: Petroleum Industry Press, 2009 4. Deng Ruijian and others, Oil Production Technology of Low Permeable Oil Deposit with Hydrocarbon Gas Injection. Beijing: Petroleum Industry Press, 2003 5. Liu Yijiang and others, Polymer and C02-driving Technology. Beijing: China Petrochemical Press, 2001
This page intentionally left blank
19 Application of Heteropolysaccharide in Acid Gas Injection Jie Zhang1, Gang Guo2 and Shugang Li3 College of Chemistry and Chemical Engineering, Xi'an Shiyou University, Xi'an,People's Republic of China 2 Changqing Oil Field Company, PetroChina, Xi'an, People's Republic of China 3 China National Offshore Oil Company, Tianjin, People's Republic of China
Abstract
Hetropolysaccharides (HPS) are an environmentally friendly class of chemicals that have several properties that make them useful for oilfield applications. These include their effect on the surface tension and their ability to reduce swelling in clays. These properties make them particularly useful for enhancing processes related to gas injection such as acid gas injection. This paper presents some laboratory results for the properties of these chemicals.
19.1
Introduction
Polysaccharides are polymers composed of saccharide (sugar) monomer units. Two common polysaccharides are starch and cellulose. Unlike common polymers, including polysaccharides, heteropolysaccharides are composed of different monomers. Unit saccharide monomers are shown in Fiure 1. The HPS of interest here have molecular weights in the range 500xl0 3 to 1200xl0 3 g/mol. The smaller polymers (less than about lOOOxlO3 g/mol) are water soluble. The intermediate sized ones form gels whereas the larger ones are only sparingly soluble in water. Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (361-374) © Scrivener Publishing LLC
361
362
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. Typical saccharide monomers.
These polymers have several properties that make them beneficial as oilfield chemicals: 1. They reduce clay swelling, 2. Reduce interfacial tension, and 3. Absorb acid gases. In addition, they are environmentally friendly chemicals first because they are natural products and second because they are biodegradable. The combination of reduced clay swelling and and surface tension effects means that these chemicals can be used to improve the injection process, bth for acid gas injection and for C 0 2 injection for enhanced oil recovery. Many natural gas corporations which produce natural gas with abundant H2S and C 0 2 in it adopt a new technique to treat the waste gas. Because the primary component is H2S and C0 2 , we called the waste gas "acid gas". The process of acid gas treatment is called acid gas reinjection system, which mainly concern with compression, transportation and injection into the subsurface reservoir. At present, the technique of acid gas reinjection is becoming a practicable way to solve the recycling of sulphur and air pollution by C 0 2 in North America. During the process of actual production, the acid gas device shows more practical valuable when treat with a small quantity of acid gas (less than 150xl0 3 m 3 /d). There are nearly 50 acid gas reinjection systems in western of Canada, and 20 have been used in America [1]. It has begun to field test in many oil fields in China, such as Xinjiang, Jiangsu, central Plains, Daqing, Shengli and so on, mainly on C 0 2 miscible phase recovery [2]. When the sulphur market is depressed, the acid gas
HETEROPOLYSACCHARIDE IN ACID GAS INJECTION
363
reinjection is also a way to treat the acid gas by big natural gas corporations, and avoid sulphur-overstocking. As the growing awareness of environment protection, it becomes a problem that how to treat a small amount of acid gas. The producer cannot discharge the acid gas into the atmosphere like before, instead, compressing and injecting it into the non-productive formation becomes a selectable method. Recently, people started to research the value of making the compressed acid gas as part of gas phase recovery, while, acid gas reinjection is also a green way to reduce greenhouse gases, which is more meaningful in the Kyoto Protocol. While the processing of acid gas separation, the main by-product is acid gas currents which contains H2S and C O r If we do not consider the water content of acid gas currents, the mol fraction of H2S and C 0 2 can overtake 95%. While researching the new environment-friendly carbohydrate oil field chemicals, we found out that amylum was the basic material of natural or modified polysaccharose. The products, glucoside or its derivatives, were all derivatives of polysaccharose. Because of the simple structure, the products usually showed low performance and poor acid resistance. Through field application in recent years, heteropolysaccharide, which contains so many advantages such as low chemical activity, high shale stability, better stability of high temperature, low toxicity and biodegradability, received good environmental, economic and social benefit. Because the acid gas reinjection working fluids must contain these advantages, the research of heteropoly-saccharide in acid gas reinjection is more valuable in industry application.
19.2 Application of Heteropolysaccharide in C0 2 Reinjection Miscible Phase Recovery As a way to enhance recovery, the C 0 2 miscible/immiscible phase recovery has been widely used in these countries which are resourceful in natural C 0 2 and received good economic benefit. It is reported that the oil recovery fraction can be increased by 10%-15%. The technology of C 0 2 alternating injection is the first thing that most oil fields must consider. As a common method, it is used to control C 0 2 fluidity and avoid C 0 2 breakthrough too early. While using water alternating gas recovery which can reduce the interfacial tension and enlarge the area, the gas slug must be more than
364
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
0.1 pore volume (PV) to make it local mixed phase. The technology of gas alternating water is required to treat the clay stability when inject water. The following tow experiments turned out that the heteropolysaccharide showed good performance of it.
19.2.1 Test of Clay Polar Expansion Rate 29.2.2.2
Test Method
According to the "SY/T 6335-1997 Evaluation Procedure of Drilling Fluids Shale Inhibitor": Preparation of bentonite sample: Drying the bentonite of 100 meshes for 2-3 h at 105°C and put it into the dryer, keep in room temperature for 20 min, weight 10g sample and put it into the testing cylinder with API filter paper, remove it after pressuring as 4 MPa for 5 min. Measure the height of the sample and record as initial (H/mm). Fill in the solution or collosol into the cylinder, measure the expansion data and record the line expansion (R t /mm) at 2 h and 16 h, we can get the expansion rate (Vt) divide the two data by initial height. That is to say V t =R t /Hxl00%. Measure the expansion rate of the inhibitor and solution of heteropolysaccharide.
Figure 2. The relationship between time and polar expansion of heteropolysaccharide solution with different concentration.
3% FS-2
6% FS-2
9% FS-2
3% FS-3
6% FS-3
9% FS-3
3% FP-2
6% FP-2
2
3
4
5
6
7
8
9
Note: The soak time of clay sample is 8h.
43.90
up
1
55.87
34.79
55.37
53.72
37.36
51.07
47.11
37.93
Polar Expansion Rate /%
Inhibitor Partitioning
Number
18
17
16
15
14
13
12
11
10
Number
Table 1. The polar expansion rate of shale with different inhibitor.
2%SJ
1%SJ+1%HPAN
9% FP-2+0.2%TIPA
6% FP-2+0.2%TIPA
3% FP-2+0.2%TIPA
18% Liquid sodium silicate
12% Liquid sodium silicate
6% Liquid sodium silicate
9% FP-2
Inhibitor Partitioning
11.56
25.15
58.84
56.28
44.55
56.61
56.45
51.65
53.31
Polar Expansion Rate 1%
HETEROPOLYSACCHARIDE IN ACID GAS INJECTIO
366
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
29.2.2.2
Testing results as the Figure 2 and Table 1 shows
From Figure 3, the shale expansion rate of heteropolysaccharide SJ is relatively less compared with most inhibitors. Thus, heteropolysaccharide SJ water-soluble glue solution has strong effects on inhibiting shale hydration swelling. Its mechanism comprises two aspects. The first one: there is much ortho cycloalcohol hydroxyl on molecular chain of heteropolysaccharide SJ. When the addition of heteropolysaccharide SJ in water-based drilling fluid reach a certain amount, the positive part of polar hydrated cycloalcohol hydroxyl can be adhered on the electronegative mud shale surface and form a layer of unintermittent semi-permeable membrane, which stop free water molecules in heteropolysaccharide drilling fluid from moving to the surface of mud shale. Thus, hydration reaction on the surface with shale from the outside to the inside was effectively prevented. The second one: soft colloidal particles in heteropolysaccharide drilling fluid can fill pore or crack of mud cake and make them denser, and then lower filtration of drilling fluids and prevent clay-hydrated dispersion of mud shale, caused by invasion of water molecular. From Figure 2, the shale expanded fast at first and near to steadiness after 7.5h. From the table 1, the shale expansion rate of heteropolysaccharide is less than other inhibitors; it shows that heteropolysaccharide has good shale expansion rejection capability.
Figure 3. Shale Expansion rate of heteropolysaccharide SJ and other common inhibitors (8 hours).
HETEROPOLYSACCHARIDE IN ACID GAS INJECTION
367
The active mechanism is as follows: (1) there are multiple hydrophilic hydroxyls on heteropolysaccharide molecular, it can adsorb onto the surface of the bentonite sample and generate a very dense semi-permeable diaphragm. It also can combine with the hydrone and generate hydrogen bond to reduce the content of free water. (2) Suspensoid, which is in the heteropolysaccharide collosol, is a good bridging agent which can block off crack and hole of bentonite sample. Under the conditions without bentonite and diversion agents, the heteropolysaccharide collosol can generate a dense mud cake fast to prevent filtrate and solid phase flowing into the bentonite sample. In this way, the hydration of bentonite sample can be reduced and protect the reservoir. 19.2.2
Test of Water A b s o r p t i o n of M u d Ball in Heteropolysaccharide Collosol
At room temperature, mix the natrium bentonite and distilled water as proportion 2:1 and make it into mud ball as 10 grams per singleton, put them into the heteropolysaccharide collosol or other inhibitors which are different concentration but the same volume for 72h respectively. Watch the mud ball and weight it at definite time, compare the mud ball before and after, shows in Figure 4 and Table 2. Draw relationship of concentration between water absorption of the mud ball and heteropolysaccharide collosol and compare with it in the solution or collosol of oil field inhibitors, shows in Figures 5 and 6. Table 2. The describe after mud ball absorbs water. Inhibitor Concentration
Water Absorption/% Time of Water Absorption /H
Appearance of Mud Ball
12
24
48
0.5%YZ-9
86.71
118.36
175.17
deep crack
0.5%YZ-10
50.42
90.69
130.90
microcracks
5.0%YZ-4
20.24
5.0%KC1
0.70
0.72
0.79
rough no crack
1.0% SJ
5.56
7.22
7.85
smooth and complete
spread out, cannot weigh
spread
368
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 4. Appearance of mud ball in heteropolysaccharide between before and after.
Figure 5. The relationship between soak time and water absorption of heteropolysaccharide solution with different concentration.
Figure 6. The water absorption of different clay stabilizer solution.
HETEROPOLYSACCHARIDE IN ACID GAS INJECTION
369
From Figure 2 1, the water absorption rate of inhibitors increased with the increasing time. The water absorption of mud ball is low and the appearance is neat after soaking in the heteropolysaccharide collosol. In Figure 5, the water absorption rate increased with the increasing soaking time of mud ball. After 39h, it is near to steadiness. When the concentration of heteropolysaccharide is more than 0.6%, the water absorption rate is the same to 0.6%. Form the experiment we have found out that after soaking in pure water and heteropolysaccharide collosol with different concentration for 72h, the shape of mud balls had changed different degrees: in the pure water, the volume of mud ball is rapidly expanded, cracked and broken, which already cannot be weighed, in the heteropolysaccharide collosol, the surface of the mud ball was smooth, no cracks and the volume had shrunk a bit, shows in Figure 3. So we could find out that heteropolysaccharide collosol has the function of semi-permeable or even non-permeable, meanwhile with the function of dehydration. In summary of the two experiments: 1. The active mechanism of heteropolysaccharide solution clay stabilization: between the cyclic alcoholichydroxyl group/ Glycosidic b o n d / a small amount of carboxyl of the heteropolysaccharide molecular and the Silicon atom of the mud ball, there is a net structure formed by the chemical bond of Si—O—Si or RO—Si. This net structure wrapped with the mud ball, formed "Silicon Sealing lock shell" on the surface layer of the mud ball. The shell prevented the water molecules into the mud ball then inhibited the clay dispersion and further hydration. At the same time, the shell also had the ability to win water molecules of some mud balls and then showed the role of "Silicon Sealing lock hydration shell to dehydration". 2. The heteropolysaccharide water-based working fluids showed a good inhibition of shale hydration expansion. Its mechanism consists of the following three aspects: "continuous semi-permeable membrane", "soft colloid particles filling and then fluid loss" and "Silicon Sealing lock shell to hydration". Therefore, the heteropolysaccharide can use as a clay stabilizer in the C 0 2 miscible phase recovery.
370
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
19.3
Application of Heteropolysaccharide in H 2 S Reinjection formation
The primary component of the acid gas is H2S except for the C0 2 . The H2S in the acid gas is devastating to the environment and the creature. It is a poisonous gas and when the concentration in the air is beyond lOOOmg/m3, it may cause acute poisoning and death of people. At present, there are many technologies used to separate the harmful components like hydrocarbon compounds and hydrogen sulphide. Many chemicals, among which most of them are similar to organic compounds called hydramine, can be used to separate H2S and C 0 2 in the natural gas. The result of our experiments showed that heteropolysaccharide also had the function to absorb H2S to some degree. So it can get good performance when the acid gas was reinjected.
19.3.1 Experiment Process, Method and Instruction 19.3.1.1
Experiment Process
After H2S created by the gas plant and gone through the gas buffer bottle, kept filling with water-based working fluids with H2S absorption and reacted sufficiently, then made the residual gas go through the tail gas absorption plant with NaCN solution and a certain acetic acid solution in proper order. Using iodimetry to titrate the S2" of zinc acetate solution in the same time separation. Evaluate the absorption effect of tail gas absorption plant by this way. At the same time, when making sure a fixed reacting time, stop the reaction. Add a certain amount of saturated NaOH solution to neutralize superfluous acid so that we can make sure the reaction is stopped. Then keep filling with nitrogen for 5 minutes to expel the H2S which remain in the device into tail gas absorption plant. Last, collect some working fluids sample which absorbed H2S and dilute it so that we can measure the contents of the S2" in water-based working fluid to evaluate the effect of the absorbent agents when it is under a static condition. The block diagram for the experimental procedure is given in Figure 7. 19.3.1.2
Experimen t Method
After H2S created by the gas plant, kept filling with working fluids with absorbent agents for 60 minutes, collect a certain working
sc lution
saturated
NaOH
•L
Gas buffer bottle
Sampling
Water-based working solution
Dilution
1 tail gas absorption
2 tail gas absorption
Figure 7. Static processing of H2S absorption by water-based working fluid. Experimental design.
Nitrogen bottle
,i
Occurrence of hydrogen sulfide gas bottle Exhaust testing
working solution
— ►
Concen tration ofH 2 S
HETEROPOLYSACCHARIDE IN ACID GAS INJECTION
372
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
fluids sample and dilute it, using iodimetry to measure the content of S2" in the drilling fluid. Then use the result to evaluate the absorbing effect of H2S absorbent agents. 193.1.2
Experiment
Results
The experiment results showed that phenolic compound and heteropolysaccharide collosol have marked function of absorbing the H2S. When the two combined in a proper proportion, the function will be strengthened markedly and the admixture can be Table 3. Results of H2S absorption by absorbent agent. pH
Measurement of Absorbent Agent
Conductivity / x 104 (fis-cnv1)
Average Content of in S2 Absorption Agent / (mg-L1)
Before
After
Before
After
Bentonite 4%
11.72
8.23
0.33
0.35
687.52
Phenol-1 0.6%
10.56
8.06
0.35
0.38
1385.01
Phenol-2 0.6%
10.48
7.98
0.34
0.4
1596.68
Phenol-3 0.6%
10.34
7.37
0.33
0.38
1527.6
Phenol-4 0.6%
10.62
7.01
0.35
0.39
1554.34
Phenol-5 0.6%
10.91
7.95
0.36
0.42
1628.26
Phenol-6 0.6%
10.34
8.08
0.32
0.34
871.64
Phenol-7 0.6%
10.26
8.55
0.32
0.35
902.75
Phenol-8 0.6%
10.33
8.94
0.31
0.33
792.45
Basic Zinc Carbonate 0.6%
12.11
9.71
0.634
0.585
2010.22
H,0
632.42
Heteropolysaccharide 2.0%
1223.43
Phenol+Heteropolysaccharide
1967.24
Phenol+Heteropolysaccharide+Basic Zinc Carbonate
2106.84
HETEROPOLYSACCHARIDE IN ACID GAS INJECTION
373
used as a new H2S absorbent agent in wet acid gas reinjection. The experimental results are summarized in Table 3.
19.4 Conclusions 1. The heteropolysaccharide can be used as a clay stabilizer of C 0 2 miscible phase recovery the mechanisms of action of it include three aspects that "continued semi-permeable diaphragm", soft "colloid particles filling and then fluid loss" and "Silicon Sealing lock shell to hydration". 2. The combination of the product formula of heteropolysaccharide and phenolic compound has good effect of absorbing H2S and can be used as absorbent agent of H2S in the process of wet acid gas reinjection.
References 1. J.J. Carroll, Wang Shouxi, and Tang Lin, "Acid gas injection: Another approach of acid gas treatment", Natural Gas Industry, 2009, vol.29, No. 10, pp. 96-100. 2. Li Mengtao,Zhang Hao,Liu Xiangui, "Chemical mechanism of C 0 2 flooding Study ", Chemistry & Bioengineering, 2005, vol. 21, No. 9, pp. 7-9. 3. Sun Lijuan,Wu Fan, "Ultra-low permeability reservoirs the feasibility of gas injection oil recovery experiment ", Henan Petroleum, 2009, vol. 19,No. 3, pp. 38-40. 4. Wang Shouxi, J.J. Carroll, Tang Lin, "Acid gas re-injection of the wellbore flow model and the phase distribution", Natural Gas Industry, 2010, vol. 30, No. 3, pp. 95-97. 5. Ding Guo,Gong Xiaoxiong,Hu Qi, "Pu Bei oilfield gas injection oil recovery process technology", Drilling Technology, 2002, vol. 25, No. 6, pp. 42-45. 6. Yan Jienian, "Drilling Huid Technology Studies", Dong Ying: China Petroleum University Press, 2001. 7. Zhang Jie, Yang Hewei, "Compatibility evaluation between polysaccharides and silicate drilling fluids", Natural Gas Industry, 2009, vol. 29, No. 3, pp. 71-73. 8. Lin Xibin, Sun Jinsheng, Su Yinao, "The study and application of semipermeable diaphragm water base drilling fluid", Drilling Fluid & Completion Fluid, 2005, vol. 22, No. 6, pp. 5-8. 9. Wang Song, Zeng Ke, Yuan Jianqiang,et al, "Research and Application of Salt-resisting and High Temperature Resisting and Water Base Drilling Fluid System", journal of Oil and Gas Technology, 2006, vol. 28, No. 3, pp. 105-107.
374
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
10. Huang Zhizhong, Yang Yuliang, Ma Shichang,et al, "Study On Water-Based Drilling Fluild Resisting High Temperature", Xinjiang Oil and Gas, 2009, vol. 5, No. 3, pp. 52-54. 11. Jiang Guancheng, Wu Xueshi, Yan Jienian,et al, "Study on the Rheology Property of Water Based Drilling Fluid at High Temperature and High Pressure", Drilling Fluid & Completion Fluid, 1994, vol 1, No. 5, pp. 19-20.
SECTION 5 GEOLOGY AND GEOCHEMISTRY
This page intentionally left blank
20
Impact of S0 2 and NO on Carbonated Rocks Submitted to a Geological Storage of C0 2 : An Experimental Study Stéphane Renard1, Jérôme Sterpenich1, Jacques Pironon1, Aurélien Randi1, Pierre Chiquet2 and Marc Lesearme2 'Nancy-University, CNRS, CREGU, UMR G2R, B.P. 239, F-54506 Vandoeuvre-lès-Nancy, France 2 TOTAL, CSTJF, Avenue Larribau, F-64018 Pau, France
Abstract Geological storage of acid gases in carbonated rocks (deep saline aquifers or oil depleted reservoirs) is one of the solutions studied to limit the emissions of greenhouse gases in the atmosphere This paper is devoted to the study of the reactivity of rocks that could be submitted to C0 2 and annex gases (S02 and NO) during the injection of a C0 2 rich gas in a geological storage. This experimental study focuses on the interactions that take place between carbonate rocks (dolomite and calcite rich) and C0 2 coinjected annex gases. The results, interpreted in terms of petrophysical and chemical impacts of the injected gases, can be used to improve thermodynamic and geochemical modelling.
20.1
Introduction
The C 0 2 capture and geological storage from high emitting sources (coal and gas power plants) is one of a panel of solutions proposed to reduce the global greenhouse gas emissions. Different pre-, postor oxy-combustion capture processes are now available to separate associated gases (SOx, NOx, etc.) and the C0 2 . However, complete
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (377-392) © Scrivener Publishing LLC
377
378
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
purification of C 0 2 is unachievable for cost reasons as well as for C 0 2 surplus of emissions due to the separation processes. By consequence, a non-negligible part of these gases could be co-injected with the C0 2 . Their impact on the chemical stability of reservoir rocks, caprocks and well has to be evaluated before any large scale injection procedure. Physico-chemical transformations could modify mechanical and injectivity properties of the site and possibly alter storage safety. The study presented here is focused on experiments of geochemical interactions between rocks and gases (S0 2 and NO) which could be co-injected with C0 2 . The rocks we studied are carbonate rocks (dolomite and calcite rich) which are some possible analogues of reservoir rocks and cap-rocks. Samples are placed in 1cm3 gold capsules together with saline water (25 NaCl g/1). Gases are hermetically transferred by cold trap into the gold reactors that are sealed by electrical welding and placed in an autoclave during one month at 150°C and 100 bar, which represent geological conditions of a depleted deep reservoir. After experiments, solid samples are observed and analysed with different techniques (SEM, TEM, Raman and XRD). Gases are also collected and analysed by Raman spectrometry whereas the aqueous solution is analysed with ICP-MS, ICP-AES and ionic chromatography. As sampling during experiments wasn't possible, we developed the synthetic fluid inclusions technique to trap and analyse the fluids under experimental conditions. This allows to characterise the different phases and the nature of dissolved species. Mass budgets are established in order to quantify the ratio of mineral transformation. This study shows the first results concerning the mineralogical transformation of rocks and well materials submitted to the chemical action of possible annex gases, NO and S0 2 . The results, interpreted in terms of petrophysical and chemical impacts of the injected gases can be used to improve thermodynamical and geochemical modelling.
20.2 Apparatus and Methods Experiments are performed on natural rock samples in batch conditions during one month at 150°C and 100 bar, which represent realistic conditions in the context of geological storage of
IMPACT OF S0 2 AND NO ON CARBONATED ROCKS
379
C 0 2 into depleted reservoir. The batch reactors are made of gold capsules hermetically welded. Gold is used because of its chemical inertia, and its ability to conduct pressure and temperature (Seyfried et al., 1987). The volume of the reactors is around 2 cm 3 (inner diameter of 0.5 cm for a length of 10 cm). After welding capsules are placed in a pressure vessel of 100 cm 3 heated by a coating device (Figure 1). The pressure is controlled by a hydraulic p u m p . The device is presented in more details in Jacquemet et al. (2005). It has been routinely employed for several experimental studies under similar pressure and temperature conditions (Landais et al., 1989; Teinturier and Pironon, 2003; - Jacquemet et al., 2005) mimicking geological environments. Mass balances are established after experiment using analytical characterization of each phase.
20.2.1 Solids and Aqueous Solution The rock samples come from cores drilled in the Aquitania basin (France) in a fractured Portlandian dolomite, namely the Mano Dolostone, and in Early Cretaceous limestones, namely the Campanian Flysch. They were sampled respectively at 4580 m and 4500 m deep and were previously analyzed by Renard (2010.) using Scanning Electron Microscopy (SEM), Electron Probe Micro Analysis (EPMA) and Transmission Electron Microscopy (TEM). The sample of Mano Dolostone is made of a dolomitic matrix crossed by a fracture filled with Fe-dolomite and with a thin layer of calcite. For the experiments, we selected samples containing both facieses separated according to a ~ 20 um-thick layer of calcite. The Campanian Flysch is mainly calcitic. The fracture of the Mano Dolostone is made of 93% Fe-dolomite (CaMg)xFe2 x (C0 3 ) 2 , 5% calcite CaC0 3 and 2% dolomite CaMg(C0 3 ) 2 . The matrix of the Mano Dolostone contains 92.2% dolomite, 4.2% illite Si343Al226Fe0 06 Mg 024 K 071 Na 007 Ca 002 and interstratified illite /smectite, 3% quartz Si0 2 , 0.5% pyrite FeS2 and 0.1% calcite. The Campanian Flysch is made of 63.2% calcite, 10.5% quartz Si0 2 , 8.3% illite Si342Al218Fe02Mg02K07Ca005, 6.5% interstratified chlorite/smectite, 4.5% chlorite Si256 AÍ27Fe356 Mg 127 , 4.6% ankerite F e ^ C a ^ M n ) , x C0 3 ,2.1 % dolomite, 0.3% pyrite. The rock samples were cut into stick fragments of around 10 mm x 2 mm x 2 mm. They were then polished on one face in order to
380
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 1. Quantities expressed in mg of rock, aqueous solution and gas used for the experiments on the reservoir rock and caprock. Solution (mg) Gas (mg)
Mass
Rock (mg)
N2
Reservoir rock
127
434
38
N2
Caprock
145
550
40
so 2 so 2
Reservoir rock
103
510
423
Caprock
130
630
220
NO
Reservoir rock
130
520
160
NO
Caprock
135
510
155
Experiment
better detect the mineralogical changes (dissolution or precipitation) on the surface. We partially filled the gold capsules with a 25g/l NaCl brine. The water/rock and water/gas mass ratios were respectively about 3 and 5 as specified in Table 1. For each experiment, a decrepited quartz was added to the system in order to trap the fluids during the experiment in synthetic fluid inclusions. At the end of experiment, gold capsules were opened to collect the gas phase, the aqueous solution and the minerals for analyses.
20.2.2
Gases
The two different types of gases selected for experiments are S0 2 and NO. A blank capsule containing the same phases (aqueous solution and solid) was filled with N 2 as an inert gas phase. The injected quantities for each experiment are displayed in Table 1. The gases are loaded in the capsules using the gas loading device adapted from Jacquemet et al., 2005 (Figure 1). During the loading procedure, the gold capsules are hermetically fixed on the capsule connector which is plugged to the loading device through the valve E. Knowing the volume of the loading line and controlling the pressure in the system, it is possible to fill the reactor with a known mass of gas thanks to a nitrogen cold trap. After experiment, cold capsules are pierced in an appropriate device plugged
IMPACT OF S0 2 AND N O ON CARBONATED ROCKS
381
Figure 1. Gas loading and sampling line used during the experimental phase, adapted from Jacquemet et al. (2005). (A-E) valves. Different devices can be connected to the line: a capsule piercing device used to collect gases after experiment, a capsule loading device used to trap gases in the capsule and a cell for the Raman analysis of the gases.
to valve C. After trapping, the gas can be driven to a Raman cell for analysis.
20.3
Results and Discussion
This section is devoted to the description of the mineral changes observed from the solid samples of reservoir and caprock aged with N 2 (blank experiments) S0 2 and NO during one month at 150°C and 100 bar.
20.3.1 Reactivity of the Blank Experiments After experiment, the samples of the reservoir rock do not present any visible transformation except a slight frosted aspect of the
382
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
initially polished face. SEM observations (Figure 2) show that the frosted aspect is due to a slight dissolution of the carbonate phases, the dolomite of the matrix and the calcite of the fracture. However the dolomite of the fracture does not seem to have undergone any significant dissolution. Pyrite and quartz keep unaltered whereas the analyses of the clay fraction (Renard 2010) show a partial leaching of Na and Ca cations. Concerning the caprock, the optical observations show a slight frosted aspect as well as the presence of a brown-orange colour on the surface. The limited reactivity is observable with SEM (Figure 3). The surface of the calcite is slightly dissolved. The grains of quartz and the framboidal pyrites seem to be unaltered. However EDS (Energy Dispersive Spectrometry) analyses show that the surface of the pyrites is oxidised explaining the brownish aspect of the sample. Clay minerals analysed by TEM before and after experiment do not react significantly during experiment.
Figure 2. SEM backscattered images of the reservoir rock sample after experiment with N 2 and saline water (25 g/1). (I) Global view of the matrix (Ma) and the fracture (Fr), (II) zoom on the matrix, (III) zoom on close to the wall of the fracture, (cal) calcite, (dol) dolomite, (Py) pyrite.
IMPACT OF S0 2 AND NO ON CARBONATED ROCKS
383
Figure 3. SEM backscattered images of the caprock sample after experiment with N 2 and saline water (25 g/1). (I) Global view of the sample, (II-III) zoom on a zoned siderite, (IV) zoom on a pyrite rich zone, (ag) clay minerals, (cal) calcite, (dol) dolomite, (qtz) quartz, (sd) siderite, (py) pyrite.
The blank experiments both with the caprock and the reservoir show a very limited reactivity of the minerals corresponding to the equilibration between the initial aqueous solution and the different minerals. The pH of the solution is rapidly buffered by carbonate minerals (dolomite and calcite). The main chemical reactions considered during the experiment that can affect the pH as well as the elemental concentrations of the solution are: C a M g ( C 0 3 ) 2 + 2H + = Ca 2+ + Mg 2 + + 2 H C 0 3
(1)
C a C Q 3 + H + = Ca 2+ + H C 0 3
(2)
(CaMg) 0 ; 1 3 Fe 0 , 7 4 CO 3 +H + = 0,13 Ca 2 + + 0,13 Mg 2 + + 0,74 Fe 2+ + H C 0 3
(3)
384
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Q u a r t z = S i 0 2 aq
(4)
A m o r p h o u s silica = S i 0 2 aq + n H 2 0
(5)
HCOj = C 0 2 g + H+
(6)
H 2 0 = H + + OH"
(7)
During the loading of the reactors, gaseous oxygen can be trapped leading to a partial oxidation of the reduced mineral such as pyrite. This phenomenon, enhanced by the framboidal shape of the mineral increasing its reactive surface area, can be resumed by the following chemical reaction leading to the formation of hematite (Fe203) and sulfates (mainly anhydrite CaS0 4 ). 2 FeS 2 + 4 H 2 0 + 7.5 0 2 = F e 2 0 3 + 4 S0 4 2 " + 8 H +
(8)
The mass balance calculated from these blank experiments confirm that the mineral dissolution is very limited with less than 5% of the initial quantity of the minerals affected by the mineral transformations. Calcite and pyrite seem to be the most sensitive minerals in our experimental conditions. 20.3.2
R e a c t i v i t y w i t h Pure S 0 2
The initial reservoir and caprock samples were completely crumbled after the experiment with S0 2 . After drying, a powder made of fibrous crystals of anhydrite and amorphous native sulphur was observed in association with an amorphous silica rich phase (Figure 4, Figure 5) containing iron sulphur, aluminium and potassium. Quartz and pyrite couldn't be detected. Large amounts of C 0 2 were released in the gas phase as a proof of the high reactivity of the carbonates towards S0 2 . Under experimental pressure, temperature and water molar ratio, respectively 100 bar, 150°C and 0.6 to 0.9, the S0 2 -H 2 0 system is monophasic with a complete dissolution of the S0 2 in the liquid water (Van Berkum et al., 1979). The effect of NaCl is not documented under the experimental conditions but the synthetic fluid
IMPACT OF S0 2 AND N O ON CARBONATED ROCKS
385
Figure 4. SEM backscattered images of the reservoir rock sample after experiment with S0 2 and saline water (25 g/1). (I, III) global view; (II) zoom on native sulfur; (IV) zoom on a zone containing native sulfur, anhydrite and amorphous silica. (S0) native sulfur, (Anh) anhydrite, (Si^) amorphous silica rich phase.
Figure 5. SEM backscattered images of the caprock sample after experiment with S0 2 and saline water (25 g/1). (S0) native sulfur, (Anh) anhydrite, (Siam) amorphous silica rich phase.
386
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
inclusions analyses show that no gaseous S0 2 was trapped during the experiments implying its quasi- total dissolution in the saline water. When S0 2 is in solution, a reaction of disproportionation occurs leading to sulphuric acid and native sulphur according to the reaction: 2 H 2 0 + 3 S0 2 / a q = 2 H + + S O / " + 1 / 8S 8
(9)
This reaction accounts for the presence of native sulphur after experiment as well as the strong alteration of minerals due to high acidic conditions. The main mineral transformations can be sum u p by the following reactions involving carbonate minerals: Dolomite + S0 4 2 " + Ca 2+ + 4 H + = 2 Anhydrite + 2 C 0 2
aq
+ Mg 2 +
(10)
The clay minerals are also strongly affected by the high acidity of the solution. They dissolved to give mainly Si and Al in solution that can combine to form an amorphous phase called amorphous silica rich phase. This gel can incorporate sulfur and a part of the alkalis and alkaline-earth elements coming from the dissolution of carbonates and silicates. If we consider muscovite as a proxy for clay minerals, the reaction could be expressed as follows: Si 3 Al 3 KO 1 0 (OH) 2 + i 0 H + = 3 S i 0 2 am + 3 A1 3+ + K + + 6 H 2 0
(11)
Pyrite is also concerned both by the acidic attack and the oxidizing power of S0 2 . Pyrite is thus transformed by the following reaction enhanced in acidic conditions: 4 H + + S 0 2 a q + 2 FeS 2 = 2 Fe 2+ + 5 / 8 S 8 + 2 H 2 0
(12)
The dissolution of clay minerals, especially ilutes, can release Fe3+ but the presence of S0 2 , as a reducing compound in this case,
IMPACT OF S0 2 AND NO ON CARBONATED ROCKS
387
leads to its reduction in Fe 2+ in agreement with Palandri et al. (2005) according to: 2 Fe 3+ + S 0 2 aq + 2 H 2 0 = 2 Fe 2+ + H S O ¡ + 3 H +
(13)
Thus, the presence of high amounts of S0 2 leads to a total dissolution of carbonates, silicates and pyrite and to the precipitation of anhydrite, native sulfur and an amorphous silica rich phase. The mass budget of the experiment was calculated thanks to the chemistry of the solution, the stoichiometry of the mineral phases and the composition of the gaseous phase (consumed S0 2 and produced C0 2 ). For both the reservoir and the caprock, the total amount of carbonates disappeared whereas it was the case only for 15 to 20% of the clayey fraction. After reaction, about 15% of the initial S0 2 gave anhydrite, 25% gave native sulfur and less than 1% gave barite (BaS04).
20.3.3
Reactivity with Pure NO
After experiment with NO the caprock and reservoir samples kept their initial shape but showed strong visible transformations on their surface. The matrix dolomite of the reservoir rock sample (Figure 6), disappeared from the surface and was only detectable deeper below the surface. Clay minerals and quartz are still present. The fracture wall calcite is altered, and the dolomite is partially dissolved according to its cleavages. The pyrites of the rock was completely oxidized into hematite. A part of the sulfur coming from the oxidation of the sulfides re-precipitated in anhydrite and in a lesser extent in barite (BaS04) from the calcium of carbonates and the barium as a trace element in the calcites. Concerning the caprock (Figure 7), the calcite was strongly dissolved. Fe-containing minerals (siderite and pyrite) were oxidized leading to the precipitation of hematite. The sulfur from pyrites partially precipitated into anhydrite and barite. The ferriferous chlorites were also oxidized. The observations of both the reservoir and the caprock show two main chemical mechanisms responsible for the mineral transformations: reactions under acidic conditions and oxydo-reduction reactions.
388
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. SEM backscattered images of the reservoir rock sample after experiment with NO and saline water (25 g/1). (I) global view; (II) zoom on the limit between matrix and fracture; (III) zoom on the matrix. (Dol) dolomite, (Anh) anhydrite, (Ba) barite, (Ag) clay minerals.
There are very few thermodynamical data for NO under the experimental pressure and temperature range. The analyses performed onto the gaseous and aqueous phase indicate that NO is not stable under these conditions. The chemistry of nitrogen oxides is complex and numerous phases appear during the experiment such as N 2 0 , N 0 2 , N 2 , 0 2 , NH 4 + , N03~. In the gaseous phase some reactions of oxydo-reduction can run such as: 3 NO = N02 + N 2 0
(14)
2NO =N2+02
(15)
2NO =y2N2+N02
(16)
2NO =y202+N20
(17)
IMPACT OF S0 2 AND NO ON CARBONATED ROCKS
389
Figure 7. SEM backscattered images of the caprock sample after experiment with NO and saline water (25 g/1). (I) global view; (II and III) successive zooms on the matrix. (Ex-Si) ex-siderite transformed in hematite, (Ag) clay minerals.
In the queous phase, the following reaction can explain the presence of N 2 0 and the nitrates and leading to a very acidic solution: 4 N O + y2 H 2 0 = 3 / 2 N 2 0 + H + + N 0 3
(18)
The dissociation of N 2 0 in N 2 and 0 2 was also descibed in the littretaure but under different conditions (Li et al., 1992, Rivallan et al., 2009). Whatever the occuring reactions, the presence of NO in an aqueous system leads to a dual reactivity due to the presence of protons H + and oxidising agents such as 0 2 . In this case, several chemical mechanisms can be written to explain the complete oxidation of iron bearing phases (pyrite and siderite) as well as the presence of ammonium or nitrates in the aqueous solution. The following reactions can be proposed athough they are not exhaustive.
390
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The presence of C 0 2 in the fluid phase after experiment proves that the carbonates phases are altered by the acidic solution due to the initial presence of NO. The same reactions as for the dissolution of carbonates in an acidic solution (reactions 1 to 3) can be also envisaged here: C a C 0 3 + 2H + = Ca 2+ + H 2 0 + C 0 2 C a M g ( C 0 3 ) 2 + 4 H + = Ca 2 + + Mg 2 + + 2 H 2 0 + 2 C 0 2
(19) (20)
The reactions lead to emission of gaseous C 0 2 and the release of Ca2+ and Ba2+. The oxidation of pyrites cans be described differently if considering either the presence of ammonium or di-nitrogen, or considering the oxidising agent to be nitrate or di-oxygen: 8FeS2+31H20 + 15N03 = 4 F e 2 0 3 + 15 NH 4 + + 16 SO 2 " + 2 H +
(21)
2 FeS 2 + H 2 0 + 6 N0 3 " = F e 2 0 3 + 3 N 2 + 4 SO 2 " + 2 H + 2 FeS 2 + 4 H 2 0 + 7,5 0 2 = F e 2 0 3 + 4 S0 4 2 " + 8 H +
(22) (23)
For each reaction used, sulphates are formed that combine with Ca and Ba to form anhydrite and barite: Ca2++S042" = CaS04
(24)
Ba 2+ + S0 4 2 " = B a S 0 4
(25)
Contrary to the experiments with S0 2 , silicate minerals, mainly quartz and clay minerals, are slightly affected by NO with only a few percents dissolved.
IMPACT OF S0 2 AND NO ON CARBONATED ROCKS
391
To sum up, experiments with NO are complex and lead to a complete oxidation of iron bearing phases (mainly pyrite and siderite), to a partial dissolution of carbonates with an enhanced reactivity of calcite by comparison with dolomite, and keep the silicates phases almost free of dissolution. For the chosen conditions of experiment, the mass budget shows that between 20 and 50% percent of the calcite is dissolved as against 15 to 20% of the dolomite. 100% of the siderite and the pyrite are oxidised in hematite. Less than a few percent of the silicates is affected by NO.
20.4
Conclusion
The experiments performed in the context of the injection of C 0 2 and co-injected gases in a geological storage have demonstrated that S0 2 and NO should play a role on the mineralogy of both the reservoir and the caprock. First, this study has shown that S0 2 and NO have a complex behaviour with a dual action, oxidising and acidic, on the minerals. Second, many disproportionation reactions can occur when S0 2 and NO are placed under geological conditions of pressure and temperature. These oxydo-reduction reactions complicate the system by multiplying the possible oxidising agents and thus the possible reactions and products of reactions. Third, the reactivity of both the reservoir rock and the caprock is strongly dependent on the nature of the mineral phases (silicates, carbonates, sulphides, etc.) but also on the nature of the reacting gas. For example it is noticeable that the presence of S0 2 should lead to the formation of sulphate mineral and native sulphur, when the presence of NO should be responsible for the strong oxidation of iron bearing phases. In any case, since the molar volumes of initial minerals are different of those of secondary products (as an example the molar volume of calcite is 36.93 cm3.mol"1 against 45.16 cm3.mol_1 for anhydrite), the minerals transformations occurring with the injection of reacting gases should be interpreted in terms of petrophysical properties (porosity and permeability) of the hosting rock. This study shows also that experiments with the gases of interest under geological conditions of storage are necessary to predict the evolution of the storage submitted to the injection of C 0 2 and co-injected gases.
392
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Acknowledgments This work is supported by TOTAL and ADEME (France). It is included in the project "Gaz Annexes" of the French National Agency for Research (ANR).
References Seyfried, W. E. J., Janecky, D. R., and Berndt, M. E. (1987). "Rocking autoclaves for hydrothermal experiments The flexible reaction-cell system." Hydrothermal experimental techniques (ed. Ulmer, G. C. and Barnes, H. L.), pp. 216-239. John Wiley & Sons. Jacquemet, N., J. Pironon and E. Caroli (2005). "Anew experimental procedure for simulation of H2S + C 0 2 geological storage. Application to well cement aging". Oil and Gas Science and Technology 60(1), pp. 193-206. Landais, P., Michels, R., and Poty, B. (1989) "Pyrolysis of organic matter in coldseal pressure autoclaves. Experimental approach and applications." journal of analytical and applied pyrolysis 16,103-115. Teinturier, S., Pironon, J. (2003). "Synthetic fluid inclusions as recorders of microfracture healing and overgrowth formation rates." American Mineralogist, 88 (8-9), pp. 1204-1208. Renard, S (2010) "Rôle des gaz annexes sur l'évolution géochimique d'un site de stockage de dioxyde de carbone. Application à des réservoirs carbonates." PhD thesis Nancy Université INPL, p. 422. Van Berkum, J.G., Diepen, G.A.M. (1979). "Phase equilibria in S0 2 + H 2 0: the sulfur dioxide gas hydrate, two liquid phases, and the gas phase in the temperature range 273 to 400 K and at pressures up to 400 MPa." The Journal of Chemical Thermodynamics, 11 (4), pp. 317-334. Palandri, J. L., R. J. Rosenbauer and Y. K. Kharaka (2005). "Ferric iron in sediments as a novel C 0 2 mineral trap: C0 2 -S0 2 reaction with hematite." Applied Geochemistry 20(11), pp. 2038-2048. Li, Y, Armor, J.N. (1992). "Catalytic decomposition of nitrous oxide on metal exchanged Zeolites." Applied Catalysis B, Environmental, 1 (3), pp. L21-L29. Rivallan, M., Ricchiardi, G., Bordiga, S., Zecchina, A. (2009). "Adsorption and reactivity of nitrogen oxides (N0 2 , NO, N 2 0) on Fe-zeolites." Journal of Catalysis, 264 (2), pp. 104-116.
21 Geochemical Modeling of Huff 'N' Puff Oil Recovery With C 0 2 at the Northwest Mcgregor Oil Field Yevhen I. Holubnyak, Blaise A.F. Mibeck, Jordan M. Bremer, Steven A. Smith, James A. Sorensen, Charles D. Gorecki, Edward N. Steadman, and John A. Harju Energy & Environmental Research Center, University of North Dakota, Grand Forks, ND, USA
Abstract The huff 'n' puff enhanced oil recovery method was used at the Northwest McGregor oil field in North Dakota as a part of a C 0 2 storage demonstration project. Specifically, 440 tons of supercritical C 0 2 was injected into a well over a 2-day period and allowed to "soak" for a 2-week period. The well was subsequently put back into production to recover incremental oil. This paper outlines the approach and current observations derived from numerical modeling and laboratory simulations of potential geochemical reactions to evaluate the short-term risks for operations (e.g., porosity and permeability decrease) and long-term implications for CÖ 2 storage via mineralization. The integration of data obtained during mineralogical analyses, fluid sampling, and laboratory experiments proved to be key for better understanding of the dynamic geochemical processes that happen in the reservoir after C 0 2 injection and was necessary for successful completion of the numerical modeling.
21.1
Introduction
In recent y e a r s , t h e m a n a g e m e n t of c a r b o n d i o x i d e (C0 2 ) e m i s sions from large i n d u s t r i a l p o i n t sources h a s b e e n identified as
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (393-406) © Scrivener Publishing LLC
393
394
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
a potential means to mitigate global climate change. Efforts to reduce C 0 2 emissions now are a significant focus for energy producers and users, including the general public, governments, industry, regulators, and nongovernmental organizations. Carbon capture and storage in geological media have been identified as important mechanisms for reducing anthropogenic C 0 2 emissions currently vented to the atmosphere. Several geologic settings for geological storage of C 0 2 are available, such as in depleted oil and gas reservoirs, deep saline formations, C 0 2 flood enhanced oil recovery (EOR) operations, and enhanced coalbed methane recovery. The Plains C 0 2 Reduction (PCOR) Partnership has conducted regional characterization activities which indicated that Williston Basin oil fields may have over 1.2 billion barrels of incremental oil that could be produced from C 0 2 EOR operations (Smith et al., 2006). While the C0 2 -based EOR operations at the Weyburn and Midale Fields in Saskatchewan, Canada, are good examples of economically and technically successful injection of C 0 2 for simultaneous EOR and sequestration, the depths of injection and, therefore, reservoir conditions in those fields are relatively shallow (ca. 4600 ft) and not necessarily representative of many large Williston Basin oil fields. One of the primary goals of the PCOR Partnership Phase II Williston Basin Field Validation Test was to evaluate the effectiveness of C 0 2 for EOR and sequestration in carbonate oil fields at depths greater than 8000 ft. To achieve that goal, a C 0 2 huff 'n' puff test was conducted in an oil-producing well from an interval of the Mississippian-age Madison Group at a depth of approximately 8050 ft in the Northwest McGregor oil field in Williams County, North Dakota. The 440 tons of supercritical C 0 2 was injected into a well over a 2-day period and allowed to "soak" for a 2-week period. The well was subsequently put back into production to recover incremental oil. The main purpose of this study is to determine the effects C 0 2 will have on the productivity of the reservoir and the carbonate formation into which C 0 2 was injected. This paper outlines the approach for the numerical modeling and laboratory simulations of potential geochemical reactions and compares them with current field observations in order to evaluate the short-term risks for operations (e.g., porosity and permeability decrease) and long-term implications for C 0 2 storage via mineralization.
HUFF 'N' PUFF O I L RECOVERY W I T H C 0 2
395
21.2 Northwest McGregor Location and Geological Setting The Northwest McGregor oil field is located in Williams County in northwestern North Dakota, approximately 20 miles north of the town of Tioga. The field covers an area of about 30 mi 2 in an area of glaciated prairie uplands. Figure 1 shows the location of the Northwest McGregor oil field within the PCOR Partnership region and the relative locations of the E. Goetz #1 Well, which served as the injection well, and the E.L. Gudvangen #1 Well, which served as a deep observation well, within the Northwest McGregor oil field. Both oil wells are owned and operated by Eagle Operating Company, an independent oil company with headquarters in Kenmare, North Dakota. The Northwest McGregor oil-producing zone is in the Mississippian-age Mission Canyon Formation (Figure 1), which represents deposition of predominantly carbonate sediments and evaporites in environments that ranged from open marine to coastal sabkha or salina (Lindsay, 1988; Kent et al., 1988). The E. Goetz #1 Well was initially drilled in 1963, with production from the Mission Canyon beginning in 1964 and continuing through and beyond the time period of this project.
21.3 The Northwest McGregor Field, E. Goetz #1 Well Operational History The Northwest McGregor oil field began producing oil in the early 1960s. Over the course of its operational lifetime, as of 2009, the Northwest McGregor oil field has produced over 2.2 million barrels of oil from 14 wells. The E. Goetz #1 Well was initially drilled in 1963, with production from the Mission Canyon beginning in 1964 and continuing through and beyond the time period of this project. Table 1 provides data on the initial reservoir conditions of the Northwest McGregor Mission Canyon Reservoir at the E. Goetz #1 location. It is important to note that the matrix permeability of the Mission Canyon Formation at the E. Goetz #1 location is very low (0.35 md), and most of the fluid movement within the reservoir happens in fractures.
396
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. Location of the Northwest McGregor site (red rectangle) within the PCOR Partnership region and the zoomed map view of Northwest McGregor oil field with relative locations of the injection and observation wells.
HUFF 'N' PUFF O I L RECOVERY W I T H C 0 2
397
Table 1. Initial conditions of the mission canyon reservoir of the northwest McGregor oil field. Reservoir Characteristics Producing Formation
Mission Canyon
Lithology
Limestone and Dolostone
Average Pay Thickness
14 ft
Average Porosity
15%
Matrix Permeability
0.35 md
Secondary Permeability
Fractures
Depth from Surface to Pay
8050 ft
Average Temperature
216°F
Original Discovery Reservoir Pressure
3127 psig
Preinjection Reservoir Pressure
2700 psig
Oil Gravity (API)
41.7°
Cumulative Oil Production
2.2 million STB
21.4
Reservoir Mineralogy
Because the Mission Canyon Formation has been one of the most prolific producers of oil in the Nesson Anticline portion of the Williston Basin, it has been the subject of numerous technical papers and academic studies. With respect to the Northwest McGregor Field and its neighboring oil fields, there are bountiful data in well files that are publicly available through the North Dakota Department of Natural Resources. These papers, studies, and well files (including historical well logs) provide a tremendous amount of data regarding lithology, mineralogy, and formation fluid chemistry. However, in order to improve the accuracy of the geochemical modeling, available cuttings, core samples, and current reservoir fluid properties were analyzed. The formation mineralogy, mineral composition, and spatial variations at the Northwest McGregor site were determined using well logs, traditional core sample analysis with x-ray diffraction
398
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
(XRD), x-ray fluorescence (XRF), and QEMSCAN® techniques. All of these techniques have certain advantages and disadvantages. For instance, XRD is usually considered to be a semiquantitative technique and is unable to identify phases below 1 to 5 wt%. Also, if solid solutions are present or amorphous phases exist, it is very difficult to interpret the mineral assemblage. Therefore, an integrative mineralogical analysis was performed utilizing linear program normative analysis (LPNORM; de Caritat et al., 1994). Using the results of these analyses, the mineral phases selected for model inputs were anhydrite, calcite, dolomite, illite, quartz, and traces of pyrite (Figure 2).
21.5
Preinjection and Postinjection Reservoir Fluid Analysis
The composition of the formation water is one of the critical inputs for geochemical modeling. However, the fluid analysis often becomes a very complicated matter because of the changing nature of gases and water at various pressures and temperatures and conditions of thermodynamic equilibrium in a changing environment. Preinjection and postinjection bottomhole samples were collected using Schlumberger's electric-line (E-line) tool and then transferred to Oilphase-DBR. The reservoir fluid and stock tank
Figure 2. Mineralogical composition and an example of a core sample from the E. Goetz #1 Well.
HUFF 'N' PUFF O I L RECOVERY W I T H C 0 2
399
water (STW) properties for the before and after injection samples are presented in Figure 3. The gas from zero flash was subjected to ion chromatography, and its composition was determined for both samples (Figure 4). Other properties such as the physical properties of the STW were calculated and are listed in Table 2. The ion concentrations and other reservoir fluid properties (e.g., pH, ionic strength) were also modeled using PHREEQC and Geochemist's
Figure 3. Extended comparison of preinjection and postinjection reservoir fluid collected using Schlumberger's E-line from the depth of 8087 ft at the E. Goetz #1 Well and analyzed with Oilphase-DBR and adjusted with the geochemical modeling software.
Figure 4. Comparison of preinjection and postinjection reservoir gas compositions from zero-flash and subjected to chromatography from the depth of 8087 ft at the E. Goetz #1 Well.
400
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 2. Comparison of Preinjection and Postinjection Reservoir Fluid Collected Using Schlumberger's E-line from the Depth of 8087 ft from the E. Goetz #1 Well and Analyzed with Oilphase-DBR. pH
Density, g/cm3
Resistivity Qm,at 77°F
Salinity, mg/kg
TDS,* mg/kg
Before Injection
5.55 (at 106 °F) 4.50 (at 216 °F - live ph) 4.23 (modeled)
1200
4.02
283855
273353
After Injection
5.4 (at 106°F) 3.1 (modeled)
1208
4.17
282925
276477
Total dissolved solids.
Workbench software packages and adjusted for correct reservoir pressure and temperature. The Oilphase-DBR live pH measurement technique uses pH-sensitive dyes that change color according to the pH of the formation water. The live water pH technique was applied for the preinjection sample analysis only. On injection of dye into the sample at reservoir pressure and temperature, it was determined that the pH value of the sample is expected to be < 4.5 units at 2600 psia and 225°F.
21.6
Major Observations and the Analysis of the Reservoir Fluid Sampling
The formation water geochemistry in the northern portion of the Williston Basin and at the Northwest McGregor oilfield in particular is characterized as high salinity NaCl brine (TDS > 250,000 mg/kg) (Jensen, 2007). Also, because of operational history and regional water properties, the water at the E. Goetz #1 Well had already low pH, ~ 4.5. The key observations identified by this study are 1) the displacing of the H2S gas by C 0 2 around wellbore; 2) the increase in total dissolved solids because of some mineral dissolution, in particular, the Ca and Sr concentration increase can be explained by the limestone dissolution; and 3) further pH decreases because of C 0 2 dissolution.
HUFF 'N' PUFF O I L RECOVERY W I T H C 0 2
21.7
401
Laboratory Experimentations
A series of laboratory experiments and numerical modeling of geochemical reactions were conducted. Core samples collected from the Mississippian Mission Canyon formation of the Williston Basin (Figure 5) were exposed for a period of 4 weeks to pure supercritical C 0 2 at 2250 psi (155 bar) and 158°F (70°C) in 10 wt% NaCl synthetic brine conditions (Hawthorne et al., 2010; Holubnyak et al., 2010). Prior to exposure, XRD and XRF mineralogical analysis demonstrated the presence of ankerite, anhydrite, calcite, dolomite, halite, illite, pyrite, and quartz. After exposure, mineralogical (XRD and QEMSCAN) and water analysis inductively coupled plasma-mass spectroscopy were also performed. The laboratory observations were later correlated with the field data and numerical modeling (Figure 6). Observations made during the laboratory experiments were in good correlation with field observations and illustrated the dissolution of the carbonate rocks. In addition, insignificant hematite precipitation as a result of iron mobilization was observed (Figure 6).
EERC YH38S42.CDR
Original sample
After exposure to C0 2
Figure 5. This Mississippian Mission Canyon sample was collected from the depth of 8140 ft (2481 m). It was saturated with synthetic NaCl brine and exposed to supercritical C 0 2 at the Northwest McGregor reservoir conditions.
402
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. The Mississippian Mission Canyon sample was saturated with synthetic NaCl brine and exposed to supercritical C 0 2 at the reservoir conditions. Changes in concentration of Ca and Mg are modeled and correlated with field and laboratory observations as shown on the left. Mineralogical changes are shown on the right.
21.8
2-D Reservoir Geochemical Modeling with GEM
The reservoir simulation model was created according to generalized uniform reservoir parameters: pressure of 3000 psi; in situ gas composition of C 0 2 - 12.5%, CH 4 - 47%, H2S - 35.5 %; porosity of 15%; permeability of 35 mD; water saturation of near 1. The permeability of 35 mD was picked to compensate for the movement in fractures, which was not implemented in this exercise for time saving purposes and is planned to be implemented in the next set of calculations. The reservoir thickness was assumed to be 30 ft. The C 0 2 was injected into a grid block that offset the boundary layer by 3 ft. Moreover, this simulation did not account for the C 0 2 production. Considering the many limitations of this model, the simulation run included calcium and dolomite minerals and did not account for hematite precipitation. The time line for the modeling exercise was picked as 10 years based on the preliminary kinetic numerical modeling with PHREEQC and Geochemist's Workbench. The distribution of the H+ ion within the formation repeats the distribution of C 0 2 plume in the reservoir and outlines the margins where the most significant mineralogical changes are predicted (Figure 7). The dissolution of carbonate minerals was illustrated, and as a result of dissolution, the increase in porosity was modeled (Figure 8).
HUFF 'N' PUFF O I L RECOVERY W I T H C 0 2
403
Figure 7. Spatial 2-D distribution of H+ in the reservoir: 1) 30 days after C 0 2 injection shut-in, 2) after 1 year, and 3) 10 years after the injection.
Figure 8. Spatial 2-D distribution of the calcite and dolomite dissolution, and insignificant porosity increase modeled 10 years after the injection.
21.9
Summary and Conclusions
The integrated investigation of field and laboratory data and numerical modeling exercises revealed that no significant changes in reservoir geochemistry have occurred. The small porosity increase might have contributed to the improved oil production from the E. Goetz #1 Well, though the magnitude of that contribution is open to speculation. Laboratory studies and numerical modeling suggests that C 0 2 trapping by mineralogical processes is minimal for the Northwest McGregor oil field EOR case. The high concentration of salts in the formation fluid and the already
404
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
very acidic environment of the Mission Canyon reservoir are likely the primary factors that contribute to the minimal geochemical response of the Northwest McGregor reservoir to the injected C0 2 . However, the full-scale reservoir geochemical modeling is the next logical step in order to determine the effects of C 0 2 movement in fracture-dominated carbonate reservoirs. Also, the precipitation of the iron-bearing minerals needs to be included in future modeling.
21.10
Acknowledgments
The authors would like to acknowledge the U.S. Department of Energy National Energy Technology Laboratory, Eagle Operating, Computer Modelling Group, Schlumberger, the North Dakota Geological Survey, and all PCOR Partnership partners for their input and support. This material is based upon work supported by the Department of Energy National Energy Technology Laboratory under Award Number DE-FC26-05NT42592.
21.11 Disclaimer This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
HUFF 'N' PUFF O I L RECOVERY W I T H C 0 2
405
References de Caritat, P.J. Bloch, and I. Hutcheon, 1994, LPNORM: "A linear programming normative analysis code." Computers and Geosciences, v. 20,313-341. Hawthorne, S.B., D.J. Miller, Y Holubnyak, B.G. Kutchko, and B.R. Strazisar, 2010, "Experimental investigations of the effects of acid gas (H 2 S/C0 2 ) exposure under geological sequestration conditions," in 10th International Conference on Greenhouse Gas Control Technologies, Amsterdam, The Netherlands, September 19-23. Holubnyak, Y.I., S.B. Hawthorne, B.A.F. Mibeck, D.J. Miller, J.M. Bremer, J.A. Sorensen, E.N. Steadman, and J.A. Harju, 2010, "Modeling C0 2 -H 2 S-water-rock interactions at Williston Basin reservoir conditions," in 10th International Conference on Greenhouse Gas Control Technologies, Amsterdam, The Netherlands, September 19-23. Jensen, S., 2007, "Fluid flow and geochemistry of the Mississippian aquifers in the Williston Basin, Canada-U.S.A.," Department of Earth and Atmospheric Sciences, Edmonton, Alberta, Canada. Kent, D.M., EM. Haidl, and J.A. MacEachern, 1988, "Mississippian oil fields in the northern Williston Basin," in Goolsby, S.M., and Longman, M.W, eds., Occurrence and petrophysical properties of carbonate reservoirs in the Rocky Mountain region: Rocky Mountain Association of Geologists, Denver, Colorado, p. 381-417. Lindsay, R.F., 1988, "Mission Canyon Formation reservoir characteristics in North Dakota," in Goolsby, S.M., and Longman, M.W., eds., Occurrence and petrophysical properties of carbonate reservoirs in the Rocky Mountain region: Rocky Mountain Association of Geologists, Denver, Colorado, p. 317-346. Smith, S.A., J.A. Sorensen, D.W. Fischer, E.M. O'Leary, W.D. Peck, E.N. Steadman, and J.A. Harju, 2006, "Estimates of C 0 2 storage capacity in saline aquifers and oil fields of the PCOR Partnership region," in 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway, June 19-22 Proceedings.
This page intentionally left blank
22
Comparison of C0 2 and Acid Gas Interactions with Reservoir Fluid and Rocks at Williston Basin Conditions Yevhen I. Holubnyak, Steven B. Hawthorne, Blaise A. Mibeck, David J. Miller, Jordan M. Bremer, Steven A. Smith, James A. Sorensen, Edward N. Steadman, and John A. Harju Energy & Environmental Research Center University of North DakotaGrand Forks, ND, USA
Abstract A series of laboratory experiments, field observations from a small-scale C0 2 enhanced oil recovery project, and numerical modeling of geochemical reactions have been conducted to determine the chemical kinetics of potential mineral dissolution and/or precipitation caused by the injection of C0 2 and of sour gas. Batch laboratory experiments were conducted using core samples from potential C0 2 and acid gas storage formations of the Williston Basin in North Dakota. Two sample sets consisting of 16 samples each, under the same experimental conditions, were "soaked" for a period of 4 weeks at 145 bar (2100 psi) and 80°C (176°F) in synthetically generated brine conditions. Over that time period, one set was exposed to pure carbon dioxide and the other to a mixture of C0 2 (88 mol%) and H2S (12 mol%). Williston Basin geological settings, sample selection, and the results of the geochemical analysis of exposed samples are discussed in this paper.
22.1
Introduction
The Plains C0 2 Reduction (PCOR) Partnership, led by the University of North Dakota Energy & Environmental Research Center (EERC), is one of seven regional partnerships in the United States funded Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (407-420) © Scrivener Publishing LLC
407
408
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
by the U.S. Department of Energy's Regional Carbon Sequestration Partnership Program. As part of its ongoing regional characterization efforts, the PCOR Partnership has conducted a detailed examination of the potential C 0 2 storage capacity of several stacked brine-saturated formations in the central North Dakota portion of the Williston Basin. The study area, referred to as the Washburn area, encompasses 15,900 km 2 (6140 square miles) and is home to six coalfired power plants and one coal gasification plant which, combined, account for annual emissions of over 32 million tonnes of C0 2 . The Williston Basin is characterized by a thick sequence of sedimentary rock formations, in excess of 4877 m (1600 ft) at the basin center, which date from the Cambrian Period to the Holocene (Fischer et al., 2005a). Deposition from the Cambrian Period through the lower Ordovician was predominantly siliciclastic (sandstone and shales). Carbonates (limestones and dolomites) and evaporites (anhydrites and salts) were the dominant lithologies from the middle Ordovician through most of the Mississippian. Siliciclastics again became the dominant lithology in the Pennsylvanian and
Figure 1. PCOR partnership area and sedimentary basins.
COMPARISON OF C 0 2 AND ACID GAS INTERACTIONS
409
remained so through the Holocene. The stratigraphy of the Williston Basin is illustrated in Figure 1. To evaluate potential chemical and physical reactions between pure C 0 2 or a mixture of C 0 2 and H2S and selected Williston Basin rock units, samples representing five different formations were tested in bench-scale laboratory experiments. The comparison of the impact of the pure C 0 2 versus acid gas (C0 2 + H2S) became a subtask for this project. Numerical modeling of geochemical reactions was performed and verified with laboratory results. The samples were chosen based on both core availability and on the likelihood of the formation being a target for future C 0 2 storage. All Williston Basin samples were obtained through the North Dakota Geological Survey's Core Library located on the campus of the University of North Dakota. A detailed description of each sample and its relevance as a potential carbon storage unit is described in the following section.
22.2
Rock Unit Selection
To evaluate potential chemical and physical reactions between C 0 2 and selected Williston Basin rock units, samples representing three different formations were tested in bench-scale laboratory experiments: Madison Group, Broom Creek Formation, and Tyler Formation (Figure 2). The Madison Group is historically the primary oil-producing unit in the Williston Basin and provides significant opportunities for C 0 2 sequestration through enhanced oil recovery (Fischer et al., 2005a). The Madison is divided into three formations, which, in ascending order, are the Lodgepole, the Mission Canyon, and the Charles (Fischer et al., 2005b). To evaluate potential interactions between C 0 2 or sour gas, brine, and the Mission Canyon Formation rocks, a sample from a core was obtained. The combined mineralogical analysis suggests that major mineral phases in this sample are calcite (~60%), dolomite (~28%), anhydrite (~6%), quartz (less than 2%), illite (less than 2%), and pyrite (less than 1%). The minor mineralogical phases (less than 1%) were represented by chlorite, fluorite, magnesite, and others. Another representative of the Madison Group is the Mississippian-Ratcliffe Interval of the Charles Formation. This light gray limestone was recovered from the depth of 1800 m
410
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. Stratigraphie column for the North Dakota portion of the Williston Basin with evaluated formations in red rectangles.
(5895 ft). This is almost uniformly light gray matrix (calcite) with minor inclusions of darker grey color. The sample is characterized by smooth, nonporous texture. The combined mineralogical analysis suggests that the dominant phase is calcite (-75%) with dolomite (-11%), ankerite (-7%), quartz (less than 4%), and anhydrite (-1%). The Pennsylvanian-Tyler Formation is another oil-producing formation within the Williston Basin. The selected sample was recovered from the depth of 2430 m (7970 ft) and it is primarily clastic with a black, nonuniform structure with veins and spots of
COMPARISON OF C 0 2 AND ACID GAS INTERACTIONS
411
lighter and darker color. This sample primarily consists of calcite (-50%) and quartz (~35%), with smaller amounts of many other minerals, such as muscovite, kaolinite, dolomite, anhydrite, albite, pyrite, and others. The mineralogical analysis of minor phases can be viewed as semiquantitative only, as the amounts of all minor phases were lower than 6%-7%. The Broom Creek Formation is the thickest and most extensive brine-saturated sandstone in the Williston Basin, representing an excellent target for large-scale C 0 2 storage. The Broom Creek Formation is the uppermost member of the three formations comprising the Minnelusa Group. The Broom Creek is characterized by porous and permeable fine- to medium-grained sands (Williams and Bluemle, 1978). A sample of the Broom Creek Formation was obtained from a core that was extracted from a wellbore in Billings County at a depth of approximately 2380 m (7800 ft). The sample appears as a white and red, subangular to rounded, fine-grained sandstone. The mineralogy analysis indicated quartz (~76%), illite (~13%), kaolinite (-6%), and pyrite (~2%) as primary mineral phases.
22.3
C 0 2 Chamber Experiments
These experiments were designed to expose the selected rock/ mineral samples to supercritical C 0 2 under relatively high pressure and temperature, specifically 145 bar (2100 psi) and 80°C (176°F), respectively (Table 1). The tests were conducted by placing a V^-in. core plug into a small scintillation vial and inserting the open vials Table 1. Experimental conditions. C 0 2 and H2S Pressure:
145 bar (2100 psi)
C 0 2 Partial Pressure:
88 mole %
H2S Partial Pressure:
12 mole %
Temperature:
80°C (176°F)
Mass of Sample:
-7-15 g (-0.25-0.53 oz)
Saturation Conditions:
Synthetic brine: NaCl, 10% by weight
Time of Exposure:
4 weeks
412
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
into a reaction chamber, which could be regulated for temperature and pressurized with a C 0 2 or combined C 0 2 and H2S atmosphere (Hawthorne et al., 2010). Each sample was simultaneously saturated with saline solution (sodium chloride - NaCl). The samples were incubated in the testing chamber for a period of 4 weeks (28 days). The 4-week exposure time was conservatively selected after initial evaluation of the control sample (magnesium silicate) indicated that a complete reaction (carbonation reaction) was achieved after approximately 2 weeks (Sorensen et al., 2008).
22.4
Mineralogical Analysis
An x-ray diffraction (XRD) analysis was performed on each sample after CÓ 2 exposure to determine the mineralogical components of the samples and to evaluate any physical or chemical changes. The XRD scans are utilized to identify mineralogical signatures and to qualitatively estimate major and minor sample constituents. In addition to analyzing the samples exposed to C0 2 , a portion of the original sample was also analyzed to identify the original mineralogy. For the QEMSCAN analysis, samples of core plugs were prepared by placing a horizontal and a vertical section into a mold, which was then filled with epoxy. After setting, the epoxy slug was cut to expose the sample and polished to an approximately 1-um finish. Surficial reactions such as salt precipitation appear as a rind on the outside edges of a sample, whereas deeper reactions may be quantified by comparison to unreacted relative area percentages. Increases in phase definition to better examine trace concentrations of suggested reactive minerals within the matrix, specifically calcite/dolomite solutions, as well as added attention to rind composition, should help to explain the reactions. An integrative mineralogical analysis was performed utilizing linear program normative analysis (LPNORM). The computer code LPNORM implements the mathematical method of linear programming to calculate the mineralogical makeup of mineral mixtures, such as rock, sediment, or soil samples, from their bulk geochemical composition and from the mineralogical (or geochemical) composition of the contained minerals. This method simultaneously solves the set of linear equations governing the distribution of oxides into
COMPARISON OF C 0 2 AND ACID GAS INTERACTIONS
413
these minerals, subject to an objective function and a set of basic constraints (de Caritat et al., 1994). Changes in brine composition as a result of mineral dissolution and precipitation were analyzed with the inductively coupled plasma mass spectrometry instrument.
22.5
Numerical Modeling
The numerical modeling was performed with PHREEQC (Parkhurst and Appelo, 1999) and Geochemist's Workbench (GWB) software packages. The kinetic rate parameters were selected from available literature sources (Palandri and Kharaka, 2004) which describe pressure and temperature conditions in close proximity for the pressure and temperature conditions of the current experiment. Some of the listed kinetic rate parameters were not found in literature sources, so data which exist for similar minerals (e.g., minerals of the same group, similar crystal structure) were used instead. The sensitivity of the modeling because of this approximation is not known and requires further investigation. For improved modeling accuracy, the thermodynamic database for PHREEQC and GWB was recalculated and adjusted for the modeled set of pressure and temperature conditions with SUPRCRT92 code (Johnson et al., 1992).
22.6
Results
After the 28 days of exposure to supercritical pure C 0 2 or C 0 2 + H 2 S mixture, most samples were visibly altered. The changes apparent to the naked eye included obvious changes in porosity, coloration, crystal growth on the surface and cracks infill, changes in water coloration, and water contamination by precipitated minerals; for instance, consider Figure 3. In some extreme cases, full or partial destruction of the sample was observed (e.g., Tyler Formation sample). There are several observations which are common for all investigated Williston Basin rocks: 1) relatively fast dissolution of carbonate minerals (calcite, dolomite, etc.), 2) mobilization of iron within carbonate, iron-bearing, and possibly clay minerals; 3) the reaction products are different for pure C 0 2 and acid gas cases.
414
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 3. The Mississippian Mission Canyon sample collected from the depth of 2480 m (8140 ft) was saturated with brine (NaCl, 10%) and exposed to pure supercritical C 0 2 and a mixture of supercritical C 0 2 (88 mole %) and H2S (12 mole %) under pressure of 145 bar (2100 psi) and temperature of 80°C (176°F). The left side of the figure represents samples that were vacuum-dried after exposure compared with the original specimen; and the right side illustrates samples saturated in fluid after the completion of the experiment.
22.7
Carbonate Minerals Dissolution
For all investigated rocks from the Williston Basin, it was apparent that carbonate mineral dissolution had occurred. The dominant and fastest reaction was evidently the calcite dissolution. Different rocks from all four formations had different rates of carbonate dissolution; however, the difference in rates did not exceed 50%. For instance, after the 28 days of exposure to pure supercritical C0 2 , the porous structure of the Mission Canyon Formation rock became more prominent; the dark gray areas remained less porous and seemed to be affected less than the white and light gray areas (Figure 3). This observation correlates with the mineralogical analysis, which indicated that the dolomite dissolution was insignificant. In contrast, both QEMSCAN and XRD analysis show the reduction in calcite content by more than 10%. In addition, the water analysis suggests that change in Ca content (1602 mg/1) has to be attributed to calcite dissolution. The magnesium concentration (189 mg/1) in water was noticeably lower if compared to calcium and can be attributed to Mg content naturally present in calcite minerals. These observations correlate with numerical modeling predictions very well (Figure 4).
COMPARISON OF C0 2 AND ACID GAS INTERACTIONS
415
Figure 4. On the left is the combined mineralogical analysis of the initial (unexposed) sample (blue color), the sample exposed to C 0 2 (dark green), and the sample exposed to C 0 2 and H2S (orange). On the right is the exposed water composition analysis for metals compared to numerical modeling.
The observed calcite dissolution reaction can be written in the following chemical equations: C0 2 dissolution: H2O^H+
+ OH-
(1)
C0 2 , sup ** COlAH
(2)
C02aq + H20 <-> H2C03 &H++ HC03-
(3)
C02ac¡+OH^HC03 HCO¡^H+
+ CO¡-
(4) (5)
Calcite dissolution: CaCOs +H+ ^Ca2+ + HCO;
(6)
CaC03 +H2C03Ca2+ 2HCO¡
(7)
CaC03 + H20 <-> Ca2+C023 + H20 <^> Ca2+ + HCO¡ + OH-
(8)
416
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Dolomite dissolution: CaC03 + 2H+ <-» Ca2+ + Mg2+ + 2HC03 CaMg(C03)2 CaMg(C03)2
+ 2H2C03
(9)
<^ Ca2+ + Mg2+ + 4 H C 0 3
(1Q)
+ 2H20 <-> Ca2+ + Mg2+ + 2C023 + 2H20
<-> Ca2+ + Mg2+ + 2HCO¡ + 2 0 H " (11) The third reaction describes the dissociation of water into hydrogen ions and hydroxyl ions. The fourth reaction represents the physical dissolution of carbon dioxide in water. The fifth and sixth reactions cover the conversion of carbon dioxide into hydrogen and bicarbonate, depending on the pH value of the solution. The seventh reaction describes the dissociation of bicarbonate into hydrogen ions and carbonate ions. The eighth to tenth reactions illustrate the dissolution of calcite, depending on the pH value. The eleventh to thirteenth reactions show the dissolution of dolomite, also depending on the pH value (Kaufmann and Dreybrodt, 2007). The visible increase in porosity and darker gray coloration are among the observed changes of rock properties after the exposure. The mineralogical analysis suggests a reduction in dolomite by more than 5%, which is supported by the water analysis where concentrations of Mg exceeded 350 mg/1 for most samples (Figure 4). Also, this observation was supported by the numerical modeling predictions, which correlates with laboratory measurements within a 10% margin of error.
22.8
Mobilization of Fe
An interesting correlation was observed for all exposed samples that contained iron; for example, it is appropriate to consider the Pennsylvanian-Tyler Formation sample. After the exposure to supercritical C0 2 , the brine, the walls of the vial, and the sample itself had noticeable rust red coloration (Figure 5). The origination
COMPARISON OF C 0 2 AND ACID GAS INTERACTIONS
417
Figure 5. The Pennsylvanian-Tyler Formation sample collected from the depth of 2430 m (7970 ft) was saturated with brine (NaCl, 10%) and exposed to pure supercritical C 0 2 and a mixture of supercritical C 0 2 (88 mole %) and H2S (12 mole %) under pressure of 145 bar (2100 psi) and temperature of 80°C (176°F). The left side of the figure represents a sample that was vacuum-dried after exposure, compared with the original specimen; and the right side illustrates samples saturated in fluid after the completion of the experiment.
of this coloration could be explained by precipitation of iron oxide (hematite F 2 0 3 ); however, the source of oxygen for this reaction (13) is still unclear. 2Fe2+ + H20 + 0.5O 2 (aq) i 2Fe203 + 4 H +
(13)
Nevertheless, these visible changes in iron content are clear indicators of transformation in rocks exposed to brine and pure supercritical C0 2 . The refined QEMSCAN technique aims to identify the origination of these changes (Figure 6); the redistribution of iron after the exposure to pure supercritical C 0 2 is evident on this image. As for the C 0 2 + H2S case, the coloration of the water, the walls of the vial, and the sample itself after the exposure is gray in contrast to rust red coloration (Figure 5). The numerical modeling suggested that magnetite precipitation had occurred. However, it was impossible to confirm reaction products for the acid gas case. Currently, the EERC is working on refined methods of identification of mineral phases with the QEMSCAN instrumentation.
418
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. The refined QEMSCAN image of the Williston Basin rock, which illustrates the spatial distribution of iron content with the sample. On the left is the unreacted sample. On the right is the sample after the exposure to C0 2 .
22.9
Summary and Suggestions for Future Developments
The analysis of obtained reaction products suggests that 1) there is no strong evidence for higher degradation of samples exposed to a mixture of C 0 2 and H2S if compared to the pure C 0 2 stream; however, 2) carbonate rocks seem to be more unstable when exposed to the acid gas if compared to pure C0 2 , 3) if H2S is present in the stream, it seems to be more dominant in the reactions; and 3) reactivity of the sample is strongly driven by its mineralogy. The mineralogical analysis performed with various analytical tools (x-ray fluorescence, XRD, and QEMSCAN) required verification with numerical modeling tools. Often, the error in instrument tolerance, small-scale sample heterogeneity, or measurement error can be corrected by thermodynamic modeling suggestions.
22.10
Acknowledgments
This material is based upon work supported by the Department of Energy National Energy Technology Laboratory under Award Number DE-FC26-05NT42592.
22.11
Disclaimer
This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees,
COMPARISON OF C 0 2 AND ACID GAS INTERACTIONS
419
makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
References de Caritat, P., J. Bloch, and I. Hutcheon, "LPNORM: Alinear programming normative analysis code." Computers and Geosciences, Vol. 20, p. 313-341,1994. Fischer, D.W., S.A. Smith, W.D. Peck, J.A. LeFever, J.A., R.D. LeFever, L.D. Helms, J.A. Sorensen, E.N. Steadman, and J.A. Harju, "Sequestration potential of the Madison of the northern Great Plains aquifer system (Madison Geological Sequestration Unit), Plains C 0 2 Reduction (PCOR) Partnership topical report for U.S. Department of Energy and multiclients," Grand Forks, ND, Energy & Environmental Research Center, September 2005,2005a. Fischer, D.W., J.A. LeFever, R.D. LeFever, S.B. Anderson, L.D. Helms, S. Whittaker, J.A. Sorensen, S.A. Smith, W.D. Peck, E.N. Steadman, and J.A. Harju, "Overview of Williston Basin geology as it relates to C 0 2 sequestration, Plains C 0 2 Reduction (PCOR) Partnership topical report for U.S. Department of Energy and multiclients," Grand Forks, ND, Energy & Environmental Research Center, May 2005,2005b. Hawthorne, S.B., D.J. Miller, Y. Holubnyak, B.G. Kutchko, B.R. Strazisar, "Experimental Investigations of the Effects of Acid Gas (H 2 S/C0 2 ) Exposure under Geological Sequestration Conditions," in International Conference on Greenhouse Gas Control Technologies, 10th, Amsterdam, Netherlands, September 19-23,2010. Johnson, J.W., E.H. Oelkers, H.C. Helgeson, "SUPCRT92 - a software package for calculating the standard molal thermodynamic properties of minerals, gases, aqueous species, and reactions from 1 to 5000 bar and 0 to 1000CC," Computational Geosciences, Vol. 18, p. 899-947,1992. Kaufmann, G., and W. Dreybrodt, "Calcite dissolution kinetics in the system C a C 0 3 - H 2 0 - C 0 2 at high undersaturation," Geochimica et Cosmochimica Acta, Vol. 71, Issue 6,15 March 2007, p. 1398-1410,2007. Palandri, J.L. and Y.K. Kharaka, "A compilation of rate parameters of watermineral interaction kinetics for application to geochemical modelling," U.S. Geological Survey Open File Report 2004-1068,70 pages, 2004. Parkhurst, D.L., and C.A.J. Appelo, "User's guide to PHREEQC (version 2) - a computer program for speciation, batch-reaction, one-dimensional transport,
420
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
and inverse geochemical calculations," Water-Resources Investigations Report 99-4259, Denver, CO, 1999. Sorensen, J.A., Y.I. Holubnyak, S.B. Hawthorne, D.J. Miller, K.E. Eylands, E.N. Steadman, and J.A. Harju, "Laboratory and numerical modeling of geochemical reactions in a reservoir used for C 0 2 storage," in International Conference on Greenhouse Gas Control Technologies, 9th, Washington, D.C., November 2008, Proceedings, in press. Williams, B.B., and M.E. Bluemle, "Status of mineral resource information for the Fort Berthold Reservation, North Dakota," Administrative Report No. 40, Bureau of Indian Affairs, p. 71.11,1978.
SECTION 6 WELL TECHNOLOGY
This page intentionally left blank
23
Well Cement Aging in Various H 2 S-C0 2 Fluids at High Pressure and High Temperature: Experiments and Modelling Well Cement Aging at High PT Conditions Nicolas Jacquemet1'2, Jacques Pironon1, Vincent Lagneau3, Jérémie Saint-Marc4 ^ancy
Université-CNRS-CREGU, France 2 BRGM, France 3 AÍINES ParisTech, France 4 TOTAL, France
Abstract The reactivity of a deep well cement analogue in contact with (1) a brine with dissolved H 2 S-C0 2 ; (2) a dry H 2 S-C0 2 supercritical phase; (3) a two-phase fluid associating a brine with dissolved H 2 S-C0 2 and a H 2 S-C0 2 supercritical phase, was investigated in batch experiments at 500 bar and 120°C. The cement was carbonated by the C 0 2 and different mineralogical alteration profiles were observed according to the different exposures cited above. H2S reacted with the minor iron-bearing minerals of cement (ferrites) to form sulfate/sulfide mixtures. The cement carbonation was maximal when occurring within the dry supercritical phase without liquid water. In all the experiments, the porosity decreased (from 40% to about 15%). Due to the low W / R ratio in the reactor, a calcium carbonates coating even precipitates on surface cement inhibiting further reactions when cement is immerged in brine. Reactive transport simulations performed with HYTEC confirmed the coating as alteration-inhibitor.
Wu/CarroU/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (423-436) © Scrivener Publishing LLC
423
424
23.1
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Introduction
The geological storage of C 0 2 and possibly H2S is a promising solution for the long-term storage of these undesirable gases (1-2). It consists in injecting them via wells into deep geological reservoirs (depleted oil and gas fields or deep aquifers). More than fifty acid gas (mixture of C 0 2 and hydrogen sulfide (H2S)) injections into geological reservoirs have been operated in western Canada since 1990 and potential failures of 315,000 oil, gas, and injection wells in the province of Alberta have been statistically described (3). Well failures correspond to surface casing vent flow, casing failure, tubing failure, packer failure, and zonal isolation failure (4). Carey et al. (5) described alteration at the seal-wellbore interface after samples of cement and shales were recovered from a well used in a long-term C 0 2 enhanced oil recovery operation during 30 years (SACROC unit, Permian Basin, Texas). There was evidence for C 0 2 migration along casing-cement and cement-shale interfaces, marked by precipitation of polymorphs of calcium carbonate (calcite, aragonite and vaterite). Transport of C 0 2 or acid gases through "anthropogenic" pathways such as wellbore, may be of the order of tens to hundreds of years whereas natural pathways can lead to very slow transport on timescales of tens of thousands of years (6, 7). Failures can be caused by geochemical reactions occuring at the well-reservoir/caprock interface between the well materials and the fluids generated by the injection of gas (e.g. H 2 S-C0 2 supercritical phase, diphasic fluids and aqueous liquids with dissolved gas, Figure 1). The chemical reactivity between the cement of the well construction and the fluids (gas, brine) occurring within acid gas geological storage can result in its alteration. Precipitation and / o r dissolution of minerals induced by the fluid-minerals interactions may change the physical properties of the material (texture, porosity, transport properties, ...). The chemical reactions between the cement of the well construction and the fluids (gas, brine) occurring within acid geological storage may change the hydraulic properties of the material (porosity, diffusivity, permeability) and its mechanical strength. These changes have been deduced from experiments in the case of Concernent interactions (8, 9,10) or C0 2 +H 2 S-cement/ steel interactions (11,12). Hence, the desirable confinement properties of the cement may be degraded and the well's sheath could
WELL CEMENT AGING IN VARIOUS
H 2 S-C0 2 FLUIDS
425
Figure 1. Scheme of the location and type of fluid-well interactions occurring in an acid gas geological storage.
provide pathway for the H 2 S-C0 2 fluids up to the surface. The geochemical reactions may potentially degrade the well with subsequent human and environmental consequences. The objective of this article is to simulate in lab aging of the well materials (cement and steel) at relevant pressures (P) and temperatures (T) to evaluate the durability of the materials. The phenomena observed were used to define a numerical model. For this purpose, we used the reactive transport code HYTEC (13). These actions provide practical experience for the operators/ institutions that are/will be in charge of the on-going and planned C0 2 /acid gas geological storages.
23.2 Experimental Equipment A specific experimental procedure was developed to study systems with high concentration of H2S up to 990 bars-450 °C with respect to safety rules (14). Experiments were conducted in batch type micro-reactors consisting in flexible gold capsules which were introduced in hydraulic pressure vessels (Figure 2, top). An original gas-loading device was designed in order to fill accurately the gold capsules with known quantities of gas (Figure 2, bottom). The H 2 S-C0 2 mixture was condensed in a cryogenic bath to ensure a low pressure gas handling.
426
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. Schematic illustration of the high pressure, high temperature experimental device (top) and of the gas loading device (bottom). Modified from [14].
23.3
Materials, Experimental Conditions and Analysis
23.3.1 Cement The initial cement was a blend of Portland-Class G-High Sulfate Resistant type and silica flour (35% by weight of cement) with W / C ratio of 0.55 and density of 1900 g/1 (standard composition for high
WELL CEMENT AGING IN VARIOUS
H 2 S-C0 2 FLUIDS
427
temperature). It was pre-cured at 210 bar-140°C for 8 days in aqueous media (cure analogue to deep well environment). It was conditioned in bars of 10 mm x 3 mm section. 23.3.2
Casing
The steel was a C22E type (or XC18) (usual composition for well casing, 98 mol% of iron). It was conditioned in cylinders of 3 mm x 1 mm diameter. 23.3.3
Environment
The brine was a 150 g/1 NaCl solution (mean salinity of usual oilfields). The gas mixture was composed of 66 mol% of H2S plus 34 mol% of C 0 2 (composition of several currently operated injection mixtures in western Canada).
23.3.4
Exposures (Figure 3):
• In the first treatment, the cement was completely immersed in the brine at equilibrium with the H 2 S-C0 2 mixture in supercritical state, • In the second treatment, the porous volume of the cement was saturated with brine and the bar was surrounded by the H 2 S-C0 2 mixture in supercritical state, • In the third treatment, the pre-dried cement (i.e. containing no porous water) was surrounded by the H 2 S-C0 2 mixture in supercritical state. Equivalent pressure and temperature conditions of Caspian Sea oil reservoirs were chosen for the experiments : 500 bar, 120°C.
23.3.5
Analyses
Several analysis techniques (optical microscopy, XRD, SEM, MicroRaman spectroscopy and water porosimetry) were used to characterize the samples after aging; only results with optical microscopy, SEM, and water porosimetry are proposed in the present paper. As microreactors (i.e. gold capsules) do not permit fluid sampling during experiment, synthetic inclusion technique (15) has been used to trap the fluids in microcavities of quartz crystals added to
428
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 3. The three types of cement exposures, showing the nature of the fluid phase in contact with cement and steel.
the reactant. Raman microspectroscopy, coupled to a microthermometric stage, allows molecular analyses reproducing PT conditions of experiment. The number of coeval fluid phases and their nature can be characterized.
23.4
Results and Discussion
23.4.1 Cement The optical microscope and SEM observations of the cement samples aged following the three scenarios are presented in the Figure 4. For cement immersed in brine, we have observed a reaction front marked by three successive layers: • a massive calcite deposit at the surface of the cement bars, • a decalcified zone with a constant thickness, • a relatively unaltered cement.
WELL CEMENT AGING IN VARIOUS
H 2 S-C0 2 FLUIDS
429
Figure 4. Optical microscope and SEM observations of the reaction fronts produced by aging in brine (left) and in supercritical phase (centre and right).
Moreover, we have observed that the progression of the reaction front is mostly stopped after 15 days (the calcite deposit and the decalcified zone have the same thickness for 15 and 60 days). The formation of the decalcified zone is explained by the dissolution of the C-S-H due to the contact with the low pH (relative to the basic pH of the porous aqueous liquid), low calcium, external aqueous liquid. The migration of the calcium and hydroxides towards the external solution, combined with the dissolved carbonates of the brines (from the dissolution-dissociation of the C0 2 ) allow the precipitation of calcite at the interface cement-solution. The massive calcite deposit can then block the diffusion and further reactions between 1 to 15 days. For cement immersed in supercritical fluid with brine-saturated pores, the reaction front was marked by successive layers: • a thin calcite deposit at the surface of the cement bars, • a carbonated zone of variable thickness (digitations), • a relatively unaltered cement. The progressive intrusion of the supercritical phase within the porous media could explain the observed digitations of carbonated
430
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
zones within the unaltered cement. Indeed the capillary forces at the water-gas interface depend on the highly heterogeneous pore structure, which can easily account for a differential intrusion of the gas. Little migration of calcium is expected in the vapour phase; hence, no important calcium re-concentration is possible at the surface of the cement (with the associated massive calcite deposit formation). For cement immersed in supercritical fluid, the alteration was marked by: • a thin aragonite or calcite external deposit, • the full carbonation of the cement. The full carbonation of the cement resulted from the optimal diffusion of the C 0 2 at supercritical (P,T) conditions (16) as from the high reactivity of cement with C 0 2 occurring in absence of liquid water (12). As in the diphasic fluid, the thinness of the external deposit is explained by the little migration of calcium in the vapour phase. Bulk porosity of starting and aged cement bars was measured by water porosimetry The porosity of the starting cement is 40%. The porosity of aged cement decreased systematically (due to CaC0 3 precipitation) in all experiments. The final porosity of the aged cement bars is about 15%.
23.4.2
Steel
The steel corrosion is characterised by the precipitation of sulphidation crust composed of pyrrhotite (Figure 5). The sulphidation of steel can be described by the following reaction: Fe° + H 2 S -> FeS + H 2
(1)
The presence of H 2 has been confirmed by the analysis of synthetic inclusions having trapped experimental fluid in gold capsule.
23.5
Reactive Transport Modelling
The purpose of the reactive transport modelling was to reproduce the alteration front observed after aging in the aqueous liquid phase; no attempts were made so far for the diphasic and / o r dry supercritical phase experiments.
WELL CEMENT AGING IN VARIOUS H 2 S-C0 2 FLUIDS
431
Figure 5. Binocular (left), optical microscope on polished sections (centre) and SEM (right) observations of sulphidation crust (pyrrhotite) developed on steel micro-cylinder.
Figure 6. ID model showing the four different simplified hydro-geochemical zones.
The numerical simulation of the cement alteration needs a reactive transport code as both chemistry and transport processes (diffusion) are involved in the cement alteration. The reactive transport code HYTEC (13), based on the geochemical code CHESS, was used to simulate the reactions between the cement and the acidic brine. The CTDP (Common Thermodynamic Data Project) thermodynamic database (17) was selected for the study and enriched with additional data about H2S and C 0 2 solubility from the literature (12, 15, 18, 19). A ID model of a cross section of the cement bar was constructed for the calculations (Figure 6). It contains four initial hydro-geochemical zones (film of brine in contact with the gas, brine, crust and cement) defined by their initial diffusion properties, porosities and chemical compositions. The results of the simulations are shown in the Figure 7. The model reproduced the successive layers observed experimentally:
432
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
• a calcite crust formation at the surface of the cement bar, • a decalcified zone marked by the presence of silica and gyrolite (CSH with Ca/Si=0.66) / • The unaltered cement mimicked by a CSH association of tobermorite 11Â (Ca/Si=0.83) and foshagite (Ca/ Si=1.33). The formation of the calcite crust efficiently reduces the diffusion at the cement-solution interface. The flux of calcium and carbonates through the interface decreases dramatically so that the progression of the front stops after 5 to 10 days. This numerical result confirmed that the diffusion blockage (caused by the calcite deposit formation) is the process responsible for the front progression stop. However the thickness of the decalcified zone predicted by the model is two to three times larger than the thickness experimentally observed. It should be improved by a better fit of the model parameters.
23.6
Conclusion
Figure 7. Numerical simulation results obtained with HYTEC at 15 days duration.
WELL CEMENT AGING IN VARIOUS
H 2 S-C0 2 FLUIDS
433
We designed a specific experimental device to study well cement aging with H 2 S-C0 2 in geological-relevant conditions (brine, high pressure and temperature). The aim of this study was the evaluation of well cement durability in acid-gas storage hydrocarbon operations and by extension in C 0 2 geological storage in DBR (Deeply Buried Reservoirs) conditions. SEM characterization and water porosimetry figured out the carbonation of cement beneficial decrease of porosity (and by extension improvement of isolation properties). While immersed in aqueous liquid with dissolved H 2 S-C0 2 , a calcite deposition occurred at cement surface acting as a passivation layer that stops further carbonation. The passivation character of the deposit was evidenced by reactive transport modelling. We suppose those beneficial carbonate deposition to be related to low W / R (Water/ Rock) ratio (~1) occurring in the reactors. In presence of supercritical fluid, the initially dried cement is completely carbonated with a thin crust of aragonite and calcite. When pores are initially brine saturated, alteration fronts are detected showing digitations. These results show how the fluid state is important to predict the behaviour of solid phases. C 0 2 is the main actor of solid transformation in cement, whereas H2S is responsible for steel sulphidation in the extreme P-T conditions applied in this work. Future experiments with refreshing of surrounding fluid e.g. plug-flow, or large W / R ratio would be necessary to assess the long-term persistence of carbonates deposits within of at surface of cement.
Acknowledgments This study has been supported by TOTAL in the form of a Ph.D. thesis of Nancy Université, France, as part of a R&D project "Residual Gas Management" dedicated to C 0 2 capture and storage as well as acid gas injection. The authors thank TOTAL for the authorization of publishing these results. The authors thank also Alain Köhler from Nancy Université for the SEM analyses. Any opinions, findings, conclusions, or recommendations expressed herein are those of the authors and do not necessarily reflect the views of the sponsors.
434
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
References 1. Chakma, A. "Acid gas re-injection - a practical way to eliminate C02 emissions from gas processing plants." Energy Convers. and Manage. 1997, 38 (Supplt), S 205-S 209. 2. Connock, L. "Acid gas re-injection reduces sulphur burden." Sulphur 2001, 272,35-41. 3. Bachu, S., Watson, T.L., "Review of Failures for Wells used for C 0 2 and Acid Gas Injection in Alberta," Canada Energy Procedía 2009,1,3531-3537. 4. Watson, T.L. and Bachu, S. 2009. "Evaluation of the Potential for Gas and C 0 2 Leakage Along Wellbores." SPE Drill & Compl 24 (1): 115-126. SPE-106817-PA. 5. Carey, J. W; Wigand, M.; Chipera, S. J.; WoldeGabriel, G.; Pawar, R.; Lichtner, P. C ; Wehner, S. C ; Raines, M. A.; Guthrie, G. D. "Analysis and performance of oil well cement with 30 years of C 0 2 exposure from the SACROC unit, West Texas, USA." International ]. of Greenhouse Gas Control 2007,1 (1), 75-85. 6. Savage D, Maul P R, Benbow S J and Stenhouse, M. (2003) "The assessment of the long-term fate of carbon dioxide in geological systems." In Coping with Climate Change, 25-27 March 2003. Geological Society of London Online Extended Abstracts. 7. Celia, M. A. and Bachu, S. (2002) "Geological sequestration of C0 2 : is leakage unavoidable and acceptable ?" GHGT-6 (6th Int. conf. on GreenHouse Gas control Technologies). 6 pages, http://www.princeton.edu/~cmi/research/ kyoto02/celia&bachu.kyoto%2002.pdf 8. Kutchko, B.G., Strazisar, B.R., Dzombak, D.A., Lowry, G.V., Thaulow, N. "Degradation of Well Cement by C 0 2 under Geologic Sequestration Conditions." Environ. Sei. Technol. 2007,41:12,4787-4792. 9. Kutchko, B., Strazisar, B., Dzombak, D., Lowry, G. "Rate of C0 2 Attack on Hydrated Class H Well Cement under Geologic Sequestration Conditions." Environ. Sei. and Technol. 2008, 42:16, 6237-6242. 10. Fabbri, A., Corvisier, J., Schubnel, A., Brunet, E, Goffé, B., Rimmele, G., Barlet-Gouédard, V. "Effect of carbonation on the hydro-mechanical properties of Portland cements," Cement and Concrete Research, Volume 39, Issue 12, December 2009, Pages 1156-1163. 11. Jacquemet N, Pironon J, Saint-Marc J (2008). « Mineralogical changes of a well cement in various H 2 S-C0 2 (-brine) fluids at high pressure and temperature." Environ. Sei. Technol. 42, 282-288. 12. Jacquemet, N., 2006, "Well materials durability in a context of carbon dioxide and hydrogen sulfide geological storage," Ph.D. Thesis, Université Henri Poincaré, Nancy, France. 13. van der Lee, J., De Windt, L., Lagneau, V. and Goblet, P, 2003, "Module-oriented of reactive transport with HYTEC." Computers and Geosciences, 29,265-275. 14. Jacquemet, N., Pironon, J. and Caroli, E., 2005, "A new experimental procedure for simulation of H 2 S+C0 2 geological storage-Application to well cement aging." Oil & Gas Science and Technology-Rev. IFP, Vol. 60, No. 1, pp. 193-206. 15. Pironon, J., Jacquemet, N., Lhomme, T., Teinturier, S. 2007 "Fluid inclusions as micro-samplers in batch experiments: a study of the system C-O-H-S-cement with application to the geological storage of industrial acid gas." Chemical Geology 237,264-273.
WELL CEMENT AGING IN VARIOUS H 2 S-C0 2 FLUIDS
435
16. Fernandez Bertos, M., Simons, S. J. R., Hills, C. D., Carey, P. J., 2004, "A review of accelerated carbonation technology in the treatment of cement-based materials and sequestration of CO,." Journal of hazardous materials, B112,193-205. 17. CTDP v. 1.0.0, available at http://ctdp.ensmp.fr/ 18. Duan, Z. et Sun, R. (2003) "An improved model calculating C 0 2 solubility in pure water and aqueous NaCl solutions from 273 to 533 k and from 0 to 2000 bar." Chemical geology, 193, 257-271. 19. Bakker, R.J. 2009 "Package FLUIDS. Part 3: correlations between equations of state, thermodynamics and fluid inclusions," Geofluids 9, 63-74.
This page intentionally left blank
24
Casing Selection and Correlation Technology for Ultra-Deep, Ultra- High Pressure, High H2S Gas Wells Yongxing Sun, Yuanhua Lin, Taihe Shi, Zhongsheng Wang, Dajiang Zhu, Liping Chen, Sujun Liu, and Dezhi Zeng CCDC Drilling & Production Technology Research Institute Guanghan People's Republic of China State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation (SWPU) Chengdu, People's Republic of China CNPC Key Laboratory for Tubular Goods Chengdu, People's Republic of China
Abstract With ultra-deep (more than 5000 m), ultra-high pressure (more than 100 MPa), ultra-high temperature (more than 150°C), high H2S (2-70%) gas wells increasing, the casing service environment tends to be critical, sour and complex in the Northeast Sichuan province. In these wells, when well temperature is below 90°C, failures due to sulfide stress cracking (SSC) of High Strength Sour Service grades for downhole applications have been reported in recent years. Furthermore, it is well known that the plastic creep formation made of rock salt, gypse, and clay shale will give rise to much higher external collapse pressure on casing, which needs much higher collapse strength OCTG grades (such as non-sulfur resistance casing 140ksi, 150ksi grades) to meet test and production criteria. All of the complex cases result in catastrophic consequences that require the sulfur resistance casing (such as C110) not only to meet the ultra-high pressure criteria (more than lOOMPa) but also to meet the high sulfur resistance criteria (H2S partial pressure is 0.3-2MPa in Sichuan). Though rules for the selection of proper materials to avoid the catastrophic consequences of SSC are gathered in the latest edition of the NACE MR0175/ISO151516 standard "Materials for use in H2S containing environments" and "Sumitomo Metal sulfur resistance casing selection Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (437-448) © Scrivener Publishing LLC
437
438
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
recommended practice", both selections' recommended practices cannot present suitable casing. So this paper discusses an economical and suitable method of string design combined with sulfur resistance packer completion test technology, and this method successfully deals with the current difficult problems of sulfur resistance casing selection and casing program design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
24.1
Introduction
Currently, more ultra-deep, ultra -high pressure, ultra-high temperature oil & gas wells are being drilled and completed in the presence of H2S and C0 2 , the need for high strength (>=110ksi) sour service-rated tubular is ever increasing. In these wells, when well temperature is below 90°C, downhole High Strength Sour Service grades failures accidents due to sulfide stress cracking (SSC) have been reported in recent years. East Sichuan gas fields contain 7.12% to 17.03% of H2S, and 3.29% to 10.41% of C0 2 . Chloride ion content is about 20429 mg/1 in the L well, as determined in the water analysis report. The corrosion problems of oil country tubular goods (OCTG) and equipments often take place due to H2S and C 0 2 [1, 2, 3, and 4]. Though rules for the selection of proper materials to avoid the catastrophic consequences of SSC are gathered in the latest edition of the NACE MR0175/ISO 151516standard15-61 "Materials for use in H2S containing environments" and "Sumitomo Metal sulfur resistance casing selection recommended practice", both selection recommended practices cannot present suitable casing. So this paper discusses an economical and suitable method of string design combined with sulfur resistance packer completion test technology, and this method successfully deals with current difficult problem of sulfur resistance casing selection and casing program design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
24.2
Material Selection Recommended Practice
The severity of the sour environment, determined in accordance with ISO 15156-1 (2009), with respect to the SSC of a carbon or low-alloy steel shall be assessed using Figure 1 [6]. In defining the
CASING SELECTION AND CORRELATION TECHNOLOGY
439
severity of the H 2 S-containing environment, the possibility of exposure to unbuffered, condensed aqueous phases of low pH during upset operating conditions or downtime, or to acids used for well stimulation and / o r the backflow of stimulation acid after reaction should be considered. When partial pressure is higher than 0.7 bar, Cr-Ni materials should be used. Carbon and low-alloy steels selection should be tested by casing and tubing serve conditions from regions 1, 2, and 3 (shown in Figure 1) [6]. If the recommended casing and tubing such as C110 has been applied downhole, whose serve environments can be referred as casing and tubing selection recommended practice. The discontinuities in the figure below 0.3 kPa (0.05 psi) and above 1 MPa (150 psi) partial pressure H 2 S reflect uncertainty with respect to the measurement of H2S partial pressure (low H2S) and the steel's performance outside these limits (for both low and high H2S). In this case, it must use fit for purpose (FFT)[7] method to evaluate the OCTG materials according actual well conditions, then, it can be determined whether or not to use the recommended OCTG materials.
Figure 1. Regions of environmental severity with respect to the SSC of carbon and low-alloy steels Key: X-H2S partial pressure, expressed in kilopascals; Y-in situ pH; 0-region 0; l-SSC region 1; 2-SSC region 2; 3-SSC region 3. Note 1 Guidance on the calculation of H2S partial pressure is given in ISOl 5156:2009, Annex C. Note 2 Guidance on the calculation of pH is given in ISO15156:2009, Annex D.
440
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2 provides a simplified guideline to choose from the material applications based on C 0 2 content, H2S content and temperature by Sumitomo Metals [8]. Reliable and optimized material selection depends upon a large set of parameters, from fluid characteristics to well conditions. The chart provides a quick way to make a pre-selection of material type. It should be used as a guideline. For a more detailed assessment, please refer to each Material Data Sheet [8]. This material selection chart features Sumitomo Metals proprietary grades. Sumitomo Metals also does manufacture API 5CT/ISO 11960 grades. It has been recognized that there are still much inadequacies to design casing program by Figures 1 and 2 for oil and gas wells with H 2 S[6,8,and9]. 1. The low grade anti-sulfur casing such as J55, T95 can just be applied in shallow oil and gas well for their low strength. 2. The low grade casing strength (burst and collapse) can be increased by adding the wall thickness, which can be done within limited conditions, because of the increase in cost and change of casing program that has been wide applied in oil and gas fields. In addition, a significant increase in wall thickness will make the casing cannot meet the tensile strength safe criteria. So it cannot still meet high collapse resistance and internal pressure strength criteria in HTHP sour gas well through adding wall thickness. 3. Stainless steel can meet anti-sulfide criteria, but it cannot meet strength design criteria in ultra-HTHP sour gas well, as well as its high-cost. 4. Though lots of anti-sulfide casing such as 110SS, 125SS have been applied in oil and gas fields at abroad and home, but all of them must be evaluated through FFT method before they are applied in oil and gas fields, otherwise, whose strength cannot still meet strength design criteria in ultra-HTHP sour gas well in east Sichuan oil fields, and 140 ksi or 150 ksi grade OCTG may be just meet the high collapse criteria [10].
CASING SELECTION AND CORRELATION TECHNOLOGY
441
Figure 2. Material Selection According to Gas (C0 2 and H2S) Partial Pressure (Note: Cl" content should be less than 30,000ppm for SMCR and SM 13CR).
24.3 Casing Selection and Correlation Technology All above discussed cannot deal with casing strength design problem for ultra-HTHP, high H2S well in east Sichuan oil fields. Based on the further study, the contradiction between the anticorrosion and high collapse criteria has been perfectly solved in following technology: 1. When well temperature is above 90°C, sulfide-stress cracking (SSC) can be neglected, but the plastic creep formation made of rock salt, gypse, clay shale will
442
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
give rise to much higher external collapse pressure on casing, for this case, it can use higher collapse strength OCTG grades such as non-sulfur resistance casing 140ksi, 150ksi grades to meet test and production criteria. 2. During completion operations, especially in the stimulation operations, the well temperature will decrease to 60°C -90°C, which results in the high strength casing such as 140 or 150 ksi grade stress SSC. So the two methods below have been used to prevent the high strength OCTG such as 140 or 150 ksi casing from SSC, which can be described as follows. a. Completion strings with an anti-sulfide packer can separate the high strength casing (140 or 150 ksi) from natural gas containing H 2 S, which can decrease the possibility of SSC and meet the strength criteria. b. For test well, it is necessary to prevent the bottom temperature from decrease and make the high strength casing (140 or 150 ksi) immune to SSC. Because the high collapse strength and anticorrosion capability of OCTG cannot be simultaneously presented by Figures 1 and 2. So this paper presents an economical and suitable method of string design combined with sulfur resistance packer completion test technology. By anti-sulfide packer, the low temperature (below 90°C) natural gas containing H2S is isolated from the casing which prevent the high strength casing from SSC. In this case, the high strength casing such as 140 ksi etc. can be applied in the plastic creep formation containing H2S to meet strength design criteria.
24.3.1
Casing Selection and Match Technology Below 90°C
During completion operations, when well temperature is below 90°C, H2S corrosion cannot be ignored. The low-temperature wells containing H2S (below 90°C) should select the low grade casing. If the casing strength cannot meet the design criteria, the suitable casing can be obtained through reasonable increase in the wall thickness.
CASING SELECTION AND CORRELATION TECHNOLOGY
443
24.3.2 Casing Selection and Match Technology Above 90°C When well temperature is above 90°C, H2S corrosion can be ignored. In this case, the high strength casing such as 140 ksi etc. can be applied in high temperature well. 1. This approach is feasible when the well temperature is constant 2. During acidizing treatment, well temperature will be below 90°C, in this case, the SSC must be considered. In order to prevent the high strength casing from corrosion due to well temperature decreasing, the following match technology measures are used: a. By sulfur resistance packer completion test technology, the gas containing H2S is isolated from low temperature casing above sulfur resistance packer. b. Because of extending acid reaction time can make the acid temperature above 100°C in formation, and then the casing below sulfur resistance packer can be protected. c. After controlling the higher wellhead pressure, because of gas volume expansion, it is necessary to keep the gas from the information in constant temperature. When the high strength casing such as 140 ksi contacts the gas with H 2 S, the well temperature remains above 90°C, the casing below sulfur resistance packer can be effectively protected.
24.4
Field Applications
This new method has been successfully applied in most of gas wells containing H2S in M gas fields in China, and the typical well programs are shown in Figures 3 and 4. Figure 3 is the original casing program of L well in M gas fields, and the OD 177.8 mmxll.51 mm wall thickness VMllOSS casing is selected for its good anti-sulfide, but whose strength is not immune to L well with formation pressure 127.79 MPa (mid-depth 5942.9 m) and the wellhead pressure 107 MPa. The OD 177.8 mmxll.51 mm wall thickness VMllOSS casing collapse and anti-pressure safe factor is respectively 0.76 and
444
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
0.68 (shown in Table 1) in air. All the safety factors cannot meet the "Drilling technical operation" [11] (collapse safety factor is 1.125, intermediate casing and production's is not less than 0.80), as well as the SY/T5322-2000[12] (collapse safety factor is from 1.00 to 1.125, anti-pressure safety factor is from 1.05 to 1.15). It is worth noting how to avoid the catastrophic consequences of casing collapsing of M gas fields, because the casing strength cannot meet the strength criteria. So the original casing program of L well has been partly changed. On the basis of guaranteeing completion and cementing quality, the OD 177.8 mmxll.51 mm wall thickness VM110SS casing is replaced by OD 193.68 mmxl9.05 mm wall thickness TP110SS casing, and the collapse strength is 15% higher than that of VM110SS, which not only meet the strength criteria but meet anti-corrosion requirement, in the meantime, all the safety factors is higher than 1.0 The high strength casing such as VM140HC (collapse strength is 117.62 MPa) is applied in the plastic creep formation zone, and the
Figure 3. The original casing program L well.
CASING SELECTION AND CORRELATION TECHNOLOGY
445
Figure 4. The changed casing program L well.
changed casing program is shown in Figure 4. This method successfully deals with current difficult problem of sulfur resistance casing selection and casing program design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
24.4
Conclusions 1. The method of string design combined with sulfur resistance packer completion test technology can successfully deal with current difficult problem of sulfur resistance casing selection and casing strength design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
VM140HC
180
1500
177.8
177.8
139.7
Production Tieback
Production suspend
WM110SS
4568
244.5
Intermediate 2 hang
3300
1448
244.5
Intermediate 2 Tieback
KO-140V
TP-95TS
TP-95ss
P-110
1698
339.7
Intermediate 1
J-55
Grade
98
Setting Depth(m)
508
(mm)
OD
Surface
String
Table 1. The original casing program of L well.
12.7
12.65
11.51
11.99
11.99
598
689
2175
2308
1012
1818
155
12.7 13.06
Weight (kN)
Wall (mm)
159.5
117.62
74
45.5
35.09
19.89
5.3
Collapse Strength (MPa)
1.19
1.16
0.76
0.70
1.77
1.86
5.25
Safety Factor
Collapse
108.5
120.18
86
56.19
56.19
34
16
Burst Strength (MPa)
1.4
1.51
0.68
0.83
0.74
1.02
15.94
Safety Factor
Burst
3866
6334
4559
5846
5846
10160
7495
Tensile (kN)
6.46
9.19
2.09
2.53
5.77
5.58
48.35
Safety Factor
Tensile
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
CASING SELECTION AND CORRELATION TECHNOLOGY
447
2. For critical corrosion environments, when well bottom temperature is below 90°C, the lower grade anti-sulfide casing such as T95, J55 etc. can be applied. If the lower grade casing anti-pressure cannot meet strength criterion, adding the pipe wall thickness.
24.5 Acknowledgments The authors are grateful to the support from Program for New Century Excellent Talents in University (NCET-08-0907) and China National Natural Science foundation and Shanghai Baosteel Group Corporation (Grant No.: 51074135).
References 1. Li Luguang, Huang Liming, Gu Tan, Li Feng. "Corrosion Characters and Inhibition Methods for Sichuan Gas Fields." Chemical Engineering of Oil & Gas, 2007,36(l):46-54 (in Chinese with English Abstr.). 2. Jiang Fang. "In Lab. Evaluation Methods of Metal Materials for High Sour Gas Fields." Natural Gas Industry, 2004, 24(10): 105-107 (in Chinese with English Abstr.). 3. Zeng Dezhi, Huang Liming, Lin Yuanhua. "Material Evaluation and Selection of OCTG and Gathering Lines for High Sour Gas Fields." SPE 131943, 2010, 6 (in Chinese with English Abstr.). 4. Liu Zhide, Huang Liming, Yang Zhongxi, et al. "Material Corrosion Factors of Ground Gathering Line in High Sulfurous Environment." Natural Gas Industry, 2004, 24(12):122-123(in Chinese with English Abstr.). 5. NACE TM0177-2005. "Laboratory Testing of Metals for Resistance to Specific Forms of Environmental Cracking." NACE International, Houston, TX, 2005. 6. IS015156/NACE MR0175. "Petroleum and Natural Gas Industries-Materials for Use in H2S Containing Environments in Oil and Gas Production[S]." Switzerland: The International Organization for Standardization, 2009. 7. Fitness-For-Services. API 579-1 /ASME FFS-1, JUNE 5, 2007(SECOND EDITION). 8. Sumitomo Metal, http://www.sumitomo-tubulars.com/materials/index. htm. 9. "Petroleum and natural gas industries-Steel pipes for use as casing or tubing for wells." ISO/FIDS 11960 (E):2009. 10. Sun, Y.X., Lin, Y.H. 2010. "Study on Casing and Tubing Design for Sour Oil & Gas Field." Corrosion Science and protection technology, (in Chinese, with English Abtsr.). 11. Q/SYCQZ 001-2008 Drilling technical operation.2008. 12. SY/T5322-2000. Design method of casing string strength[S]. Beijing: China Machine Press, 2001.
This page intentionally left blank
25
Coupled Mathematical Model of Gas Migration in Cemented Annulus with Mud Column in Acid Gas Well Hongjun Zhu, Yuanhua Lin, Yongxing Sun, Dezhi Zeng, Zhi Zhang, and Taihe Shi State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation Southwest Petroleum University, Chengdu, People's Republic of China.
Abstract Sustained casing pressure (SCP) in an acid gas well brings serious threat to worker safety and environmental protection. In this paper, the mathematical model of gas flow through porous media in cement sheath was obtained based on the generalized Darcy percolation theory. And the establishment of buoyancy model for gas migration in mud columns was based on multi-phase fluid dynamics theory. On this basis, the coupled mathematical model of gas migration in cemented annulus with mud column has been improved. Then the value of SCP changing with time in an acid gas well in a field of China has been calculated by this model. Calculation results coincided well with the actual field data, which provide some reference for the following security evaluation and solution measures of SCP.
25.1
Introduction
Well cement problems such as small cracks or channels can result in gas migration and lead to sustained casing pressure (SCP) at casing heads. Several other reasons for casing pressure buildup are tubular corrosion or mechanical failures, packer failures, and connection leaks. In some cases, the pressure can reach dangerously Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (449-462) © Scrivener Publishing LLC
449
450
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
high values. Moreover, SCP in acid gas wells brings serious threat to worker safety and environmental protection. SCP is the universal existing problem in gas wells in China. After the 12 May 2008 Wenchuan earthquake, the possibility of forming SCP in gas wells in northeast of Sichuan is much greater than previous. While acid gas reservoirs widely locate in Sichuan-Chongqing region in China, which contain high content of acidic components such as carbon dioxide and hydrogen sulfide. The material of current intermediate casing and surface casing in gas well is usually the common carbon steel. Acid gas migrating into annulus of protection casing and production casing or annulus of protection casing and surface casing may cause corrosion on the strings and endanger the safety of wellbore. The gas escaping from high-sulfur gas well will specially give rise to significant personnel casualty and property damage. Therefore, we are in urgent need of an explicit reason for acid gas migration and a model to calculate sustained casing pressure. The work presented here focuses on the coupled mathematical model of gas migration in cemented annulus with mud column in acid gas well, which provide some reference for the following security evaluation and solution measures of SCP.
25.2 Coupled Mathematical Model As shown in Figure 1, two possible configurations of the cement column in the annulus are commonfl]: cemented to the surface or a mud column above the cement. And gas column or gas-liquid column may present in the cement if the cementing job is poor. In wells cemented to the surface, gas migration can be considered as a one dimensional flow through a medium have some conductivity [2]. After bleed-down, the buildup behavior is controlled by cement properties, such as permeability and porosity, and by gas formation pressure. While in wells with a mud column above the cement, gas migration occurs in two stages. Firstly, gas flow follows Darcy's Law in the cement column. Then gas rises as bubbles through stagnant non-Newtonian drilling fluids, in which the gas migration is affected by the characteristics of mud and the status of the top gas cap. We focus our attention on the coupled mathematical model of gas migration in cemented annulus with mud column, which includes gas migration in cement column. Considering acid gas migration in cement and stagnant mud, the coupled model was obtained by improving Xu's SCP model[3,4].
COUPLED MATHEMATICAL MODEL
451
Figure 1. Configurations of the cement column in the annulus (a) Annulus cemented to the surface (b) Annulus with a mud column above cement.
25.2.1
Gas Migration in Cement
Gas migration in cement column can be considered as a one-dimensional flow through a medium having some conductivity, which related the cement properties, interface pressure, interface flow rate, gas formation pressure and elapsed time. The following assumptions were made for establishing the mathematical model[5]. Firstly, the gas formation pressure is constant because permeability of gas zone is much higher than that of cement. Secondly, the pseudo gas pressure concept is used. Finally, gas is vented out from the well at a small constant rate at the end of bleed-down. Then with a constant flow rate qn during the n-th period, the pressure in cement can be obtained as:
^
+
dz2
i i ^ T 4 ^
p{dz)
if
p< 15 MPa
and
v
Çdt
z
•
Í l = If^ dz2 Ç dt
if
p>l5MPa
and
z
(1)
452
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
where: = —
£
Z~
(2)
The boundary and initial conditions may be stated as follows: a. p = pc at t = 0 for z (0 < z < Lc) b. p = p, at z = 0 for all f 3D
^ z=r
forf>0 ^
25.2.2 Gas Migration in Stagnant Mud Gas migration in stagnant mud can be modeled as dispersed twophase flow that can be described by a drift-flux model[6]. Basic assumptions are concentric annulus, equal phase pressure, uniform phase densities normal to the flow direction, constant temperature profile and thermodynamic equilibrium. Due to slow phase segregation after the bleed-down, it is assumed that the relative velocity term is negligible. Under the assumptions, the one-dimensional two-equation drift-flux model is summarized in the foliowing[3]: Continuity equation for dispersed phase (gas) is: d(apg)
^(ccpgvg)
dt
=Q
(3)
dz
And mixture momentum equation is:
-^+pmg+jj-pym=o
(4)
where: V
v
g=
ü
s+C0vm
(5)
m=(^+<7L)M = ^sg+^L
(6)
dh=d0Pm=Pxa
di
+ PLtt-a)
(7) (8)
COUPLED MATHEMATICAL MODEL
453
In order to complete the model, slip velocity (v) and fluid fraction factor (/) must be algebraically specified. Based on the analysis of the SCP field data, it can be safely assumed that two-phase flows happened in SCP problem are bubble and slug flow patterns. According to Hasan and Kabir method[7], the value of the distribution factor C„ for bubble flow can be described as "0 '
Cn
[1.2
if
dh < 0.12m
or
2.0
if
dh > 0.12m
and
Vr, > 0.02m/ s Vc, < 0 . 0 2 m / s
(9)
For slug flow, C0 is equal to 1.2. For bubble flow, slip velocity can be calculated as[8]
vc = 0.06
g(pL-pJ(j
(I-a)2
(10)
p\ And for slug flow in vertical annulus, it is described as[9] v = 0.03 + 0 . 0 0 8 8 ^
l ^ - P M
(11)
Using apparent Newtonian concept, equations for flow in Newtonian fluid can be used to calculate Fanning friction factor. For laminar flow, friction factor is / =
16
(12)
Re.
where, Re„
PmVmdh
(13)
The mixture viscosity for two-phase flow is: (14) where: ßa=K
An 3n + l
(15) du
454
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
4=-^—
(16)
For turbulent flow, friction factor is J
Re„"m°'25
Analytical solution is generally not possible to be obtained for most practical problems in two-phase flow. Numerical methods based on finite-difference concepts provide an alternate and powerful solution approach. The two equations, (3) and (4), have been solved using finite difference method with computational cells shown in Figure 2, the semi-implicit manner of which isas follows [4].
Mapg)^[(ap¡rvr]_o At Az +(pmgr+(+-pymr=o Az 2du A
m
(i9)
25.2.3 Gas Unloading and Accumulation at Wellhead At wellhead, gas is released from the top when needle valve is open. For SCP build-up, gas accumulates at the top with closed needle valve. Thus, two different upper boundary conditions are considered. Gas or gas-liquid flowing through needle valve to the atmosphere can be considered as single-phase gas or multiphase flow through a choke. During bleed-down, gas usually flows at sonic velocity and its flow rate is easy to record. If gas flow rate cannot be
Figure 2. Computational cells for finite difference solution.
COUPLED MATHEMATICAL MODEL 455
measured directly, it can be computed from the following iterative procedure [3,4]. 1. First, assumed that the critical pressure ratio y* is 0.5. Then the pressure ratio (y=p2/p1) was calculated, where, p2 is downstream pressure (atmospheric pressure), and p1 is upstream pressure. 2. If y is less than y*, the critical flow exists. The gas flow rate could be calculated using following equation[10]. (
a=
rfi.89
V- 83
s
2.79x10 x
„
(20)
0.454
where, dch is the diameter of choke, and qLm is the liquid flow rate which can be calculated by the volume of liquid collected and bleed-down time recorded in SCP diagnostic tests. If not, then sub-critical flow exists. The gas flow rate could be calculated using following equation[ll].
^öxlO-x^p^-^
(21)
where, pml is the mixture density at pr 3. Then gas/liquid ratio R1 and yc were obtained from following equation. 7R
7+1
±— yj +2R1(Y+l)yc {7-1)
7+1
+ ry/ -2Rt-2R
2 7
- ¿ - =0 7-1 (22)
where, yis ratio of specific heats. Then set y*= yc. 4. Steps (2) to (3) are repeated until there is no change in the flow condition. The gas chamber in the wellhead is treated separately from all other cells, which is completely filled with gas. When needle valve is closed, it does not lose gas to any other cell but receives gas from the cell immediately below it. The volume of this gas chamber
456
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
changes with time, at n+1 time step, which is related to the n-th step volume, in the following manner[3,4],
v::1=v^+X(vi)/"+X(vt); -x^)^ 1 -twr j=i
j=\
j=\
j=i
-( 9g ) 0 " +1 Aí + (9am)"N+1Aí
(23)
where: N is the number of equal-size cells in well. The two summation terms for Vg are volumes of gas accumulated in the remaining mud column. The two summation terms for VL are the volume increase caused by increased liquid column pressure during this time step. The second to last term is the volume reduction due to gas flow from cement below the liquid column. The last term indicates the increased volume caused by gas flow out from the annulus during the bleed-down, which is zero during casing pressure buildup. The wellhead pressure at n+1 time step is[3,4] n+l-j/n vy" + l n nn+iyn n
V
wh
(24)
Z
Once the volume of gas chamber is estimated by Eq. 23, the wellhead pressure can be calculated from Eq. 24. This step, in turn, allows calculation of interface pressure from the wellhead pressure by Eq. 19.
25.2.4
Coupled Gas Flows in Cement and Mud
The most important problem for coupling gas flows in cement and mud is calculating the interface pressure. The calculation procedure is iterative, which involves a simultaneous solution of the mass and momentum equation for pressure and velocity in all cells in mud column except for gas chamber. The procedure is as follows[3,4]. 1. The gas flow rate at the interface (q ) was defined and the interface pressure (p*) was assumed. 2. For the first step (i=l), the average pressure (p.) and the gas and liquid densities (p , pL) were calculated with an assumed pressure (pi+1/2*)3. The void fraction (a) and gas velocity (w ) were obtained from the equations (18) and (19).
COUPLED MATHEMATICAL MODEL
457
4. Then the mixture density (pm) and pi+1/2 were calculated by the equations (8) and (19) respectively. 5. If the absolute value of difference of pMj2 and pi+1/2* is less than 10 3 , proceed to the next step (i=i+l). If not, a new pi+1/2* is estimated as the average of pi+1/2 and previous pi+1/2*- Then return to step (2). 6. Steps (2) to (6) are repeated until calculations at all cells below gas chamber converge. 7. Then the volume of gas chamber (V^) and wellhead pressure (pwh) were solved by the equations (23) and (24). 8. The interface pressure (pc) was calculated by the equation 19. 9. If the absolute value of difference of pc and p* is less than 10~3, proceed to the next time step. If not, a new p* is estimated as the average of pc and previous p*. Then return to step (2). 10. Steps (2) to (9) are repeated until the calculations converge. The coupling procedure begins with the pressure bleed-down because SCP diagnostic test begins with it. At initial time, wellhead pressure, size of gas chamber and the gas concentration in mud column are known. Constant pressure gradient throughout the entire mud column and zero flow rate at the interface between cement and mud were assumed. Then the pressure and gas distribution in mud as well as pressure at interface can be calculated. Guessing the cement filled with gas, the pressure is uniform and equal to interface pressure in entire cement, except at the point of gas-source formation. Gas flow rate at each recorded time interval was determined from bleed-down pressure history. With initial condition and two boundary conditions (known flow rate at wellhead and constant pressure at formation), pressure distribution in mud can be calculated by the iterative procedure described above. The pressure distribution in cement can be calculated by the Eq. 1. Iteration stops until the flow rate and pressure at either side of interface are equal. When a good match for bleed-down is obtained, initial condition at the instant of shun-in is known. The buildup just right begins with the variable distribution at the end of bleed-down. Assuming zero flow rate at the interface, the pressure distribution in mud and
458
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
cement can be calculated respectively. If there is a pressure difference across the interface, a force will drive gas flowing through the interface. Therefore, the pressure distribution should be recalculated using a guessed flow rate. The computation is repeated until pressure at each side of interface is equal. Time is then incremented, boundary conditions defined again, and the process is repeated until the desired time is reached.
25.3
Illustration
The mathematic model was verified with selected data from one acid gas well in a field of China. The geological conditions, wellbore structure and cementing data of this well are shown in Figure 3. And information about gas, mud, cement and formation are listed in Table 1. According to the coupled model, the casing pressure of B annulus in the illustration well has been calculated. The annulus between 139.7 mm production casing and 244.5 mm intermediate casing has been bled for 12 minutes before the needle valve is closed and followed casing buildup lasts 24 hours. Shown from Figure 4, the calculation result coincides well with the actual data, which presents long breed-down and normal buildup pattern.
Figure 3. Structural parameters of the well.
COUPLED MATHEMATICAL MODEL
459
Table 3. Information of the illustration well. Category
gas
mud
Property
Value
viscosity
10"5Pa-s
compressibility
1.112xlO-7Pa-1
specific gravity
0.7
apparent viscosity
0.056Pa-s
interface tension
0.068N/m
density
2270kg/m 3
permeability
1.8xlO u m 2
porosity
0.01
pressure
46MPa
cement formation
Figure 4. Calculation result of casing pressure in illustration well.
25.4
Conclusions
Coupled mathematical model of gas migration in cemented annulus with mud column has been improved based on annular percolation and gas-liquid two-phase flow theories, which was verified
460
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
with selected data from one actual gas well. This calculation method may provide some reference for the following security evaluation and solution measures of SCP.
25.5
Nomenclature
A = wellbore area, m 2 C0 = distribution factor cg = gas compressibility, Pa"1 dch = diameter of choke, m dh = hydraulic diameter of annulus, m d. = diameter of inner string, m dg = diameter of outer string, m / = friction factor g - acceleration of gravity, m / s 2 k = cement permeability, m 2 K = consistency coefficient Lc - length of cement column, m n = liquidity factor n' = moles pc = pressure on the top pf the cement, MPa p, = reservoir pressure, MPa pwh - wellhead pressure, MPa p1 = choke upstream pressure p2 = choke downstream pressure (atmospheric pressure) q = gas flow rate qL = liquid flow rate q m = gas flow rate from the choke qLm = liquid flow rate from the choke R7 = gas/liquid ratio Rem = Reynold number of two-phase flow t = time, s v = gas velocity, m / s vm = mixture velocity, m / s vs = gas slip velocity, m / s vs = superficial velocity of gas, m / s vsL = superficial velocity of liquid, m / s V = volume of gas, m 3 VL - volume of liquid, m 3 Vwh = volume of gas chamber, m 3
COUPLED MATHEMATICAL MODEL
461
yc = critical pressure ratio z = distance from the gas-source formation, m Z = gas-law deviation factor a = void fraction y- ratio of specific heats AL = no slip liquid holdup Ha = apparent viscosity for power-law mud, Pa-s ¡x = gas viscosity, Pa-s fim = mixture viscosity, Pa-s p = gas density, k g / m 3 pL = liquid density, k g / m 3 pm = mixture density, k g / m 3 pml = mixture density at p2, k g / m 3 (7= surface tension of liquid, N / m 0 = cement porosity
25.6 Acknowledgment Research work was co-financed by the China National Natural Science foundation and Shanghai Baosteel Group Corporation (grant No.: 51074135), and supported by program for New Century Excellent Talents in University (NCET-08-0907). Without their support, this work would not have been possible.
References 1. R. Xu, A.K. Wojtanowicz, "Diagnostic Testing of Wells with Sustained Casing Pressure-An Analytical Approach/' SPE paper 2003-221 presented at the 2003 Petroleum Society's Canadian International Petroleum Conference, Calgary, June 10-12,2003. 2. S. Nishikawa, "Mechanism of Gas Migration after Cement Placement and Control of Sustained Casing Pressure," Dissertation submitted to the graduate faculty of the Louisiana State University for the degree of Master, May, 1999. 3. R. Xu, "Analysis of Diagnostic Testing of Sustained Casing Pressure in Wells," Dissertation submitted to the graduate faculty of the Louisiana State University for the degree of Doctor, December, 2002. 4. R. Xu, A.K. Wojtanowicz, "Diagnosis of Sustained Casing Pressure from Bleedoff/Buildup Testing Patterns," SPE paper 67194 presented at the 2001 SPE Production and Operations Symposium, Oklahoma, March 24-27, 2001. 5. A.K. Wojtanowicz, S. Nishikawa, and R. Xu, "Diagnosis and Remediation of Sustained Casing Pressure in Wells," Final report submitted to MMS, July 31, 2001.
462
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
6. N. Zuber, J.A. Finadlay, "Average Volumetric Concentration in Two-Phase Systems," Trans, ASME,}. of Heat Transfer, Vol.87, p.453-468,1965. 7. A.R. Hasan, C.S. Kabir, "A Mechanistic Approach to Understanding Wellbore Phase Redistribution," SPE paper 26483 presented at the 68th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers, Houston, October 3-6,1993. 8. E.F. Caetano, O. Shoham, and J.P. Brill, "Upward Vertical Two-Phase Flow through an Annulus, Part 1 : Single-Phase Friction Factor, Taylor Bubble Rise Velocity and Flow Pattern Prediction," /. Energy Res. Tech., Vol.114, p.l, 1992. 9. A.R. Hasan, C.S. Kabir, "Two-Phase Flow in Vertical and Inclined Annuli," Int. J. Multiphase Flow, Vol.18, p.279,1992. 10. W.E. Gilbert, "Flowing and Gas-Lift Well Performance," Drill. & Prod. Prac, Vol.9, p.126,1954. 11. R. Sachdeva, Z. Schmidt, J.P. Brill, and R.M. Biais, "Two-Phase Flow through Chokes," SPE paper 15657 presented at the 61 th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers, New Orleans, October 5-8,1986.
SECTION 7 CORROSION
This page intentionally left blank
26 Study on Corrosion Resistance of L245/825 Lined Steel Pipe Welding Gap in H 2 S+C0 2 Environment Dezhi Zeng1-3, Yuanhua Lin3, Liming Huang1, Daijiang Zhu3, Tan Gu1, Taihe Shi1, and Yongxing Sun 2 1
State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation Chengdu, People's Republic of China 2 CNPC Key Laboratory for Tubular GoodsChengdu, People's Republic of China 3 Southwest Oil and Gas Field Company Chengdu, People's Republic of China
Abstract Gas fields in Luojiazai, Sichuan contain about 17% of hydrogen sulfide, and 10% of carbon dioxide. The chloride ion concentration is about 20 mg/1 in the gas field water. The pressure at the surface is about 9 MPa. The surface lines face severe corrosion challenges; therefore lined steel pipe is a reliable and cost-effective anticorrosion measure for this application. However, lined steel pipe welding involves dissimilar steel welding, which is difficult as the anticorrosion alloy will be diluted by the external carbon steel in this process, and anticorrosion performance will be affected. So it is necessary to make evaluations for welding technology of lined steel pipe. In this paper, taking welding gap of L245/825 lined steel pipe as an example, the anti-SSC performances of straight and ring welding gaps of L245/825 lined steel pipe in NACE A solution are studied, and stress corrosion cracking and electrochemical corrosion behaviors of them are evaluated by laboratory experiments. Then the experiments are rerun in the corrosive environment found in Tiandong 5-1 high sour gas well. Laboratory investigations and field observations show that straight and ring welding gaps of L245/825 lined steel pipe have a good antienvironment and anti-cracking performance of electrochemical corrosion in H 2 S+C0 2 environment. The welding technology selected in the paper is Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (465-478) © Scrivener Publishing LLC
465
466
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
reliable. Research results provide reference for surface lines selection and welding operation.
26.1
Introduction
The metal materials are facing the problems of anti-environment cracking and electrochemical corrosion in H 2 S+C0 2 environment. Gas fields in Sichuan are rich in H2S and C0 2 . Chloride ion content in the gas field water is also very high. Exploitation of these high acid gas fields is facing severe corrosion challenges [1-3]. In particular, the unexpected metal sulfide stress cracking will not only result in huge economic losses, but its toxicity will also threat personal safety, therefore requiring particular attention. In order to ensure safe and efficient development, it is very reliable to use nickel-base alloy in high acid gas fields, but it is costly. In this case, the lined steel pipe is an economical selection for surface transportation lines [4-7]. However, lined steel pipe welding involves dissimilar steel welding, and anticorrosion performances will be affected if the welding process is unreasonable. In the paper, taking welding gap of L245/825 lined steel pipe as example, the electrochemical corrosion and anti-environment cracking performances of L245/825 lined steel pipe welding gaps are studied in H 2 S+C0 2 environments to provide reference for welding operation of L245/825 lined steel pipe in high acid gas fields.
26.2
Welding Process of Lined Steel Pipe
The 825 thin wall inner tube for experiment is welded by laser without filler before forming of L245/825 lined steel pipe. The specification is: 9 114.3x (6+2) mm. The L245/825 lined steel pipe is welded by TIG method, the welding process is: sealing welding —>root welding —»filler welding —»cover welding. The solder is TGS-61 produced by Tian Tai, whose material is nickel-base alloy 625. The temperature between welding layers varies from room temperature to 150°C; Welding parameters are shown in Table 1. The quality of root welding gap is correlated with the anticorrosion performances of lined steel pipe. The elemental composition of root welding gap, 825 inner tube and solder is shown in Table 2.
CORROSION RESISTANCE OF
L245/825 LINED STEEL PIPE
467
Table 1. Welding parameters of lined steel pipe. No.
Weld Bead
Solder Brand
Heat Input
(KJ/mm)
Welding Velocity (cm/min)
1
Seal welding
TGS-61
0.47-0.53
9.0
2
Root welding
TGS-61
0.97-1.03
5.5
3
Filler welding
TGS-61
0.77-1.57
4.0-7.0
4
Cover welding
TGS-61
1.31-1.46
4.3
It clear form Table 2 that the alloy element composition of root welding gap and 825 inner tube is almost the same due to the use of 625 solder that has high alloy element and optimized welding parameters.
26.3 Corrosion Test Method of Straight and Ring Welding Gaps of L245/825 Lined Steel Pipe We should first pay attention to the anti-environment cracking performances of metal materials used in gas fields rich in H2S and C 0 2 to avoid unexpected metal sulfide stress cracking, then to the problem of electrochemical corrosion in H 2 S-containing environments [8, 9]. The outer pipe supports mechanically for the lined steel pipe, and the inner pipe is anti-corrosion layer. Corrosion performances of lined steel pipe depend on that of anti-corrosion inner layer and root welds. So the corrosion performances of straight and ring welding gaps of L245/825 lined steel pipe after forming are studied in the paper. In view of the elements argued above, corrosion performance measurements are performed for straight and ring welding gaps of L245/825 lined steel pipe. On the basis of results issued from many laboratory investigations, it is more reliable to use NACE TM0177 CR method to evaluate anti-corrosion performances of straight and ring welding gaps of L245/825 lined steel pipe. All anti-environment cracking performances of L245/825 lined steel pipe in H 2 S-containing environment are evaluated by CR method.
0.030
0.054
Root weld
<0.05
C
Solder
825 inner tube
Element
19.04
39.18
61.5
0.001 <0.003
0.24
0.211
-
P
0.02
<1.0
Mn
0.09
<0.5
39.17
20.44
22.4
Si
Ni
Cr
Table 2. Elementa composition of cladding, solder and root welding gap-
-
0.017
0.002 <0.002
1.5-3.0
Cu
<0.03
S
8.51
9.13
2.5-3.5
Mo
-
0.0321
-
V
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
CORROSION RESISTANCE OF
L245/825 LINED STEEL PIPE
469
The key point is to obtain the tensile stress of specimen's working face when measuring the anti-environment cracking performances of metal materials using CR method. There are two methods recommended in NACE TM0177, namely to calculate the loading stress from empirical formula by measuring loading deflection of specimen or measure the loading stress by resistance-strain test method directly. Due to the strict requirements for operators and the high measuring error when using deflection test method, the resistance-strain test method is adopted in the paper to measure the loading stress of welding gap of lined steel pipe. As there are differences between microstructure and elemental composition of welding gap and 825 tube body, the mechanical properties of them are also different. Furthermore, the yield strength of material may differ as the plastic strengthening produced in the forming process of straight welding gap is different. According to IS015156, the material should be loaded at a 1OO%0 stress level when evaluates stress corrosion of anticors
rosion alloy. In the paper, we will apply loads and then unload to see whether there is residual strain, in the case that there is small residual strain, the specimen's working face is just loaded at 100%o stress level. s
The loading and testing process is shown in Figure 1. The specimen is loaded with feed screw nut. A Teflon gasket is used between loading clamps for insulation. Testing accessories are removed from specimen after loading. Cleaning and drying the specimens for later use. Prepared specimens are shown in Tables 3 and 4. From Tables 3 and 4, it can be seen that yield strength of 825 cladding straight welding gap of L245/825 lined steel pipe is approximately 197 MPa, yield strength of root welding zone of ring welding gap is 208 MPa. Loading stresses of Z7-Z12, H7, H9, H10, H l l specimens are bigger than normal.
Figure 1. Stress loading and testing process for C-ring specimen.
14 14
2.02
2.02
100.72
100.78
Zll
Z12
7
7
7
14
1.96
100.9
Z10
7
14
1.96
101.02
Z9
7
14
2.00
100.96
Z8
7
14
1.96
100.78
Z7
7
14
1.98
101.18
Z6
7
14
2.00
101.44
Z5
7
14
2.00
101.08
Z4
7
14
1.96
100.88
Z3
7
14
1.98
101.28
Z2
7
14
1.96
101.36
Zl
Pore Diameter (mm)
Width (mm)
Outer Diameter (mm)
No.
Wall Thickness (mm)
Table 3. Straight welding gap specimens.
1341
1420
1363
1407
1299
1265
994
998
1018
1010
1028
1298
Strain (106)
261.46
276.90
265.79
274.37
253.31
246.68
193.83
194.61
198.51
196.95
200.47
253.11
Loading Stress (MPa)
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Outer Diameter (mm)
100.30
101.98
97.40
101.48
101.38
100.70
101.86
96.28
101.36
101.30
101.40
107.96
No.
HI
H2
H3
H4
H5
H6
H7
H8
H9
H10
mi
H12
0.9
1.68
1.24
1.58
7 7
14
7
14
14
7
7
14 14
7
14
7
7
14
14
7
14
7
7
14
14
7
Pore Diameter (mm)
14
Width (mm)
1010
1574
1500
1675
994
1538
1048
1119
1051
1059
1096
1052
Strain (10"6)
195.39
303.81
292.50
326.63
193.83
308.69
204.36
218.21
204.95
206.51
213.72
205.14
Loading Stress (MPa)
L245/825
0.98
1.54
1.54
1.68
1.40
1.18
1.74
1.88
Wall Thickness (mm)
Table 4. Ring welding gap specimens. CORROSION RESISTANCE OF LINED STEEL PIP
472
26.4
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Corrosion Test Results of Straight and Ring Welding Gaps of 1245/825 Lined Steel Pipe
26.4.1 Atmospheric Corrosion Test Results The corrosion test process follows the criterions of NACE TM0177 and IS015156. All the specimens are loaded at a 100%o"s stress level. NACE A solution is used in the test, with test temperature of 24±3°C, and test period of 720 hours. Taking out the specimens and clean them after the experiment. Firstly, observing the macrostructures of specimens to see whether there are ruptures in these samples, and then observe the microstructures of them after being magnified by 50 multiples using a stereo microscope to see whether there are cracks in the working faces. Atmospheric corrosion test results of straight and ring welding gaps of L245/825 lined steel pipe are shown in Table 5. There are no cracks in straight and ring welding gaps of L245/825 lined steel pipe after a 720 hours test period, revealing good antiSSC performances.
26.4.2 Corrosion Test Results at High Pressure According to the environment of surface gathering and transportation pipelines in high H2S gas fields in northeastern Sichuan, a simulated solution of X gas field is used in the experiment. The average concentration of chloride ion is 20 g/1, partial pressure of hydrogen sulfide at 1.5 MPa, partial pressure of carbon dioxide at 1.0 MPa, total test pressure at 9 MPa, test temperature of 70°C, and test period of 720 hours. Examining all the specimens to see whether there are cracks. Corrosion test results at high pressure are shown in Table 6. There are no cracks in straight and ring welding gaps of L245/825 lined steel pipe after a 720 hours test period, showing good antienvironment cracking performances. The specimens of straight welding gaps after experiment are bright, which indicates that all of them have excellent resistance to electrochemical corrosion in simulated operating conditions. One specimen of ring welding gaps has localized corrosion, the other two are bright, the reason for this phenomenon remains to be proved.
CORROSION RESISTANCE OF
L245/825
LINED STEEL PIPE
473
Table 5. Atmospheric corrosion test results. No.
Zl
Loading Stress
Solutions
Results
Note
100%o
NACE A simulation
720 h no fracture
Straight welds
NACE A simulation
720 h no fracture
Straight welds
NACE A simulation
720 h no fracture
Straight welds
NACE A simulation
720 h no fracture
Ring welds
NACE A simulation
720 h no fracture
Ring welds
NACE A simulation
720 h no fracture
Ring welds
s
Z2
100%a S
Z3
100%a S
HI
100%a s
H2
100%a s
H3
100%a s
Table 6. Corrosion test results at high pressure. Solutions
Results
Note
X gas field simulation water
720h, bright, no fracture
Straight welds
X gas field simulation water
720h, bright, no fracture
Straight welds
s
X gas field simulation water
720h, bright, no fracture
Straight welds
S
X gas field simulation water
720h, bright, no fracture
Straight welds
S
X gas field simulation water
720h, no fracture, local corrosion
Straight welds
s
X gas field simulation water
720h, bright, no fracture
Straight welds
s
No.
Loading Stress
Z4
100%CT S
Z5 Z6 H4 H5
H6
100%c 100%o 100%a 100%o
100%o
474
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
26.4.3
Field Corrosion Test Results
To verify corrosion resistance of L245/825 lined steel pipe in field conditions and its difference from L360NCS carbon steel, eight rectangular coupon specimens from L360NCS pipeline steel are used for corrosion experiment. The specimens of straight and ring welding gaps are positioned on two brackets as shown in figure 2, in which from the left hand to the right hand are Z7, Z8, Z9, H7, H8, H9, Z10, Z l l , Z12, H10, H l l , H12 specimens respectively. All specimens are divided into two groups and placed in the horizontal tank of on-line corrosion installation in Tiandong 5-1 high sour gas well for corrosion evaluation. The test period is 720 hours; Average concentration of chloride ion is 22289 mg/1 in solution with the average temperature of 33.23°C, total pressure of 7.43 MPa, partial pressure of hydrogen sulfide at 0.53 MPa and partial pressure of carbon dioxide at 0.19 MPa. X-ray diffraction analysis is performed for sediments in horizontal tank after experiment, inspection indicates that little amount of elemental sulfur exists in the corrosive materials. Field corrosion test results of straight and ring welding gaps of L245/825 lined steel pipe are shown in Tables 7 and 8. From Tables 7 and 8, we can see that carbon steel corrosion is serious. The corrosion rate in the liquid phase is higher than that in the gas phase. The highest corrosion rate in the liquid phase is up to 0.75 m m / a . All specimens of straight and ring welding gaps that placed on the upper and lower layers of horizontal tank are bright and have no cracks during a 720 hours test period. They show good anti-environment cracking performances and excellent resistance to electrochemical corrosion in field operating conditions in Tiandong 5-1 well.
Figure 2. Specimens of composite pipe welding gaps for field corrosion test.
-
42.6186 39.1271 43.2328 32.4637 17.5421 33.5229
42.6196 39.1287 43.2348 32.4641 17.5427 33.5239
L245/825 lined steel pipe straight welds
L245/825 lined steel pipe straight welds
L245/825 lined steel pipe straight welds
L245/825 lined steel pipe ring welds heat affected zone
L245/825 lined steel pipe ring welds heat affected zone
L245/825 lined steel pipe ring welds
Z7
Z8
Z9
H7
H8
H9
-
Bright No surface crack
Bright No surface crack
Bright No surface crack
Bright No surface crack
Bright No surface crack
Bright No surface crack
Uniform corrosion + local corrosion
Uniform corrosion + local corrosion
Uniform corrosion + local corrosion
Uniform corrosion + local corrosion
Surface Conditions
L245/825
-
0.45
L360NCS
L4 9.5755
0.33
9.6662
9.9402
L360NCS
L3 9.9456
0.28
9.7432
9.9773
L360NCS
0.32
9.6473
9.9115
L2
Corrosion Rate (mm/a)
Weight After Test (g)
Weight Before Test(g)
L360NCS
Materials
LI
No.
Table 7. Materials corrosion test results, in the upper horizontal tank of on-line corrosion installation in Tian Dong 5-1 well. CORROSION RESISTANCE OF LINED STEEL PIPE
-
40.9843 42.1028 33.4043 32.1311 21.4154
40.9851 42.1036 33.4046 32.1314
L245/825 lined steel pipe straight welds
L245/825 lined steel pipe straight welds
L245/825 lined steel pipe ring welds heat affected zone
L245/825 lined steel pipe ring welds heat affected zone
L245/825 lined steel pipe ring welds
Zll
Z12
H10
Hll
H12
21.4175
-
42.4369
42.4378
L245/825 lined steel pipe straight welds
Z10
-
Bright No surface crack
Bright No surface crack
Bright No surface crack
Bright No surface crack
Bright No surface crack
-
Bright No surface crack
Uniform corrosion + local corrosion
-
0.69
9.3690
9.9444
L360NCS
Uniform corrosion + local corrosion
0.69
L8
9.5216
10.0950
L360NCS
Uniform corrosion + local corrosion
0.59
L7
9.5309
10.0237
Uniform corrosion + local corrosion
0.75
L360NCS
9.3643
9.9860
Surface Conditions
Corrosion Rate (mm/a)
L6
Weight After Test (g)
Weight Before Test (g)
L360NCS
Materials
L5
No.
Table 8. Materials corrosion test results in the lower horizontal tank of on-line corrosion installation in Tian Dong 5-1 well.
n
4^
M
a
O
zo
X
n
S?
% a O
a r->
o z > z o
w
M H
d
en O
NO
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
CORROSION RESISTANCE OF L245/825 LINED STEEL PIPE
26.5
477
Conclusions
Corrosion performance measurements are performed for straight and ring welding gaps of L245/825 lined steel pipe in NACE A solution, simulated operating conditions and field conditions in Tiandong 5-1 well respectively. Experimental results show that L245/825 lined steel pipe welded joints have good anti-environment cracking performances and excellent resistance to electrochemical corrosion. The technology of welding process selected in the paper is reliable and can provide reference for field welding operation.
26.6
Acknowledgments
The authors are grateful to the support from Program for New Century Excellent Talents in University (NCET-08-0907) and China National Natural Science foundation and Shanghai Baosteel Group Corporation (Grant No.: 51074135).
References 1. Li Luguang, Huang Liming, Gu Tan, Li Feng. "Corrosion Characters and Inhibition Methods for Sichuan Gas Fields." Chemical Engineering of Oil & Gas, 2007, 36(l):46-54. 2. Jiang Fang. "In Lab. Evaluation Methods of Metal Materials for High Sour Gas Fields." Natural Gas Industry, 2004,24(10):105-107. 3. Liu Zhide, Gu Tan, Tang Yongfan, et al. "Researches on the Electrochemical Corrosion of the Gathering Lines in Sour Gas Fields." Chemical Engineering of Oil & Gas, 2007,36(1)55-58. 4 . TOJIOBHH C . B . MeTajIJIMCKHe Tpy6onpOBOÄM C IIJiaKHpOBKOH H3 KOpp03HOHHO-
CToiÍKoro crmaBa. ra3.npoM-CTb, 1992,(9): 30-31. 5. Specification for CRA Clad or Lined Steel Pipe. API Specification 5LD Second Edition, 1998. 6. GOMEZ X, ECHEBERRIA J. "Microstructure and mechanical properties of Carbon steel A2102 superalloy Sanicro 28 bimetallic tubes." []]. Material Science and Engineering, 2003, 348(122):180-191. 7. Rommerskirchen I. "New progress caps 10 years of work with BuBi pipes." World Oil, 2005, 226(7): 69-70. 8. NACE TM 0177, "Laboratory testing of metals for resistance to sulfide stresses cracking and stress corrosion cracking in H2S environments." 9. IS015156, "Petroleum and natural gas industries—Materials for use in H 2 S-containing environments in oil and gas production."
This page intentionally left blank
Index Alkanolamine 126-128 Amine (see alkanolamine) Ammonia 56,59-60, 62 AQUAlibrium 133,138,141,144 Benzene 3-11,31,108,161 Bicarbonate (ion) 124-125,202, 323, 416 BTEX 3-11,108,110 Bubble point 7-9,11,13-21,117, 149,158 Butane 44, 69,141,146, 252,287 C3 (see propane) C4 (see butane) C5 (see pentane) Capital cost 160,247, 248, 310,312 Carbon capture xix-xxiii, 121-131, 155,157,171,204,394 Carbon sequestration 208,408 Carbonate (ion) 384,386,416 Carbonate (rock) 190,197, 377-392, 394,395,401, 402,404,408, 413-416,418,424 Casing 97, 98,103-104,183-186, 189,194-199,202,229,342-343, 424,427,437-447,449-150,456, 458-459 Cement 103,156,183-186,188-189, 194-195,198, 202,423-433, 449-461 Changyuan 329-349
Chemical reaction 124, 383-384, 393-394 Compressibility factor 65-84, 91 Compressor xxi, 3-4,11, 30, 89-105,119,133-152,155,157, 159-160,164-169,202 Corrosion 126-127,135-137, 185-186,326,430,438,441-444, 449-450,465-477 Crude oil 13,15-21,233,249,251, 255-256, 261,321-325,330-331, 339, 346,352 Dehydration xxi, 107-120,133-152, 165, 369 Densimeter 24-27,30-33,37 Density 13-14,16-17,19,23-37, 41-43,45^9, 52, 66,91-95, 97, 101,167,188,210,213-214,218, 228-229, 237,256,261,263, 330-331, 338, 339,400, 426,455, 457,459 Dew point 3-11,110,113-114,118, 136,138-140,147-151,158 Dolomite 193-194, 377-379, 382-383, 386-388, 391,398, 401-403,408-416 Dun vegan 247-317 Elemental sulfur 227-242, 385, 387,474 Enhanced oil recovery xix-xxiii, 14,156-157,162,168,247-317,
479
480
INDEX
319-327,329-350, 351-359, 362, 393-404,407, 409, 424 EOR (see enhanced oil recovery) EOS (see equation of state) Equation of state 3, 5-6, 32-36, 41-45, 49, 55-58, 66,108, 261 Ethane 41,43-44 Ethyl benzene 3-11,108
Nitrogen oxides xx, 377-391 North Dakota 393-404,407-418 Northwest McGregor 393^04 NOx (see nitrogen oxides)
Gas hydrates 136,145,147,149, 159,209-224, Gas well 181-183, 239-240,437-447 Geothermal gradient 89, 91,93, 94, 98, 99,213-215, 339
Pentane 28-29, 69,142,146,252 Permeability 90-91,182,187-188, 190,193,194,197-198,201, 210,216-217,227,230,232, 237-238, 240-241,255,268, 269-272,273, 276-277, 280-282, 287,290,315,317,319-326, 329-349,351-352, 359,391, 394-395,397,402,424, 450-451,459 Permian basin 136,175-176, 190-199,203,410, 424 Pipeline xx, xxii, 103-104,135,139, 155,157,159-160,164,176,194, 196, 472,474 Polysaccharides (see heteropolysaccharides) PR (see equation of state) Propane 44, 69,141,146, 252, 399 Pump xxi-xxii, 30, 67-68,123,137, 155-172 Pyrite 379, 382-384,386-387, 389-391,398,401, 409,411
Heteropolysaccharides 361-373 Horizontal well 247,291-292, 307-308,310,317,326 Huff 'n puff xxii, 393^04 Hydrates (see gas hydrates) Hydrostatic gradient 89-91, 95-96, 99,195 Injection pressure 140,155,157, 170,182,196,199,290,323, 332-333,340, 349,352-354 Interfacial tension 13-14,19-21, 227,232,307, 319, 321-322, 348, 361-363 Jilin 14,17-18,20-21,320-321 Limestone 104,193,199-201, 379, 397,400,408-409 Mass transfer 107,113,115-120, 227-228 MDEA 4, 6,8,10, (also see alkanolamine) MEA (see alkanolamine) MR0175 437-438 New Mexico 175-207, NH3 (see ammonia)
Oil well 341, 348, 351-359, 395 Operating cost 157,164,171, 248,312
Saccharides (see heteropolysaccharides) San Juan basin 175-177,192 199-204 Sandstone 104,194,199-200,248, 269, 334-336, 338,340, 343, 408,411 Siderite 383, 387, 389,391 Sn0hvit 168 Soave-Redlich-Kwong (see equation of state)
INDEX
Sour gas 4-5, 23-37,55, 58, 60-62, 66,134,176-177,407, 409,440, 465,474 SOx (see sulfur oxides) SRK (see equation of state) Sulfur (see elemental sulfur) Sulfur oxides xx, 377-392 Surface tension (see interfacial tension) Sweet gas 66,118-119,134,205 TEG (see triethylene glycol) Texas 136,175-176,190-199,203, 410,424 Toluene 3-11,108 Triethylene glycol 107-120,137
113,128,162,230,233, 249,251, 255-256,261-262, 307,319, 321-322,324,330-331,338-339, 346,352,453,459 Water content xxi, 51,95,108-111, 113-119,139-140,144-146, 148-149,320, 363 Water flood 344 Wellhead pressure (see injection pressure) Weyburn-Midale 168,394 Williston basin 394, 397, 400^01, 407-^18 Wyoming 206 Xylene 3-11,108
Viscometer 17,23-37, 67 Viscosity 13-14,17-19,21,23-37, 41-52, 66, 90-91, 93-94,104-105,
481
Z-factor (see compressibility factor)
This page intentionally left blank
Also of Interest
Check out these forthcoming related titles coming soon from Scrivener Publishing Now Available from the Same
Author:
Acid Gas Injection and Carbon Dioxide Sequestration, by John J. Carroll, ISBN 9780470625934. Provides a complete overview and guide on the hot topics of acid gas injection and C 0 2 sequestration. Carbon Dioxide Thermodynamic Properties Handbook, by Sara Anwar and John J. Carroll, ISBN 9781118012987. The most comprehensive collection of carbon dioxide (C0 2 ) data ever compiled. Advances in Natural Gas Engineering Series, Volumes 2 and 3: Carbon Dioxide Sequestration and Related Technologies, by Ying Wu and John J. Carroll, ISBN 9780470938768. Volume two focuses on one of the hottest topics in any field of engineering, carbon dioxide sequestration. PUBLISHING MAY 2011 Sour Gas and Related Technologies, by Ying Wu and John J. Carroll, ISBN 9780470948149. This third volume in the series focuses on sour gas, one of the most important issues facing chemical and process engineers. PUBLISHING MAY 2012 Other Books in Energy and Chemical
Engineering:
Emergency Response Management for Offshore Oil Spills, by Nicholas P. Cheremisinoff, PhD, and Anton Davletshin, September 2010, ISBN 9780470927120. The first book to examine the Deepwater Horizon disaster and offer processes for safety and environmental protection.
Advanced Petroleum Reservoir Simulation, by M.R. Islam, S.H. Mousavizadegan, Shabbir Mustafiz, and Jamal H. Abou-Kassem, ISBN 9780470625811. The state of the art in petroleum reservoir simulation. NOW AVAILABLE! Energy Storage: A New Approach, by Ralph Zito, ISBN 9780470625910. Exploring the potential of reversible concentrations cells, the author of this groundbreaking volume reveals new technologies to solve the global crisis of energy storage. NOW AVAILABLE! Ethics in Engineering, by James Speight and Russell Foote, December 2010, ISBN 9780470626023. Covers the most thought-provoking ethical questions in engineering. Zero-Waste Engineering, by Rafiqul Islam, February 2011, ISBN 9780470626047. In this controvercial new volume, the author explores the question of zero-waste engineering and how it can be done, efficiently and profitably. Fundamentals of LNG Plant Design, by Saeid Mokhatab, David Messersmith, Walter Sonne, and Kamal Shah, August 2011. The only book of its kind, detailing LNG plant design, as the world turns more and more to LNG for its energy needs. Flow Assurance, by Boyun Guo and Rafiqul Islam, September 2011, ISBN 9780470626085. Comprehensive and state-of-the-art guide to flow assurance in the petroleum industry.
